CADTH Reimbursement Review

Axicabtagene Ciloleucel (Yescarta)

Sponsor: Gilead Sciences Canada Inc.

Therapeutic area: Large B-cell lymphoma

This multi-part report includes:

Clinical Review

Pharmacoeconomic Review

Ethics Review

Stakeholder Input

Clinical Review

Executive Summary

An overview of the submission details for the drug under review is provided in Table 1.

Table 1: Submitted for Review

ASCT = autologous stem cell transplant; CAR = chimeric antigen receptor; NOC = Notice of Compliance.

Introduction

Non-Hodgkin lymphoma (NHL) comprises a diverse group of closely related cancers of the lymphocytes.¹ It is the most prevalent hematological malignancy and the fifth most common cancer diagnosed in Canada.¹ In 2022 it was estimated that 11,400 people in Canada would be diagnosed with NHL and 3,000 would die from it.² NHL can affect any organ in the body with a wide range of clinical presentations.³ Most patients with NHL present with painless enlarged lymph nodes with or without systemic symptoms (e.g., fevers, night sweats, weight loss, and fatigue). Large B-cell lymphoma (LBCL) is the most common subtype of NHL diagnosed in Canada, constituting 30% to 40% of all NHL cases in Canada.⁴ LBCL is an aggressive but potentially curable NHL, and is typically diagnosed at an advanced stage (stage III or IV).⁵ The 5-year progression-free survival (PFS) in patients with limited disease is 80% to 85%, whereas those with advanced disease have a 5-year PFS of approximately 50%.⁶

The standard of care (SOC) first-line treatment for patients with newly diagnosed LBCL is CHOP — cyclophosphamide, hydroxydaunorubicin (doxorubicin), oncovin (vincristine), and prednisone — often used in combination with rituximab (R-CHOP).^4,7-10 However, 30% to 50% of patients are refractory to or relapse after first-line therapy.^4,11 Overall survival (OS) in patients with primary refractory disease is very poor, with only 15% to 20% surviving at 5 years.¹² Patients with partial response (PR) or complete response (CR) to first-line treatment also have poor survival at relapse, with 38% and 42% surviving at 5 years, respectively.¹² In this population, 82% progressed within 12 months of first-line chemoimmunotherapy, and these patients had significantly worse 3-year OS rate compared with patients who relapsed after 12 months (10% versus 39%, respectively).¹³

For patients who relapse or whose disease is refractory to first-line chemoimmunotherapy, second-line treatment comprises salvage chemotherapy, and if responsive to salvage therapy, this is followed by high-dose therapy (HDT) and autologous stem cell transplant (ASCT). However, only about half of the patients with relapsed or refractory LBCL are fit enough for transplant (i.e., have adequate organ function with no major comorbidities), and only half of transplant-eligible patients respond to salvage chemotherapy and can proceed to ASCT. Treatment options for patients with relapsed or refractory LBCL who are ineligible for ASCT, do not respond to salvage chemotherapy, or relapse post-ASCT include palliative chemotherapy, radiotherapy, clinical trials, or third-line chimeric antigen receptor (CAR) T-cell therapy if the patient meets the eligibility criteria.

Axicabtagene ciloleucel (axi-cel) is a CD19 antigen–directed genetically modified autologous T-cell immunotherapy (i.e., CAR T-cell therapy). The objective of this review was to evaluate the efficacy and safety of axi-cel (IV infusion, target dose of 2 × 10⁶ anti-CD19 CAR T cells/kg body weight) for the treatment of adult patients with relapsed or refractory LBCL. Of note, this CADTH reimbursement review was conducted before issuance of Health Canada Notice of Compliance (NOC) and the scope was based on the anticipated indication.

Stakeholder Perspectives

The information in this section is a summary of input provided by the patient groups who responded to CADTH’s call for patient input and from clinical expert(s) consulted by CADTH for the purpose of this review.

Patient Input

One patient group, Lymphoma Canada, submitted patient input for this review. This group gathered information from patients with diffuse large B-cell lymphoma (DLBCL) through 2 online surveys, 1 in 2018 and the other in 2022. Respondents from both surveys reported that fear of progression/relapse was the most common psychosocial impact (67%) affecting quality of life, followed by anxiety (37%), memory loss (37%), and concentration problems (36%). The majority of respondents (83%) were treated with CHOP with or without rituximab as their first line of treatment after diagnosis. The respondents reported that long-term treatment side effects (i.e., lasting longer than 2 years or appearing at least 2 years after the end of treatment) included fatigue (52%) and chemo brain (42%). Respondents from the first survey reported that their lymphoma treatment had the greatest negative impact on their work, travel, and other activities. Patients rated longer survival, longer remission, better quality of life, and fewer side effects as important outcomes expected from their treatment. About half of the patients stated that they would choose a treatment with potentially serious side effects if their doctor recommended it as the most effective option for DLBCL. Respondents also indicated that treatment choice was an important factor, and 91% of patients felt a need for more therapy options for patients with DLBCL.

Three patients reported receiving axi-cel as third-line therapy. All 3 patients were away from home for more than 3 months while receiving the treatment, pointing to the challenges in accessing CAR T-cell therapy. All 3 patients reported thrombocytopenia as a side effect of axi-cel treatment, with fever, anemia, nausea/vomiting, neutropenia, diarrhea, joint or muscle pain, and fatigue reported by 2 patients. Fear of progression/relapse and difficulty sleeping were reported by all 3 patients as psychosocial impacts related to their CAR T-cell therapy.

Clinician Input

Input From Clinical Experts Consulted by CADTH

Four clinical experts from across Canada contributed input to the CADTH review. The clinical experts consulted by CADTH noted the unmet treatment needs of patients who are refractory to or who relapsed after front-line therapy. While high-dose therapy and autologous stem cell transplant (HDT-ASCT) has curative potential for patients with relapsed or refractory LBCL, many are ineligible for ASCT or do not respond to salvage chemotherapy. The clinical experts indicated that axi-cel would fit well earlier in the treatment lines and could be used for most patients instead of ASCT. The clinical experts noted that patient outcomes are expected to improve if patients receive a potentially curative therapy earlier in the course of disease as some patients deteriorate rapidly and may be less likely to survive if definitive treatment is delayed.

The clinical experts noted that while the ZUMA-7 trial recruited only patients who were eligible for ASCT, in standard clinical practice there is no clinical rationale for restricting axi-cel only to those who are candidates for ASCT, and that any patient with adequate organ function and good performance status (Eastern Cooperative Oncology Group Performance Status [ECOG PS] score ≤ 2) who, based on the clinician’s judgment, can tolerate the known toxicities of CAR T-cell therapy (e.g., cytokine release syndrome [CRS]) would be suitable for axi-cel treatment. The clinical experts noted that axi-cel treatment can be provided by oncologists or hematologists in a hospital setting with adequate infrastructure for cell therapy and with access to highly specialized multidisciplinary clinical care including critical/intensive care and specialist care (e.g., neurology, nephrology) to manage toxicities, as well as laboratory support to handle and process samples. The clinical experts also pointed out that the 13-day median manufacturing time reported in the ZUMA-7 study is rapid and may not be reproducible outside the clinical trial setting; longer wait times may compromise patient outcomes.

Clinician Group Input

Clinician group input was received from 3 groups: Lymphoma Canada, Ontario Health (CCO) Hematology Cancer Drug Advisory Committee, and Cell Therapy Transplant Canada (CTTC). The clinician groups agreed that there are unmet needs in the current second-line treatment for patients with relapsed or refractory LBCL. The clinician groups indicated that there may be limited eligibility or tolerability to further salvage chemotherapy for some patients (e.g., patients with primary refractory disease or early relapse, and older patients). The clinician groups also noted that toxicities such as febrile neutropenia, bacteremia and other infections, gastrointestinal toxicity, mucositis, need for transfusion support, and secondary malignancies associated with ASCT treatment have made it unsuitable for patients who are high risk, who are refractory to treatment, or who relapse within 12 months of diagnosis.

Drug Program Input

Input was obtained from the drug programs that participate in the CADTH reimbursement review process. The following were identified as the key factors that could potentially impact the implementation of a CADTH recommendation for axi-cel:

• Considerations for initiation of therapy

• Considerations for prescribing of therapy

• Care provision issues

• System and economic issues

The clinical experts consulted by CADTH provided advice on the potential implementation issues raised by the drug programs.

Clinical Evidence

Pivotal Studies and Protocol-Selected Studies

Description of Studies

A single sponsor-submitted pivotal study was included in this review.^14,15 The ZUMA-7 study is a phase III, multicentre, randomized, open-label study evaluating the efficacy of axi-cel compared with SOC (salvage chemoimmunotherapy followed by HDT-ASCT) as a second-line therapy in patients with relapsed or refractory LBCL after first-line rituximab and anthracycline-based chemotherapy. The trial was conducted in 14 countries; 20 patients from 8 centres were recruited in Canada. The first patient was enrolled (randomized) on January 25, 2018, and enrolment was completed on October 4, 2019; ZUMA-7 is currently ongoing. All patients had either primary refractory disease or relapse within 12 months of completing first-line therapy, were potentially eligible for ASCT, and had not yet received second-line treatment. The data cut-off date for the primary analysis was March 18, 2021. For patients in the axi-cel arm (N = 180), treatment consisted of lymphodepleting chemotherapy followed by a single IV infusion of axi-cel. Bridging therapy consisting of corticosteroids was allowed before lymphodepleting chemotherapy for patients with high disease burden, at the discretion of the investigator. For patients in the SOC arm (N = 179), treatment consisted of a single protocol-defined, platinum-based salvage chemotherapy regimen for 2 to 3 cycles as selected by the treating investigator. Patients who responded to salvage chemotherapy were to proceed to HDT followed by ASCT. The mean age of patients was 57 years (standard deviation [SD] = 12 years); 30% of the patients were aged 65 years or older. Overall, 74% of the study population had primary refractory disease and 26% had early relapse. Approximately one-quarter of patients in both treatment arms had achieved a best response of CR to first-line treatment.

Efficacy Results

Key efficacy results from the ZUMA-7 study are summarized in Table 2.

Overall Survival

OS was a key secondary outcome in the ZUMA-7 study. The OS data remain immature at the time of this review. At the time of the data cut-off date (March 18, 2021), 72 patients (40%) in the axi-cel arm had died and 81 patients (45%) in the SOC arm had died. The primary OS analysis will occur at approximately 210 deaths or 5 years after the first patient was enrolled. At the interim OS analysis, the hazard ratio (HR) for death was 0.730 (95% confidence interval [CI], 0.530 to 1.007; 1-sided stratified log-rank P value = 0.0270).

Event-Free Survival

Event-free survival (EFS) as determined by blinded central assessment was the primary outcome of the ZUMA-7 study. At the time of the data cut-off, 252 EFS events as determined by blinded central assessment had occurred for 108 patients (60%) in the axi-cel arm and for 144 patients (80%) in the SOC arm. The median EFS was 8.3 months (95% CI, 4.5 to 15.8 months) in the axi-cel arm and 2.0 months (95% CI, 1.6 to 2.8 months) in the SOC arm. The stratified HR for event or death was 0.398 (95% CI, 0.308 to 0.514; P < 0.0001).

Health-Related Quality of Life

HRQoL was assessed by changes from screening in the global health status scale and the physical functioning domain of the European Organisation for Research and Treatment of Cancer Quality of Life Questionnaire Core 30 (EORTC QLQ-C30) and the 5-Level EQ-5D (EQ-5D-5L) index and visual analogue scale (VAS) scores were secondary outcomes. There was a clinically meaningful difference (based on the trial-specified threshold of ± 10 points)¹⁶ in mean change of scores from baseline to study day 100 for the EORTC QLQ-C30 global health status and physical functioning, and EQ-5D-5L and VAS scores with axi-cel compared with SOC. However, the high attrition rate, which was imbalanced between groups at all follow-up time points, limits interpretation of these data.

For the EORTC QLQ-C30 global health status there was a clinically meaningful difference in mean change of scores from screening at study day 100 for patients treated with axi-cel compared with SOC (estimated difference = 18.1 points; 95% CI, 12.3 to 23.9; adjusted P < 0.0001). At study day 150, the estimated difference was 9.8 points (95% CI, 2.6 to 17.0; adjusted P = 0.0124).

For the EORTC QLQ-C30 physical functioning subscale, there was a clinically meaningful difference in mean change of scores from screening to study day 100 for patients treated with axi-cel compared with SOC (estimated difference = 13.1; 95% CI, 8.0 to 18.2; adjusted P < 0.0001).

For the EQ-5D-5L VAS there was a clinically meaningful difference in mean change of scores from screening in the axi-cel arm compared with SOC at study day 100 (estimated difference = 13.7; 95% CI, 8.5 to 18.8; adjusted P < 0.0001) and study day 150 (estimated difference = 11.3; 95% CI, 5.4 to 17.1; adjusted P = 0.0004).

Progression-Free Survival

PFS was a secondary outcome. At the data cut-off date, the median duration of PFS based on investigator disease assessments was 14.9 months (95% CI, 7.2 months to not estimable [NE]) in the axi-cel arm and 5.0 months (95% CI, 3.4 months to 8.5 months) in the SOC arm, with a stratified HR of 0.562 (95% CI, 0.414 to 0.762).

Objective Response Rate

Objective response rate (ORR) as determined by blinded central assessment was a key secondary outcome in ZUMA-7. The ORR (CR or PR) as determined by blinded central assessment was 83% in the axi-cel arm and 50% in the SOC arm (difference in ORR = 33.1%; 95% CI, 23.2% to 42.1%; P < 0.0001).

The CR rates in the axi-cel arm and the SOC arm were 65% (95% CI, 57.6% to 71.9%; n = 117) and 32% (95% CI, 25.6% to 39.8%; n = 58), respectively. The PR rates were 18% (95% CI, 13.0% to 24.8%; n = 33) in the axi-cel arm and 18% (95% CI, 12.6%, 24.3%; n = 32) in the SOC arm.

Duration of Response

Duration of response (DOR) was a secondary outcome. For the 150 patients in the axi-cel arm and the 90 patients in the SOC arm who achieved an objective response of CR or PR as determined by blinded central assessment, the Kaplan-Meier estimated median DOR was 26.9 months (95% CI, 13.6 months to NE) in the axi-cel arm compared with 8.9 months (95% CI, 5.7 months to NE) in the SOC arm (stratified HR = 0.736; 95% CI, 0.488 to 1.108).

The Kaplan-Meier estimates of the percentage of patients who remained in response at 12 and 24 months from first objective response were 60.9% (95% CI, 52.4% to 68.4%) and 54.0% (95% CI, 45.1% to 62.0%), respectively, in the axi-cel arm compared with 47.6% (95% CI, 35.2% to 58.9%) and 45.6% (95% CI, 33.2% to 57.1%), respectively, in the SOC arm.

Time to Next Treatment

Time to next treatment was an exploratory outcome. Time to next treatment events occurred for 99 patients (55%) in the axi-cel arm and 135 patients (75%) in the SOC arm. The median time to next treatment was 14.7 months (95% CI, 6.5 months to NE) and 3.4 months (95% CI, 3.1 months to 4.4 months), respectively (stratified HR = 0.430; 95% CI, 0.329 to 0.560).

Health Care Resource Utilization

A total of 42 patients (25%) in the axi-cel arm and 9 patients (5%) in the SOC arm were admitted to the intensive care unit (ICU). Median duration of ICU hospitalization was 5 days (range, 1 to 12 days) in the axi-cel arm and 3 days (range, 2 to 17 days) in the SOC arm. Median duration of hospitalization for axi-cel infusion was 16 days (range, 5 to 103 days); median duration of inpatient hospitalization for stem cell transplant in the SOC arm was 21 days (range, 1 to 53 days).

Table 2: Summary of Key Efficacy Results From the ZUMA-7 Study (Full Analysis Set)

axi-cel = axicabtagene ciloleucel; CI = confidence interval; CR = complete response; DOR = duration of response; EFS = event-free survival; NA = not applicable; NE = not estimable; NR = not reached; ORR = objective response rate; OS = overall survival; PFS = progression-free survival; PR = partial response; SOC = standard of care.

^aAlpha allocated to this analysis: 1-sided P = 0.004.

^bThe stratification factors are responses to first-line therapy (primary refractory vs. relapse ≤ 6 months of first-line therapy vs. relapse > 6 and ≤ 12 months of first-line therapy) and second-line age-adjusted International Prognostic Index (0 to 1 vs. 2 to 3) as collected via interactive voice/web response system.

^cEstimated using the Kaplan-Meier method.

^dEstimated using the reverse Kaplan-Meier approach.

^eHazard ratio and 95% CIs were estimated using Cox regression models for axi-cel relative to SOC. The Breslow method was used to handle ties for the Cox regression models.

^f95% CI for the difference in ORR (with the SOC arm as reference group) was from the Wilson score method with continuity correction.

^gOne-sided P value from the Cochran-Mantel-Haenszel test.

^hEstimated using the Clopper-Pearson method.

ⁱP value is based on the log-rank test.

^jAnalysis not part of the statistical hierarchy.

Source: Clinical Study Report for Yescarta.¹⁵

Harms Results

Key harms results from the ZUMA-7 study are summarized in Table 3.

All patients in the axi-cel arm (170 patients; 100%) and the SOC arm (168; 100%) had at least 1 treatment-emergent adverse event (TEAE); 155 patients (91%) in the axi-cel arm and 140 patients (83%) in the SOC arm had grade 3 or higher TEAEs. The most frequently reported TEAEs of worst grade 3 or higher (reported in ≥ 20% of patients in either treatment arm) were neutropenia (73 patients; 43%); anemia (51 patients; 30%); neutrophil count decreased (49 patients; 29%); and white blood cell count decreased (43 patients; 25%) in the axi-cel arm, and anemia (65 patients; 39%); platelet count decreased (60 patients; 36%); decreased neutrophil count (47 patients; 28%); febrile neutropenia (46 patients; 27%); and thrombocytopenia (37 patients; 22%) in the SOC arm.

A total of 85 patients (50%) in the axi-cel arm and 77 patients (46%) in the SOC arm had at least 1 serious adverse event (SAE). The most frequently (in ≥ 5% of patients) reported SAEs of any grade in each treatment arm were pyrexia (27 patients; 16%), encephalopathy (17 patients; 10%), hypotension (15 patients; 9%), aphasia (9 patients; 5%), and pneumonia (8 patients; 5%) in the axi-cel arm, and febrile neutropenia (22 patients; 13%) and acute kidney injury and pyrexia (8 patients each; 5%) in the SOC arm.

No patient discontinued treatment due to TEAEs in the axi-cel arm. Two patients in the SOC arm discontinued treatment due to TEAEs of grade 4 acute kidney injury and grade 1 blood stem cell harvest failure.

In the axi-cel arm, 64 patients (38%) had died by the data cut-off date due to progressive disease (PD) (n = 47; 28%), TEAEs (n = 6; 4%), and other reasons (n = 10; 6%) and 1 patient (1%) had died from an event reported by the investigator as a “secondary malignancy” (lung adenocarcinoma). In the SOC arm, 78 patients (46%) had died by the data cut-off date due to PD (n = 64; 38%), TEAEs (n = 2; 1%), or other reasons (n = 12; 7%).

In total, 102 patients (60%) in the axi-cel arm and 33 patients (20%) in the SOC arm had at least 1 treatment-emergent neurologic event, including 36 patients (21%) and 1 patient (1%) in the axi-cel and SOC arms, respectively, with worst grade 3 or higher neurologic events. Of these, 10 patients (6%) in the axi-cel arm had worst grade 4 neurologic events; no patients in either treatment arm had a grade 5 neurologic event. Serious treatment-emergent neurologic events of any grade were reported for 34 patients (20%) in the axi-cel arm and 1 patient (1%) in the SOC arm, including 26 patients (15%) in the axi-cel arm with a serious worst grade 3 or higher neurologic event and 1 patient (1%) in the SOC arm with a serious worst grade 2 neurologic event.

CRS of any grade was reported for 157 patients (92%), including 11 patients (6%) who had worst grade 3 or higher CRS, in the axi-cel arm. No patient had grade 5 CRS.

The most common cytopenias of any grade reported in the axi-cel arm were thrombocytopenia (50 patients; 29%), neutropenia (122 patients; 72%), and anemia (73 patients; 43%). The most common cytopenias of any grade reported in the SOC arm were thrombocytopenia (101 patients; 60%), neutropenia (92 patients; 55%), and anemia (92 patients; 55%).

Seventy patients (41%) in the axi-cel arm and 51 patients (30%) in the SOC arm had at least 1 treatment-emergent infection, including 24 patients (14%) and 19 patients (11%), respectively, with worst grade 3 or higher infections. Three patients (2%) in the axi-cel arm and 6 patients (4%) in the SOC arm had worst grade 4 infections. Five patients (3%) in the axi-cel arm had a grade 5 TEAE of infection (2 patients with COVID-19, 1 patient with progressive multifocal leukoencephalopathy, 1 patient with hepatitis B reactivation, and 1 patient with sepsis). No patients in the SOC arm had a grade 5 TEAE of infection.

Critical Appraisal

The ZUMA-7 trial was open label. Despite the open-label design, there is low risk of bias in the measurement of outcomes such as EFS, PFS, and ORR, which were assessed via independent blinded radiologic review of disease response, or OS, which is an objective outcome. Although independent blinded radiologic review of disease response was performed as the primary analysis to minimize investigator bias, patients’ knowledge of their assigned treatment could impact HRQoL data and any subjective harm (which are particularly susceptible to bias from a lack of blinding of patients to their treatment). The HRQoL data were also at high risk of attrition bias because there were large amounts of missing data for all follow-up time points and the amount of missing data was not balanced across treatment arms. The HRQoL tools were not validated in patients with LBCL. The primary and secondary efficacy end points of EFS, ORR, and OS are considered appropriate for the disease setting. The OS data were immature and, because the results are based on an interim analysis, there is a risk that the effect of axi-cel compared with SOC on survival is overestimated.^17,18 Although there is some evidence to suggest EFS is a robust surrogate end point for OS in hematological malignancies,¹⁹ it is unknown whether this could be extended to CAR T-cell therapies in relapsed or refractory LBCL.

The ZUMA-7 trial included patients with relapsed or refractory LBCL with a wide range of clinical presentations, but patients with HIV and those with an ECOG PS score of 2 or more were excluded. The clinical experts consulted by CADTH indicated that these patient groups should be eligible for CAR T-cell therapies including axi-cel. While the SOC treatment including salvage chemotherapy regimens used in the SOC arm of the ZUMA-7 trial are reflective of clinical practice in Canada, the clinical experts noted the challenges in reproducing the same treatment processes for axi-cel treatment in the real-world setting, notably the rapid manufacturing time (13 days in the ZUMA-7 trial). According to the clinical experts, delays in manufacturing and access times to axi-cel treatment would potentially compromise patient outcomes as the probability of disease progression or other complications increase with longer axi-cel manufacturing wait times. Data for all outcomes considered important to patients, clinicians, and drug plans, as per the systematic review protocol, were collected and reported; however, conclusions could not be drawn for effects of axi-cel compared with SOC on HRQoL due to limitations in the data.

Other Relevant Evidence

The sponsor provided long-term (≥ 4 year and ≥ 5 year) data from ZUMA-1, the pivotal multicentre, single-arm, registrational phase I and II study of axi-cel in patients with relapsed or refractory LBCL.^20-22

In the 2-year analysis of ZUMA-1 (n = 101; median follow-up from axi-cel dosing to data cut-off = 27.1 months), the ORR was 83%; 58% of patients achieved a CR. The 2-year OS rate was 50.5%.

The most recently updated survival results from the phase II ZUMA-1 study showed a 5-year OS rate of 42.6% (95% CI, 32.8% to 51.9%) among patients treated with axi-cel after 5 years of follow-up. The 5-year OS rate among complete responders was 64.4% (95% CI, 50.8% to 75.1%). The median survival time among complete responders was not reached (95% CI, 63.4% to NE). Of 59 patients who achieved CR, 37 (63%) were alive at the 5-year data cut-off date. Since the 4-year data cut-off date, 1 death at month 63 (CR) and 1 PD at month 54 (PR) were observed.

Supportive safety data comparing axi-cel-treated populations in ZUMA-7 and ZUMA-1 suggest comparable TEAEs between the 2 trials.

Critical Appraisal

The ZUMA-1 trial was critically appraised in the 2019 CADTH Optimal Use Report report.²³ ZUMA-1 was a single-arm, phase I and II study of axi-cel in patients with relapsed or refractory LBCL who had received at least 2 previous systemic therapies. The primary limitation of ZUMA-1 was the absence of a comparator group against which the treatment benefits and harms of axi-cel could be compared. As such, causal effects cannot be inferred. In addition, patients in ZUMA-1 received axi-cel as third- or later-line treatment. It is unknown whether results are generalizable to patients treated with axi-cel as second-line treatment, which is the indication under review.

Conclusions

Based on data from the ZUMA-7 trial, the anti-CD19 CAR T-cell axi-cel led to an improvement in EFS and PFS compared with SOC HDT-ASCT as second-line treatment in ASCT-eligible patients with relapsed or refractory LBCL within 12 months of first-line chemoimmunotherapy. It is not yet clear whether EFS benefits will translate to improved OS as the data remain immature and follow-up is ongoing. Because the OS results are based on an interim analysis, there is a risk that the effect of axi-cel compared with SOC on survival is overestimated. The risks associated with axi-cel treatment are consistent with its mechanism of action and acceptable given the high-risk patient population, but it must be administered in specialized centres equipped to manage CAR T-cell therapy–related toxicities. Although the most common toxicities were similar in the ZUMA-7 and ZUMA-1 studies, the lack of long-term safety and efficacy data from the ZUMA-7 trial remains an important limitation. A longer follow-up time is needed to better establish the survival benefit and long-term safety of axi-cel, as relapse of lymphoma and potential late toxicities of CAR T-cell therapy such as secondary malignancies may emerge with longer follow-up. Uncertainties remain regarding the implementation of CAR T-cell therapy and the support system needed to optimize timely access and deliverability of axi-cel in the real-world setting.

Introduction

Disease Background

NHL comprises a diverse group of at least 60 closely related cancers of the lymphocytes.¹ About 85% to 90% of NHLs are derived from B lymphocytes; the remaining lymphomas are derived from T lymphocytes or natural killer cells.³ NHL is the most prevalent hematological malignancy and the fifth most common cancer diagnosed in Canada.¹ An estimated 11,400 Canadians were diagnosed with NHL in 2022 and 3,000 died from it.² Risk factors for NHL include immune disorders (e.g., rheumatoid arthritis, Sjögren syndrome, systemic lupus erythematosus), use of immunosuppressive therapies, bacterial and viral infections (e.g., Helicobacter pylori, Epstein-Barr virus, and hepatitis C virus), family history, and genetic, occupational, and lifestyle risk factors.³ NHL can affect any organ in the body and have a wide range of clinical presentations.³ Most patients with NHL present with painless enlarged lymph nodes with or without systemic symptoms (e.g., fevers, night sweats, weight loss, and fatigue); however, because NHL can involve any organ in the body, a wide range of presentations that mimic other conditions are possible.³ Diagnosis is established based on a biopsy sample with immunohistochemical and genetic testing, ideally from an excisional biopsy of an involved lymph node or a tumour in another organ, or a cutting-needle biopsy if an excisional biopsy is impractical.³ The differential diagnosis and classification of the many subtypes of NHL are complex and have evolved with advances in clinical science and understanding of their distinctive molecular and immunophenotypic features.

LBCL is the most prevalent subtype of NHL in Canada, constituting 30% to 40% of all NHL cases.⁴ LBCL comprises a group of cancers of B lymphocytes that develop and mature in the bone marrow and lymph nodes. The B-cell lineage-specific antigens CD19 and CD20 are highly and consistently expressed on the cell surface of most B-cell lymphomas. DLBCL not otherwise specified (NOS) is the most common subtype of LBCL, accounting for more than 80% of cases.²⁴ DLBCL comprises a group of LBCL cases that remain biologically and clinically heterogeneous, for which there are no clear and accepted criteria for subclassification. Other subtypes of LBCL include DLBCL transformed from follicular lymphoma or chronic lymphocytic leukemia; high-grade B-cell lymphoma (HGBL); primary DLBCL of the central nervous system (CNS); primary cutaneous DLBCL, leg type; Epstein-Barr virus–positive DLBCL; DLBCL associated with chronic inflammation; and T-cell/histiocyte-rich LBCL.^24-27 The median age at diagnosis of LBCL is 65 years, with 30% of patients older than 75 years. LBCL is slightly more common among males.⁵ LBCL often manifests as a rapidly enlarging painless mass at a nodal or extranodal site, and other general symptoms such as enlarged lymph nodes, fatigue, cough, itchy skin, and loss of appetite.⁵ In addition, 30% to 40% of patients present with “B symptoms,” including fevers, night sweats, and unexplained weight loss.^5,24 LBCL is an aggressive but potentially curable NHL, and is typically diagnosed at an advanced stage (stage III or IV).⁵ The 5-year PFS in patients with limited disease is 80% to 85%, whereas those with advanced disease have a 5-year PFS of about 50%.⁶ Patients with LBCL experience a high symptom burden and poorer HRQoL compared with a general cancer reference population, with further decline in HRQoL observed with disease progression and subsequent lines of therapy.^28,29

Patients with newly diagnosed LBCL are routinely treated with the chemotherapy regimen CHOP in combination with the anti-CD20 monoclonal antibody rituximab (R-CHOP). Treatment with R-CHOP is widely recognized as the first-line SOC treatment for LBCL based on Canadian and international clinical practice guidelines.^4,7-10 Although many patients respond well to R-CHOP and achieve long-lasting remission, a large proportion (30% to 50%) will either be refractory to or relapse following front-line treatment.^4,11 Most relapses occur within the 2 to 3 years after initial treatment.⁴ OS in patients with primary refractory disease is very low, with only 15% to 20% surviving at 5 years.¹² Patients with partial or complete response to first-line treatment also have poor survival at relapse, with 38% and 42% surviving at 5 years, respectively.¹² Patients with primary refractory disease or early relapsed disease (less than 12 months) have worse outcomes than patients who relapse more than 12 months after first-line treatment.^30,31 Patients with relapsed or refractory LBCL who relapsed within 12 months of initial treatment had worse response rates to second-line salvage chemotherapy than patients who relapsed after 12 months.³² Another recent Canadian real-world study of patients with relapsed or refractory LBCL found a median OS of 5.1 months, with 15% of patients alive at 3 years. In this population, 82% progressed within 12 months of first-line chemoimmunotherapy, and these patients had significantly worse 3-year OS than those who relapsed after 12 months (10% versus 39%, respectively).¹³

Standards of Therapy

For patients who relapse or who have disease that is refractory to first-line chemoimmunotherapy, second-line treatment comprises salvage chemotherapy, followed by curative-intent HDT and ASCT, if the patients meet the eligibility criteria and have chemosensitive disease. The goal of salvage chemotherapy is to reduce tumour burden and demonstrate chemosensitivity before proceeding to HDT-ASCT. Achieving better response to salvage chemotherapy is associated with better outcomes following ASCT, but no salvage regimen has been shown to be superior to another.^4,33 Commonly used salvage chemotherapy regimens in Canada include rituximab plus gemcitabine, dexamethasone, and cisplatin (R-GDP), rituximab plus ifosfamide, carboplatin, and etoposide (R-ICE), rituximab plus dexamethasone, high-dose cytarabine, and cisplatin (R-DHAP), and rituximab plus dexamethasone, cyclophosphamide, etoposide, and cisplatin (R-DICEP).⁴ Patients who have a CR or PR to salvage chemotherapy will then have their stem cells mobilized and collected by leukapheresis and be administered HDT consisting of a combination chemotherapy such as etoposide plus melphalan or carmustine (BCNU), etoposide, cytarabine (Ara-C), and melphalan (BEAM). The stem cells are then infused back into the patient as part of the ASCT. Although HDT-ASCT has curative potential, many patients are not fit enough for transplant, do not respond to salvage chemotherapy (i.e., do not have chemosensitive disease), or relapse following ASCT. Only about half of the patients with relapsed or refractory LBCL are fit enough for transplant (i.e., have adequate organ function with no major comorbidities), and only half of transplant-eligible patients respond to salvage chemotherapy and can proceed to ASCT.^32,34 Of the patients who receive ASCT (25% of patients with relapsed or refractory LBCL), approximately half remain event-free 4 years after transplant.^32,35

Patients with relapsed or refractory LBCL who are ineligible for ASCT, do not respond to salvage chemotherapy, or relapse post-ASCT have limited treatment options and a poor prognosis.⁴ Treatment options for these patients include palliative chemotherapy, radiotherapy, clinical trials, or third-line CAR T-cell therapy. Polatuzumab vedotin plus bendamustine and rituximab (Pola-BR) may provide a future treatment option for patients who are ineligible for ASCT as it is currently Health Canada–approved in this setting and has received a positive reimbursement recommendation from CADTH.^4,36 The clinical experts consulted by CADTH indicated that provincial funding for CAR T-cell therapy currently varies and some patients are required to travel out of province or country to access it. For patients who are ineligible for CAR T-cell therapy or fail treatment, there is no standard treatment approach and further treatment depends on access to clinical trials.

Drug

Axi-cel (brand name Yescarta) is a CD19-directed genetically modified autologous T-cell immunotherapy that binds to CD19-expressing cancer cells and normal B cells.³⁷ Following anti-CD19 CAR T-cell engagement with CD19-expressing target cells, the CD28 and CD3-zeta costimulatory domains activate downstream signalling cascades that lead to T-cell activation, proliferation, acquisition of effector functions, and secretion of inflammatory cytokines and chemokines. This sequence of events leads to killing of CD19-expressing cells.³⁷ To prepare axi-cel, a patient’s own T cells are harvested and genetically modified ex vivo by retroviral transduction to express a CAR protein comprising a murine anti-CD19 single-chain variable fragment (scFV) linked to CD28 and CD3-zeta costimulatory domains. The anti-CD19 CAR T cells are expanded and infused back into the patient, where they recognize and eliminate CD19-expressing target cells.

Axi-cel was first approved by the FDA (FDA) in October 2017 for the treatment of adult patients with relapsed or refractory LBCL after 2 or more lines of systemic therapy, including DLBCL NOS, primary mediastinal B-cell lymphoma (PMBCL), HGBL, and DLBCL arising from follicular lymphoma.³⁸ Approval by the European Medicines Agency followed in August 2018 as a treatment for adult patients with relapsed or refractory DLBCL and primary mediastinal B-cell lymphoma after 2 or more lines of systemic therapy.³⁹

In Canada, axi-cel is currently approved for the treatment of patients with relapsed or refractory LBCL who have received at least 2 previous systemic therapies.³⁷ Of note, this CADTH reimbursement review was conducted before issuance of Health Canada NOC and the scope was based on the anticipated indication. The anticipated indication during the time of this review was for the treatment of adult patients with LBCL that is refractory to first-line chemoimmunotherapy or that relapses within 12 months of first-line chemoimmunotherapy. The sponsor’s reimbursement request is the same as the anticipated Health Canada indication, except that it narrows the population to patients who are candidates for ASCT. NOC was received on December 6, 2022, and the final approved indication is for the treatment of adult patients with DLBCL or HGBL that is refractory to first-line chemoimmunotherapy or that relapses within 12 months of first-line chemoimmunotherapy. In 2019, CADTH published an Optimal Use Report evaluating the beneficial and harmful effects of axi-cel for eligible types of relapsed or refractory B-cell lymphomas in adult patients.²³

Axi-cel is a single-dose, 1-time treatment in a patient-specific infusion bag. Axi-cel should be administered by experienced health professionals at specialized treatment centres. Each patient-specific, single-infusion bag contains a suspension of anti-CD19 CAR-positive viable T cells in approximately 60 mL for a target dose of 2 × 10⁶ anti-CD19 CAR T cells/kg body weight (range, 1 × 10⁶ cells/kg to 2.4 × 10⁶ cells/kg), with a maximum of 2 × 10⁸ anti-CD19 CAR T cells for patients weighing 100 kg and above.³⁷

Stakeholder Perspectives

Patient Group Input

This section was prepared by CADTH staff based on the input provided by patient groups. The full original patient input(s) received by CADTH have been included in the stakeholder section at the end of this report.

One patient advocacy group, Lymphoma Canada, provided input on the treatment of adult patients with relapsed or refractory LBCL. This group gathered information from patients with DLBCL living in Canada through 2 online surveys, 1 in 2018 and another between April 26 and June 20, 2022. A total 97 patients with DLBCL completed the first survey and 23 patients with DLBCL completed the first and second survey, respectively. Patients from the second survey reported that fatigue (83%) was the most common physical symptom affecting their quality of life, followed by body aches and pains (67%), thrombocytopenia (67%), enlarged lymph nodes (58%), anemia (58%), neutropenia (58%) and night sweats (58%). Patients from both surveys reported that fear of progression/relapse was the most common psychosocial impact (67%) affecting quality of life, followed by anxiety (37%), memory loss (37%), and concentration problems (36%).

The majority of respondents (83%) were treated with CHOP with or without rituximab as their first-line treatment. When describing their experiences with their current or most recent therapy managing DLBCL symptoms, 6 out of 11 patients (55%) from the second survey agreed with the statement that it was able to manage their symptoms. While describing the side effects caused by the treatment, patients from both surveys reported that hair loss (93%) was the most common side effect, followed by fatigue (85%), neutropenia (70%), “chemo brain” or memory problems and confusion (67%), and nausea (60%). The most difficult-to-tolerate side effect was reported to be fatigue (40%). Long-term treatment side effects (lasting longer than 2 years or appearing 2 years or later after the end of treatment) were reported to be fatigue (52%) and chemo brain (42%). Patients from the first survey reported that their lymphoma treatment had the greatest negative impact on their work, travel, and other activities.

Regarding the importance of outcomes of new treatments on a scale of 1 (“not important”) to 5 (“very important”), patients from the first survey rated all of the following as important: longer survival (mean = 4.88), longer remission (mean = 4.84), better quality of life (mean = 4.64), and fewer side effects (mean = 4.12). Moreover, 51% of patients from both surveys stated that they would choose a treatment with potentially serious side effects if their doctor recommended it to be the most effective treatment option for DLBCL, whereas 46% of patients were not sure. Respondents from the first survey indicated treatment choice as important, with an average score of 4.54 out of 5. In addition, 91% patients from the second survey felt a need for more therapy options for patients with DLBCL.

Three patients from the second survey reported receiving axi-cel as third-line therapy. All 3 patients were away from home for more than 3 months while receiving the treatment, pointing to the challenges in accessing treatment currently. All 3 patients reported thrombocytopenia as a side effect, and 2 patients reported fever, anemia, nausea or vomiting, neutropenia, diarrhea, joint or muscle pain, and fatigue as side effects. Fear of progression/relapse and difficulty sleeping were reported by all 3 patients as psychosocial impacts related to their CAR T-cell therapy.

Clinician Input

Input From Clinical Experts Consulted by CADTH

All CADTH review teams include at least 1 clinical specialist with expertise in the diagnosis and management of the condition for which the drug is indicated. Clinical experts are a critical part of the review team and are involved in all phases of the review process (e.g., providing guidance on the development of the review protocol; assisting in the critical appraisal of clinical evidence; interpreting the clinical relevance of the results; and providing guidance on the potential place in therapy). In addition, as part of the axi-cel review, a panel of 4 clinical experts from across Canada was convened to characterize unmet therapeutic needs, assist in identifying and communicating situations where there are gaps in the evidence that could be addressed through the collection of additional data, promote the early identification of potential implementation challenges, gain further insight into the clinical management of patients living with a condition, and explore the potential place in therapy of the drug (e.g., potential reimbursement conditions). A summary of this panel discussion is presented below.

Unmet Needs

The clinical experts highlighted that the most important goal of treatment for patients with LBCL is a cure. The clinical experts reported that about half of patients will be cured with front-line therapy. For patients who do not respond to front-line therapy or who relapse, intensive therapy with curative intent such as stem cell transplant is considered in second line. However, the clinical panel reported that only half of patients who relapse after front-line therapy or are refractory to front-line therapy will respond to salvage chemotherapy and proceed to transplant, per the clinical experts. The clinical experts noted that some patients do not tolerate or do not respond to salvage chemotherapy and therefore cannot continue to transplant. Others, who are responsive to salvage therapy may not continue to transplant for example those whose hematopoietic progenitor cells cannot be collected. Some patients may not be eligible for transplant, including patients with comorbidities or inadequate organ function. The clinical panel highlighted that patients with active CNS lymphoma and those who progress or relapse with CNS disease are a particularly difficult to treat population. In addition, patients with genetically high-risk disease (double-hit or triple-hit lymphoma) for whom the likelihood of responding to conventional chemotherapy is very low have limited treatment options, per the clinical experts. The clinical experts indicated that patients with relapsed or refractory LBCL who cannot tolerate or are ineligible for transplant have the largest unmet treatment need. The clinical panel noted that salvage chemotherapy and ASCT is an arduous treatment regimen and is associated with serious toxicities. The clinical panel indicated that more effective and safe treatment options are needed that are better tolerated than current conventional treatment and that can be used soon after failure of front-line therapy to improve patient outcomes.

Place in Therapy

The clinical experts noted that axi-cel is already in the current treatment paradigm as it is approved for the treatment of relapsed or refractory LBCL in third line. The clinical experts indicated that axi-cel fits well, earlier in the line of treatment and can move up to the second line and replace ASCT for most patients. The clinical experts noted that they expect patient outcomes to be better when the patients receive a potentially curative therapy earlier in the course of disease. The clinical experts noted that some patients deteriorate rapidly and thus may be less likely to survive if definitive treatment is delayed. The clinical panel thought that patients with primary refractory LBCL would benefit from CAR T-cell therapy in the second-line setting, as current salvage therapies can be ineffective and toxic in many patients, thereby preventing accessing or decreasing the efficacy of curative therapy. The clinical experts believed that a significant problem of the current treatment paradigm is that CAR T-cell therapy is only approved for third-line treatment, as this limits the number of patients with relapsed or refractory LBCL who can potentially benefit from a therapy that has demonstrated efficacy over SOC.

Patient Population

The clinical experts noted that although the ZUMA-7 trial recruited only patients who were eligible for ASCT, in clinical practice there is no clinical rationale for restricting axi-cel only to those who are candidates for ASCT. The clinical experts reported that there are patients who would be ineligible for ASCT based on fitness (i.e., comorbidities), whom the clinical experts thought should be eligible for CAR T-cell therapy. The clinical experts indicated that any patient with adequate organ function and good performance status (ECOG ≤ 2) who, based on the clinician’s judgment can tolerate the known toxicities of CAR T-cell therapy (e.g., CRS) would be suitable for axi-cel treatment. They also indicated that there are no uniformly implemented, easily reproducible or objective criteria for ASCT eligibility in Canada. Suitability for ASCT thus varies across provinces or even treatment centres depending on resources. For example, although some centres may use age as an eligibility criterion for ASCT, other centres may not exclude older patients if they have adequate organ function and performance status.

Assessing Response to Treatment

The clinical panel reported that response to treatment is typically assessed by CT scan or PET scan at 1 month and then at 3 months after initiation of therapy. The clinical panel noted that the 3-month assessment is important for assessing prognosis because patients who achieve CR at 3 months tend to have favourable outcomes, whereas those who have a PR or no response at the 3-month assessment tend to have poor long-term outcomes. The clinical experts highlighted that HRQoL is also important to assess to determine effectiveness of treatment.

Discontinuing Treatment

The clinical panel noted that in some cases patients may go through pretreatment (leukapheresis, bridging therapy, and conditioning chemotherapy), but do not receive axi-cel. These situations include lymphoma progression despite salvage and/or bridging therapy to the extent that performance status or vital organ function becomes a major concern (e.g., CNS disease, pulmonary disease), serious opportunistic infections or serious end-organ damage, and less commonly, therapy-related organ impairment (e.g., renal impairment) that limits patients’ predicted ability to tolerate CAR T-cell therapy. The clinical experts pointed out that the probability of disease progression or other complications increase with longer CAR T cell manufacturing wait times. The clinical experts noted that the 13-day manufacturing time reported in the ZUMA-7 study is rapid and may not be reproducible outside the clinical trial setting. According to the clinical experts, management of patients in such situations depends on patients’ disease-related characteristics; in most circumstances, lymphoma progression at this stage would require palliation and abandonment of axi-cel treatment. Organ toxicity could be managed, and reversed in rare circumstances, while infections could be managed for those who are amenable to treatment in the short term, as most patients are very immune deficient at this stage of the disease. The clinical experts reiterated that CAR T-cell therapy given as a second-line treatment (compared to third line or fourth line) would reduce the likelihood of patients progress or become too unwell after salvage treatment and before axi-cel infusion and are unable to complete CAR T-cell therapy.

For the 50% to 60% of patients who are expected to fail CAR T-cell therapy, subsequent treatment options currently include pola-BR (polatuzumab vedotin plus bendamustine and rituximab) or further chemotherapy, often in a palliative setting.

Prescribing Conditions

The clinical experts noted that axi-cel treatment can be provided by oncologists or hematologists in a hospital setting with adequate infrastructure for cell therapy and with access to highly specialized multidisciplinary clinical care including critical/intensive care and specialist care (e.g., neurology, nephrology) to manage toxicities, as well as laboratory support to handle and process samples.

Clinician Group Input

This section was prepared by CADTH staff based on the input provided by clinician groups. The full original clinician group input(s) received by CADTH have been included in the stakeholder section at the end of this report.

Clinician group input was received from 3 groups: Lymphoma Canada, Ontario Health (CCO) Hematology Cancer Drug Advisory Committee, and Cell Therapy Transplant Canada (CTTC).

The clinician groups agreed that, based on Canadian and international real-world evidence, there are unmet needs and challenges in the current second-line chemoimmunotherapy and ASCT treatment for patients with relapsed or refractory LBCL. The clinician groups indicated that there may be limited eligibility or tolerability to further salvage chemotherapy for some patients (e.g., patients with primary refractory disease or early relapse, and older patients). Moreover, toxicities such as febrile neutropenia, bacteremia and other infections, gastrointestinal toxicity, mucositis, need for transfusion support, and secondary malignancies associated with ASCT treatment have made it unsuitable for patients who are high risk, who are refractory to treatment or who relapse within 12 months of diagnosis.

The clinician groups also pointed out that they thought axi-cel could be expanded to specific groups of patients (e.g., older patients, patients with comorbidities, patients with ECOG PS score of 2) who may be unable to tolerate HDT and ASCT but could tolerate axi-cel. The clinician groups emphasized that CAR T-cell therapy needs to be prescribed and administered only by specialized, accredited centres for cell therapy programs, given the unique toxicities (notably CRS and immune effector cell-associated neurotoxicity syndrome [ICANS]) associated with this treatment.

Drug Program Input

The drug programs provide input on each drug being reviewed through CADTH’s reimbursement review processes by identifying issues that may impact their ability to implement a recommendation. The implementation questions and corresponding responses from the clinical experts consulted by CADTH are summarized in Table 4.

Table 4: Summary of Drug Plan Input and Clinical Expert Response

allo = allogeneic; ASCT = autologous stem cell transplant; axi-cel = axicabtagene ciloleucel; CAR = chimeric antigen receptor; CLL = chronic lymphocyte leukemia; CNS = central nervous system; CR = complete response; CRS = cytokine release syndrome; DLBCL = diffuse large B-cell lymphoma; ECOG = Eastern Cooperative Oncology Group; FL = follicular lymphoma; ICU = intensive care unit; IL = interleukin; LBCL = large B-cell lymphoma; PAG = provincial advisory group; pERC = CADTH pan-Canadian Oncology Drug Review Expert Review Committee; PMBCL = primary mediastinal B-cell lymphoma; R-CHOP = rituximab, cyclophosphamide, doxorubicin hydrochloride, vincristine, and prednisone; R-DHAP = rituximab, dexamethasone, cytarabine, and cisplatin; R-ESHAP = rituximab, etoposide, solu-medrone, high-dose cytarabine, and cisplatin; R-GDP = rituximab, gemcitabine, dexamethasone, and cisplatin; R-GemOx = rituximab, gemcitabine, and oxaliplatin; R-ICE = rituximab, ifosfamide, carboplatin, and etoposide; pola-BR = polatuzumab vedotin, bendamustine, and rituximab.

Clinical Evidence

The clinical evidence included in the review of axi-cel (Yescarta) is presented in 3 sections. The first section, the systematic review, includes pivotal studies provided in the sponsor’s submission to CADTH and Health Canada, as well as those studies that were selected according to an a priori protocol. The second section includes indirect evidence from the sponsor (if submitted) and indirect evidence selected from the literature that met the selection criteria specified in the review. The third section includes sponsor-submitted long-term extension studies and additional relevant studies that were considered to address important gaps in the evidence included in the systematic review.

Systematic Review (Pivotal and Protocol-Selected Studies)

Objectives

To perform a systematic review of the efficacy and safety of axi-cel (IV infusion, target dose of 2 × 10⁶ anti-CD19 CAR T cells/kg body weight) for the treatment of adult patients with relapsed or refractory LBCL.

Methods

Studies selected for inclusion in the systematic review included pivotal studies provided in the sponsor’s submission to CADTH and Health Canada, as well as those meeting the selection criteria presented in Table 5. Outcomes included in the CADTH review protocol reflect outcomes considered to be important to patients, clinicians, and drug plans. Of note, the systematic review protocol presented below was established before the granting of a Notice of Compliance from Health Canada.

Table 5: Inclusion Criteria for the Systematic Review

ASCT = autologous stem cell transplant; CAR = chimeric antigen receptor; ECOG PS = Eastern Cooperative Oncology Group Performance Status; HRQoL = health-related quality of life; ICU = intensive care unit.

The literature search for clinical studies was performed by an information specialist using a peer-reviewed search strategy according to the PRESS Peer Review of Electronic Search Strategies checklist.⁴⁰

Published literature was identified by searching the following bibliographic databases: MEDLINE All (1946—) via Ovid and Embase (1974—) via Ovid. All Ovid searches were run simultaneously as a multifile search. Duplicates were removed using Ovid deduplication for multifile searches, followed by manual deduplication in Endnote. The search strategy comprised both controlled vocabulary, such as the National Library of Medicine’s MeSH (Medical Subject Headings), and keywords. The main search concept was Yescarta (axicabtagene ciloleucel). The following clinical trials registries were searched: the US National Institutes of Health’s clinicaltrials.gov, the WHO’s International Clinical Trials Registry Platform (ICTRP) search portal, Health Canada’s Clinical Trials Database, and the European Union Clinical Trials Register.

No filters were applied to limit the retrieval by study type. Retrieval was not limited by publication date or by language. Conference abstracts were excluded from the search results. Refer to Appendix 1 for the detailed search strategies.

The initial search was completed on July 7, 2022. Regular alerts updated the search until the meeting of the CADTH pan-Canadian Oncology Drug Review Expert Review Committee (pERC) on November 9, 2022.

Grey literature (literature that is not commercially published) was identified by searching relevant websites from the Grey Matters: A Practical Tool for Searching Health-Related Grey Literature checklist.⁴¹ Included in this search were the websites of regulatory agencies (US FDA and the European Medicines Agency). Google was used to search for additional internet-based materials. Refer to Appendix 1 for more information on the grey literature search strategy.

These searches were supplemented by reviewing bibliographies of key papers and through contacts with appropriate experts. In addition, the manufacturer of the drug was contacted for information about unpublished studies.

Two CADTH clinical reviewers independently selected studies for inclusion in the review based on titles and abstracts, according to the predetermined protocol. Full-text articles of all citations considered potentially relevant by at least 1 reviewer were acquired. Reviewers independently made the final selection of studies to be included in the review, and differences were resolved through discussion.

Findings From the Literature

A total of 484 records were identified via the searches (Figure 1). Of these, 483 records were excluded; 1 study was retained. An additional 2 reports provided by the sponsor were included, as was an FDA review report. In total, 4 reports presenting data from a single study, ZUMA-7, were included in this review.^14,15,42 Details of the ZUMA-7 study are summarized in Table 6.

Figure 1: Flow Diagram for Inclusion and Exclusion of Studies

Table 6: Details of the ZUMA-7 Study

ABC = activated B cell; AE = adverse event; ASCT = autologous stem cell transplant; axi-cel = axicabtagene ciloleucel; CAR = chimeric antigen receptor; CR = complete response; CRS = cytokine release syndrome; CNS = central nervous system; DLBCL = diffuse large B-cell lymphoma; DOR = duration of response; ECOG = Eastern Cooperative Oncology Group; EFS = event-free survival; EORTC QLQ-C30 = European Organisation for Research and Treatment of Cancer Quality of Life Questionnaire Core 30; EQ-5D-5L = 5-Level EQ-5D; GCB = germinal centre B cell; HDT = high-dose therapy; HGBL = high-grade B-cell lymphoma; HRQoL = health-related quality of life; LBCL = large B-cell lymphoma; ORR = objective response rate; OS = overall survival; PCR = polymerase chain reaction; PD = progressive disease; PFS = progression-free survival; PR = partial response; R-DHAP/X = rituximab plus dexamethasone, high-dose cytarabine, and cisplatin/oxaliplatin; R-ESHAP = rituximab plus etoposide, methylprednisolone, cytarabine, and cisplatin; R-GDP = rituximab plus gemcitabine, dexamethasone, and cisplatin/carboplatin; R-ICE = rituximab plus ifosfamide, carboplatin, and etoposide; SD = stable disease; SOC = standard of care; TBI = total body irradiation; VAS = visual analogue scale; WDAE = withdrawal due to adverse event.

Description of the ZUMA-7 Trial

ZUMA-7 is a phase III, multicentre, randomized, open-label trial evaluating the efficacy of axi-cel compared with SOC (salvage chemoimmunotherapy followed by HDT-ASCT) as a second-line therapy in patients with relapsed or refractory LBCL after first-line rituximab and anthracycline-based chemotherapy (Figure 2). The trial was conducted in 14 countries; 20 patients were recruited in Canada. The first patient was enrolled (randomized) on January 25, 2018, and enrolment was completed on October 4, 2019. The ZUMA-7 is ongoing at the time of this review. The data cut-off date for the primary analysis was March 18, 2021.

For patients in the axi-cel arm, treatment consisted of lymphodepleting chemotherapy followed by a single IV infusion of axi-cel. Bridging therapy of corticosteroids was allowed before lymphodepleting chemotherapy for patients with high disease burden, at the discretion of the investigator. For patients in the SOC arm, treatment consisted of a single protocol-defined, platinum-based salvage chemotherapy regimen as selected by the treating investigator. Patients who responded to salvage chemotherapy were to proceed to HDT with or without total body irradiation (TBI), followed by ASCT.

Randomization and treatment allocation: After all screening activities were completed, an Interactive Voice/Web (x) Response System (IxRS) was used to randomly assign patients in a 1:1 ratio to receive axi-cel or SOC. Randomization was stratified by response to first-line therapy (primary refractory, relapse ≤ 6 months of first-line therapy, or relapse > 6 and ≤ 12 months of first-line therapy) and second-line age-adjusted International Prognostic Index (sAAIPI) (low: 0 to 1; or high: 2 to 3), as assessed at the time of screening.

Blinding: This was an open-label trial; patients were aware of their assigned treatment group and all staff at each investigative site involved in treating and caring for study patients had full knowledge of the patient’s treatment assignment. An open-label design was necessary due to the difference in the nature of treatment in the 2 arms. Radiologic evaluations of response (i.e., for EFS, PFS, ORR, and DOR) were performed both by an unblinded investigator and by an independent central reviewer who was unaware of the patient’s treatment arm.

Study phases: The screening phase started on the date the patient signed the informed consent form and continued through confirmation of eligibility and randomization into the study. Screening procedures were completed within 14 days before enrolment (i.e., randomization). Informed consent was obtained before completion of any study-specific procedures that were not SOC. To establish a baseline and to confirm eligibility, screening fluorodeoxyglucose (FDG)-PET from skull base to mid-thighs, diagnostic quality contrast-enhanced (unless contraindicated) CT from skull base through lesser trochanters (PET-CT), and the appropriate imaging of all other sites of disease were required to be performed within 28 days before randomization and as close to randomization as possible. Patients who failed to meet the eligibility criteria were allowed to rescreen once. The assessment that initially resulted in the failed screening was repeated, including any other procedures that fell outside of the designated screening window (i.e., laboratory assessments or PET-CT scans).

The treatment phase started with the first dose of treatment following randomization. In the axi-cel arm, treatment day 0 was defined as the day the patient received the first axi-cel infusion. End of study will occur when all patients in the axi-cel arm have been followed up for 15 years in the axi-cel arm and 5 years in the SOC arm after randomization, have withdrawn consent, have been lost to follow-up, or have died.

Protocol Amendments

The original protocol, dated May 22, 2017, was amended 6 times in the US; for all other regions, the protocol was amended 5 times. No patients were treated until the second amendment. Important protocol amendments included the following: the primary EFS analysis event trigger was modified from 270 to approximately 250 EFS events, with an acceptable lower limit for the observed total EFS events of 225, to maintain the power for the primary analysis to within 5% of the targeted 90%; the duration of follow-up for the primary analysis of EFS was increased from the study day 150 assessment to the month 9 assessment; and a second interim OS analysis and a sensitivity analysis of OS were added. The second interim analysis of OS was to occur when approximately 160 deaths were observed or no later than 4 years after the first patient was randomized. The sensitivity analyses were added to address the confounding effect from treatment switching (date of analysis June 25, 2020).

Some changes were made to the planned analyses after the finalization of the Statistical Analysis Plan. The primary EFS analysis had been planned to occur after 250 EFS events were observed in the study. Because the time to reach 250 EFS events was longer than estimated, the first interim OS analysis was conducted at 153 events instead of the planned 110 events. As a result, the interim OS analysis conducted at 153 events meets the criteria for both originally planned interim OS analyses at 110 and 160 events. The only subsequent planned OS analysis will be the primary (final) OS analysis, expected to occur when 210 events are observed or no later than 5 years after the first patient was randomized. Other minor changes included modifications within some categories of baseline characteristics and subgroup covariates and clarification to definitions and reporting of some end points.

Figure 2: ZUMA-7 Study Treatment Schema

Auto-SCT = autologous stem cell transplant; CAR = chimeric antigen receptor; HDT = high-dose therapy; R-ESHAP = rituximab+etoposide-methylprednisolone-cytarabine-cisplatin; R-GDP = rituximab+gemcitabine-dexamethasone-cisplatin/carboplatin; R-DHAP = rituximab+dexamethasone-cytarabine-cisplatin; R-ICE = rituximab+ifosfamide-carboplatin-etoposide; SOC = standard of care; SOCT = standard of care therapy.

Notes: Study day = number of days from the day of randomization; treatment day = number of days from the day of axicabtagene ciloleucel treatment.

^a At the discretion of the investigator, corticosteroid bridging therapy could have been considered for patients with high disease burden at screening.

^b Minimum observation period: 7 days unless otherwise required by country regulatory agencies (e.g., 10 days for patients treated in Germany, Switzerland, and France).

^c Disease assessments were to be calculated from the date of randomization and not the date of dosing with axicabtagene ciloleucel or SOCT. Independent of the treatment arm, study procedures and disease assessments were to occur at the same protocol-defined time points.

Source: Clinical Study Report for Yescarta.¹⁵

Populations

Inclusion and Exclusion Criteria

Eligible patients were aged at least 18 years and had histologically confirmed LBCL, according to the WHO 2016 classification criteria,²⁷ that was refractory to first-line treatment or that had relapsed from complete remission. Refractory disease was defined as a lack of CR to first-line therapy and relapsed disease was defined as biopsy-proven disease relapse occurring no more than 12 months after the completion of first-line chemoimmunotherapy. Patients must have received adequate first-line therapy, including at a minimum an anti-CD20 monoclonal antibody (unless their tumour was CD20 negative) and an anthracycline-containing chemotherapy regimen, and must have intended to proceed to HDT-ASCT if there was a response to salvage chemotherapy. Patients with a history of malignancy other than nonmelanoma skin cancer or carcinoma in situ (unless disease-free for at least 3 years), those with a history of Richter transformation of chronic lymphocytic leukemia or primary mediastinal B-cell lymphoma (PMBCL) were ineligible. Patients with a history of autologous or allogeneic stem cell transplant, or those who received prior CD19 targeted therapy, CAR T-cell therapy or other genetically modified T-cell therapy, or systemic immunostimulatory agents within 6 weeks or 5 half-lives of the drug before the first dose of the study drug were also excluded (Table 6).

Baseline Characteristics

The mean age of patients was 57 years (SD = 12); 30% of the patients were aged 65 years or older. The majority of patients were male (66%) and White (83%). Most patients were randomized in North America (75%), of whom the majority were in the US (70%). Approximately half of the patients had an sAAIPI score of 0 or 1 (52% and 56% in the axi-cel and SOC arms, respectively). Overall, 74% of the study population had primary refractory disease and 26% had early relapse. Based on central laboratory assessment, patients in the axi-cel and SOC arms were categorized as DLBCL NOS/without further classification possible (70% and 67%, respectively); HGBL with MYC, BCL2, and/or BCL6 rearrangements (17% and 14%, respectively); or HGBL, NOS, (1% in the SOC arm). The remaining patients were categorized under not confirmed, missing, or other. Approximately one-quarter of patients in both treatment arms had achieved a best response of CR to first-line treatment (Table 7). Overall, baseline characteristics were generally comparable between the 2 treatment arms, but a difference of at least 10 percentage points was observed between axi-cel and SOC arms for sex (61% versus 71% male, respectively) and extranodal disease (57% versus 67%, respectively).

Table 7: Summary of Baseline Characteristics — ZUMA-7 Study

axi-cel = axicabtagene ciloleucel; ECOG = Eastern Cooperative Oncology Group; IHC = immunohistochemistry; SD = standard deviation; SOC = standard of care.

^aThe definition of diffuse large B-cell lymphoma per central laboratory assessment included cases of incomplete evaluation due to inadequate sample amount or type for which further classification of the subtype was not possible.

^bCD19 IHC-positive is defined as the H-score of staining ≥ 5.

^cMissing CD19 H-scores are mainly due to quantity not sufficient, biopsy not available at central laboratory, CD19 negative, or tumour tissue not present in sample.

^dBone marrow involvement is collected on the diagnosis history case report form.

^eAn elevated lactate dehydrogenase level was defined as a level that was above the upper limit of the normal range according to the local laboratory.

^fTumour burden was determined on the basis of the sum of product diameters of the target lesions, according to the Cheson criteria, and was assessed by the central laboratory.

Sources: Clinical Study Report for Yescarta¹⁵ and Locke et al. (2022).¹⁴

Interventions

Axi-Cel Arm

Bridging Therapy

Administration of bridging therapy was optional. Bridging therapy consisted of corticosteroids (e.g., dexamethasone at a dose of 20 mg to 40 mg or equivalent, either orally [PO] or IV daily for 1 to 4 days) administered after leukapheresis through 5 days before administration of axi-cel. Bridging therapy was sourced by the investigative site. Choice of corticosteroid and dosing could be adjusted for age, comorbidities, or per clinical judgment.

Lymphodepleting Chemotherapy

Lymphodepleting chemotherapy (conditioning chemotherapy) consisted of fludarabine and cyclophosphamide used for lymphodepletion before administration of axi-cel. Five days before the planned axi-cel infusion date, patients were treated with lymphodepleting chemotherapy. The 3-day lymphodepleting chemotherapy regimen consisted of fludarabine 30 mg/m² per day and cyclophosphamide 500 mg/m² per day on treatment days −5 to −3 followed by 2 rest days before axi-cel infusion on treatment day 0.

Axi-Cel

Any screening assessments or procedures that were repeated between confirmation of eligibility and the start of axi-cel infusion that had results outside the eligibility criteria were resolved before proceeding with the infusion (except for peripheral blood cell counts that had been impacted by lymphodepleting chemotherapy). The patient must not have shown any evidence or suspicion of an infection. If the axi-cel infusion was delayed for more than 2 weeks, protocol guidelines regarding the need for repeat lymphodepleting chemotherapy were to be followed.

Approximately 1 hour before the axi-cel infusion, the preinfusion medications acetaminophen (650 mg PO or equivalent) and diphenhydramine (12.5 mg PO or IV or equivalent) were administered. Axi-cel was administered intravenously as a single infusion of CAR-transduced autologous T cells. Axi-cel was supplied cryopreserved in cryostorage bags and the product was required to remain frozen until the patient was ready for treatment to assure viable live autologous T cells were administered. Axi-cel was administered at a target dose of 2 × 10⁶ anti-CD19 CAR T cells/kg body weight but may have been dosed at a minimum of 1 × 10⁶ anti-CD19 CAR T cells/kg. For patients weighing more than 100 kg, a maximum flat dose of axi-cel at 2 × 10⁸ anti-CD19 CAR T cells was administered.

All patients received axi-cel infusion at a health care facility, followed by daily monitoring at a health care facility for at least 7 days to monitor for signs and symptoms of CRS and neurologic events. Alternatively, if deemed appropriate by the investigator, patients may have been hospitalized to receive their infusion and were observed for CRS and neurologic events in the hospital setting.

Re-Treatment

Axi-cel re-treatment was allowed based on experience from the ZUMA-1 study, in which 6 of 10 patients who were re-treated subsequently achieved a PR or CR.⁴³ Patients in ZUMA-7 who achieved a PR or CR at the study day 50 disease assessment and subsequently experienced disease progression had the option to receive a second course of lymphodepleting chemotherapy and axi-cel if they met the following criteria: their disease was not known to be CD19 negative; they continued to meet the study eligibility criteria; they had no life-threatening toxicity related to axi-cel during the original course of treatment and any toxicities related to lymphodepleting chemotherapy (except for alopecia) had resolved to grade 1 or lower or baseline before re-treatment; they had not received subsequent therapy for lymphoma; and they did not have known axi-cel neutralizing antibodies. A maximum of 1 re-treatment per patient was permitted. Sites were strongly encouraged to collect and submit a biopsy to the central laboratory before initiating re-treatment for confirmation of disease progression and CD19 expression. The second dose of axi-cel could have been manufactured at the same time as the first axi-cel dose with existing cryopreserved peripheral blood mononuclear cells (PBMCs), or the patient could have undergone a second leukapheresis to collect T cells for manufacture of the second dose of axi-cel.

SOC Arm

Salvage Chemotherapy

Patients randomized to the SOC arm received a salvage chemotherapy regimen, and those who responded were to proceed to HDT-ASCT. The SOC treatment was sourced by and administered at the investigative site in a setting with emergency medical facilities. Multiple salvage chemotherapy regimens were allowed; dosing was recommended but not required due to regional and institutional differences. It was anticipated that patients who did not respond to salvage chemotherapy would receive subsequent, off-protocol treatment that could include commercially available CAR T-cell therapies.

Salvage chemotherapy consisted of 1 of the following platinum-based second-line combination chemotherapy regimens and recommended dosing:

• R-ICE

  • rituximab: 375 mg/m² before chemotherapy

  • ifosfamide: 5 g/m² 24-hour continuous infusion on cycle day 2 with mesna

  • carboplatin: area under the curve (AUC) 5 on cycle day 2, maximum dose 800 mg

  • etoposide: 100 mg/m² per day on cycle days 1 to 3

• R-ESHAP:

  • rituximab: 375 mg/m² on cycle day 1

  • etoposide: 40 mg/m²/day IV on cycle days 1 to 4

  • methylprednisolone: 500 mg/day IV, on cycle days 1 to 4 or 5

  • cytarabine: 2 g/m² on cycle day 5

  • cisplatin: 25 mg/m²/day continuous infusion, on cycle days 1 to 4

• R-GDP:

  • rituximab: 375 mg/m² on cycle day 1 (or day 8)

  • gemcitabine: 1 g/m² on cycle days 1 and 8

  • dexamethasone: 40 mg on cycle days 1 to 4

  • cisplatin: 75 mg/m² on cycle day 1 (or carboplatin, AUC = 5)

• R-DHAP or R-DHAX:

  • rituximab: 375 mg/m² before chemotherapy

  • dexamethasone: 40 mg/d on cycle days 1 to 4

  • high-dose cytarabine: 2 g/m² every 12 hours for 2 doses on cycle day 2 following platinum treatment

  • cisplatin: 100 mg/m² 24-hour continuous infusion on cycle day 1 (or oxaliplatin 100 mg/m²)

Salvage chemotherapy was recommended to start within approximately 5 days after randomization. Patients received 2 or 3 cycles of a single permitted salvage chemotherapy regimen, with 1 cycle administered every 2 to 3 weeks. After salvage chemotherapy was administered, the patient received the same chemotherapy regimen for subsequent cycles. Dose modifications due to toxicity were allowed, but a change of regimen was noted as a new lymphoma therapy.

HDT and ASCT

If a patient was deemed transplant eligible (i.e., demonstrated adequate disease response with CR or PR after 2 or 3 cycles of salvage chemotherapy and collected a sufficient number of CD34+ stem cells), HDT and ASCT could be initiated. Given the lack of a standard high-dose regimen, HDT and ASCT procedures were performed according to regional and institutional standards. Before HDT, granulocyte colony-stimulating factor (G-CSF) was administered to mobilize stem cells from the bone marrow to the periphery, after which peripheral blood progenitor cells were collected by leukapheresis to a minimum target of 2 × 10⁶ CD34+ hematopoietic stem cells/kg body weight.

The HDT conditioning regimen consisted of combination high-dose chemotherapy with or without TBI. Commonly used high-dose regimens include BEAM (carmustine, etoposide, ara-C, and melphalan) or cyclophosphamide-BCNU-etoposide (CBV). After HDT, the CD34+ hematopoietic stems cells were reinfused to rescue hematopoiesis.

Discontinuation

Patients could be removed from protocol-required investigational treatments or procedures due to AEs, disease progression, patient request or noncompliance, unavailability of product, loss to follow-up, death, or decision by sponsor.

Subsequent Therapy

Subsequent therapy that was administered after the protocol-specified study treatment (axi-cel or SOC) and was necessary to treat a patient’s PD, such as chemotherapy not specified in the study, immunotherapy, targeted agents, as well as allo-SCT or ASCT and radiation therapy, were to be recorded for all patients randomized until 1 of the following occurred: the patient completed the long-term follow-up period, was considered lost to follow-up, withdrew full consent, or died.

Prior and Concomitant Therapy

Investigators were allowed to prescribe any concomitant medications or treatment deemed necessary to provide adequate supportive care, including growth factor support (e.g., G-CSF) and routine antiemetic prophylaxis, with the following exceptions:

• After treatment with axi-cel or SOC, lymphoma treatment such as chemotherapy, immunotherapy, targeted agents, radiation (TBI for HDT was allowed for the SOC arm), high-dose corticosteroids (other than those allowed in the protocol for either arm), and other investigational agents were prohibited, except as needed for the treatment of PD.

• For patients in the axi-cel arm, corticosteroid therapy at a pharmacologic dose (≥ 5 mg/day of prednisone or equivalent doses of other corticosteroids) and other immunosuppressive drugs were to be avoided for 7 days before leukapheresis and 5 days before axi-cel administration.

• Patients for whom CT scans with contrast are contraindicated (i.e., patients with contrast allergy or impaired renal clearance) were to undergo an MRI with contrast and noncontrast CT scans instead of premedicating with systemic corticosteroids.

• For 3 months after axi-cel administration, corticosteroids and other immunosuppressive drugs were to be avoided unless used to manage axi-cel-related toxicities. Other medications that might have interfered with the evaluation of axi-cel, such as nonsteroidal anti-inflammatory agents, were also to be avoided for the same period unless medically necessary.

Outcomes

The efficacy end points identified in the CADTH review protocol that were assessed in the clinical trial included in this review are summarized below.

Efficacy

The primary end point was EFS according to blinded central assessment. The key secondary efficacy end points were ORR according to blinded central assessment, and OS. Other secondary end points included PFS based on investigator assessment, DOR according to blinded central assessment, and HRQoL. Time to next treatment was an exploratory end point. Health resource utilization was not measured as an efficacy end point but was reported supplementary to the assessment of safety.

Throughout the study, patients in both treatment groups were assessed for response and progression at the same times relative to randomization (study day 0), including study days 50, 100, and 150, month 9, and every 3 months thereafter until month 24, and then every 6 months from months 30 to 60. Patients were to have their first posttreatment planned PET-CT tumour assessment within the study day 50 assessment period (range, −7 to + 21 days, from study day 50). The patient’s disease was to be assessed at regular intervals through month 9 after randomization or until a change in lymphoma therapy or disease progression, whichever came first. If the patient’s disease had not progressed by month 9, disease assessments were to be evaluated by CT scans where a CR was suspected and by PET-CT scans where a PR was suspected. Patients with symptoms suggestive of disease progression were to be evaluated for progression at the time symptoms occurred. PET-CT scan could be performed at any time disease progression was suspected.

The fluorodeoxyglucose (FDG)-PET assessment was to take precedence over the CT assessment for the time points when both were available. In addition to the investigator’s assessment, PET-CT scans of all patients evaluated for disease response were to be submitted to and reviewed by an independent central reviewer who was blinded to the patient’s treatment arm. If PD was reported by the investigator in the absence of radiologic evidence, objective evidence of clinical progression (e.g., tumour tissue or bone marrow biopsy results, physical examination findings, lactate dehydrogenase elevation, weight loss, or fever) were to be submitted to and reviewed by an independent central reviewer. For patients who discontinued protocol therapy due to an assessment of PD but for whom there was no change in lymphoma therapy, any additional imaging data subsequent to the image in question were to be submitted to the central reviewer to confirm disease response. If the patient was eligible for re-treatment with axi-cel, the last scan before re-treatment was to be considered the baseline for the purpose of evaluating the response to re-treatment.

The primary end point of EFS was defined as the time from randomization to the earliest date of disease progression per the Lugano Classification,⁴⁴ as determined by blinded central assessment; commencement of new lymphoma therapy (including stem cell transfer in the axi-cel arm without axi-cel–induced response or re-treatment of axi-cel); or death from any cause. Patients with established PR, CR, or stable disease who subsequently began a new lymphoma therapy in the absence of documented PD were to have an EFS event time defined as the time from randomization to the last evaluable disease assessment (or last time stable disease was established) before commencing the new therapy. For patients who started a new lymphoma therapy in the absence of an evaluable disease assessment, the EFS date was imputed as the randomization date. Patients with a best response of stable disease up to and including study day 150 were considered to have an EFS event, defined as the time from randomization to the time of the first establishment of stable disease.

The following criteria were used to define censoring times for EFS: (1) patients alive, in response, and with no new therapy were censored at the last evaluable disease assessment; (2) patients with no evaluable disease assessment by study day 150 were not considered to have an EFS event and event time was censored at the time of randomization; (3) for patients in the axi-cel arm who underwent ASCT without documented progression or new lymphoma therapy were censored at the time of ASCT; (4) for patients in the SOC arm, TBI, HDT, and ASCT that occurred while in response were not considered an EFS event; (5) at the time of the interim EFS analysis, patients who had not been followed for 150 days and who did not have an EFS event were censored at the last evaluable disease assessment.

ORR was defined as the incidence of either a CR or PR by the Lugano Classification⁴⁴ as defined by blinded central assessment. Patients who did not meet the criteria for an objective response by the analysis cut-off date were considered as nonresponders.

OS was defined as the time from randomization to death from any cause. Patients who did not die by the analysis data cut-off date had survival time censored at their last date known to be alive. For patients alive or dead after the data cut-off date, survival time was to be censored at the data cut-off date.

PFS was defined as the time from randomization to disease progression per the Lugano Classification⁴⁴ as determined by the investigator assessment or death from any cause. The following criteria were used to define censoring times for PFS: (1) patients who were alive and did not meet the criteria for progression at the analysis cut-off date had PFS censored a the last evaluable disease assessment, (2) patients who received subsequent lymphoma therapy (except HDT, TBI for HDT, and ASCT while in protocol therapy-induced response) in the absence of documented PD were censored at their last evaluable disease assessment date before starting the subsequent therapy, (3) the receipt of ASCT and allo-SCT while a patient was in response was not considered a PFS event and these patients were censored at the last evaluable disease assessment before ASCT or allo-SCT.

DOR was defined as the time from first response to disease progression per the Lugano Classification⁴⁴ or death from any cause. DOR was derived only in patients who experienced an objective response per the Lugano Classification as determined by blinded central assessment. The following criteria were used to define censoring times for DOR: (1) patients not meeting criteria for progression or death by the analysis cut-off date were censored at the last evaluable disease assessment date; (2) patients who received subsequent new lymphoma therapy (except HDT, TBI for HDT, or ASCT while in protocol therapy-induced response) in the absence of documented PD were censored at the last evaluable disease assessment before starting the new therapy; (3) for the primary DOR analysis, DOR was censored at the last evaluable disease assessment date before ASCT or allo-SCT for patients undergoing ASCT or allo-SCT while in protocol-defined therapy response.

Time to next treatment was defined as the time from the randomization date to the start of the subsequent new lymphoma therapy (including re-treatment or subsequent stem cell transfer for patients in the axi-cel arm) or death from any cause. Patients who had not received subsequent new lymphoma therapy and were still alive were to be censored at the last contact date.

Patient-reported outcomes included HRQoL assessed using the EORTC QLQ-C30 version 3.0 and the EQ-5D-5L questionnaires. The EORTC QLQ-C30 is a 30-item cancer-specific questionnaire designed to provide a multidimensional assessment of HRQoL with a recall period of 1 week. The EORTC QLQ-C30 version 3.0 includes the following scales: 5 multi-item functional scales (assessing physical, role, emotional, cognitive, and social functioning) and 3 symptom scales (fatigue, nausea and vomiting, and pain); 6 single-item symptom scales (dyspnea, insomnia, appetite loss, constipation, diarrhea, and financial difficulties); and 2 global scales (global health status and HRQoL). No items are shared between the scales. The first 28 of 30 items have 4-level ordinal responses from 1 (“not at all”) to 4 (“very much”). The final 2 items, comprising the global health status and HRQoL scale, have 7-level ordinal responses from 1 (“very poor”) to 7 (“excellent”). Each scale is measured from 0 to 100 after a linear transformation. Higher scores for functioning scales and for the global health status or quality of life scales indicate a higher level of functioning and a better HRQoL, respectively, whereas higher scores in symptom scales represent a higher level of symptoms. The minimal important difference (MID) of the QLQ-C30 scores was 10 points. This MID was based on a study of patients with lung cancer.¹⁶

The EQ-5D-5L questionnaire is a generic and preference-weighted measure of health status captured on the day of assessment. Two components comprise the EQ-5D-5L: the EQ-5D descriptive system and the EQ VAS. The descriptive system has 5 dimensions (mobility, self-care, usual activities, pain/discomfort, and anxiety/depression) and is divided into 5 levels of severity: no problems, slight problems, moderate problems, severe problems, and extreme problems. The EQ VAS is a vertical visual analogue scale for recording self-rated HRQoL state reported from 0, described as “the worst health you can imagine,” to 100, described as “the best health you can imagine.” The MID was defined as a change of 0.06 points in the EQ-5D-5L index and a change of 7 points from screening in the EQ VAS score. This MID was based on a study of patients with general cancer.⁴⁵

The HRQoL assessments were conducted at screening (within 14 days before randomization), 5 days after randomization, at study day 50, treatment period (day of transplant), and during posttreatment follow-up, that is, study day 100 and study day 150 (± 14 days).

Safety

Safety assessments included monitoring of AEs and concomitant medications, clinical laboratory analyses (serum chemistries, complete blood count with differential, alanine transaminase, aspartate transaminase, total bilirubin, lactate dehydrogenase, C-reactive protein, ferritin, urine/serum pregnancy tests, and viral serologic tests [for European sites only]), vital sign measurements, physical examinations, ECOG PS, neurologic examinations, cardiac function (echocardiography and electrocardiography), and testing for replication-competent virus in the blood, as well as for antibodies against the axi-cel CAR in the serum.

The severity of AEs and SAEs were graded using the National Cancer Institute’s Common Terminology Criteria for Adverse Events (CTCAE) version 4.03.⁴⁶ The investigator was to assess and record whether the AE or SAE was possibly related to axi-cel, leukapheresis, or lymphodepleting chemotherapy (in the axi-cel arm); salvage chemotherapy, HDT, CD34+ leukapheresis, or CD34+ infusion (in the SOC arm); disease progression; concurrent disease; concomitant medication; or other possible reasons.

An AE was defined as any untoward medical occurrence in a clinical study experienced by a participant; the event did not necessarily have a relationship to the study treatment. Investigators were responsible for ensuring that any AEs observed by the investigator or reported by the patient were recorded in the patient’s medical record. The definition of an AE included the worsening of a pre-existing medical condition. Worsening indicated that the pre-existing medical condition had increased in severity, frequency, and/or duration or was associated with a worse outcome. When recording such events, descriptions that the pre-existing condition had changed (e.g., more frequent headaches for a patient with pre-existing headaches or increased blood pressure in a patient with pre-existing hypertension) were to be provided. A pre-existing condition that had not worsened during the study or involved an intervention, such as elective cosmetic surgery or a medical procedure while on study, was not to be considered an AE. Hospitalization for study treatment infusions or precautionary measures per institutional policy were not to be considered AEs. TEAEs were defined as any AE with onset on or after the axi-cel infusion for the axi-cel arm, and as any AE with onset on or after the first dose of salvage chemotherapy for the SOC arm.

An SAE was defined as an event that met at least 1 of the following: fatal; life-threatening; required inpatient hospitalization or prolongation of planned hospitalization; resulted in persistent or significant disability or incapacity; resulted in congenital anomaly or birth defect. Other medically important serious events could be classified as an SAE if an investigator considered them to be clinically important, but they did not meet any of the serious criteria.

AEs of special interest for axi-cel treatment included important identified risks (CRS, neurologic events [including cerebral edema], cytopenias [thrombocytopenia, neutropenia, and anemia], infections, and hypogammaglobulinemia) and important potential risks (secondary malignancies, immunogenicity, replication-competent virus, tumour lysis syndrome, and aggravation of graft-versus host disease). CRS is an identified risk for axi-cel but not for SOC; as such, CRS as a syndrome was only recorded for axi-cel. The term “disease progression” as assessed by measurement of malignant lesions on radiographs or other methods was not to be considered an AE. However, for situations when an AE or SAE was due to the disease under investigation, the signs and symptoms or worsening of signs and symptoms of the malignancy under study were considered to be AEs.

Investigators were to report all AEs observed by the investigator or reported by the patient and that occurred from enrolment (i.e., randomization) through the study day 150 visit or until a change in the patient’s lymphoma therapy, whichever occurred first. All SAEs that occurred from the date of the screening through the study day 150 visit or until initiation of new lymphoma therapy (whichever occurred first) were to be reported. Targeted SAEs (neurologic events, hematologic events, infections, autoimmune disorders, and secondary malignancies) were to be reported for up to 5 years for the SOC arm, 15 years for the axi-cel treatment arm, or until disease progression, whichever occurred first. SAEs assessed by the investigator as related to axi-cel were to be reported regardless of the time period. Patients who did not meet the screening criteria or who were enrolled but not dosed with axi-cel were to be followed for AEs and SAEs for 30 days after the last study-specific procedure (e.g., screening procedure, leukapheresis, or lymphodepleting chemotherapy).

Deaths were to be reported as SAEs if they occurred after randomization and before the study day 150 visit or until the initiation of new lymphoma therapy (whichever occurred first), regardless of attribution to treatment; the event or condition leading to death was to be recorded on the AE form. The term “disease progression” as assessed by measurement of malignant lesions on radiographs or other methods was not to be reported as an AE. Deaths attributed solely to progression of the underlying lymphoma and in the absence of signs and symptoms were to be recorded as SAEs and on the AE electronic case report form with the preferred term “B-cell lymphoma.” Deaths that occurred after the study day 150 visit during the long-term follow-up period or after initiation of new lymphoma therapy that were due to the underlying malignancy were to be recorded on the Survival Status electronic case report form and Death Summary Page.

Statistical Analysis

Sample Size Calculations

The anticipated enrolment in this study was approximately 350 patients. The primary analysis was planned to occur when all patients had had the opportunity to be followed for the month 9 disease assessment (i.e., the month 9 time point had passed for all patients) and 250 EFS events by blinded central assessment had been observed; the study was sized to achieve approximately 90% power at the 1-sided 2.5% significance level to detect a 50% improvement in EFS (the hypothesis for the study). The minimum effect size that could be determined to be statistically significant was an EFS HR of 0.79, or a 27% relative improvement in EFS. Assuming a concave accrual distribution with 50% of accrual in the last 33% of the accrual period of 24 months and a 10% rate of loss to follow-up (5% by month 1 and cumulative 10% by month 8) in the axi-cel arm and 15% rate of loss to follow-up (10% by month 1 and cumulative 15% by month 8) in the SOC arm, it was anticipated that the event goal would be achieved if 350 patients were randomized (175 patients per arm) and would occur approximately 31 months after the first patient was randomized.

Analyses of Outcomes

For the primary end point of EFS according to blinded central assessment, a stratified log-rank test was used for the primary comparison of EFS. Stratified Cox regression models were used to provide the estimated EFS HR and 2-sided 95% CIs for axi-cel relative to SOC. IxRS values for response to first-line therapy and sAAIPI were used as the stratification factors for the log-rank test and Cox regression model. The median EFS event time and event-free rates at 3-month intervals were calculated. The Breslow method and the Efron method were used to handle the ties for the Cox regression models. Nonproportionality among the treatment groups was assessed by comparing the standardized martingale residuals over time to a normal distribution at the 5% level. If the comparison of the standardized martingale residuals over time was significant, a piece-wise Cox model was used for the analysis. For the stratified piece-wise Cox model, 2 or more equal length intervals of 12 weeks was considered. This was to include 1 scheduled tumour assessment in each interval. These models allowed estimation of the overall as well as within-interval treatment HR. The overall log HR was calculated from the weighted average of the interval log HRs. The weight for each interval log HR estimate was to be inversely proportional to the variance of the interval log HR estimate. If there were intervals where 1 or more strata was too sparse, the stratified piece-wise Cox model was to be collapsed as specified for the primary analysis. Kaplan-Meier estimates and 2-sided 95% CIs were calculated for event-time quartiles, for event-free rates at 3-month intervals, and for the difference in event-free rates between treatment arms at these times.

For the analysis of ORR based on blinded central assessment, the incidence of objective response and best response was summarized. An exact binomial 2-sided 95% CI was generated for the ORR and best response rates for each treatment arm. The Wilson score method with continuity correction was used to calculate 95% CIs for the difference in ORRs between treatment arms.

The analysis of OS was conducted using the same methods as for the analysis of EFS. A stratified log-rank test was used for the primary comparison of OS. In addition, stratified Cox regression models were used to provide the estimated OS HR and 2-sided 95% CIs for the axi-cel arm relative to the SOC arm.

The analyses of PFS, DOR, and time to next treatment used the same methods as the analysis of EFS, with P values from the log-rank test descriptive.

Changes in the EORTC QLQ-C30 domains, EQ-5D-5L index, and EQ-5D-5L VAS scores were summarized with descriptive statistics. Each of the 3 HRQoL scores was examined via a series of mixed models for repeated measures (MMRM) for the individual change in HRQoL scores over time and systematic differences between treatment arms. The base MMRM was fit to each patient-reported outcome score with terms for treatment arm, time point (treated as discrete), and treatment by time interaction. An unstructured correlation structure was assumed for measurements collected across visits. In case the estimation procedure did not converge, the compound symmetry variance structure was considered. The fit of the model to the data was assessed via Akaike information criterion (AIC) and Bayesian information criterion (BIC). Using MMRM, planned comparisons were conducted between treatment arms at predetermined specific time points of interest (study days 100 and 150, and months 9, 12, 15, 18, 21, and 24). The MMRM was considered sufficient for handling missing data under the missing-at-random assumption. The base MMRM was run with variables for treatment, time, and treatment time by interaction (primary analysis) and controlling for Interactive voice/web (x) response IxRS values of response to first-line therapy and age-adjusted International Prognostic Index.

Descriptive statistics were used to summarize safety data. A summary of the patient incidence of key AEs (any, worst severity, serious, related, CRS, neurologic event) was provided. The incidence of AEs, SAEs, and grade 3 or higher AEs was provided for all TEAEs, all AEs related to lymphodepleting chemotherapy (axi-cel arm only), all TEAEs related to axi-cel (axi-cel arm only), all TEAEs related to salvage chemotherapy, HDT, or ASCT (SOC arm only), fatal AEs, and TEAEs of interest (including identified and potential risks). Summary statistics for the onset time, duration, and resolution of TEAEs of interest were reported.

Statistical Considerations for Multiple Comparisons

The study was planned to have an overall alpha (significance level) of 2.5% with 1-sided testing. To preserve the overall significance level, statistical testing of the primary and key secondary efficacy end points was to follow a hierarchical scheme: first, EFS was to be tested at the primary analysis. An EFS HR (test versus control arm) of 0.67 was hypothesized. Assuming an exponential distribution for EFS and a median EFS of 4 months in the SOC arm, this implied a 50% relative improvement in EFS, corresponding to a median EFS of 4 months versus 6 months (control arm versus test arm, respectively). A log-rank test stratified by randomization factors was to be used to test the null hypothesis of no difference in EFS using an overall 1-sided alpha level of 2.5%.

Conditional on a statistically significant improvement in EFS, ORR was to be tested at the 1-sided alpha level of 2.5% at the time of the primary EFS analysis. ORR was to be tested with a stratified Cochran-Mantel-Haenszel test using randomization factors. Conditional on a statistically significant improvement in EFS and ORR, OS was to be tested up to 3 times at an overall alpha level of 2.5%. The primary analysis of OS is to occur when approximately 210 deaths have been observed or no later than 5 years after the first patient was randomized. A first interim analysis of OS was to occur at the time of the primary EFS analysis and a second interim analysis when approximately 160 deaths have been observed or no later than 4 years after the first patient is randomized. A spending function of the rho family was to be used to allocate the alpha between the 2 interim analyses of OS and the primary analysis of OS. Log-rank tests stratified by randomization factors were to be used to test the null hypothesis of no difference in OS. If a statistically significant improvement in EFS was not demonstrated at the time of the primary EFS analysis, hierarchical testing of ORR and OS was not to occur. If a statistically significant improvement in EFS was demonstrated, but a statistically significant improvement in ORR was not demonstrated at the time of the primary EFS analysis, hierarchical testing of OS was not to occur.

One primary analysis of EFS was planned and was to be performed after 250 EFS events as determined by blinded central assessment had been observed and all patients had had the opportunity to be followed for 9 months (i.e., the month 9 disease assessment time point for all patients had passed at the time of the primary analysis, which allowed patients to be included in the analysis if they did not have their month 9 visit [e.g., due to death, disease progression, loss to follow-up, consent withdrawal, or the COVID-19 pandemic]). The primary analysis of ORR was to occur at the same time. The primary analysis for OS was planned to be performed after approximately 210 deaths had occurred or no later than 5 years after the first patient was randomized.

Interim Analyses

The study planned for 1 interim and 1 primary analysis for EFS. The interim analysis was to be a safety and futility analysis of EFS after 135 EFS events (by blinded central assessment) had occurred. A total of 135 EFS events occurred approximately 19 months after the first patient was randomized. Two interim analyses of OS were planned. A first interim analysis of OS was to occur at the time of the primary EFS analysis by when approximately 110 deaths were anticipated to occur, conditional upon statistically significant tests of EFS and ORR. A second interim analysis of OS was to occur when approximately 160 deaths had occurred or no later than 4 years after the first patient was randomized.

To allocate the type II error between the interim and primary EFS analyses, an O’Brien-Fleming spending function of the Lan-DeMets family was to be used. Under the null hypothesis, the probability of stopping for futility at the interim analysis was approximately 60%.

To allocate the alpha between the 2 interim analyses of OS and the primary analysis of OS, a spending function of the rho family with parameter (rho = 6) was to be used. The 1-sided alpha of 0.1% and 0.4% was to be allocated at the first and second interim analysis of OS, respectively.

Sensitivity Analyses

Sensitivity analyses of EFS were performed to assess ascertainment time bias in disease progression. An additional sensitivity analyses of EFS, PFS, DOR were performed in which patients in the axi-cel arm who underwent ASCT while in an axi-cel-induced response were imputed to have an event at the time the ASCT was performed.

Sensitivity analyses of OS were performed to address the confounding effects of treatment switching in the SOC arm (i.e., patients randomized to the SOC arm who did not respond or relapsed after SOC and subsequently received commercial or investigational cell therapy) using the rank-preserving structural failure time (RPSFT) model with g-estimation⁴⁷ and inverse probability of censoring weights (IPCW) adjustment methods.⁴⁸

Subgroup Analyses

The consistency of treatment effect on the EFS HR for axi-cel versus SOC was examined by estimating Cox model HRs within preplanned subgroups according to age, sex, ECOG PS, geographic region, race or ethnicity, response to first-line therapy, sAAIPI, prognostic marker (HGBL, double- or triple-hit and double expressor lymphoma), molecular subgroup (germinal centre B cell–like, activated B cell–like, unclassified), and disease type according to investigator and according to the central laboratory. ECOG PS score and response to first-line therapy were subgroups of interest identified in the CADTH systematic review protocol. A similar statistical analysis approach to the main EFS analysis was used. There was no multiplicity control. As such, all subgroup analyses are exploratory in nature.

Analysis Populations

The full analysis set (intention-to-treat population; N = 359) includes all randomized patients. Patients were analyzed by the protocol therapy to which they were randomized. The primary analysis of all efficacy end points, unless noted otherwise, was an intention-to-treat analysis conducted using the full analysis set. The safety analysis set (N = 338) was the subset of all randomized patients who received at least 1 dose of axi-cel as protocol therapy or SOC salvage chemotherapy as protocol therapy. The safety analysis set–ASCT (N = 62) comprised the subset of patients randomized to the SOC arm who underwent transplant as part of protocol therapy. The QoL analysis set (N = 296) comprised the subset of patients in the full analysis set who had a baseline and at least 1 completed postrandomization measurement through study day 150. The re-treatment analysis set (N = 9) comprised the subset of patients treated with axi-cel as study treatment who received any dose of axi-cel as re-treatment.

Results

Patient Disposition

Of the 437 patients screened, 359 (82%) patients met the inclusion criteria and underwent randomization (axi-cel arm: n = 180; SOC arm: n = 179). As of the data cut-off date of March 18, 2021, the median follow-up from randomization to the data cut-off date was 24.9 months (interquartile range [IQR], 21.0 to 28.7 months).

In the axi-cel arm, 178 of 180 randomized patients (99%) underwent leukapheresis. Of the 2 patients (1%) who did not undergo leukapheresis, 1 patient had rapid disease progression that required immediate antilymphoma therapy before the planned study treatment, and 1 patient was found to be ineligible after randomization due to cardiac lymphoma involvement. A total of 172 patients (96%) received lymphodepleting chemotherapy and 170 patients (94%) subsequently received axi-cel. Of the 178 patients who underwent leukapheresis, 6 patients (3%) received neither lymphodepleting chemotherapy nor axi-cel and 2 patients (1%) received lymphodepleting chemotherapy but not axi-cel due to AEs, disease progression or death (Table 8 and Figure 3). In the axi-cel arm, all 170 patients (100%) who received a single dose of axi-cel completed treatment. After axi-cel treatment, 9 patients who had had a response and later progressed were re-treated with axi-cel.

In the SOC arm, 168 of 179 randomized patients (94%) received at least 1 dose of salvage chemotherapy. In the SOC arm, completion of treatment was defined as completion of 2 or 3 cycles of salvage chemotherapy and either stable disease at study day 50 or PR or CR at study day 50 and completion of HDT-ASCT. Treatment was reported to be completed for 89 (53%) of the 168 patients who received at least 1 dose of salvage chemotherapy. Treatment was not completed for the remaining 79 (47%) patients for the following reasons: disease progression (71 patients), AEs (2 patients: 1 patient with grade 4 acute kidney injury related to SOC, and 1 patient with blood stem cell harvest failure), and other reasons (6 patients: 3 patients could not continue the initial salvage chemotherapy regimen, 1 patient was initiated on an alternative protocol, 1 patient achieved PR but the investigator deemed the response insufficient to proceed to ASCT, and 1 patient did not have a complete metabolic response and could not receive ASCT per institutional guidelines) (Table 8 and Figure 3).

According to the investigator assessment, 80 patients (45%) responded to salvage chemotherapy, of whom 69 patients (86%) underwent leukapheresis of CD34+ cells for ASCT. The other 11 patients (14%) who responded did not undergo leukapheresis of CD34+ cells for ASCT due to PD after the initial response (9 patients), an AE (1 patient; blood stem cell harvest failure), and investigator assessment of insufficient response to proceed to ASCT (1 patient). Another 5 patients underwent leukapheresis of CD34+ cells for ASCT but did not respond to salvage chemotherapy and could not proceed with ASCT. Thus, a total of 74 patients (41%) in the SOC arm underwent leukapheresis, of whom 69 patients responded, 64 patients (36%) received HDT and reached HDT-ASCT, and 62 patients (35%) subsequently received CD34+ stem cell rescue (ASCT) on-protocol. An additional 2 patients received CD34+ stem cell rescue off-protocol (1 patient was inadvertently started on an alternative protocol and 1 patient had disease progression 1 day after HDT was started and later received off-protocol HDT-ASCT).

At the data cut-off date, 66 patients (39%) in the axi-cel arm who received axi-cel and 86 patients (51%) in the SOC arm who received at least 1 dose of salvage chemotherapy had discontinued participation in the study. Reasons for discontinuation after receiving study therapy were death (64 patients in the axi-cel arm and 75 patients in the SOC arm), loss to follow-up (2 patients in each arm), withdrawal of consent (7 patients in the SOC arm), investigator decision (1 patient in the SOC arm), and other (1 patient in the SOC arm). In addition, 8 patients in the axi-cel arm who did not receive axi-cel were discontinued from the study due to death, and 7 patients in the SOC arm were discontinued from the study due to death (1 patient), full consent withdrawal (5 patients), or loss to follow-up (1 patient).

Table 8: Patient Disposition — ZUMA-7 Study (Full Analysis Set)

axi-cel = axicabtagene ciloleucel; NA = not applicable; SOC = standard of care.

Notes: In the SOC arm, patients who completed treatment are those who completed salvage chemotherapy (2 or 3 cycles), high-dose therapy, and stem cell transplant, or those who completed salvage chemotherapy and were assessed as stable disease at study day 50 without proceeding to stem cell transplant.

^aFor 2 patients, the study day 50 disease assessments were updated by the clinical site to stable disease, but the corresponding end-of-treatment electronic case report form was not updated.

Source: Clinical Study Report for Yescarta.¹⁵

Figure 3: Patient Disposition — ZUMA-7

AE = adverse event; ASCT = autologous stem cell transplant; auto-SCT = autologous stem cell transplant; FAS = full analysis set; HDT = high-dose therapy; PD = progressive disease; SOCT = standard of care therapy; SD = stable disease.

^a Patient was ineligible.

^b One patient had an AE of increased alanine transaminase; 1 patient had an AE of hyperbilirubinemia.

^c Patient in false progression at baseline; reassessment showed they were not progressing.

^d One patient had an AE of cerebrovascular accident; 1 patient had an AE of small intestinal perforation.

^e Three patients because of reasons related to insurance, 1 patient due to rapid progression, and 1 patient opted out.

^f One patient had a negative disease biopsy; 1 patient had a false-positive PET-CT and no refractory double-hit lymphoma after R-EPOCH x 5 (rituximab+etoposide- prednisolone -vincristine-cyclophosphamide-doxorubicin).

^g Withdrawals: 5 patients withdrew with full consent due to patient request. Patients are also included in the categories of reasons not received.

^h Includes 4 patients with PD who were leukapheresed. PD represents best response to salvage chemotherapy.

ⁱ Includes 1 patient with SD who was leukapheresed. SD represents best response to salvage chemotherapy.

^j Patient had an AE of acute kidney injury.

^k Includes 1 patient with lack of response to salvage chemoimmunotherapy with rituximab+ifosfamide-carboplatin-etoposide (R-ICE); 1 patient who did not tolerate rituximab+gemcitabine-dexamethasone-cisplatin/carboplatin (R-GDP) and switched to R-ICE; 1 patient who changed treatment after 1 cycle of rituximab+dexamethasone-cytarabine-cisplatin (R-DHAP) due to renal impairment; and 1 patient with insufficient overall response to proceed to ASCT per investigator.

^l Withdrawals: Patient withdrew with full consent; 4 patients completed therapy but no response; 3 patients with PD. Patients are also included in the categories of reasons for not proceeding.

^m As determined by the investigator.

ⁿ PD represents disease progression after an initial response to salvage chemotherapy.

^o Patient had an AE of blood stem cell harvest failure.

^p As determined by the investigator.

^q Patient was inadvertently enrolled on an alternative protocol.

Source: Clinical Study Report for Yescarta.¹⁵

Protocol Deviations

Important protocol deviations were reported for 56 patients (16%) during the study, including 31 patients (17%) in the axi-cel arm and 25 patients (14%) in the SOC arm. These included patients receiving prohibited investigational agents or radiation (2%), exclusion criteria (1%), inclusion criteria (4%), and missing data (8%). Each category of important protocol deviations occurred for less than 10% of patients overall. The most frequent important protocol deviation in both the axi-cel and SOC arms was missing data (20 patients [11%] and 10 patients [6%], respectively).

Exposure to Study Treatments

Axi-Cel Arm

Bridging Therapy

Of the 170 patients treated with axi-cel, 60 patients (35%) received bridging therapy, most of whom (48 patients, 80%) received at least 1 dose of dexamethasone.

Lymphodepleting Chemotherapy

Of the 180 patients randomized to the axi-cel arm, 172 (96%) received lymphodepleting chemotherapy. Of the 169 patients in the safety analysis set with available body surface area adjusted dose information, the planned total BSA-adjusted dose (± 10%) of cyclophosphamide (1,500 mg/m²) and fludarabine (90 mg/m²) was administered to 165 patients (98%) and 164 patients (97%), respectively. Six patients (3%) did not receive lymphodepleting chemotherapy, primarily due to AEs or disease progression.

Axi-Cel

Of the 180 patients randomized to the axi-cel arm, 170 patients (94%) received axi-cel infusion. The median weight-adjusted dose of axi-cel was 2.0 × 10⁶ anti-CD19 CAR T cells/kg body weight (range, 1.0 × 10⁶ cells/kg to 2.1 × 10⁶ cells/kg). In total, 166 patients (98%) received within 10% of the target dose of 2 × 10⁶ anti-CD19 CAR T cells/kg (or a flat dose of 200 × 10⁶ anti-CD19 CAR T cells for patients weighing > 100 kg). The median total number of anti-CD19 CAR T cells in the axi-cel infusion was 170 × 10⁶ cells (range, 58 × 10⁶ cells to 200 × 10⁶ cells), and the median total number of T cells infused was 301.5 × 10⁶ cells (range, 88 × 10⁶ cells to 633 × 10⁶ cells).

Two patients who received lymphodepleting chemotherapy did not receive axi-cel due to AEs unrelated to leukapheresis or lymphodepleting chemotherapy. Manufacturing failure or delay in product availability was not reported as a limitation to any patient not receiving axi-cel. Nine patients (5%) who achieved a PR or CR at the study day 50 disease assessment and subsequently experienced disease progression were re-treated with axi-cel. The second dose of axi-cel could have been manufactured at the same time as the first axi-cel dose or manufactured later from existing cryopreserved peripheral blood mononuclear cells or a second leukapheresis.

Manufacturing Turnaround Time

For the 170 patients who received axi-cel ████ ████████ ██ █████ ███████ ██████ ███ ███ ███████ ███ ██ ████████ ██ ███ ████ ██ ███ █████ ██████ ███████ ███████ ███ ████████████ the median time from leukapheresis to the product’s release from the manufacturing site was 13 days (range, 10 to 24 days) overall, ██ ████ ███████ ██ ██ ██ █████ ███ █████ ███████ ███ ██ ████ ███████ ██ ██ ██ █████ ███ █████ ████████ Axi-cel was delivered to the study site after a median of 18 days (range, 13 to 49 days) overall, ██ ████ ███████ ██ ██ ██ █████ ███ █████ ████████ ███ ██ ████ ███████ ██ ██ ██ █████ ███ █████ ████████

███ ██████ ████ ████ █████████████ ██ █████████████ ███ ███ ████ ███████ █ ██ ██ ██████ The median time from leukapheresis to axi-cel administration was 26 days (range, 16 to 52 days) ████████ ██ ████ ███████ ██ ██ ██ █████ ███ ████████ ██ █████ ████████ ███ ██ ████ ███████ ██ ██ ██ █████ ███ ████████ ██ █████ ████████ ███ ████████ ██ ████ ████ █████████████ ██ ███████ ██████████████ ████████ ███████ █████████████ ███ ██████████ ██ ███████████████ █████████████ ████ █ ██████ ████ ██ ██ ████ ███████ ██ ██ ██ █████ ████████ ██ ████ ███████ ██ ██ ██ █████ ███ ████████ ██ █████ ████████ ███ ██ ████ ███████ ██ ██ ██ █████ ███ ████████ ██ ███ █████ ████████ ███ ████ ███████ ██████████ ██ ███████████████ ████████████ ███ ███████ ████████ ███ █ ████ ███ ███ ████████

Of the 170 patients who received axi-cel, 166 patients (98%) received within 10% of the planned dose of 2 × 10⁶ anti-CD19 CAR T cells/kg body weight (or a flat dose of 200 × 10⁶ anti-CD19 CAR T cells for patients weighing over 100 kg). For the 137 patients who received axi-cel and weighed 100 kg or less, the median weight-adjusted dose was 2 × 10⁶ anti-CD19 CAR T cells/kg body weight (range, 1.0 × 10⁶ cells/kg to 2.1 × 10⁶ cells/kg). The 33 patients who received axi-cel and who weighed more than 100 kg received the planned flat dose of 200 × 10⁶ cells. For all 170 patients, the median total number of anti-CD19 CAR T cells in the axi-cel infusion was 170 × 10⁶ cells (range, 58 × 10⁶ cells to 200 × 10⁶ cells) and the median total number of T cells infused was 301.5 × 10⁶ (range, 88 × 10⁶ cells to 633 × 10⁶ cells).

SOC Arm

Of the 168 patients who received SOC salvage chemotherapy, 84 patients (50%) received R-ICE, 42 patients (25%) received R-GDP, 37 patients (22%) received R-DHAP/R-DHAX, and 5 patients (3%) received rituximab+etoposide-methylprednisolone-cytarabine-cisplatin (R-ESHAP). Just over half of the patients (91 patients; 54%) received 2 cycles of salvage chemotherapy and 61 patients (36%) received 3 cycles; 16 patients (10%) received only 1 cycle. A total of 62 patients (37% of the SOC – safety analysis set) who achieved a CR or PR with SOC salvage chemotherapy went on to receive HDT-ASCT. Of the 62 patients who received HDT plus on-protocol ASCT, 42 achieved a response after 3 cycles and 20 patients achieved a response after 2 cycles.

Prior Therapy

All patients received chemotherapy with rituximab in first line. Exposure to prior chemotherapy regimens was generally balanced between the treatment arms. The majority of patients in both arms (59%) had previously been treated with CHOP or a CHOP-like regimen with rituximab only. Prior chemoimmunotherapy regimens are detailed in Table 9.

Table 9: Prior Therapy for Primary Study Disease — ZUMA-7 Study (Full Analysis Set)

axi-cel = axicabtagene ciloleucel; CHOP = cyclophosphamide+doxorubicin-vincristine-prednisone; CNS = central nervous system; EPOCH = etoposide-prednisone-vincristine-cyclophosphamide-doxorubicin; HyperCVAD = hyperfractionated therapy of cyclophosphamide-vincristine sulphate-doxorubicin (Adriamycin)-dexamethasone; SOC = standard of care.

Note: Multiple prior therapies within the same line are counted as 1 incidence.

Source: Clinical Study Report for Yescarta.¹⁵

Concomitant Medications and Procedures

Among patients who received axi-cel, 77 patients (45%) received corticosteroids (with or without tocilizumab), 112 patients (66%) were treated with tocilizumab (with or without corticosteroids), and 68 patients (40%) were treated with corticosteroids and tocilizumab. Reasons for corticosteroid treatment included neurologic events (54 patients; 32%), CRS (40 patients; 24%), for other AEs (18 patients; 11%), or other uses that included prophylaxis and medical history (36 patients; 21%). Reasons for tocilizumab treatment included CRS (111 patients; 65%), neurologic events (17 patients; 10%), for other AEs (11 patients, 6%), or for other uses (1 patient, 1%). Nineteen patients (11%) were treated with vasopressors, and 28 patients (16%) were treated with immunoglobulins.

Subsequent Anticancer Therapy

Although treatment switching between the treatment groups was not planned, patients who did not respond to SOC salvage chemotherapy or who relapsed after ASCT could receive cell therapy (including CAR T-cell therapy) outside the protocol as per local clinical decision.

Of the 70 patients (41%) in the axi-cel arm who received subsequent lymphoma therapy, none received a cell therapy. Of the 120 patients (71%) in the SOC arm who received subsequent lymphoma therapy, 75 patients (63%) received axi-cel, 16 patients (13%) received tisagenlecleucel, 5 patients (4%) were reported to receive a CAR T-cell therapy without further explanation of type, and 1 patient (0.8%) received CD19/CD22 CAR T-cell therapy.

Nineteen patients (11%) in the axi-cel arm received subsequent stem cell transfer, including 11 patients who received ASCT and 8 patients who received allo-SCT, and 11 patients (7%) in the SOC arm received subsequent stem cell transfer, including 5 patients who received ASCT, 5 patients who received allo-SCT, and 1 patient who received both.

Efficacy

Only those efficacy outcomes and analyses of subgroups identified in the review protocol are reported below.

All efficacy analyses reported below are based on the full analysis set (intention-to-treat population; N = 359).

Overall Survival

The OS data remain immature. The primary OS analysis will occur at approximately 210 deaths or 5 years after the first patient was enrolled, whichever is earlier. The interim analysis of OS occurred following demonstration of a statistically significant improvement in EFS and ORR for axi-cel compared with SOC. Based on the a priori alpha spending function, a 1-sided alpha of 0.004 was allocated for the interim analysis. The HR for death was 0.730 (95% CI, 0.530 to 1.007; 1-sided stratified log-rank P = 0.027) (Table 10 and Figure 4).

Table 10: Overall Survival — ZUMA-7 Study (Full Analysis Set)

axi-cel = axicabtagene ciloleucel; CI = confidence interval, NE = not estimable; NR = not reached; OS = overall survival; SOC = standard of care.

Note: Data cut-off date: March 18, 2021.

^aA 1-sided alpha of 0.004 was allocated to this analysis.

^bThe stratification factors are responses to first-line therapy (primary refractory vs. relapse ≤ 6 months of first-line therapy vs. relapse > 6 and ≤ 12 months of first-line therapy) and second-line age-adjusted International Prognostic Index (0 to 1 vs. 2 to 3).

^cEstimated using Cox regression models.

^dEstimated using Kaplan-Meier estimation.

^eEstimated using the reverse Kaplan-Meier approach.

Source: Clinical Study Report for Yescarta.¹⁵

Figure 4: Kaplan-Meier Plot of Overall Survival (Full Analysis Set)

axi-cel = axicabtagene ciloleucel; CI = confidence interval; HR = hazard ratio; NE = not estimable; NR = not reached; OS = overall survival; SOC = standard of care; SOCT = standard of care therapy.

Notes: The stratification factors are responses to first-line therapy (primary refractory vs. relapse ≤ 6 months of first-line therapy vs. relapse > 6 and ≤ 12 months of first-line therapy) and second-line age-adjusted International Prognostic Index (0 to 1 vs. 2 to 3) as collected via interactive voice/web response system.

Stratified Cox regression models were used to provide the estimated HR and 2-sided 95% CIs for axi-cel relative to SOC. The Breslow method was used to handle the ties for the Cox regression models. One-sided P value from log-rank test is presented. A 1-sided alpha of 0.004 was allocated to this interim analysis.

Data cut-off date: March 18, 2021.

Source: Clinical Study Report for Yescarta.¹⁵

Addendum to OS Data

An addendum to the ZUMA-7 Clinical Study Report was made to provide data generated in response to a Health Authority request to obtain additional survival follow-up, including from public records, for patients discontinued from ZUMA-7 that was not available at the time of the interim OS analysis (which was conducted at the time of the primary EFS analysis) for completeness of the interim OS data.⁴⁹

By the EFS primary analysis data cut-off date (March 18, 2021), 14 patients had discontinued from the study and were either lost to follow-up, had withdrawn consent, or had been withdrawn by the investigator. A subsequent search of public records identified additional survival data for 8 of the discontinued patients, including 4 patients (all in the SOC arm) who had died before the primary analysis data cut-off date, and 4 patients (3 in the SOC arm and 1 in the axi-cel arm) confirmed as being alive at the primary analysis cut-off date. Additional survival data for the remaining 6 patients (5 in the SOC and 1 in the axi-cel arm) could not be obtained. The interim OS analysis data were updated (with the same data cut-off date of March 18, 2021) to include these additional data.

By the primary analysis data cut-off date, 153 OS events (i.e., deaths) were observed in the full analysis set and included in the primary analysis. The additional deaths of 4 patients in the SOC arm that were identified after the primary analysis resulted in a total of 157 OS events (72 deaths in the axi-cel arm and 85 deaths in the SOC arm).

The update to the interim OS analysis with additional survival data were consistent with the OS interim analysis originally conducted. The stratified HR was 0.708 (95% CI, 0.515 to 0.972; P = 0.0159). While the median OS in the axi-cel arm was consistent with the median OS reported in the original interim OS analysis (not reached; 95% CI, 28.3 months to NE), the median OS in the SOC arm was reduced from 35.1 months (95% CI, 18.5 months to NE) in the original interim OS analysis to 25.7 months (95% CI, 17.6 months to NE) in the update to the interim OS analysis.

Sensitivity Analysis of OS

The results of the sensitivity analysis to address the confounding effects of treatment switching in the SOC arm were consistent with the original interim analysis of OS (RPSFT model stratified HR = 0.575 [95% CI, 0.413 to 0.800] and IPCW model stratified HR = 0.618 [95% CI, 0.417 to 0.916]).

Table 11: Overall Survival With Additional Survival Data for Discontinued Patients — ZUMA-7 Study (Full Analysis Set)

axi-cel = axicabtagene ciloleucel; CI = confidence interval, NA = not applicable; NE = not estimable; NR = not reached; OS = overall survival; SOC = standard of care.

Note: Data cut-off date: March 18, 2021.

^aThe stratification factors are responses to first-line therapy (primary refractory vs. relapse ≤ 6 months of first-line therapy vs. relapse > 6 and ≤ 12 months of first-line therapy) and second-line age-adjusted International Prognostic Index (0 to 1 vs. 2 to 3).

^bEstimated using Cox regression models.

^cEstimated using the Kaplan-Meier estimation.

^dEstimated using the reverse Kaplan-Meier approach.

Source: Addendum to the primary analysis clinical study report (internal sponsor’s report).⁴⁹

Figure 5: Kaplan-Meier Plot of Overall Survival With Additional Survival Data for Discontinued Patients (Full Analysis Set)

axi-cel = axicabtagene ciloleucel; CI = confidence interval; HR = hazard ratio; NE = not estimable; NR = not reached; OS = overall survival; SOCT = standard of care therapy.

Notes: The stratification factors are responses to first-line therapy (primary refractory vs. relapse ≤ 6 months of first-line therapy vs. relapse > 6 and ≤ 12 months of first-line therapy) and second-line age-adjusted International Prognostic Index (0 to 1 vs. 2 to 3) as collected via interactive voice/web response system.

Stratified Cox regression models were used to provide the estimated HR and 2-sided 95% CIs for axi-cel relative to standard of care. The Breslow method was used to handle the ties for the Cox regression models. One-sided P value from log-rank test is presented.

Data cut-off date: March 18, 2021.

Source: Addendum to the primary analysis clinical study report (Sponsor’s internal report).⁴⁹

OS Sensitivity Analyses

A prespecified sensitivity analysis of OS conducted to address the confounding effects of subsequent cell therapy in the SOC arm showed a stratified HR of 0.580 (95% CI, 0.416 to 0.809) using the RPSFT model and 0.695 (95% CI, 0.461 to 1.049) using the IPCW model. There was no multiplicity adjustment for these analyses.

Event-Free Survival

At the time of the data cut-off (March 18, 2021), 252 EFS events as determined by blinded central assessment had occurred for 108 patients (60%) in the axi-cel arm and 144 patients (80%) in the SOC arm. The median EFS was 8.3 months (95% CI, 4.5 to 15.8 months) in the axi-cel arm and 2.0 months (95% CI, 1.6 to 2.8 months) in the SOC arm. The stratified HR for event or death was 0.398 (95% CI, 0.308 to 0.514; P < 0.0001). The Kaplan-Meier estimates of the percentage of patients who remained event-free at 12 and 24 months after randomization in the axi-cel arm were 47.2% (95% CI, 39.8% to 54.3%) and 40.5% (95% CI, 33.2% to 47.7%), respectively, compared with 17.6% (95% CI, 12.3% to 23.6%) and 16.3% (95% CI, 11.1% to 22.2%), respectively, in the SOC arm (Table 12 and Figure 6).

The most common EFS events in either the axi-cel or SOC arm were disease progression (46% and 42%, respectively), new lymphoma therapy (5% and 35%, respectively), and death from any cause (6% and 3%, respectively).

Twelve patients (2 in the axi-cel and 10 in the SOC arm) initiated a new lymphoma therapy in the absence of an evaluable post-baseline disease assessment and had EFS event dates imputed as the randomization date as predefined in the statistical analysis plan. No patients in the axi-cel arm underwent ASCT while in response.

Table 12: Event-Free Survival Per Blinded Central Assessment (Full Analysis Set)

axi-cel = axicabtagene ciloleucel; CI = confidence interval; EFS = event-free survival; HR = hazard ratio; NA = not applicable; NE = not estimable; SOC = standard of care.

Note: Data cut-off date: March 18, 2021.

^aThe stratification factors are responses to first-line therapy (primary refractory vs. relapse ≤ 6 months of first-line therapy vs. relapse > 6 and ≤ 12 months of first-line therapy) and second-line age-adjusted International Prognostic Index (0 to 1 vs. 2 to 3) as collected via interactive voice/web response system.

^bEstimated using the Kaplan-Meier method.

^cEstimated using the reverse Kaplan-Meier approach.

Source: Clinical Study Report for Yescarta.¹⁵

Figure 6: Kaplan-Meier Plot of Event-Free Survival Per Central Assessment (Full Analysis Set)

axi-cel = axicabtagene ciloleucel; CI = confidence interval; HR = hazard ratio; SOCT = standard of care therapy.

Note: Data cut-off date: March 18, 2021.

Source: Clinical Study Report for Yescarta.¹⁵

Subgroup Analysis

The improvement in EFS with axi-cel compared with SOC was consistent in the subgroups of interest (Table 13).

Table 13: Event-Free Survival Per Blinded Central Assessment by Subgroup

axi-cel = axicabtagene ciloleucel; CI = confidence interval; ECOG = Eastern Cooperative Oncology Group; HR = hazard ratio; SOC = standard of care.

Note: Data cut-off date: March 18, 2021.

Source: Clinical Study Report for Yescarta.¹⁵

Sensitivity Analysis

A sensitivity analysis wherein patients who underwent ASCT while in axi-cel-induced response were imputed to have an EFS event at the time of ASCT was consistent with the main assessment of EFS (HR = 0.406; 95% CI, 0.313 to 0.525). An assessment of nonproportional hazards was also performed and the corresponding sensitivity analysis was consistent with the primary analysis (HR = 0.444; 95% CI, 0.333 to 0.590).

The FDA’s Assessment of EFS in Patients in the SOC Arm Who Underwent HDT-ASCT

To understand the outcome of the subgroup of patients that underwent ASCT, the FDA analyzed the EFS of the 62 patients who underwent transplant in the SOC arm. The objective of this exploratory subgroup analysis was not to compare the outcome of the transplant subgroup with the axi-cel arm given that this is not a randomized comparison and the 2 groups are not comparable in terms of baseline characteristics; patients who underwent ASCT were more likely than the patients in the axi-cel arm to have relapsed disease (37% versus 26%) and ECOG PS score of 0 (68% versus 53%). Rather, the purpose of the subgroup analysis was to evaluate the efficacy of ASCT in the subset of chemosensitive patients within the primary refractory and early relapsed population. Results of this exploratory analysis showed a median EFS of 12.1 months (95% CI, 8.5 months to NE) with a 1-year EFS of 51% (95% CI, 38% to 64%).⁴²

Health-Related Quality of Life

Of the 359 patients enrolled in the ZUMA-7 study, 296 patients (165 [92%] patients in the axi-cel arm and 131 [73%] patients in the SOC arm) had baseline HRQoL data and at least 1 follow-up assessment and were included for analysis in the HRQoL Analysis set. By study day 100, data were available for 146 patients (81%) in the axi-cel arm and 62 patients (35%) in the SOC arm. By study day 150, data were available for 110 patients (61%) in the axi-cel arm and 56 patients (31%) in the SOC arm. Samples sizes continued to fall and by month 15, data were available for 67 patients (37%) in the axi-cel arm and 26 patients (15%) in the SOC arm.

EORTC QLQ-C30 Global Health Status/QoL

At screening, the mean EORTC QLQ-C30 global health status scores for evaluable patients were comparable in the axi-cel and SOC arms (mean = 68.6 [SD = 19.9] and mean = 70.1 [SD = 23.1], respectively). At study day 50, almost half of evaluable patients (41.1%) reported worsening (exceeding MID threshold of ± 10 points used in the trial) scores in both treatment arms (mean scores: 60.9 [95% CI, 57.7 to 64.2] and 61.3 [95% CI, 57.3 to 65.3], respectively). Starting at study day 100, scores in the axi-cel arm rebounded while those in the SOC arm declined, with 26.7% of evaluable patients who received axi-cel and 11.3% of patients who received SOC reporting improved scores compared with screening (mean scores: 70.4; [95% CI, 67.0 to 73.9] and 57.1 [95% CI, 51.6 to 62.7], respectively).

Figure 7 shows changes from screening in global health status/QoL. The summary of differences in change from screening is provided in Table 14.

There was a clinically meaningful difference in mean change of scores from screening at study day 100 (estimated difference = 18.1; 95% CI, 12.3 to 23.9; adjusted P < 0.0001) in favour of patients treated with axi-cel compared with patients treated with SOC. At study day 150, the estimated difference was 9.8 (95% CI, 2.6 to 17.0; adjusted P = 0.0124) in favour of patients treated with axi-cel compared with patients treated with SOC. The difference did not reach the trial-specified MID of ± 10 points. By study month 9, the estimated difference between treatment arms was 4.4 (95% CI, –3.3 to 12.0; adjusted P = 0.2655). At study months 12 and 15, the estimated differences between groups were –1.5 (95% CI, –9.6 to 6.6) and –4.9 (95% CI, –13.0 to 3.1), respectively. There was no statistical testing at either time point due to previous failure of the hierarchical testing procedure.

Figure 7: Change From Screening by Visit in EORTC QLQ-C30 Global Health Status/Quality of Life Scores (QoL Analysis Set)

CI = confidence interval; EORTC QLQ-C30 = European Organisation for Research and Treatment of Cancer Quality of Life Questionnaire Core 30; QoL = quality of life.

Notes: Study day = number of days since the day of randomization.

Changes from screening in global health status/quality of life scores were estimated using mixed models for repeated measures.

Results populated only through month 15 due to lack of model convergence when using time points. Horizontal lines represent the minimal important difference thresholds for meaningful change and are provided for clarity of interpretation. Model included variables for treatment, time, and treatment by time interaction (primary analysis) and controlled for response to first-line therapy (primary refractory, relapse ≤ 6 months of first-line therapy vs. relapse > 6 and ≤ 12 months of first-line therapy) and age-adjusted International Prognostic Index (0 to 1 vs. 2 to 3) at time of screening.

Data cut-off date: March 18, 2021.

Source: Clinical Study Report for Yescarta.¹⁵

Table 14: Summary of Difference in Change From Screening in EORTC QLQ-C30 Global Health Status/QoL Scores (QoL Analysis Set)

axi-cel = axicabtagene ciloleucel; CI = confidence interval; EORTC QLC-C30 = European Organisation for Research and Treatment of Cancer Quality of Life Questionnaire Core 30; NR = not reported; QoL = quality of life; SOC = standard of care.

Note: Data cut-off date: March 18, 2021.

^aEstimated using mixed models for repeated measures. Between-group difference is between the axi-cel and the SOC arms.

^bP values only presented for day 100 and only for subsequent visits when the previous visit was statistically significant (P < 0.05).

^cAdjusted P values were calculated using the false discovery rate methodology.

Source: Clinical Study Report for Yescarta.¹⁵

EORTC QLQ-C30 Physical Functioning

At screening, the mean and median physical functioning scores for evaluable patients in the QoL analysis set were comparable in the axi-cel and SOC arms (mean [SD]: 83.5 [95% CI, 80.8 to 86.2] and 85.3 [95% CI, 82.0 to 88.6], respectively). Mean scores ranged from 70.7 (95% CI, 66.9 to 74.5) on study day 50 to 88.8 (95% CI, 83.0 to 94.6) at month 21 for the axi-cel and between 74.0 (95% CI, 68.7 to 79.4) at study day 100 to 97.2 (95% CI, 94.4 to 100.1) at month 24 in the SOC arm. At study day 50, 48.1% of evaluable patients in the axi-cel arm and 37.3% of patients in the SOC arm reported worsening scores. Starting at study day 100, scores for both treatment arms rebounded, with 19.9% of evaluable patients who received axi-cel and 4.7% of patients who received SOC reporting improved scores compared with screening. Both worsening and improved scores were based on a change of ± 10 points from screening (Figure 8).

For patients in the QoL analysis set treated with axi-cel versus SOC, there was a clinically meaningful difference in mean change of scores from screening to study day 100 (estimated difference = 13.1; 95% CI, 8.0 to 18.2; adjusted P < 0.0001) in favour of axi-cel. At day 150, the estimated difference between groups was 5.1 (95% CI, −0.9 to 11.0; adjusted P = 0.1253). At months 9, 12, and 15, the estimated differences between groups were 3.6 (95% CI, −2.7 to 9.8), −2.7 (95% CI, −9.8 to 4.5), and −2.9 (95% CI, −9.7 to 4.0), respectively. There was no statistical testing at these time points due to previous failure of the hierarchical testing procedure.

Figure 8: EORTC QLQ-C30 Physical Functioning Mixed Model With Repeated Measures for Change From Screening (QoL Analysis Set)

CI = confidence interval; EORTC QLC-C30 = European Organisation for Research and Treatment of Cancer Quality of Life Questionnaire Core 30; MMRM = mixed model with repeated measures; QoL = quality of life.

Notes: Study day denotes number of days since the day of randomization. Results populated only through month 15 due to lack of model convergence when using time points. Horizontal lines represent the minimal important difference thresholds for meaningful change and are provided for clarity of interpretation. This model included variables for treatment, time, and treatment by time interaction (primary analysis) and controlled for response to first-line therapy (primary refractory, relapse ≤ 6 months of first-line therapy vs. relapse > 6 and ≤ 12 months of first-line therapy) and age-adjusted International Prognostic Index (0 to 1 vs. 2 to 3) at time of screening.

Data cut-off date: March 18, 2021.

Source: Clinical Study Report for Yescarta.¹⁵

EQ-5D-5L VAS

The mean EQ VAS score reported by evaluable patients in the axi-cel and SOC arms were comparable at screening (72.4 [95% CI, 69.5 to 75.2] and 74.4 [95% CI, 70.9 to 77.9], respectively). Based on the MMRM models, there was a clinically meaningful difference in mean change of scores for the EQ-5D-5L VAS from screening in the axi-cel arm compared with SOC at study day 100 (estimated difference = 13.7; 95% CI, 8.5 to 18.8; adjusted P < 0.0001) and study day 150 (estimated difference = 11.3, 95% CI, 5.4 to 17.1; adjusted P = 0.0004). At month 9, the estimated difference between groups was 3.8 (95% CI, –2.3 to 10.0; adjusted P = 0.2549). At months 12 and 15, the estimated differences between groups were 0.3 (95% CI, –6.3 to 6.8) and –3.4 (95% CI, –10.4 to 3.6), respectively. There was no statistical testing at these time points due to previous failure of the hierarchical testing procedure. Figure 9 shows changes from screening in EQ VAS. The summary of difference in change from screening is provided in Table 15.

Figure 9: Change From Screening by Visit in EQ-5D-5L VAS Scores (QoL Analysis Set)

CI = confidence interval; EQ-5D-5L = 5-Level EQ-5D; QoL = quality of life; VAS = visual analogue scale.

Notes: Study day denotes number of days since the day of randomization.

Changes from screening in EQ-5D-5L VAS scores were estimated using mixed models for repeated measures.

Results populated only through month 15 due to lack of model convergence when using time points. Horizontal lines represent the minimal important difference thresholds for meaningful change and are provided for clarity of interpretation. Model included variables for treatment, time, and treatment by time interaction (primary analysis) and controlled for response to first-line therapy (primary refractory, relapse ≤ 6 months of first-line therapy vs. relapse > 6 and ≤ 12 months of first-line therapy) and age-adjusted International Prognostic Index (0 to 1 vs. 2 to 3) at time of screening.

Note: Data cut-off date: March 18, 2021.

Source: Clinical Study Report for Yescarta.¹⁵

Table 15: Summary of Difference in Change From Screening in EQ-5D-5L VAS Scores (QoL Analysis Set)

axi-cel = axicabtagene ciloleucel; CI = confidence interval; EQ-5D-5L = 5-Level EQ-5D; NR = not reported; QoL = quality of life; SOC = standard of care; VAS = visual analogue scale.

Note: Data cut-off date: March 18, 2021.

^aEstimated using mixed models for repeated measures. Between-group difference is between the axi-cel and the SOC arms.

^bP values only presented for study day 100 and only for subsequent visits when the previous visit was statistically significant (P < 0.05).

^cAdjusted P values were calculated using the false discovery rate methodology.

Source: Clinical Study Report for Yescarta.¹⁵

Progression-Free Survival

At the data cut-off date, the median follow-up time was 21.6 months (95% CI, 20.7 to 23.7 months) in the axi-cel arm and 12.1 months (95% CI, 3.7 months to 18.3 months) in the SOC arm. The median duration of PFS was 14.9 months (95% CI, 7.2 months to NE) in the axi-cel arm and 5.0 months (95% CI, 3.4 to 8.5 months) in the SOC arm, with a stratified HR of 0.562 (95% CI, 0.414 to 0.762) (Table 16).

The Kaplan-Meier estimates of PFS rates at 12 and 24 months from randomization were 53.6% (95% CI, 45.8% to 60.7%) and 46.3% (95% CI, 38.5% to 53.7%), respectively, in the axi-cel arm and 32.3% (95% CI, 23.5% to 41.4%) and 30.9% (95% CI, 22.2% to 40.1%), respectively, in the SOC arm.

Table 16: Progression-Free Survival Per Central Assessment (Full Analysis Set)

axi-cel = axicabtagene ciloleucel; CI = confidence interval; HR = hazard ratio; NE = not estimable; PFS = progression-free survival; SOC = standard of care.

Note: Data cut-off date: March 18, 2021.

^aThe stratification factors are responses to first-line therapy (primary refractory vs. relapse ≤ 6 months of first-line therapy vs. relapse > 6 and ≤ 12 months of first-line therapy) and second-line age-adjusted International Prognostic Index (0 to 1 vs. 2 to 3) as collected via interactive voice/web response system.

^bHR and 95% CIs were estimated using Cox regression models for axi-cel relative to SOC. The Breslow method was used to handle ties for the Cox regression models.

^cEstimated using the Kaplan-Meier method.

^dEstimated using the reverse Kaplan-Meier approach.

Source: Clinical Study Report for Yescarta.¹⁵

Objective Response Rate

The difference in ORR between arms was tested statistically following demonstration of a significant improvement in EFS favouring axi-cel compared with SOC. Based on the a priori alpha spending function, an alpha of 0.025 was allocated to the analysis. The ORR (CR or PR) was 83% (95% CI, 77.1% to 88.5%) in the axi-cel arm and 50% (95% CI, 42.7% to 57.8%) in the SOC arm (difference in ORR = 33.1%; 95% CI, 23.2% to 42.1%; P < 0.0001).

The CR rates in the axi-cel arm and the SOC arm were 65% (95% CI, 57.6% to 71.9%; n = 117) and 32% (95% CI, 25.6% to 39.8%; n = 58), respectively. The PR rates were 18% (95% CI, 13.0% to 24.8%; n = 33) and 18% (95% CI, 12.6% to 24.3%; n = 32), respectively.

Table 17: Objective Response Rate and Best Overall Response Per Central Assessment (Full Analysis Set)

axi-cel = axicabtagene ciloleucel; CI = confidence interval; CR = complete response; ORR = objective response rate; PR = partial response; SOC = standard of care.

Note: Data cut-off date: March 18, 2021.

^a95% CI for the difference in ORR (with the SOC arm as reference group) was from the Wilson score method with continuity correction.

^bThe stratification factors are responses to first-line therapy (primary refractory vs. relapse ≤ 6 months of first-line therapy vs. relapse > 6 and ≤ 12 months of first-line therapy) and second-line age-adjusted International Prognostic Index (0 to 1 vs. 2 to 3) as collected via interactive voice/web response system.

^cOne-sided P value from the Cochran-Mantel-Haenszel test. An alpha of 0.025 was allocated to this analysis.

^dEstimated using the Clopper-Pearson method.

Source: Clinical Study Report for Yescarta.¹⁵

Duration of Response

The median time from randomization to the first objective response for patients who achieved a CR or PR per central assessment in the axi-cel arm (150 of 180 patients) or the SOC arm (90 of 179 patients) was 2.04 months (range, 1.5 months to 18.0 months) and 1.58 months (range, 1.1 to 5.3 months), respectively.

DOR was analyzed for the 150 patients in the axi-cel arm and the 90 patients in the SOC arm who achieved an objective response of CR or PR. Median follow-up time was 19.5 months (95% CI, 18.2 to 21.7 months) in the axi-cel arm and 17.3 months (95% CI, 12.7 to 19.6 months) in the SOC arm. The Kaplan-Meier median DOR for the axi-cel arm was 26.9 months (95% CI, 13.6 months to NE) compared with 8.9 months (95% CI, 5.7 months to NE) for the SOC arm (stratified HR = 0.736; 95% CI, 0.488 to 1.108).

The Kaplan-Meier estimates of the percentage of patients who remained in response at 12 and 24 months from first objective response were 60.9% (95% CI, 52.4% to 68.4%) and 54.0% (95% CI, 45.1% to 62.0%), respectively, in the axi-cel arm compared with 47.6% (95% CI, 35.2% to 58.9%) and 45.6% (95% CI, 33.2% to 57.1%), respectively, in the SOC arm.

Table 18: Duration of Response Per Central Assessment (Full Analysis Set)

axi-cel = axicabtagene ciloleucel; CI = confidence interval; CR = complete response; DOR = duration of response; HR = hazard ratio; NE = not estimable; PR = partial response; SOC = standard of care.

Notes: Percentages are based on number of patients in the analysis set with objective response.

Data cut-off date: March 18, 2021.

^aThe stratification factors are responses to first-line therapy (primary refractory vs. relapse ≤ 6 months of first-line therapy vs. relapse > 6 and ≤ 12 months of first-line therapy) and second-line age-adjusted International Prognostic Index (0 to 1 vs. 2 to 3) as collected via interactive voice/web response system.

^bHR and 95% CIs were estimated using Cox regression models for axi-cel relative to SOC.

^cOne-sided P value is based on log-rank test. No multiplicity adjustment was performed.

^dEstimated using the Kaplan-Meier method.

^eEstimated via the reverse Kaplan-Meier approach.

Source: Clinical Study Report for Yescarta.¹⁵

Time to Next Treatment

Time to next treatment events occurred for 99 patients (55%) in the axi-cel arm and 135 patients (75%) in the SOC arm. The median follow time was 24.6 months (95% CI, 22.0 to 25.4 months) in the axi-cel arm and 22.8 months (95% CI, 21.0 to 24.6 months) in the SOC arm. The median time to next treatment was 14.7 months (95% CI, 6.5 months to NE) in the axi-cel arm and 3.4 months (95% CI, 3.1 to 4.4 months) in the SOC arm (stratified HR = 0.430; 95% CI, 0.329 to 0.560; descriptive log-rank P value < 0.0001).

The Kaplan-Meier estimates of the percentage of patients with events 12 and 24 months from randomization were 51.7% (95% CI, 44.1% to 58.7%) and 45.0% (95% CI, 37.6% to 52.2%), respectively, in the axi-cel arm, and 24.2% (95% CI, 18.1% to 30.8%) and 21.1% (95% CI, 15.4% to 27.6%), respectively, in the SOC arm.

Table 19: Time to Next Treatment Per Central Assessment (Full Analysis Set)

axi-cel = axicabtagene ciloleucel; CI = confidence interval; HR = hazard ratio; NE = not estimable; SOC = standard of care.

Note: Data cut-off date: March 18, 2021.

^aThe stratification factors are responses to first-line therapy (primary refractory vs. relapse ≤ 6 months of first-line therapy vs. relapse > 6 and ≤ 12 months of first-line therapy) and second-line age-adjusted International Prognostic Index (0 to 1 vs. 2 to 3) as collected via interactive voice/web response system.

^bHR and 95% CIs were estimated using Cox regression models for axi-cel relative to SOC.

^cBased on the log-rank test. Not adjusted for multiple comparisons.

^dEstimated using the Kaplan-Meier method.

^eEstimated via the reverse Kaplan-Meier approach.

Source: Clinical Study Report for Yescarta.¹⁵

Health Care Resource Utilization

A total of 42 patients (25%) in the axi-cel arm and 9 patients (5%) in the SOC arm were admitted to the ICU. Median duration of ICU hospitalization was 5 days (range, 1 to 12 days) in the axi-cel arm and 3 days (range, 2 to 17 days) in the SOC arm. Median duration of hospitalization for axi-cel infusion was 16 days (range, 5 to 103 days); median duration of inpatient hospitalization for stem cell transplant in the SOC arm was 21 days (range, 1 to 53 days).¹⁴

Harms

Adverse Events

All patients in the axi-cel arm (170 patients [100%]) and the SOC arm (168 patients [100%]) had at least 1 TEAE, and 155 patients (91%) in the axi-cel arm and 140 patients (83%) in the SOC arm had worst grade 3 or higher TEAEs.

TEAEs of any grade that were most frequently (in ≥ 30% of patients) reported in each treatment arm were as follows:

• Axi-cel arm: pyrexia (158 patients; 93%); neutropenia and hypotension (75 patients each; 44%); anemia, fatigue, and diarrhea (71 patients each; 42%); headache (70 patients; 41%); nausea (69 patients; 41%), sinus tachycardia (58 patients; 34%); and neutrophil count decreased (52 patients; 31%)

• SOC arm: nausea (116 patients; 69%); anemia (91 patients; 54%); fatigue (87 patients; 52%); diarrhea (66 patients; 39%); platelet count decreased (64 patients; 38%); constipation (58 patients; 35%); and vomiting (55 patients; 33%).

The most frequently reported TEAEs of worst grade 3 or higher (reported in ≥ 20% of patients in either treatment arm) were as follows:

• Axi-cel arm: neutropenia (73 patients; 43%); anemia (51 patients; 30%); neutrophil count decreased (49 patients; 29%); and white blood cell count decreased (43 patients; 25%)

• SOC arm: anemia (65 patients; 39%); platelet count decreased (60 patients; 36%); neutrophil count decreased (47 patients; 28%); febrile neutropenia (46 patients; 27%); and thrombocytopenia (37 patients; 22%).

Table 20: Summary of Adverse Events (Safety Analysis Set)

AE = adverse event; axi-cel = axicabtagene ciloleucel; CRS = cytokine release syndrome; NA = not applicable; PD = progressive disease; SOC = standard of care; TEAE = treatment-emergent adverse event.

Note: TEAE includes all AEs with an onset on or after the axi-cel infusion date in the axi-cel arm or the first dose of salvage chemotherapy in the SOC arm.

^aAEs are graded per Common Terminology Criteria for Adverse Events Version 4.03, and CRS events are graded according to a modified grading system proposed by Lee et al. (2014).⁵⁰

Source: Clinical Study Report for Yescarta.¹⁵

Serious Adverse Events

A total of 85 patients (50%) in the axi-cel arm and 77 patients (46%) in the SOC arm had at least 1 SAE. The most frequently (in ≥ 5% of patients) reported SAEs of any grade in each treatment arm were as follows:

• Axi-cel arm: pyrexia (27 patients; 16%), encephalopathy (17 patients; 10%), hypotension (15 patients; 9%), aphasia (9 patients; 5%), and pneumonia (8 patients; 5%)

• SOC arm: febrile neutropenia (22 patients; 13%), and acute kidney injury and pyrexia (8 patients each; 5%)

Seventy-two patients (42%) in the axi-cel arm and 67 patients (40%) in the SOC arm had worst grade 3 or higher SAEs. The most frequently (in > 2% of patients) reported worst grade 3 or higher SAEs, excluding B-cell lymphoma, in each treatment arm were as follows:

• Axi-cel arm: encephalopathy (15 patients; 9%), aphasia (8 patients; 5%), hypotension (7 patients; 4%), and pneumonia (6 patients, 4%)

• SOC arm: febrile neutropenia (22 patients; 13%) and platelet count decreased (5 patients; 3%).

Withdrawals Due to Adverse Events

No patient discontinued treatment due to TEAEs in the axi-cel arm. Two patients in the SOC arm discontinued treatment due to TEAEs of grade 4 acute kidney injury and grade 1 blood stem cell harvest failure.

AEs Related to Either Leukapheresis or Lymphodepleting Chemotherapy (Axi-Cel Arm)

At least 1 AE related to leukapheresis (for T cell collection) was reported for 27 of 170 patients (16%) who later received axi-cel, including 5 patients (3%) with worst grade 3 or higher leukapheresis-related AEs. One patient (1%) had a worst grade 4 AE, and no patients had a grade 5 AE related to leukapheresis. At least 1 AE related to lymphodepleting chemotherapy was reported for 151 of 170 patients (89%) who later received axi-cel, including 130 patients (76%) with worst grade 3 or higher lymphodepleting chemotherapy-related AEs. Worst grade 4 lymphodepleting chemotherapy-related AEs were reported for 109 patients (64%), and 1 patient (1%) had a grade 5 AE of progressive multifocal leukoencephalopathy. The 3 most common AEs of any grade reported as related to lymphodepleting chemotherapy were nausea and neutropenia (63 patients each; 37%), and anemia (59 patients; 35%). The 3 most common worst grade 3 or higher lymphodepleting chemotherapy-related AEs were neutropenia (61 patients; 36%), anemia (45 patients; 26%), and neutrophil count decreased (42 patients; 25%).

TEAEs Related to Either Salvage Chemotherapy, HDT, or ASCT (SOC Arm)

At least 1 TEAE related to salvage chemotherapy was reported for 160 of 168 patients (95%) in the SOC arm, including 121 patients (72%) with a worst grade 3 or higher salvage chemotherapy-related TEAE. Worst grade 4 TEAEs were reported for 79 patients (47%), and no patients had a grade 5 TEAE. The 3 most common salvage chemotherapy-related TEAEs of any grade were nausea (95 patients; 57%), fatigue (73 patients; 43%), and anemia (70 patients; 42%). The 3 most common worst grade 3 or higher salvage chemotherapy-related TEAEs were anemia (50 patients; 30%), platelet count decreased (41 patients; 24%), and neutrophil count decreased (32 patients; 19%). Following salvage chemotherapy, 74 patients underwent leukapheresis for CD34+ stem cell collection. At least 1 TEAE related to leukapheresis was reported for 7 patients, including 2 patients with a worst grade 3 or higher leukapheresis-related TEAE. Worst grade 4 TEAEs were reported for 1 patient, and no patient had a grade 5 TEAE that was deemed related to leukapheresis for CD34+ stem cell collection.

Following leukapheresis, 64 patients received HDT. At least 1 HDT-related TEAE was reported for 58 patients, including 52 patients with a worst grade 3 or higher TEAE. Worst grade 4 HDT-related TEAEs were reported for 45 patients, and 2 patients had a grade 5 TEAE of cardiac arrest and acute respiratory distress syndrome. Following HDT, 62 patients received on-protocol CD34+ stem cell infusion. At least 1 CD34+ stem cell infusion–related TEAE was reported for 13 of the 62 patients (21%) who received ASCT in the SOC arm, including 6 patients (10%) with a worst grade 3 or higher TEAE. Worst grade 4 CD34+ stem cell infusion–related TEAEs were reported for 5 patients (8%). No patients had a grade 5 TEAE.

Mortality

In the axi-cel arm, 64 patients (38%) had died at the data cut-off date due to reasons including PD (n = 47; 28%), TEAEs (n = 6; 4%), and other reasons (n = 10; 6%) and 1 patient (1%) died from an event reported by the investigator as a “secondary malignancy” (lung adenocarcinoma) although deemed unrelated to axi-cel treatment (this event was reported as a grade 5 TEAE). In the SOC arm, 78 patients (46%) had died at the data cut-off date due to reasons including PD (n = 64; 38%), TEAEs (n = 2; 1%), and other reasons (n = 12; 7%).

Of the 64 patients in the axi-cel arm who died, 6 died more than 30 days but less than 3 months after the axi-cel infusion, and 58 died more than 3 months after the axi-cel infusion. Of the 78 patients in the SOC arm who died, 3 died more than 30 days but less than 3 months after the first dose of salvage chemotherapy, and 75 died more than 3 months after the first dose of salvage chemotherapy.

Six patients in the axi-cel arm and 2 patients in the SOC arm died due to TEAEs which included myocardial infarction, progressive multifocal leukoencephalopathy, hepatitis B reactivation, sepsis, and COVID19 in the axi-cel arm, and cardiac arrest and acute respiratory distress syndrome in the SOC arm (Table 22).

Table 21: Summary of Deaths (Safety Analysis Set)

axi-cel = axicabtagene ciloleucel; SOC = standard of care.

Note: Data cut-off date: March 18, 2021.

^aOne death was reported by the investigator as a “secondary malignancy” although deemed unrelated to study therapy. The sponsor’s medical review considered the event to be a new malignancy unrelated to study therapy. (Note: this event was reported as a grade 5 treatment-emergent adverse event.)

^bCOVID-19 infection was reported as an adverse event if it occurred during the adverse event reporting window and as “other” if it occurred outside the adverse event reporting window.

Source: Clinical Study Report for Yescarta.¹⁵

Table 22: Deaths Due to Adverse Events (Safety Analysis Set)

axi-cel = axicabtagene ciloleucel; PML = progressive multifocal leukoencephalopathy; SOC = standard of care.

Notes: Deaths were to be reported as serious adverse events if they occurred after randomization and before the study day 150 visit or until the initiation of new lymphoma therapy (whichever occurred first), regardless of attribution to treatment. Targeted SAEs (neurologic events, hematologic events, infections, autoimmune disorders, and secondary malignancies) were to be reported for up to 5 years for the SOC arm and up to 15 years for the axi-cel arm, or until disease progression.

Data cut-off date: March 18, 2021.

Source: Clinical Study Report for Yescarta.¹⁵

Notable Harms

Neurologic Events

In the safety analysis set (N = 338), 102 patients (60%) in the axi-cel arm and 33 patients (20%) in the SOC arm had at least 1 treatment-emergent neurologic event, including 36 patients (21%) and 1 patient (1%), respectively, with worst grade 3 or higher neurologic events. Of these, 10 patients (6%) in the axi-cel arm had worst grade 4 neurologic events; no patients in either treatment arm had a grade 5 neurologic event. The most frequently (in ≥ 10% of patients) reported treatment-emergent neurologic events of any grade in the axi-cel arm were tremor (44 patients; 26%), confusional state (40 patients; 24%), aphasia (36 patients; 21%), encephalopathy (29 patients; 17%), and somnolence (19 patients; 11%). No neurologic events occurred with a patient incidence higher than 10% in the SOC arm; the most frequently (in ≥ 2% of patients) reported treatment-emergent neurologic events in the SOC arm were paresthesia (14 patients; 8%), delirium (5 patients; 3%), and confusional state (4 patients; 2%).The most frequently (in ≥ 5% of patients) reported worst grade 3 or higher treatment-emergent neurologic events in the axi-cel arm were encephalopathy (20 patients, 12%), aphasia (12 patients; 7%), and confusional state (9 patients, 5%). One patient (1%) in the SOC arm had a grade 3 or higher treatment-emergent neurologic event of delirium.

Serious treatment-emergent neurologic events of any grade were reported for 34 patients (20%) in the axi-cel arm and 1 patient (1%) in the SOC arm, including 26 patients (15%) in the axi-cel arm with a serious worst grade 3 or higher neurologic event and 1 patient (1%) in the SOC arm with a serious worst grade 2 neurologic event. The 3 most common serious treatment-emergent neurologic events of any grade in the axi-cel arm were encephalopathy (17 patients; 10%), aphasia (9 patients; 5%), and confusional state (6 patients; 4%). The only serious neurologic event in the SOC arm was encephalopathy.

Among patients who had a treatment-emergent neurologic event, the median time to onset was 7.0 days (range, 1 to 133 days) after the axi-cel infusion and 23.0 days (range, 1 to 108 days) after the first dose of SOC salvage chemotherapy. At the data cut-off date, neurologic events had resolved in 96 of the 102 patients in the axi-cel arm and in 32 of the 33 patients in the SOC arm; the median duration of neurologic events was 8.5 days (range, 1 to 817 days) for patients in the axi-cel arm and 23.0 days (range, 1 to 253 days) for the patients in the SOC arm.

A total of 6 patients in the axi-cel arm and 1 patient in the SOC arm had ongoing neurologic events at the data cut-off date or unresolved neurologic events at the time of death.

Cytokine Release Syndrome

All cases of CRS were considered related to axi-cel. CRS of any grade was reported for 157 patients (92%), including 11 patients (6%) who had worst grade 3 or higher CRS. No patient had grade 5 CRS. The most frequently reported CRS symptoms (in ≥ 30% of patients with CRS) were pyrexia (155 patients; 99%), hypotension (68 patients; 43%), and sinus tachycardia (49 patients; 31%). The most frequently reported worst grade 3 or higher CRS symptoms (in ≥ 5% of patients with CRS) were hypotension (18 patients; 11%), pyrexia (14 patients; 9%), and hypoxia (13 patients; 8%).

Cytopenias

The most common cytopenias of any grade reported in the axi-cel arm were thrombocytopenia (50 patients; 29%), neutropenia (122 patients; 72%), and anemia (73 patients; 43%). The most common cytopenias of any grade reported in the SOC arm were thrombocytopenia (101 patients; 60%), neutropenia (92 patients; 55%), and anemia (92 patients; 55%). Grade 3 or higher cytopenias reported in the axi-cel arm were thrombocytopenia (25 patients; 15%), neutropenia (119 patients; 70%), and anemia (51 patients; 30%). Grade 3 or higher cytopenias reported in the SOC arm were thrombocytopenia (95 patients; 57%), neutropenia (91 patients; 54%), and anemia (65 patients; 39%).

Infections

Seventy patients (41%) in the axi-cel arm and 51 patients (30%) in the SOC arm had at least 1 treatment-emergent infection, including 24 patients (14%) and 19 patients (11%), respectively, with worst grade 3 or higher infections. Three patients (2%) in the axi-cel arm and 6 patients (4%) in the SOC arm had worst grade 4 infections. Five patients (3%) in the axi-cel arm had a grade 5 TEAE of infection (2 patients with COVID-19, 1 patient with progressive multifocal leukoencephalopathy, 1 patient with hepatitis B reactivation, and 1 patient with sepsis). No patients in the SOC arm had a grade 5 TEAE of infection.

The most frequently (in ≥ 5% of patients) reported infection categories were unspecified (44 patients; 26%), viral infections (26 patients; 15%), bacterial infections (16 patients; 9%), upper respiratory tract infections (11 patients; 6%), and opportunistic infections (8 patients; 5%) in the axi-cel arm and unspecified (40 patients; 24%), bacterial infections (15 patients; 9%), and viral infections (8 patients; 5%) in the SOC arm.

The most frequently (in ≥ 2% of patients) reported worst grade 3 or higher treatment-emergent infections by preferred term (excluding COVID-19) were pneumonia (6 patients; 4%) and upper respiratory tract infection (3 patients; 2%) in the axi-cel arm, and pneumonia and sepsis (4 patients each; 2%) in the SOC arm.

Secondary Malignancies

No new malignancies were considered by the sponsor to be secondary to axi-cel treatment (new malignancies considered not to be secondary malignancies by the sponsor included events of the primary malignancy, nonmalignant events such as tumour pain, nonmelanoma skin cancer, events of treatment-related acute myeloid leukemia (t-AML) and treatment-related myelodysplastic syndrome (t-MDS) considered related to underlying genetic predisposition or prior chemotherapy [i.e., chemotherapy administered before axi-cel infusion or the first dose of SOC], and other new malignancies with alternative etiologies).

Critical Appraisal

Internal Validity

In ZUMA-7, treatment assignment was based on a central randomization scheme, which would ensure concealment of the randomized groups until allocation. Randomization was stratified according to response to first-line therapy and sAAIPI, which are important prognostic factors in this setting and ensured that the 2 treatment arms were balanced in the proportion of patients with primary refractory versus relapsed lymphoma and low versus high sAAIPI total score. Overall, the 2 treatment arms were balanced in terms of baseline demographic and disease characteristics. The 10% difference in the proportion of male patients between the 2 arms is unlikely to present a high risk of bias due to the randomization.

The trial was open label. Despite the open-label design, there is low risk of bias in the measurement of outcomes such as EFS, PFS, and ORR, which were assessed via independent blinded radiologic review of disease response, as well as OS, which is an objective outcome. For subjective outcomes like HRQoL and some harms outcomes, the patient’s knowledge of their assigned treatment could result in overestimation or underestimation. Of note, 4% of the enrolled patients in the SOC arm declined to proceed with the study compared to none in the axi-cel arm which could indicate patient bias against SOC therapy, possibly due to perceived lack of efficacy. In addition, although the EORTC QLQ-C30 and EQ-5D-5L are comprehensive and widely used instruments designed to measure HRQoL in the general population and in patient groups with diverse chronic diseases, both are currently not validated in patients with LBCL. Further, at most time points for the analysis of HRQoL, relatively few patients completed the assessment and the amount of missing data was imbalanced between the groups, so the results are at a high risk of attrition bias.

The study was powered (based on the full analysis set, N = 359) to test its primary end point, EFS. The statistical approach of sequentially testing the primary and secondary end points was acceptable to account for multiple testing across these analyses. Of note, the subgroup analyses, although prespecified, were not adjusted for multiple comparisons, there was no test of between subgroup differences, and the trial was not powered to detect subgroup differences. Strong conclusions should not be drawn from these data. The primary and secondary efficacy end points of EFS, ORR, and OS are considered appropriate for the disease setting. There is some evidence from the literature that EFS is a robust surrogate end point for OS in hematological malignancies; a systematic review of patients with DLBCL suggested that PFS and EFS could be regarded as earlier efficacy end points in patients with DLBCL primarily treated with rituximab-containing immunochemotherapy.¹⁹ However, whether this could be extended to CAR T-cell therapies for relapsed or refractory LBCL is unknown. For the purpose of regulatory decision-making, FDA has accepted EFS as primary end point in relapsed refractory lymphoma.⁴²

OS results should be considered in the context of subsequent therapies as patients who did not respond to SOC could receive any subsequent treatment for lymphoma off-protocol deemed appropriate by the investigator (including non–study-specific chemotherapy, immunotherapy, targeted agents, and anti-CD19 CAR T-cell therapy). The treatment switching rate was more than 50%, with 56% of patients randomized to the SOC arm receiving new lymphoma therapy after SOC. Prespecified sensitivity analyses of OS were performed to address the confounding effects from subsequent cell therapy in the SOC arm. The sensitivity analysis results using RPSFT and IPCW were consistent with the main interim OS analysis. These sensitivity analyses, designed to address the confounding effect of subsequent cell therapy in the SOC arm on OS, were also updated to include the additional survival data for discontinued patients, and results also consistent with the main interim OS analysis. Since the OS data are immature and the current results are based on an interim analysis, there is a risk that the benefit of axi-cel compared with SOC is overestimated, although the degree to which this may be true is uncertain.^17,18

External Validity

The ZUMA-7 trial included a heterogenous population of patients with relapsed or refractory LBCL. Although a wide range of clinical presentations were represented, a few patient groups were excluded from the trial. The clinical experts consulted by CADTH noted that patients with an ECOG PS score of 2 and those who are HIV positive should be eligible for treatment with CAR T-cell therapies including axi-cel. In addition, ZUMA-7 enrolled patients who were eligible for ASCT. Patients in need for urgent therapy due to tumour mass effect were excluded from the study limiting the applicability of the data in patients with bulky disease needing urgent treatment. It was also noted that the median age of 59 years is younger than the median age of patients with LBCL in Canada, reflecting the age of the study population that is transplant eligible.⁵¹ Therefore, efficacy results in the trial may differ from those in the real-world context. Most patients in the ZUMA-7 trial were treated with CHOP regimen and all patients received rituximab in first-line, which is the SOC in Canada. Thus, it can be assumed that patients were adequately treated in first line before being characterized as having relapsed or refractory LBCL. Although both the clinical experts and the clinician groups indicated that ASCT eligibility is highly variable in clinical practice and should not be used to determine eligibility for CAR T-cell therapy, given that the clinical evidence from ZUMA-7 is based solely on patients who are eligible for ASCT, generalizability of results to patients who are not considered eligible for ASCT remains uncertain.

The clinical experts consulted by CADTH indicated that the SOC treatment including salvage chemotherapy regimens used in the SOC arm of the ZUMA-7 trial are reflective of Canadian clinical practice. However, some challenges were noted for axi-cel treatment. The manufacturing turnaround time for axi-cel in the ZUMA-7 trial (i.e., the median time from leukapheresis to the product’s release from the manufacturing site) was reported as 13 days (range, 10 to 24 days). The clinical experts questioned whether the same highly standardized processes can be replicated in most centres across Canada and whether and to what extent delays and restrictions in cross-border transport may affect treatment delays in the real-world. In addition, outside the highly controlled clinical trial setting, circumstances encountered more often in the real world may complicate timely treatment of patients and lead to outcomes different than those observed in the trial. The clinical experts noted that while a 13-day manufacturing turnaround for axi-cel was rapid, it may not be reproducible outside the clinical trial setting because manufacturing will be conducted in the US according to the sponsor’s implementation plan. Longer delays from leukapheresis to axi-cel infusion could adversely affect patient outcomes if, for example, patients experience disease progression during this delay. In addition, in the trial setting of ZUMA-7, risk of life-threatening and fatal AEs attributed to axi-cel treatment was mitigated by stringent training of study site staff and investigators and careful monitoring to ensure timely detection and management of the most serious complications. In real-world settings, site preparation, patient and clinician education, and risk mitigation strategies with emphasis on early recognition and treatment of serious and life-threatening complications including CRS and neurologic toxicities will be crucial to ensure patient safety.

Data for all outcomes considered important to patients, clinicians, and drug plans, as outlined in the systematic review protocol, were collected and analyzed in the trial. However, data for HRQoL were collected using a tool not validated among patients with LBCL, and data were available for few patients at most time points, precluding strong conclusions for this outcome. There remains an evidence gap regarding the impact of treatment with axi-cel compared with SOC on the HRQoL of patients with relapsed or refractory LBCL.

Indirect Evidence

A focused literature search for indirect treatment comparisons dealing with LBCL was run in MEDLINE All (1946–) on July 7, 2022. No limits were applied to the search. The search retrieved 32 reports, but none met the eligibility criteria for the systematic review. No indirect treatment comparisons were submitted by the sponsor.

Other Relevant Evidence

Long-Term Data From the ZUMA-1 Study

The sponsor provided long-term (≥ 4 year and ≥ 5 year) data from ZUMA-1, the pivotal multicentre, single-arm, registrational phase I and II study of axi-cel in patients with relapsed or refractory LBCL who had received at least 2 previous systemic therapies. These data were presented as conference abstracts or posters.^20-22

In the 2-year analysis of ZUMA-1 (n = 101; median follow-up from axi-cel dosing to data cut-off = 27.1 months; interquartile range [IQR], 25.7 to 28.8 months), the ORR was 83%; 58% of patients achieved a CR. The 2-year OS rate was 50.5%. The median OS was not reached).⁵²

Year 4²⁰

As of August 11, 2020, median follow-up was 51.1 months. With more than 4 years of follow-up, median OS was 25.8 months, and 4-year OS rate was 44% . Median EFS in patients treated with axi-cel was 5.7 months (range NR), with a 12-month EFS rate of 43% (95% CI, 33% to 52%) and a 24-month EFS rate of 38% (95% CI, 28% to 47%).

Overall, 26 (26%) patients received subsequent anticancer therapy; median time to next therapy was 8.7 months (range, 0.3 to 53.8 months). Immunoglobulin therapy was administered to 38 (38%) patients.

Of the treated patients, 58 (57%) had died, primarily due to progressive disease (46%; n = 46), other reasons (7%; n = 7), AEs (4%; n = 4), and secondary malignancy (myelodysplastic syndrome, with onset on day 574) unrelated to axi-cel (1%; n = 1).

Year 5²²

The most recently updated survival results from phase II ZUMA-1, after 5 years of follow-up, showed a 5-year OS rate of 42.6% (95% CI, 32.8% to 51.9%) among patients treated with axi-cel. The 5-year OS rate among complete responders was 64.4% (95% CI, 50.8% to 75.1%). The median survival time among complete responders was not reached (95% CI, 63.4 months to NE). Of 59 patients who achieved CR, 37 (63%) were alive at the 5-year data cut-off date. Since the 4-year data cut-off date, 1 death at month 63 (CR) and 1 progressive disease at month 54 (PR) were observed.

Median time to next anticancer therapy from axi-cel infusion was 8.7 months (range, 0.3 to 63.4 months), unchanged from the previous (4-year) analysis. By the year 5 data cut-off, 34 patients (34%) were still alive and received no subsequent therapy (excluding stem cell transfer) or re-treatment with axi-cel. Compared with the year 4 data, 2 patients (2%) who had previously progressed received new anticancer therapy.

Exploratory (Post Hoc) Analysis of OS by EFS Status at 12 and 24 Months²²

Among all treated patients, the 12-month and the 24-month EFS rates were 42.8% (95% CI, 33.0% to 52.3%) and 37.7% (95% CI, 28.3% to 47.2%), respectively (Figure 10).

Among patients with (n = 57) and without (n = 44) an EFS event by month 12, 5-year OS rates were 5.3% (95% CI, 1.4% to 13.2%) and 90.9% (95% CI, 77.6% to 96.5%), respectively.

Among patients with (n = 62) and without (n = 39) an EFS event by month 24, 5-year OS rates were 11.3% (95% CI, 5.0% to 20.5%) and 92.3% (95% CI, 78.0% to 97.5%), respectively.

Figure 10: Exploratory Analysis of Overall Survival by Event-Free Survival at 12 and 24 Months in the ZUMA-1 Study

EFS = event-free survival; NE = not estimable; NR = not reached; OS = overall survival.

Source: Jacobson et al. (2021) (sponsor’s internal report).²²

Supportive Safety Data: Comparison of Treated Populations Treated With Axi-Cel in the ZUMA-7 and ZUMA-1 Studies

In addition to data from the ZUMA-7 study, supportive safety data from the ongoing ZUMA-1 study (phase I and II) were included for consideration in the current CADTH review. Both studies treated patients with relapsed or refractory LBCL, including DLBCL and HGBL. The data cut-off date for both studies was March 18, 2021. TEAEs were comparable between patients treated with axi-cel in the ZUMA-1 and ZUMA-7 studies (Table 23). All patients in the ZUMA-7 and ZUMA-1 studies had at least 1 TEAE, including 91% and 96%, respectively, who had grade 3 or less TEAEs and 8% who had grade 5 TEAEs. Approximately half of the patients (50% in ZUMA-7 and 54% in ZUMA-1) had SAEs. The proportion of patients treated with axi-cel who experienced neurologic events or CRS was also similar in the 2 trials.

Table 23: Overall Summary of TEAEs — SOC From the ZUMA-7 Study and the Pooled Axi-Cel Population of the ZUMA-7 and ZUMA-1 Studies (Safety Analysis Set)

axi-cel = axicabtagene ciloleucel; ASCT = autologous stem cell transplant; CRS = cytokine release syndrome; CTCAE = Common Terminology Criteria for Adverse Events; NA = not applicable; SOC = standard of care; TEAE = treatment-emergent adverse event.

Notes: Adverse events (AEs) are coded using Medical Dictionary for Regulatory Activities (MedDRA) Version 23.1. A TEAE is defined as any AE occurring on or after the first axi-cel infusion or the first dose of salvage chemotherapy in the SOC arm in ZUMA-7. TEAEs that occurred during the re-treatment period are excluded.

For patients treated with axi-cel, treatment-related TEAEs are those related to axi-cel. For the SOC arm of ZUMA-7, treatment-related TEAEs are those related to salvage chemotherapy, total body irradiation, high-dose therapy, or autologous stem cell transplant. CRS events were graded according to a modification of the criteria of Lee et al.⁵⁰ Other events were graded according to Common Terminology Criteria for Adverse Events Version 4.03. Seriousness of CRS was not collected in ZUMA-1. Neurologic events are identified using a modified search strategy based on Topp et al. (2015). Multiple incidences of the same AE in 1 patient are counted once at the highest grade for each patient.

ZUMA-1 refers to ZUMA-1 phase I and phase II cohorts 1 and 2.

Data cut-off date is March 18, 2021.

Source: Clinical Summary Common Technical Document Section 2.7.4.⁵³

Critical Appraisal

Internal Validity

The ZUMA-1 trial was critically appraised in the 2019 CADTH Optimal Use Report report.²³ ZUMA-1 was a single-arm, phase I and II study of axi-cel in patients with relapsed or refractory LBCL who had received at least 2 previous systemic therapies. The primary limitation of ZUMA-1 was the absence of a comparator group against which the treatment benefits and harms of axi-cel could be compared. As such, causal effects cannot be inferred. Several limitations related to outcome assessment in the ZUMA-1 trial were also noted. These include the use of investigator disease assessment for the primary analysis (ORR based on independent review committee assessment, was reported as a secondary outcome), which can lead to detection bias particularly in open-label, single-arm trials, and inconsistency in censoring of patients throughout the analyses that could lead to more favourable estimates of the treatment effect of axi-cel. Notably, OS was not censored at the time of additional treatments (i.e., re-treatment with axi-cel, treatment with chemotherapy, or allo-SCT after relapse following axi-cel). Given the different study designs and patient populations in ZUMA-1 and ZUMA-7, the CADTH review team determined that data from the 2 trials cannot be reliably pooled. Therefore, the additional pooled safety data from the 2 trials were not interpreted for this review.

External Validity

ZUMA-1 was conducted in patients with relapsed or refractory LBCL who had received at least 2 previous systemic therapies (i.e., patients received axi-cel as third- or later-line treatment). It is unknown whether results are generalizable to patients treated with axi-cel as second-line treatment, which is the indication under review.

Discussion

Summary of Available Evidence

The ZUMA-7 trial that forms the main evidence base for this review is an ongoing phase III, open-label, randomized controlled trial that compares the efficacy and safety of axi-cel to SOC chemoimmunotherapy in patients with LBCL. Patients were those with disease that was refractory or had relapsed within 12 months of first-line chemoimmunotherapy, including an anti-CD20 monoclonal antibody and an anthracycline-containing chemotherapy regimen, who intended to proceed to HDT and ASCT. The primary end point was EFS by blinded central assessment. Secondary end points included ORR, OS, PFS, DOR, and HRQoL. Time to next treatment was an exploratory end point, and health care resource utilization was also reported. A total of 359 patients were enrolled; 180 were randomized to receive axi-cel, and 179 were randomized to receive SOC therapy consisting of salvage chemotherapy followed by HDT-ASCT. Axi-cel was manufactured for all 178 patients who underwent leukapheresis and administered to 170 (96%) patients. The primary efficacy analysis (data cut-off: March 18, 2021) was performed after 250 EFS events by blinded central assessment had been observed. The median follow-up was 23.0 months in the axi-cel arm and 21.2 months in the SOC arm. The mean age of patients was 57 years (SD = 12 years); 30% of the patients were aged 65 years or older. Overall, 74% of the study population had primary refractory disease and 26% had early relapse. Approximately one-quarter of patients in both treatment arms had achieved a best response of CR to first-line treatment.

The sponsor also provided long-term (≥ 4 year and ≥ 5 year) data from ZUMA-1, the pivotal multicentre, single-arm, registrational phase I and II study of axi-cel in patients with relapsed or refractory LBCL who had received at least 2 previous systemic therapies. Supportive safety data from ZUMA-1 were also considered.

Interpretation of Results

Efficacy

The ZUMA-7 trial demonstrated evidence of efficacy of axi-cel compared with SOC (salvage chemotherapy followed by HD-ASCT) in adult patients with primary refractory and early relapsed LBCL (< 12 months). EFS was improved for the axi-cel arm compared with the SOC arm. Secondary outcomes were supportive, with improvements observed for PFS and ORR. According to the clinical experts, the magnitude of the treatment effect is clinically meaningful. The interim analysis of OS did not reach statistical significance; the 95% CI was wide and included the possibility of appreciable OS benefit compared with SOC, as well as little-to-no difference. As the results are based on an interim analysis, there is the potential that the beneficial effect of axi-cel compared with SOC is overestimated. Results from the final analysis are required to determine with certainty the presence and magnitude of OS benefit. More mature OS data from the phase II cohort of the ZUMA-1 study for patients with LBCL treated with axi-cel in third line provide supporting evidence for the OS benefit of axi-cel in patients who had received at least 2 previous systemic therapies.^21,22 However, the single-arm design of ZUMA-1 and different patient population precludes drawing causal inferences.

The clinical experts consulted by CADTH indicated that OS is the most important efficacy outcome to assess treatment effect between axi-cel and the SOC arms, and longer survival was an important priority indicated in the patient input. The interim analysis of OS is premature given the heavy censoring around 18 months. Overall, the difference in OS was not statistically significant. As the FDA review has also noted, while the difference in OS between the 2 arms is not statistically significant, the direction of the observed treatment effect is consistent with the EFS and PFS data.⁴² It is also noted that results of the OS analysis may be confounded by the fact that 55% of the patients in the SOC arm received autologous CD19-directed CAR T therapy after experiencing an event as crossover, which was not built into the main statistical analyses.⁴² As OS data are immature and the current results are based on an interim analysis, it is possible that the benefit of axi-cel compared with SOC is overestimated, although the degree to which this may be true is uncertain.^17,18

HRQoL was also an outcome indicated as important by patients. However, findings were uncertain due to the large amount of missing data. It is thus unclear, based on these data, if axi-cel affords better QoL compared with SOC in patients with relapsed or refractory LBCL. Data on resource utilization, another important outcome as indicated by the clinical experts, were limited to acute care, and thus may not completely capture resource use associated with axi-cel.

Despite high initial response to first-line therapy, about 40% of patients with LBCL are refractory or will relapse and require second-line therapy.⁴ The current SOC for relapsed or refractory LBCL is salvage chemotherapy followed by HDT and ASCT if there is an adequate response to chemotherapy. However, a considerable proportion of patients do not benefit from current SOC second-line therapy. Many patients, even if fit enough for ASCT, ultimately cannot undergo HDT-ASCT for reasons including chemotherapy-resistant disease, treatment-emergent toxicities, or inability to mobilize stem cells. Poor outcomes have been observed in these patients, with 5-year survival below 50%. Although higher risk is associated with disease that is refractory or relapses within 12 months of first-line therapy, only 10% of all patients with relapsed or refractory LBCL are estimated to have long-term survival following ASCT.⁴ Therefore, there is an important need for effective alternative second-line therapies, including those with a mechanism of action independent of chemotherapy sensitivity, whether the patient relapses before or after 12 months of first-line therapy or is deemed eligible or ineligible to receive ASCT before starting salvage chemotherapy. In the ZUMA-7 trial, the most common reason for not proceeding with HD-ASCT in the SOC arm was due to the lack of response to salvage chemoimmunotherapy, indicating the chemorefractory nature of the study population. Axi-cel has a mechanism of action independent of chemotherapy sensitivity that may benefit patients with LBCL who require second-line therapy. No therapies are currently approved for the definitive management of LBCL in the second-line setting for potentially transplant-eligible patients who are either primary refractory or have early treatment failure to front-line therapy. The clinical experts consulted by CADTH also indicated that they thought axi-cel could fit well as a second-line treatment and replace ASCT for most patients. The clinical experts indicated that they would expect patient outcomes to be better when they receive this therapy earlier in the course of disease.

Of note, this CADTH Reimbursement Review was conducted before issuance of Health Canada NOC and the scope was based on the anticipated indication. The anticipated Health Canada indication when the submission was received by CADTH was for the treatment of adult patients with relapsed or refractory LBCL. The final approved indication is for the treatment of adult patients with DLBCL or HGBL that is refractory to first-line chemoimmunotherapy or that relapses within 12 months of first-line chemoimmunotherapy. The sponsor’s requested reimbursement criteria are for the treatment of adult patients with relapsed or refractory LBCL who are candidates for ASCT, which aligns with the trial population in ZUMA-7. However, the clinical experts expressed concern about using ASCT eligibility as a criterion for CAR T-cell therapy eligibility. The clinical experts noted that there are no standard criteria to determine ASCT eligibility; criteria vary widely across treatment centres depending on local clinical practices and resources. According to the clinical experts, only approximately 40% of patients are intended for ASCT. A recent Canadian study of real-world eligibility for second-line CAR T-cell therapy in patients with LBCL found that real-world eligibility for second-line CAR T-cell therapy may be as high as 65%.¹³ Although only 14% of patients met all the ZUMA-7 study inclusion criteria, as many as 65% of patients progressing within 12 months of chemoimmunotherapy had adequate performance status to be considered potentially eligible for second-line CAR T-cell therapy.¹³ Therefore, restricting axi-cel eligibility only to transplant-eligible patients would exclude a large proportion of patients. However, the proportion of patients who can be treated with axi-cel in real-world practice is uncertain and will likely depend on both patient-related and system-related factors, and although the definition of CAR T-cell therapy eligibility might differ from the definition of ASCT eligibility, not all patients would be considered suitable for axi-cel treatment in the second-line setting.

In the ZUMA-7 trial, only 35% of the patients randomized to the SOC arm underwent transplant. In comparison, almost 3 times as many patients (94%) randomized to the axi-cel arm underwent definitive treatment with CAR T-cell infusion. In addition to lack of chemotherapeutic response as a main reason for ineligibility to proceed with HDT-ASCT, the SOC regimen (preceded by salvage chemotherapy) is a multistep and arduous treatment regimen that most patients find difficult to tolerate. According to the clinical experts consulted by CADTH, single-infusion CAR T-cell therapy represents a less burdensome treatment option for patients and its use in the second line, particularly for patients with prognostic factors indicative of poor response to second-line SOC, could substantially increase the proportion of patients able to receive curative-intent treatment at first progression of LBCL.

Harms

All patients in the axi-cel and SOC arms had at least 1 TEAE. Serious TEAEs were reported in similar proportions of patients in the axi-cel and SOC arms. Grade 3 or higher TEAEs occurred in 91% and 83% of patients in the axi-cel and SOC arms, respectively, including 7 patients and 2 patients, respectively, with grade 5 TEAEs (excluding PD). Evidence for the safety of axi-cel was also reported from ZUMA-1. Safety data reported from ZUMA-1 were based on 108 patients treated with axi-cel; at the data cut-off date of March 18, 2021, patients in ZUMA-1 had the opportunity to be followed up for at least 54 months after their infusion of axi-cel. Nine patients (5%) in the axi-cel arm were re-treated. No new safety concerns were identified after re-treatment with axi-cel.

Differences in the TEAE profile between the 2 treatment arms were generally consistent with the AE profiles of the 2 treatments (e.g., CRS is a well-recognized toxicity of CAR T-cell therapies). TEAEs noted after axi-cel infusion in ZUMA-7 are consistent with those observed with the AE profile of other anti-CD19 CAR T therapies.⁵⁴ Patients treated with CAR T-cell therapies commonly experience CRS, associated with the use of monoclonal antibodies and adoptive cell therapies that activate lymphocytes. The condition results from the release of cytokines from cells targeted by antibodies, immune effector cells recruited to the tumour area, and the patient’s immune cells activated during this process. CRS is characterized by high fevers, hemodynamic instability, multiorgan dysfunction, and neurologic events such as encephalopathy, delirium, seizures, and cerebral edema, with most patients requiring treatment in hospital with highly specialized and intensive care.⁵⁴ In ZUMA-7, CRS occurred in 92% of patients (grade ≥ 3, 6%), and was generally manageable and resolved for all patients, with a median duration of 7.0 days. While the overall incidence of CRS observed in ZUMA-7 is similar to that observed in ZUMA-1 (92% and 94%, respectively), the incidence of grade 3 or greater CRS appeared lower in ZUMA-7 compared with ZUMA-1 (7% and 13%, respectively). This difference in the rate of grade 3 or greater CRS may be due to earlier intervention in ZUMA-7, or it may be related to differences in the study population (second-line versus third- or later-line population in the ZUMA-7 versus ZUMA-1 trials, respectively).⁴²

Another AE of special interest was neurotoxicity. The ZUMA-7 trial reported 102 patients (60%) in the axi-cel arm and 33 patients (20%) in the SOC arm who had at least 1 treatment-emergent neurologic event, including 36 patients (21%) and 1 patient (1%), respectively, with grade 3 or higher neurologic events. However, the FDA reanalysis of the safety data found that 124 (74%) patients had 1 or more neurologic toxicity events considered to be related to axi-cel, and that 42 patients (25%) experienced grade 3 or higher events.⁴² Of the 124 patients who experienced neurotoxicity, neurologic toxicity occurred before CRS in 1 (0.8%) patient, during CRS in 110 (89%) patients, and after CRS in 11 (9%) patients. Two (2%) patients had neurotoxicity without CRS.⁴² Most of the neurotoxicity events reported in the trial were manageable and reversible. The clinical experts indicated that in clinical practice treating physicians would be vigilant to the early signs of neurotoxicity and can manage these in most cases.

Axi-cel is administered as a single infusion; no patients discontinued treatment due to TEAEs in the axi-cel arm. Two patients in the SOC arm discontinued treatment (due to TEAEs of grade 4 acute kidney injury and grade 1 blood stem cell harvest failure). Overall, 3 patients in the SOC arm were unable to tolerate protocol-specified chemotherapy due to toxicities, resulting in treatment discontinuation: 2 patients discontinued R-ICE and R-DHAP, respectively, due to renal impairment, and 1 patient was unable to tolerate R-GDP. One patient was unable to complete HSCT due to stem cell harvest failure. Similarly, because axi-cel is administered as a single infusion, no patients in the axi-cel arm had a dose interruption or reduction due to TEAEs. In the SOC arm, 30 patients had at least 1 TEAE that required dose interruption or reduction to salvage chemotherapy.

The patient groups indicated that having access to treatments with fewer side effects was very important. Although the safety profile of axi-cel is consistent with its known mechanism of action, the lack of long-term safety data from ZUMA-7 remains an important limitation as the risk of certain AEs that may develop over a longer time period after treatment, such as secondary malignancies, is not well-characterized.

Conclusions

Based on data from the ZUMA-7 trial, treatment with the anti-CD19 CAR T-cell axi-cel led to an improvement in EFS and PFS compared with SOC HDT-ASCT as second-line treatment in ASCT-eligible patients with relapsed or refractory LBCL within 12 months of first-line chemoimmunotherapy. It is not yet clear whether EFS benefits will translate to improved OS as the data remain immature, and follow-up is ongoing. As the OS results are based on an interim analysis, there is a risk that the effect of axi-cel compared with SOC on survival is overestimated. The risks associated with axi-cel treatment are consistent with its mechanism of action and acceptable given the high-risk patient population, but it must be administered in a specialized centres equipped to manage toxicities related to CAR T-cell therapy. Although the most common toxicities were similar in the ZUMA-7 and ZUMA-1 trials, the lack of long-term safety and efficacy data from the ZUMA-7 trial remains an important limitation. A longer follow-up time is needed to better establish the survival benefit and long-term safety of axi-cel as relapse of lymphoma and potential late toxicities of CAR T-cell therapy, such as secondary malignancies, may emerge with longer follow-up. Uncertainties remain regarding the implementation of CAR T-cell therapy and the support system needed to optimize timely access and deliverability of axi-cel in the real-world setting.

References

1.Canadian Cancer Society. Non-Hodgkin Lymphoma. 2022; https://www.lymphoma.ca/lymphoma/non-hodgkin-lymphoma/. Accessed 2022 Aug 12.

2.Canadian Cancer Society. Non-Hodgkin Lymphoma statistics. 2022; https://cancer.ca/en/cancer-information/cancer-types/non-hodgkin-lymphoma/statistics. Accessed 2022 Sep 1.

3.Armitage JO, Gascoyne RD, Lunning MA, Cavalli F. Non-Hodgkin lymphoma. Lancet. 2017;390(10091):298-310. PubMed

4.Shafey M, Savage KJ, Skrabek P, Elsawy M, Bosch M, Kuruvilla J. Canadian Evidence-based Guideline for the Treatment of Relapsed/Refractory Diffuse Large B-Cell Lymphoma. Mississauga (ON): Lymphoma Canada; 2021: https://www.lymphoma.ca/resources/healthcare-professionals/dlbcl-diffuse-large-b-cell-lymphoma/. Accessed 2022 Aug 15.

5.Lymphoma Canada. Diffuse Large B cell Lymphoma (DLBCL). 2022; https://www.lymphoma.ca/diffuse-large-b-cell-lymphoma/. Accessed 2022 Aug 12.

6.Martelli M, Ferreri AJ, Agostinelli C, Di Rocco A, Pfreundschuh M, Pileri SA. Diffuse large B-cell lymphoma. Crit Rev Oncol Hematol. 2013;87(2):146-171. PubMed

7.NCCN Guidelines for Patients: Diffuse Large B-cell Lymphoma. Plymouth Meeting (PA): National Comprehensive Cancer Network; 2020.

8.Lymphoma Clinical Practice Guidelines. Edmonton (AB): Alberta Health Services; 2021: https://www.albertahealthservices.ca/assets/info/hp/cancer/if-hp-cancer-guide-lyhe002-lymphoma.pdf. Accessed 2022 Aug 11.

9.Prica A, Baldassarre F, Hicks LK, et al. Rituximab in Lymphoma and Chronic Lymphocytic Leukemia: A Clinical Practice Guideline. (Guideline 6-8 Version 3). Toronto (ON): Cancer Care Ontario; 2015: https://www.cancercareontario.ca/en/guidelines-advice/types-of-cancer/721. Accessed 2022 Aug 15.

10.BC Cancer Protocol Summary for Treatment of Lymphoma with DOXOrubicin, Cyclophosphamide, vinCRIStine, predniSONE and riTUXimab (CHOP-R). [Victoria (BC)]: BC Cancer; 2022: http://www.bccancer.bc.ca/chemotherapy-protocols-site/Documents/Lymphoma-Myeloma/LYCHOPR_Protocol.pdf. Accessed 2022 Aug 15.

11.Coiffier B, Sarkozy C. Diffuse large B-cell lymphoma: R-CHOP failure-what to do? Hematology Am Soc Hematol Educ Program. 2016;2016(1):366-378. PubMed

12.Rovira J, Valera A, Colomo L, et al. Prognosis of patients with diffuse large B cell lymphoma not reaching complete response or relapsing after frontline chemotherapy or immunochemotherapy. Ann Hematol. 2015;94(5):803-812. PubMed

13.Puckrin R, Stewart DA, Shafey M. Real-World Eligibility for Second-Line Chimeric Antigen Receptor T Cell Therapy in Large B Cell Lymphoma: A Population-Based Analysis. Transplant Cell Ther. 2022;28(4):218.e211-218.e214. PubMed

14.Locke FL, Miklos DB, Jacobson CA, et al. Axicabtagene Ciloleucel as Second-Line Therapy for Large B-Cell Lymphoma. N Engl J Med. 2022;386(7):640-654. PubMed

15.Clinical Study Report: KTE-C19-107. A phase 3, randomized, open-label study evaluating the efficacy of axicabtagene ciloleucel versus standard of care therapy in subjects with relapsed/refractory diffuse large B-cell lymphoma (ZUMA-7) [internal sponsor's report]. Santa Monica (CA): Kite Pharma, Inc.; 2021 Sep 14.

16.Maringwa JT, Quinten C, King M, et al. Minimal important differences for interpreting health-related quality of life scores from the EORTC QLQ-C30 in lung cancer patients participating in randomized controlled trials. Support Care Cancer. 2011;19(11):1753-1760. PubMed

17.Guyatt GH, Briel M, Glasziou P, Bassler D, Montori VM. Problems of stopping trials early. BMJ. 2012;344:e3863. PubMed

18.Montori VM, Devereaux PJ, Adhikari NK, et al. Randomized trials stopped early for benefit: a systematic review. JAMA. 2005;294(17):2203-2209. PubMed

19.Zhu J, Yang Y, Tao J, et al. Association of progression-free or event-free survival with overall survival in diffuse large B-cell lymphoma after immunochemotherapy: a systematic review. Leukemia. 2020;34(10):2576-2591. PubMed

20.Jacobson C, Locke FL, Ghobadi A, et al. Long-Term (>4 Year and >5 Year) Overall Survival (OS) By 12- and 24-Month Event-Free Survival (EFS): An Updated Analysis of ZUMA-1, the Pivotal Study of Axicabtagene Ciloleucel (Axi-Cel) in Patients (Pts) with Refractory Large B-Cell Lymphoma (LBCL) [conference abstract]. Blood. 2021;138(Suppl 1):1764.

21.Jacobson C, Locke FL, Ghobadi A, et al. 10 - Long-Term (5 Year) Overall Survival in Zuma-1, the Pivotal Study of Axicabtagene Ciloleucel (Axi-Cel) in Patients with Refractory Large B-Cell Lymphoma (LBCL) [conference abstract]. Transplant Cell Ther. 2021;28(3S):S9-S10.

22.Jacobson C, Locke FL, Ghobadi A, et al. Long-Term (4- and 5- Year) Overall Survival in ZUMA-1, the Pivotal Study of Axicabtagene Ciloleucel in Patients with Refractory Large B-Cell Lymphoma [conference poster]. In: Drug Reimbursement Review sponsor submission: Yescarta® (axicabtagene ciloleucel), cell suspension in patient-specific single infusion bag, target of 2 × 106 chimeric antigen receptor (CAR)-positive viable T cells per kg body weight with a maximum of 2 × 108 CAR-positive viable T cells, for intravenous infusion. Mississauga (ON): Gilead Sciences Canada, Inc.; 2021.

23.Axicabtagene ciloleucel for large B-cell lymphoma: Clinical Report. (CADTH optimal use report vol. 9, no. 1c). Ottawa (ON): CADTH; 2019: https://www.cadth.ca/sites/default/files/pdf/car-t/ct0002-clinical-report-redacted.pdf. Accessed 2022 Sep 1.

24.Sehn LH, Salles G. Diffuse Large B-Cell Lymphoma. N Engl J Med. 2021;384(9):842-858. PubMed

25.Rosenwald A, Bens S, Advani R, et al. Prognostic Significance of MYC Rearrangement and Translocation Partner in Diffuse Large B-Cell Lymphoma: A Study by the Lunenburg Lymphoma Biomarker Consortium. J Clin Oncol. 2019;37(35):3359-3368. PubMed

26.Willenbacher E, Willenbacher W, Weger R, Dominik W, Manzl C, Brunner A. Patients with double/triple copy number gains on C-MYC, BCL2, and/or BCL6 treated with standard chemotherapy have a similarly poor prognosis than those with high-grade B cell lymphoma with C-MYC and BCL2 and/or BCL6 rearrangements: a single-center experience on a consecutive cohort of large B cell lymphomas. Ann Hematol. 2020;99(9):2125-2132. PubMed

27.Swerdlow SH, Campo E, Pileri SA, et al. The 2016 revision of the World Health Organization classification of lymphoid neoplasms. Blood. 2016;127(20):2375-2390. PubMed

28.van der Poel MW, Oerlemans S, Schouten HC, et al. Quality of life more impaired in younger than in older diffuse large B cell lymphoma survivors compared to a normative population: a study from the population-based PROFILES registry. Ann Hematol. 2014;93(5):811-819. PubMed

29.Ma Q, Bailey A, Milloy N, Butcher J, Quek RG, Johnson PC. Real-World Health-Related Quality of Life in Patients with Diffuse Large B-Cell Lymphoma: Comparisons with Reference Populations and By Line of Therapy. Blood. 2021;138(S1):4111.

30.Gisselbrecht C, Glass B, Mounier N, et al. Salvage regimens with autologous transplantation for relapsed large B-cell lymphoma in the rituximab era. J Clin Oncol. 2010;28(27):4184-4190. PubMed

31.van Imhoff GW, McMillan A, Matasar MJ, et al. Ofatumumab Versus Rituximab Salvage Chemoimmunotherapy in Relapsed or Refractory Diffuse Large B-Cell Lymphoma: The ORCHARRD Study. J Clin Oncol. 2017;35(5):544-551. PubMed

32.Crump M, Kuruvilla J, Couban S, et al. Randomized comparison of gemcitabine, dexamethasone, and cisplatin versus dexamethasone, cytarabine, and cisplatin chemotherapy before autologous stem-cell transplantation for relapsed and refractory aggressive lymphomas: NCIC-CTG LY.12. J Clin Oncol. 2014;32(31):3490-3496. PubMed

33.Kondo E. Autologous Hematopoietic Stem Cell Transplantation for Diffuse Large B-Cell Lymphoma. J Clin Exp Hematop. 2016;56(2):100-108. PubMed

34.Friedberg JW. Relapsed/refractory diffuse large B-cell lymphoma. Hematology Am Soc Hematol Educ Program. 2011;2011:498-505. PubMed

35.Gisselbrecht C, Schmitz N, Mounier N, et al. Rituximab maintenance therapy after autologous stem-cell transplantation in patients with relapsed CD20(+) diffuse large B-cell lymphoma: final analysis of the collaborative trial in relapsed aggressive lymphoma. J Clin Oncol. 2012;30(36):4462-4469. PubMed

36.CADTH. Drug reimbursement review: Polatuzumab Vedotin (Polivy) for diffuse large B-cell lymphoma (DLBCL). 2021; https://www.cadth.ca/polatuzumab-vedotin-polivy-dlbcl-details. Accessed 2022 Sep 1.

37.Yescarta® (axicabtagene ciloleucel): cell suspension in patient-specific single infusion bag, target of 2 × 106 chimeric antigen receptor (CAR)-positive viable T cells per kg body weight with a maximum of 2 × 108 CAR-positive viable T cells, for intravenous infusion [product monograph]. Mississauga (ON): Gilead Sciences Canada, Inc.; 2021 Dec 15.

38.Summary Basis for Regulatory Action: Yescarta (axicabtagene ciloleucel). Silver Spring (MD): U.S. Food and Drug Administration; 2017: https://www.fda.gov/media/108788/download. Accessed 2022 Aug 10.

39.European Public Assessment Report (EPAR): Yescarta (axicabtagene ciloleucel). Amsterdam (NL): European Medicines Agency; 2022: https://www.ema.europa.eu/en/medicines/human/EPAR/yescarta. Accessed 2022 Aug 10.

40.McGowan J, Sampson M, Salzwedel DM, Cogo E, Foerster V, Lefebvre C. PRESS Peer Review of Electronic Search Strategies: 2015 guideline statement. J Clin Epidemiol. 2016;75:40-46. PubMed

41.Grey matters: a practical tool for searching health-related grey literature. Ottawa (ON): CADTH; 2019: https://www.cadth.ca/grey-matters. Accessed 2022 Jun 28.

42.BLA 125643/394 Clinical Review and Evaluation: Axicabtagene ciloleucel (Yescarta). Silver Spring (MD): U.S. Food and Drug Administration; 2022: https://fda.report/media/157687/April-1-2022-Clinical-Review-and-Evaluation-YESCART_0.pdf. Accessed 2022 Oct 18.

43.Neelapu SS, Locke FL, Bartlett NL, et al. Axicabtagene Ciloleucel CAR T-Cell Therapy in Refractory Large B-Cell Lymphoma. N Engl J Med. 2017;377(26):2531-2544. PubMed

44.Cheson BD, Fisher RI, Barrington SF, et al. Recommendations for initial evaluation, staging, and response assessment of Hodgkin and non-Hodgkin lymphoma: the Lugano classification. J Clin Oncol. 2014;32(27):3059-3068. PubMed

45.Pickard AS, Neary MP, Cella D. Estimation of minimally important differences in EQ-5D utility and VAS scores in cancer. Health Qual Life Outcomes. 2007;5:70. PubMed

46.National Cancer Institute Division of Cancer Treatment and Diagnosis. Common Terminology Criteria for Adverse Events (CTCAE) v4.0. [2010]; https://ctep.cancer.gov/protocoldevelopment/electronic_applications/ctc.htm#ctc_40. Accessed 2022 Sep 1.

47.Robins JM, Tsiatis AA. Correcting for Non-Compliance in Randomized Trials Using Rank Preserving Structural Failure Time Models. Commun Statist Theory Meth. 1991;20(8):22.

48.Robins JM, Finkelstein DM. Correcting for noncompliance and dependent censoring in an AIDS Clinical Trial with inverse probability of censoring weighted (IPCW) log-rank tests. Biometrics. 2000;56(3):779-788. PubMed

49.Addendum to the Primary Analysis Clinical Study Report: A phase 3, randomized, open-label study evaluating the efficacy of axicabtagene ciloleucel versus standard of care therapy in subjects with relapsed/refractory diffuse large B-cell lymphoma (ZUMA-7) (CSR KTE-C19-107) [internal sponsor's report]. Santa Monica (CA): Kite Pharma, Inc.; 2022 Mar 03.

50.Lee DW, Gardner R, Porter DL, et al. Current concepts in the diagnosis and management of cytokine release syndrome. Blood. 2014;124(2):188-195. PubMed

51.Ye X, Mahmud S, Skrabek P, Lix L, Johnston JB. Long-term time trends in incidence, survival and mortality of lymphomas by subtype among adults in Manitoba, Canada: a population-based study using cancer registry data. BMJ Open. 2017;7(7):e015106. PubMed

52.Locke FL, Ghobadi A, Jacobson CA, et al. Long-term safety and activity of axicabtagene ciloleucel in refractory large B-cell lymphoma (ZUMA-1): a single-arm, multicentre, phase 1-2 trial. Lancet Oncol. 2019;20(1):31-42. PubMed

53.Section 2.7 Clinical Summary. Section 2.7.4: Summary of Clinical Safety Axicabtagene Ciloleucel Yescarta [internal sponsor's report]. Santa Monica (CA): Kite Pharma Inc.; 2021 Sep 23.

54.Neelapu SS. Managing the toxicities of CAR T-cell therapy. Hematol Oncol. 2019;37 Suppl 1:48-52. PubMed

55.EORTC Quality of Life. EORTC QLQ-C30 FAQ Categories: Scoring. https://qol.eortc.org/faq-category/scoring/. Accessed 2021 Oct 01.

56.Cocks K, King MT, Velikova G, et al. Evidence-based guidelines for interpreting change scores for the European Organisation for the Research and Treatment of Cancer Quality of Life Questionnaire Core 30. Eur J Cancer. 2012;48(11):1713-1721. PubMed

57.Osoba D, Rodrigues G, Myles J, Zee B, Pater J. Interpreting the significance of changes in health-related quality-of-life scores. J Clin Oncol. 1998;16(1):139-144. PubMed

58.van Reenen M, Janssen B, Oppe M, Kreimeier S, Greiner W, Stolk E. EQ-5D-Y User Guide. Version 2.0. 2020; https://euroqol.org/publications/user-guides. Accessed 2022 Mar 06.

59.McClure NS, Sayah FA, Xie F, Luo N, Johnson JA. Instrument-Defined Estimates of the Minimally Important Difference for EQ-5D-5L Index Scores. Value Health. 2017;20(4):644-650. PubMed

60.Elsawy M, Chavez JC, Avivi I, et al. Patient-reported outcomes in ZUMA-7, a phase 3 study of axicabtagene ciloleucel in second-line large B-cell lymphoma. Blood. 2022;15:15. PubMed

61.Shah J, Shacham S, Kauffman M, et al. Health-related quality of life and utility outcomes with selinexor in relapsed/refractory diffuse large B-cell lymphoma. Fut Oncol. 2021;17(11):1295-1310. PubMed

62.Giesinger JM, Kieffer JM, Fayers PM, et al. Replication and validation of higher order models demonstrated that a summary score for the EORTC QLQ-C30 is robust. J Clin Epidemiol. 2016;69:79-88. PubMed

63.Cochrane T, Enrico A, Gomez-Almaguer D, et al. Impact of venetoclax monotherapy on the quality of life of patients with relapsed or refractory chronic lymphocytic leukemia: results from the phase 3b VENICE II trial. Leuk Lymphoma. 2022;63(2):304-314. PubMed

Appendix 1: Literature Search Strategy

Note that this appendix has not been copy-edited.

Clinical Literature Search

Overview

Interface: Ovid

Databases:

• MEDLINE All (1946 to present)

• Embase (1974 to present)

Note: Subject headings and search fields have been customized for each database. Duplicates between databases were removed in Ovid.

Date of search: July 7, 2022

Alerts: Bi-weekly search updates until project completion

Search filters applied: No filters were applied to limit the retrieval by study type

Limits: Conference abstracts: excluded

Table 24: Syntax Guide

Multi-Database Strategy

1. (axicabtagene* or Yescarta* or axi-cel or axicel or KTEC19* or KTE-C19* or fkc 876* or fkc876* or U2I8T43Y7R).ti,ab,ot,kf,hw,nm,rn.

2. 1 use medall

3. *Axicabtagene Ciloleucel/

4. (axicabtagene* or Yescarta* or axi-cel or axicel or KTEC19* or KTE-C19* or fkc 876* or fkc876*).ti,ab,kf,dq.

5. 3 or 4

6. 5 not (conference review or conference abstract).pt.

7. 6 use oemezd

8. 2 or 7

9. remove duplicates from 8

Clinical Trials Registries

ClinicalTrials.gov

Produced by the US National Library of Medicine. Targeted search used to capture registered clinical trials.

• [Search -- axicabtagene OR axicel OR “axi cel” OR yescarta OR ktec19 OR “kte c19” OR “fkc 876” OR fkc876]

WHO ICTRP

International Clinical Trials Registry Platform, produced by the WHO. Targeted search used to capture registered clinical trials.

• [Search terms -- axicabtagene* OR axicel* OR “axi cel” OR yescarta* OR ktec19 OR “kte c19” OR “fkc 876” OR fkc876]

Health Canada’s Clinical Trials Database

Produced by Health Canada. Targeted search used to capture registered clinical trials.

• [Search terms – axicabtagene, axicel, “axi cel”, yescarta, ktec19, “kte c19”, “fkc 876”, fkc876]

EU Clinical Trials Register

European Union Clinical Trials Register, produced by the European Union. Targeted search used to capture registered clinical trials.

• [Search terms -- axicabtagene* OR axicel* OR “axi cel” OR yescarta* OR ktec19 OR “kte c19” OR “fkc 876” OR fkc876]

Grey Literature

Search dates: June 24 to July 6, 2022

Keywords: Yescarta, axicabtagene, axi-cel, axicel, KTE-C19, KTEC19, CAR-T, chimeric antigen, lymphoma

Limits: None

Updated: No update

Relevant websites from the following sections of the CADTH grey literature checklist Grey Matters: A Practical Tool for Searching Health-Related Grey Literature were searched:

• Health Technology Assessment Agencies

• Health Economics

• Clinical Practice Guidelines

• Drug and Device Regulatory Approvals

• Advisories and Warnings

• Drug Class Reviews

• Clinical Trials Registries

• Databases (free)

• Internet Search

Appendix 2: Description and Appraisal of Outcome Measures

Note this appendix has not been copy-edited.

Aim

To describe the following outcome measures and review their measurement properties (validity, reliability, responsiveness to change, and MID):

• European Organisation for Research and Treatment of Cancer Quality of Life Questionnaire Core 30 (EORTC QLQ-C30)

• EQ-5D-5L

Findings

A focused literature search was conducted to identify the psychometric properties and minimally important difference (MID) of each of the stated outcome measures.

The findings about validity, reliability, responsiveness, and MID of each outcome measure are summarized in Table 25.

Table 25: Summary of Outcome Measures and Their Measurement Properties

EORTC QLQ-C30 = European Organisation for Research and Treatment of Cancer Quality of Life Questionnaire Core 30; EQ VAS = EuroQol-visual analogue scales; GHS = global health status; HRQoL = health-related quality of life; MCID = minimal clinically important difference; MID = minimal important difference; NHL = non-Hodgkin lymphoma; QoL = quality of life; SD = standard deviation; SF-36 = Short Form-36; VAS = visual analogue scale.

European Organisation for Research and Treatment of Cancer Quality of Life Questionnaire-C30

Description

The European Organization for Research and Treatment of Cancer (EORTC) Quality of Life Questionnaire Core 30, or EORTC QLQ-C30, is 1 of the most commonly used patient-reported outcomes (PRO) measures in oncology clinical trials.⁶² It is a multidimensional, cancer-specific, evaluative measure of health-related quality of life (HRQoL). It was designed specifically for the purpose of assessing changes in participants’ HRQoL in clinical trials, in response to treatment. The core questionnaire of the EORTC QLQ-C30 consists of 30 questions that are scored to create 5 multi-item functional scales, 3 multi-item symptom scales, 6 single-item symptom scales, and a 2-item quality of life (QoL) scale, as outlined in Table 26. Version 3.0 of the questionnaire is the most current version and has been in use since December 1997. It is available in 90 different languages and is intended for use in adult populations only. Notably, the global QoL scale is also known as the Global Health Status (GHS).⁵⁵

Table 26: Scales of EORTC QLQ-C30

EORTC QLQ-C30 = European Organisation for Research and Treatment of Cancer Quality of Life Questionnaire Core 30.

Scoring

The EORTC QLQ-C30 uses a 1-week recall period in assessing function and symptoms. Most questions have 4 response options (“not at all,” “a little,” “quite a bit,” “very much”), with scores on these items ranging from 1 to 4. For the 2 items form the global QoL scale, however, the response format is a 7-point Likert-type scale, with anchors between 1 (very poor) and 7 (excellent).⁵⁵

Raw scores for each scale are computed as the average of the items that contribute to a particular scale. This scaling approach is based upon the assumption that it is appropriate to provide equal weighting to each item that comprises a scale. There is also an assumption that, for each item, the interval between response options is equal (for example, the difference in score between “not at all” and “a little” is the same as “a little” and “quite a bit,” at a value of 1 unit). All the scales and single-item measures range in score from 0 to 100. Higher score for the functioning scales and global health status denotes a better level of functioning (i.e., a better state of the patient), while higher scores on the symptom and single-item scales indicate a higher level of symptoms (i.e., a worse state of the patient). Each raw scale score is converted to a standardized score that ranges from 0 to 100 using a linear transformation, with a higher score reflecting better function on the function scales, higher symptoms on the symptom scales, and better quality of life (i.e., higher scores simply reflect higher levels of response on that scale). According to the EORTC QLQ-C30 scoring algorithm, if there are missing items for a scale (i.e., the participant did not provide a response), the score for the scale can still be computed if there are responses for at least one-half of the items. In calculating the scale score, the missing items are simply ignored — an approach that assumes that the missing items have values equal to the average of those items for what the respondent completed.⁵⁵

Psychometric properties of the EORTC QLQ-C30 have not been assessed in patients with lymphoma, therefore its validity, reliability, and responsiveness to change in this patient population remain unknown.

Minimal Important Difference

Cocks et al.⁵⁶ used a systematic review of the literature and experts’ opinion to evaluate meaningful differences and magnitude of change in the QLQ-C30 scores. In a meta-analysis of 118 relevant papers (13.6% from US and Canada, 5.1% about hematological diseases) with timescales ranging from 4 days to 5 years, authors estimated trivial, small, medium, and large size classes for meaningful change in the scales. Medium and large changes could not be estimated for all scales due to insufficient data and response shift (i.e., psychological adaption of patients to their changing health status). Small differences were defined as subtle but nevertheless clinically meaningful changes (Table 27).

Table 27: Small Mean Differences of EORTC QLQ-C30 Subscales

EORTC QLQ-C30 = European Organisation for Research and Treatment of Cancer Quality of Life Questionnaire Core 30.

Among patients with lung, ovarian, or breast cancer, a 5-point change in the GHS/QoL subscale has been estimated as a clinically meaningful difference, whereas for other domains, a 5-point change from baseline and 10-point change in mean scores were estimated as a clinically meaningful difference.^57,63 In the ZUMA-7 trial, clinically meaningful changes were defined as 10 points for QLQ-C30.^16,56,57,60

EuroQoL 5-Dimensions 5-Levels Questionnaire

The EQ-5D-5L is a generic preference-based HRQoL instrument that has been applied to a wide range of health conditions and treatments. The EQ-5D-5L was developed by the EuroQoL Group as an improvement to the EQ-5D 3 level (i.e., EQ-5D-3L) to measure small and medium health changes and reduce ceiling effects. The EQ-5D-5L consists of the EQ-5D descriptive system and the EQ visual analogue scale (EQ VAS). The descriptive system comprises 5 dimensions: mobility, self-care, usual activities, pain/discomfort, and anxiety/depression. Each dimension is rated on 5 levels: level 1 “no problems,” level 2 “slight problems,” level 3 “moderate problems,” level 4 “severe problems,” and level 5 “extreme problems” or “unable to perform.” A total of 3,125 unique health states are possible, with 55555 (extreme problems in all of the dimensions) representing the worst health state and 11111 (no problems in any of the dimensions) representing the best state. The corresponding scoring of EQ-5D-5L health states is based on a scoring algorithm that is derived from preference data obtained from interviews using choice-based techniques (e.g., time trade-off) and discrete choice experiment tasks. The lowest score varies depending on the scoring algorithm used. The anchors are 0 (dead) and 1 (full health); however, negative values are also allowed to represent health states that a society considers worse than death. As an example, a Canadian scoring algorithm results in a score of –0.148 for health state 55555 (worst health state).⁵⁸

Another component of the EQ-5D-5L, the EQ VAS, asks respondents to rate their health on a visual analogue scale from 0 (worst health imaginable) to 100 (best health imaginable). The EQ VAS records the respondent’s self-rated health on a vertical VAS where the end points are labelled 0 and 100. The respondents are asked to mark an X on the point of the VAS that best represents their health on that day.

The EQ-5D-5L index and VAS scores can be summarized and analyzed as continuous data. Hence, the EQ-5D-5L produces 3 types of data for each respondent:

• a profile indicating the extent of problems on each of the 5 dimensions represented by a 5-digit descriptor, such as 11121 or 21143

• a population preference-weighted health index score based on the descriptive system

• a self-reported assessment of health status based on the EQ VAS

The EQ-5D-5L has been extensively validated across countries around the world and in various conditions.⁵⁸ However, the psychometric properties of the EQ-5D-5L have not been assessed in patients with lymphoma, therefore its validity, reliability, and responsiveness to change in this patient population remain unknown.

Minimal Important Difference

To estimate the MID values of the EQ-5D-5L for each country-specific scoring algorithm, a simulation-based approach based on instrument-defined single-level transitions has been used. The simulation-based instrument-defined generally accepted MID estimate (mean ± SD) for Canada is 0.056 ± 0.011.⁵⁹ In the ZUMA-7 trial, clinically meaningful changes were defined as 7 points for EQ-5D-5L VAS score, and 0.06 points for the EQ-5D-5L index for patients with lymphoma, based on relevant literature.^45,60 In a study of 239 patients with multiple myeloma, the MID was estimated as 0.08 for improvement and 0.10 for deterioration.⁶¹

Pharmacoeconomic Review

Executive Summary

The executive summary comprises 2 tables (Table 1 and Table 2) and a conclusion.

Table 1: Submitted for Review

CAR = chimeric antigen receptor; NOC = Notice of Compliance.

^aThe previous CAR T-cell therapy review of axicabtagene ciloleucel went through the interim CAR T-cell therapy review process in which recommendations were issued by the CADTH Health Technology Expert Review Panel (HTERP).

Table 2: Summary of Economic Evaluation

axi-cel = axicabtagene ciloleucel; ASCT = autologous stem cell transplant; BEAM = carmustine, etoposide, cytarabine, and melphalan; CAR = chimeric antigen receptor; EM = etoposide and melphalan; HDT = high-dose therapy; ICER = incremental cost-effectiveness ratio; LBCL = large B-cell lymphoma; OS = overall survival; PSM = partitioned survival model; QALY = quality-adjusted life-year; R-DHAP = rituximab + dexamethasone, cytarabine, and cisplatin; R-GDP = rituximab + gemcitabine, dexamethasone, and cisplatin; R-ICE = rituximab + ifosfamide, carboplatin, and etoposide; SOC = standard of care.

Conclusions

Evidence from the ZUMA-7 trial suggests that axicabtagene ciloleucel (axi-cel) is associated with clinically meaningful improvements in event-free survival (EFS) compared to standard of care (SOC) as second-line treatment in autologous stem cell transplant (ASCT)-eligible patients with relapsed or refractory large B-cell lymphoma (LBCL) within 12 months of first-line chemoimmunotherapy. It is not yet clear whether this EFS benefit will translate to improved overall survival (OS) as the data remain immature and follow-up is ongoing. As noted in the CADTH Clinical Review, as OS results are based on an interim analysis, there is a risk that the effect of axi-cel compared with SOC on survival is overestimated. This uncertainty is propagated into the submitted model given that the sponsor’s partitioned survival model (PSM) is informed by OS curves. The clinical trial population upon which the economic analysis was based consists of patients who are relatively stable and may not be generalizable to patients who are less stable or who do not meet the specific inclusion criteria of the trial.

In addition to the aforementioned limitations with the clinical evidence, CADTH identified several limitations with the sponsor’s economic submission. As part of the base-case reanalysis, CADTH addressed the uncertainties associated with long-term treatment efficacy by selecting an alternative extrapolation curve to inform OS for axi-cel; revising the pre-event (for the first 149 days of treatment) and postevent utility values; adjusting the distribution of subsequent therapies in the axi-cel arm to exclude investigational therapies; and applying a higher hazard of death among long-term survivors based on North American data. In CADTH’s base-case reanalysis, the incremental cost-effectiveness ratio (ICER) of axi-cel is $404,418 per QALY (incremental costs: $235,146; incremental QALYs: 0.58) compared with SOC, assuming 50% of patients who fail SOC would be treated with CAR T-cell therapies in the third-line setting. A price reduction of 45% would be necessary for axi-cel to be considered cost-effective at a willingness-to-pay (WTP) threshold of $50,000 per QALY.

The cost-effectiveness of axi-cel was sensitive to the assumptions regarding the selected parametric model distribution used to extrapolate OS for axi-cel. The CADTH base-case reanalysis estimated a smaller OS benefit with axi-cel when compared to the sponsor’s base case (i.e., incremental life-years: 1.50 [sponsor’s base case] versus 0.24 [CADTH’s reanalysis]), and this translated to a smaller difference in QALYs between axi-cel and SOC. Nonetheless, similar to the sponsor’s analysis, the majority of the QALY gains (71%) conferred by axi-cel in the CADTH reanalysis were derived from the period beyond which there are observed trial data. Furthermore, clinical uncertainties in the extrapolation period could not be adequately explored due to the inflexible modelling approach. This is particularly concerning when considering the uncertainties in the distribution of subsequent therapy. The cost-effectiveness results from both the sponsor and CADTH’s base-case reanalysis are informed by underlying OS curves from ZUMA-7 trial in which 56% of patients on SOC therapy received subsequent CAR T-cell therapies in the third-line setting. It is uncertain whether this reflects the proportion of patients who will have access to CAR T-cell therapy for third-line treatment.

The cost-effectiveness of axi-cel as second-line treatment in patients who are ineligible for ASCT and in patients who may not meet the specific inclusion or exclusion criteria of the ZUMA-7 trial, which reflects a component of the Health Canada indication, is unknown.

Stakeholder Input Relevant to the Economic Review

This section is a summary of the feedback received from the patient groups, registered clinicians, and drug plans that participated in the CADTH review process.

One patient group, Lymphoma Canada, provided input through data collected from 2 online patient surveys — the first gathered information from 97 patients with diffuse large B-cell lymphoma (DLBCL) in 2018 and the second gathered information from 23 patients with DLBCL in 2022. Overall, patients’ disease experience was influenced by the physical symptoms associated with LBCL (e.g., fatigue, body aches, thrombocytopenia, enlarged lymph nodes, neutropenia, and night sweats) and the psychosocial effect (e.g., fear of disease progression and relapse). The most important outcomes for patients included delaying disease progression and achieving long-term remission, with the ultimate objective of improving survival; reducing side effects from treatments; and maintaining quality of life. Of the 3 patients who had experience with the drug under review as third-line therapy, 1 had relapsed following treatment, 1 had been in remission for less than 6 months, and 1 had been in remission for 1 to 2 years. All 3 patients were away from home for more than 3 months to access treatment with axi-cel, and treatment access was described as a significant challenge. All 3 respondents reported thrombocytopenia as a treatment side effect, with 2 respondents reporting fever, neutropenia, anemia, nausea/vomiting, diarrhea, joint or muscle pain, and fatigue.

Registered clinician input was received from 3 groups: Lymphoma Canada, Ontario Health Hematology Cancer Drug Advisory Committee, and Cell Therapy Transplant Canada. According to clinician input, the current pathway of care for patients with relapsed or refractory LBCL is salvage chemotherapy, with an intent to proceed to ASCT in patients who are chemosensitive. The most common salvage chemotherapy regimen used in Canada is a combination of rituximab, gemcitabine, dexamethasone, and cisplatin (R-GDP). The most commonly used high-dose therapy (HDT) regimens were reported to be a combination of etoposide and melphalan (EM) or a combination of carmustine (BCNU), etoposide, cytarabine (Ara-C), and melphalan (BEAM). Clinicians noted that although ASCT is the best current curative option for patients with relapsed or refractory LBCL, it is an intensive intervention performed at a limited number of centres with known rates of relapse and multiple late toxicities. Clinicians indicated that axi-cel is currently available in Canada for patients with relapsed or refractory LBCL after 2 lines of systemic therapy and that ZUMA-7 explores the earlier use of axi-cel as second-line treatment. Across the feedback received, axi-cel was unanimously expected to shift the current treatment paradigm by replacing intensive chemoimmunotherapy and ASCT as the new preferred second-line treatment for patients with relapsed or refractory LBCL. Furthermore, it was noted that axi-cel should be eligible to the group of population (e.g., older age, comorbidities, Eastern Cooperative Oncology Group Performance Status [ECOG PS] score of 2) who might not be expected to tolerate ASCT. To assess treatment response to axi-cel, the clinician groups noted a variety of response outcomes may be used (e.g., overall response rate, complete response rate, EFS, progression-free survival, OS) and may involve restaging CT and PET/CT scans.

CADTH-participating drug plans highlighted several implementation and economic considerations for axi-cel. Plans noted concerns that the existing capacity may not be able to meet the anticipated demand in Canada. Given the requirement for specialized and accredited centres where the therapy can be administered, accessing CAR T-cell therapy may require interprovincial travel and, without full coverage of interprovincial reimbursement, may impact equitable access across Canada. The evidence surrounding CAR T-cell re-treatment was noted to be another area of uncertainty.

Several of these concerns were addressed in the sponsor’s model:

• The PSM was informed by the EFS and OS curves, 2 outcomes that are valued by patients.

• The impact of disease and treatment on patients’ quality of life was captured with utility values. Adverse events (AEs) were incorporated as disutilities within the analyses.

• The SOC modelled by the sponsor reflected the current treatments available to patients with relapsed or refractory LBCL who undergo ASCT (including salvage chemotherapy and HDT regimens).

In addition, CADTH addressed some of these concerns as follows:

• A scenario analysis exploring re-treatment based on the fact that 5% of patients who received axi-cel in the ZUMA-7 trial were re-treated.

CADTH was unable to address the following concerns raised from stakeholder input:

• Clinical data were not available for the broader Health Canada–indicated population (i.e., not eligible for ASCT) who may benefit from CAR T-cell therapy. As such, the modelled population in the cost-effectiveness analysis and budget impact analysis (BIA) excludes this subgroup of patients.

• Capacity is not explicitly considered in the model. The sponsor assumes all patients with relapsed or refractory LBCL have access to axi-cel if required and that the manufacturing time is similar to that observed in the trial.

• Accessing axi-cel may require interprovincial travel. These costs were not considered in the analysis given heterogeneity across provinces in terms of their policy for interprovincial billings. Furthermore, given the public payer perspective, patient-borne interprovincial travel costs were not considered as it was considered outside the scope of this review’s perspective.

Economic Review

The current review is for axi-cel (Yescarta) for adult patients with relapsed or refractory LBCL who are candidates for ASCT.

Economic Evaluation

Summary of Sponsor’s Economic Evaluation

Overview

Axi-cel is indicated for the treatment of adult patients with DLBCL or high-grade B-cell lymphoma (HGBL) that is refractory to first-line chemoimmunotherapy or that relapses within 12 months of first-line chemoimmunotherapy. The sponsor submitted a cost-utility analysis of axi-cel compared with SOC in the treatment of adult patients with relapsed or refractory LBCL who are candidates for ASCT.¹ The modelled population aligned with the reimbursement request criteria but differed from its Health Canada indication. SOC consisted of salvage chemoimmunotherapy followed by HDT and ASCT in responders. Salvage chemoimmunotherapy was defined as a basket of chemotherapy regimens including R-GDP, R-ICE (rituximab plus ifosfamide, carboplatin, and etoposide), and R-DHAP (rituximab plus dexamethasone, high-dose cytarabine, and cisplatin. HDT was defined as a basket of drug regimens including EM and BEAM.¹ For patients who failed their second-line treatment, the sponsor assumed 90% would proceed to subsequent therapy. For patients who received axi-cel, subsequent treatment consisted of enrolment in a clinical trial for investigational treatment (40%), salvage chemoimmunotherapy (i.e., R-GDP, R-DHAP, and R-ICE; 20%), radiotherapy (20%), or ASCT (10%), whereas patients who failed SOC proceeded to CAR T-cell therapies (50%) or radiotherapy (40%).

Axi-cel is a CD19-directed genetically modified autologous T-cell immunotherapy individually prepared from patients’ peripheral blood mononuclear cells. It is available as a cell suspension for infusion containing a target dose of 2 × 10⁶ CAR T cells/kg body weight (range, 1 × 10⁶ cells/kg to 2.4 × 10⁶ cells/kg), with a maximum of 2 × 10⁸ CAR T cells for patients weighing 100 kg or more.² It is provided as a single-dose, one-time treatment.² The sponsor’s submitted price of axi-cel is $485,021 per infusion,¹ not including costs associated with pretreatment (i.e., leukapheresis, bridging therapy, conditioning chemotherapy) and postinfusion management (i.e., ICU and non-ICU inpatient stay). The comparator for this analysis consisted of SOC, with treatment costs divided into 3 phases: salvage chemoimmunotherapy, HDT, and ASCT. Based on the proportion of patients receiving R-GDP (80%), R-ICE (10%), and R-DHAP (10%), the drug cost of salvage chemotherapy over a mean of 2.3 treatment cycles was estimated by the sponsor to be $10,011. The drug cost of HDT, derived from either 1 cycle of EM (80%) or 1 cycle of BEAM (20%), was estimated to be $3,529.¹ The cost of ASCT was $77,956 based on the Ontario interprovincial billing rates for an adult ASCT with a hospital stay longer than 72 hours, which includes all facility costs including inpatient and diagnostic costs.³

The outcomes of interest were QALYs and life-years. The analysis takes the perspective of a public health care payer. The time horizon in the base case was specified by the sponsor as a lifetime time horizon (50 years) with monthly cycles. The discount rate for costs and outcomes was 1.5% per year.¹

Model Structure

The sponsor used a PSM to capture all costs and outcomes associated with axi-cel and SOC. The model included 3 health states: event-free, postevent, and death (Figure 1).¹ The model considered the pretreatment period with patients who are considering second-line treatment entering the PSM in the event-free health state. The proportion of patients with event-free, postevent, and death were estimated over time based on the OS and EFS curves, informed by the ZUMA-7 trial.¹ The proportion of patients with progressed disease (i.e., postevent state) was estimated as the difference between the proportion of patients who are living (estimated from the OS curve) and the proportion of patients who are event-free (estimated from the EFS curve).¹ In other words, from the event-free state, patients can transition to the death state or progress to the postevent state where they can remain until transitioning to the death state (i.e., patients cannot go back to the event-free health state). Events were defined as disease progression per Lugano Classification, commencement of new lymphoma therapy (including stem cell transplant in the axi-cel arm without axi-cel–induced response or re-treatment of axi-cel), or death from any cause.⁴ Patients who experienced an event that was not death may transition to third-line therapy as determined by the time-to-next-treatment curve derived from the ZUMA-7 trial data.¹ If patients were in the EFS state for more than 5 years, they were assumed to be in long-term remission.¹

Model Inputs

Baseline patient characteristics were derived from ZUMA-7, a phase III, randomized, open-label multicentre clinical trial evaluating the efficacy of axi-cel versus SOC in patients with relapsed or refractory LBCL (i.e., age = 57 years, weight = 84 kg; proportion male = 66%).¹

The clinical outcomes informing the model (i.e., OS, EFS) were those reported by the ZUMA-7 trial, based on the intention-to-treat population. Survival data from the ZUMA-7 trial (data cut-off date: March 18, 2021) were extrapolated to derive long-term survival estimates of OS, EFS, and time to next treatment to inform the economic model.¹ The data were fitted by different statistical methods (i.e., parametric curve, mixture cure model [MCM], spline model) and different parametric functions. Under the assumption that both axi-cel treatment and ASCT may be curative, the MCM was selected for use in the sponsor’s base case. This statistical method estimates a likelihood of “cure” (i.e., cure fraction) in which the survival outcomes for this group, compared to the group who does not achieve long-term remission, would be different. The OS for the “cured” group was informed by adjusting pre-2020 age- and sex-matched Canadian mortality rates,¹ with an additional risk of excess mortality related to the disease (hazard ratio [HR] = 1.09).⁵ The survival function for the noncured group was evaluated with a diversity of parametric models with the statistical model selected based on clinical plausibility and statistical fit. The sponsor used a generalized gamma hazard to model the OS for the uncured fraction (i.e., patients who do not achieve long-term remission) in both the axi-cel and SOC arm. An MCM was also selected to estimate EFS in which, rather than estimating the cure fraction, the model predicted the likelihood of remaining event-free (i.e., event-free fraction). In the sponsor’s base case, log-logistic and exponential distributions were selected for the EFS for axi-cel and SOC, respectively.¹

Utility data for the event-free health state were derived directly from the EQ-5D-5L index data collected in the ZUMA-7 trial with Canadian tariffs applied.¹ Treatment-specific utilities were applied to the event-free state during the first month for axi-cel and the first 3 months for SOC. The sponsor assumed the utility for the postevent state would be equal to the utility derived from the JULIET trial.⁶ Moreover, the sponsor assumed that the utility values for patients who survived for at least 5 years would be equal to the sex-matched general population values, with an age-related decrement incorporated.^1,7 Refer to Table 11 for the health state utility values used in the sponsor’s base case.

The model included AEs grade 3 or higher as well as those graded 2 or higher for AEs of special interest, including cytokine release syndrome (CRS), neurologic events, and hypogammaglobulinemia.¹ Treatment-specific AE rates were informed from the ZUMA-7 trial.⁸ The value of the utility decrements was informed from published literature and applied as a one-time decrement to the first model cycle.^9-12 It was assumed that AEs would have no further impact on costs as these would be managed during the initial hospitalization period.

Costs captured in the model included drug acquisition, administration, monitoring, disease management, subsequent therapy, and end-of-life care. Drug acquisition costs were sourced from the Ontario Drug Benefit Formulary,¹³ CADTH’s previous pharmacoeconomic reviews,^14-19 and the Canadian Association of Provincial Cancer Agencies,²⁰ with dosing schedules based on the chemotherapy regimen monographs from Cancer Care Ontario²¹ and the Alberta Health Services clinical practice guideline on lymphoma.²² Costs specific to axi-cel included leukapheresis, bridging therapy, conditioning chemotherapy, and drug acquisition and administration (Table 12). The sponsor considered the costs typically associated with the ongoing monitoring, which were obtained from the Canadian Institute for Health Information’s Patient Cost Estimator for an inpatient hospitalization for malignant lymphoma.²³ For patients receiving SOC, treatment costs included salvage chemoimmunotherapy, HDT, and ASCT (Table 12). The salvage chemotherapy and HDT regimen were assumed to be administered on an outpatient basis. The total weighted costs of subsequent therapy differed by prior treatment (e.g., those who received axi-cel or SOC as second line) and were applied as a one-off cost at the time of initiation of third-line therapy based on the TTNT curve (Table 14). Treatment monitoring costs and health care resource use costs were sourced from the Ontario Ministry of Health Schedule of Benefits for laboratory and physician services,²⁴ whereas end-of-life costs were based on the average cost of care for patients with terminal lymphoma from the Ontario Cancer Registry.²⁵

Summary of Sponsor’s Economic Evaluation Results

The sponsor conducted a probabilistic analysis based on 1,000 simulations. Discrepancies were observed between the deterministic (Table 13) and probabilistic results (Table 3). This discrepancy was due to a divergence in the predicted life-years as the deterministic analysis estimated fewer life-years associated with SOC compared to the probabilistic analysis (deterministic analysis: 9.96 versus probabilistic analysis: 10.28). Upon further investigation, CADTH noted that the source of this discrepancy lay in the definition of the probabilistic distributions associated with the OS parametric distributions. The probabilistic findings are presented below.

Base-Case Results

In the sponsor’s requested reimbursement population, axi-cel was associated with an incremental QALY gain of 1.64 and an incremental cost of $232,469, resulting in an ICER of $141,757 per QALY compared with SOC (Table 3) under the assumption that 50% of patients who fail SOC would be treated with CAR T-cell therapies in third line.

The sponsor’s analysis predicted that axi-cel was associated with a longer duration of life than SOC (i.e., incremental life-years: 1.50). Given the duration of the ZUMA-7 trial (i.e., 37 months) in contrast to the model’s time horizon (i.e., 50 years), it is important to note that the majority of the QALY benefit (87%) realized by patients receiving axi-cel was derived from the period beyond which there are observed trial data (i.e., extrapolated period). Most of the QALYs gained by patients receiving axi-cel were realized in the event-free state (74%), whereas patients receiving SOC realized most of their QALY gains in the postevent state (60%). The key cost driver among patients receiving axi-cel was the cost of axi-cel acquisition, accounting for 91% of the total cost. Given the sponsor’s assumption that patients treated with SOC would be receiving CAR T-cell therapies in subsequent treatment lines (≥ third line), the main cost driver in the SOC arm that accounted for 85% of its total expected cost was the acquisition cost of CAR T-cell therapy. The sponsor’s submitted analysis is based on the publicly available prices for all drug treatments.

Table 3: Summary of the Sponsor’s Economic Evaluation Results (Probabilistic)

axi-cel = axicabtagene ciloleucel; ICER = incremental cost-effectiveness ratio; QALY = quality-adjusted life-year; SOC = standard of care.

Source: Sponsor’s pharmacoeconomic submission.¹

Additional results and calculations from the sponsor’s submitted economic evaluation base case are presented in Appendix 3.

Scenario Analysis Results

The sponsor assessed several model parameters and assumptions in probabilistic scenario analyses. The sponsor’s findings were generally sensitive to the scenario analyses tested. When adjusting the model for treatment switching to cell immunotherapy in the subsequent line of SOC, the ICER decreased to $40,554 per QALY. In contrast, when selecting the log-logistic distribution to extrapolate the OS curves for axi-cel and SOC, the ICER increased to $514,797 per QALY. All other scenarios resulted in ICERs between $58,898 and $402,350 per QALY.

CADTH Appraisal of the Sponsor’s Economic Evaluation

CADTH identified several key limitations to the sponsor’s analysis that have notable implications on the economic analysis:

• Exclusion of ASCT-ineligible patient population: The sponsor’s economic evaluation was specifically conducted using adult patients with relapsed or refractory LBCL who are candidates for ASCT, which is a subset of the Health Canada indication. The sponsor submitted a request for deviation to focus their analysis on this subpopulation (i.e., their reimbursement request). During the review process, the clinical expert panel convened by CADTH noted that the expected place in therapy for axi-cel would align more closely with its Health Canada indication (i.e., second-line treatment of adult patients with DLBCL or HGBL that is refractory to first-line chemoimmunotherapy or that relapses within 12 months of first-line chemoimmunotherapy, regardless of ASCT eligibility). From their perspective, patients who could potentially benefit from CAR T-cell therapy should not be excluded on the basis of ASCT eligibility. This is because the current SOC requires chemosensitivity which cannot be established up front; as such, patients must start salvage chemotherapy with the intent of receiving ASCT before a response can be observed, and those without a response would switch to receive CAR T cells in the third-line setting. They further noted challenges in implementing an initiation criterion based on ASCT eligibility (refer to the Issue for Consideration section for more information). The clinical experts indicated that any patient with adequate organ function and good performance status (i.e., Eastern Cooperative Oncology Group Performance Status [ECOG PS] ≤ 2) who could tolerate the known toxicities of CAR T-cell therapy (e.g., CRS) would be suitable for axi-cel treatment in real-world clinical practice. However, the clinical evidence informing the economic evaluation was based on the ZUMA-7 trial in which ASCT eligibility was an inclusion criterion. To evaluate the broader Health Canada population, clinical evidence on the population ineligible for ASCT would be required to inform the clinical effectiveness inputs within the economic model and alternative comparators would need to be considered (e.g., salvage chemotherapy, polatuzumab in combination with bendamustine and rituximab [Pola-BR]).

  • CADTH could not address this limitation given the lack of clinical data on the efficacy of axi-cel as second-line treatment in ASCT-ineligible patients. As such, the cost-effectiveness of axi-cel in this population is unknown.

• Inflexible modelling approach: The sponsor used a PSM to capture all costs and outcomes associated with axi-cel and SOC. Although the PSM has routinely been used to model oncology treatments, there are numerous issues documented with its use.^26,27 As there is no explicit modelling of transitions between states, subsequent therapy had no impact on OS within the submitted model. This is concerning given access to CAR T-cell therapy varies across jurisdictions. The OS curves are informed by the ZUMA-7 trial in which 56% of patients in the SOC group received subsequent CAR T-cell therapy in the third-line setting, while the sponsor assumed that 50% of patients who fail SOC would incur the cost of CAR T-cell therapy. Although it was possible to alter the distribution of subsequent therapies, doing so would only impact the cost of subsequent therapy and not of the underlying OS estimates, with the sole exception of the scenario analysis that incorporated the crossover adjustment (i.e., this would address a scenario whereby no patients receiving SOC would receive subsequent third-line CAR T-cell therapies). The clinical effects for alternate distribution of CAR T cells in third-line treatment is unknown.

  The PSM approach further introduces structural assumptions about the relationship between EFS and OS (i.e., non–mutually exclusive curves), which is potentially problematic since EFS and OS are likely dependent outcomes. Clinical expert opinion suggested that survival is linked to the occurrence of an event and thus the transition probability to death should vary for patients within the event-free state compared to those in the postevent state.

  • CADTH could not fully address this limitation given the structure of the model.

  • Both axi-cel and tisagenlecleucel are indicated in the third-line setting in Canada, although access to CAR T-cell therapy varies across jurisdictions. To reflect the jurisdictions that do not presently fund CAR T-cell therapy in the third-line setting, CADTH conducted scenario analyses. These scenario analyses made the following adjustment: (1) incorporated the sponsor’s crossover adjustments (i.e., applied an HR to adjust the OS curve for SOC); and (2) reassigned the proportion of patients who were expected to be on subsequent third-line therapy with CAR T-cell therapy to chemotherapy (i.e., excluded CAR T-cell therapies).

• Uncertainty in the extrapolation of OS for axi-cel: The sponsor selected an MCM based on a study showing the long-term OS of axi-cel in ZUMA-1 (a phase II, single-arm study of patients with LBCL [N = 101] treated with axi-cel in the third line), as well as a study that compared the fit of mixture cure and standard parametric models to the 12-month data cut-off of ZUMA-1.^1,28 Although the clinical expert feedback sought by CADTH noted the plausibility of cure in both interventions, there remains uncertainty in the sponsor’s predicted cure rate for axi-cel. The sponsor’s base case estimated a 53% cure fraction at 5 years for axi-cel. This is highly uncertain given the limited follow-up time of the ZUMA-7 trial. As noted in the CADTH clinical report, the sponsor's OS results are immature and, because it is based on an interim analysis, there is a risk that the effect of axi-cel compared with SOC on survival is overestimated. Indeed, considering that the OS data were based on an interim analysis with large confidence intervals that included appreciable benefit as well as little-to-no difference, longer follow-up is required to ascertain the presence and magnitude of OS benefit. The ZUMA-1 trial reported a cure fraction of 52% at year 2, and existing real-world evidence have reported a 5-year OS for CAR T-cell therapies of between 30% and 50%.²⁹ Although these estimates are derived in a later line of therapy (third line), CADTH clinical expert feedback noted that any estimates of more than 50% would be considered highly optimistic.

  There is further uncertainty regarding the expected OS benefit in a Canadian setting following the implementation of axi-cel to a broader population that goes beyond the selective patient population recruited within the clinical trial. ZUMA-7 consisted exclusively of patients with an ECOG PS of 1 or less with a mean age of 57 years. The clinical experts noted that the average age for patients with relapsed or refractory LBCL is expected to be higher in Canada. Indeed, a Canadian real-world population-based study of a cancer registry estimated the median age of patients with incident LBCL to be 67 years in males and 71 in females.³⁰ As such, if axi-cel were to become available in a clinical practice where patients are likely to have more diverse clinical and demographic profiles, there remains uncertainty about the presence and the magnitude of the OS benefit in a real-world setting.

  In the sponsor’s base case, a generalized gamma hazard was used as the parametric distribution of OS for both axi-cel and SOC. According to the sponsor, their model selection was based on several factors including best statistical fit, clinical expert opinion and the requirement that long-term survivors in both arms should follow the same functional form for the remainder of the model time horizon. However, the generalized gamma distribution selected by the sponsor to extrapolate OS for axi-cel yielded the worst goodness of fit, based on Bayesian information criterion (BIC), of the available distributions and lacked face validity as noted above when CADTH sought clinical expert feedback on the OS extrapolations. The sponsor’s model allowed exploration of several parametric distributions. However, significant uncertainty remains around these and their impact on the extrapolation of overall comparative costs and effectiveness. The appropriateness of any extrapolation depends on the quantity of observed data available. Given the lack of long-term evidence and the high amount of censoring observed in the interim analysis of OS at 18 months within the ZUMA-7 trial, it is not possible to conclude that axi-cel is a curative therapy. Thus, although an OS benefit with axi-cel was deemed plausible, the magnitude of such a benefit was uncertain in the absence of more robust long-term evidence.

  • This limitation is particularly concerning giving that the majority of QALY benefit (87%) realized by patients receiving axi-cel in the sponsor’s base-case analysis was derived from the period beyond which there are observed trial data (i.e., extrapolated period). To address some of the uncertainties in long-term OS estimates, CADTH changed the distributional assumption to log-logistic to derive a more plausible OS curve for axi-cel. This distribution had the best statistical fit of the 7 parametric distributions provided by the sponsor (BIC = 714.9 [generalized gamma] versus 709.6 [log-logistic], respectively). According to clinical expert feedback, this change further produced a more plausible OS curve in the absence of long-term evidence (i.e., 5-year cure fraction = 44%), while still conferring a survival benefit with axi-cel.

• Utility values associated with uncertainty: The sponsor derived utility estimates for the on- and off-treatment event-free states directly from the raw EQ-5D index data collected in the ZUMA-7 trial and transformed them to utility weights based on Canadian population tariffs. However, in selecting this approach, patients preparing to receive and receiving the axi-cel infusion (in the first month) or undergoing SOC including ASCT (in the first 3 months) would have similar or better quality of life than the reported general age- and sex-adjusted Canadian population (index score = 0.827).⁷ Clinical expert feedback sought by CADTH further noted that time-dependent utilities may better reflect the impact of treatment in the short-term given that patients are expected to undergo an array of taxing treatments with varying timelines as part of axi-cel infusion or SOC. However, after 5 months posttreatment, the quality of life would be similar for patients in the event-free state, regardless of the second-line treatment received.

  In addition, the sponsor sourced the utility value associated with the postevent state from the JULIET trial data for tisagenlecleucel in patients with relapsed or refractory LBCL following 2 or more lines of systemic therapy. In doing so, the sponsor assumed that the postevent utility for this patient population (i.e., progressed disease following axi-cel or SOC in second line) would be equal to the postevent utility derived from the JULIET trial (i.e., further progressed LBCL following failure of CAR T-cell therapy in the third-line setting). The clinical experts consulted by CADTH indicated that the quality of life among patients who progress following CAR T-cell therapy in second line would not be equivalent to that of patients who progress following CAR T-cell therapy in later lines because this reflects a more biologically aggressive disease.

  • To address these limitations, CADTH applied time-dependent utility estimates based on shorter 50-day intervals, based on the ZUMA-7 trial, for the first 149 days. Thereafter, CADTH assumed that utility values did not exceed age-adjusted population utility norms. Utilities based on shorter time intervals were considered more valid (i.e., utility estimates did not exceed age- and sex-matched Canadian utility estimates) and yielded higher sensitivity to detect differences in the quality-of-life impacts between therapies. For the postevent health state, in line with clinical expert opinion, a utility value of 0.73 was assumed. This value had improved face validity as it remained lower than the utility estimates in the pre-event health states but was higher than the original postevent estimates derived from the JULIET trial.

  • CADTH acknowledges that the trial-based values may overestimate the event-free utilities giving that the ZUMA-7 population was biologically fitter by design (refer to Key Limitations in Table 2). In light of the uncertainties in the utility estimates, CADTH conducted a scenario analysis sourcing utilities from alternate sources.^31-33 The utility value assigned to patients while on treatment with axi-cel in the first month (0.74) was sourced from an ad hoc analysis from a phase I/II safety management study of axi-cel in combination with atezolizumab (ZUMA-6) for the treatment of patients with refractory LBCL,^33,34 whereas the utility estimate for patients while on treatment with SOC in the first 3 months (0.673) was sourced from a published study.³¹ Finally, the event-free utility assigned to patients following treatment with either axi-cel or SOC and before an event (0.823) was sourced from a study reporting preference-based EQ-5D index scores for chronic conditions in the US.³²

• Inconsistent modelling of subsequent treatment mix: CADTH noted that the estimated treatment mix for subsequent therapies varied substantially between patients receiving axi-cel and patients receiving SOC. Of note, 40% of patients receiving the axi-cel infusion were assumed to seek subsequent treatment through clinical trials for investigational therapies (at no additional cost to drug plans), while this assumption was extended to zero patients in the SOC arm. This implicitly assumes patients receiving axi-cel would have access to other treatment options that are not available to those on SOC, which lacks face validity. Furthermore, by assuming that patients receiving axi-cel were disproportionately more likely than patients receiving SOC to seek clinical trials, the sponsor effectively reduced the costs associated with subsequent treatment in the axi-cel arm.

  • In light of this limitation, CADTH conducted a reanalysis assuming that zero patients in the axi-cel arm receive subsequent therapies through clinical trials. The proportion of axi-cel clinical trial recipients was redistributed to chemotherapy, radiotherapy, and no active treatment based on clinical expert feedback.

• Mortality of long-term survivors: The sponsor applied a standardized mortality ratio (SMR) of 1.09 to increase the hazard of death relative to the general population for those who are assumed to achieve long-term remission (i.e., the cure fraction). The sponsor relied on a previous study of newly diagnosed DLBCL that reported the SMR of patients in a French cohort (N = 820) who were event-free at 24 months following immunochemotherapy.⁵ The same study reported higher excess mortality (SMR = 1.18) in the American cohort of patients with DLBCL (N = 767). CADTH considers that the higher estimate would better capture the excess mortality of an adult cohort with relapsed or refractory LBCL who have received extensive prior treatments including first-line chemotherapy followed by second-line axi-cel or ASCT. A higher risk of mortality is expected due to potential late toxicities of therapy. CADTH sought clinical expert feedback that supported the use of the higher SMR in the long term surviving population, and further noted that it would be more reasonable to align the potential excess mortality experienced by a Canadian population of patients to estimates derived from a North American population rather than a European population.

  • To address this limitation, CADTH performed a reanalysis that adjusted the hazard of death to 1.18%.

In addition, the following key assumptions were made by the sponsor and have been appraised by CADTH (refer to Table 4).

Table 4: Key Assumptions of the Submitted Economic Evaluation

AE = adverse event; axi-cel = axicabtagene ciloleucel; CAR = chimeric antigen receptor; CRS = cytokine release syndrome; SOC = standard of care.

CADTH Reanalyses of the Economic Evaluation

Base-Case Results

CADTH’s reanalysis addressed several limitations within the submitted economic model. The CADTH base case was derived by making changes in the model parameter values and assumptions, in consultation with clinical experts. The following changes were applied: adjusting the parametric distribution to log-logistic to derive a more plausible OS curve for axi-cel; employing time-dependent utility estimates based on shorter 50-day intervals, based on the ZUMA-7 trial, for the first 149 days, while assuming that utility values did not exceed age- and sex-adjusted population utility norms thereafter; revising the postevent utility value; assuming that zero patients in the axi-cel group receive subsequent therapies through clinical trials; and adjusting the hazard of death among long-term survivors. These changes are summarized in Table 5.

Table 5: CADTH Revisions to the Submitted Economic Evaluation

ASCT = autologous stem cell transplant; axi-cel = axicabtagene ciloleucel; MCM = mixture cure model; OS = overall survival; SE = standard error; SMR = standardized mortality ratio.

^aTreatment-specific utilities were applied to the event-free state during the first month for axi-cel and the first 3 months for SOC.

^bCADTH assumed that beyond 149 days, utility values did not exceed age- and sex-adjusted utility norms. For instance, the age-adjusted utility values for the age range of 55 years to 64 years is 0.830 for men and 0.820 for women.

In the reimbursement requested population, CADTH’s base-case reanalysis estimated that axi-cel was $235,146 more expensive and yielded 0.58 more QALYs when compared to SOC for adult patients with relapsed or refractory LBCL who are candidates for ASCT. This resulted in an ICER for axi-cel of $404,418 per QALY when compared to SOC (Table 6). The probability that axi-cel was cost-effective at a WTP threshold of $50,000 per QALY was 0%. A detailed breakdown of the disaggregated results is available in Table 15. It is important to note that, similar to the sponsor’s analysis, the CADTH reanalysis was based on the assumption that 50% of patients who fail SOC would be treated with CAR T-cell therapies in the third-line setting.

The estimated ICER was higher than the sponsor’s base-case value, driven primarily by adjusting the parametric distribution of OS to log-logistic. This produced a more plausible OS curve for axi-cel in the absence of long-term evidence, while still conferring a survival benefit with axi-cel. All of axi-cel’s QALY gain compared with SOC was found to be derived in the event-free health state. Furthermore, 71% of the QALY benefit was derived from the extrapolated period in which there are no OS data to confirm this modelled benefit.

The majority of the total costs in the axi-cel group (80%) and the SOC group (68%) were similarly related to CAR T-cell treatment acquisition costs.

Table 6: Summary of the Stepped Analysis of the CADTH Reanalysis Results