CADTH Reimbursement Review

Axicabtagene Ciloleucel (Yescarta)

Sponsor: Gilead Sciences Canada Inc.

Therapeutic area: Relapsed or refractory follicular 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: Background Information of Application Submitted for Review

CAR = chimeric antigen receptor; NOC/c = Notice of Compliance with conditions.

Introduction

Non-Hodgkin lymphoma (NHL) encompasses a heterogenous group of more than 80 closely related cancers.¹ It is characterized by the abnormal and uncontrolled proliferation of cells (i.e., T cells, B cells, and natural killer cells) of the lymphatic system.^2,3 Follicular lymphoma (FL), a subtype of NHL, is an indolent B-cell lymphoma originating from the germinal centre of lymphoid tissues^2-6 and characterized by slow growth and spread.⁷ It makes up 20% to 30% of all NHL cases.¹ The sponsor-calculated overall incidence rate of FL in Canada (based on the NHL age-standardized incidence rates [25.7 per 100,000] and the proportion of FL among NHL cases [25%])^4,6 reported was 7.21 per 100,000.^4,8,9 Although responsive to initial first- or second-line therapies, FL is characterized by a relapsing and remitting disease course, especially in advanced disease stages. Patients will eventually require multiple treatments to manage or slow disease progression throughout their lifetime as response to treatments decline upon repeated therapy.^3,10-12 The clinical experts consulted by the sponsor reported that approximately 30.95% of incident FL patients would progress to third-line treatment, of whom 60% would proceed to receive active therapy.^13-15 FL can be further classified into 3 grades (1, 2, and 3 [a and b]) based on cell structures under the microscope, specifically, the number of large FL cells (centroblasts) observed.¹⁶ Grades 1, 2, and 3a diseases are generally considered low grade or slow growing compared to grade 3b, which grows fast and is considered high-grade lymphoma. According to the Canadian Cancer Society, 91% of patients considered “low risk” at diagnosis as per the Follicular Lymphoma Internal Prognostic Index (FLIPI) score have a 5-year survival rate and 71% have a 10-year survival rate; 78% of patients considered “intermediate risk” have a 5-year survival rate and 51% have a 10-year survival rate; and 53% of patients considered “high risk” have a 5-year survival rate and 36% have a 10-year survival rate.¹⁷

The objective of this report is to review and critically appraise the evidence submitted by the sponsor on the beneficial and harmful effects of axicabtagene ciloleucel, at a target dose of 2 × 10⁶ chimeric antigen receptor (CAR)-positive T cells/kg body weight, to a maximum of 2 × 10⁸ CAR-positive viable T cells, by IV infusion, in the treatment of adult patients with relapsed or refractory (r/r) FL after 2 or more lines of systemic therapy.

Stakeholder Perspectives

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

Patient Input

One patient advocacy group, Lymphoma Canada (LC), provided input for this review. LC is a national Canadian registered charity whose mission is to empower patients and the lymphoma community through education, support, advocacy, and research. The LC patient group expressed the need for accessible treatment options for patients, emphasizing that local access to treatments would significantly improve patients' quality of life and experience by reducing fear and the risk of getting sick while travelling.

LC gathered information for this input via online surveys completed anonymously by patients between April 21, 2022, and April 3, 2023. Of the 143 responses submitted, 3 respondents reported prior experience with axicabtagene ciloleucel. Respondents indicated that fatigue (50%), body aches and pain (33%), enlarged lymph nodes (33%), indigestion (32%), and bodily swelling (21%) were the most challenging symptoms that impacted their quality of life at the time of diagnosis. Respondents described FL symptoms as challenges in their daily lives that impacted their ability to travel (46%), spend time with family or friends (41%), exercise (37%), concentrate (36%), and work or complete school or volunteer activities (35%) . About half (49%) of the respondents reported that they went through a period of “watchful waiting” before commencing treatment. Most respondents (43%) had received 1 line of treatment. The most common treatments reported by respondents who had received 1 or 2 lines of systemic therapy included chemotherapy, chemoimmunotherapy, rituximab with or without bendamustine, or radiation. The posttreatment symptoms that most significantly negatively impacted respondents included treatment-related fatigue (28%), immediate side effects of treatment (26%), and low activity level (23%). Fatigue (69%), hair loss (41%), and constipation (38%) were the most common side effects reported by respondents. The most important outcomes highlighted by respondents included long life (84%), longer disease remission (82%), improved quality of life and ability to perform daily activities (69%), ability to control disease symptoms (63%), and ability to normalize blood counts (58%). Two-thirds of the respondents indicated that they were willing to tolerate nonsevere side effects for a short-term as a trade-off for a novel treatment. Two respondents who had previously received axicabtagene ciloleucel reported having access to the drug via a clinical trial. Reported side effects were cytokine release syndrome (CRS), neutropenia, febrile neutropenia, thrombocytopenia, constipation, and swelling. Some of the challenges the respondents associated with receiving axicabtagene ciloleucel included the frequent monitoring of side effects postinfusion, the inability to perform daily activities, and being away from family and friends. Both respondents said that they had a good or very good experience with axicabtagene ciloleucel and would recommend this treatment to other patients with r/r FL.

Clinician Input

Input From Clinical Experts Consulted by CADTH

A panel of 4 experts with experience treating r/r FL were consulted to determine the unmet needs, place in therapy, the patient population identified as most and least likely to benefit from treatment, when to start treatment, how best to assess response to treatment, and guidance for discontinuing treatment. The clinical experts indicated that the most important goals for treatment are to prolong life, and that the greatest unmet needs exist in patients with cancer that progresses within 2 years after their initial therapy, the patients who have already received autologous stem cell transplant (auto-SCT) or are ineligible for auto-SCT, or those who have been double refractory to earlier line treatments (implying limited treatment options available to them). The clinical panel suggested that axicabtagene ciloleucel be used as third- or later-line treatments for patients with r/r FL. These patients usually have a treatment response that lasts less than 6 months after their last treatment (medication or SCT).

The clinical panel indicated that, in practice, CAR T-cell therapy is used in a patient population that is broader than the population selected and recruited for clinical trials. The panel indicated that in clinical practice, patients are evaluated and followed in a manner similar to that described in the clinical trials of FL treatments. Remission and survival are measured. Physical exams and imaging exams are routinely conducted to assess the patient’s response to CAR T-cell therapy. The panel suggested that meaningful responses to treatment with axicabtagene ciloleucel would include a high complete remission rate, durability of treatment response, and long-term progression-free survival (PFS) and overall survival (OS). The panel indicated that in the event of treatment failure after infusion with axicabtagene ciloleucel, patients may participate in a clinical trial. In the absence of clinical trial, they may try a different chemoimmunotherapy that they have not been exposed to or undergo auto-SCT if they have not already received this therapy. The panel emphasized the importance of an accredited multidisciplinary team involving hematologists, infectious disease specialists, neurologists, an intensive care unit team, and other specialists to diagnose, treat, and monitor the patients receiving axicabtagene ciloleucel and to ensure the safe and effective delivery of this treatment.

Clinician Group Input

Input from 1 clinician group, the Ontario Health — Cancer Care Ontario (OH-CCO) Hematology Cancer Drug Advisory Committee, was summarized for this review. The disease course of FL varies for every patient. Some patients may present with long remissions between therapies while others would have refractory disease. Current treatment goals for patients with FL, according to the clinician group include palliative care and, in some scenarios, treatment with curative intent using allogeneic stem cell transplant (allo-SCT). The most important goals outlined were to delay disease progression, improve patient health-related quality of life, and alleviate symptoms. The OH-CCO Hematology Cancer Drug Advisory Committee acknowledged that current treatment options do not meet the needs of patients with r/r FL. The clinicians in the committee mentioned that patients who become refractory to chemotherapy have no other treatment options to delay the disease. In addition, the clinicians highlighted that repeated administration of cytotoxic therapy may be associated with marrow damage (myelodysplastic syndrome), which further limits the ability to treat patients, and adversely affects quality of life. Hence, there is a need for treatment options that patients can tolerate. Treatment with CAR T-cell therapy, according to the committee members, is not anticipated to cause long-term marrow damage issues. The clinicians noted that a third-line therapy with a CAR T-cell therapy would be appropriate, given that current therapy provides less benefit to patients with r/r FL disease. Patients eligible to receive axicabtagene ciloleucel in clinical practice would be similar to patients included in the clinical trial, according to the experts. However, patients with severe organ dysfunction, poor performance status, and uncontrolled infections would be ineligible. The clinicians pointed out that patients who had received prior CD19-directed therapy should be considered for treatment with CAR T-cell therapy and highlighted the need for flexibility around patients’ Eastern Cooperative Oncology Performance Status (ECOG PS) or Karnofsky Performance Status (KPS) scores. The committee members noted that some patients might become ineligible to receive CAR T-cell therapy during manufacturing, which may lead to discontinuation.

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 drug plans identified implementation issues related to initiation, prescribing, generalizability, funding algorithm, care provision, system issues, and economic considerations. The clinical experts consulted by CADTH for this review weighed evidence from the included study and other clinical considerations to provide responses to the drug plan’s implementation questions.

Clinical Evidence

Systematic Review

Description of Studies

ZUMA-5 is a multicentre, international, open-label, single-arm phase II trial.¹⁸ The study objective was to determine the efficacy and safety of axicabtagene ciloleucel in patients with r/r FL or marginal zone lymphoma (MZL) after 2 or more lines of systemic therapy. Between |||| |||| ||| |||| ||||, 127 FL patients were enrolled at 15 sites in the US and 2 in France. There were no study sites in Canada. Prior to receiving any treatments, patients underwent leukapheresis to obtain T cells as part of the manufacturing process for axicabtagene ciloleucel. Patients were then treated with cyclophosphamide and fludarabine lymphodepleting chemotherapy between 5 and 3 days before axicabtagene ciloleucel infusion. After 2 days of rest, patients received axicabtagene ciloleucel through IV infusion with a target dose of 2 × 10⁶ anti-CD19 CAR T cells/kg body weight. Analyses were conducted at 18 months, 24 months, and 36 months. The statistical analysis plan prespecified that tests on the inferential analysis set be conducted at 18 months, that is, on the date when 80 patients had been followed for at least 18 months. Using all enrolled patients, analyses were conducted at 18 months, 24 months (not presented), and 36 months. The data cut-off for the 18-month analysis was September 14, 2020; the data cut-off for the 36-month analysis was March 31, 2022.

At the 36-month time point for analysis, the median age was 60 years (range, 34 years to 79 years) and 62% of patients had an ECOG PS score of 0. Of the enrolled patients, 69% were refractory, defined as progressing within 6 months of their most recent treatment. Most patients enrolled in ZUMA-5 had received 2 prior therapies |||||, 26% had received 3 prior therapies, 20% had received 4 prior therapies, and 17% had received 5 or more prior therapies. The proportion of patients who had received a prior auto-SCT was 24%, while the proportion of patients with high bulk tumour was 51%. The proportion of patients who had progressed within 24 months of anti-CD20 chemotherapy combination therapy (i.e., progression of disease within 24 months [POD24]) was 55%.

Efficacy Results

A summary of the efficacy results in the ZUMA-5 trial is shown in Table 2.

Overall Survival

The proportion of patients who had died due to any cause was ||| after 36 months of follow-up. The median OS had not been reached. Clinical experts considered OS to be the ideal survival end point for decision-making, but acknowledged that due to the extended survival periods seen in r/r FL, immature OS results are common. The Kaplan-Meier (KM) survival probability at 18 months was ||||| ||||| || |||||, at 24 months was ||||| ||||| || |||||, and at 36 months was 75.5% (95% confidence interval [CI], 66.9% to 82.2%).

Progression-Free Survival

The proportion of patients who experienced a progression event was ||| after 36 months of follow-up. The median PFS was 40.2 months (95% CI, 28.9 months to not evaluable [NE]). The KM PFS probability at 18 months was ||||| ||||| || |||||, at 24 months was ||||||||| ||| ||||| ||||| || |||||, and at 36 months was 54.4% (95% CI, 44.2% to 63.5%).

Objective Response Rate

At the 36-month time point, the estimated objective response rate (ORR) as per investigator assessment was a clinically meaningful 94% (95% CI, 88% to 97%) in the full analysis set (FAS), while the complete response rate (CRR) was 79% ||| || |||. According to clinical experts, and within the context of the extended survival periods in r/r FL, ORR and CRR are considered acceptable surrogate end points for more important survival end points.

The primary end point in the ZUMA-5 trial was ORR at the 18-month analysis in the inferential analysis set, with a prespecified threshold of 40% for ORR and 15% for CRR. The estimated ORR as per central assessment in the 18-month inferential analysis set was 94% ||| || |||| | ||||| ||||||| and the CRR was 79% ||| || |||| | ||||| |||||||. Subgroup analyses conducted on prespecified baseline characteristics were consistent with the overall results.

Duration of Response

At the 36-month time point for analysis, ||| of patients with a response had experienced a loss-of-response event. The estimated median duration of response (DOR) was 38.6 months (95% CI, 29.0 months to NE), which was considered clinically meaningful by the clinical experts consulted by CADTH. The KM-estimated event-free probability at 18 months was ||||| ||||| || |||||, at 24 months was ||||| ||||| || |||||, and at 36 months was ||||| ||||| || |||||.

Time to Next Treatment

At the 36-month time point for analysis, ||| of patients had experienced a time-to-next-treatment (TTNT) event. The median TTNT was NE months (95% CI, 37.8 months to NE). The KM-estimated event-free probability at 18 months was ||||| ||||| || |||||, at 24 months was ||||| ||||| || |||||, and at 36 months was 59.5% (95% CI, 50.2% to 67.6%).

Harms Results

At the 36-month time point for analysis, a total of ||| of patients in the safety analysis set experienced a treatment-emergent adverse event (TEAE) with pyrexia |||||, hypotension |||||, headache |||||, and fatigue ||||| the most frequently reported TEAEs. A total of ||| of patients in the safety analysis experienced a serious adverse event (SAE), with pyrexia ||||| and pneumonia |||| the most frequently reported SAEs. At the 36-month time point, ||| of patients in the safety analysis set had died. The most common reason for death was progressive disease ||||, following by an adverse event (AE) due to reasons other than progressive disease or subsequent therapy |||| and secondary malignancy ||||.

Notable harms identified included CRS, neurologic events, cytopenias, infection, and hypogammaglobulinemia. At the 36-month analysis ||| of patients in the safety analysis set had experienced CRS, with || experiencing a grade 3 or higher CRS. Neurologic events were reported in ||| of patients, with ||| reporting a grade 3 or higher neurologic event. Cytopenias were reported in ||| of patients, with ||| of patients reporting a grade 3 or higher cytopenia. Infections were reported in ||| of patients, with ||| reporting a grade 3 or higher infection. Hypogammaglobulinemia was reported in ||| of patients, with || of patients reporting a grade 3 or higher hypogammaglobulinemia.

Table 2: Summary of Key Results From Pivotal Studies and RCT Evidence

AESI = adverse event of special interest; CI = confidence interval; CRR = complete response rate; CRS = cytokine release syndrome; DOR = duration of response; FAS = full analysis set; NA = not applicable; NE = not evaluable; ORR = objective response rate; OS = overall survival; PFS = progression-free survival; RCT = randomized controlled trial; SAE = serious adverse event; TEAE = treatment-emergent adverse event; TTNT = time to next treatment.

Note: ORR, CRR, PFS, and DOR are reported according to central assessment at the 18-month analysis and as per investigator assessment for the 36-month analysis.

Source: ZUMA-5 Clinical Study Report.¹⁸

Critical Appraisal

The ZUMA-5 trial, the only eligible study identified by the sponsor, was a phase II, single-arm, open-label clinical trial. The lack of comparative data is a key limitation to the interpretation of the results from a single-arm trial, as it is difficult to distinguish between the effect of the intervention, a placebo effect, or the effect of natural history. Due to the open-label design of the trial, the response outcomes measures (i.e., ORR, CRR, DOR, PFS) and subjective harms are at risk of measurement or reporting bias, though the direction of this bias is unclear. It is noted that these limitations are partly addressed through the use of a prespecified threshold for ORR and CRR end points and the use of central assessment.

Another important limitation of the ZUMA-5 trial is related to the insufficient follow-up time to draw strong conclusions on the long-term survival impacts of axicabtagene ciloleucel for patients with r/r FL. The clinical experts consulted by CADTH noted that r/r FL is a disease that can have very long periods of PFS and survival, suggesting that the follow-up duration was not long enough to fully capture the effects on OS and PFS. In addition, subsequent treatments could confound the long-term survival results of the ZUMA-5 trial.

According to the clinical experts consulted by CADTH, the ZUMA-5 trial patient population is overall representative of the patients in the population with r/r FL in Canada who would be receiving axicabtagene ciloleucel. However, the clinical experts noted that patients seen in clinical practice would include those with poorer performance status (the ZUMA-5 trial only included patients with an ECOG PS score of 0 or 1, whereas clinical experts suggest that an ECOG PS score of 2 may be treated in the clinical setting), and patients with more comorbidities. The clinical experts had different opinions regarding patients who received prior CD19-targeted therapy; some suggested that any prior CD19-targeted therapy would preclude the use of axicabtagene ciloleucel, whereas others suggested that only patients who are refractory to CD19-targeted therapy (did not respond or relapsed within 6 months) would not be suitable candidates for treatment with axicabtagene ciloleucel. According to the clinical experts consulted by CADTH, the efficacy outcomes used in this study are clinically relevant and important for the clinical trials in r/r FL, with the notable exception of health-related quality of life (HRQoL) outcomes, which are important to patients but were excluded from the ZUMA-5 trial. As such, it is not possible to determine how the introduction of axicabtagene ciloleucel will impact the HRQoL of patients in Canada.

Long-Term Extension Studies

No long-term extension studies were submitted as part of this review.

Indirect Comparisons

No indirect treatment comparisons were submitted as part of this review.

Studies Addressing Gaps in the Evidence From the Systematic Review

The aim of the sponsor was to provide an estimate of relative efficacy against standard of care therapies in patients with r/r FL who have received 2 or more lines of therapy.¹⁹

Description of Studies

The relative efficacy of axicabtagene ciloleucel versus standard of care estimated among the ZUMA-5 treated population using propensity scores with standardized mortality ratio (SMR) weights. The SCHOLAR-5 trial, the standard of care cohort, is a retrospective, observational, multicentre, database study of patients with r/r FL (grades 1 to 3a) who have received 2 or more systemic therapies. Patient-level data for the ZUMA-5 and SCHOLAR-5 trials were used to inform the comparative analysis. Propensity scores were calculated for each patient in the pooled analysis set to account for differences in baseline characteristics across populations. Selection of variables for the propensity score model was determined in a hierarchal manner and based on the advice of investigators/clinical experts with the goal of minimizing the imbalance in prognostically important covariates.

Efficacy Results

The ORR in the ZUMA-5 population was 93.7% compared to 54.0% in the propensity score–weighted SCHOLAR-5 population (odds ratio [OR] = 12.66; 95% CI, 5.24 to 30.57). The CRR in the ZUMA-5 population was 78.7% compared to 34.9% in the propensity score–weighted SCHOLAR-5 population (OR = 6.90; 95% CI, 3.62 to 13.18). The median DOR in the ZUMA-5 population was 38.64 months (95% CI, 29.04 months to NE) compared to ||||| |||||| ||||| || |||||| in the propensity score–weighted SCHOLAR-5 population (hazard ratio [HR] = |||| ||||| || |||||).

The median PFS in the ZUMA-5 population was 40.21 months (95% CI, 28.94 months to NE) compared to 12.97 months (95% CI, 7.75 months to 15.47 months) in the propensity score–weighted SCHOLAR-5 population (HR = 0.27; 95% CI, 0.18 to 0.41). The median OS in the ZUMA-5 population was NE (95% CI, NE to NE) compared to NE (38.40 months to NE) in the propensity score–weighted SCHOLAR-5 population (HR = 0.56; 95% CI, 0.33 to 0.95). The median TTNT in the ZUMA-5 population was NE (95% CI, 37.85 months to NE) compared to 26.61 months (95% CI, 12.65 months to NE) in the propensity score–weighted SCHOLAR-5 population, with HR of 0.60 (95% CI, 0.39 to 0.93).

Harms Results

Safety end points were not included in the analysis.

Critical Appraisal

Due to differences between the ZUMA-5 and SCHOLAR-5 cohorts in treatment allocation, it is possible that the treatment effect estimate is confounded by imbalances in prognostic covariates across populations. The sponsor identified and adjusted for several important variables, resulting in a suitable balance of these characteristics across both populations; however, important characteristics such as FLIPI score could not be adjusted for due to missing data. Characteristics such as ECOG PS, FL grade, and whether patients were double refractory differed significantly between populations after propensity score weighting. The clinical experts consulted by CADTH suggested that differences in ECOG PS score and the proportion of patients who are double refractory could impact how patients respond to treatment. The direction of this impact is uncertain, with some differences (e.g., double refractory status and FL grade) potentially favouring the SCHOLAR-5 comparator over axicabtagene ciloleucel, and some differences (e.g., ECOG PS score) potentially favouring axicabtagene ciloleucel over the SCHOLAR-5 comparator.

There is additional uncertainty in the results due to the low effective sample sizes in both the ZUMA-5 trial and the SCHOLAR-5 cohort. The removal of the DELTA patients from the SCHOLAR-5 cohort resulted in a statistically significant change in the mean number (standard deviation [SD]) of lines of prior therapy: |||| |||||| in the SCHOLAR-5 trial compared to |||| |||||| in the ZUMA-5 trial. Differences between populations in the number of lines of prior therapy in particular affect determining how patients would be expected to respond to treatment. The proportion of patients with POD24 and the proportion of patients who were refractory to their most recent treatment was also reduced with the exclusion of the DELTA cohort, indicating that their removal from the SCHOLAR-5 cohort resulted in a population with a lower risk prognosis.

Conclusions

Evidence from a single-arm study (the ZUMA-5 trial) suggests that treatment with axicabtagene ciloleucel affects clinically important tumour responses, including complete remission, in adult patients with r/r FL after 2 or more lines of systemic therapies. Due to the single-arm design of the trial and limited duration of follow-up, there is insufficient evidence to determine the magnitude of the effect of axicabtagene ciloleucel on OS and PFS. HRQoL outcomes were not included in the ZUMA-5 trial and therefore the impact of axicabtagene ciloleucel on patients HRQoL is unknown. The harms associated with the axicabtagene infusion are as expected given the mechanism of action and prior experience in other indications. The comparison of the ZUMA-5 trial to the retrospective SCHOLAR-5 external control was limited by heterogeneity across study designs and populations. Specifically, the inability to adjust for ECOG PS and double refractory status can introduce bias to the estimation procedure within the comparative populations. Generalizability to individuals that do not meet the ZUMA-5 trial criteria is also in question. Therefore, the magnitude of the comparative efficacy estimates for axicabtagene ciloleucel against standard of care in the Canadian setting is uncertain.

Introduction

The objective of this report is to review and critically appraise the evidence submitted by the sponsor on the beneficial and harmful effects of axicabtagene ciloleucel (target of 2 × 10⁶ CAR-positive viable T cells/kg body weight, for IV use; range, 1 × 10⁶ cells/kg to 2.4 × 10⁶ cells/kg) in the treatment of adult patients with r/r grade 1, 2, or 3a FL after 2 or more lines of systemic therapy.

Disease Background

Contents within this section have been informed by materials submitted by the sponsor and clinical expert input. The following have been summarized and validated by the CADTH review team.

NHL encompasses a heterogenous group of more than 80 closely related cancers.¹ It is characterized by the abnormal and uncontrolled proliferation of cells (i.e., T cells, B cells, and natural killer cells) of the lymphatic system.^2,3 About 85% to 90% of NHLs develop from B lymphocytes, and the rest from T lymphocytes or natural killer cells.^2,20 NHL is the fifth most common cancer diagnosed in Canada, and is commonly diagnosed in adults aged 50 years or older.¹ In 2022, an estimated 11,400 new diagnoses of NHL and 3,000 deaths from NHL were projected for Canada.²¹ The estimates were higher in males (6,600 new cases and 1,700 deaths) than in females (4,800 new cases and 1,250 deaths).²¹ Risk factors for all NHL-related cancers include immune disorders (e.g., rheumatoid arthritis, Sjögren syndrome, systemic lupus erythematosus), the use of immunosuppressive therapies, bacterial and viral infections (e.g., Helicobacter pylori, Epstein-Barr virus, and hepatitis C virus), family history and genetics, and occupational and lifestyle risk factors.^1-3,20 NHL can affect any organ in the body and has a wide range of clinical presentations.^3,20

FL, a subtype of NHL, is an indolent B-cell lymphoma originating from the germinal centre of lymphoid tissues,^2-6 and characterized by slow growth and spread.⁷ FL makes up 20% to 30% of all NHL cases¹ and up to 70% of all indolent NHL cases. There are limited epidemiological data on FL in Canada compared with other NHL-associated cancers. One study that looked at trends (incidence and mortality) of FL across Canada using data from 3 registries (the Canadian Cancer Registry, the Registre québécois du cancer, and the Canadian Vital Statistics database) reported that there were about 22,625 new cases of FL between 1992 and 2010. The mean age of patients was 60.8 years at diagnosis, with equal incidence rates observed in males and females (50% in each sex) in the study population.⁶ The authors also reported a variability in incidence rates across provinces, with rates in Manitoba, New Brunswick, Nova Scotia, and Prince Edward Island notably higher than the national average.^6,22 The sponsor-calculated overall incidence rate of FL in Canada (based on the age-standardized incidence rates of NHL [25.7 per 100,000] and the reported proportion of FL among NHL cases [25%])^4,6 was 7.21 per 100,000.^4,8,9

Although responsive to initial first- or second-line therapies, FL is characterized by a relapsing and remitting disease course, especially in advanced disease stages. Patients will eventually require multiple treatments to manage or slow disease progression throughout their lifetime as response to treatments declines with repeated therapy.^3,10-12 The clinical experts consulted by the sponsor reported that approximately |||||| of incident FL patients would progress to third-line treatment, of which ||| would proceed to receive active therapy.^13-15 In Canada, FL is common in people aged 50 years or older, and most are diagnosed at an advanced stage (i.e., 66% to 70% of patients are diagnosed at stage III or IV).²³ FL can be further classified into 3 grades (1, 2, and 3 [a and b]) based on cell structures visible under the microscope, specifically the number of large FL cells (centroblasts).¹⁶ Grades 1, 2, and 3a diseases are generally considered low grade or slow growing compared to grade 3b disease, which grows fast and is considered high grade. FL may relapse or recur to more aggressive or high-grade forms of NHL, such as diffuse large B-cell lymphoma (DLBCL) during the disease, thus requiring other treatment options to manage the disease.^3,24 In a UK study that prospectively followed 325 patients diagnosed with FL between 1972 and 1999, the risk of histologic conversion to more aggressive forms was reported to be 28% (95% CI, 23% to 34%) by 10 years of diagnosis.^3,25 Some FL patients present with no symptoms, while others present with a range of symptoms, most of which are not specific to FL. Typical symptoms include adenopathy, splenomegaly, locally obstructing symptoms, fever, night sweats, and weight loss.^1,3,26

Survival depends on the prognostic factors used in the FLIPI score.²⁴ The FLIPI prognostic score system takes into account the following factors: age (> 60 years versus ≤ 60 years), Ann Arbor stage (III to IV versus I to II), number of involved nodal areas (> 4 versus ≤ 4), hemoglobin level (< 120 g/L versus ≥ 120 g/L), and serum lactate dehydrogenase concentration (above normal versus normal or below).²⁷ Patients with “good” prognostic FLIPI scores respond well to treatment compared to patients with “poor” FLIPI scores, in whom cancer may recur following treatment.²⁴ According to the Canadian Cancer Society, 91% of patients considered “low risk” at diagnosis according to the FLIPI score have a 5-year survival rate and 71% have a 10-year survival rate; 78% of patients considered “intermediate risk” have 5-year survival rate and 51% have a 10-year survival rate; and 53% of patients considered high risk have a 5-year survival rate and 36% have a 10-year survival rate.¹⁷ Casulo and colleagues (2015)²⁸ reported that patients with POD24 after first-line rituximab therapy experienced particularly poor outcomes. In a previous study, the same authors reported that patients with POD24 had a 5-year survival rate of only 50% compared to 90% for patients without POD24.²⁸

NHL, including FL, is diagnosed using immunohistochemical and genetic testing of tissue samples biopsied from lymph nodes.²⁰ Patients undergo diagnostic testing to confirm r/r FL grade 1, 2, or 3a disease after 2 or more lines of systemic therapies. Testing may include a history and physical examination, tissue biopsy to confirm relapse or rule out transformation to aggressive lymphoma, imaging tests (PET scans or CT scans), and laboratory tests (e.g., complete blood counts). The timing of diagnostic tests relative to receiving axicabtagene ciloleucel may vary between patients.²⁹

Standards of Therapy

Contents within this section have been informed by materials submitted by the sponsor and clinical expert input. The following have been summarized and validated by the CADTH review team.

The treatment goals for FL vary depending upon the stage of the FL and individual patient factors. In general, available treatments for stage I to II FL have curative potential; however, for most patients with stage III to IV FL, no curative therapies are available. Therefore, the main goals of treatment are to cure the lymphoma in patients with stage I to II FL, and to extend remission in patients with stage III to IV FL.³⁰

Once a diagnosis of FL is confirmed, the gold standard for the management of asymptomatic patients with indolent FL is watchful waiting, also known as “watch and wait.”^26,31 According to the clinical experts consulted by CADTH, watch and wait is a common practice for many patients with FL, even after disease relapse.

First-Line Treatments

For small, localized symptomatic FL, radiotherapy is considered the standard of care according to North American and European guidelines.^30,32,33 This is supported by several provincial guidelines in Canada.^34,35 For grade 1, 2, and 3a FL, the preferred chemoimmunotherapy regimen is bendamustine plus rituximab, based on high-level evidence of efficacy and favourable tolerability in this population.^26,30,36 In frail and older patients, rituximab monotherapy, a chemotherapy-free approach, is the preferred first-line regimen according to European and North American guidelines.^37,38 However, some treatment centres in Canada do not have access to rituximab monotherapy. Instead, physicians keep patients on bendamustine plus rituximab for several treatment cycles for as long as possible. Beyond first-line treatment, there is currently no gold standard of care for the r/r FL population.

Second-Line Treatments

Treatment options for second-line regimens for r/r FL depend on several factors, including level of fitness, prior treatment, and length of time to relapse.³⁰ The preferred treatment strategy in this patient population is combined immunochemotherapy, such as O-CHOP (obinutuzumab, cyclophosphamide, doxorubicin hydrochloride [hydroxydaunorubicin], vincristine sulphate [Oncovin], and prednisone).³⁰ SCT may be considered in young and fit patients with no comorbidities in the second-line setting. Auto-SCT is given more often than allo-SCT in this population. However, only a small subset of FL patients would be eligible for transplant in the second line. SCT is limited by highly selective eligibility criteria and is typically reserved for younger, medically fit patients with chemotherapy-sensitive disease.^30,37,39,40 The clinical experts consulted by CADTH agreed with these strategies.

Third-Line Treatments

FL is a relapsing disease with continued unmet need in adult patients with r/r FL after 2 or more lines of therapy despite the availability of established therapies. Patients with r/r FL in the third- and later-line treatment setting represent a heavily pretreated and advanced-stage patient population. The standard of care in Canada for the third-line treatment of r/r FL is heterogenous and varies across regions. Based on Canadian clinician input, a heterogenous mix of immunochemotherapy (for most patients) and SCT (for a minority of patients) are the current treatment options in this hard-to-treat population.

Treatments in this setting may include SCT; however, there is controversy with regard to clinical benefit of SCT in patients with r/r FL, and both allo-SCT and auto-SCT may be associated with significant mortality and morbidity.^30,37,39-41 While SCT has been included in the third-line treatment algorithm diagram for r/r FL, few people with r/r FL are expected to be eligible for SCT in the third line.

According to clinical experts consulted by the sponsor, most patients in the third-line treatment setting will continue recycling combined immunochemotherapy that might have been used in previous lines, including the following regimens (which are the most frequently used options in Canada): rituximab plus cyclophosphamide, vincristine, and prednisone (R-CVP); rituximab plus cyclophosphamide, hydroxydaunorubicin (doxorubicin hydrochloride), oncovin (vincristine sulphate), and prednisone (R-CHOP); obinutuzumab, cyclophosphamide, doxorubicin hydrochloride [hydroxydaunorubicin], oncovin (vincristine sulphate), and prednisone (O-CHOP); rituximab plus gemcitabine, dexamethasone, and cisplatin (R-GDP); bendamustine plus rituximab; and rituximab plus ifosfamide, carboplatin, and etoposide (R-ICE). Although it is used by some physicians in their clinical practice, lenalidomide plus rituximab regimen is currently not officially indicated for treatment of FL in Canada, as per the most recent Canadian product monograph for lenalidomide.⁴² Furthermore, although no official submission has been made by the sponsor to Canadian health technology assessment agencies, CADTH conducted a health technology review of lenalidomide plus rituximab in r/r B-cell NHL and concluded that available evidence remains limited.⁴³ According to Canadian clinicians’ input, access to lenalidomide plus rituximab is somewhat limited in some provinces. In the third-line setting and beyond, idelalisib, a phosphoinositide 3-kinase (PI3K) inhibitor, is indicated for the treatment of patients with r/r FL in Canada.⁴⁴ However, idelalisib is not publicly reimbursed across Canada and is inaccessible at some sites, according to several clinicians.⁴⁵ Furthermore, it is generally at the bottom of the treatment list due to serious side effects and is, therefore, used as a palliative treatment. In recent years, CAR T-cell therapy has emerged as another form of immunotherapy for the treatment of blood cancers, including lymphomas. While promising results have been reported for the CAR T-cell therapies for advanced-stage lymphomas, severe adverse effects related to CAR T-cell therapy, such as CRS and neurologic toxicities, have also been reported.⁴⁶ One CAR T-cell product, tisagenlecleucel, has been recently approved by Health Canada (December 2022) with a Notice of Compliance with conditions for the treatment of adult patients with r/r FL grade 1, 2, or 3a after 2 or more lines of systemic therapy.⁴⁷ Note that at the time of writing this report, tisagenlecleucel has not been publicly funded for this indication.

In clinical practice, a patient’s response to treatment is commonly assessed using the Lugano classification,⁴⁸ in which fluorodeoxyglucose-PET/CT is incorporated into the initial Ann Arbor Staging System for fluorodeoxyglucose-avid lymphomas.

Drug Under Review

Axicabtagene ciloleucel (Yescarta) is a CD19-directed genetically modified autologous T-cell immunotherapy (i.e., CAR T-cell therapy) that binds to CD19-expressing cancer cells and normal B cells.⁴⁹ Following anti-CD19 CAR T-cell engagement with CD19-expressing target cells, the CD28 and CD3-zeta co-stimulatory 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.⁴⁹ Axicabtagene ciloleucel is a single-dose, 1-time treatment in a patient-specific infusion bag. Axicabtagene ciloleucel should be administered by experienced health professionals in 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), to a maximum of 2 × 10⁸ anti-CD19 CAR T cells for patients weighing 100 kg or more.⁴⁹

Axicabtagene ciloleucel was approved in Canada on September 28, 2022, for the treatment of adult patients with r/r grade 1, 2, or 3a FL after 2 or more lines of systemic therapy. Axicabtagene ciloleucel has been previously reviewed by CADTH for 2 indications. On August 15, 2019, CADTH issued a positive recommendation for axicabtagene ciloleucel for adult patients with r/r large B-cell lymphoma after 2 or more lines of systemic therapy, including DLBCL not otherwise specified, primary mediastinal large B-cell lymphoma, high-grade B-cell lymphoma, and DLBCL arising from FL. On January 5, 2023, another positive recommendation was posted for the treatment of adult patients with DLBCL or high-grade B-cell lymphoma that is refractory to first-line chemoimmunotherapy or that relapses within 12 months of first-line chemoimmunotherapy, who are eligible for auto-SCT.

Key characteristics of axicabtagene ciloleucel are summarized in Table 3 with other treatments available for FL.

Table 3: Key Characteristics of Axicabtagene Ciloleucel

AE = adverse event; CAR = chimeric antigen receptor; CRS = cytokine release syndrome; FL = follicular lymphoma.

^aHealth Canada–approved indication.

Source: Yescarta Product Monograph.⁴⁹

Stakeholder Perspectives

Patient Group Input

This section was prepared by the CADTH review team based on the input provided by patient groups. The full original patient input received by CADTH has been included in the Stakeholder section of this report.

One patient advocacy group, LC, provided input for this review. LC is a national Canadian registered charity whose mission is to empower patients and the lymphoma community through education, support, advocacy, and research. LC collaborates with patients, caregivers, health care professionals, other organizations, and stakeholders to promote early detection, find new and better treatments for lymphoma patients, help patients access treatments, learn about the causes of lymphoma, and work together to find a cure. The LC patient group expressed the need for accessible treatment options for patients, highlighting that local access to treatments will significantly improve patients' experience by reducing fear and the risk of getting sick while travelling and patient quality of life. The LC patient group expressed the need for accessible treatment options for patients, highlighting that local access to treatments will significantly improve patients' quality of life and experience by reducing fear and the risk of getting sick while travelling.

LC gathered information for this input via online surveys completed anonymously by patients between April 21, 2022, and April 3, 2023. Of the 143 responses submitted, 3 respondents reported having experience with axicabtagene ciloleucel. Of the total surveyed, 86% of respondents lived in Canada, 71% were aged between 55 and 64 years, 64% were female, and 34% had received a FL diagnosis within the last 3 to 5 years. Respondents indicated that fatigue (50%), body aches and pain (33%), enlarged lymph nodes (33%), indigestion (32%), and bodily swelling (21%) were the most challenging symptoms with the biggest impact on their quality of life at the time of diagnosis. Respondents described FL symptoms as challenges in their daily lives that impacted their ability to travel (46%), spend time with family or friends (41%), exercise (37%), concentrate (36%), and work or complete school or volunteer activities (35%). Anxiety or worry (84%), stress from diagnosis (77%), fear of progression (70%), and difficulty sleeping (48%) were the most common psychosocial symptoms that impacted patients.

About half (49%) of the respondents reported that they went through a period of watchful waiting before commencing treatment. Most respondents (43%) had received 1 line of treatment. The most common treatments reported by respondents who had received 1 or 2 lines of therapy included chemotherapy, chemoimmunotherapy, rituximab with or without bendamustine, or radiation. When asked to describe their treatment experience, 57% of respondents indicated that they were “satisfied” or “very satisfied” with the treatment options in the front-line setting. Only 22% of respondents indicated that they were “satisfied” or “very satisfied” with the current treatment options in the r/r setting. While describing how their current therapy (or most recent therapy) was able to manage their FL symptoms on a scale of 1 (strongly disagree) to 5 (strongly agree), 40% of respondents strongly agreed and 20% strongly disagreed. According to the respondents, the post-treatment symptoms that most significantly negatively impacted them included treatment-related fatigue (28%), immediate side effects of treatment (26%), and low activity level (23%). Fatigue (69%), hair loss (41%), and constipation (38%) were the most common side effects reported by respondents.

The most important outcomes (rated 5 on a scale of 1 to 5) included delaying disease progression (84%), longer disease remission (82%), improved quality of life, ability to perform daily activities (69%), ability to control disease symptoms (63%), and ability to normalize blood counts (58%). In total, 68% of respondents indicated a willingness to tolerate nonsevere side effects over the short-term period when undertaking a novel therapy, 42% expressed the importance of having a choice in deciding treatment options based on known side effects and expected outcomes, and 79% noted the need for more accessible treatment options that were proven to be effective for FL.

Two respondents completed all questions about axicabtagene ciloleucel in the survey. One survey respondent who confirmed that they had prior treatment experience with axicabtagene ciloleucel did not complete all other treatment questions. The 2 respondents who completed the survey reported having access to treatment via a clinical trial. One patient received treatment in the second-line and the other in the fifth-line setting. Reported side effects included CRS, neutropenia, febrile neutropenia, thrombocytopenia, constipation, and swelling. According to the respondents, the challenges that significantly impacted their physical and mental health during treatment included the frequent monitoring of side effects postinfusion, the inability to perform daily activities, and being away from family and friends. One respondent indicated they were away from home for 1 to 3 months, while the other was away for more than 3 months. Other highlighted challenges included financial issues due to absence from work and travel accommodation expenses accumulated during the clinical trial. Both respondents stated that they had a good or very good experience with axicabtagene ciloleucel and would recommend the treatment to other patients with r/r FL.

Clinician Input

Input From Clinical Experts Consulted by CADTH

All CADTH review teams include at least 1 clinical specialist with expertise regarding 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 review of axicabtagene ciloleucel, 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 follows.

Unmet Needs

The clinical panel indicated that for patients with FL, the most important goals for an ideal treatment are to prolong survival (both OS and PFS) and to improve their quality of life. However, patients with r/r FL relapse after the front-line therapies or are refractory to the available treatments, which impacts their long-term PFS and quality of life. In addition, current treatments may not be well tolerated by some patients due to the related adverse events or complications associated with SCT. The clinical panel indicated that the greatest unmet needs for the current treatments exist for patients with cancer that progresses within 2 years after their initial therapy, for those who have already received auto-SCT or are ineligible for auto-SCT, and for those who have been double refractory to earlier line treatments (implying they have limited treatment options available to them). Patients who are eligible for CAR T-cell therapy but ineligible for auto-SCT would generally be older and less able to tolerate the auto-SCT process.

Place in Therapy

The treatment algorithm for adult patients with r/r FL is complicated. Many factors (e.g., patient characteristics, previous treatments, treatment effects and toxicity, drug plan coverage, disease progression and transformation, and patient preference) need to be considered before making decisions. Watch and wait is a common first-line approach for many patients with FL, and is also considered, though less commonly, after disease relapse. Patients who need active treatments typically receive bendamustine or rituximab-based therapies such as bendamustine plus rituximab, R-CVP, R-CHOP, or lenalidomide plus rituximab in the first- and second-line setting. PI3K inhibitors are rarely used. Patients with relapsed disease after treatment with chemoimmunotherapy, particularly those who progress within 2 years, may receive auto-SCT if they are suitable candidates. The clinical panel noted that some evidence suggests that auto-SCT cures half of the patients with POD24 and with PFS in the 20- to 30-year range. After all these treatments, some patients maintain the indolent status and some transform to large cell lymphoma; therefore, the proportion of patients who may be considered for treatment with CAR T-cell therapy is small. The clinical panel suggested that axicabtagene ciloleucel is most appropriate for use as a third or later line of treatment for patients with r/r FL. These patients usually have a treatment response that lasts less than 6 months from their last treatment (medication or SCT). There are not many options available for the patients at this stage. CAR T-cell therapy would be considered because it has a different mechanism of action.

The clinical panel noted that an auto-SCT is not mandatory before axicabtagene ciloleucel can be given, since this is not standard of care in Canada. The clinical panel suggested that an auto-SCT before axicabtagene ciloleucel would be recommended if the patient had access to auto-SCT and was eligible, noting that 80% of auto-SCT are performed after the first or second relapse.

For patients who have received previous CD19-targeted therapy, there is a lack of evidence to suggest whether the use of axicabtagene ciloleucel is appropriate.

Patient Population

The clinical panel indicated that, in practice, CAR T-cell therapy is used in a patient population that is broader than that in clinical trials, to which a more select population is recruited. For example, the panel noted that suitable candidates for treatment with axicabtagene ciloleucel would be patients with acceptable rather than excellent organ function, which is what is generally required for an auto-SCT. When determining whether axicabtagene ciloleucel treatment is suitable for a particular patient, bulk of disease and rapid disease progression are among the factors that need to be taken into account. In clinical practice, patients’ suitability can be determined based on clinical judgment, which combines medical history, laboratory and imaging findings, and often a lymph node biopsy.

The panel also noted that the patients who did not meet the eligibility criteria in the ZUMA-5 clinical trial (e.g., because they had certain comorbidities or disease status) would be the least suitable candidates for treatment with axicabtagene ciloleucel. Patients with an ECOG PS score of 3 or higher are also least suitable for treatment with axicabtagene ciloleucel.

The panel noted that there is no specific patient characteristic that can be used to predict who would respond better to axicabtagene ciloleucel than other patients.

Assessing the Response Treatment

The panel indicated that in clinical practice, patients are evaluated and followed in a manner similar to that described in the clinical trials of FL treatments. Remission and survival are measured. Physical exams and imaging exams are routinely conducted to assess the patient’s response to CAR T-cell therapy.

The panel suggested that meaningful responses to treatment with axicabtagene ciloleucel would include a high complete remission rate, durability of treatment response, and long-term PFS and OS. Ideally, a successful treatment would show a plateau in the PFS and OS curves. In addition, the clinicians were interested in knowing if the treatment is cost-effective. Ideally, the treatment benefits of axicabtagene ciloleucel can be compared to the other treatments.

The panel noted that after CAR T-cell therapy, clinicians would assess the treatment response (e.g., via CT scan) every 3 months, or sooner if needed.

When Patients Go Through Pretreatment but Do not Receive Axicabtagene Ciloleucel

The panel noted that situations when patients go through pretreatment but do not receive axicabtagene ciloleucel are rare. However, this can happen because of rapid disease progression in the interim or because the patient has major complications such as a new myocardial infarction or stroke. Manufacturing failure is another reason for this situation, although it is not expected to be an issue with axicabtagene ciloleucel.

If patients do not receive axicabtagene ciloleucel after undergoing pretreatment, most of them (in particular, high-risk patients) can progress within 6 months of their last treatment, and limited treatment options are then available for them. Palliative chemotherapy can be given. Other options may include radiation therapy, more chemotherapy, novel agents, or a clinical trial, depending on each patient’s clinical status.

Therapy Post Axicabtagene Ciloleucel Failure

The panel indicated that after infusion with axicabtagene ciloleucel and failure of treatment, patients may participate in a clinical trial. In the absence of a clinical trial, patients may try a different chemoimmunotherapy that they have not been exposed to or undergo auto-SCT if they have not already received this therapy.

Prescribing Considerations

The panel emphasized the importance of an accredited multidisciplinary team involving hematologists, infectious disease specialists, neurologists, an ICU team, and other specialists to diagnose, treat, and monitor the patients receiving axicabtagene ciloleucel and to ensure the safe and effective delivery of this treatment.

Clinician Group Input

This section was prepared by the CADTH review team based on the input provided by clinician groups. The full original clinician group input received by CADTH has been included in the Stakeholder section of this report.

Input from 1 clinician group, the OH-CCO Hematology Cancer Drug Advisory Committee, was summarized for this review. The OH-CCO Hematology Cancer Drug Advisory Committee provides timely, evidence-based clinical and health system guidance on drug-related issues supporting the OH-CCO's mandate, including the Provincial Drug Reimbursement Programs and the Systemic Treatment Program. Information in this input was obtained via video conferencing and email.

According to the clinicians consulted, the disease course of FL varies for every patient. Some patients may present with long remissions between therapies while others would have refractory disease. Current treatment goals for patients with FL, according to the clinician group, include palliative care and, in some scenarios, treatment with curative intent using allo-SCT. The most important goals outlined were to delay disease progression, improve patient HRQoL, and alleviate symptoms. Current standard of care for FL patients identified by the clinician group included chemotherapy, chemoimmunotherapy, auto-SCT, allo-SCT (for a selected group of patients), and radiation (to control symptoms and for palliative care scenarios). The OH-CCO Hematology Cancer Drug Advisory Committee acknowledged that current treatment options do not meet the needs of patients with r/r FL. The OH-CCO Hematology Cancer Drug Advisory Committee mentioned that patients who become refractory to chemotherapy have no other treatment options to delay the disease. In addition, the committee members highlighted that repeated administration of cytotoxic therapy may be associated with marrow damage (myelodysplastic syndrome), which further limits the ability to treat patients, and adversely affects quality of life. Hence, there is a need for treatment options that patients can tolerate. Treatment with a CAR T-cell therapy, according to the clinical group, is not anticipated to cause long-term marrow damage issues. The committee members said that a third-line therapy with a CAR T-cell therapy would be appropriate, given that current therapy provides less benefit to patients with r/r FL. The OH-CCO Hematology Cancer Drug Advisory Committee clinicians could not ascertain whether CAR T-cell therapy would replace auto-SCT; however, they said that they suspect that CAR T-cell therapy might be tried first, rather than auto-SCT, for patients with a history of chemotherapy-refractory forms of FL. The clinician group noted that there would be a prevalent FL patient population that will become eligible for axicabtagene ciloleucel at the time of its implementation. According to the clinicians, patients eligible to receive axicabtagene ciloleucel in clinical practice should be similar to patients included in the clinical trial. However, patients with severe organ dysfunction, poor performance status, and uncontrolled infections would be ineligible. The clinicians pointed out that patients who had received prior CD19-directed therapy (these patients were excluded from the trial) should be considered for treatment with CAR T-cell therapy and highlighted the need for flexibility around patients’ ECOG PS or KPS scores. The experts also noted that some patients might become ineligible to receive CAR T-cell therapy during the manufacturing process, which may lead to discontinuation. The clinician group noted that axicabtagene ciloleucel may have a toxicity profile that is different from that of tisagenlecleucel (Kymriah), another CAR T-cell therapy currently under review. The clinician group input aligned with the input provided by the clinical experts consulted during the CADTH review.

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

CAR = chimeric antigen receptor; CNS = central nervous system; CRS = cytokine release syndrome; DLBCL = diffuse large B-cell lymphoma; ECOG PS = Eastern Cooperative Oncology Group Performance Status; FL = follicular lymphoma; MALT = mucosa-associated lymphoid tissue; MZL = marginal zone lymphoma; pERC = CADTH pan-Canadian Oncology Drug Review Expert Review Committee; r/r = relapsed or refractory.

Clinical Evidence

The objective of this Clinical Review was to review and critically appraise the clinical evidence submitted by the sponsor on the beneficial and harmful effects of axicabtagene ciloleucel, target dose of 2 × 10⁶ CAR T cells/kg body weight, to a maximum of 2 × 10⁸ CAR-positive viable T cells, by IV infusion, in the treatment of adult patients with r/r FL after 2 or more lines of systemic therapy. The focus is to compare axicabtagene ciloleucel to relevant comparators and identify gaps in the current evidence.

A summary of the clinical evidence included by the sponsor in the review of axicabtagene ciloleucel is presented in 2 sections, with CADTH’s critical appraisal of the evidence included at the end of each section. The first section, the systematic review, includes a pivotal study that was selected according to the sponsor’s systematic review protocol. The second section includes 1 additional study that was considered by the sponsor to address important gaps in the systematic review evidence.

Included Studies

Clinical evidence included in the CADTH review and appraised in this document:

• 1 pivotal study identified in systematic review

• 1 additional study addressing gaps in evidence.

Systematic Review

Contents within this section have been informed by materials submitted by the sponsor. The following have been summarized and validated by the CADTH review team.

Description of Studies

Details of the included studies are summarized in Table 5.

Table 5: Details of Studies Included in the Systematic Review

allo-SCT = allogeneic stem cell transplant; auto-SCT = autologous stem cell transplant; CAR = chimeric antigen receptor; CNS = central nervous system; CR = complete response; CRR = complete response rate; CSF = cerebrospinal fluid; DOR = duration of response; DVT = deep vein thrombosis; ECOG PS = Eastern Cooperative Oncology Group Performance Status; FL = follicular lymphoma; iNHL = indolent non-Hodgkin lymphoma; MZL = marginal zone lymphoma; ORR = objective response rate; OS = overall survival; ORR = objective response rate; PFS = progression-free survival; PR = partial response; r/r = relapsed or refractory; TEAE = treatment-emergent adverse event; TTNT = time to next treatment.

Details included in this table are from the sponsor’s Summary of Clinical Evidence.⁵⁹

Source: ZUMA-5 Clinical Study Report.¹⁸

The ZUMA-5 trial is a multicentre, international, open-label, single-arm phase II trial. The study objective was to determine the efficacy and safety of axicabtagene ciloleucel in patients with r/r FL or MZL after 2 or more lines of systemic therapy. In line with the submitted reimbursement request and anticipated Health Canada indication, the r/r FL patient group will be the focus of this review. The ZUMA-5 study design, shown in Figure 1, consisted of screening, enrolment/leukapheresis, a conditioning chemotherapy period, a 1-time IV infusion of axicabtagene ciloleucel, posttreatment assessment, and long-term follow-up.

Between |||| |||| ||| |||| ||||, 127 FL patients were enrolled at 15 sites in the US and 2 in France. There were no study sites in Canada. Prior to receiving any treatments, patients underwent leukapheresis to obtain T cells as part of the manufacturing process for axicabtagene ciloleucel. Patients were then treated with cyclophosphamide and fludarabine lymphodepleting chemotherapy between 5 and 3 days before axicabtagene ciloleucel infusion. After 2 days of rest, patients received axicabtagene ciloleucel through IV infusion at a target dose of 2 × 10⁶ anti-CD19 CAR T cells/kg body weight. PET-CT scans for disease assessment were conducted at week 4, month 3, month 6, month 9, month 12, month 15, month 18, month 24, and at any subsequent scheduled or unscheduled visit if disease progression was a clinical concern, as per standard of care.

Analyses were conducted at 18 months, 24 months (not presented), and 36 months. The statistical analysis plan prespecified that tests be conducted on the inferential analysis set at 18 months; this was defined as the point when 80 patients had been followed for at least 18 months. Using all enrolled patients, analyses were conducted at 18 months, 24 months, and 36 months; this population of patients is referred to as the FAS. The data cut-off for the 18-month analysis was September 14, 2020, and the data cut-off for the 36-month analysis was March 31, 2022.

Figure 1: ZUMA-5 Study Design

CAR = chimeric antigen receptor.

Source: Reproduced as is from ZUMA-5 Clinical Study Report.¹⁸

Populations

Inclusion and Exclusion Criteria

Patients eligible for inclusion in the ZUMA-5 trial were aged 18 years or older, with a confirmed diagnosis of r/r indolent NHL, with FL limited to grade 1, 2, or 3a, and with r/r disease after 2 or more lines of systemic therapy. Prior therapy must have included an anti-CD20 monoclonal antibody combined with an alkylating agent. Eligible patients were required to have least 1 measurable lesion according to the Lugano Response Criteria for Malignant Lymphoma,⁴⁸ an ECOG PS score of 0 or 1, and no known history of central nervous system involvement. Patients were excluded from the ZUMA-5 trial if they had transformed disease or if they had received auto-SCT within 6 weeks of the planned axicabtagene ciloleucel infusion.

Interventions

Five days before axicabtagene ciloleucel infusion, patients received an IV conditioning chemotherapy regimen of fludarabine 30 mg/m²/day and cyclophosphamide 500 mg/m²/day. Patients were treated with these chemotherapies over 3 days to induce lymphocyte depletion. Following 2 days of rest, patients received axicabtagene ciloleucel infusion at a target dose of 2 × 10⁶ anti-CD19 CAR T cells/kg body weight. Patients who achieved a partial response at the 3-month assessment but subsequently experienced disease progression were eligible for an optional course of re-treatment with conditioning chemotherapy and axicabtagene ciloleucel.

Corticosteroid therapy at a pharmacologic dose (≥ 5 mg/day of prednisone or equivalent doses of other corticosteroids) and other immunosuppressive drugs were restricted for 7 days before leukapheresis and 5 days before axicabtagene ciloleucel administration. Systemic corticosteroids were restricted as premedication to patients for whom CT scans with contrast were contraindicated (i.e., patients with contrast allergy or impaired renal clearance). Corticosteroids and other immunosuppressive drugs were restricted for 3 months after axicabtagene ciloleucel administration unless used to manage axicabtagene ciloleucel–related or other severe toxicities (e.g., anaphylaxis). Treatments for the patient’s lymphoma other than what was defined/allowed in the protocol, such as chemotherapy, immunotherapy, targeted agents, radiation, high-dose corticosteroid, and other investigational agents, were prohibited, except as needed for treatment of disease progression after axicabtagene ciloleucel infusion.

Outcomes

A list of the efficacy end points assessed in this review is provided in Table 6. Summarized end points are based on outcomes included in the sponsor’s Summary of Clinical Evidence as well as any outcomes identified as important according to the clinical experts consulted by CADTH and stakeholder input from patient and clinician groups and public drug plans. Using the same considerations, the CADTH review team selected end points they considered to be most relevant for informing the CADTH Canadian Drug Expert Committee (CDEC) in its deliberations and finalized this list of end points in consultation with members of the expert committee.

The primary end point of the ZUMA-5 trial was ORR, defined as the incidence of a complete response (CR) or a partial response, as determined by central assessment. These end points were defined by the Lugano classification criteria. Key secondary end points determined by central assessment included the CRR, defined as the incidence of CR as the best response to treatment, and the ORR and CRR in patients who had 3 or more lines of prior therapy. Other secondary end points included the best overall response, defined as the incidence of CR, partial response, stable disease, progressive disease, or NE as the best response to treatment, as adjudicated by central assessment. The DOR was measured in patients who had an objective response and was defined as the time from the first objective response to disease progression or death. PFS was defined as the time from the date of axicabtagene ciloleucel infusion for the inferential or safety analysis sets (or date of leukapheresis for the FAS) to the date of disease progression or death, while OS was defined as the time from the date of axicabtagene ciloleucel infusion for the inferential or safety analysis sets (or date of leukapheresis for the FAS) to the date of death. TTNT was defined as the time from the date of axicabtagene ciloleucel infusion to the start of new lymphoma therapy or death. Patient-reported outcomes were not reported in the ZUMA-5 trial.

Safety outcomes that occurred with the onset or after infusion of axicabtagene ciloleucel were reported as TEAEs. The incidence of TEAEs and SAEs were summarized by preferred term, according to the Medical Dictionary for Regulatory Activities (MedDRA) Version 25.0. The severity of AEs was graded using the National Cancer Institute’s Common Terminology Criteria for Adverse Events Version 4.03 or more recent editions. AEs of special interest were also reported and include CRS, neurologic events, cytopenias, infection, and hypogammaglobulinemia.

Table 6: Outcomes Summarized From the Studies Included in the Systematic Review

CR = complete response; CRR = complete response rate; DOR = duration of response; ORR = objective response rate; OS = overall survival; PFS = progression-free survival; PR = partial response; TTNT = time to next treatment.

^aAll other outcomes were analyzed in patients with 2 or more lines of prior therapy unless otherwise specified.

Details included in this table are from the sponsor’s Summary of Clinical Evidence.⁵⁹

Source: ZUMA-5 Clinical Study Report.¹⁸

Statistical Analysis

The FAS of the ZUMA-5 trial consists of all enrolled patients (n = 127). The inferential analysis set was defined as the first 80 patients followed for 18 months after axicabtagene ciloleucel treatment. For any patient who could have but did not attend the 18-month study visit, an additional patient was added, resulting in 86 patients included in the inferential analysis set. This sample size (n = 86) provided 93% power to reject the null hypothesis of an ORR of less than or equal to 40% at the alpha level of 0.0237 under an assumed alternative ORR of 60%.

An interim analysis was conducted for efficacy outcomes when 30 patients had been followed for 6 months after axicabtagene ciloleucel infusion. The nominal alpha level for the assessment of efficacy for this analysis was 0.0003. Another interim analysis was conducted when 80 patients had been followed for 6 months after axicabtagene ciloleucel infusion. The nominal alpha level for the assessment of efficacy for this analysis was 0.0005. Another interim analysis was conducted when at least 80 patients had been followed for at least 9 months after the first disease response assessment. The nominal alpha level for the assessment of efficacy for this analysis was 0.0005. Primary analysis was to be performed when at least 80 patients in the inferential analysis set had been followed for 12 months after the first disease response assessment, reserving alpha of 0.0237 for the final significance test.

At each interim analysis, 4 hypotheses were tested in the inferential analysis set using a fixed sequence procedure to control overall type I error at 1-sided alpha level of 0.025. In order, the hypotheses tested were ORR as determined by central assessment, CRR as determined by central assessment, ORR as determined by central assessment in the patients who had had 3 or more lines of prior therapy, and CRR as determined by central assessment in the patients who had had 3 or more lines of prior therapy. Each significance test used the alpha allocated at the time of interim analysis.

Analyses conducted on the FAS at the 18-month and 36-month cut-off dates were intended to be descriptive, and power calculations were not conducted.

Subgroup Analyses

Prespecified baseline subgroups used to examine key efficacy and safety analyses include age (< 65 years, ≥ 65 years), sex, race, ethnicity, FLIPI total score, ECOG PS score (0, 1), meeting the criteria for high tumour bulk load as per the Groupe d’Étude des Lymphomes Folliculaires (GELF) versus not meeting these criteria, relapsed versus refractory at study entry, time to relapse from initiation of first anti-CD20 chemotherapy combination therapy (≥ 24 months, < 24 months), prior treatment with PI3K inhibitor, number of lines of prior therapy, and double refractory status.

Table 7: Statistical Analysis of Efficacy End Points

CI = confidence interval; CR = complete response; DOR = duration of response; FAS = full analysis set; KM = Kaplan-Meier; NA = not applicable; ORR = objective response rate; OS = overall survival; PFS = progression-free survival; TTNT = time to next treatment.

^aPer central assessment.

Details included in this table are from the sponsor’s Summary of Clinical Evidence.⁵⁹

Source: ZUMA-5 Clinical Study Report.¹⁸

Analysis Populations

A summary of the analysis populations in the ZUMA-5 trial is shown in Table 8.

Results

Patient Disposition

A summary of patient disposition in the ZUMA-5 trial is presented in Table 9. At the 36-month analysis, ||| patients with either FL or MZL had been screened, with 127 patients enrolled in the FL group and included in the FAS. Discontinuation from the study occurred for ||| of patients, with the majority discontinuing due to death (|||||). Median follow-up at the 36-month analysis was ||||| |||||||||||||| || ||| || |||| ||||||). The safety analysis set included only the 124 patients who received an axicabtagene ciloleucel infusion. Patient disposition for the inferential 18-month analysis was not reported.

Table 8: Analysis Populations in the ZUMA-5 Trial

FAS = full analysis set.

Details included in this table are from the sponsor’s Summary of Clinical Evidence.⁵⁹

Source: ZUMA-5 Clinical Study Report.¹⁸

Table 9: Summary of Patient Disposition From Studies Included in the Systematic Review

FAS = full analysis set; PP = per protocol; SD = standard deviation.

^aPatient disposition was not reported for the 18-month inferential analysis set.

^bReported number of patients screened includes patients with follicular lymphoma (FL) or marginal zone lymphoma. The number of patients with FL only was not reported.

Details included in this table are from the sponsor’s Summary of Clinical Evidence.⁵⁹

Source: ZUMA-5 Clinical Study Report.¹⁸

Baseline Characteristics

A summary of baseline characteristics in the ZUMA-5 trial is shown in Table 10. In the FAS, the median age was 60 years (range, 34 years to 79 years) and 62% of patients had an ECOG PS score of 0. Of the enrolled patients, 69% were refractory, defined as progressing within 6 months of their most recent treatment. Most patients enrolled in the ZUMA-5 trial had received 2 prior therapies (|||), 26% had received 3 prior therapies, 20% had received 4 prior therapies, and 17% had received 5 or more prior therapies. The proportion of patients who had received a prior auto-SCT was 24%, while the proportion of patients with high bulk tumour was 51%. The proportion of patients who had progressed within 24 months of anti-CD20 chemotherapy combination therapy (i.e., POD24) was 55%.

The baseline characteristics outlined in Table 10 are limited to those most relevant to this review as prognostic or effect-modifying variables.

Table 10: Summary of Baseline Characteristics From Studies Included in the Systematic Review

auto-SCT = autologous stem cell transplant; CR = complete response; ECOG PS = Eastern Cooperative Oncology Group Performance Status; FAS = full analysis set; FLIPI = Follicular Lymphoma International Prognostic Index; NE = not evaluable; PI3K = phosphoinositide 3-kinase; POD24 = progression of disease within 24 months; PR = partial response.

^aPatients with FL who progressed within 6 months of completion of the most recent prior treatment are defined as refractory. Patients with FL who progressed 6 months after completing the most recent prior treatment are defined as relapsed.

Details included in this table are from the sponsor’s Summary of Clinical Evidence.⁵⁹

Source: ZUMA-5 Clinical Study Report.¹⁸

Exposure to Study Treatments

As axicabtagene ciloleucel is a single infusion, treatment exposure measures such as total patient-weeks, duration, and adherence are not applicable. Patients in the ZUMA-5 trial were eligible for re-treatment (n = 11 in the 18-month FAS and n = 13 in the 36-month FAS). It is not anticipated that axicabtagene ciloleucel will be eligible for use in re-treatment in Canada.

Table 11: Summary of Concomitant and Subsequent Treatment in the ZUMA-5 Trial

Details included in this table are from the sponsor’s Summary of Clinical Evidence.⁵⁹

Source: ZUMA-5 Clinical Study Report.¹⁸

Efficacy

Overall Survival

A summary of the OS results from the ZUMA-5 trial is shown in Table 12. At the 36-month FAS time point, ||| of patients had died due to any cause, and the median OS had not been reached. The KM-estimated survival probability at 18 months was ||||| ||||| || ||||||, at 24 months was ||||| ||||| || |||||, and at 36 months was 75.5% (95% CI, 66.9% to 82.2%). Results for the 18-month inferential analysis set and FAS are shown for context. The estimated OS curve in the ZUMA-5 trial is shown in Figure 2.

Table 12: OS in the ZUMA-5 Trial

CI = confidence interval; FAS = full analysis set; NE = not evaluable; OS = overall survival.

^aSurvival probabilities are according to Kaplan-Meier estimates.

Source: ZUMA-5 Clinical Study Report.¹⁸

Figure 2: Redacted

CI = confidence interval; FL = follicular lymphoma; NE = not evaluable.

This figure has been redacted.

Source: ZUMA-5 Clinical Study Report.¹⁸

Progression-Free Survival

A summary of the PFS results from the ZUMA-5 trial is shown in Table 13. At the 36-month FAS time point, ||| of patients had experienced a progression event. The median PFS was 40.2 months (95% CI, 28.9 months to NE). The KM-estimated PFS rate at 18 months was ||||| ||||| || |||||, at 24 months was ||||| ||||| || |||||, and at 36 months was 54.4% (95% CI, 44.2% to 63.5%). Results for the 18-month inferential analysis set and 18-month FAS are shown for context. The estimated PFS curve in the ZUMA-5 trial is shown in Figure 3.

Table 13: PFS in the ZUMA-5 Trial

CI = confidence interval; FAS = full analysis set; NE = not evaluable; PFS = progression-free survival.

Note: PFS is reported according to central assessment at the 18-month analysis and as per investigator assessment for the 36-month analysis.

^aSurvival probabilities are according to Kaplan-Meier estimates.

Source: ZUMA-5 Clinical Study Report.¹⁸

Figure 3: Redacted

CI = confidence interval; FAS = full analysis set; NE = not evaluable; PFS = progression-free survival.

This figure has been redacted.

Note: PFS is reported as per investigator assessment for the 36-month analysis.

Source: ZUMA-5 Clinical Study Report.¹⁸

Objective Response Rate

A summary of the response results for the ZUMA-5 trial is shown in Table 14. At the 36-month FAS time point, the ORR as per investigator assessment was 94% (95% CI, 88% to 97%), while the CRR was 79% (|| || ||). The primary end point in the ZUMA-5 trial was ORR at the 18-month analysis in the inferential analysis set. The estimated ORR as per central assessment was 94% (|| || ||) and the CRR was 79% (|| || ||), sufficient to reject the null hypotheses of 40% for ORR and 15% for CRR.

Table 14: ORR in the ZUMA-5 Trial

CI = confidence interval; CRR = complete response rate; FAS = full analysis set; NA = not applicable; NE = not evaluable; ORR = objective response rate; PR = partial response.

^a95% CI from the Clopper-Pearson method.

^bHypothesis testing was conducted on an earlier 18-month data cut-off with 84 patients yielding significant P values (P < 0.0001) for both ORR and CRR.

^cPatients who were treated with axicabtagene ciloleucel but died before first disease assessment were recorded as having a “not done” response.

Note: ORR is reported according to central assessment at the 18-month analysis and as per investigator assessment for the 36-month analysis.

Source: ZUMA-5 Clinical Study Report.¹⁸

The ORR results for selected subgroups, based on discussion with clinical experts consulted by CADTH, are presented in Table 15. The results appear to be consistent across all subgroups, with the ORR from |||| ||| || |||| in patients who received 1 prior line of therapy to ||| ||| || ||| in patients whose disease had progressed within less than 24 months.

The CRR results for selected subgroups, based on discussions with clinical experts consulted by CADTH, are presented in Table 16. The results appear to be consistent across all subgroups, with the CRR from |||| ||| || | in patients who received 1 prior line of therapy to ||| ||| || ||| in patients who received 3 prior therapies.

Table 15: Subgroup Analysis of ORR as per Investigator Assessment — 36-Month FAS

CI = confidence interval; FAS = full analysis set; GELF = Groupe d’Étude des Lymphomes Folliculaires; ORR = objective response rate.

Note: ORR is reported as per investigator assessment for the 36-month analysis.

^a95% CI is from the Clopper-Pearson method.

Source: ZUMA-5 Clinical Study Report.¹⁸

Table 16: Subgroup Analysis of CRR as per Investigator Assessment — 36-Month FAS

CI = confidence interval; CRR = complete response rate; FAS = full analysis set; GELF = Groupe d’Étude des Lymphomes Folliculaires.

Note: CRR is reported as per investigator assessment for the 36-month analysis.

^a95% CI is from the Clopper-Pearson method.

Source: ZUMA-5 Clinical Study Report.¹⁸

Duration of Response

A summary of the DOR results from the ZUMA-5 trial is shown in Table 17. At the 36-month FAS time point, ||| of patients no longer demonstrated a response, and the median DOR was 38.6 months (95% CI, 29.0 months to NE). The KM-estimated event-free probability at 18 months was ||||| ||||| || |||||, at 24 months was ||||| ||||| || |||||, and at 36 months was 57.1% (95% CI, 46.6% to 66.3%).

Table 17: DOR in the ZUMA-5 Trial

CI = confidence interval; FAS = full analysis set; NE = not evaluable; NR = not reported.

Note: DOR is reported according to central assessment at the 18-month analysis and as per investigator assessment for the 36-month analysis.

^aReasons for censoring for DOR may include: response ongoing, lost to follow-up, investigator decision, started new anticancer therapy, stem cell transplant, re-treatment.

^bSurvival probabilities are according to Kaplan-Meier estimates.

^cMedian follow-up time derived using the reverse Kaplan-Meier approach.

Source: ZUMA-5 Clinical Study Report.¹⁸

Time to Next Treatment

A summary of the TTNT results from the ZUMA-5 trial is shown in Table 18. At the 36-month FAS time point, ||| of patients had experienced a TTNT event; the median TTNT was NE months (95% CI, 37.8 months to NE). The KM-estimated event-free probability at 18 months was ||||| ||||| || |||||, at 24 months was ||||| ||||| || |||||, and at 36 months was 59.5% (95% CI, 50.2% to 67.6%).

Table 18: TTNT in the ZUMA-5 Trial

CI = confidence interval; FAS = full analysis set; NE = not evaluable; NR = not reported; TTNT = time to next treatment.

^aReasons for censoring for TTNT may include: alive and without new anticancer therapy, lost to follow-up, withdrawal of consent, investigator decision, and end of study due to other reason.

^bSurvival probabilities are according to Kaplan-Meier estimates.

^cMedian follow-up time derived using the reverse Kapan-Meier approach.

Source: ZUMA-5 Clinical Study Report.¹⁸

Harms

Refer to Table 19 for a summary of harms data in the ZUMA-5 trial.

Adverse Events

At the 36-month time point for analysis, a total of ||| of patients in the safety analysis set experienced a TEAE, with pyrexia (|||||) hypotension (|||), headache (|||), and fatigue (|||) the most frequently reported TEAEs.

Serious Adverse Events

At the 36-month analysis, a total of ||| of patients in the safety analysis set experienced an SAE, with pyrexia (|||||), pneumonia (||||), encephalopathy (||||), and confusional state (||||) the only SAEs reported by at least 5% of patients.

Withdrawals Due to Adverse Events

As axicabtagene ciloleucel is a 1-time infusion, withdrawals due to adverse events are not applicable.

Mortality

At the 36-month time point of analysis, ||| of patients in the safety analysis set had died. The most common reason was due to progressive disease (||), following by AE due to reasons other than progressive disease or subsequent therapy (||) and secondary malignancy (||). Most “other” classifications of death were related to various infections.

Notable Harms

Notable harms identified by CADTH in consultation with the clinical experts consulted on this review included CRS, neurologic events, cytopenias, infection, and hypogammaglobulinemia. At the 36-month analysis, ||| of patients in the safety analysis set had experienced CRS, with || experiencing a grade 3 or higher CRS. Neurologic events were reported in ||| of patients, with ||| reporting a grade 3 or higher neurologic event. Cytopenias were reported in ||| of patients, with ||| of patients reporting a grade 3 or higher cytopenia. Infections were reported in ||| of patients, with ||| reporting a grade 3 or higher infection. Hypogammaglobulinemia was reported in ||| of patients, with || of patients reporting a grade 3 or higher hypogammaglobulinemia.

Table 19: Summary of Harms Results From Studies Included in the Systematic Review

AE = adverse event; AESI = adverse event of special interest; CRS = cytokine release syndrome; GVHD = graft vs. host disease; NA = not applicable; SAE = serious adverse event; TEAE = treatment-emergent adverse event.

^aPresent in ≥ 20% of patients.

^bPresent in > 5% of patients.

Details included in this table are from the sponsor’s Summary of Clinical Evidence.⁵⁹

Source: ZUMA-5 Clinical Study Report.¹⁸

Critical Appraisal

Internal Validity

The ZUMA-5 trial, the only eligible study identified by the sponsor, was a phase II, single-arm, open-label clinical trial. The lack of comparative data is a key limitation to the interpretation of the results from a single-arm trial, as it is difficult to distinguish between the effect of the intervention, a placebo effect, or the effect of natural history.^54,55 It is acknowledged that there may be practical limitations to conducting a randomized controlled trial in patients with r/r FL (beyond first-line treatment), such as decreasing population size with subsequent lines of therapy and lack of a standard treatment in these later lines of treatment. However, the sponsor’s submission included information on the phase III ZUMA-22 trial,⁵⁶ which is studying axicabtagene ciloleucel compared to standard of care, and which is currently recruiting patients with results expected in 2027;⁵⁶ this suggests that it is possible to conduct a phase III comparative trial in this population. The hypothesized historical control of 40% ORR and 15% CRR was used to determine a clinically meaningful benefit. The prespecified thresholds were established against the response rate derived from studies of available PI3K inhibitors for the treatment of r/r FL.⁵⁷ The clinical experts consulted by CADTH supported using a CRR of 15% and an ORR of 40% as clinically relevant thresholds.

Due to the open-label design of the trial, the response outcomes measures (i.e., ORR, DOR, PFS) and subjective harms are at risk of measurement or reporting bias, though the direction of this bias is unclear. The primary end point of ORR was assessed by central assessment at the 18-month primary analysis, which reduces the likelihood that the open-label nature of the trial impacted ORR. Although response rates at the 36-month time point were only assessed by the investigator, which can impact reporting of AEs, response rates as measured by central and investigator assessments were similar at the 18-month analysis, which suggests accurate reporting of the ORR by investigators.

Follow-up time was deemed sufficient for assessing tumour response and safety outcomes associated with axicabtagene ciloleucel. The clinical experts consulted by CADTH noted that r/r FL is a disease that can have very long periods of PFS and survival, suggesting that the follow-up duration was not long enough to fully capture the effects on OS and PFS. In addition to the duration of the study and the noncomparative design, subsequent treatments make it difficult to interpret the OS results. After the infusion of axicabtagene ciloleucel, ||| of the patients received at least 1 subsequent antineoplastic medication, and ||| of them received SCT. The estimated clinical end point curve (OS) should be interpreted in cases where subsequent treatments are given as the magnitude of patient benefit due to treatment is difficult to quantify.

As treatment with axicabtagene ciloleucel is a 1-time infusion, adherence to treatment is not a concern for internal validity. Of the 127 patients included in the 36-month FAS, ||| patients (|||) had discontinued the study before receiving axicabtagene ciloleucel. Clinical experts consulted by CADTH considered this to be a realistic proportion of patients who are unable to complete the treatment process, and representative of clinical practice.

External Validity

According to the clinical experts consulted by CADTH, the ZUMA-5 study population overall represents the patients in the population with r/r FL in Canada who would be receiving axicabtagene ciloleucel. However, the clinical experts noted that patients seen in clinical practice would include those with poorer performance status (the ZUMA-5 trial only included patients with an ECOG PS score of 0 or 1, whereas clinical experts suggest that patients with an ECOG PS score of 2 may be treated in the clinical setting), as well as patients with more comorbidities. The clinical experts differed in their opinions regarding patients who received prior CD19-targeted therapy; some suggested that any prior CD19-targeted therapy would preclude the use of axicabtagene ciloleucel, whereas others suggested that only patients who were refractory to CD19-targeted therapy (did not respond or relapsed within 6 months) would not be suitable candidates for treatment with axicabtagene ciloleucel.

After screening, the procedures and co-interventions (including manufacturing process, depleting chemotherapy, bridging therapy, and post–axicabtagene ciloleucel interventions) were consistent with those adopted in the Canadian setting, although some minor discrepancies exist. The ZUMA-5 trial results can nevertheless be generalized to the population of patients in Canada.

Subgroup analyses were conducted based on prespecified patient characteristics. Due to insufficient follow-up time to observe a large number of survival end points, the OS and PFS subgroup analyses were considered unstable and therefore uninformative. Based on the subgroup analyses conducted for ORR and CRR, efficacy appears to be present for each patient population based on prior therapies, POD24, tumour bulk, or r/r status.

According to the clinical experts consulted by CADTH, the efficacy outcomes used in this study are clinically relevant and important for the clinical trials in r/r FL, with the notable exception of HRQoL outcomes, which are important to patients but were excluded from the ZUMA-5 trial. As such, it is not possible to determine how the introduction of axicabtagene ciloleucel will impact the HRQoL of patients in Canada.

Lack of long-term data on patients’ survival and response rates is another limitation, given that FL is an indolent and slowly progressing disease. Clinical benefits of the treatment need to be evaluated over a longer follow-up time to increase confidence in the durability of response expected from the use of axicabtagene ciloleucel in Canada.

Studies Addressing Gaps in the Systematic Review Evidence

Contents within this section have been informed by materials submitted by the sponsor. The following has been summarized and validated by the CADTH review team.

Objectives for the Summary of Other Relevant Evidence

The ZUMA-5 trial is a single-arm trial and therefore does not provide evidence of efficacy against standard of care for patients with r/r FL who have received 2 or more lines of therapy. Treatment choice for r/r FL is dependent on a variety of factors, including prior therapies, duration of remission, patient-related factors such as age, and clinician/patient preferences. As such, the comparator patients receive can vary considerably. The sponsor submitted evidence of relative efficacy of axicabtagene ciloleucel against standard of care therapies in patients with r/r FL who have received 2 or more lines of therapy.

Description of Other Relevant Evidence

The relative efficacy of axicabtagene ciloleucel versus standard of care was estimated in the ZUMA-5 population using propensity scores with SMR weights.¹⁹ SMR weighting estimates the treatment effect in a population with an equal distribution of risk factors to that of the treatment study participants only.⁵⁸ SCHOLAR-5, the standard of care cohort, is a retrospective, observational, multicentre, database study of patients with r/r FL (grades 1, 2, or 3a) who have received 2 or more systemic therapies. The SCHOLAR-5 cohort was derived from 3 international cohorts: IQVIA, Vanderbilt, and DELTA. The IQVIA and Vanderbilt cohorts were created from electronic medical records, while the DELTA cohort represented patients from the DELTA clinical trial (NCT01282424) who proceeded to receive therapy after idelalisib treatment. The DELTA cohort was added to increase statistical power for OS, TTNT, and response outcomes. However, the DELTA cohort was not used to inform PFS as tumour assessment dates were not provided for the line of therapy subsequent to idelalisib (i.e., the index line of therapy). The index line of therapy was chosen at random from all lines of therapy received by a patient after they met all the eligibility criteria for the study. The index date for the primary analysis was defined as the initiation date of the patient’s index line. A summary of the SCHOLAR-5 study is provided in Table 20.

Table 20: Study Selection Criteria and Methods for Other Relevant Evidence Submitted by the Sponsor

allo-SCT = allogeneic stem cell transplant; auto-SCT = autologous stem cell transplant; CNS = central nervous system; CR = complete response; DOR = duration of response; ECOG PS = Eastern Cooperative Oncology Group Performance Status; FL = follicular lymphoma; iNHL = indolent non-Hodgkin lymphoma; ORR = objective response rate; OS = overall survival; PFS = progression-free survival; PI3K = phosphoinositide 3-kinase; PR = partial response; r/r = relapsed or refractory; TTNT = time to next treatment.

^aEligibility criteria were not restricted by date for patients from the DELTA clinical trial.

^bPatients who pass all other inclusion/exclusion criteria but transform at a later date were eligible until transformation.

Details included in this table are from the sponsor’s Summary of Clinical Evidence.⁵⁹

Source: Sponsor-provided analysis report.¹⁹

Analysis Methods

Patient-level data for the ZUMA-5 and SCHOLAR-5 studies were used to inform a comparative analysis. Propensity scores were calculated for each patient in the pooled analysis set to account for differences in baseline characteristics across populations. These calculated scores were then used to apply SMR weighting, with differences required to be less than 0.1. All ZUMA-5 patients were included in the analysis. Selection of variables for the propensity score model was determined in a hierarchal manner and based on the advice of investigators or clinical experts, with the goal of minimizing the imbalance in prognostically important covariates. For high and medium priority rank variables with less than 40% missing data, multiple imputation was performed.

The variables included in the propensity score model were:

• POD24

• number of lines of prior therapy

• r/r to prior line of therapy

• prior SCT (yes/no)

• tumour bulk

• time since last treatment

• response to prior line of therapy

• age

• prior anti-CD20 + alkylator combination treatment.

Weighted logistic regression was used to estimate ORs of the ORR and CRR across cohorts. A weighted KM estimator of the risk probabilities and 95% CIs for OS, PFS, TTNT, and DOR were estimated at 3-month intervals.¹⁹ Time-to-event outcomes were summarized via hazard ratios, estimated using a Cox proportional hazard model. Statistical sensitivity analyses of ORR and OS were conducted to support the robustness of the findings. Additional sensitivity analyses performed for all outcomes included: exclusion of DELTA trial patients from the SCHOLAR-5 cohort, using the date of axicabtagene ciloleucel infusion as the start point for time-to-event variables, and using the safety and inferential analysis sets. Prespecified subgroup analyses included investigations on patient disease status and prior therapies. A summary of the methods is shown in Table 21.

Table 21: Analysis Methods

CI = confidence interval; CR = complete response; DOR = duration of response; FAS = full analysis set; HR = hazard ratio; ORR = overall response rate; OS = overall survival; PFS = progression-free survival; POD24 = progression of disease within 24 months; PR = partial response; PS = propensity score; PSM = propensity score matching; SCT = stem cell transplant; SMR = standardized mortality ratio; TTNT = time to next treatment; vs. = versus.

Note: Low priority variables were not included in the propensity score weighting.

Details included in this table are from the sponsor’s Summary of Clinical Evidence.⁵⁹

Source: Sponsor-provided analysis report.¹⁹

Results

Summary of Included Studies

The comparative analysis contrasted 2 populations, those in the ZUMA-5 and SCHOLAR-5 studies, to derive estimates of relative efficacy for axicabtagene ciloleucel and standard of care. The ZUMA-5 trial is a single-arm, phase II study evaluating the efficacy and safety of axicabtagene ciloleucel in patients with r/r grade 1, 2, or 3a FL (n = 127). The SCHOLAR-5 cohort was constructed retrospectively from 3 international observational cohorts: IQVIA, Vanderbilt, and DELTA. The DELTA cohort included patients from the DELTA clinical trial (NCT01282424) who proceeded to receive therapy after idelalisib treatment. The DELTA cohort in the SCHOLAR-5 trial did not report progression assessment dates and had to be removed from the PFS and DOR analysis. Table 22 summarizes the assessment of homogeneity between the ZUMA-5 trial and the SCHOLAR-5 cohort.

Table 22: Assessment of Homogeneity

CAR = chimeric antigen receptor; CNS = central nervous system; CR = complete response; DOR = duration of response; ECOG PS = Eastern Cooperative Oncology Group Performance Status; iNHL = indolent non-Hodgkin lymphoma; FL = follicular lymphoma; LOT = line of therapy; ORR = ; OS = overall survival; PFS = progression-free survival; POD24 = progression of disease within 24 months; PR = partial response; r/r = relapsed or refractory; TTNT = time to next treatment.

Details included in this table are from the sponsor’s Summary of Clinical Evidence.⁵⁹

Source: Sponsor-provided analysis report.¹⁹

Table 23 summarizes the index treatment patterns in the SCHOLAR-5 cohort before propensity score weighting. The most common index treatments, in descending order, were anti-CD20 agent plus bendamustine (16.8%), experimental therapy (16.1%), immunomodulatory agent (12.6%), PI3K inhibitor (9.8%), auto-SCT (8.4%), anti-CD20 agent plus other chemotherapy (8.4%), and other chemotherapy (8.4%).

A full description of the ZUMA-5 trial can be found in the Systematic Review section of this report.

Table 23: Index Treatment Patterns for the SCHOLAR-5 (n = 143) Sample of Patients With 2 or More Lines of Prior Therapy

allo-SCT = allogeneic stem cell transplant; auto-SCT = autologous stem cell transplant; CHOP = cyclophosphamide + doxorubicin-vincristine-prednisone; PI3K = phosphoinositide 3-kinase; SCT = stem cell transplant.

Source: Sponsor-provided analysis report.¹⁹

Results — Propensity Score Weighting

The SCHOLAR-5 cohort was reweighted using the propensity scores to further align with the ZUMA-5 patient population. After propensity score weighting, the SCHOLAR-5 effective sample size was reduced from 143 patients to 128 patients. The largest difference between cohorts was in time from last treatment; after weighting the average was reduced from the original ||||| months in the full SCHOLAR-5 population to |||| months, and the proportion of patients with POD24 increased from 35.7% in the unadjusted population to 57.1% in the propensity score–weighted population. After propensity score weighting, variables were mostly similar to the ZUMA-5 population. The removal of the DELTA cohort from the SCHOLAR-5 cohort also resulted in relatively balanced variables; however, the proportion of patients with POD24 and the proportion who were refractory to their most recent treatment were both lower.

A summary of the patient characteristics not included in the propensity score model for both the ZUMA-5 and SCHOLAR-5 trials before and after weighting is shown in Table 25. The largest differences between the post-weighting SCHOLAR-5 and ZUMA-5 trials are in the proportion of patients with an ECOG PS score of 0, that is, 32.6% versus 62.2%, respectively. Other large differences between the SCHOLAR-5 and ZUMA-5 trials are in the proportion of patients who are double refractory (||||| versus |||||, respectively) and have bone marrow involvement (||||| versus |||||, respectively).

Table 24: Propensity Score Variables in the ZUMA-5 and SCHOLAR-5 Trials

CR = complete response; POD24 = progression of disease within 24 months; PR = partial response; SCT = stem cell transplant; SD = standard deviation; SMD = standardized mean difference.

Source: Sponsor-provided analysis report.¹⁹

Table 25: Patient Characteristics not Included in the Propensity Score Model in the ZUMA-5 and SCHOLAR-5 Trials

ECOG PS = Eastern Cooperative Oncology Group Performance Status; FL = follicular lymphoma; FLIPI = Follicular Lymphoma Internal Prognostic Index; PI3K = phosphoinositide 3-kinase; SD = standard deviation; SMD = standardized mean difference.

Source: Sponsor-provided analysis report.¹⁹

A summary of the estimated relative efficacy is provided in Table 26. The ORR in the ZUMA-5 population was 93.7% compared to 54.0% in the propensity score–weighted SCHOLAR-5 population (OR = 12.66; 95% CI, 5.24 to 30.57; P < 0.001). The CRR in the ZUMA-5 population was 78.7% compared to 34.9% in the propensity score–weighted SCHOLAR-5 population (OR = 6.90; 95% CI, 3.62 to 13.18; P < 0.001).

The median PFS in the ZUMA-5 population was 40.21 months (95% CI, 28.94 months to NE) compared to 12.97 months (95% CI, 7.75 months to 15.47 months) in the propensity score–weighted SCHOLAR-5 population, with HR of 0.27 (95% CI, 0.18 to 0.41; P < 0.0001. The KM curve for the PFS analysis is shown in Figure 4. The median OS in the ZUMA-5 population was NE (95% CI, NE to NE) compared to NE (95% CI, 38.40 months to NE) in the propensity score–weighted SCHOLAR-5 population (HR = 0.56; 95% CI, 0.33 to 0.95; P = 0.0303). The KM curve for the OS analysis is shown in Figure 5. The median TTNT in the ZUMA-5 population was NE (95% CI, 37.85 months to NE) compared to 26.61 months (95% CI, 12.65 months to NE) in the propensity score–weighted SCHOLAR-5 population (HR = 0.25; 95% CI, 0.15 to 0.41; P < 0.0001).

The median DOR in the ZUMA-5 population was 38.64 months (95% CI, 29.04 months to NE) compared to ||||| |||||| ||||| || |||||| in the propensity score–weighted SCHOLAR-5 population, with HR of |||| ||||| || |||||; P ||||||.

Table 26: Summary of Estimated Relative Efficacy in r/r FL

CI = confidence interval; CRR = complete response rate; DOR = duration of response; HR = hazard ratio; NE = not evaluable; OR = odds ratio; ORR = objective response rate; OS = overall survival; PFS = progression-free survival; TTNT = time to next treatment.

Note: OR > 1 and HR < 1 demonstrate a relative benefit of axicabtagene ciloleucel vs. the comparator basket.

Source: Sponsor-provided analysis report.¹⁹

Figure 4: PFS in the ZUMA-5 Trial (36-Month FAS) Compared to the SCHOLAR-5 Study (Propensity Score Weighted)

FAS = full analysis set; PFS = progression-free survival.

Source: Sponsor-provided analysis report.¹⁹

Figure 5: OS in the ZUMA-5 Trial (36-Month FAS) Compared to the SCHOLAR-5 Trial (Propensity Score Weighted)

FAS = full analysis set; OS = overall survival.

Source: Sponsor-provided analysis report.¹⁹

Critical Appraisal of Other Relevant Evidence

Internal Validity

The submitted comparative analysis provides an estimate of relative efficacy of axicabtagene ciloleucel, using evidence from the ZUMA-5 clinical trial, against a basket of therapies representing standard of care, using evidence from the SCHOLAR-5 cohort, a combination of retrospective real-world evidence and a subset of the DELTA clinical trial.

Due to differences between the ZUMA-5 and SCHOLAR-5 cohorts in treatment allocation, it is possible that the treatment effect estimate is confounded by imbalances in prognostic covariates across populations. The sponsor identified and adjusted for several important variables, resulting in a suitable balance of these characteristics across both populations. However, characteristics deemed critical by the clinical experts — FLIPI score, ECOG PS score, and double refractory status — were omitted; FL grade was also omitted. Of the variables collected in both cohorts, there were observed imbalances across treatment groups even after reweighting the populations. The variables were acknowledged by the sponsor, but were omitted due to concerns regarding missing data. The clinical experts consulted by CADTH suggested that differences in ECOG PS scores and the proportion of patients who are double refractory could affect how patients would be expected to respond to treatment. The direction of this impact is uncertain, with some differences potentially favouring axicabtagene ciloleucel over the SCHOLAR-5 comparator. There is additional uncertainty in the results because of the low effective sample sizes in both the ZUMA-5 trial and the SCHOLAR-5 cohort. The removal of the DELTA cohort resulted in a statistically significant change in the mean number of lines of prior therapy, that is, |||| |||||| in the SCHOLAR-5 study compared to |||| |||||| in the ZUMA-5 trial. Differences between populations in the number of lines of prior therapy have a particularly large impact when determining how patients would be expected to respond to treatment. The proportion of patients with POD24 and the proportion of patients who were refractory to their most recent treatment were also both reduced with the exclusion of the DELTA cohort, indicating that the removal of this cohort results in a population with a lower risk prognosis. The ZUMA-5 trial and the SCHOLAR-5 study had different follow-up time requirements: the ZUMA-5 cohort used a 36-month follow-up, while the SCHOLAR-5 cohort required at least 12 months of follow-up. It is possible that individuals’ characteristics or issues with treatment adherence determined patients’ decisions to exit the cohort early and introduced informative censoring, which bias the results.

External Validity

The clinical experts confirmed that the distribution of therapies in the SCHOLAR-5 cohort were representative of the standard of care for patients in Canada; however, it is unclear how this distribution was affected by the propensity score weighting or the exclusion of the DELTA cohort, as the comparator therapies used were not reported in the post-weighting population. The inclusion and exclusion criteria for the SCHOLAR-5 cohort was similar to the ZUMA-5 trial, although it is uncertain if the results can be generalized beyond this selected group of patients.

Future Planned Studies

Contents within this section have been informed by materials submitted by the sponsor. The following have been summarized and validated by the CADTH review team.

Table 27: Summary of Gaps in the Systematic Review Evidence

FL = follicular lymphoma; HRQoL = health-related quality of life; NA = not applicable; POD24 = progression of disease within 24 months; r/r = relapsed or refractory.

Source: Sponsor’s Summary of Clinical Evidence.⁵⁹

Description of Studies

The ZUMA-22 trial⁵⁶ is a prospective, interventional, open-label, randomized, phase III study assessing the efficacy and safety of axicabtagene ciloleucel versus standard of care in patients with r/r FL after 2 or more lines of systemic therapy or r/r FL after first-line chemoimmunotherapy in patients with POD24. The study began in September 2022 and is actively recruiting participants. Final study results of the ZUMA-22 trial are expected by Health Canada in |||| as a requirement for the issued Notice of Compliance for adult patients with r/r grade 1, 2, or 3a FL after 2 or more lines of systemic therapy.

Discussion

Summary of Available Evidence

One clinical trial was included in the systematic review conducted by the sponsor. The ZUMA-5 study is a phase II, open-label, single-arm study that evaluated the efficacy and safety of axicabtagene ciloleucel in patients with r/r grade 1, 2, or 3a FL after 2 or more lines of systemic therapy. The primary end point was ORR per central assessment in the inferential analysis set at the 18-month time point. Secondary end points included CRR, OS, PFS, DOR, and TTNT. A total of 127 patients with r/r FL were enrolled, with 86 included in the 18-month primary analysis. Data for the full analysis up to the 36-month time point were available at the time of this review (the data cut-off was March 31, 2022). The median age in the FAS was 60 years (range, 34 years to 79 years). More males (59%) were enrolled than females (41%). Most patients had an ECOG PS score of 0 (62%), and the most common number of prior therapies was 2 (|||||). The proportion of patients with POD24 was 55%.

The sponsor provided an additional study in which patient-level data from the retrospective SCHOLAR-5 study were reweighted using propensity scores to be comparable with the ZUMA-5 populations. Given that the ZUMA-5 trial is a single-arm noncomparative study, adjusting the SCHOLAR-5 patient population to be more similar to the ZUMA-5 population allows for an estimate of comparative efficacy against standard of care within the ZUMA-5 population. The outcomes analyzed were ORR, CRR, DOR, OS, PFS, and TTNT. The treatments that made up the basket of therapies in the SCHOLAR-5 study were any available treatment for r/r FL, including approved and experimental therapies, and autologous and allogeneic transplant.

Interpretation of Results

Efficacy

In the ZUMA-5 trial, based on the primary end point ORR, 94% (95% CI, 88% to 97%) of patients achieved a partial response to axicabtagene ciloleucel treatment 36 months following infusion, and based on the secondary end point CRR, 79% (95% CI, NE to NE) of patients achieved a complete response to treatment. The results of subgroup analyses were consistent with those of the FAS. The clinical experts consulted by CADTH indicated that the ORR and CRR results were clinically important and acceptable surrogates for survival outcomes (PFS and OS) based on their clinical experience in treating patients with r/r FL where extended survival is common, and therefore trials with mature survival follow-up are rare.

Survival outcomes were identified by CADTH with input from patient groups and clinicians as the most important efficacy outcomes to assess treatment effect in patients with r/r FL. Prolonged survival may be correlated with high response rates (e.g., ORR and CRR). In a meta-analysis evaluating the relationship between response rates and median PFS in patients with NHL (including FL, which accounted for 23% of the study population), strong correlation between response rates and PFS was found, the coefficient of determination (R²) was 0.78 for ORR versus median PFS and 0.74 for CRR versus median PFS. The results were similar in the subgroup of patients with r/r FL and treatment-naive FL.⁶⁰ In another meta-analysis examining the correlation between response and survival outcomes, a moderate correlation was observed between CRR and median PFS in patients with FL (R² = 0.69). In this study, the authors noted that since the median OS was usually not reached in clinical trials of NHL, none of the median OS-related correlation analysis results were evaluable due to the limited data.⁶¹ The clinical experts supported the use of surrogate outcomes by noting that, in their experience, patients who achieve complete remission after CAR T-cell therapy have better prognosis (e.g., more favourable survival) compared to those who do not respond well, but this is not always the case.

At the 36-month analysis time point in the ZUMA-5 trial, the survival probability for patients treated with axicabtagene ciloleucel was 75.5% (95% CI, 66.9% to 82.2%) and the median OS had not been reached. PFS was measured from the time of leukapheresis to the date of disease progression or death due to any cause. The proportion of patients who were progression free at the 36-month time point was 54.4% (95% CI, 44.2% to 63.5%), the median PFS was 40.2 months (95% CI, 28.9 months to NE). According to the clinical experts consulted by CADTH, the survival results from the ZUMA-5 trial are consistent with the expected clinical benefit; however, it must be noted that due to the slowly progressing nature of r/r FL, the OS and PFS data are immature and therefore the effect on long-term survival is uncertain. Results from a retrospective analysis conducted in a single centre in the US showed that median OS was 11.7 years, 8.8 years, and 5.3 years for patients who received second-line, third-line, and fourth-line treatments, respectively. In this study, recurrent uses of single-agent rituximab (9% to 31%), alkylator-based chemotherapy (22% to 26%), and radiotherapy (alone or radioimmunotherapy, 10% to 18%) were common in second- to sixth-line treatment therapy. Ten percent of the treated patients received SCT during their course of therapy (auto-SCT 6%; allo-SCT 4%). Investigational therapies (not specified, and unclear whether CAR T-cell products were used) ranged from 8% to 22% when second or later lines of therapy were required.⁶² Furthermore, when considering the long survival periods in r/r FL, it is important to take into account the impact of any beneficial therapies that could be introduced in the future and the impact that they may have on the long-term survival of patients who received axicabtagene ciloleucel in the ZUMA-5 trial.

HRQoL was identified by patient groups and clinicians as an important outcome. The ZUMA-5 trial did not include any HRQoL end points, and therefore the effects of axicabtagene ciloleucel on HRQoL in patients with r/r FL is unknown.

Given the single-arm design of the ZUMA-5 trial, there is no head-to-head evidence against standard of care for patients with r/r FL. A comparison of axicabtagene ciloleucel against an external standard of care control arm from the retrospective SCHOLAR-5 cohort found that axicabtagene ciloleucel is associated with improved OS and PFS. The interpretation of the comparative efficacy estimates is limited by the potential for selection bias when accepting patients into the ZUMA-5 clinical trial, and residual imbalances in important prognostic and effect-modifying patient characteristics, despite propensity score weighting.

It is also noted that tisagenlecleucel is another CAR T-cell therapy currently under review at CADTH for use in patients with r/r FL after 2 or more lines of therapy.⁶³ No comparative evidence between axicabtagene and tisagenlecleucel was identified, and therefore the comparative efficacy is unknown.

Harms

At the 36-month time point, ||| of patients in the ZUMA-5 trial had reported at least 1 AE. The most common were pyrexia (|||), hypotension (|||||), headache (|||), and fatigue (|||||). SAEs were reported in ||| of patients, with the most common being pyrexia (|||||). It was also noted that given that axicabtagene ciloleucel is administered as a 1-time infusion, the ZUMA-5 trial follow-up is sufficient to characterize the safety profile. The clinical experts highlighted the adverse events of special interest, specifically CRS that occurred in ||| of patients (|| at grade ≥ 3) and neurologic events that occurred in ||| of patients (||| at grade ≥ 3). The clinical experts noted that it appears that patients have different rates of CRS and immune effector cell–associated neurotoxicity syndrome. The experts indicated that patients’ characteristics, such as performance status and comorbidities that might predict their ability to tolerate an episode of CRS or immune effector cell–associated neurotoxicity syndrome, might influence the choice of product. However, given the lack of head-to-head evidence or an indirect treatment comparison, the comparative safety profiles are unknown.

The SCHOLAR-5 comparison did not include safety as an end point, and therefore the safety profile of axicabtagene ciloleucel compared to standard of care is unknown. The clinical experts consulted by CADTH considered the safety profile of axicabtagene ciloleucel to be as expected given the mechanism of action and prior experience in other indications; therefore, side effects are expected to be manageable with proper monitoring.

Conclusion

Evidence from a single-arm study (ZUMA-5) suggests that treatment with axicabtagene ciloleucel affects clinically important tumour responses, including complete remission, in adult patients with r/r FL after 2 or more lines of systemic therapies. Due to the single-arm design of the trial and limited duration of follow-up, there is insufficient evidence to determine the magnitude of the effect of axicabtagene ciloleucel on OS and PFS. HRQoL outcomes were not included in the ZUMA-5 trial and therefore the impact of axicabtagene ciloleucel on patients’ HRQoL is unknown. The harms associated with the axicabtagene infusion are as expected given the mechanism of action and prior experience in other indications. The comparison of the ZUMA-5 trial to the retrospective SCHOLAR-5 external control was limited by heterogeneity across study designs and populations. Specifically, the inability to adjust for ECOG PS and double refractory status can introduce bias to the estimation procedure within the comparative populations. Generalizability to individuals who do not meet the ZUMA-5 trial criteria is also in question. Therefore, the magnitude of the comparative efficacy estimates for axicabtagene ciloleucel against standard of care in the Canadian setting is uncertain.

References

1.Lymphoma Canada. Understanding NHL: A patient's guide to Non-Hodgkin Lymphoma. 2023; https://z9a2f9c2.rocketcdn.me/wp-content/uploads/2020/05/2020430-NHL-Interior-WEB.pdf. Accessed 2023 May 4.

2.Canadian Cancer Society. What is Non-Hodgkin lymphoma? 2023; https://cancer.ca/en/cancer-information/cancer-types/non-hodgkin-lymphoma/what-is-non-hodgkin-lymphoma. Accessed 2023 May 4.

3.National Cancer Institute. Non-Hodgkin lymphoma treatment (PDQ®)–health professional version. 2023; https://www.cancer.gov/types/lymphoma/hp/adult-nhl-treatment-pdq#_1. Accessed 2023 May 4.

4.Hubel K, Ghielmini M, Ladetto M, Gopal AK. Controversies in the treatment of follicular lymphoma. Hemasphere. 2020;4(1):e317. PubMed

5.Jaffe ES. Follicular lymphomas: a tapestry of common and contrasting threads. Haematologica. 2013;98(8):1163-1165. PubMed

6.Le M, Ghazawi FM, Alakel A, et al. Incidence and mortality trends and geographic patterns of follicular lymphoma in Canada. Curr Oncol. 2019;26(4):e473-e481. PubMed

7.Pongas G, Cheson B. Recent Advances in the Management of Patients with Relapsed/Refractory Follicular Lymphoma. Blood Lymphat Cancer. 2021;11:55-66. PubMed

8.Ellison LF, Wilkins K. Canadian trends in cancer prevalence. Health Rep. 2012;23(1):7-16. PubMed

9.Canadian Cancer Society. Canadian cancer statistics. 2021; https://cdn.cancer.ca/-/media/files/research/cancer-statistics/2021-statistics/2021-pdf-en-final.pdf. Accessed 2022 Nov 30.

10.Horning SJ. Natural history of and therapy for the indolent non-Hodgkin's lymphomas. Semin Oncol. 1993;20(5 Suppl 5):75-88. PubMed

11.Horning SJ, Rosenberg SA. The natural history of initially untreated low-grade non-Hodgkin's lymphomas. NEJM. 1984;311(23):1471-1475. PubMed

12.Johnson PW, Rohatiner AZ, Whelan JS, et al. Patterns of survival in patients with recurrent follicular lymphoma: a 20-year study from a single center. J Clin Oncol. 1995;13(1):140-147. PubMed

13.Liu J, Stewart D, Fontaine A, et al. Characteristics and outcomes of patients with relapsed follicular lymphoma following retreatment with second-line rituximab-containing chemotherapy. Leuk Lymphoma. 2020;61(10):2492-2496. PubMed

14.Rivas-Delgado A, Magnano L, Moreno-Velazquez M, et al. Response duration and survival shorten after each relapse in patients with follicular lymphoma treated in the rituximab era. Br J Haematol. 2019;184:753-759. PubMed

15.Erblich T, Montoto S. Treating relapsed follicular lymphoma. Expert Rev Hematol. 2018;11(5):403-410. PubMed

16.Cancer Reasearch UK. Non-Hodgkin lymphoma- follicular lymphoma. 2020; https://www.cancerresearchuk.org/about-cancer/non-hodgkin-lymphoma/types/follicular-lymphoma#:~:text=Follicular%20lymphoma%20is%20a%20type,normally%20helps%20us%20fight%20infections. Accessed 2023 May 4.

17.Canadian Cancer Society. Survival statistics for non-Hodgkin lymphoma. 2023; https://cancer.ca/en/cancer-information/cancer-types/non-hodgkin-lymphoma/prognosis-and-survival/survival-statistics. Accessed 2023 May 4.

18.Clinical Study Report KTE-C19-105 (ZUMA-5): A phase 2 multicenter study of axicabtagene ciloleucel in subjects with relapsed/refractory indolent non-Hodgkin lymphoma. NCT03105336 [internal sponsor's report]. Santa Monica (CA): Kite Pharma, Inc; 2021.

19.A comparison of clinical outcomes from ZUMA-5 (axicabtagene ciloleucel) and the international SCHOLAR-5 External Control Cohort In Relapsed/Refractory Follicular Lymphoma (R/R FL): 36-month follow-up for patients with three or more prior lines of therapy and patients with two or more prior lines of therapy, addendum to modules 5.3.5.2 ZUMA-5 clinical study report. [internal sponsor's report]. In: Drug Reimbursement Review sponsor submission: axicabtagene ciloleucel, 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, IV. Mississauga (ON): Gilead Sciences Canada; 2023 Apr 13.

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

21.Canadian Cancer Society. Non-Hodgkin lymphoma statistics. 2022; https://cancer.ca/en/cancer-information/cancer-types/non-hodgkin-lymphoma/statistics. Accessed 2023 May 4.

22.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

23.Sarkozy C, Maurer MJ, Link BK, et al. Cause of Death in Follicular Lymphoma in the First Decade of the Rituximab Era: A Pooled Analysis of French and US Cohorts. J Clin Oncol. 2019;37(2):144-152. PubMed

24.Canadian Cancer Society. Follicular lymphoma. 2023; https://cancer.ca/en/cancer-information/cancer-types/non-hodgkin-lymphoma/what-is-non-hodgkin-lymphoma/follicular-lymphoma. Accessed 2023 May 4.

25.Montoto S, Davies AJ, Matthews J, et al. Risk and clinical implications of transformation of follicular lymphoma to diffuse large B-cell lymphoma. J Clin Oncol. 2007;25(17):2426-2433. PubMed

26.Kuruvilla J, Assouline S, Hodgson D, et al. A Canadian evidence-based guideline for the first-line treatment of follicular lymphoma: joint consensus of the Lymphoma Canada Scientific Advisory Board. Clin Lymphoma Myeloma Leuk. 2015;15(2):59-74. PubMed

27.Solal-Céligny P, Roy P, Colombat P, et al. Follicular Lymphoma International Prognostic Index. Blood. 2004;104(5):1258-1265. PubMed

28.Casulo C, Byrtek M, Dawson KL, et al. Early relapse of follicular lymphoma after rituximab plus cyclophosphamide, doxorubicin, vincristine, and prednisone defines patients at high risk for death: An analysis from the National LymphoCare Study. J Clin Oncol. 2015;33(23):2516-2522. PubMed

29.UHN Princess Margaret Cancer Centre. Lymphoma: indolent lymphomas. (Clinical Practice Guidelines) 2020; https://www.uhn.ca/PrincessMargaret/Health_Professionals/Programs_Departments/Documents/CPG_Lymphoma_IndolentLymphomas.pdf. Accessed 2023 Jan 5.

30.B-Cell lymphomas. (NCCN Guidelines). Plymouth Meeting (PA): National Comprehensive Cancer Network; 2022 Jul 12: https://www.nccn.org/professionals/physician_gls/pdf/b-cell.pdf. Accessed 2022 Oct 4.

31.Follicular lymphoma. (NCCN Guidelines for Patients.). Plymouth Meeting (PA): National Comprehensive Cancer Network; 2022: https://www.nccn.org/patients/guidelines/content/PDF/nhl-follicular-patient.pdf. Accessed 2022 Oct 4.

32.López-Guillermo A, Caballero D, Canales M, Provencio M, Rueda A, Salar A. Clinical practice guidelines for first-line/after-relapse treatment of patients with follicular lymphoma. Leuk Lymphoma. 2011;52 Suppl 3:1-14. PubMed

33.Dreyling M, Ghielmini M, Rule S, et al. Newly diagnosed and relapsed follicular lymphoma: ESMO Clinical Practice Guidelines for diagnosis, treatment and follow-up. Ann Oncol. 2021;32(3):298-308. PubMed

34.Alberta Health Services Guideline Resource Unit. Lymphoma. (Clinical Practice Guideline LYHE-002 V16.) 2021; https://www.albertahealthservices.ca/assets/info/hp/cancer/if-hp-cancer-guide-lyhe002-lymphoma.pdf. Accessed 2022 Oct 5.

35.Lymphoma, chronic leukemia, myeloma: Hodgkin lymphoma. (Cancer Management Manual). Vancouver (BC): BC Cancer Agency; 2013: http://www.bccancer.bc.ca/health-professionals/clinical-resources/cancer-management-manual/lymphoma-chronic-leukemia-myeloma/hodgkin-lymphoma#Treatment. Accessed 2022 Oct 5.

36.Lymphoma Canada. Follicular lymphoma guidelines. 2015; http://www.lymphoma.ca/medpath/#md:cid=1&md:viewname=dashboardTreeView&ca:viewname=CADefault&ca:id=404&car:viewname=CourseActivityResourcesEmpty&cat:viewname=CourseActivityTreeActivityByType. Accessed 2022 Oct 5.

37.Robinson SP, Canals C, Luang JJ, et al. The outcome of reduced intensity allogeneic stem cell transplantation and autologous stem cell transplantation when performed as a first transplant strategy in relapsed follicular lymphoma: an analysis from the Lymphoma Working Party of the EBMT. Bone Marrow Transplant. 2013;48(11):1409-1414. PubMed

38.Buske C, Hutchings M, Ladetto M, et al. ESMO Consensus Conference on malignant lymphoma: general perspectives and recommendations for the clinical management of the elderly patient with malignant lymphoma. Ann Oncol. 2018;29(3):544-562. PubMed

39.Klyuchnikov E, Bacher U, Woo Ahn K, et al. Long-term survival outcomes of reduced-intensity allogeneic or autologous transplantation in relapsed grade 3 follicular lymphoma. Bone Marrow Transplant. 2016;51(1):58-66. PubMed

40.Smith SM, Godfrey J, Ahn KW, et al. Autologous transplantation versus allogeneic transplantation in patients with follicular lymphoma experiencing early treatment failure. Cancer. 2018;124(12):2541-2551. PubMed

41.Sarkozy C, Trneny M, Xerri L, et al. Risk Factors and Outcomes for Patients With Follicular Lymphoma Who Had Histologic Transformation After Response to First-Line Immunochemotherapy in the PRIMA Trial. J Clin Oncol. 2016;34(22):2575-2582. PubMed

42.Teva-Lenalidomide (lenalidomide) 2.5 mg, 5 mg, 10 mg, 15 mg, 20 mg, and 25 mg capsules [product monograph]. Toronto (ON): Teva Canada Ltd; 2022 Sep 13.

43.Lenalidomide plus rituximab chemotherapy for relapsed or refractory indolent B-cell Non-Hodgkin lymphomas. (CADTH Health Technology Review). Can J Health Technol. Vol 12021: https://www.cadth.ca/sites/default/files/pdf/htis/2021/RC1392-R2-RR-NHL_Final.pdf. Accessed 2023 Oct 25.

44.Zydelig (idelalisib) 150 mg and 100 mg tablets [product monograph]. Mississauga (ON): Gilead Sciences Canada Inc; 2015 Feb 19, revised 2020 Aug 18.

45.CADTH Pan-Canadian Oncology Drug Review (pCODR) Expert Review Committee (pERC) final recommendation: Idelalisib (Zydelig - Gilead Sciences). Ottawa (ON): CADTH; 2016: https://www.cadth.ca/sites/default/files/pcodr/pcodr_idelalisib_zydelig_fl_fn_rec.pdf. Accessed 2022 Oct 31.

46.National Cancer Institute. CAR T cells: engineering patients' immune cells to treat their cancers. 2022; https://www.cancer.gov/about-cancer/treatment/research/car-t-cells. Accessed 2023 Apr 8.

47.Kymriah (tisagenlecleucel) cell suspension in infusion bag, 1.2 x 106 to 6.0 x 108 CAR-positive viable T cells for intravenous use [product monograph] Montreal (QC): Novartis Pharmaceuticals Canada; 2018, revised 2022: https://www.ask.novartispharma.ca/download.htm?res=kymriah_scrip_e.pdf&resTitleId=1520. Accessed 2022 Dec 19.

48.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

49.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.

50.Jacobson C, Chavez J, Sehgal A, et al. Interim analysis of ZUMA-5: A phase 2 study of axicabtagene ciloleucel (Axi-cel) in patients with relapsed/refractory indolent non-Hodgkin lymphoma. Am J Hematol. 2020;95(SUPPL 1):S20-S21.

51.Jacobson C, Chavez JC, Sehgal AR, et al. Primary Analysis of Zuma-5: A Phase 2 Study of Axicabtagene Ciloleucel (Axi-cel) in Patients with Relapsed/Refractory (R/R) Indolent Non-Hodgkin Lymphoma (iNHL). Blood. 2020;136(Supplement 1):40-41.

52.Neelapu SS, Chavez JC, Sehgal AR, et al. Long-term follow-up analysis of zuma-5: A phase 2 study of axicabtagene ciloleucel (AXI-CEL) in patients with relapsed/refractory (R/R) indolent non-Hodgkin lymphoma (INHL). Blood. 2021;138(SUPPL 1):93.

53.Jacobson CA, Chavez JC, Sehgal AR, et al. Axicabtagene ciloleucel in relapsed or refractory indolent non-Hodgkin lymphoma (ZUMA-5): a single-arm, multicentre, phase 2 trial. Lancet Oncol. 2022;23(1):91-103. PubMed

54.Evans SR. Clinical trial structures. J Exp Stroke Transl Med. 2010;3(1):8-18. PubMed

55.Sun GW, Shook TL, Kay GL. Inappropriate use of bivariable analysis to screen risk factors for use in multivariable analysis. J Clin Epidemiol. 1996;49(8):907-916. PubMed

56.NCT05371093: Study of axicabtagene ciloleucel versus standard of care therapy in participants with relapsed/refractory follicular lymphoma (ZUMA-22). ClinicalTrials.gov. Bethesda (MD): U.S. National Library of Medicine; 2023: https://classic.clinicaltrials.gov/ct2/show/study/NCT05371093. Accessed 2023 Jul 14.

57.Dreyling M, Santoro A, Mollica L, et al. Phosphatidylinositol 3-kinase inhibition by copanlisib in relapsed or refractory indolent lymphoma. J Clin Oncol. 2017;35(35):3898-3905. PubMed

58.Sato T, Matsuyama Y. Marginal structural models as a tool for standardization. Epidemiology. 2003;14(6):680-686. PubMed

59.Clinical Evidence Template Yescarta (axicabtagene ciloleucel [internal sponsor's report]. Gilead Sciences Canada, Inc.; 2022 In: Drug Reimbursement Review sponsor submission: axicabtagene ciloleucel, 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, IV. Mississauga (ON): Gilead Sciences Canada; 2023 Apr 13.

60.Mangal N, Salem AH, Li M, Menon R, Freise KJ. Relationship between response rates and median progression-free survival in non-Hodgkin's lymphoma: A meta-analysis of published clinical trials. Hematol Oncol. 2018;36(1):37-43. PubMed

61.Zhu R, Lu D, Chu YW, et al. Assessment of correlation between early and late efficacy endpoints to identify potential surrogacy relationships in non-Hodgkin lymphoma: A literature-based meta-analysis of 108 phase II and phase III studies. AAPS J. 2017;19(3):669-681. PubMed

62.Batlevi CL, Sha F, Alperovich A, et al. Follicular lymphoma in the modern era: survival, treatment outcomes, and identification of high-risk subgroups. Blood Cancer J. 2020;10(7):74. PubMed

63.Tisagenlecleucel for relapsed or refractory follicular lymphoma. (CADTH Health Technology Review). Ottawa (ON): CADTH; 2023: https://www.cadth.ca/tisagenlecleucel. Accessed 2023 Jul 14.

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; DLBCL = diffuse large B-cell lymphoma; FL = follicular lymphoma; HGBL = high-grade B-cell lymphoma; LBCL = large B-cell lymphoma; NOC = Notice of Compliance; NOC/c = Notice of Compliance with conditions; r/r = relapsed or refractory.

^aThis review of axicabtagene ciloleucel went through the interim review process for CAR T-cell therapies, in which recommendations were issued by the CADTH Health Technology Expert Review Panel.

Table 2: Summary of Economic Evaluation

axi-cel = axicabtagene ciloleucel; CAR = chimeric antigen receptor; CHOP = cyclophosphamide, doxorubicin, vincristine, and prednisone; CVP = cyclophosphamide, vincristine, and prednisone; FL = follicular lymphoma; GDP = gemcitabine, dexamethasone, and cisplatin; ICER = incremental cost-effectiveness ratio; KM = Kaplan-Meier; OS = overall survival; PD = progressed disease; PFS = progression-free survival; QALY = quality-adjusted life-year; R2 = lenalidomide plus rituximab; R-CVP = rituximab plus cyclophosphamide, vincristine, and prednisone; r/r = relapsed or refractory; SOC = standard of care; SCT = stem cell transplant; WTP = willingness to pay.

Conclusions

Evidence from the ZUMA-5 single-arm trial suggests that treatment with axi-cel may be associated with clinically important tumour responses, including complete remission, in adult patients with relapsed or refractory (r/r) FL after 2 or more lines of systemic therapies. However, there is insufficient evidence — due to the single-arm design of the trial as well as limited follow-up duration — to determine the effects of axi-cel on overall survival (OS) and progression-free survival (PFS). In addition, the CADTH clinical assessment identified limitations with the sponsor’s comparison of the ZUMA-5 trial to the SCHOLAR-5 trial, which substantially restricted the ability to interpret the treatment effects of axi-cel relative to that of SOC. Overall, the Clinical Review concluded that there is a high degree of uncertainty around the comparative treatment effects of axi-cel relative to SOC.

Given the magnitude of uncertainty to do with the effect of axi-cel treatment on OS, its comparative efficacy against SOC, and the durability of such a benefit, CADTH was unable to derive a robust base-case estimate of cost-effectiveness. Moreover, given the duration of the ZUMA-5 trial (36 months) in contrast to the model’s time horizon (50 years), it is important to note that the majority of the quality-adjusted life-year (QALY) benefit was derived from the time period beyond which there are observed trial data (i.e., extrapolated period). To address this, CADTH conducted separate reanalyses involving different parametric assumptions of treatment effect waning for OS: (A) using the exponential distribution that assumes 15.3 years of treatment effect postinfusion before waning; and (B) using the log-normal distribution that assumes sustained treatment effect for the entire time horizon of the model.

In CADTH reanalysis A (assuming 15.3 years of treatment effect postinfusion before waning), axi-cel was associated with 0.93 incremental QALYs gained and additional costs of $505,223 relative to SOC, resulting in an incremental cost-effectiveness ratio (ICER) of $544,875 per QALY gained. In CADTH reanalysis B (assuming that treatment effect would be maintained for the entire time horizon of the model), axi-cel was associated with 2.07 incremental QALYs gained and additional costs of $505,885 relative to SOC, resulting in an ICER of $243,879 per QALY gained. The estimated ICERs were higher than the sponsor’s base-case value, driven primarily by removing the cure assumption. In line with clinical expert advice, these reanalyses achieved more plausible OS curves in the absence of robust long-term evidence, while still conferring a benefit with axi-cel. CADTH noted that both reanalyses assume life expectancy increases for patients receiving axi-cel relative to current SOC (2.68 and 1.10 years of life gained in reanalysis A and B, respectively). However, the true impact of axi-cel on OS relative to SOC remains uncertain in the absence of evidence from randomized studies. The CADTH reanalyses assume that the impacts of residual confounding that could influence the nonrandomized comparison of the ZUMA-5 and SCHOLAR-5 studies are limited and that their findings could be replicated in real-world clinical practice. Both assumptions are highly uncertain. Given the available evidence, the estimates presented within the CADTH reanalyses may represent the upper bounds of the incremental gains that may be realized from this therapy.

Assuming OS and PFS outcomes from the nonrandomized comparison of the ZUMA-5 and SCHOLAR-5 studies can be replicated in real-world settings and extensions in lifespan occur relative to current SOC, a price reduction of between 82% and 95% would be required for axi-cel to be cost-effective at a willingness-to-pay (WTP) threshold of $50,000 per QALY gained. This would reduce the 1-time price of axi-cel from $485,021 to $86,334 (an 82% price reduction) or $26,676 (a 95% price reduction). This range reflects the uncertainty around long-term survival extrapolation as analyzed in CADTH reanalyses A and B.

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 via an online survey. The survey, conducted from April 2022 to April 2023, included 143 patients with lymphoma (of whom 34% were diagnosed with follicular lymphoma [FL] and 86% lived in Canada). Of note, 3 respondents reported having experience with axicabtagene ciloleucel (axi-cel). The most important outcomes, according to the respondents, included delaying disease progression and achieving long-term remission, with the ultimate objective of improving survival; reducing side effects from treatments; preserving independence to minimize the burden on caregivers; and maintaining quality of life. Based on survey responses, 49% of patients underwent a period of “watchful waiting” before starting treatment. The majority of the patients surveyed (43%) had received 1 line of therapy, while 20% had received 2 and 16% had received 3, where chemoimmunotherapy was the most commonly prescribed treatment. The majority of patients surveyed received 2 regimens: rituximab plus cyclophosphamide, hydroxydaunorubicin (doxorubicin), oncovin (vincristine), and prednisone; and cyclophosphamide, vincristine, and prednisone. Important side effects of chemotherapy included fatigue, low activity level, and hair loss. The patients emphasized the need for therapies that can be administered at a hospital located near home to minimize travel time and burden on caregivers, as well as improve quality of life by keeping patients close to their support systems. Patients who had experience with axi-cel accessed this therapy via enrolment in a clinical trial. They reported travelling out of province to receive treatment and being away from home for up to 3 months to do so. Patients who received axi-cel experienced side effects that included cytokine release syndrome (CRS), neutropenia, febrile neutropenia, thrombocytopenia, constipation, and swelling.

Registered clinician input was received from Ontario Health-Cancer Care Ontario Hematology Cancer Drug Advisory Committee. According to the clinicians, the current pathway of care for patients with r/r FL after 2 or more lines of systemic therapy is chemoimmunotherapy and autologous stem cell transplant (auto-SCT). The clinicians noted that in select patients for whom the treatment goal is mostly palliative; allogeneic stem cell transplant (allo-SCT) and radiotherapy may also be considered. The clinicians suggested chemoimmunotherapy would likely be less efficacious among re-treated patients, thus pointing to the need for additional therapy options. The clinicians suggested that axi-cel may shift the current treatment paradigm by replacing chemoimmunotherapy in third line, but will likely not replace auto-SCT among eligible patients. Although it is uncertain whether axi-cel could replace auto-SCT in third line, the clinicians noted the potential for axi-cel to be prescribed in advance of auto-SCT among patients who are refractory to chemotherapy. Furthermore, it was noted that axi-cel should only be considered in relatively fit patients without significant comorbidities. Patients with uncontrolled infections, severe organ dysfunction, and poor performance status should be excluded. Clinician input further noted that axi-cel should be considered in select patients who had received prior CD19-directed therapy or allo-SCT, despite their exclusion from the pivotal trial.

Participating drug 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 treatment centres where the therapy can be administered, access may require interprovincial travel and, without full coverage of interprovincial reimbursement, may impact equitable access across Canada. Finally, drug plans queried whether, and how, potential manufacturing delays may impact the clinical effectiveness of axi-cel.

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

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

• The standard of care (SOC) modelled by the sponsor reflected the current treatments available to patients with r/r FL.

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

• Noting that r/r FL is a disease that can have extended periods of survival when treated with current SOC, CADTH revised the sponsor’s long-term extrapolation of OS for patients receiving SOC in line with real-world evidence and clinical plausibility.

• In line with clinicians’ expectations that axi-cel will likely shift the current treatment paradigm by replacing chemoimmunotherapy as a new preferred treatment for patients with r/r grade 1, 2, or 3a FL after 2 or more lines of systemic therapy, CADTH revised the sponsor’s projected market share in the budget impact analysis (BIA).

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

• Capacity is not explicitly considered in the model. The sponsor assumes all patients with r/r FL 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 included 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 r/r grade 1, 2, or 3a FL after 2 or more lines of systemic therapy.

Economic Evaluation

Summary of Sponsor’s Economic Evaluation

Overview

The sponsor submitted a cost-utility analysis of axi-cel compared with SOC. Aligned with Health Canada’s indicated population, the modelled population comprised adult patients with r/r grade 1, 2, or 3a FL after 2 or more lines of systemic therapy.¹

Axi-cel is a CD19-directed genetically modified autologous T-cell immunotherapy individually prepared from a patient’s peripheral blood mononuclear cells.¹ It is available as a cell suspension for infusion containing a target dose of 2 × 10⁶ chimeric antigen receptor (CAR) T-cells/kg body weight (range, 1 × 10⁶ cells/kg to 2.4 × 10⁶ cells/kg), to a maximum of 2 × 10⁸ CAR T cells for patients weighing 100 kg or more.² It is provided as a single-dose, 1-time infusion. The sponsor’s submitted price for axi-cel is $485,021 per infusion,¹ not including costs associated with pretreatment (i.e., leukapheresis, bridging chemotherapy, and lymphodepleting chemotherapy), hospitalization related to inpatient administration, and postinfusion stay in an intensive care unit (ICU) stay.

SOC, the comparator for this analysis, encompassed a basket of therapies commonly used in Canadian clinical practice. SOC was composed of 50% chemotherapy, 12% SCT, 5% phosphoinositide 3-kinase inhibitors (i.e., idelalisib), and 33% investigational therapies offered though clinical trials.¹ The treatment cost associated with chemotherapy was estimated as a weighted average of 6 chemotherapy regimens: rituximab plus bendamustine; cyclophosphamide, hydroxydaunorubicin, vincristine, and prednisone; cyclophosphamide, vincristine, and prednisone; obinutuzumab plus bendamustine; gemcitabine, dexamethasone, and cisplatin; rituximab plus cyclophosphamide, vincristine, and prednisone.¹

Based on clinician-informed proportions of patients on each chemotherapy regimen and average number of cycles per regimen, the sponsor estimated the total weighted drug cost of chemotherapy to be $15,650.¹ The sponsor estimated the cost of auto-SCT and allo-SCT to be $70,434 and $91,992, respectively, which incorporated the cost of stem cell transplant (SCT) procedures (including high-dose chemotherapy, stem cell harvest, and infusion), as well as the cost of inpatient stay associated with administration.¹ The SOC composite cost was estimated as a weighted average of the drug acquisition and administration costs associated with chemotherapy, SCT, phosphoinositide 3-kinase inhibitors, and investigational therapies accessed through clinical trials ($32,073).¹ Vial-sharing was not incorporated by the sponsor.

The clinical outcomes modelled were PFS and OS.¹ The economic outcomes of interest were QALYs and life-years. The economic evaluation was conducted over a lifetime time horizon (50 years), from the perspective of the Canadian public health care payer. Costs and outcomes were discounted at 1.5% per annum.¹

Model Structure

The sponsor used a partitioned survival model to capture all costs and outcomes associated with axi-cel and SOC. The model included 3 health states: progression-free, progressed disease (PD), and death, with transitions between health states occurred on a monthly cycle length (Figure 1). The proportion of patients in progression-free, PD, and death states was estimated over time based on OS and PFS curves, which were informed by the ZUMA-5 single-arm trial, as well as the SCHOLAR-5 retrospective cohort study. The proportion of patients with PD (i.e., post-progression state) was estimated as the difference between the proportion of living patients (estimated from the OS curve) and the proportion of progression-free patients (estimated from the PFS curve). PFS was defined as the time from the date of enrolment to the date of first documented progression or death due to any cause. Patients began in the progression-free health state and, over time, could transition to either the PD health state or the death state. Patients in the PD health state could remain either in this health state or transition to the death state (i.e., patients could not return to the progression-free health state).

Model Inputs

Baseline patient characteristics were derived from the ZUMA-5 trial, a phase II, single-arm, multicentre trial investigating the efficacy and safety of axi-cel among patients with r/r grade 1, 2, or 3a FL (n = 127).¹ The typical patient in the modelled cohort, which the sponsor assumed reflected the patient population in Canada, was aged | | years and weighed | | kg, and was more likely to be male (59%).¹ These characteristics were those of the patient population enrolled in the ZUMA-5 trial and were used to inform the drug dosage regimens and the age- and sex-specific distribution of the general population mortality risk.¹

Clinical efficacy parameters used to characterize axi-cel and SOC, including OS and PFS, were derived from various data sources. For axi-cel, inputs were based on the 36-month follow-up analysis of the ZUMA-5 single-arm trial (data cut-off: March 31, 2022). For SOC, inputs were informed by the SCHOLAR-5 study, a multicentre, international, observational, retrospective study that constructed a historical control cohort of patients with r/r FL treated with 2 or more prior lines of usual therapies in routine practice (n = 128). The SCHOLAR-5 cohort was derived from 3 international cohorts: IQVIA, Vanderbilt, and DELTA. The IQVIA and Vanderbilt cohorts were created from electronic medical records, while the DELTA cohort represented patients from the DELTA clinical trial who proceeded to receive therapy subsequent to idelalisib. The DELTA cohort was added to increase statistical power for OS. In the absence of head-to-head evidence, the sponsor conducted an indirect treatment comparison to assess the relative efficacy of axi-cel versus SOC therapies. Propensity score methods, specifically standardized mortality ratio weighting, were applied to account for the imbalance of confounders between the ZUMA-5 trial and the SCHOLAR-5 external control cohort.³ Specifically, the OS and PFS for axi-cel and SOC were estimated using a weighted Kaplan-Meier (KM) estimator, whereby patients from the SCHOLAR-5 external cohort were weighted to be comparable with the ZUMA-5 trial population across baseline covariates using propensity score matching.³

Survival data from the indirect comparison of the ZUMA-5 and SCHOLAR-5 study results were extrapolated to derive the long-term survival estimates of OS and PFS informing the economic model.¹ Standard parametric models were fit independently to the individual patient data from the ZUMA-5 and SCHOLAR-5 studies. Under the assumption that axi-cel was curative, the sponsor derived piecewise cure models to estimate OS and PFS, which accounted for the proportion of patients receiving axi-cel who may experience long-term survival. The piecewise cure models estimated a likelihood of cure (i.e., “statistically cured” fraction), wherein the survival outcomes for the “cured” group, relative to the “uncured” group, which does not achieve long-term remission, are assumed to be better. Cured patients experience a slightly worse survival than the general population (hazard ratio = 1.09),⁴ while uncured patients experience poorer survival outcomes relative to cancer-free individuals of the same age and sex. In the submitted base case, 40% of patients receiving axi-cel were expected to achieve long-term remission (i.e., the cure fraction), with long-term survival applied 5 years from the start of the model (i.e., the cure time point). Hence, in the sponsor’s piecewise cure model, (1) standard parametric extrapolations for OS and PFS were applied until the 5-year cure time point; and (2) OS and PFS were calculated thereafter as the weighted product of background survival (for the cured population) and cancer-specific survival (for the uncured fraction). For patients receiving axi-cel, the sponsor used a piecewise cure model based on the exponential distribution to model OS, and a piecewise cure model based on the log-normal distribution to model PFS. Given that the sponsor did not extend the assumption of cure to patients receiving SOC therapies, the sponsor selected the standard exponential distribution to model both the OS and PFS of SOC. Survival distributions were selected based on Akaike information criterion and Bayesian information criterion, as well as visual inspection.¹

As health-related quality of life data were not collected in the ZUMA-5 trial, health state utility values were obtained from the literature. Utility values were derived from 2 sources that reported health state utilities for indolent non-Hodgkin lymphoma using the EQ-5D results of 222 patients in the UK. The sponsor sourced utility values of 0.805 for the progression-free state, 0.620 for the PD on-treatment state, and 0.736 for the PD off-treatment state. Utility values were adjusted using age-related utility decrements based on an algorithm developed by Ara et al. (2010)⁵ to account for the natural decline in quality of life associated with age. The model incorporated disutilities associated with AEs categorized as grade 3 or greater in any of the treatments considered.¹ Disutilities were applied as a 1-time decrement to the first model cycle; thus, it was assumed that AEs would have no further impact on costs beyond the initial hospitalization period. For axi-cel, AE rates were derived from those occurring in at least 5% of patients in the ZUMA-5 trial (36-month data cut-off);⁶ while for SOC, rates were calculated as a weighted average of AEs reported in clinical trials of SOC regimens.^7-9

Costs captured in the model included pretreatment cost (i.e., drug acquisition and drug administration associated with pretreatment), treatment cost (i.e., drug acquisition, drug administration, and hospitalization associated with treatment), follow-up medical costs before progression (i.e., physician visits, PET/CT scans, and laboratory tests), subsequent treatment costs in fourth line, follow-up medical costs in postprogression, AE management costs, and terminal care costs.¹ Drug acquisition costs for axi-cel were based on the sponsor’s submitted price.¹ The dosing modelled for axi-cel is consistent with that described in the overview section. Drug acquisition costs were sourced from the Ontario Drug Benefit Formulary,¹⁰ with dosing schedules based on the chemotherapy regimen monographs from Ontario Health-Cancer Care Ontario.¹¹

Prior to infusion, patients receiving axi-cel underwent leukapheresis, bridging chemotherapy, and conditioning chemotherapy. Given that data from the ZUMA-5 trial were used to determine the proportion of patients that would receive each phase of treatment, weighted costs were modelled for axi-cel. Pretreatment costs were applied in the first cycle of the model. The cost of leukapheresis ($2,688) was applied to all patients receiving axi-cel.¹² The cost of bridging chemotherapy ($25) was estimated as the weighted 1-cycle cost of the following therapies: 60% dexamethasone ($3); 10% radiotherapy ($157); and 30% gemcitabine, dexamethasone, and cisplatin regimen ($652). This cost was applied to 3.1% of patients receiving axi-cel.¹ Moreover, all patients receiving axi-cel were assumed to receive fludarabine plus cyclophosphamide administered daily for 3 days as conditioning chemotherapy ($2,300).¹ Hospitalization and ICU inputs for axi-cel were estimated based on the average length of stay (| || || |) and proportion of patients in ICU (||||||) observed in the ZUMA trial as these were assumed to be reflective of Canadian clinical practice. Costs typically associated with the ongoing monitoring were obtained from the Canadian Institute for Health Information’s Patient Cost Estimator for an inpatient hospitalization for malignant lymphoma.¹³

For patients receiving SOC, treatment costs included chemotherapy (50%), SCT (12%), idelalisib (5%), and clinical trials (33%), based on a review of Canadian clinical practice guidelines and clinical expert opinion. The treatment costs associated with the SOC basket of therapies were as previously described. All regimens were assumed to be given in an outpatient setting, except for SCT. The cost of outpatient administration was based on the cost of complex chemotherapy administration from the Ontario Schedule of Benefits for Physician Services,¹⁴ and the cost associated with chair time was estimated based on the CCO regimen monographs.¹¹ SCT procedure costs were obtained from the Interprovincial Health Insurance Agreements Coordinating Committee,¹⁵ and the Ontario Case Costing Initiative (OCCI).¹⁶

The total weighted costs of subsequent therapy differed by prior treatment (i.e., those who received CAR T cells versus SOC as third-line therapy) and were applied as a one-off cost in the first model cycle after progression. The weighted cost associated with subsequent therapy among patients receiving axi-cel and SOC in third line was estimated to be $14,714 and $6,871, respectively. 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.¹⁷ Treatment-emergent AE costs were estimated based on the data from the OCCI and applied as a 1-time cost in the first model cycle. All patients who transitioned to the death state were assumed to incur terminal care costs ($68,703) in the last cycle before death, based on the average cost for patients with terminal lymphoma in the Ontario Cancer Registry.¹⁹

Summary of Sponsor’s Economic Evaluation Results

The sponsor conducted the base case via a probabilistic sensitivity analysis with 1,000 simulations. The deterministic and probabilistic results were similar. The probabilistic findings are presented below.¹

Base-Case Results

Compared with SOC, axi-cel was associated with an incremental QALY gain of 4.38 and an incremental cost of $505,565, resulting in an ICER of $115,543 per QALY (Table 3).¹ Notably, the sponsor’s analysis predicted that axi-cel was associated with a longer duration of life than SOC (i.e., incremental life-years: 5.83).

Given the duration of the ZUMA-5 trial (i.e., 36 months) in contrast to the model’s time horizon (i.e., 50 years), it is important to note that the majority of the QALY and life-year benefit (96% and 97%, respectively) realized by patients receiving axi-cel was derived from the time 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 progression-free state (78%), whereas patients receiving SOC realized most of their QALY gains in the PD state (78%). The key cost driver among patients receiving axi-cel was the cost of drug acquisition, accounting for 81% of the total cost incurred by patients. The main cost driver among patients receiving SOC was end-of-life care, which accounted for 55% of the total estimated cost.

Axi-cel was not cost-effective at a WTP threshold of $50,000 per QALY in any of the iterations when compared to SOC. The sponsor’s submitted analysis is based on the publicly available prices for all drug treatments. Additional results from the sponsor’s submitted economic evaluation base case are presented in Appendix 3.

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

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

Source: Sponsor’s pharmacoeconomic submission.¹

Sensitivity and Scenario Analysis Results

The sponsor assessed several model parameters and assumptions in probabilistic scenario analyses. These included: applying a shorter model time horizon; using another analysis set from the ZUMA-5 trial; applying different health state utility values; assessing the impact of different cure fractions; applying alternative parametric distributions to model the OS and PFS of axi-cel and SOC; and aligning the axi-cel re-treatment proportion with that observed in the ZUMA-5 trial. The parameters with the greatest influence were other assumptions regarding efficacy, particularly selection of OS extrapolations and cure fraction, as well as shorter time horizons. When applying a 10% cure fraction for axi-cel, the ICER increased to $168,794 per QALY gained. Moreover, when applying a 10-year time horizon, the ICER increased to $375,698 per QALY gained. All other scenarios resulted in ICERs ranging between $92,912 and $162,936 per QALY gained.¹

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:

• Assumption of cure of patients with FL is inappropriate: The sponsor derived piecewise cure models to extrapolate the OS and PFS for axi-cel beyond the ZUMA-5 trial period based on the assumption that a proportion of the population receiving axi-cel would be cured. Specifically, in the submitted base case, 40% of patients receiving axi-cel were expected to achieve long-term remission (i.e., the cure fraction), with long-term survival applied 5 years from the start of the model (i.e., the cure time point). The sponsor suggested that a cure model would be better suited to capture the expected long-term survival of patient populations receiving CAR T-cell therapies with curative intent. CADTH noted that due to the limited duration of follow-up in the ZUMA-5 trial (36 months), there is a great deal of uncertainty as to whether there truly is a statistically cured fraction. CADTH additionally noted that both the cure fraction and cure time point used in the submitted model were sourced from a single-centre trial with a relatively small number of patients with FL (n = 14), for whom the median duration of response had not been reached.²⁰ Given this small sample size, CADTH noted that the evidence sourced by the sponsor to inform relevant cure input parameters is uncertain and subject to imprecision. This is a concern in view of the fact that estimating survival (and whether this includes a statistically cured fraction) plays a key role in determining the cost-effectiveness of axi-cel relative to that of SOC. Moreover, the clinical expert panel convened by CADTH during the review process noted that long remissions are common for patients with FL, and it would thus be premature to determine the fraction of patients and the time point when patients with r/r FL who remain progression-free are considered clinically cured. The clinical panel highlighted that the evidence submitted by the sponsor does not support a proportion of patients being cured; moreover, it was noted that cure in this disease area is rare. CADTH further noted that clinical expert opinion is aligned with the emerging scientific literature on the uncertainty around the permanence of CAR T-cell treatment efficacy among patients with hematologic malignancies; lack of permanence, in turn, can result in late relapse.²¹ In alignment with clinical expert advice, CADTH noted that the assumption of cure of patients with r/r FL remains uncertain given the available evidence.

  • CADTH conducted reanalyses using standard parametric models based on KM data from the ZUMA-5 trial to extrapolate the OS and PFS of axi-cel for the entire duration of the model.

• Impact of axi-cel on long-term OS is highly uncertain: The sponsor’s base case predicted a survival advantage with axi-cel compared to SOC (incremental life-years: 5.83), of which, 96% were accrued in the time period beyond which there are observed trial data. At the 36-month follow-up analysis of the ZUMA-5 trial reviewed by the CADTH clinical team (data cut-off: March 31, 2022), the OS data for axi-cel were immature as both the median OS and the upper bound of the 95% confidence interval had not been reached.²² The clinical experts consulted by CADTH noted that r/r FL is a disease that can have extended periods of survival, suggesting that the duration of follow-up may not be sufficiently long to capture the effects of axi-cel on OS. As noted by the CADTH Clinical Review, in addition to the duration of the ZUMA-5 trial and the noncomparative design, the prevalence of subsequent therapy complicates the interpretation of the OS findings. Following axi-cel infusion, | | of patients received at least 1 subsequent antineoplastic medication (of whom | | received SCT). Hence, the trial’s OS results should be considered in the context of subsequent therapies, given that it may be difficult to distinguish which treatment (primary, or subsequent) had greater impact on survival. This is especially relevant considering the lack of comparative data in the ZUMA-5 trial. Furthermore, as the ZUMA-5 trial is ongoing, the data analyzed at the aforementioned cut-off date represent an interim analysis. Evidence suggests that the clinical benefit of cancer agents demonstrated in primary publications is often different from updated mature data. A recent study comparing the predicted survival of parametric extrapolations with observed survival based on updated data of 32 trials identified through US FDA oncology approvals revealed that extrapolations based on initial KM curves had low precision compared with updated KM curves.²³ Therefore, CADTH contends that, in the absence of mature OS data and in tandem with the noncomparative design of the ZUMA-5 trial, the extent to which the parametric distributions of OS for axi-cel overestimate or underestimate the true incremental life-years is uncertain.

  CADTH also noted that there is uncertainty regarding the expected OS benefit of axi-cel to a broader population beyond the select patient population recruited for the clinical trial. The ZUMA-5 trial consisted exclusively of patients with an Eastern Cooperative Oncology Group Performance Status of less than or equal to 1 and with a mean age of | | years. According to the clinical experts consulted by CADTH, the ZUMA-5 study population is generally representative of patients in the Canadian population with r/r FL who would receive axi-cel. However, the clinical experts noted that patients seen in clinical practice would also include those with poorer performance status, prior experience with CD19-targeted therapy, and higher prevalence of comorbidities. The clinical experts also noted that the average age of patients with r/r FL is expected to be higher. A Canadian real-world population-based study of a cancer registry estimated the median age of patients with incident FL to be 64 and 61 years in women and men, respectively.²⁴ As such, if axi-cel were to become available in clinical practice, where patients are likely to have more diverse clinical and demographic profiles than in a clinical trial, there remains uncertainty regarding the expected presence and magnitude of the OS benefit in the real-world setting.

  The sponsor selected parametric distributions based on goodness-of-fit criteria, visual inspection, and clinical plausibility. CADTH agrees that appropriate models should be compared based on their statistical fit to the observed trial data, as well as based on their ability to generate clinically plausible long-term extrapolations.²⁵ However, statistical fit pertains only to the observed trial period, not to the extrapolation period. The weight given to the comparative fit of alternative parametric models to the observed data depends on the extent to which extrapolation is required and the degree of censoring present. Given the length of time required for extrapolation and the size of the censored population in the ZUMA-5 trial, the clinical plausibility of the extrapolated portion of alternative models is of greater importance than the statistical fit to the observed data.²⁵ According to the clinical experts consulted by CADTH for this review, although the ZUMA-5 trial findings appeared favourable and clinically important (and a survival benefit with axi-cel was deemed plausible), the magnitude and durability of such a benefit was highly uncertain in the absence of longer-term or head-to-head evidence. Clinical experts indicated that given the degree of uncertainty to do with the durability of the treatment effect of axi-cel relative to SOC, it is clinically plausible for the OS curve of axi-cel to converge with that of SOC within the model’s lifetime horizon.

  • Considering these limitations, CADTH conducted 2 reanalyses with different assumptions of treatment effect durability for axi-cel, which sought to address the uncertainty in long-term clinical outcomes.

  • In reanalysis A, the OS of patients receiving axi-cel was modelled using the exponential distribution; as such, the OS for axi-cel was capped by the OS for SOC at 15.3 years, where the survival curves would have otherwise crossed.

  • In reanalysis B, the OS of patients receiving axi-cel was modelled using the log-normal distribution; as such, it was assumed that the treatment effect of axi-cel, relative to that of SOC, would be maintained for the entire model time horizon. The log-normal distribution was maintained for the PFS extrapolation of axi-cel in each reanalysis. In line with clinical expert advice, these reanalyses achieved more plausible OS curves in the absence of long-term evidence, while still conferring a benefit with axi-cel.

• Impact of SOC on long-term OS is underestimated: The sponsor derived clinical efficacy data used to characterize SOC from the SCHOLAR-5 study, a retrospective study that constructed a historical control cohort of patients with r/r FL treated with 2 or more prior lines of usual therapies in routine practice. In the absence of head-to-head evidence, the SCHOLAR-5 data used to extrapolate OS for SOC were weighted to be comparable with the ZUMA-5 trial population across baseline covariates. The sponsor selected the exponential distribution based on the weighted KM data from the SCHOLAR-5 trial to model the long-term OS for patients receiving SOC. CADTH noted that the 5-year OS estimate predicted by the exponential distribution for SOC (49.8%) underestimated the observed 5-year survival rate informed by KM data from the SCHOLAR-5 trial (| || || || || || || || || | |).³ CADTH further noted that a recent retrospective analysis conducted by Batlevi et al. (2020)²⁶ in a single centre in the US reported a median OS of 8.8 years for patients with r/r FL treated with third-line SOC therapies. In this study, usage of single-agent rituximab (22.4%), rituximab plus alkylator chemotherapy (16.7%), rituximab plus anthracycline chemotherapy (11.7%), investigational therapy (10.0%), SCT (8.3%), and radioimmunotherapy (8.0%) were common in the third-line setting.²⁶ Therefore, the exponential distribution selected by the sponsor to extrapolate long-term survival for SOC underestimated the 9-year OS by 20.8 percentage points (29.2% versus 50%). Hence, the parametric distribution selected by the sponsor to model long-term OS for patients receiving SOC in the economic model underestimated both the KM estimates informed by the sponsor-submitted SCHOLAR-5 control cohort and the median OS derived from real-world evidence.

  According to clinical expert judgment, the OS curve for SOC generated with the exponential distribution produced 10-, 20-, and 30-year extrapolations that were misaligned with the expectation of survival in current clinical practice. The clinical experts consulted by CADTH reiterated that although long-term survival evidence for SOC therapies is not yet obtainable, clinically plausible extrapolations of OS for patients receiving SOC should align with the available real-world evidence. Therefore, the clinical experts noted, the log-logistic distribution was likely the most appropriate way to extrapolate the OS for SOC given that: midterm projections were within range of the survival estimates observed in the SCHOLAR-5 trial and by Batlevi et al. (2020); and 10-year, 20-year, and 30-year long-term extrapolations were aligned with clinical plausibility.

  • CADTH conducted a reanalysis using the log-logistic parametric model to extrapolate the OS of SOC.

• Exclusion of costs related to CAR T-cell therapy: In the sponsor’s base case, costs related to CAR T-cell therapy eligibility assessment were excluded. Feedback from the clinical experts consulted by CADTH noted that the upfront costs of assessment of CAR T-cell therapy eligibility would include those associated with MRIs, PET scans, bone marrow transplants, lumbar punctures, and bloodwork. Should axi-cel be reimbursed, this assessment cost would be incurred by all adult patients with r/r grade 1, 2, or 3a FL after 2 or more lines of systemic therapy regardless of whether they would go on to receive CAR T-cell therapy. In addition, the pretreatment cost of leukapheresis considered by the sponsor ($2,688) for patients receiving CAR T-cell therapy was underestimated. CADTH consulted the OCCI and was informed that the cost of stem cell apheresis (code 1.LZ.58.HX; data not available for leukapheresis specifically due to limited number of events) was $5,426.¹⁶ This underestimation of the costs incurred by patients receiving axi-cel relative to their receiving SOC biased the results of the economic analysis in favour of the drug under review.

  • CADTH conducted a reanalysis that included the additional CAR T-cell therapy eligibility assessment cost and the updated cost associated with apheresis.

• Health state utility estimates are uncertain. Given that health-related quality of life data were not collected in the ZUMA-5 trial, the sponsor sourced health state utility values from the accrued literature. Values were derived from 2 sources that reported health state utilities for 222 patients with indolent non-Hodgkin lymphoma in the UK using the EQ-5D questionnaire.^27,28 CADTH noted that the utility values adopted in the economic analysis should reflect the preferences of the general population in Canada. The sponsor sourced utility values of 0.805 for the progression-free state, 0.620 for the PD on-treatment state, and 0.736 for the PD off-treatment state. CADTH noted that these estimates were applied without distinguishing which interventions were received (i.e., axi-cel, SOC). The clinical experts consulted by CADTH indicated that, while declining utility in the PD state, relative to the progression-free state, appeared reasonable, the difference between on- and off-treatment utilities accrued in the PD state did not align with their experiences in current clinical practice. In fact, the clinical experts noted that patients treated with currently available subsequent therapies (compared with those available in 2007, as per Pettengell et al., 2008) are unlikely to experience substantial differences in quality of life resulting from treatment status. As such, clinical experts reiterated that, for patients with r/r FL who have PD, quality of life is not expected to differ between those who are on subsequent treatment and those who are off subsequent treatment.

  • CADTH conducted a scenario analysis assuming that the utility value for patients with PD would not differ according to treatment status.

• Selection of comparators is not supported by current clinical practice. The sponsor omitted the lenalidomide plus rituximab (R2) regimen, from the base-case analysis, despite evidence that this therapy is prescribed by some oncologists in current Canadian clinical practice.²⁹ During the review process, both the participating drug plans and the clinical expert panel convened by CADTH indicated that the R2 regimen is used off-label for the treatment of adult patients with r/r grade 1, 2, or 3a FL after 2 or more lines of systemic therapy in Canada. According to CADTH’s Reimbursement Review procedures, the base case must comprise all relevant comparators, including (1) treatments currently reimbursed by at least 1 participating drug plan for the indication under review; (2) reimbursed treatments that are currently used off-label in Canadian practice; or (3) treatments that have previously received a recommendation in favour of reimbursement from CADTH for the indication under review. CADTH noted that the off-label use of the R2 regimen for this indication is infrequent, and, as such, it is unlikely that its omission impacted the cost-effectiveness results.

  • CADTH was unable to address this limitation as the relevant evidence for the R2 regimen was not incorporated in the economic model submitted by the sponsor.

• Poor modelling practices were employed. The sponsor’s submitted model included numerous “IFERROR” statements, resulting in situations where the parameter value was overwritten with another value without alerting the user to the automatized overwriting. The systematic use of IFERROR statements rendered thorough auditing of the sponsor’s model impractical, as it remains unclear whether the model ran inappropriately by overriding errors. In addition, CADTH noted that selecting the generalized gamma parametric distribution to extrapolate the OS for axi-cel within a prespecified dropdown list yielded calculation errors across several iterations, which complicated the validation process.

  • CADTH was unable to address this limitation, noting that a thorough validation of the submitted model was not possible.

Additionally, the following key assumptions were made by the sponsor and have been appraised by CADTH (See Table 5).

Table 4: Key Assumptions of the Submitted Economic Evaluation

AE = adverse event; axi-cel = axicabtagene ciloleucel; FL = follicular lymphoma; OS = overall survival; PFS = progression-free survival; r/r = relapsed or refractory; SOC = standard of care.

CADTH Reanalyses of the Economic Evaluation

Base-Case Results

CADTH reanalyses were derived by making changes to model parameter values and assumptions, in consultation with clinical experts. The following changes were made to address several limitations within the economic model: using standard parametric models based on KM data from the ZUMA-5 trial to extrapolate the OS and PFS of axi-cel for the entire duration of the model; using alternative parametric models to extrapolate the OS of SOC and axi-cel; and including a CAR T-cell therapy eligibility assessment cost and updating the pretreatment cost associated with apheresis. However, given the magnitude of uncertainty to do with the comparative clinical efficacy and durability of treatment effect of axi-cel relative to SOC, CADTH was unable to derive a robust base-case estimate of cost-effectiveness. CADTH conducted separate reanalyses involving different parametric assumptions of treatment effect for axi-cel. In reanalysis A, the treatment effect of axi-cel on OS was modelled using the exponential distribution; in reanalysis B, the treatment effect of axi-cel on OS was modelled using the log-normal distribution. These changes are summarized in Table 5.

Table 5: CADTH Revisions to the Submitted Economic Evaluation

axi-cel = axicabtagene ciloleucel; CAR = chimeric antigen receptor; OS = overall survival; PFS = progression-free survival; SOC = standard of care.

^aCADTH reanalyses 4 and 5 (which change the axi-cel parametric distribution of OS to exponential and log-normal, respectively), require that reanalyses 1 and 2 be performed concurrently, that is, the axi-cel extrapolation method must be standard parametric (reanalysis 1) and the OS for SOC modelled using the log-logistic distribution (reanalysis 2).

Results from CADTH reanalyses A and B were generally aligned: axi-cel is not cost-effective at a $50,000 WTP threshold compared to SOC. In CADTH reanalysis A (assuming 15.3 years of treatment effect postinfusion before waning), axi-cel was associated with an ICER of $544,875 per QALY gained compared to SOC (incremental costs: $505,223; incremental QALYs: 0.93). In CADTH reanalysis B (assuming that treatment effect would be maintained for the entire model time horizon), axi-cel was associated with an ICER of $243,879 per QALY gained compared to SOC (incremental costs: $505,885; incremental QALYs: 2.07). The probability that axi-cel was cost-effective at a WTP threshold of $50,000 per QALY was 0% in both reanalyses A and B.

The estimated ICERs were higher than the sponsor’s base-case value, driven primarily by the use of standard parametric models based on KM data from the ZUMA-5 trial to extrapolate the OS and PFS of axi-cel (i.e., rejecting the assumption of a statistically cured fraction). In line with clinical expert advice, these reanalyses achieved more plausible OS curves in the absence of long-term evidence, while still conferring a survival benefit with axi-cel. In both reanalyses, most incremental QALYs were due to improvements in life-years. Furthermore, 76% (reanalysis A) and 89% (reanalysis B) of QALYs gained by patients receiving axi-cel were derived from the extrapolated period in which there are no observed trial data. The majority of the total cost among patients receiving axi-cel was associated with drug acquisition costs (78%), while the key cost driver among patients receiving SOC was related to terminal care costs (51%).

CADTH reanalyses are based on the publicly available prices for all drug treatments. Full results are available in Appendix 4.

Table 6: Summary of the Stepped Analysis of the CADTH Reanalysis Results (Deterministic)

axi-cel = axicabtagene ciloleucel; CAR = chimeric antigen receptor; ICER = incremental cost-effectiveness ratio; OS = overall survival; QALY = quality-adjusted life-year; SOC = standard of care.

Scenario Analysis Results

CADTH undertook price reduction analyses based on the sponsor’s results and CADTH’s reanalyses. Results of CADTH reanalysis A suggested a price reduction of 95% would be required to achieve cost-effectiveness of axi-cel relative to SOC at a $50,000 per QALY threshold. In CADTH reanalysis B, a price reduction of 82% would be required (Table 7).

Table 7: CADTH Price Reduction Analyses

axi-cel = axicabtagene ciloleucel; ICER = incremental cost-effectiveness ratio; SOC = standard of care; vs. = versus.

CADTH undertook 1 scenario analysis on each of the CADTH reanalyses. In line with clinical expert advice, CADTH assumed that for patients with r/r FL who have PD, quality of life would not differ according to subsequent treatment status. Specifically, CADTH assumed utility equivalence between patients who are on and off subsequent treatment (0.736). In reanalysis A, the ICER increased to $564,126 (incremental costs: $505,223; incremental QALYs: 0.90). In reanalysis B, the ICER increased to $246,379 (incremental costs: $505,885; incremental QALYs: 2.05).

The results of this analysis are presented in Table 14.

Issues for Consideration

• Travel-associated costs: Travel costs and the requirement for time spent away from work were not taken into account in the economic model. The sponsor’s implementation plan indicated that not all provinces and territories will have a site to provide axi-cel within the next 3 years.³⁰ Patients in jurisdictions that do not currently have a site for providing axi-cel will need to travel out of province for treatment. Furthermore, the clinical experts noted that some provinces do not have capacity to assess patients’ eligibility for CAR T-cell therapy. The implementation plan suggests that the sponsor will coordinate travel and lodging for the patient and their caregiver, who need to remain close to the qualified treatment centre following infusion.³⁰ The sponsor stated that the program is intended to support adherence to axi-cel’s monitoring requirements by providing financial support to cover transportation-related expenses and lodging costs for the patient and their caregiver during the pretreatment and treatment periods when they are required to stay close to the qualified treatment centre. The sponsor also noted that, for the patient to be eligible for support, their primary residence must be at least 2 hours or 200 km away from the authorized treatment centre. If this patient support program is not operationalizable and travel expenses (e.g., travel, lodging, food) are absorbed by the patient or public payer, this may impact access to axi-cel. Disparities in funding and treatment access may vary depending on the province or territory. Hence, the requirement for access to a tertiary care centre for delivery of axi-cel may have equity of access implications, which were not substantively considered in the economic submission.

• Manufacturing delays: The sponsor’s implementation plan indicated that, in the ZUMA-5 trial, the median time from leukapheresis collection to release from the manufacturing site was | || |, and the median time to final product delivery (i.e., time axi-cel is ready to be infused back into the patient) was observed to be 17 days.²² The sponsor further noted that comparable findings were demonstrated in a retrospective analysis of the manufacturing and supply experience in Canada, where the median time between leukapheresis and product delivery was reported to be 21 days.³¹ Moreover, a recent real-world study conducted in the US based on 3 commercial claims databases revealed that the median time from leukapheresis to CAR T-cell infusion was 26 to 27 days.³² However, CADTH clinical expert feedback noted the potential for greater variability in manufacturing time in the real-world setting given apheresis collection would be conducted across a broad network of pan-Canadian sites with manufacturing conducted across the border in the US.³⁰ Clinical expert feedback emphasized that manufacturing delays are a significant clinical problem, especially among patients who progress relatively fast and have a higher disease burden. In the context of manufacturing delays, axi-cel would not be as effective among fast-progressors as it would among patients who do not have disease progression.

• Manufacturing failures: Issues pertaining to manufacturing are important to the successful delivery of CAR T-cell therapies. Manufacturing failure may occur due to inadequate number of T-cells in the apheresed product, poor selection of T-cells on day 0 of manufacturing, or irreversibly impaired T-cells (i.e., no response to stimulation in culture), microbial contamination, equipment-related cell loss, high endotoxin levels, and accidents. The sponsor noted that the axi-cel manufacturing success rate in the ZUMA-5 trial was 100%; hence, the impact of manufacturing failure was not considered in the submitted economic model. However, manufacturing failure of CAR T-cell therapies is not uncommon and has been observed in trials for axi-cel³³ and other CAR T-cell products.³⁴ There may be additional costs associated with manufacturing failure including increased hospital stay while a second sample is prepared, if at all possible. In addition, manufacturing failure may impact patient outcomes due to treatment delays or compromised doses.

• Re-treatment with axi-cel: In ZUMA-5, patients who achieved a partial response at the 3-month assessment and subsequently experienced disease progression were eligible for an optional course of re-treatment with conditioning chemotherapy and axi-cel. Hence, 10.2% of enrolled patients were eligible for re-treatment (n = 13 in the 36-month full analysis set). Aligned with the ZUMA-5 trial, the sponsor conducted a scenario assuming that 10.2% of patients receiving axi-cel would be eligible for an optional course of re-treatment. Of note, in the economic model, the sponsor assumed that re-treated patients would not incur the additional drug acquisition cost of axi-cel and leukapheresis. CADTH noted that the sponsor had not offered additional information in the implementation plan regarding the potential costs associated with re-treatment with axi-cel (i.e., neither confirming nor denying that drug acquisition costs for axi-cel would be incurred only once in the event of re-treatment). The clinical expert panel convened by CADTH indicated that it is unlikely that axi-cel would be eligible for use in re-treatment of patients with r/r grade 1, 2, or 3a FL after 2 or more lines of systemic therapy in Canada, as there is no evidence to support re-treatment in this population. CADTH noted that the cost-effectiveness of re-treatment with axi-cel in patients with r/r FL is unknown.

• Capacity constraints: The sponsor’s implementation plan indicated the capacity for annual production of axi-cel for patients in Canada. The sponsor stated that it currently has the capacity to produce therapy for 4,500 patients per year, with capacity increasing to 7,100 patients per year by December 2023.³⁰ The sponsor did not take into account potential capacity constraints in the submitted economic evaluation (i.e., additional costs arising from treatment delays and/or adverse clinical outcomes due to capacity issues).

• Long-term clinical impacts: The evidence for the effectiveness of CAR T-cell therapy is still in its early stages. In addition, evidence about the rate of late treatment-related toxicities, duration of treatment effect, and what comprises follow-up for patients receiving CAR T-cell therapy in Canada is emerging. Furthermore, if axi-cel is curative as claimed, patients with r/r FL would be expected to live a longer life, and as such, the health care system may incur additional costs.

• High resource and administrative burden: To be able to treat patients with axi-cel, personnel at specialized treatment centres need to be trained and the centres accredited by the manufacturer. Both obtaining and maintaining this accreditation process, including the development of various protocols and supporting yearly audits, can result in a high resource burden. There is also the added complexity of needing to coordinate patient care and product preparation with an external manufacturer. Since there will likely be multiple CAR T-cell therapies administered by the specialized treatment centres, various protocols for preparation will need to be managed along with delivery of each product type, which would increase the overall administrative burden.

• Shortage of drugs to manage CRS: CRS tends to be managed with tocilizumab, which is in relatively short supply in Canada. Tocilizumab has been on the Drug Shortages Canada website list because of its use in the treatment of COVID-19 treatment. Health Canada has previously declared a “Tier 3” shortage of tocilizumab, a designation reserved for shortages with the greatest potential impact on Canada’s health care system.³⁵ The use of siltuximab has been considered by some clinicians if there is a severe shortage of tocilizumab, but this treatment is currently only publicly funded via the Alberta drug formulary.³⁶ Shortage of treatments for CRS may impact axi-cel use because of the risk of CRS associated with CAR T-cell therapy. This is especially relevant as 78% of patients in the ZUMA-5 trial experienced CRS.

• Future treatments: Clinical experts indicated that Canadian hospitals are in the midst of expanding their capacity to manufacture CAR T-cell products in the future. In particular, British Columbia’s BC Cancer Immunotherapy Program has developed in-house expertise and infrastructure to manufacture CAR T cells for the treatment of blood cancers. More than 50 patients with end-stage leukemia and lymphoma in British Columbia and Ontario have received CAR T-cell therapy manufactured by the BC Cancer Immunotherapy Program as part of the CLIC-01 clinical trial, a phase I/II pan-Canadian clinical trial.³⁷ Clinical expert feedback noted that the price of CAR T-cell therapy produced in this setting would be substantially lower than the CAR T-cell therapy developed by the pharmaceutical industry.

Overall Conclusions

Evidence from the ZUMA-5 single-arm trial suggests that treatment with axi-cel may be associated with clinically important tumour responses, including complete remission, in adult patients with r/r FL after 2 or more lines of systemic therapies. However, there is insufficient evidence — due to the single-arm design of the trial as well as limited follow-up duration — to determine the effects of axi-cel on OS and PFS. The CADTH clinical assessment identified limitations with the sponsor’s comparison of the ZUMA-5 trial to the SCHOLAR-5 study, which substantially restricted the ability to interpret the relative treatment effects of axi-cel and SOC. Overall, the CADTH Clinical Review concluded that there is a high degree of uncertainty around the treatment effects of axi-cel relative to SOC.

Given the magnitude of uncertainty to do with the effect of axi-cel treatment on OS, axi-cel’s efficacy compared to SOC, and the durability of any benefit, CADTH was unable to derive a robust base-case estimate of cost-effectiveness. Moreover, given the duration of the ZUMA-5 trial (36 months) in contrast to the model’s time horizon (50 years), it is important to note that the majority of the QALY benefit realized by patients receiving axi-cel was derived from the time period beyond which there are observed trial data (i.e., extrapolated period). To address this, CADTH conducted separate reanalyses involving different parametric assumptions of treatment effect: in reanalysis A, the OS of patients receiving axi-cel was modelled using the exponential distribution (thus, assuming 15.3 years of treatment effect postinfusion before waning); and in reanalysis B, the OS of patients receiving axi-cel was modelled using the log-normal distribution (thus, assuming that the treatment effect of axi-cel, relative to SOC, would be maintained for the entire time horizon of the model). In addition, the following changes were made consistently across reanalyses A and B to address limitations within the economic model: using standard parametric models based on KM data from the ZUMA-5 trial to extrapolate the OS and PFS of axi-cel for the entire duration of the model; using alternative parametric models to extrapolate the OS of SOC and axi-cel; and including a CAR T-cell eligibility assessment cost and updating the pretreatment cost associated with apheresis.

Results from CADTH reanalyses A and B were generally aligned: axi-cel is not cost-effective at a $50,000 WTP threshold compared to SOC. In CADTH reanalysis A (assuming 15.3 years of treatment effect postinfusion before waning), axi-cel was associated with an ICER of $544,875 per QALY gained compared to SOC (incremental costs: $505,223; incremental QALYs: 0.93). In CADTH reanalysis B (assuming that treatment effect would be maintained for the entire time horizon of the model), axi-cel was associated with an ICER of $243,879 per QALY gained compared to SOC (incremental costs: $505,885; incremental QALYs: 2.07). The estimated ICERs were higher than the sponsor’s base-case value, driven primarily by the use of standard parametric models based on KM data from the ZUMA-5 trial to extrapolate the OS and PFS of axi-cel (i.e., rejecting the assumption of a statistically cured fraction). In line with clinical expert advice, these reanalyses achieved more plausible OS curves in the absence of robust long-term evidence, while still conferring a benefit with axi-cel. CADTH noted that both reanalyses assume life expectancy increases for patients receiving axi-cel relative to current SOC (2.68 and 1.10 years of life gained in reanalysis A and B, respectively). However, the true impact of axi-cel on OS relative to SOC remains uncertain in the absence of evidence from randomized trials. The CADTH reanalyses assume that the impacts of residual confounding that could influence the nonrandomized comparison of the ZUMA-5 and SCHOLAR-5 studies are limited and that their findings could be replicated in real-world clinical practice. Both assumptions are highly uncertain. Hence, given the available evidence, the estimates presented in the CADTH reanalyses likely represent the upper bounds of the incremental gains that may be realized from this therapy.

A price reduction of 95% or 82% would be required for axi-cel to be cost-effective at a WTP threshold of $50,000 per QALY gained, conditional on axi-cel’s long-term impact on OS relative to SOC. This would mean a reduction in the 1-time price of axi-cel from $485,021 to $86,334 (an 82% price reduction) or $26,676 (a 95% price reduction). This range reflects the uncertainty around long-term survival extrapolation as analyzed in CADTH reanalyses A and B. Finally, CADTH undertook 1 scenario analysis on each of the CADTH reanalyses. In line with clinical expert advice, CADTH assumed that among patients with r/r FL who have PD, quality of life would not differ according to subsequent treatment status. In reanalysis A, the ICER increased to $564,126 (incremental costs: $505,223; incremental QALYs: 0.90). In reanalysis B, the ICER increased to $246,379 (incremental costs: $505,885; incremental QALYs: 2.05).

References

1.Pharmacoeconomic evaluation [internal sponsor's report]. In: Drug Reimbursement Review sponsor submission: axicabtagene ciloleucel, 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, IV. Mississauga (ON): Gilead Sciences Canada; 2023 Apr 13.

2.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; 2019 Feb 13.

3.A comparison of clinical outcomes from ZUMA-5 (axicabtagene ciloleucel) and the international SCHOLAR-5 external control cohort in relapsed/refractory follicular lymphoma (R/R FL): 36-month follow-up for patients with three or more prior lines of therapy and patients with two or more prior lines of therapy, addendum to modules 5.3.5.2 ZUMA-5 clinical study report. [internal sponsor's report]. In: Drug Reimbursement Review sponsor submission: axicabtagene ciloleucel, 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, IV. Mississauga (ON): Gilead Sciences Canada; 2023 Apr 13.

4.Maurer MJ, Ghesquières H, Jais JP, et al. Event-free survival at 24 months is a robust end point for disease-related outcome in diffuse large B-cell lymphoma treated with immunochemotherapy. J Clin Oncol. 2014;32(10):1066-1073. PubMed

5.Ara R, Brazier JE. Populating an economic model with health state utility values: moving toward better practice. Value Health. 2010;13(5):509-518. PubMed

6.Kite Pharma. Common Technical Document section 2.7.4: Summary of clinical safety [internal sponsor's report]. In: Drug Reimbursement Review sponsor submission: axicabtagene ciloleucel, 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, IV. Mississauga (ON): Gilead Sciences Canada; 2023 Apr 13.: Gilead; 2021.

7.Marcus R, Davies A, Ando K, et al. Obinutuzumab for the first-line treatment of follicular lymphoma. N Engl J Med. 2017;377(14):1331-1344. PubMed