Reimbursement Review

Pembrolizumab (Keytruda)

Sponsor: Merck Canada Inc.

Therapeutic area: Non-small cell lung cancer

This multi-part report includes:

Clinical Review

Pharmacoeconomic Review

Clinical Review

Abbreviations

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

NOC = Notice of Compliance; NSCLC = non–small cell lung cancer; T2a = tumour stage 2a.

Introduction

Lung cancer is the most diagnosed cancer and leading cause of cancer deaths in Canada.^1,2 Lung cancer survival in all stages and histologies are poor, with an overall 5-year net survival of 22%,^2,3 and only 3% for those diagnosed with stage IV disease.² In 2024, it was estimated that there would be 32,100 new cases of lung cancer diagnosed and 20,700 deaths from lung cancer.² Non–small cell lung cancer (NSCLC) accounts for approximately 88% of all cases of lung cancer in Canada.¹ It is estimated that 30% to 35% of NSCLCs are diagnosed at an early stage (I to IIIA)^4-6 and approximately 20% to 25% of patients with NSCLC have surgically resectable disease.⁷ After surgery, 45% of patients with stage IB disease and 76% of patients with stage III disease will experience disease recurrence and subsequently die over a median follow-up of 5 years, regardless of the use of adjuvant chemotherapy (ACT) or immunotherapy.⁸

Standard treatment for patients with stage IB to IIIA NSCLC is surgical resection.⁹ Perioperative treatments (neoadjuvant or adjuvant) are used depending on stage. According to the clinical experts consulted by the review team, the goal of any adjuvant therapy following complete resection in early-stage NSCLC is to improve cure rates by reducing the risk of relapse, after which no curative therapies are currently available. In the perioperative setting, the current treatment standard for patients with resectable NSCLC without actionable oncogenic alterations is neoadjuvant platinum-doublet chemotherapy in combination with immunotherapy (nivolumab), or adjuvant platinum-doublet chemotherapy followed by immunotherapy (atezolizumab) for patients with a PD-L1 tumour proportion score (TPS) of 50% or greater.^10-13 However, the clinical experts consulted for this review indicated that not all patients receive immunotherapy or platinum-based chemotherapy in the perioperative setting as they may decline, not be offered a referral to medical oncology, or are ineligible. The clinical experts noted that there is a need to reduce the recurrence rates in patients with early-stage NSCLC, and provide options for those who have not had neoadjuvant immunochemotherapy and are ineligible for adjuvant atezolizumab (e.g., patients with a PD-L1 TPS of less than 50%) to cure disease, delay disease recurrence, improve survival, and maintain quality of life (QoL).^10,14,15 According to the clinical experts consulted by the review team, PD-L1 TPS testing is currently performed as standard of care for patients with NSCLC in Canada.

This report reviews and critically appraises the evidence submitted by the sponsor on the beneficial and harmful effects of pembrolizumab (Keytruda) administered by IV infusion as either 200 mg every 3 weeks or 400 mg every 6 weeks for the adjuvant treatment of adult patients with stage IB (tumour stage 2a [T2a] ≥ 4 cm), II, or IIIA NSCLC who have undergone complete resection and platinum-based chemotherapy and whose tumours have a PD-L1 TPS of less than 50%, as determined by a validated test.

Perspectives of Patients, Clinicians, and Drug Programs

The information in this section is a summary of input provided by the patient and clinician groups who responded to a call for input from Canada’s Drug Agency (CDA-AMC) and from the clinical experts consulted for the purpose of this review.

Patient Input

This review received a joint submission by 3 patient groups: the Lung Health Foundation (LHF), Lung Cancer Canada (LCC), and the Canadian Cancer Survivor Network. The input was based on information collected by the LHF from individuals living with lung cancer, 3 interviews with patients from Ottawa, Vancouver, and Toronto who had experience with pembrolizumab, and 33 responses to an online survey available between June 2023 to June 2024.

Survey respondents reported similar symptoms and challenges due to their lung cancer, some of which included fatigue (53%), shortness of breath (50%), coughing (23%), and pain (20%). Most respondents indicated that living with lung cancer negatively affects their emotional well-being through feelings of isolation, challenges with symptom management, and perceived burdens on caregivers and family. Disease aspects that were most important to responders to control included symptoms and pain and side effects from therapy. Patients who previously received surgery reported experiencing deconditioning and chronic fatigue, and medication side effects that included extreme itching affecting sleep, brain fog, fatigue, nausea, vomiting, mood changes, diminished appetite, weight loss, hair loss, anemia, and neuropathy. These patients also experienced challenges accessing some therapies due to high treatment costs, as well as difficulty navigating the health care system and locating disease information and support. The input also noted concerns from patients on targeted therapy as to their ability to access the next line of treatment if or when their current therapy stops working.

Respondents indicated that key treatment outcomes to consider when evaluating new therapies include stopping or slowing disease progression with minimal side effects, and effectiveness in advanced disease. Three LHF interviewees had experience with pembrolizumab, although these patients were not part of the eligible population of the current indication under review, which is limited to those with stage IB (T2a ≥ 4 cm), II, or IIIA NSCLC. One interviewee with epidermal growth factor receptor (EGFR)-positive stage 4 lung cancer was taking pembrolizumab in combination with chemotherapy, although it was unclear if the interviewees with NSCLC were taking pembrolizumab similarly or as monotherapy. One patient discontinued pembrolizumab after 19 months due to progression, while another experienced sufficient tumour shrinkage and inactivity to discontinue pembrolizumab after 3 years, before re-initiating shortly after upon tumour reactivation. After re-initiation of pembrolizumab, the patient’s tumours once again decreased in size. Side effects reported by these patients included nausea, fatigue, muscle soreness, constipation, diarrhea, and worsening of their diabetes, eczema, and liver-enzyme levels. Patients did not report that these side effects impeded their ability to participate in daily activities or exercise, and overall they reported experiencing improved QoL while on therapy.

Clinician Input

Input From Clinical Experts Consulted by CDA-AMC

According to the clinical experts consulted by the review team, the key treatment goals for patients with early-stage NSCLC who receive adjuvant therapy following complete surgical resection is to improve cure rates by reducing the risk of relapse. The clinical experts noted that there are currently no adjuvant immunotherapy options for the patients in Canada with stage IB to IIIA NSCLC who have a PD-L1 TPS of less than 50%, creating an unmet need for this group. The clinical experts noted that adjuvant pembrolizumab would fill a treatment gap, as currently patients can only access this treatment after relapse. The clinical experts pointed out that overall survival (OS) is the outcome most important to patients. The clinical experts indicated that pembrolizumab treatment should be discontinued if 1 of the following has been met: a total of 18 cycles (1 year) of adjuvant immunotherapy has been completed, disease progression has been detected, or unacceptable toxicity develops. The clinical experts noted that pembrolizumab should be administered in a specialty setting that has surgical and medical oncology multidisciplinary staff with the expertise to provide systemic therapy, monitor the patient, and manage treatment-related toxicities.

Clinician Group Input

This review received input from the Ontario Health (Cancer Care Ontario) Lung Cancer Drug Advisory Committee and the LCC Clinician Group. Six clinicians from the Drug Advisory Committee and 35 clinicians from LCC provided input for this review. The clinician groups supplied input aligning with input from the clinical experts consulted for this review with respect to the unmet needs, patient population, key outcomes, discontinuing treatment, and prescribing considerations. The LCC group emphasized that, in patients with early-stage resected IB to IIIA NSCLC, current therapies do not adequately achieve high cure rates or prevent recurrences. The LCC group noted that this is particularly important for patients with NSCLC, as the risk of relapse increases substantially with each subsequent disease stage. The LCC group also pointed out that patient relapse and metastatic disease impose substantial costs on patient health, QoL, utilization of health care resources, and economic loss of productivity, and overall costs to society. The LCC group expected that pembrolizumab would shift the current treatment paradigm, as it represents the first adjuvant immunotherapy option for this patient population. The LCC group also anticipated that treatment with pembrolizumab for eligible patients with a sensitizing EGFR mutation would require case-by-case consideration by treating clinicians, weighing the risks and benefits of adjuvant sequential chemotherapy and immunotherapy versus those of adjuvant osimertinib. The clinician groups agreed that treatment benefits in the adjuvant setting are primarily determined by disease recurrence, which the LCC group noted typically occurs within 2 to 3 years for patients with stage IB to IIIA NSCLC. The LCC group indicated that cure rates, as measured by 5-year OS, can also determine response, but typically requires additional years of follow-up. The LCC group suggested implementing clinical and laboratory follow-ups every 3 weeks to evaluate toxicity and disease recurrence, as well as imaging scans at 3- to 4-month intervals, given pembrolizumab is administered over 1 year. The LCC group noted that, overall, immunotherapies are well tolerated by patients, and autoimmune side effects can often be readily managed by oncologists.

Drug Program Input

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

• consideration for initiation of therapy

• consideration for discontinuation of therapy

• generalizability.

The clinical experts consulted for this review provided advice on the potential implementation issues raised by the drug programs. Table 5 provides more details.

Clinical Evidence

Systematic Review

Description of Studies

One ongoing, multicentre, triple-blind, phase III randomized controlled trial (RCT), KEYNOTE-091 (N = 1,177) submitted by the sponsor was included for this review. The study compared pembrolizumab (200 mg every 3 weeks or 400 mg every 6 weeks, by IV infusion, for up to 1 year or until disease recurrence or unacceptable toxicity) with placebo as adjuvant therapy for completely resected stage IB to IIIA NSCLC. Eligible patients were adults with pathologically confirmed NSCLC (any histology) of stage IB (T2a ≥ 4 cm), II or IIIA as defined by the seventh edition of the American Joint Committee on Cancer (AJCC) staging system¹⁶ after complete surgical resection (lobectomy, sleeve lobectomy, bi-lobectomy, or pneumonectomy) and negative margins (R0). Eligible patients had an available tumour sample obtained during resection for PD-L1 assessment, a known PD-L1 expression status, no evidence of disease on clinical examination and radiographic assessment according to the Response Evaluation Criteria in Solid Tumors Version 1.1 (RECIST 1.1)¹⁷ as determined by local investigator review after surgery but within 12 weeks before randomization, an Eastern Cooperative Oncology Group Performance Status (ECOG PS) of 0 or 1, and adequate organ function within 10 days of treatment initiation. Previous neoadjuvant or adjuvant radiotherapy for the current malignancy was not permitted. ACT was not mandatory but was to be considered for patients with stage IB disease and strongly recommended for those with stage II or IIIA disease according to local practice and national guidelines. Patients with prior treatment with an anti–PD-1, anti–PD-L1 or anti–PD-L2, anti-CD137, CTLA-4 modulators, or any other immune-modulating drugs were excluded from the KEYNOTE-091 study. The subpopulation of interest (the reimbursement request population [n = 363 in the pembrolizumab group, and n = 363 in the placebo group]), were patients receiving ACT and a PD-L1 TPS of less than 50%. The outcomes relevant to this review included OS, disease-free survival (DFS), health-related quality of life (HRQoL), and safety.

The reimbursement request population of the KEYNOTE-091 study had a median age of 64 to 65 years. The proportion of male patients (65% to 68%) was higher than that of female patients (32% to 35%). Most patients (77% to 78%) were white, followed by Asians (17% to 18%), among others. A relatively small proportion of patients (11% to 12%) had stage IB disease and more than half (55% to 59%) had stage II disease. Most patients (68% to 73%) were former smokers, followed by those who had never smoked (13% to 19%), and current smokers (13% to 14%). More patients had an ECOG PS of 0 (56% to 62%) compared to those with an ECOG PS of 1 (38% to 44%). Most patients (41%) had a lymph-node stage of N0 (no regional lymph-node involvement) and 38% had a lymph-node stage of N1 (nearby lymph-node involvement). More patients had a nonsquamous histology (62.5% to 72.2%) compared with those with squamous histology (27.8% to 37.5%). The proportion of patients who had an EGFR mutation (5.8% to 8.5%) or anaplastic lymphoma kinase (ALK) translocation (0.8% to 1.7%) was low. More than half of the patients had an unknown status for the EGFR mutation (51.8% to 57.4%) or ALK translocation (56.7% to 66.4%).

Efficacy Results

The key efficacy results from the KEYNOTE-091 trial are summarized in Table 2 in order from the most important to the less important outcomes as suggested by the clinical experts consulted for this review. The efficacy and harms outcomes of the KEYNOTE-091 study reported in this review were based on the protocol-prespecified third interim analysis (IA3), for which the data cut-off date was January 24, 2023.

Overall Survival

As of the data cut-off date (January 24, 2023) and among the reimbursement request population (N = 726), the median duration of follow-up was 46.6 months (range = 0.6 to 84.2). The median OS was not reached in either group ███████ █████ █ █████ ███ ██████████ ████████ █████ ████ ██ █████ | | ██████. The Kaplan-Meier estimates of the probability of OS in the pembrolizumab and placebo groups were █████ ████ ███ ████ ██ █████ versus █████ ████ ███ ████ ██ █████ at 36 months; and █████ ████ ███ ████ ██ █████ versus █████ ████ ███ ████ ██ █████ at 48 months, respectively.

Disease-Free Survival

As of the data cut-off date among the reimbursement request population, the median DFS was 51.7 months (95% confidence interval [CI], 39.0 to 70.4) for patients treated with pembrolizumab and 34.5 months (95% CI, 23.3 to 46.4) for patients who received placebo (hazard ratio [HR] = 0.72; 95% CI, 0.58 to 0.89; P < 0.001). The Kaplan-Meier estimates of the probability of DFS in the pembrolizumab and placebo groups were 67.2% (95% CI, 61.9 to 71.9) versus 55.0% (95% CI, 49.7 to 60.0) at 24 months; and 51.2% (95% CI, 45.2 to 56.9) versus 42.4% (95% CI, 36.7 to 47.9) at 48 months, respectively.

Health-Related Quality of Life

The HRQoL outcomes assessed with the European Organisation for Research and Treatment of Cancer Quality of Life Questionnaire Core 30 (EORTC QLQ-C30) Global Health Status (GHS)/QoL measure and the European Organisation for Research and Treatment of Cancer Quality of Life Questionnaire Lung Cancer Module (EORTC QLQ-LC13) symptom scales at 48 weeks were available for the patient-reported outcome (PRO) full analysis set (FAS) population (N = 1,161). The compliance rates for the EORTC QLQ-C30 and EORTC QLQ-LC13 were 98.6% at baseline and 85.8% at week 48 in the pembrolizumab group, and 99.8% at baseline and 90.0% at week 48 in the placebo group, respectively. At week 48, the questionnaire completion rates were 77.9% and 84.9% in the pembrolizumab and placebo groups, respectively.

The proportion of patients with a deteriorated score in EORTC QLQ-C30 GHS/QoL (a 10-point or greater deterioration from baseline at any time during the trial when the criteria for improved or stable is not met) in the pembrolizumab group was higher than in the placebo group (18.1% and 12.9%, respectively; difference = 5.2%; 95% CI, 1.0 to 9.4; P = 0.015).

At week 48, the proportion of patients with a deteriorated score in EORTC QLQ-LC13 symptom scales was similar between the 2 groups for chest pain (difference = −1.7%, 95% CI, −4.9 to 1.5; P = 0.295), coughing (difference = −0.1%, 95% CI, −3.9 to 3.6; P = 0.945), and dyspnea (difference = 3.2%, 95% CI, −1.5 to 7.8; P = 0.181).

Harms Results

The harms outcomes were available among the all participants as treated (APaT) population in the KEYNOTE-091 study (N = 1,161). Adverse events (AEs) were reported in 96% and 91% of patients in the pembrolizumab and placebo groups, respectively. The most common AEs were increased weight (23% in the pembrolizumab group and 29% in the placebo group, respectively), pruritus (22% and 13%), and hypothyroidism (21% and 5%). The incidence was higher in the pembrolizumab group than in the placebo group for grade 3 to 5 AEs (34% and 26%), serious adverse events (SAEs) (25% and 16%), AEs resulting in treatment discontinuation (20% and 6%), AEs resulting in death (2% and 1%), and adverse events of special interest (AEOSIs) such as immune-mediated events and infusion-related reactions (39% and 13%). The most frequently reported AEOSIs in the pembrolizumab and placebo groups were hypothyroidism (21% and 5%, respectively), hyperthyroidism (11% and 3%), and pneumonitis (7% and 3%).

Critical Appraisal

In the KEYNOTE-091 study, the review team and the clinical experts consulted for this review did not identify major issues that would affect the validity of the study results as a consequence of presenting the DFS and OS outcomes through ad hoc analyses in the reimbursement request population, based on the fact that the PD-L1 TPS category was a stratification variable. The patient demographic and disease characteristics appeared to be generally balanced between the treatment groups in both overall population and reimbursement request population, suggesting that the benefits of the randomization were reasonably maintained in the subpopulation for reimbursement request. The review team noted that histologic status was unbalanced between the 2 groups (27.8% squamous in the pembrolizumab group versus 37.5% in the placebo group). To what extent this imbalance could bias the results is unknown. In the reimbursement request population, a higher proportion of patients (48%) in the pembrolizumab group discontinued from the study medication compared with the placebo group (37%), mainly due to AEs. The clinical experts commented that the between-group imbalance and the reasons for the study medication discontinuation were reasonable and in line with the safety outcomes, in that larger proportions of patients in the pembrolizumab group experienced AEOSIs compared with those in the placebo group. In the reimbursement request population, nearly all of the study patients (95%) received at least 1 concomitant medication and the proportions of patients with the use of most medications were similar between treatment groups. However, there was a larger proportion of some concomitant medications (e.g., antihistamines, corticosteroids, and thyroid replacement therapy) in the pembrolizumab group compared to the placebo group, which may have affected the assessment of HRQoL and biased the results in favour of pembrolizumab, as these concomitant drug uses were most likely for the control or treatment of drug-related side effects associated with pembrolizumab. The proportion of patients receiving subsequent anticancer treatment during the trial was smaller in the pembrolizumab group than in the placebo group, for both antineoplastic therapy and immunotherapies. Although these uneven uses of anticancer therapies may have biased the efficacy results against pembrolizumab as compared to the placebo group for OS, the extent of any important impact on interpretation of the observed effect could not be determined. The triple-blind design of the trial likely mitigated the risks associated with knowledge of group assignment for these outcomes. The risk of bias due to missing outcome data for OS, DFS, and safety outcomes appeared to be low as losses to follow-up for reasons other than death were low, and sensitivity analysis with the different censoring rule for DFS in the reimbursement request population was consistent. The risk of bias due to missing outcome data for the HRQoL outcome is low as only a small proportion of patients (1% to 6%) had “no assessment” for the select measures in the PRO FAS population (N = 1,161). OS and DFS were tested by applying a multiplicity hierarchical testing procedure to account for the potential inflated type I error rates across multiple end points and interim analyses. However, OS and DFS results were based on interim analyses, which may have overestimated the treatment-effect estimates.^18,19 The presence and extent of any overestimate that may have been introduced could not be determined.

Patients in the KEYNOTE-091 study were recruited from multiple countries, including Canada. The clinical experts described the eligibility criteria of patients in the KEYNOTE-091 study as appropriate and noted that the demographic diversity of the patients in the study was mostly in line with what is seen in clinical practice in Canada. The clinical experts noted that pembrolizumab is often offered to patients with an ECOG PS of up to 2 in clinical practice in Canada, and these patients might benefit from pembrolizumab, even though only patients with an ECOG PS of 0 to 1 were enrolled in the KEYNOTE-091 study as specified by the study inclusion criteria. The clinical experts noted that presenting the survival outcomes among the subgroup of patients in the KEYNOTE-091 study who had a PD-L1 TPS of less than 50% and had received ACT was appropriate for this review, aligns with the reimbursement request, and addresses the unmet therapeutic needs.

GRADE Summary of Findings and Certainty of the Evidence

Methods for Assessing the Certainty of the Evidence

The Grading of Recommendations Assessment, Development and Evaluation (GRADE) framework was used to assess the certainty of the evidence in the pivotal studies and RCTs identified in the sponsor’s systematic review regarding outcomes considered most relevant to deliberations of the CDA-AMC expert committee, and a final certainty rating was determined as outlined by the GRADE Working Group.^20,21 Following the GRADE approach, evidence from RCTs started as high certainty and could be rated down for concerns related to study limitations (which refers to internal validity or risk of bias), inconsistency across studies, indirectness, imprecision of effects, and publication bias.

The selection of outcomes for the GRADE assessment was based on the sponsor’s Summary of Clinical Evidence, consultation with clinical experts, and input received from patient and clinician groups and public drug plans. The following list of outcomes was finalized in consultation with expert committee members:

• probability of OS at months 36 and 48

• probability of DFS at months 24 and 48

• HRQoL as measured by the EORTC QLQ-C30 (GHS/QoL) and EORTC QLQ-LC13 symptomatic scales (chest pain, coughing, and dyspnea) at week 48

• grade 3 to 5 AEs.

When possible, certainty was rated in the context of the presence of an important (nontrivial) treatment effect; if this was not possible, certainty was rated in the context of the presence of any treatment effect (i.e., the clinical importance is unclear). In all cases, the target of the certainty-of-evidence assessment was based on the point estimate and where it was located relative to the threshold for a clinically important effect (when a threshold was available) or to the null.

Results of GRADE Assessments

Table 2 presents the GRADE summary of findings for pembrolizumab versus placebo in adult patients with stage IB (T2a ≥ 4 cm), II, or IIIA NSCLC who have undergone complete resection and platinum-based chemotherapy and whose tumours have a PD-L1 TPS of less than 50%.

Table 2: Summary of Findings for Pembrolizumab Versus Placebo for Adult Patients With Stage IB, II, or IIIA NSCLC With a PD-L1 TPS of Less Than 50%

AE = adverse event; CDA-AMC = Canada’s Drug Agency; CI = confidence interval; DFS = disease-free survival; EORTC QLQ-C30 = European Organisation for Research and Treatment of Cancer Quality of Life Questionnaire Core 30; EORTC QLQ-LC13 = European Organisation for Research and Treatment of Cancer Quality of Life Questionnaire Lung Cancer Module; GHS = Global Health Status; HRQoL = health-related quality of life; MID = minimal important difference; NR = not reported; NSCLC = non–small cell lung cancer; OS = overall survival; QoL = quality of life; RCT = randomized controlled trial; TPS = tumour proportion score.

Notes: Data presented in this table were based on analyses at clinical cut-off date of January 24, 2023. The OS data were not mature as of January 24, 2023. The between-group differences for all the outcomes in this table were requested from the sponsor to aid in interpretation and were not part of the sponsor’s analysis plan. Study limitations (which refer to internal validity or risk of bias), inconsistency across studies, indirectness, imprecision of effects, and publication bias were considered when assessing the certainty of the evidence. All serious concerns in these domains that led to the rating down of the level of certainty are documented in the table footnotes.

^aRated down 2 levels for very serious imprecision. At the data cut-off date, the OS results data were immature. An empirically derived and validated between-group MID for OS was not identified. Based on a threshold that is usually used by the CDA-AMC for effect assessment of an adjuvant treatment in patients with NSCLC of similar severity or stage, a between-group difference in the probability of OS of 5% may be clinically meaningful. The 95% CI of the absolute effect included the “no effect” threshold of 0 as well as the clinical importance threshold of 5%.

^bRated down 1 level for serious imprecision. At the data cut-off date, the OS results data were immature. An empirically derived and validated between-group MID for OS was not identified. Based on a threshold that is usually used by the CDA-AMC for effect assessment of an adjuvant treatment in patients with NSCLC of similar severity or stage, a between-group difference in the probability of OS of 5% may be clinically meaningful. At 48 months, the 95% CIs of the absolute effect excluded the “no effect” threshold of 0, and the point estimate and the upper bound of the 95% CI suggest a clinical important increase in OS based on the threshold of 5%.

^cRated down 1 level for serious imprecision. An empirically derived and validated between-group MID for DFS was not identified. Based on the threshold usually used by the CDA-AMC for effect assessment of an adjuvant treatment in patients with NSCLC of similar severity or stage, a between-group difference in the probability of DFS of 10% may be clinically meaningful. At both 24 months and 48 months, the 95% CIs of the absolute effect excluded the “no effect” threshold of 0, and the upper bound of the 95% CIs suggests a clinical important increase in DFS based on the threshold of 10%.

^dIndirectness was not rated down. The outcome data for HRQoL and harms were based on the total population in the KEYNOTE-091 study. Although there may be uncertainties regarding the presence and magnitude of any potential differences in these outcomes between the total population and the reimbursement request population, the review team and the clinical experts consulted for this review did not identify major issues that would affect the study results as the PD-L1 TPS category was a stratification variable, and the patient characteristics appeared balanced between the treatment groups in both overall population and reimbursement request population, suggesting that the benefits of the randomization were reasonably maintained in the subpopulation for reimbursement request.

^eImprecision did not result in the level of certainty being rated down, as the 95% CI of the absolute effect excluded the null threshold of 0. The clinical experts consulted for this review could not provide a threshold of important difference; however, the review team judged that the point estimate and 95% CI of the absolute effect were unlikely to include any important difference.

^fRated down 2 levels for very serious imprecision. The review team was unable to identify the MID to assess a between-group difference from literature or the clinical experts consulted for this review; the null was therefore used to assess certainty. The 95% CI of the absolute effect included the “no effect” threshold of 0.

Sources: BARDS Health Technology Assessment Statistical Report: Baseline Characteristics and Efficacy (April 9, 2024),²² Clinical Study Report: P091V02MK3475,²³ and sponsor’s submission.²⁴

Long-Term Extension Studies

The KEYNOTE-091 study is ongoing for OS follow-up. No other long-term extension studies are currently ongoing or completed.

Indirect Comparisons

In the absence of direct head-to-head trials evaluating the comparative efficacy of pembrolizumab and atezolizumab for the adjuvant treatment of adult patients with early-stage NSCLC who have undergone complete resection and platinum-based chemotherapy, the sponsor conducted 1 indirect treatment comparison (ITC) including only the subpopulation of participants with a TPS of 50% or greater. The findings from this ITC were used to support the sponsor’s reimbursement request and a request for a deviation from pharmacoeconomic requirements that excludes this subpopulation.

Description of Studies

The sponsor included 2 studies, KEYNOTE-091 and IMpower010, in its ITC. For the KEYNOTE-091 study, the sponsor included only the ongoing trial patients after excluding patients who discontinued treatment. The sponsor did not report the median follow-up duration for this subpopulation; however, the median follow-up time for the intention-to-treat (ITT) population of patients with a PD-L1 TPS of 50% or greater who were treated with pembrolizumab was 46.8 months (range = 3.4 to 83.5). For the IMpower010 study, from the published data, the median follow-up time for patients with a PD-L1 TPS of 50% or greater who were treated with atezolizumab was 35.98 months (range = 0.2 to 54.2). The sponsor did not report any assessment of homogeneity or any handling of potential effect modifiers.

Efficacy Results

The ITC of pembrolizumab versus atezolizumab in patients with a TPS of 50% or greater and stage II to IIIA cancer and who had received prior adjuvant chemotherapy showed that pembrolizumab appears to be less effective than atezolizumab in this subpopulation, with an HR of ████ ████ ███ ████ ██ ██████.

Harms Results

Harms outcomes were not reported.

Critical Appraisal

The sponsor-submitted ITC was used to support its reimbursement request and request for deviation from pharmacoeconomic requirements that excludes this subpopulation of adult patients with early-stage NSCLC who have undergone complete resection and platinum-based chemotherapy with a TPS of 50% or greater. The sponsor did not conduct an additional ITC for the reimbursement request population (patients with a TPS of less than 50%). It did not conduct a systematic literature review for this ITC. The apparent 10-month difference in median follow-up time between both trials may have an impact on the time to event outcomes. The clinical experts consulted for this review commented that the baseline patient characteristics from both trials appeared to be well matched. However, the sponsor did not report or appear to assess homogeneity between the 2 studies, and could only include published aggregate level data from the Impower010 study. It is therefore unclear if sources of clinical or methodological heterogeneity biased effect estimates of ITC.

Harms outcomes and other outcomes of relevance to patients (e.g., HRQoL) were not reported.

The clinical experts noted that, while no therapy options are currently available after adjuvant chemotherapy for patients with resected stage IB to IIIA NSCLC who have a PD-L1 TPS of less than 50%, atezolizumab is currently the treatment of choice for patients with a TPS of 50% or greater. They added that the results from this ITC support the current therapy guidelines and the sponsor’s reimbursement request.

Studies Addressing Gaps in the Evidence From the Systematic Review

No studies addressing gaps in the pivotal and RCT evidence were identified for this review.

Conclusions

One triple-blind, phase III RCT comparing the efficacy and safety of adjuvant pembrolizumab and placebo in adult patients with stage IB to IIIA NSCLC who have undergone complete resection and platinum-based chemotherapy and whose tumours have a PD-L1 TPS of less than 50% showed a clinically meaningful benefit of adjuvant pembrolizumab in the probability of DFS at 24 and 48 months. Uncertainty remains in the OS results because the data were immature (the median OS was not reached in either group, with OS events observed in 23% and 30% of patients in the pembrolizumab and placebo groups, respectively), even though there is a trend toward improved OS in favour of pembrolizumab. The HRQoL of the patients in both the pembrolizumab and placebo groups was relatively stable over 48 weeks. According to the clinical experts consulted for this review, the safety profile of pembrolizumab was consistent with their expectations for this drug.

The sponsor submitted an ITC comparing pembrolizumab and atezolizumab in patients with a TPS of 50% or greater to support its reimbursement request for patients with a TPS of less than 50%. The indirect comparative evidence suggests that atezolizumab is superior in patients with a TPS of 50% or greater; however, it is unclear if sources of clinical or methodological heterogeneity biased the effect estimates in the ITC as no assessment of homogeneity was conducted. The sponsor did not submit an ITC comparing pembrolizumab and atezolizumab in patients with a TPS of less than 50% because, according to the clinician input, atezolizumab is not indicated for this subpopulation of patients.

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 pembrolizumab administered by IV infusion as either 200 mg every 3 weeks or 400 mg every 6 weeks for the adjuvant treatment of adult patients with stage IB (T2a ≥ 4 cm), II, or IIIA NSCLC who have undergone complete resection and platinum-based chemotherapy and whose tumours have a PD-L1 TPS of less than 50%, as determined by a validated test.

Disease Background

Contents within this section were informed by materials submitted by the sponsor and clinical expert input. The following summary was validated by the CDA-AMC review team.

Lung cancer is the most diagnosed cancer and the leading cause of cancer deaths in Canada.^1,2 Lung cancer survival at all stages and with all histologies is poor, with an overall 5-year net survival of 22%,^2,3 and only 3% for those diagnosed with stage IV disease.² In 2024, it was estimated that there would be 32,100 new cases of lung cancer diagnosed and 20,700 deaths from lung cancer.² It is estimated that 1 in 21 Canadians (4.8%) will die from lung cancer.²

Lung cancer is classified into NSCLC or small-cell lung cancer, with NSCLC accounting for approximately 88% of all lung cancer cases in Canada.¹ NSCLC is further classified by histology into squamous and nonsquamous (including adenocarcinoma, large-cell carcinoma, and other less common histologies).¹ To determine a patient’s prognosis and treatment, NSCLC is staged using the AJCC staging criteria, which involves tumour-node-metastasis classification of the disease based on the size and spread of the primary tumour, lymph-node involvement, and occurrence of metastasis.²⁵ The clinical experts consulted for this review pointed out that such changes in staging were made relatively recently as there has been some movement between stage I and II. As a consequence, comparing a more recent study’s results with previous trials may be problematic.

It is estimated that 30% to 35% of patients with NSCLC are diagnosed at an early stage (I to IIIA)^4-6 and approximately 20% to 25% of patients with NSCLC have surgically resectable disease.⁷ After surgery, 45% of patients with stage IB disease and 76% of patients with stage III disease will experience disease recurrence and subsequently die over a median follow-up of 5 years, regardless of the use of ACT or immunotherapy.⁸ As NSCLC may be asymptomatic or minimally symptomatic, patients may have a late presentation.²⁶ The most common symptoms include coughing, chest and shoulder pain, hemoptysis, weight loss, dyspnea, hoarseness, bone pain, fever, and recurring infections with bronchitis and pneumonia.^26,27 Diagnostic procedures include lung imaging and tissue biopsy.²⁸

Standards of Therapy

Contents within this section were informed by materials submitted by the sponsor and clinical expert input. The following summary was validated by the CDA-AMC review team.

Standard treatment for patients with stage IB to IIIA NSCLC, as staged by the AJCC,¹⁶ is surgical resection.⁹ Perioperative treatments (neoadjuvant or adjuvant) are used depending on the stage. According to the clinical experts consulted by the review team, the goal of any adjuvant therapy following complete resection for early-stage NSCLC is to improve cure rates by reducing the risk of relapse, after which no curative therapies are currently available.

In the perioperative setting, the current treatment standard for patients with resectable NSCLC without actionable oncogenic alterations is neoadjuvant platinum-doublet chemotherapy in combination with immunotherapy (nivolumab), or adjuvant platinum-doublet chemotherapy followed by immunotherapy (atezolizumab for patients with a PD-L1 TPS of 50% or greater).^11,13 However, the clinical experts consulted for this review indicated that there is currently no consensus on whether neoadjuvant or adjuvant therapy is best, or if a combination of the 2 is beneficial. The clinical experts noted that the recommended neoadjuvant regimen for patients with resectable stage IB to IIIA NSCLC is 3 cycles of platinum-doublet chemotherapy in combination with nivolumab immunotherapy, and the adjuvant regimen is platinum-doublet chemotherapy, followed by atezolizumab immunotherapy if eligible. Neoadjuvant nivolumab is indicated for patients with tumours measuring 4 cm or greater or those that are node-positive, while adjuvant atezolizumab is indicated for patients with a PD-L1 TPS of 50% or greater.^29,30 However, the clinical experts indicated that not all patients receive immunotherapy or platinum-based chemotherapy in the perioperative setting as they may decline, not be offered a referral to medical oncology, or are ineligible. Additionally, the experts noted that patients with a PD-L1 TPS of less than 50% are not eligible for adjuvant atezolizumab, and these patients currently have no adjuvant immunotherapy options. There is a need to improve the rate of recurrences in early-stage NSCLC, as well as provide options for those who have not had neoadjuvant immunochemotherapy and are ineligible for adjuvant atezolizumab (as their tumours have a PD-L1 TPS of less than 50%) to cure disease, delay disease recurrence, improve survival, and maintain QoL.^10,14,15

Drug Under Review

The key characteristics of pembrolizumab are summarized in Table 3, with other treatments available as monotherapy for the adjuvant treatment of adult patients with stage IB (T2a ≥ 4 cm), II, or IIIA NSCLC who have undergone complete resection and platinum-based chemotherapy. The recommended dose of pembrolizumab in adults is 200 mg every 3 weeks or 400 mg every 6 weeks, administered as an IV infusion over 30 minutes.³¹

Pembrolizumab is a high-affinity antibody against PD-1, an immune checkpoint receptor that limits T-lymphocyte activity in peripheral tissues.³¹ Pembrolizumab reactivates tumour-specific cytotoxic T lymphocytes in the tumour microenvironment by exerting a dual-ligand blockade of the PD-1 pathway, including both PD-L1 and PD-L2, on antigen-presenting or tumour cells.³¹

Pembrolizumab has been approved by Health Canada as monotherapy for the adjuvant treatment of adult patients with stage IB (T2a ≥ 4 cm), II, or IIIA NSCLC who have undergone complete resection and platinum-based chemotherapy.³¹ The sponsor has requested reimbursement for pembrolizumab as monotherapy for the adjuvant treatment of adult patients with stage IB (T2a ≥ 4 cm), II, or IIIA NSCLC, and with a PD-L1 TPS of less than 50% who have undergone complete resection and platinum-based chemotherapy.

Table 3: Key Characteristics of Pembrolizumab and Atezolizumab

HSCT = hematopoietic stem-cell transplant; NSCLC = non–small cell lung cancer.

^aHealth Canada–approved indication.

Sources: Product monographs.^29,31

Testing Procedure Considerations

Tumour PD-L1 expression is determined using immunohistochemistry, which uses antibodies to detect the presence of specific biomarkers. PD-L1 testing results are commonly reported as the percentage of tumour cells that stain in the presence of an antibody, known as the TPS.^32,33 Approximately 72% to 85% of patients with early-stage NSCLC have a PD-L1 TPS of 49% or greater.³⁴ Multiple immunohistochemistry testing platforms are available for detecting and measuring PD-L1 expression in NSCLC, each co-developed as a companion or complementary diagnostic for a specific immune checkpoint inhibitor.^32,33,35 According to the clinical experts consulted by the review team, the 22C3 clone, used to determine PD-L1 expression status in the KEYNOTE-091 study, is a standard assay used in Canada.⁹

Testing for PD-L1 TPS at diagnosis is the recommended standard of care for all patients with NSCLC.^13,36 The clinical experts consulted by the review team verified that, in Canada, PD-L1 testing is routinely conducted from the biopsy sample. If no biopsy is performed, PD-L1 expression can be assessed using tissue from the surgically resected tumour. The clinical experts estimated that 95% of patients in Canada with NSCLC are tested for PD-L1 expression, with this proportion approaching 100% for patients seen in cancer centres.

We considered the potential impacts of PD-L1 TPS testing to determine eligibility for adjuvant treatment with pembrolizumab for patients with stage IB to IIIA NSCLC following complete tumour resection and platinum-based chemotherapy. We considered impacts on health systems and patients (including families and caregivers), and costs; these impacts are not anticipated to be substantial. Key considerations and relevant information available from materials submitted by the sponsor, input from the clinical experts consulted by the review team, and sources from the literature were validated by the review team when possible and are summarized in Table 4.

Table 4: Considerations for PD-L1 Testing for Establishing Treatment Eligibility for Pembrolizumab in Stage IB to IIIA NSCLC with Complete Resection and Platinum-Based Chemotherapy

NSCLC = non–small cell lung cancer; TPS = tumour proportion score.

^aCanada’s Drug Agency has not evaluated or critically appraised this evidence to determine its validity or reliability.

Perspectives of Patients, Clinicians, and Drug Programs

Patient Group Input

This section was prepared by the CDA-AMC review team based on the input provided by patient groups. The full original patient inputs received by CDA-AMC have been included in the Perspectives of Patients, Clinicians, and Drug Programs section of this report.

This review received a joint submission by 3 patient groups, the LHF, LCC, and the Canadian Cancer Survivor Network. The input was based on information collected by the LHF from individuals living with lung cancer, conducting 3 interviews with patients from Ottawa, Vancouver and Toronto who had experience with pembrolizumab, and gathering 33 responses to an online survey available between June 2023 to June 2024.

When asked about their disease experience, interviewees reported substantial challenges obtaining an accurate and timely diagnosis, which resulted in substantial declines in quality of life until diagnosis. Most had difficulty verbally communicating during this period due to interruptions by coughing fits, which resolved for some upon treatment initiation. Survey respondents reported similar symptoms and challenges due to their lung cancer, some of which included: fatigue (53%), shortness of breath (50%), cough (23%), and pain (20%). Respondents also noted chest tightness, wheezing, weight loss, diminished appetite, low mood and/or depressive periods, and challenges with physical and emotional intimacy. When asked how lung cancer negatively affects their day-to-day life, respondents highlighted their inability to work (48%), inability to participate in physical activities (33%), do housework (21%), use stairs (22%), or engage in hobbies (21%). Most respondents indicated that living with lung cancer negatively affects their emotional well-being through feelings of isolation, challenges with symptom management, and perceived burden on caregivers and family. Disease aspects that were most important to responders to control included improved management of disease symptoms, as well as pain and side effects from therapy.

Respondents reported some benefit from previous treatment with alectinib, lorlatinib, metoclopramide, gefitinib, entrectinib, osimertinib, and brigatinib, such as reduced cough and shortness of breath, increased participation in daily activities, ability to exercise, prolonged life, delayed disease progression and a reduction in the severity of other disease-related symptoms. The input also noted that patients on oral therapies value the flexibility such therapies provide in allowing them to work and travel without restrictions. However, respondents reported struggling with lingering side effects. Patients who previously received surgery reported experiencing deconditioning and chronic fatigue, and medication side effects included extreme itching that affects sleep, brain fog, fatigue, nausea, vomiting, mood changes, diminished appetite, weight loss, hair loss, anemia, and neuropathy. These patients also experienced challenges accessing some therapies due to high treatment costs, as well as difficulty navigating the health care system and locating disease information and support. The input also noted concerns from patients on targeted therapy as to their ability to access the next line of treatment if or when their current therapy stops working.

Respondents indicated that key treatment outcomes to consider when evaluating new therapies include stopping or slowing disease progression with minimal side effects, and effectiveness in advanced disease. Respondents also highlighted efficacy as an outcome of interest, with 1 respondent noting they would be more receptive to side effects if there was robust evidence that the medication would stop or slow their lung cancer progression.

Three LHF interviewees had experience with the drug under review. These patients (2 of whom had NSCLC with PD-L1 expression) were diagnosed with stage IV lung cancer and were initiated on pembrolizumab shortly after surgical resection. However, these patients were not part of the eligible population of the current indication under review, which is limited to stage IB (T2a ≥ 4 cm), II, or IIIA NSCLC. One patient was diagnosed with stage I lung cancer, received surgery, and was initiated on pembrolizumab after progression to stage IV following a 3-year disease-free period. One interviewee with EGFR-positive, stage IV lung cancer was taking pembrolizumab in combination with chemotherapy, although it was unclear if the interviewees with NSCLC were taking pembrolizumab similarly or as monotherapy. One patient discontinued pembrolizumab after 19 months due to progression, while another experienced sufficient tumour shrinkage and inactivity to discontinue pembrolizumab after 3 years, before re-initiating shortly after tumour reactivation. After re-initiation of pembrolizumab, the patient’s tumours once again decreased in size. Side effects reported by these patients included nausea, fatigue, muscle soreness, constipation, diarrhea, and worsening of their diabetes, eczema, and liver enzymes. Patients did not report that these side effects impeded their ability to participate in activities of daily living or exercise and, overall, described an improved QoL while on therapy.

Clinician Input

Input From Clinical Experts Consulted by CDA-AMC

All CDA-AMC review teams include at least 1 clinical specialist with expertise in the diagnosis and management of the condition for which the drug is indicated. Clinical experts are a critical part of the review team and are involved in all phases of the review process (e.g., providing guidance on the development of the review protocol, assisting in the critical appraisal of clinical evidence, interpreting the clinical relevance of the results, and providing guidance on the potential place in therapy). The following input was provided by 2 clinical specialists with expertise in the diagnosis and management of lung cancer.

Unmet Needs

According to the clinical experts consulted by the review team, the key treatment goals for patients with early-stage NSCLC who receive adjuvant therapy following complete surgical resection is to improve cure rates through reducing risk of relapse.

The clinical experts noted that a risk of recurrence remains even after complete resection. The experts highlighted 2 approaches to curative-intent treatment: neoadjuvant therapy and adjuvant therapy. For patients with resectable NSCLC without actionable mutations, the current patient management in Canada is neoadjuvant platinum-based chemotherapy in combination with nivolumab, or adjuvant platinum-based chemotherapy, followed by atezolizumab if they have a PD-L1 TPS of 50% or greater. The clinical experts added that no adjuvant immunotherapy options are available for patients with stage IB to IIIA NSCLC who have a PD-L1 TPS of less than 50% in Canada, defining an unmet need for this group.

Place in Therapy

According to the clinical experts consulted by the review team, pembrolizumab would represent a new adjuvant immunotherapy option for patients with resected stage IB to IIIA NSCLC who have a PD-L1 TPS of less than 50% and who have not had access to neoadjuvant immunochemotherapy. The clinical experts agreed that adjuvant pembrolizumab would fill a treatment gap, as currently patients can only access this treatment after relapse.

Patient Population

The clinical experts consulted for this review indicated that the patient population most suited for treatment with adjuvant pembrolizumab are those with resectable NSCLC and a PD-L1 TPS of less than 50% using a validated PD-L1 assay.

Assessing the Response Treatment

Both clinical experts consulted by the review team agreed that the gold-standard outcome to use when determining a patient’s response to treatment in clinical practice will be overall survival.

Discontinuing Treatment

Clinical experts consulted for this review indicated that pembrolizumab treatment should be discontinued if 1 of the following has been met: a total of 18 cycles (1 year) of adjuvant immunotherapy has been completed, disease progression has been detected, or unacceptable toxicity develops. One expert noted that toxicity related to gastrointestinal, skin, lung, heart, central nervous system, or endocrine functions is particularly important when determining whether to discontinue therapy.

Prescribing Considerations

The clinical experts noted that pembrolizumab should be administered in a specialty setting by multidisciplinary surgical and medical oncology staff with the expertise to administer systemic therapy, monitor the patient, and manage treatment-related toxicities.

Clinician Group Input

This section was prepared by the CDA-AMC review team based on the input provided by clinician groups. The full original clinician group inputs received by the CDA-AMC review team have been included in the Perspective of Patients, Clinicians, and Drug Programs section of this report.

Clinician group input on this review was received from 2 clinician groups: The Ontario Health (Cancer Care Ontario Drug Advisory Committee and LCC. Six clinicians from the Drug Advisory Committee and 35 clinicians from LCC provided input for this review.

For the treatment of resected stage IB to IIIA NSCLC, the LCC group indicated that chemotherapy, immunotherapy, radiation, and targeted drugs play a key role in improving outcomes and cure rates. Aligning with clinical expert input, the Drug Advisory Committee emphasized that, in the subset of patients with resected stage IB to IIIA NSCLC who have a PD-L1 of less than 50%, there are currently no other therapy options available after adjuvant chemotherapy, creating an urgent unmet need for this patient population. The LCC group also highlighted that, in patients with early-stage resected IB to IIIA NSCLC, current therapies do not achieve sufficiently high cure rates or prevent recurrences. The LCC group described this as particularly important for patients with NSCLC, as the risk of relapse substantially increases with each subsequent disease stage. The LCC group also emphasized that patient relapse and metastatic disease also come with substantial costs to patient health, QoL, utilization of health care resources, economic loss of productivity, and overall costs to society.

Aligning with clinical expert input, the clinicians agreed that pembrolizumab is an option for patients with stage IB to IIIA NSCLC with a PD-L1 TPS of less than 50% and proposed using it as a second adjuvant therapy following adjuvant chemotherapy. The LCC group expected that pembrolizumab will shift the current treatment paradigm, as it represents the first adjuvant immunotherapy option for this patient population. Clinician groups agreed that all patients with stage IB to IIIA NSCLC who have undergone complete resection and who have a PD-L1 TPS of less than 50% would be suitable for pembrolizumab in the absence of contraindications to immunotherapy. The LCC group expected that clinicians will continue to treat patients with a PD-L1 TPS of 50% or greater or ALK translocations with other available drugs that have demonstrated superior efficacy in these subgroups. The LCC group also pointed out that treatment with pembrolizumab for eligible patients with a sensitizing EGFR mutation should be considered on case-by-case basis by treating clinicians, weighing the risks and benefits of adjuvant sequential chemotherapy and immunotherapy versus adjuvant osimertinib. The Ontario Health Drug Advisory Committee and clinical experts agreed that pembrolizumab should be considered for patients who are ineligible for chemotherapy, as study patients were permitted to receive it even without chemotherapy.

The groups agreed that treatment benefit in the adjuvant setting is primarily determined by disease recurrence, which the LCC group noted typically occurs within 2 to 3 years for patients with stage IB to IIIA NSCLC. The LCC group indicated that cure rates, as measured by 5-year OS, can also determine response, but typically require even more years of follow up. The LCC group suggested implementing clinical and laboratory follow up every 3 weeks to evaluate toxicity and disease recurrence, as well as imaging scans at 3- to 4-month intervals, as pembrolizumab is administered over 1 year. The LCC group noted that, overall, immunotherapies are well tolerated by patients, and autoimmune side effects can often be readily managed by oncologists. Both clinician groups agreed that treatment could be delivered in an outpatient setting under the supervision of a medical oncologist; however, the Drug Advisory Committee suggested that a pulmonologist experienced in managing thoracic malignancies could also diagnose, treat, and monitor patients on pembrolizumab. Both clinician groups, as well as the consulted clinical experts, agreed that treatment would be discontinued in the event of disease recurrence or progression, SAEs, or completion of therapy after 18 cycles.

Drug Program Input

The drug programs provide input on each drug being reviewed through the CDA-AMC reimbursement review process by identifying issues that may affect their ability to implement a recommendation. The implementation questions and corresponding responses from the clinical experts consulted by the CDA-AMC review team are summarized in Table 5.

Table 5: Summary of Drug Plan Input and Clinical Expert Response

ACT = adjuvant chemotherapy; AE = adverse event; ECOG = Eastern Cooperative Oncology Group Performance Status; NAT = neoadjuvant treatment; NSCLC = non–small cell lung cancer; pERC = pan-Canadian Oncology Drug Review Expert Review Committee; vs. = versus.

Clinical Evidence

The objective of this CDA-AMC Clinical Review Report is to review and critically appraise the clinical evidence submitted by the sponsor on the beneficial and harmful effects of pembrolizumab (administered as an IV infusion of either 200 mg every 3 weeks or 400 mg every 6 weeks) for the adjuvant treatment of adult patients with stage IB (T2a ≥ 4 cm), II, or IIIA NSCLC who have undergone complete resection and platinum-based chemotherapy and whose tumours have a PD-L1 TPS of less than 50%, as determined by a validated test. The focus will be placed on comparing pembrolizumab to relevant comparators and identifying gaps in the current evidence.

A summary of the clinical evidence included by the sponsor in the review of pembrolizumab is presented in 4 sections, with the CDA-AMC critical appraisal of the evidence included at the end of each section. The first section, the systematic review, includes pivotal studies and RCTs that were selected according to the sponsor’s systematic review protocol. The CDA-AMC assessment of the certainty of the evidence in this first section using the GRADE approach follows the critical appraisal of the evidence. The second section includes sponsor-submitted long-term extension studies. The third section includes indirect evidence from the sponsor. The fourth section includes additional studies that were considered by the sponsor to address important gaps in the systematic review evidence. No long-term extension studies (Section 2) or additional studies to address important gaps in the systematic review evidence (Section 4) were submitted by the sponsor.

Included Studies

Clinical evidence from the following is included in the CDA-AMC review and appraised in this document:

• 1 pivotal placebo-controlled RCT

• 1 indirect treatment comparison.

Systematic Review

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

Description of Studies

Characteristics of the included studies are summarized in Table 6.

Table 6: Details of Studies Included in the Systematic Review

DFS = disease-free survival; ECOG PS = Eastern Cooperative Oncology Group Performance Status; HRQoL = health-related quality of life; IHC = immunohistochemistry; NSCLC = non–small cell lung cancer; OS = overall survival; RCT = randomized controlled trial; TPS = tumour proportion score; UICC v7 = Union for International Cancer Control version 7.

Note: One additional report was included: O’Brien et al. (2022).⁹

Source: Clinical Study Report for P091V02MK3475.²³

One pivotal, multicentre, triple-blind, placebo-controlled randomized phase III trial (KEYNOTE-091) met the inclusion criteria for the sponsor’s systemic review, in which a total of 206 sites randomized 1,177 patients (Table 6). The KEYNOTE-091 trial investigated the efficacy and safety of pembrolizumab compared with placebo as adjuvant therapy for completely resected stage IB to IIIA NSCLC.^23,39 Eligible patients were randomized 1:1 to receive pembrolizumab or placebo using a central interactive voice-response system (Figure 1). Randomization was stratified by disease stage (IB versus II versus IIIA), receipt of ACT (no versus yes), PD-L1 TPS (< 1% [negative] versus 1% to 49% versus ≥ 50%) and geographic region (Asia versus Eastern Europe versus Western Europe versus the rest of the world). Patients, investigators, and individuals who collected or analyzed the data, including sponsor representatives, were masked to treatment assignment.^39,40

Scans of the chest and upper abdomen were performed by CT within 12 weeks before randomization, 12 weeks after the first dose of study treatment, and every 12 weeks thereafter during year 1, every 6 months during years 2 and 3, annually during years 4 and 5, and according to local standard of care thereafter until disease recurrence or withdrawal of consent. CT or MRI scans of the brain were performed within 12 weeks before randomization, and only if clinically indicated thereafter. After treatment discontinuation, survival was assessed every 12 weeks through year 5 and every 6 months thereafter.^39,40

The KEYNOTE-091 study is ongoing, and the results reported in this review are based on the protocol-prespecified IA3, for which the data cut-off date was January 24, 2023.^23,39

Figure 1: Study Design for the KEYNOTE-091 Study

ACT = adjuvant chemotherapy; ADA = antidrug antibody; AJCC v7 = American Joint Committee on Cancer version 7; DFS = disease-free survival; ECOG PS = Eastern Cooperative Oncology Group Performance Status; EORTC QLQ-C30 = European Organisation for Research and Treatment of Cancer Quality of Life Questionnaire Core 30; IHC = immunochemistry; LCSS = lung cancer–specific survival; NACT = neoadjuvant chemotherapy; NSCLC = non–small cell lung cancer; OS = overall survival; PK = pharmacokinetics; PRO = patient-reported outcome; Q3W = every 3 weeks; Q6M = every 6 months; Q12M = every 12 months; Q12W = every 12 weeks; QLQ-LC13 Quality of Life Questionnaire Lung Cancer Module; TPS = tumour proportion score.

Source: Sponsor’s submission.³⁹

Populations

Inclusion and Exclusion Criteria

In the KEYNOTE-091 study, eligible patients were aged 18 years or older, had pathologically confirmed NSCLC (any histology) of stage IB (T2a ≥ 4 cm), II, or IIIA according to the AJCC staging system¹⁶ after complete surgical resection (lobectomy, sleeve lobectomy, bi-lobectomy, or pneumonectomy) and negative margins (R0). Eligible patients had an available tumour sample obtained during resection for PD-L1 assessment, known PD-L1 expression status, no evidence of disease on clinical examination and radiographic assessment according to RECIST 1.1)¹⁷ as determined by local investigator review after surgery but within 12 weeks before randomization, an ECOG PS of 0 or 1, and adequate organ function within 10 days of treatment initiation. Previous neoadjuvant or adjuvant radiotherapy for the current malignancy was not permitted. ACT was not mandatory but was to be considered for patients with stage IB disease and strongly recommended for those with stage II or IIIA disease according to local practice and national guidelines. Patients without ACT had to receive their first study treatment dose within 12 weeks of surgery. Patients who received ACT had to initiate it within 12 weeks of surgery and receive a maximum of 4 chemotherapy cycles; the first dose of study treatment was administered at least 3 weeks but no more than 12 weeks after the last dose of chemotherapy. Patients with prior treatment with an anti–PD-1, anti–PD-L1 or anti–PD-L2, anti-CD137, CTLA-4 modulators, or any other immune-modulating drugs were excluded from the study.^39,40

Interventions

Pembrolizumab 200 mg or placebo was administered by IV infusion on day 1 of every 3-week cycle until recurrence (as determined by the investigator according to RECIST 1.1),¹⁷ new malignancy, unacceptable toxicity, withdrawal of consent, completion of 18 cycles (approximately 1 year) of treatment or other reason (intercurrent illness, failure to comply with study treatment or procedure requirements, pregnancy, investigator’s decision, or administrative reasons).^39,40

Outcomes

A list of efficacy end points assessed in this Clinical Review is provided in Table 7. Summarized end points are based on outcomes included in the sponsor’s Summary of Clinical Evidence as well as any outcomes identified as important to this review by the clinical experts consulted for this review and input from patient and clinician groups and public drug plans. Using the same considerations, the review team selected end points that were considered to be most relevant to inform the expert committee deliberations and finalized this list of end points in consultation with members of the expert committee. Select efficacy and harms outcomes that were considered important for informing the expert committee deliberations were assessed using GRADE.

Table 7: Outcomes Summarized From the KEYNOTE-091 Study

AE = adverse event; DFS = disease-free survival; EORTC QLQ-C30 = European Organisation for Research and Treatment of Cancer Quality of Life Questionnaire Core 30; EORTC QLQ-LC13 = European Organisation for Research and Treatment of Cancer Quality of Life Questionnaire Quality of Life Questionnaire Lung Cancer Module; OS = overall survival.

^aStatistical testing for these end points was adjusted for multiple comparisons.

^bAEs were recorded according to the Common Terminology Criteria for Adverse Events version 4.03.

Source: Clinical Study Report for P091V02MK3475.²³

In the KEYNOTE-091 study, the dual primary end points were DFS in the PD-L1 strong positive (TPS ≥ 50%) population or in the overall population. Secondary and exploratory end points included DFS in the population with a PD-L1 TPS of 1% or greater, OS in different populations, HRQoL, and safety, among others.^39,40

Overall survival was defined as the time from randomization to death due to any cause.

Disease-free survival was defined as the time from randomization to locoregional or metastatic recurrence assessed under RECIST 1.1¹⁷ by investigator review, appearance of a second primary NSCLC or other malignancy, or death from any cause.

A summary of EORTC QLQ-C30 and EORTC QLQ-LC13 results is provided in Table 8. All HRQoL questionnaires were filled out at the hospital during scheduled visits. Pre-treatment questionnaires were completed within 12 weeks before randomization. Subsequent questionnaires were filled in every 12 weeks during the first year after randomization (starting on day 1 cycle 1), every 6 months during the second year and then yearly until the fifth year. HRQoL data were collected regardless of the patient’s progression status, before the pembrolizumab infusion.

Any AEs, irrespective of causality, were reported from the time of treatment randomization through 30 days after the last dose of study treatment (intensity was assessed by the investigator according to the National Cancer Institute Common Terminology Criteria for Adverse Events, version 4.03).

Immune-mediated events and infusion-related reactions associated with pembrolizumab were considered AEOSIs. A predefined list of preferred terms was developed by the sponsor to consistently characterize the nature and frequency of each AEOSI regardless of causality as reported by investigators. These preferred terms are considered medically equivalent to the immune-mediated events and infusion-related reactions. AEOSIs were reported from the time of randomization through 90 days after the last dose of study treatment or 30 days following discontinuation of study treatment if the patient initiated a new anticancer treatment, whichever occurred first.

Considerations that informed the selection of efficacy outcomes to be summarized and assessed using GRADE include the following:

• Survival outcomes were identified by the patient and clinician group input, and specified by the clinical experts consulted for this review to include OS and DFS. Of the Kaplan-Meier estimates of probability of the survival outcomes at multiple time points, those at 36 months and 48 months for OS and 24 months and 48 months for DFS were specified by the clinical experts.

• HRQoL outcomes were identified by the patient and clinician group input, and specified by the clinical experts to include the EORTC QLQ-C30 GHS/QoL and the select EORTC QLQ-LC13 symptom scales (chest pain, coughing, and dyspnea).

Table 8: Summary of Outcome Measures and Their Measurement Properties

BPI = Brief Pain Inventory; ECOG PS = Eastern Cooperative Oncology Group Performance Status; EORTC = European Organisation for Research and Treatment of Cancer; FACT-G = Functional Assessment of Cancer Therapy–General; GHS = Global Health Status; HADS = Hospital Anxiety and Depression Scale; HRQoL = health-related quality of life; MID = minimal important difference; NSCLC = non–small cell lung cancer; QLQ-C30 = Quality of Life Questionnaire Core 30; QLQ-LC13 = Quality of Life Questionnaire Lung Cancer Module; SSQ = subjective significance questionnaire; vs. = versus.

Statistical Analysis

Clinical Trial End Points

Estimates of the median OS, and DFS were obtained by the Kaplan-Meier technique. The 95% CIs for the medians were calculated using the reflected CI. Estimates of the event-free rate at a fixed time point were obtained using the Kaplan-Meier technique and 95% CIs were calculated by the Greenwood formula for standard deviation. Estimates of HRs and their 95% CI were obtained by Cox regression.^39,40 For patient-reported HRQoL outcomes, no PRO analyses were planned in any subpopulation.³⁹ Details about the statistical analyses for each of the selected end points are presented in Table 9.

Table 9: Statistical Analysis of Efficacy End Points for the KEYNOTE-091 Study (All Patients)

DFS = disease-free survival; EORTC QLQ-C30 v3 = European Organisation for Research and Treatment of Cancer Quality of Life Questionnaire Core 30 version 3; EORTC QLQ-LC13 = European Organisation for Research and Treatment of Cancer Quality of Life Questionnaire Lung Cancer Module; HRQoL = health-related quality of life; NR = not reported; OS = overall survival; vs. = versus.

Sources: Clinical Study Report: P091V02MK3475²³ and sponsor’s submission.²⁴

Sample Size and Power Calculation

The KEYNOTE-091 study was designed to enroll approximately 1,180 patients randomized 1:1 into the experimental and control arms. For DFS, the study has approximately 86% power at an alpha of 1.25% (1-sided) and approximately 92% power at an alpha of 2.5% (1-sided) in the overall study population based on a target number of approximately 551 events at final analysis. It was expected that there would be approximately 334 patients with a PD-L1 TPS of 50% or greater in the study to achieve about 90% power for DFS at an alpha of 1.25% (1-sided) and approximately 94% power at an alpha of 2.5% (1-sided) in the final DFS analysis in this subgroup.

Statistical Testing

The KEYNOTE-091 study was designed with 2 primary end points: DFS in the whole population and DFS in the subgroup with a PD-L1 TPS of 50% or greater. The 1-sided family-wise error rate for DFS and OS hypotheses was strongly controlled at 2.5%.²³ The study used the graphic method of Maurer and Bretz to provide strong multiplicity control for multiple hypotheses as well as interim efficacy analyses. The alpha was equally split to test DFS in the overall population and in patients in the subgroup with a TPS of 50% or greater (Figure 2). If the null hypothesis of DFS was rejected in that subgroup, its alpha would be fully reallocated to the DFS hypothesis in the subgroup with a TPS of 1% or greater. The OS hypotheses testing in the overall population and the subgroup with a TPS of 50% or greater would not occur until the DFS hypothesis was rejected in the overall populations. The actual boundaries were updated based on the number of events observed at the time of analysis using the Hwang-Shih-DeCani spending function with a gamma of −4.⁴⁰

The protocol specified 2 event-driven interim analyses for DFS. Analyses of OS were planned in all analyses conducted for DFS. Additional OS analyses were also planned after the final DFS analysis. The final OS analysis was scheduled to be conducted no later than 120 months after the first randomized patient if the accumulation of OS events was much slower than expected and the final target OS events could not be reached by 120 months.⁴⁰ In this situation, all remaining alpha would be used in the final OS analysis. An alpha reallocation strategy was used to address multiplicity issues across multiple end points and interim analyses (Figure 2).⁴⁰

Analyses of GHS/QoL scores of the EORTC QLQ-C30 were prespecified key exploratory PRO end points. Analyses of the following disease-related symptoms of the EORTC QLQ-C30 and EORTC QLQ-LC13 were prespecified supportive exploratory PRO end points: pain, fatigue, appetite loss, shortness of breath, cough, chest pain, and dyspnea.³⁹ In the statistical analysis plan, week 48 was selected as the primary time point for the mean change from baseline analysis to ensure that completion rates reached or exceeded 60% and compliance rates reached or exceeded 80% across the treatment groups.^23,39

The results reported in this review are based on IA3, which is the final analysis for DFS and an interim analysis for OS. The data cut-off date for IA3 was January 24, 2023 (first patient first visit, November 10, 2015; database lock date, March 15, 2023).³⁹

Figure 2: Multiplicity Graph for Alpha Reallocation Strategy in the KEYNOTE-091 Study

DFS = disease-free survival; OS = overall survival; PD-L1+ = PD-L1 TPS of 1% or greater; PD-L1++ = PD-L1 TPS of 50% or greater; TPS = tumour proportion score.

Source: KEYNOTE-091 study protocol.⁴⁰

Subpopulation of Interest

Efficacy analyses for OS and DFS were carried out on an ad hoc basis in the subpopulation of patients with ACT and a PD-L1 TPS of less than 50% (the reimbursement request population). The models used in these analyses are the same as those described in Table 9.

Analysis Populations

Analyses of OS and DFS were conducted on all randomized patients according to the ITT principle, as well as in the subpopulation of interest in this review, i.e., all randomized patients who received prior ACT and had a PD-L1 TPS of less than 50% (the reimbursement request population). Safety analyses were based on all patients who received at least 1 dose of study treatment (APaT).⁴⁰ Key analysis populations are presented in Table 10.

Table 10: Analysis Populations in the KEYNOTE-091 Study

ACT = adjuvant chemotherapy; APaT = all participants as treated; FAS = full analysis set; HRQoL = health-related quality of life; ITT = intention-to-treat; PRO = patient-reported outcome; TPS = tumour proportion score.

Sources: Clinical Study Report for P091V02MK3475²³ and BARDS Health Technology Assessment Statistical Report: Baseline Characteristics and Efficacy (April 9, 2024).²²

Results

Patient Disposition

Patient screening in all patients and patient disposition as of the IA3 data cut-off date (January 24, 2023) in the reimbursement request population is summarized in Table 11. In the pembrolizumab (N = 363) and placebo (N = 363) groups, 171 (48%) and 133 (37%) patients from these 2 groups discontinued treatment before completing 18 cycles, respectively. The most common reasons for treatment discontinuation were recurrence, relapse, or death due to NSCLC (13% in the pembrolizumab group and 23% in the placebo group, respectively) and toxicity related to study medication (18% and 4%, respectively). The median numbers of treatment cycles administered were 17 in the pembrolizumab group and 18 in the placebo group.

Table 11: Summary of Patient Disposition in the KEYNOTE-091 Study

ACT = adjuvant chemotherapy; ECOG PS = Eastern Cooperative Oncology Group Performance Status; FAS = full analysis set; ITT = intention to treat; NA = not available; PRO = patient-reported outcome; TPS = tumour proportion score.

^aThe PRO results were only analyzed in randomized patients who received at least 1 dose of study treatment and who completed at least 1 PRO assessment. These analyses were not conducted in the reimbursement request population (i.e., the subpopulation of patients who received prior ACT and had a PD-L1 TPS of less than 50%).

^bSafety data were reported in the overall APaT population (i.e., in all patients randomized who received at least 1 dose of study treatment) to provide an exhaustive picture of pembrolizumab safety in the adjuvant setting (i.e., larger population exposed to treatment than the subgroup constituting the reimbursement request population).

Sources: BARDS Health Technology Assessment Statistical Report: Baseline Characteristics and Efficacy (April 9, 2024)²² and sponsor’s submission.²⁴

Baseline Characteristics

The baseline characteristics outlined in Table 12 are limited to those that are most relevant to this review or were deemed to affect the outcomes or interpretation of the study results.

In the KEYNOTE-091 study, the median age across the groups in the overall ITT population and the reimbursement request population was 64 to 65 years. The proportion of male patients (64.5% to 68.7%) was higher than that of female patients (31.3% to 35.5%). Most enrolled patients were white (76.3% to 77.7%), followed by Asian (17.4% to 18.2%), of multiple ethnicities (0 to 1.1%), other (0.3% to 1.0%), or Black or African American (0 to 0.8%). Most patients in the overall ITT population (86%) had received 3 or 4 cycles of prior ACT with a cisplatin- or carboplatin-based regimen or both. All patients in the reimbursement request population (100%) had received ACT. A relatively small proportion of patients had stage IB disease (11.0% to 14.8% across the groups in both populations of the study) and more than half (55.4% to 59.0%) had stage II disease. Most patients (68.0% to 73.4%) were former smokers, followed by those who had never smoked (11.2% to 19.0%), and current smokers (12.7% to 15.3%). More patients (56.2% to 64.4%) had an ECOG PS of 0 compared to those with an ECOG PS of 1 (35.6% to 43.8%). The proportions of patients who had the EGFR mutation (5.8% to 8.5%) or ALK translocation (0.8% to 1.7%) were low. More than half of the patients (51.8% to 57.4%) had an unknown status for EGGR mutations or ALK translocations (56.7% to 66.4%). A total of 333 patients (28%) in the overall ITT population had a PD-L1 TPS of 50% or greater and 844 (72%) had a PD-L1 TPS of less than 50%. Most patients in the reimbursement request population had a lymph-node stage of N0 (no regional lymph-node involvement, 41.3%) or N1 (nearby lymph-node involvement, 37.5%), followed by N2 (involvement of lymph nodes in the mediastinum, 21.2%).⁴⁸ Overall, the baseline characteristics were well balanced between both treatment groups in the overall 2 ITT populations. Despite that, there was a slightly higher variation in histology in the reimbursement request population: 27.8% in the pembrolizumab group versus 37.5% in the placebo group.⁴⁸

Table 12: Summary of Baseline Characteristics in the KEYNOTE-091 Study

ACT = adjuvant chemotherapy; AJCC = American Joint Committee on Cancer version 7; ALK = anaplastic lymphoma kinase; ECOG PS = Eastern Cooperative Oncology Group Performance Status; EGFR = epidermal growth factor receptor; ITT = intention-to-treat; NR = not reported; SD = standard deviation; TPS = tumour proportion score.

^aThe lymph nodes stages in the KEYNOTE-091 study included N0 (no regional lymph-node involvement), N1 (nearby lymph-node involvement), and N2 (involvement of lymph nodes located in the mediastinum).

Sources: Clinical Study Report for P091V02MK3475,²³ BARDS Health Technology Assessment Statistical Report: Baseline Characteristics and efficacy (April 9, 2024),²² and sponsor’s submission.⁴⁸

Exposure to Study Treatments

In the reimbursement request population, the mean durations of treatment exposure were 266 days in the pembrolizumab group and 302 days in the placebo group; and the median duration of treatment was similar in the 2 treatment arms (Table 13). Among patients who received at least 1 dose of their assigned study treatment, 244 (68%) and 292 (81%) in the pembrolizumab and placebo groups, respectively, remained on study treatment for 6 months or longer (Table 13).

Table 13: Summary of Patient Exposure in the KEYNOTE-091 Study (Reimbursement Request Population)

ACT = adjuvant chemotherapy; SD = standard deviation; TPS = tumour proportion score.

^aEach patient is counted once on each applicable duration category row. Duration of exposure is the time from the first dose date to the last dose date.

^bOne month = 30.4367 days.

Source: BARDS Health Technology Assessment Statistical Report: Baseline Characteristics and efficacy (April 9, 2024).²²

Concomitant Medications and Cointerventions

In the reimbursement request population, 346 patients (95.3%) in the pembrolizumab group and 342 patients (94.2%) in the placebo group received 1 or more concomitant medications during the study.⁴⁹ The concomitant medications were generally balanced between the treatment groups for most reported medications.⁴⁹ The use of the following concomitant medications were higher in the pembrolizumab group compared with the placebo group: “antidiarrheals, intestinal anti-inflammatory/anti-infective agents” (42.7% versus 26.4%, respectively), systemic antihistamines (28.1% versus 17.6%), systemic corticosteroids (43.3% versus 27.3%), and thyroid replacement therapy (23.7% versus 12.1%).⁴⁹

Subsequent Treatment

In the reimbursement request population, a smaller proportion of patients initiated subsequent antineoplastic therapy in the pembrolizumab group (19.3%) compared with the placebo group (28.1%). Similarly, a smaller proportion of patients initiated subsequent immunotherapies in the pembrolizumab group (4.1%) compared with the placebo group (14.6%) (Table 14). In the reimbursement request population, the most prescribed subsequent systemic anticancer therapies in the pembrolizumab group were carboplatin or pemetrexed (4.7%), docetaxel (3.0%), carboplatin or paclitaxel (2.8%), and pemetrexed (2.5%). In the placebo group, the most prescribed subsequent therapies were pembrolizumab (6.6%), carboplatin or pemetrexed (5.2%), atezolizumab (5.0%), pemetrexed (3.9%), docetaxel (3.3%), carboplatin (3.0%), and nivolumab (2.2%) (Table 14).

Table 14: Summary of Subsequent Treatment in the KEYNOTE-091 Study (Reimbursement Request Population)

ACT = adjuvant chemotherapy; TPS = tumour proportion score.

Source: BARDS Health Technology Assessment Statistical Report: Baseline Characteristics and Efficacy (April 9, 2024).²²

Efficacy

Survival Outcomes

The OS and DFS outcomes for the reimbursement request population are presented in Table 15, Figure 3, and Figure 4. As of the data cut-off date of January 24, 2023, the median duration of follow-up was 46.6 months (range = 0.6 to 84.2) for the reimbursement request population.²⁴

Overall Survival

As of the data cut-off date of January 24, 2023, the median OS was not reached in either group ███████ █████ █ █████ █████████████ ████████ █████ ████ ██ █████ | | ██████. The Kaplan-Meier estimates of the probability of OS in the pembrolizumab and placebo groups were █████ ████ ███ ████ ██ █████ versus █████ ████ ███ ████ ██ █████ at 36 months; and █████ ████ ███ ████ ██ █████ versus █████ ████ ███ ████ ██ █████ at 48 months, respectively.

The results from the prespecified sensitivity analyses of OS in the overall ITT population were generally consistent with those of the primary analysis. No results are available for the reimbursement request population.²³

In the reimbursement request population, the OS results were consistent for point estimates of the HR across subgroups; however, the 95% CIs of HRs of nearly all categories (except for age) in the subgroups included the null value (P values for interaction tests for all the subgroups > 0.05) (Table 21).

Disease-Free Survival

As of the data cut-off date of January 24, 2023, the median DFS values were 51.7 months (95% CI, 39.0 to 70.4) for patients treated with pembrolizumab and 34.5 months (95% CI, 23.3 to 46.4) for patients who received placebo (HR = 0.72; 95% CI, 0.58 to 0.89; P < 0.001). The Kaplan-Meier estimates of the probability of DFS in the pembrolizumab and placebo groups were 67.2% (95% CI, 61.9 to 71.9) versus 55.0% (95% CI, 49.7 to 60.0) at 24 months; and 51.2% (95% CI, 45.2 to 56.9) versus 42.4% (95% CI, 36.7 to 47.9) at 48 months, respectively.

The results from the prespecified sensitivity analysis of DFS using actual stages according to the AJCC were consistent with the primary analysis results for the reimbursement request population.²⁴

In the reimbursement request population, the DFS results were generally consistent across subgroups, including those for age, sex, race, region, disease stage, smoking status (never smoker, former smoker, or current smoker), histology (squamous or nonsquamous), ECOG PS (0 or 1), EGFR mutation status (no, yes, or unknown), and PD-L1 status (< 1% or 1% to 49%), with the 95% CIs of HRs of a few categories in subgroups marginally crossing 1 (P values for interaction test for all the subgroups > 0.05). The subgroup of patients with disease stage IB showed a wide 95% CI crossing the null value (HR = 0.66; 95% CI, 0.32 to 1.35) compared with patients with stage II to IIIA NSCLC (HR = 0.79; 95% CI, 0.63 to 0.98) (Table 22).

Table 15: Summary of Key Efficacy Results in the KEYNOTE-091 Study (Reimbursement Request Population)

ACT = adjuvant chemotherapy; CI = confidence interval; DFS = disease-free survival; HR = hazard ratio; NR = not reported; NSCLC = non–small cell lung cancer; OS = overall survival; TPS = tumour proportion score; vs. = versus.

^aBased on the multivariate Cox regression model with treatment adjusted by the following covariates: stage (IB vs. II vs. IIIA), region (Western Europe vs. Eastern Europe vs. rest of world vs. Asia), histology (squamous vs. nonsquamous), and smoking status (never vs. former or current).

^bOne-sided P value based on the Wald test with a multivariate Cox regression model.

^cA summary of disease status, showing number and percentage of participants by treatment arm per type of first event in DFS analysis (no event vs. event [not disease-free at baseline/local and/or regional recurrence/distant metastasis/both/new malignancy/death]). DFS was defined as the time from randomization to the first documented loco-regional recurrence, occurrence of distant metastasis(es), a second primary NSCLC or second malignancy, based on investigator assessment or death due to any cause, whichever occurred first, expressed in months.

^dNew malignancy included the second primary and second malignancies.

Sources: BARDS Health Technology Assessment Statistical Report: Baseline Characteristics and Efficacy (April 9, 2024)²² and sponsor’s submission.⁴⁹

Figure 3: Kaplan-Meier Estimates of Overall Survival in Patients With Prior ACT and PD-L1 TPS Less Than 50% in the KEYNOTE-091 Study (Reimbursement Request Population) [Redacted]

Figure 4: Kaplan-Meier Estimates of Disease-Free Survival in Patients With Prior ACT and PD-L1 TPS Less Than 50% in the KEYNOTE-091 Study (Reimbursement Request Population)

ACT = adjuvant chemotherapy; TPS = tumour proportion score.

Source: BARDS Health Technology Assessment Statistical Report: Baseline Characteristics and Efficacy (April 9, 2024).²²

Health-Related Quality of Life

According to the sponsor, HRQoL outcome data for the reimbursement request population were not available. The key HRQoL outcomes among the PRO FAS population in the KEYNOTE-091 study (data cut-off date: January 24, 2023) at 48 weeks are summarized in Table 16. Additional HRQoL outcomes (select EORTC QLQ-C30 subscales) are summarized in Table 23.

In the PRO FAS population of the KEYNOTE-91 study, compliance rates for the EORTC QLQ-C30 and EORTC QLQ-LC13 at baseline through week 48 were high and similar between the treatment groups (98.6% at baseline and 85.8% at week 48 in the pembrolizumab group, 99.8% at baseline and 90.0% at week 48 in the placebo group for EORTC QLQ-C30 and EORTC QLQ-LC13, respectively). At week 48, the overall questionnaire completion rates were 77.9% and 84.9%, in the pembrolizumab and placebo groups, respectively.^23,39

EORTC QLQ-C30 GHS/QoL

At week 48, the majority of participants achieved an “improved” or “stable” GHS/QoL score over time in the pembrolizumab (72.4%; 95% CI, 68.6 to 76.0) and placebo (82.4%; 95% CI, 79.1 to 85.5) groups. The proportion of patients with a deteriorated score (greater than or equal to a 10-point deterioration from baseline at any time during the trial when the criteria for improved or stable is not met) in the pembrolizumab group was higher than in the placebo group (18.1% and 12.9%, respectively; difference = 5.2%; 95% CI, 1.0 to 9.4; P = 0.015) (Table 16).

The empirical mean change from baseline in GHS/QoL scores was stable over time (Figure 5). According to the sponsor, there were no clinically meaningful differences across the treatment groups at different time points, including week 48.^23,39

EORTC QLQ-LC13 Symptom Scales

At week 48, the proportion of patients with a deteriorated score in EORTC QLQ-LC13 symptom scales was similar between the 2 groups for chest pain (difference = −1.7%, 95% CI, −4.9 to 1.5; P = 0.295), coughing (difference = −0.1%, 95% CI, −3.9 to 3.6; P = 0.945), and dyspnea (difference = 3.2%, 95% CI, −1.5 to 7.8; P = 0.181) (Table 16).

Table 16: Key HRQoL Outcomes in the KEYNOTE-091 Study (PRO FAS Population at 48 Weeks)

CI = confidence interval; EORTC QLQ-C30 = European Organisation for Research and Treatment of Cancer Quality of Life Questionnaire Core 30; EORTC QLQ-LC13 = European Organisation for Research and Treatment of Cancer Quality of Life Questionnaire Lung Cancer Module; FAS = full analysis set; HRQoL = health-related quality of life; PRO = patient-reported outcome; vs. = versus.

^aDeteriorated is defined as a deterioration of 10 points or greater in the score from baseline at any time during the trial when the criteria for improved or stable is not met.

^bBased on binomial exact confidence interval method.

^cBased on unstratified Miettinen and Nurminen method.

^dTwo-sided P value for testing. H0: difference in % = 0 vs. H1: difference in % ≠ 0.

Source: Clinical Study Report: P091V02MK3475.²³

Figure 5: Empirical Mean Change From Baseline and 95% CI for the EORTC QLQ-C30 GHS/QoL Over Time by Treatment (PRO FAS Population)

CI = confidence interval; EORTC QLQ-C30 = European Organisation for Research and Treatment of Cancer Quality of Life Questionnaire Core 30; FAS = full analysis set; GHS Global Health Status; PRO = patient-reported outcome.

Source: Clinical Study Report for P091V02MK3475.²³

Harms

According to the sponsor, the harms outcome data among the reimbursement request population were not available. The harms observed among all patient population as treated in the KEYNOTE-091 study (data cut-off date of January 24, 2023) are summarized in Table 17.

Adverse Events

Adverse events were reported in 95.9% and 91.0% of patients in the pembrolizumab and placebo groups, respectively. The most common AEs in the pembrolizumab and placebo groups were increased weight (22.8% and 28.9%, respectively), pruritus (21.6% and 12.7%), and hypothyroidism (20.7% and 4.6%).

Grade 3 to 5 Adverse Events

The incidence of grade 3 to 5 AEs was higher in the pembrolizumab group (34.1%) than in the placebo group (25.8%). The most common grade 3 to 5 AEs in the pembrolizumab and placebo groups were hypertension (6.0% and 5.5%, respectively), and pneumonia (2.1% and 1.2%).

Serious Adverse Events

The incidence of SAEs was higher in the pembrolizumab group (24.5%) than in the placebo group (15.5%). The most common SAEs in the pembrolizumab and placebo groups were pneumonia (2.2% and 1.5%, respectively), and pneumonitis (2.1% and 0.7%).

Withdrawals due to Adverse Events

The incidence of AEs resulting in treatment discontinuation was higher in the pembrolizumab group (20.0%) than that in the placebo group (5.9%). The most common AEs leading to discontinuation were pneumonitis (3.6%) and diarrhea (1.2%) in the pembrolizumab group, and no AE led to treatment discontinuation in at least 1% of patients in the placebo group.

Mortality

Eleven patients (1.9%) in the pembrolizumab group and 6 patients (1.0%) in the placebo group had AEs resulting in death. Based on the study investigators’ assessment, 4 patients (0.7%) in the pembrolizumab group and 0 patients in the placebo group died due to a drug-related AEs. The reasons for these 4 deaths were cardiogenic shock and myocarditis, septic shock and myocarditis, pneumonia, and sudden death.

Notable Harms

The incidence of AEOSIs, which are considered to be medically equivalent to the immune-mediated events and infusion-related reactions, was higher in the pembrolizumab group (39.1%) compared with the placebo group (13.1%). Most AEOSIs were grade 1 or 2 in severity, nonserious and generally consistent with the known safety profile of pembrolizumab. Grade 3 to 5 AEOSIs occurred in 7.9% of pembrolizumab-treated patients compared to 1.9% in the placebo group.^23,39 The most frequently reported AEOSIs in the pembrolizumab and placebo groups were hypothyroidism (20.7% and 4.6%, respectively), hyperthyroidism (10.7% and 2.9%), and pneumonitis (6.9% and 2.9%).^23,39

Table 17: Summary of Harms Results From the KEYNOTE-091 Study (Overall APaT Population; Data Cut-Off: January 24, 2023)

APaT = all participants as treated; AE = adverse event; AEOSI = adverse event of special interest; CI = confidence interval; MedDRA = Medical Dictionary for Regulatory Activities; NCI CTCAE = National Cancer Institute Common Terminology Criteria for Adverse Events; RD = risk difference.

Note: AEs were monitored throughout the study and graded in severity according to the guideline outlined in the NCI CTCAE, Version 4.03.

^aEvery patient is counted once for each applicable specific AE.

^bNonserious AEs up to 30 days of last dose, SAEs up to 90 days of last dose and AEOSIs up to 90 days of last dose are included.

^cThe MedDRA-preferred terms “neoplasm progression,” “malignant neoplasm progression,” and “disease progression” not related to the drug are excluded.

^dA specific AE appears in this report only if its incidence in 1 or more groups meets the incidence criterion in the table section title, after rounding.

^eSAEs up to 90 days after the last dose are included.

^fFour deaths were deemed related to pembrolizumab by the investigator: 1 death was due to both cardiogenic shock and myocarditis, 1 death was due to both septic shock and myocarditis, 1 death was due to pneumonia, and 1 death was due to sudden death.

^gTreatment-emergent AEs were not reported in KEYNOTE-091. The reported drug-related AEs were determined by the investigator to be related to the drug.

^hAEOSIs are potentially immune-mediated AEs and infusion reactions based on a list of terms prepared by the sponsor and were considered regardless of attribution to trial treatment by the investigator. In addition to the specific preferred terms listed, related terms were included.

Source: Clinical Study Report: P091V02MK3475.²³

Critical Appraisal

Internal Validity

The KEYNOTE-091 study is a randomized, multicentre, parallel-group, placebo-controlled, triple-blinded, phase III trial. Randomization was performed using an appropriate methodology with adequate allocation concealment (a central interactive voice-response system). Randomization stratification was prespecified and was based on relevant prognostic factors of patients as well as factors related to practice and access to health care resources (i.e., disease stage, receipt of ACT, PD-L1 TPS, and geography).

In the sponsor’s reimbursement request for pembrolizumab as adjuvant treatment for patients with a PD-L1 of less than 50%, the survival outcomes (OS and DFS) submitted for this review were based on a subpopulation of the original study (ITT population, N = 1,177) among those with a PD-L1 of less than 50% and had received an ACT (N = 726). The review team and the clinical experts consulted for this review did not identify major issues that would affect the validity of study results with presenting DFS and OS outcomes through ad hoc analyses in the reimbursement request population, based on the fact that the PD-L1 TPS category was a stratification variable. The patient demographic and disease characteristics appeared to be generally balanced between the treatment groups in both the overall population and reimbursement request population, suggesting that the benefits of the randomization were maintained reasonably well in the subpopulation for reimbursement request. However, the review team noted that histologic status was slightly unbalanced between the 2 groups (27.8% squamous in the pembrolizumab group versus 37.5% in the placebo group). To what extent that this imbalance could bias the results is unknown.

In the reimbursement request population, a higher proportion of patients discontinued from the trial in the placebo group (33%) compared with those in the pembrolizumab group (28%), mainly due to death. A higher proportion of patients in the pembrolizumab group (48%) discontinued from the study medication compared with those in the placebo group (37%), mainly because of toxicity associated with the study medication (18% in the pembrolizumab group versus 4% in the placebo group). The clinical experts noted that the between-group imbalance and the reasons for the discontinuation from the study medication were reasonable and in line with the safety outcomes, specifically that higher proportions of patients in the pembrolizumab group experienced AEOSIs compared with those in the placebo group. However, these intercurrent events, particularly the high and uneven proportion of early discontinuation from study medication, would add challenges in an appropriate interpretation of efficacy results.⁵⁰ As such, the effect observed in this subpopulation was primarily from those patients who could have continued to use the study drug throughout the entire treatment period, introducing uncertainty in the study results.

In the reimbursement request population, nearly all of the study patients (95%) received at least 1 concomitant medication and the proportion of patients using the most medications was similar between treatment groups. The review team noted that there was a higher proportion of patients using some concomitant medications in the pembrolizumab group compared to the placebo group, which might have affected the assessment of HRQoL and biased the results in favour of pembrolizumab, as these concomitant drug uses could control or treat adverse effects associated with pembrolizumab, including “antidiarrheals, intestinal anti-inflammatory/anti-infective agents” (43% versus 26%), systemic antihistamines (28% versus 18%), systemic corticosteroids (43% versus 27%), and thyroid replacement therapy (24% versus 12%).

In the reimbursement request population, the proportion of patients receiving subsequent anticancer treatment during the trial was lower in the pembrolizumab group than in the placebo group, for both antineoplastic therapy (19% versus 28%, respectively) and immunotherapies (4% versus 15%). Although these uneven uses of anticancer therapies may have biased the efficacy results against pembrolizumab compared to placebo for OS, the extent of any important impact on interpretation of the observed effect could not be determined.

In the KEYNOTE-091 study, the triple-blind approach, which involved masking patients, investigators, and individuals who collected or analyzed the data regarding treatment allocation as prescribed in the study protocol, were appropriate. OS is considered an objective outcome and is not prone to bias due to knowledge of group assignment. DFS and intensity of AEs were assessed by the investigators blinded to treatment assignment. The HRQoL outcome assessed with EORTC QLQ-C30 and EORTC QLQ-LC13 was a patient-reported measure. Although these subjective outcomes may be influenced by knowledge of treatment assignment, the triple-blind design of the trial likely mitigated this risk. The sponsor’s use of a prespecified cut-off threshold of a 10-point or greater deterioration in score from baseline to define the proportion of patients with a deteriorated outcome for EORTC QLQ-C30 scales and EORTC QLQ-LC13 symptom scales was regarded as appropriate by the clinical experts. The risk of bias due to missing outcome data for OS, DFS, and safety outcomes appeared to be low as losses to follow-up for reasons other than death were low and sensitivity analysis with the different censoring rule for DFS in the reimbursement request population was consistent. The risk of bias from missing outcome data for the HRQoL outcome was low as only a small proportion of patients had a category of “no assessment” for the select measures (1% to 6%) in the PRO FAS population (N = 1,161) (Table 23).

Analysis of efficacy results for the reimbursement request population followed the same defined statistical plan and employed appropriate censoring criteria as in the overall population. The efficacy end points of OS and DFS were tested by applying a multiplicity hierarchical testing procedure to account for the potential inflated type I error rates across multiple end points and interim analyses. Both OS and DFS were modelled using a proportional hazards assumption (the multivariate Cox regression model with treatment adjusted by several covariates). Although the hazards assumption underlying the HRs for OS and DFS was not evaluated, in a visual inspection, the curves appeared to be relatively parallel after approximately 12 months for OS and 8 months for DFS. Of note, OS and DFS results were based on interim analyses, which may have overestimated the treatment-effect estimates.^18,19 The presence and extent of any overestimate that may have been introduced could not be determined.

External Validity

Patients in the KEYNOTE-091 study were recruited from multiple countries, including Canada. More than half of the overall or reimbursement request population were from Western Europe. In the reimbursement request population, 2 patients in the placebo group (0.6%) with stage IV disease were enrolled and included in the analyses, despite the inclusion criteria specifying that only patients with IB to IIIA stages were eligible. Problems with diagnoses led to screening failure for between 1% and 18% of patients. The clinical experts did not anticipate that this would influence the generalizability of study’s results. The clinical experts considered the eligibility criteria of patients in the KEYNOTE-091 study to be appropriate, and the demographic characteristics of the patients from the diversity aspect in the study were mostly in line with those of patients seen in clinical practice in Canada. Moreover, the clinical experts noted that, even though only patients with an ECOG PS of 0 to 1 were enrolled in the KEYNOTE-091 study as specified by the inclusion criteria, patients with an ECOG PS of 2 may benefit from pembrolizumab. The clinical experts pointed out that pembrolizumab is often offered to patients with an ECOG PS of up to 2 in clinical practice in Canada.

The clinical experts noted that presenting the survival outcomes among the subgroup of patients in the KEYNOTE-091 study who had a PD-L1 TPS of less than 50% and had received ACT was appropriate for this review, aligns with the reimbursement request, and addresses unmet therapeutic needs.

The clinical experts consulted for this review noted that the dosing and schedule of pembrolizumab in the KEYNOTE-091 study (200 mg every 3 weeks) as well as that specified in the drug’s product monograph (either 200 mg every 3 weeks or 400 mg every 6 weeks for up to 1 year or until disease recurrence or unacceptable toxicity) is in line with clinical practice in Canada.

The KEYNOTE-091 study included outcomes that were important to patients. OS, DFS, HRQoL, and safety were considered appropriate outcomes by the clinical experts and the clinician group. A study showed that, from the perspective of lung cancer survivors, DFS is a meaningful end point and addresses patients’ expectation for rapid approval of treatments that have been shown to improve DFS.⁵¹ The use of DFS as a surrogate end point for OS is accepted by the FDA for both accelerated and regular approval.⁵² The clinician groups noted that DFS is generally considered the most common surrogate survival end point for adjuvant treatments and described it as an appropriate end point for evidence in the current review.

GRADE Summary of Findings and Certainty of the Evidence

Methods for Assessing the Certainty of the Evidence

For pivotal studies and RCTs identified in the sponsor’s systematic review, GRADE was used to assess the certainty of the evidence for outcomes considered most relevant to inform CDA-AMC expert committee deliberations, and a final certainty rating was determined as outlined by the GRADE Working Group:^20,21

• High certainty: We are very confident that the true effect lies close to that of the estimate of the effect.

• Moderate certainty: We are moderately confident in the effect estimate — The true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. We use the word “likely” for evidence of moderate certainty (e.g., “X intervention likely results in Y outcome”).

• Low certainty: Our confidence in the effect estimate is limited — The true effect may be substantially different from the estimate of the effect. We use the word “may” for evidence of low certainty (e.g., “X intervention may result in Y outcome”).

• Very low certainty: We have very little confidence in the effect estimate — The true effect is likely to be substantially different from the estimate of effect. We describe evidence of very low certainty as “very uncertain.”

Following the GRADE approach, evidence from RCTs started as high-certainty evidence and could be rated down for concerns related to study limitations (which refers to internal validity or risk of bias), inconsistency across studies, indirectness, imprecision of effects, and publication bias.

When possible, certainty was rated in the context of the presence of an important (nontrivial) treatment effect; if this was not possible, certainty was rated in the context of the presence of any treatment effect (i.e., the clinical importance is unclear). In all cases, the target of the certainty assessment was based on the point estimate and where it was located relative to the threshold for a clinically important effect (when a threshold was available) or to the null. An empirically derived and validated between-group minimal important difference (MID) for OS or DFS was not identified. Based on the thresholds typically used by the CDA-AMC to assess the effects of an adjuvant treatment in patients with NSCLC of similar severity or stage, between-group differences of 5% for OS and 10% for DFS were considered clinically meaningful. Due to the lack of a formal MID estimate, the target of the certainty of evidence assessment was the presence or absence of any (non-null) effect for HRQoL and grade 3 to 5 AEs.

Results of GRADE Assessments

Table 2 presents the GRADE summary of findings for pembrolizumab versus placebo in adult patients with stage IB (T2a ≥ 4 cm), II, or IIIA NSCLC who have undergone complete resection and platinum-based chemotherapy and whose tumours have a PD-L1 TPS of less than 50%.

Long-Term Extension Studies

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

Description of Studies

The KEYNOTE-091 study is still ongoing for OS follow-up. No other long-term extension studies are ongoing or completed.³⁹

Indirect Evidence

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

Objectives for the Summary of Indirect Evidence

In the absence of direct head-to-head trials evaluating the comparative efficacy of pembrolizumab and atezolizumab in the adjuvant treatment of adult patients with early-stage NSCLC who have undergone complete resection and platinum-based chemotherapy, the sponsor conducted 1 ITC including only the subpopulation of participants with a TPS of 50% or greater. The findings from this ITC are used to support the sponsor’s reimbursement request and its request for a deviation from pharmacoeconomic requirements that excludes this subpopulation.

Indirect Treatment Comparison Design

Objectives

The objective of the sponsor-submitted ITC was to evaluate the comparative efficacy of pembrolizumab versus atezolizumab for the adjuvant treatment of stage II to IIIA NSCLC in adult patients with a TPS of 50% or greater who had received prior adjuvant chemotherapy.

Study Selection Methods

The sponsor did not conduct a systematic literature review, and included data from 2 phase III RCTs in its ITC:

• The KEYNOTE-091 study, which is a 2-arm, multicentre, international, triple-blind, placebo-controlled, randomized, phase III trial with anti–PD-1 monoclonal antibody pembrolizumab (MK-3475) versus placebo for participants with early-stage NSCLC after resection and completion of standard adjuvant therapy (PEARLS). The database cut-off date was January 24, 2023.

• The IMpower010 study, which is a randomized, multicentre, open-label, phase III trial of atezolizumab versus best supportive care after adjuvant cisplatin-based chemotherapy in participants with completely resected stage IB to IIIA NSCLC. The database cut-off date was January 21, 2021.

Indirect Treatment Comparison Analysis Methods

Indirect comparison of pembrolizumab and atezolizumab was conducted using the Bucher method.⁵³ The sponsor calculated the HR and its 95% CI as an exponent of the treatment effect and its corresponding 95% CI, and the standard error was calculated using the regular variance formula for 2 additive normal distributions.

For the KEYNOTE-091 study, the sponsor used individual participant data, and DFS survival curves were estimated using the Kaplan-Meier method. A Cox proportional hazards model was used with treatment as a covariate and the Efron method of tie handling was used to assess the magnitude of the treatment effect.

For the IMpower010 study, the sponsor used the treatment effect and the hazard ratio and its corresponding standard error (derived from the 95% CI) as extracted from the published literature.

Results of Indirect Treatment Comparison

Summary of Included Studies

A summary of the KEYNOTE-091 and IMpower010 studies is presented in Table 18. A summary of the baseline characteristics of patients in both studies with a PD-L1 TPS of 50% or greater is presented in Table 19.

For the KEYNOTE-091 study, the sponsor included only the ongoing trial patients after excluding patients who discontinued treatment. The sponsor did not report the median follow-up duration for this subpopulation; however, the median follow-up time for the ITT population of patients with a PD-L1 TPS of 50% or greater who were treated with pembrolizumab was ████ months (range = ███ ██ █████. For the IMpower010 study, from the published data, the median follow-up time for patients with a PD-L1 TPS of 50% or greater who were treated with atezolizumab was █████ months ███████ ███ ██ █████.

The sponsor did not report any assessment of homogeneity or any handling of potential effect modifiers.

Table 18: Summary of Included Randomized Controlled Trials

DFS = disease-free survival; NSCLC = non–small cell lung cancer.

Source: Sponsor’s network meta-analysis study report.⁵⁴

Table 19: Summary of Baseline Characteristics of Patients in the KEYNOTE-091 and IMpower010 Studies With a PD-L1 TPS of 50% or Greater

ALK = anaplastic lymphoma kinase; BSC = best supportive care; ECOG PS = Eastern Cooperative Oncology Group Performance Status; EGFR = epidermal growth factor receptor; SD = standard deviation; TPS = tumour proportion score.

^aOngoing trial patients after excluding patients who discontinued treatment.

Sources: Sponsor’s network meta-analysis study report⁵⁴ and CADTH Reimbursement Review Report for atezolizumab (Tecentriq).⁵⁵

Results

The ITC of pembrolizumab versus atezolizumab in patients with a TPS of 50% or greater and stage II to IIIA and prior adjuvant chemotherapy is presented in Table 20. Pembrolizumab appears to be less effective (██ █ █████ ███ ███ ████ ██ ████ when compared to placebo) than atezolizumab (██ █ █████ ███ ███ ████ ██ ████ when compared to best supportive care) on DFS in this subpopulation with indirect comparison by applying the Bucher method, which resulted in ██ █ █████ ███ ███ ████ ██ ████.

Table 20: Analysis of DFS Based on an Investigator-Assessment ITC of Pembrolizumab vs. Atezolizumab With a TPS of 50% or Greater, Stage II to IIIA NSCLC, and Prior Adjuvant Chemotherapy (ITT Population)

CI = confidence interval; DFS = disease-free survival; ITC = indirect treatment comparison; ITT = intention to treat; TPS = tumour proportion score; vs. = versus.

^aNumber of participants: ITT, participants with a TPS of 50% or greater and stage II to IIIa and prior adjuvant chemotherapy. For pembrolizumab this includes only ongoing trial patients after excluding patients who discontinued treatment.

^bBased on a Cox regression with treatment as a covariate.

^cBucher methodology using separate study results (estimate and its standard error) with a common control arm to perform an indirect comparison of the effect of pembrolizumab (KEYNOTE-091) vs. atezolizumab (IMpower010).

Source: Sponsor’s Network Meta-Analysis Study Report.⁵⁴

Critical Appraisal of Indirect Treatment Comparison

The sponsor-submitted ITC was used to support its reimbursement request and request for deviation from pharmacoeconomic requirements that excludes this subpopulation of adult patients with early-stage NSCLC who have undergone complete resection and platinum-based chemotherapy with a TPS of 50% or greater. The sponsor did not conduct an additional ITC for the reimbursement request population (those with a TPS of less than 50%); therefore, this ITC does not provide evidence regarding the comparative efficacy of pembrolizumab versus atezolizumab in the population that is the subject of this review. The sponsor did not conduct a systematic literature review for this ITC and included only the 2 pivotal trials that included this subpopulation. There appears to be a 10-month difference in median follow-up times between both trials, which may have an impact on the time to event outcomes. The clinical experts consulted for this review pointed out that the baseline patient characteristics from both trials appeared to be well matched. However, the sponsor did not report or appear to assess homogeneity between the 2 studies, and could only include published aggregate level data from the IMpower010 study. It is therefore unclear if sources of clinical or methodological heterogeneity biased the effect estimates of ITC, and the results should be interpreted with caution.

Harms outcomes and other outcomes of relevance to patients (e.g., HRQoL) were not reported.

The clinical experts noted that, while no therapy options are currently available after adjuvant chemotherapy for patients with resected stage IB to IIIA NSCLC who have a PD-L1 TPS of less than 50%, atezolizumab is now the treatment of choice for patients with a TPS of 50% or greater. They noted that the results from this ITC support both the current therapy guidelines and the sponsor’s reimbursement request.

Studies Addressing Gaps in the Systematic Review Evidence

No studies addressing gaps in the pivotal and RCT evidence were identified for this review.

Discussion

Summary of Available Evidence

This review included 1 phase III RCT and 1 ITC, both conducted by the sponsor.

The KEYNOTE-091 study is an ongoing, multicentre, triple-blind, phase III RCT investigating the efficacy and safety of pembrolizumab versus placebo as adjuvant therapy for completely resected, ECOG PS of 0 or 1, stage IB to IIIA NSCLC (any histology). Patients were randomized in a 1:1 ratio to receive 200 mg of pembrolizumab by IV infusion every 3 weeks or 400 mg every 6 weeks for up to 1 year or until disease recurrence or unacceptable toxicity, or placebo. This review focused on the population of patients enrolled in the KEYNOTE-09 trial that aligned with the sponsor’s reimbursement request (N = 726), which is for the adjuvant treatment of adult patients with stage IB (T2a ≥ 4 cm), II, or IIIA NSCLC who have undergone complete resection and received platinum-based chemotherapy and whose tumours have a PD-L1 TPS of less than 50%, as determined by a validated test. The outcomes assessed in the reimbursement review included OS, DFS, HRQoL with EORTC QLQ-C30 and EORTC QLQ-LC13, and safety.

The reimbursement request population had a median age of 64 to 65 years. The proportion of male patients (65% to 68%) was higher than that of female patients (32% to 35%). A relatively small proportion of patients had stage IB disease (11% to 12%) and more than half had stage II disease (55% to 59%). Most patients were former smokers (68% to 73%), followed by those who had never smoked (13% to 19%) and current smokers (13% to 14%). More patients had an ECOG PS of 0 (56% to 62%) compared to those with an ECOG PS of 1 (38% to 44%). Most patients had the lymph node stage of N0 (41%) or N1 (38%). More patients had a nonsquamous histology (62.5% to 72.2%) compared with a squamous histology (27.8% to 37.5%). The proportion of patients who had the EGFR mutation (5.8% to 8.5%) or ALK translocation (0.8% to 1.7%) was low.

No direct comparative evidence between pembrolizumab and atezolizumab were identified. The sponsor submitted an ITC comparing pembrolizumab and atezolizumab only in the subpopulation of patients with a TPS of 50% or greater to support its reimbursement request for patients with a TPS of less than 50%. The sponsor did not submit an ITC comparing pembrolizumab and atezolizumab in patients with a TPS of less than 50%, as atezolizumab is not indicated for this group of patients and not a comparator for the reimbursement request population for pembrolizumab.

Interpretation of Results

Efficacy

The patient groups indicated that stopping or slowing disease progression is the most important outcome. The clinical experts consulted for this review and the clinician groups noted that the key treatment goals in early-stage NSCLC following complete resection is to improve cure rates. In Canada, pembrolizumab was approved on April 19, 2023, for the adjuvant treatment of adult patients with stage IB (T2a ≥ 4 cm), II, or IIIA NSCLC who have undergone complete resection and platinum-based chemotherapy.³¹

The sponsor identified that the analysis in subgroups of the overall population of KEYNOTE-091 showed a larger effect of pembrolizumab in patients with a PD-L1 TPS of less than 50% than in those with a TPS of 50% or greater with respect to DFS and OS. According to the sponsor, the difference in treatment effect for DFS and OS between the populations is best explained by the overperformance of placebo in patients with a PD-L1 TPS of 50% or greater, who were excluded from the total ITT population in the survival analyses for reimbursement request population.³⁹ Also, a sponsor-submitted ITC study⁵⁴ reported results that the DFS benefit of pembrolizumab in patients with a PD-L1 TPS of 50% or greater is not as significant as the benefit of atezolizumab, the current standard of care for these patients.³⁹ In addition, the clinical experts and clinician groups consulted for this review noted that the current practice in Canada is neoadjuvant platinum-based chemotherapy in combination with nivolumab, or adjuvant platinum-based chemotherapy, followed by atezolizumab if patients have a PD-L1 TPS of 50% or greater; that is, no adjuvant immunotherapy options area available for patients in Canada with stage IB to IIIA NSCLC who have a PD-L1 TPS of less than 50%, creating an unmet need for this group.

Given these data and the unmet treatment need, the sponsor’s reimbursement request for this submission is for the adjuvant treatment of adult patients with stage IB (T2a ≥ 4 cm), II, or IIIA NSCLC who have undergone complete resection and platinum-based chemotherapy and whose tumours have a PD-L1 TPS of less than 50%, as determined by a validated test. The review team and the clinical experts consulted for this review did not identify any major issues with presenting survival outcomes through an ad hoc analyses in the reimbursement request population (N = 726, which accounted for approximately 62% patients of the total KEYNOTE-091 ITT population³⁹). The review team also did not identify major issues with the using placebo as the comparator group in the KEYNOTE-091 study, as it is not mandatory in clinical practice in most countries to use adjuvant chemotherapy in patients similar to those in the reimbursement request population.⁹

The clinical experts consulted for this review were unable to suggest MID thresholds for between-group differences in the proportions of patients with OS or DFS, as an MID likely depends on patients’ individual values and preferences, and therefore may vary among different patients and at different follow-up time points. The assessment of precision in the GRADE certainty of evidence was based on a threshold usually used by CDA-AMC to assess the effects of a treatment in patients with NSCLC of similar condition (i.e., 5% for OS and 10% for DFS). In the reimbursement request population, median OS was not reached in any treatment group at interim phase. The risk of death was reduced by 27% with pembrolizumab compared to placebo ████ █████ ███ ███ ████ ██ █████ ███████ | | ██████. There were uncertainties in the GRADE assessment for OS as its data were immature, based on the prespecified analysis plan by the sponsor for the KEYNOTE-091 trial.^23,40 Specifically, the median OS was not reached in either group (the observed Kaplan-Meier curves did not meet their medians), with OS events observed in 23.1% of patients in the pembrolizumab group and in 30.3% of patients in the placebo group at IA3. The median DFS at IA3 was 17 months longer in the pembrolizumab group than in the placebo group (51.7 months versus 34.5 months; HR = 0.72; 95% CI, 0.58 to 0.89; nominal P < 0.001), with a 28% lower risk of disease recurrence, other malignancy, or death. Evidence with a moderate level of certainty showed that pembrolizumab likely results in a clinically important increase in the probability of DFS at both 24 months and 48 months.

Two subgroup analyses in the reimbursement request population — patient smoking status and EGFR mutation at baseline — are relevant to the pharmacoeconomic analysis of this review. For OS, interpretation of subgroup results (current smoker versus former smoker versus nonsmoker; and EGFR mutation status, yes versus no) is limited, because of immature data for OS. For DFS, the review team noted a numerically larger effect of pembrolizumab among the patients who were current smokers (HR = 0.51; 95% CI, 0.27 to 0.97) compared to former smokers (HR = 0.80; 95% CI, 0.62 to 1.02) as well as nonsmokers (HR = 0.71; 95% CI, 0.44 to 1.16), for the point estimate of the HR, seemingly suggesting the effect of pembrolizumab in DFS increases with the presence and recency of smoking (current smokers > former smokers > nonsmokers). However, no conclusion could be drawn based on this subgroup results from the KEYNOTE-091 study data as the 95% CIs around the HR values were relatively wide because of small subgroup sample sizes. In addition, the interaction P value was 0.549, indicating no statistically significant difference among the categories for different smoking status (Table 22). Pembrolizumab showed a larger effect in DFS among patients with an EGFR mutation (event rates of 55% versus 84% in the pembrolizumab and placebo groups, respectively; HR = 0.44; 95% CI, 0.23 to 0.84) than in those without an EGFR mutation (event rates of 51% versus 62%; HR = 0.76; 95% CI, 0.56 to 1.04) in terms of the differences in the point estimate and 95% CI of HRs. However, the interaction P value was 0.334, indicating no statistically significant difference among the 3 categories of EGFR mutation status (Table 22). The review team noted that a possible limitation of the KEYNOTE-091 study could be that EGFR testing was not required.⁴⁰ As a result, more than half of patients had “unknown” information for EGFR mutation status at baseline (Table 12). Although an EGFR mutation could be an important prognostic factor in the reimbursement request population, our ability to interpret the consequence of this assumption is limited because of the small sample size. Moreover, all the subgroup results in the reimbursement request population were based on ad hoc analyses, making interpretation of any subgroup effects findings uncertain.

The patient and clinician groups also regarded HRQoL as an outcome of importance. Both the EORTC QLQ-C30 and EORTC QLQ-LC13 are validated tools commonly used to measure HRQoL among patients with NSCLC. The clinical experts consulted for this review acknowledged that the sponsor’s use of a 10-point cutoff threshold in change from baseline for these 0-to-100 measurements in defining categorical outcomes (improved, stable, deteriorated, and unconfirmed) was appropriate. Of these categories, the clinical experts indicated the proportion of patients who had a deteriorated outcome in HRQoL was important. The missing data for HRQoL outcomes important to this review was low as only a small proportion of patients (1% to 6%) had no assessment in the PRO FAS population at week 48. The between-group difference in the EORTC QLQ-C30 GHS/QoL change was not regarded as clinically meaningful. The 95% CIs of risk differences in the proportion of patients with a deteriorated outcome in EORTC QLQ-LC13 symptoms of chest pain, coughing, and dyspnea all included the null threshold of 0. These outcome data, together with the empirical mean change from baseline data of the EORTC QLQ-C30 GHS/QoL (Figure 5), suggest that both the patients who received pembrolizumab and those who received placebo had comparable and relatively stable HRQoL over 48 weeks.

In the ITC of pembrolizumab versus atezolizumab in patients with a TPS of 50% or greater and stage II to IIIA and prior adjuvant chemotherapy, pembrolizumab appears to be less effective (██ █ ████ ████ ███ ████ ██ █████ when compared to placebo) than atezolizumab (HR = ████ ████ ███ ████ ██ █████ when compared to best supportive care) with respect to DFS in this subpopulation, with an indirect comparison by the Bucher method resulting in a HR of ████ ████ ███ ████ ██ ████). However, the sponsor did not report or appear to assess homogeneity between the 2 included studies and could only include published aggregate level data from study on atezolizumab. It is therefore unclear if the sources of clinical or methodological heterogeneity biased the effect estimates of the ITC. The sponsor did not submit an ITC comparing pembrolizumab and atezolizumab in patients with a TPS of less than 50%, which is the population of patients included in the sponsor’s reimbursement request.

Harms

In the overall APaT population of the KEYNOTE-091 study, the incidence of harms was higher in the pembrolizumab group than in the placebo group for grade 3 to 5 AEs (34% versus 26%, respectively), SAEs (25% versus 16%), and AEs resulting in treatment discontinuation (20% versus 6%). The clinical experts consulted for this review noted that, in a placebo-controlled study, it is expected that patients in the pembrolizumab group would experience AEs more frequently than would patients in the placebo group. The most common AEOSIs in the pembrolizumab group, including hypothyroidism (21%) and hyperthyroidism (11%), were potentially immune-mediated AEs that may be associated with immune checkpoint inhibitors such as pembrolizumab.⁵⁶ The clinical experts noted that, overall, the safety outcomes in the KEYNOTE-091 study were consistent with their expectations for pembrolizumab. They added that the difference in incidence of grade 3 to 5 AEOSIs between the pembrolizumab group (7.9%) and the placebo group (1.9%) was small, indicating that the tolerability of pembrolizumab is acceptable. The clinical experts also noted that the incidence of AEs important to patients or AEs associated with immune checkpoint inhibitor therapy,⁵⁷ including thyroiditis (1.0% versus 0.2%), myocarditis (0.9% versus 0.2%), and type 1 diabetes mellitus (0.2% versus 0), was low.

Harms outcomes were not reported in the sponsor-submitted ITC.

Conclusion

One triple-blind, phase III RCT comparing the efficacy and safety of adjuvant pembrolizumab and placebo in adult patients with stage IB to IIIA NSCLC who have undergone complete resection and platinum-based chemotherapy and whose tumours have a PD-L1 TPS of less than 50% showed a clinically meaningful benefit of adjuvant pembrolizumab in the probability of DFS at 24 and 48 months. The OS results are uncertain due to the immaturity of the data (the median OS was not reached in either group, with OS events observed in 23% and 30% of patients in the pembrolizumab and placebo groups, respectively), although there is a trend toward improved OS in favour of pembrolizumab. Patients who received either pembrolizumab or placebo had comparable and relatively stable HRQoL over 48 weeks. According to the clinical experts consulted for this review, the safety profile of pembrolizumab was consistent with their expectations for this drug.

The sponsor submitted an ITC comparing pembrolizumab and atezolizumab in patients with a TPS of 50% or greater to support its reimbursement request for patients with a TPS of less than 50%. Indirect comparative evidence suggests that atezolizumab is superior in patients with a TPS of 50% or greater; however, it is unclear if sources of clinical or methodological heterogeneity biased the effect estimates of the ITC, because no assessment of homogeneity was conducted. The sponsor did not submit an ITC comparing pembrolizumab and atezolizumab in patients with a TPS of less than 50% because, according to clinician input, atezolizumab is not indicated for this subpopulation of patients.

References

1.Canadian Cancer Statistics Advisory Committee. Canadian Cancer Statistics: A 2020 special report on lung cancer. Toronto (ON): Canadian Cancer Society; 2020: https://cdn.cancer.ca/-/media/files/cancer-information/resources/publications/2020-canadian-cancer-statistics-special-report/2020-canadian-cancer-statistics-special-report-en.pdf. Accessed 2024 Aug 15.

2.Canadian Cancer Statistics Advisory in collaboration with the Canadian Cancer Society, Statistics Canada, and the Public Health Agency of Canada. Canadian Cancer Statistics 2023. Toronto (ON): Canadian Cancer Society; 2023: https://cdn.cancer.ca/-/media/files/research/cancer-statistics/2023-statistics/2023_PDF_EN.pdf. Accessed 2024 Aug 15.

3.Cancer-Specific Stats 2022: Snapshot of incidence, mortality and survival estimates by cancer type. Toronto (ON): Canadian Cancer Society; 2022: https://cdn.cancer.ca/-/media/files/research/cancer-statistics/2022-statistics/2022_cancer-specific-stats.pdf. Accessed 2024 Aug 15.

4.Ehrenstein V, Eriksen K, Taylor A, Servidio L, Jakobsen E. Characteristics and overall survival of patients with early-stage non–small cell lung cancer: A cohort study in Denmark. Cancer Med. 2023;12(1):30-37. PubMed

5.Molina JR, Yang P, Cassivi SD, Schild SE, Adjei AA. Non-small cell lung cancer: Epidemiology, risk factors, treatment, and survivorship. Mayo Clin Proc. 2008;83(5):584-594. PubMed

6.Muthusamy B, Patil PD, Pennell NA. Perioperative systemic therapy for resectable non–small cell lung cancer. J Natl Compr Canc Netw. 2022;20(8):953-961. PubMed

7.Forde PM, Spicer J, Lu S, et al. Neoadjuvant nivolumab plus chemotherapy in resectable lung cancer. N Engl J Med. 2022;386(21):1973-1985. PubMed

8.Pignon JP, Tribodet H, Scagliotti GV, et al. Lung adjuvant cisplatin evaluation: A pooled analysis by the LACE Collaborative Group. J Clin Oncol. 2008;26(21):3552-3559. PubMed

9.O'Brien M, Paz-Ares L, Marreaud S, et al. Pembrolizumab versus placebo as adjuvant therapy for completely resected stage IB-IIIA non–small-cell lung cancer (PEARLS/KEYNOTE-091): An interim analysis of a randomised, triple-blind, phase III trial. Lancet Oncol. 2022;23(10):1274-1286. PubMed

10.CADTH Reimbursement Recommendation: Atezolizumab (Tecentriq) [accessed by sponsor]. Can J Health Technol. 2022;2(9). https://www.cda-amc.ca/sites/default/files/DRR/2022/PC0269%20Tecentriq%20for%20NSCLC%20-%20CADTH%20Final%20Recommendation-Final-meta.pdf.

11.CADTH Reimbursement Review Provisional Funding Algorithm: Non-small cell lung cancer without actionable oncogenic alterations. Ottawa (ON): CADTH; 2023: https://www.cadth.ca/sites/default/files/DRR/2023/PH0029-Algorithm_secured.pdf. Accessed 2024 Mar 12.

12.Felip E, Altorki N, Zhou C, et al. Adjuvant atezolizumab after adjuvant chemotherapy in resected stage IB-IIIA non–small-cell lung cancer (IMpower010): A randomised, multicentre, open-label, phase III trial. Lancet. 2021;398(10308):1344-1357. PubMed

13.Tankel J, Spicer J, Chu Q, et al. Canadian consensus recommendations for the management of operable stage II/III non–small-cell lung cancer: Results of a modified Delphi process. Current Oncology. 2023;30(12):10363-10384. PubMed

14.CADTH Reimbursement Recommendation: Osimertinib (Tagrisso) [accessed by sponsor]. Can J Health Technol. 2022;2(1). https://www.cadth.ca/osimertinib.

15.CADTH Reimbursement Recommendation: Nivolumab (Opdivo). Can J Health Technol. 2023;3(4). https://www.cadth.ca/nivolumab-2. Accessed 2024 Mar 12.

16.Edge SB, Compton CC. The American Joint Committee on Cancer: The 7th edition of the AJCC cancer staging manual and the future of TNM. Ann Surg Oncol. 2010;17(6):1471-1474.

17.Eisenhauer EA, Therasse P, Bogaerts J, et al. New response evaluation criteria in solid tumours: Revised RECIST guideline (version 1.1). Eur J Cancer. 2009;45(2):228-247. PubMed

18.Bassler D, Briel M, Montori VM, et al. Stopping randomized trials early for benefit and estimation of treatment effects: systematic review and meta-regression analysis. JAMA. 2010;303(12):1180-1187. PubMed

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

20.Balshem H, Helfand M, Schunemann HJ, et al. GRADE guidelines: 3. Rating the quality of evidence. J Clin Epidemiol. 2011;64(4):401-406. PubMed

21.Santesso N, Glenton C, Dahm P, et al. GRADE guidelines 26: Informative statements to communicate the findings of systematic reviews of interventions. J Clin Epidemiol. 2020;119:126-135. PubMed

22.BARDS Health Technology Assessment Statistical Report. Pembrolizumab for Adjuvant Treatment of Early Stage Non-Small Cell Lung Cancer After Resection and Completion of Standard Adjuvant Therapy (PEARLS). Protocol 091 IA3 Canada. Baseline Characteristics and Efficacy [internal sponsor's report]. In: Keytruda (pembrolizumab), solution for infusion 100 mg/4 mL vial. Rahway (NJ): BARDS, Merck & Co., Inc.; 2024.

23.Clinical Study Report: P091V02MK3475. A randomized, phase III trial with anti PD-1 monoclonal antibody pembrolizumab (MK 3475) versus placebo for patients with early-stage NSCLC after resection and completion of standard adjuvant therapy (PEARLS) KEYNOTE 091 [internal sponsor's report]. Rahway (US): Merck Sharp & Dohme LLC; 2023 Jul 3.

24.Merck Canada Inc. response to August 21, 2024 CADTH request for additional information regarding Keytruda (pembrolizumab) CADTH review [internal additional sponsor's information]. Kirkland (QC): Merck Canada Inc.; 2024 Aug 23.

25.Brenner DR, Gillis J, Demers AA, et al. Projected estimates of cancer in Canada in 2024. CMAJ. 2024;196(18):E615-E623. PubMed

26.Canadian Cancer Society. What is lung cancer? https://cancer.ca/en/cancer-information/cancer-types/lung/what-is-lung-cancer. Accessed 2024 Aug 15.

27.Birring SS, Peake MD. Symptoms and the early diagnosis of lung cancer. Thorax. 2005;60(4):268-269. PubMed

28.BC Cancer. Lung. 2013; http://www.bccancer.bc.ca/health-professionals/clinical-resources/cancer-management-manual/lung/lung. Accessed 2024 Aug 15.

29.Tecentriq (atezolizumab for injection), concentrate for solution for infusion, 60 mg/mL, 1200 mg/20 mL and 840 mg/14 mL single use vials [product monograph]. Mississauga (ON): Hoffmann-La Roche Limited; 2024 Mar 13: https://pdf.hres.ca/dpd_pm/00074930.PDF. Accessed 2024 Aug 15.

30.Opdivo (nivolumab for injection), intravenous infusion, 10 mg/mL, 40 mg and 100 mg single use vials [product monograph]. . Saint-Laurent (QC): Bristol-Myers Squibb Canada; 2024 Jun 28: https://pdf.hres.ca/dpd_pm/00076152.PDF. Accessed 2024 Aug 22.

31.Keytruda (pembrolizumab), solution for infusion 100 mg/4 mL vial [product monograph]. Kirkland (QC): Merck Canada Inc.; 2024 Jun 20.

32.Hirsch FR, McElhinny A, Stanforth D, et al. PD-L1 immunohistochemistry assays for lung cancer: Results from Phase 1 of the Blueprint PD-L1 IHC Assay Comparison Project. J Thorac Oncol. 2017;12(2):208-222. PubMed

33.Tsao MS, Kerr KM, Kockx M, et al. PD-L1 immunohistochemistry comparability study in real-life clinical samples: Results of Blueprint Phase 2 project. J Thorac Oncol. 2018;13(9):1302-1311. PubMed

34.Institut national d'excellence en santé et en services sociaux. Test compagnon de Tecentriq: Cancer du poumom non à petites cellules. Avis transmis au ministre en janvier 2023. Montreal (QC): INESSS; 2023: https://www.inesss.qc.ca/fileadmin/doc/INESSS/Inscription_medicaments/Avis_au_ministre/Mars_2023/TC_Tecentriq_CPNPC_2023_01.pdf. Accessed 2024 Aug 6.

35.Lantuejoul S, Sound-Tsao M, Cooper WA, et al. PD-L1 testing for lung cancer in 2019: Perspective from the IASLC Pathology Committee. J Thorac Oncol. 2020;15(4):499-519. PubMed

36.National Comprehensive Cancer Network. NCCN Clinical Practice Guidelines in Oncology: Non-small cell lung cancer. v6.2024. 2024; https://www.nccn.org/professionals/physician_gls/pdf/nscl.pdf. Accessed 2024 Jun 25.

37.Canadian Cancer Statistics Advisory Committee. Canadian Cancer Statistics. A 2020 special report on lung cancer. Toronto (ON): Canadian Cancer Society; 2020: https://cdn.cancer.ca/-/media/files/research/cancer-statistics/2020-statistics/canadian-cancer-statistics/2020-special-report/res-cancerstatistics-canadiancancerstatistics-2020_special-report_en.pdf. Accessed 2024 Jun 12.

38.Ionescu DN, Downes MR, Christofides A, Tsao MS. Harmonization of PD-L1 testing in oncology: A Canadian pathology perspective. Curr Oncol. 2018;25(3):e209-e216. PubMed

39.Keytruda (pembrolizumab) for patients with early-stage NSCLC after resection and completion of standard adjuvant therapy: Clinical summary [internal sponsor's report]. In: Drug Reimbursement Review sponsor submission: Keytruda (pembrolizumab), solution for infusion 100 mg/4 mL vial. Kirkland (QC): Merck Canada Inc.; 2024.

40.Intergroup Study (EORTC-1416-LCG). A randomized, phase III trial with anti-PD-1 monoclonal antibody pembrolizumab (MK-3475) versus placebo for patinets with early stage NSCLC after resection and completion of standard adjuvant therapy (PEARLS) KEYNOTE-091 - Study protocol v5.0 [internal sponsor's report]. Brussels (BE): EORTC Lung Cancer Group, Merck & Co., Inc.; 2022 Sep 23.

41.Aaronson NK, Ahmedzai S, Bergman B, et al. The European Organisation for Research and Treatment of Cancer QLQ-C30: A quality-of-life instrument for use in international clinical trials in oncology. J Natl Cancer Inst. 1993;85(5):365-376. PubMed

42.Teckle P, Peacock S, McTaggart-Cowan H, et al. The ability of cancer-specific and generic preference-based instruments to discriminate across clinical and self-reported measures of cancer severities. Health Qual Life Outcomes. 2011;9:106. PubMed

43.Nicklasson M, Bergman B. Validity, reliability and clinical relevance of EORTC QLQ-C30 and LC13 in patients with chest malignancies in a palliative setting. Qual Life Res. 2007;16(6):1019-1028. PubMed

44.Ozturk A, Sarihan S, Ercan I, Karadag M. Evaluating quality of life and pulmonary function of long-term survivors of non–small cell lung cancer treated with radical or postoperative radiotherapy. Am J Clin Oncol. 2009;32(1):65-72. PubMed

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

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

47.Bergman B, Aaronson NK, Ahmedzai S, Kaasa S, Sullivan M. The EORTC QLQ-LC13: A modular supplement to the EORTC Core Quality of Life Questionnaire (QLQ-C30) for use in lung cancer clinical trials. EORTC Study Group on Quality of Life. Eur J Cancer. 1994;30A(5):635-642. PubMed

48.Merck Canada Inc. response to September 10, 2024 CADTH request for additional information regarding Keytruda (pembrolizumab) CADTH review [internal additional sponsor's information]. Kirkland (QC): Merck Canada Inc.; 2024 Sept 12.

49.Merck Canada Inc. response to August 30, 2024 CADTH request for additional information regarding Keytruda (pembrolizumab) CADTH review [internal additional sponsor's information]. Kirkland (QC): Merck Canada Inc.; 2024 Sept 5.

50.ICH E9 (R1) addendum on estimands and sensitivity analysis in clinical trials to the guideline on statistical principles for clinical trials. EMA/CHMP/ICH/436221/2017. Amsterdam (NL): European Medicines Agency (EMA); 2020 Feb 17: https://www.ema.europa.eu/en/documents/scientific-guideline/ich-e9-r1-addendum-estimands-and-sensitivity-analysis-clinical-trials-guideline-statistical-principles-clinical-trials-step-5_en.pdf. Accessed 2024 Sep 20.

51.Bever A, Manthorne J, Rahim T, Moumin L, Johnston K, Szabo S. The importance of the disease-free survival (DFS) endpoint to survivors of lung cancer. Eur Respir J. 2021;58(suppl 65):PA2190.

52.Clinical trial endpoints for the approval of cancer drugs and biologics: Guidance for industry. Silver Spring (MD): U.S. Food and Drug Administration; 2018: https://www.fda.gov/media/71195/download. Accessed 2024 Aug 23.

53.Bucher HC, Guyatt GH, Griffith LE, Walter SD. The results of direct and indirect treatment comparisons in meta-analysis of randomized controlled trials. J Clin Epidemiol. 1997;50(6):683-691. PubMed

54.BARDS Health Technology Assessment Statistical Report. Compound: Pembrolizumab (MK-3475). Indication: Adjuvant Treatment of Early Stage Non-Small Cell Lung Cancer after Resection and Completion of Standard Adjuvant Therapy (PEARLS). Protocol 091 IA3. Indirect Treatment Comparison: DFS [internal sponsor's report]. In: Keytruda (pembrolizumab), solution for infusion 100 mg/4 mL vial. Rahway (NJ): BARDS, Merck & Co., Inc.; 2024.

55.CADTH Reimbursement Review: Atezolizumab (Tecentriq). Can J Health Technol. 2022;2(12). https://www.cda-amc.ca/sites/default/files/DRR/2022/PC0269-Tecentriq%20NSCLC_combined.pdf. Accessed 2024 Sep 20.

56.Postow MA, Sidlow R, Hellmann MD. Immune-related adverse events associated with immune checkpoint blockade. N Engl J Med. 2018;378(2):158-168. PubMed

57.Rubio-Infante N, Ramirez-Flores YA, Castillo EC, Lozano O, Garcia-Rivas G, Torre-Amione G. Cardiotoxicity associated with immune checkpoint inhibitor therapy: A meta-analysis. Eur J Heart Fail. 2021;23(10):1739-1747. PubMed

Appendix 1: Detailed Outcomes Data

Please note that this appendix has not been copy-edited.

Additional data of the KEYNOTE-091 study are presented in the following.

Table 21: Subgroup Analysis Results of Overall Survival in the KEYNOTE-091 Study (Reimbursement Request Population) [Redacted]

███ █ ████████ █████████████ ████ █ ████████ █████ █████████ ██ ███████ ███ █ ██████████ ████████ ███████ ████ █ ███████ ███████████ ████████ ██████ ████ █ █████████ ██████ ██████ █████████ ███ █ ███████████████████ ██ █ ███ █████████ █████ █ █████████ ████ ████ ███████ ████ █ ██████████ ████ █████ ███████ ██ ██████ ██████████ ████ ████████████ ██ ██ █ ████████ ██████████ ███ █ █████ ██████████ ████████████ ████ █████████ ██ ████ █████ ████ █████ ██ ████████ ██ ████████ ███████ ████ ██ ███████ ███ █████ ██ ███ ███ ██████████ ██ ████████ ████ █████ ███ █ ███ ███ █████ █████|████ █████████████ ██████████████ ██████ ███ ████████ ██████|███ █████ █████████ ████████ ███ █████ ██ ████████████ ███ ██████████ █████ ████ ██████████ ████████ ██ ███ █████████ ███████████ █████ ███ ███ ██ ███ ██████ █████ ██████ ██ ███ ███ █████ ███ █ ████ ██████ ████████ ██████ ███ ███████ ██████ ███ ████ ██ █████ ███ ██████ █████████ █████████ ███ ██████████████ ███ ███████ ██████ ██████ ███ ████████████████ █████ ████ ██████████ █████████ ███ █████ ██████████ ████████ ███ █████ ██ ███ ██████████ █████ ████ █████████ ██ █ █████████ █████ ████ ██████████ ██████████|███ █████ █████████ ████████ ███ █████ ██ ████████████ ███ ██████████ █████ ████ █████████ ████████ ██ ███ █████████ ███████████ █████ ███ ███ ██ ███ ██████ █████ ██████ ██ ███ ███ █████ ███ █ ████ ██████ ████████ ██████ ███ ███████ ██████ ███ ████ ██ █████ ███ ██████ █████████ █████████ ███ ██████████████ ███████ ██████ ██████ ███ ████████████████ ███ █████████████████████ ███████████ ██ █████ ██ ██████████ █████ ████ ███ ███████████ ██████ ███ █████ ██████████ ████████ ███ █████ ██ ███ ██████████ █████ ████ █████████ ███ ████████ ██ ███████████ ███ █████████████████████ ███████████ ██ █████ ██ ██████████ █████ ████ ███ ███████████ ██████████████ █████████ █████████████

Table 22: Subgroup Analysis Results of Disease-Free Survival in the KEYNOTE-091 Study (Reimbursement Request Population)

ACT = adjuvant chemotherapy; AJCC = American Joint Committee on Cancer; CI = confidence interval; DFS = disease-free survival; ECOG = Eastern Cooperative Oncology Group; EGFR = epidermal growth factor receptor; ITT = intention-to-treat; NR = not reported; NSCLC = non–small cell lung cancer; TPS = tumour proportion score.

Note: Data presented in this table were based on analyses at clinical cut-off date of January 24, 2023, in the ITT population of patients with a PD-L1 TPS of less than 50% and prior ACT.

^aFrom product-limit (Kaplan-Meier) method for censored data.

^bFor PD-L1 subgroup, analysis was based on multivariate Cox regression model with treatment, adjusted by the following covariates: stage (IB vs. II vs. IIIA), PD-L1 status (≥ 50% vs. 1% to 49% vs. < 1%), ACT (yes vs. no), region (Western Europe vs. Eastern Europe vs. Rest of World vs. Asia), histology (squamous vs. non-squamous), and smoking status (never vs. former/current), using Wald confidence interval. For other subgroups, analysis was based on Cox regression model with treatment as a covariate using Wald confidence interval.

^cFor PD-L1 subgroup, analysis was based on multivariate Cox regression model with treatment adjusted by the following covariates: stage (IB vs. II vs. IIIA), PD-L1 status (≥ 50% vs. 1% to 49% vs. < 1%), ACT (yes vs. no), region (Western Europe vs. Eastern Europe vs. Rest of World vs. Asia), histology (squamous vs. non-squamous), smoking status (never vs. former/current), and treatment-by-subgroup interaction (P value of likelihood ratio test for interaction term). For other subgroups, analysis was based on Cox regression model with treatment and subgroup as covariates, and treatment-by-subgroup interaction (P value of likelihood ratio test for interaction term).

Source: Sponsor’s submission.²⁴

Table 23: Additional HRQoL Outcomes in the KEYNOTE-091 Study (PRO FAS Population)

CI = confidence interval; EORTC = European Organisation for Research and Treatment of Cancer; FAS = full analysis set; PRO = patient-reported outcomes; QLQ-C30 = Quality of Life Questionnaire-Core 30; QLQ-LC13 = Quality of Life Questionnaire Lung Cancer module; QoL = quality of life; vs. = versus.

Note: Categories of the outcome definitions are as follows:

Improved is defined as a 10-point or more increase in score (in the positive direction) at any time during the trial and confirmed by a 10-point or more increase in score at the next consecutive visit.

Stable is defined as a 10-point or more increase (in the positive direction) or less than 10-point change in score (in the positive or negative direction) from baseline and confirmed by a less than 10-point change in score at the next consecutive visit; OR a less than 10-point change in score and a 10-point or more increase in score at the next consecutive visit.

Improved + stable is defined as the composite of the improved and stable.

Deteriorated is defined as a ≥ 10-point deterioration in score from baseline at any time during the trial when the criteria for improved or stable is not met.

Unconfirmed is defined as when the criteria for improved or stable with confirmation or deterioration is not met.

No assessment is defined as participants who do not have baseline or post-baseline assessments available.

^aBased on binomial exact confidence interval method.

^bBased on unstratified Miettinen and Nurminen method

^c2-sided P value for testing. H0: difference in % = 0 vs. H1: difference in % ≠ 0.

Source: Clinical Study Report for P091V02MK3475.²³

EORTC QLQ-C30

The least squares (LS) mean change from baseline in global health status/QOL score did not show a 10-point change in either treatment group with little to no separation between the treatment groups at Week 48, indicating stable scores with no clinically meaningful differences across the treatment groups. The LS mean change from baseline in functioning scores were generally stable from baseline to Week 48 in both treatment groups (Figure 6).^23,39

Figure 6: Summary LS Mean Change From Baseline to Week 48 and 95% Confidence Interval in EORTC QLQ-C30 Functional Scales/Global Health Status/QoL (PRO FAS Population)

EORTC = European Organisation for Research and Treatment of Cancer; FAS = full analysis set; LS = least squares; PRO = patient-reported outcomes; QLQ-C30 = Quality of Life Questionnaire Core 30; QoL = quality of life.

Source: Clinical Study Report: P091V02MK3475.²³

Pharmacoeconomic Review

Abbreviations

Executive Summary

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

Table 1: Submitted for Review