CADTH Reimbursement Recommendation

Pembrolizumab (Keytruda)

Indication: In combination with trastuzumab, fluoropyrimidine- and platinum-containing chemotherapy for the first-line treatment of adult patients with locally advanced unresectable or metastatic HER2-positive gastric or GEJ adenocarcinoma, whose tumours express PD-L1 (CPS ≥ 1) as determined by a validated test

Sponsor: Merck Canada Inc.

Final recommendation: Reimburse with conditions

Recommendation

What Is the CADTH Reimbursement Recommendation for Keytruda?

CADTH recommends that Keytruda should be reimbursed by public drug plans for the treatment of locally advanced unresectable or metastatic human epidermal growth factor receptor 2 (HER2) positive gastric or gastroesophageal junction (GEJ) adenocarcinoma whose tumours express programmed cell death-ligand 1 (PD-L1) if certain conditions are met.

Which Patients Are Eligible for Coverage?

Keytruda should only be covered to treat adult patients who have not received previous treatment for advanced or metastatic esophageal or GEJ cancer and have good performance status.

What Are the Conditions for Reimbursement?

Keytruda should only be reimbursed if prescribed in combination with trastuzumab, fluoropyrimidine, and platinum-containing chemotherapy by a clinician with expertise and experience in treating gastric and GEJ cancer.

Why Did CADTH Make This Recommendation?

• Evidence from 1 clinical trial demonstrated that patients treated with Keytruda, when added to trastuzumab and fluoropyrimidine- and platinum-containing chemotherapy, resulted in improved survival and could delay cancer progression.

• Keytruda meets patient needs of delaying disease progression and prolonging survival and was unlikely to worsen health-related quality of life (HRQoL).

• Based on CADTH’s assessment of the health economic evidence, Keytruda does not represent good value to the health care system at the public list price. A price reduction is therefore required.

• Based on public list prices, Keytruda, in combination with trastuzumab, fluoropyrimidine, and platinum-containing chemotherapy, is estimated to cost the public drug plans approximately $38.1 million over the next 3 years.

Additional Information

What Is Gastric or GEJ Cancer?

Gastric and GEJ cancers occur in the stomach, where the esophagus and stomach join, respectively. Most gastric and GEJ cancers are adenocarcinomas. The cancer is considered locally advanced if it spreads in the stomach or GEJ and metastatic if it spreads to another part of the body. The 5-year survival rate for patients diagnosed with gastric and GEJ cancer living in Canada is 29%. For patients with metastatic gastric or GEJ cancer, the 5-year survival rate is 6.6%.

Unmet Needs in Gastric or GEJ Cancer

Patients diagnosed with HER2-positive gastric or GEJ cancer have more aggressive disease, and many patients do not respond to available treatment options. Even in patients who do respond to treatment, their survival is quite limited.

How Much Does Keytruda Cost?

Treatment with Keytruda is expected to cost approximately $11,733 per patient per cycle when used in combination with trastuzumab and fluoropyrimidine- and platinum-containing chemotherapy.

Recommendation

The pCODR Expert Review Committee (pERC) recommends that pembrolizumab, in combination with trastuzumab, fluoropyrimidine- and platinum-containing chemotherapy, be reimbursed for adult patients with locally advanced unresectable or metastatic human HER2-positive gastric or GEJ adenocarcinoma, whose tumours express programmed cell death-ligand 1 (PD-L1) (combined positive score [CPS] ≥ 1) as determined by a validated test, only if the conditions listed in Table 1 are met.

Rationale for the Recommendation

Evidence from 1 phase III double-blind randomized controlled trial (KEYNOTE-811; N = 698) demonstrated that pembrolizumab, when added to standard of care (SOC) therapy with trastuzumab and platinum-fluoropyrimidine doublet chemotherapy may result in added clinical benefit in patients with locally advanced unresectable or metastatic HER2-positive gastric or GEJ adenocarcinomas, whose tumours express PD-L1 (CPS ≥ 1). The KEYNOTE-811 trial demonstrated that in patients whose tumours express PD-L1 (N = 594) treatment with pembrolizumab plus SOC therapy was associated with statistically significant improvements in progression-free survival (PFS; hazard ratio [HR] = 0.71; 95% confidence interval [CI], 0.59 to 0.86, P = 0.0002) and overall survival (OS) (HR = 0.81; 95% CI, 0.67 to 0.98; P = 0.0142), when compared with placebo plus SOC therapy. The median OS was 20.0 months (95% CI, 17.9 to 22.7 months) in the pembrolizumab plus SOC group and 15.7 months (95% CI, 13.5 to 18.5 months) in the placebo plus SOC group. pERC agreed with the clinical experts that the net improvement in median OS of 4.3 months may be considered clinically meaningful in this patient population, given the poor prognosis of the disease in patients diagnosed at a locally advanced or metastatic stage. Immunotherapy-mediated adverse events were more frequent in the pembrolizumab group; however, pERC considered the safety profile of pembrolizumab in combination with SOC therapy to be manageable and consistent with the known safety profile of pembrolizumab. Conclusions on HRQoL outcomes could not be drawn due to the exploratory nature of these outcomes in the trial, the absence of minimally important difference (MID) estimates in patients with gastric or GEJ cancer and high proportions of missing data. However, the trial results suggested that HRQoL was not worse in the pembrolizumab plus SOC therapy group when compared to the placebo plus SOC group.

Patients identified a need for more effective and accessible treatments that prolong survival, reduce the risk of disease progression, improve quality of life, allow for more convenient therapy administration, and minimize side effects. pERC noted that pembrolizumab when added to SOC therapy with trastuzumab and platinum-fluoropyrimidine doublet chemotherapy, met some of the needs identified by patients because it may prolong survival, delay disease progression, result in little or no deterioration in HRQoL, and have a manageable safety profile.

Using the sponsor-submitted price for pembrolizumab and publicly listed prices for all other drug costs, the incremental cost-effectiveness ratio (ICER) for pembrolizumab plus trastuzumab and chemotherapy was $425,549 per quality-adjusted life-year (QALY) gained compared with trastuzumab and chemotherapy alone using a fixed-dose regimen for pembrolizumab ($297,169 per QALY gained using weight-based dosing). Using either a fixed-dose or weight-based regimen, pembrolizumab is not cost-effective at a $50,000 per QALY willingness to pay (WTP) threshold for adults with locally advanced unresectable or metastatic HER2 positive gastric or GEJ adenocarcinoma whose tumours express PD-L1. A price reduction for pembrolizumab is required for pembrolizumab plus trastuzumab and chemotherapy to be considered cost-effective at a $50,000 per QALY threshold.

Table 1: Reimbursement Conditions and Reasons

5-FU = 5 fluorouracil; CAPOX = capecitabine plus oxaliplatin; CNS = central nervous system; CPS = combined positive score; ECOG = Eastern Cooperative Oncology Group; HER2 = human epidermal growth factor receptor 2; PD-L1 = programmed cell death-ligand-1; QALY = quality-adjusted life-year.

Discussion Points

• Aligned with the input from the patient and clinician group, pERC acknowledged that there is an unmet need for effective and safe therapy options in this patient population, as patients who are diagnosed with advanced-stage disease have poor prognoses. pERC additionally noted that HER2 overexpression in patients with gastric or GEJ adenocarcinomas may be associated with poorer outcomes and more aggressive disease.

• pERC deliberated on the results of the KEYNOTE-811 trial. It noted that the trial’s dual primary end points (i.e., OS and PFS) were statistically significant in favour of adding pembrolizumab to trastuzumab plus chemotherapy. pERC discussed that, over a median of 38 months of follow-up, treatment with pembrolizumab plus SOC resulted in 3.6 months and 4.3 months incremental improvements in median PFS and OS, respectively, over standard treatment alone. pERC discussed that the clinical importance of the observed improvement in PFS was unclear and that there was uncertainty about the validity of PFS as a surrogate outcome in predicting the long-term treatment effect on OS. However, pERC agreed with the clinical experts that the net improvement in median OS of 4.3 months would be considered clinically meaningful in this patient population with poor prognosis. pERC further noted that the observed difference in OS benefit at 12, 18, and 36 months (point estimates) met the expert-identified threshold for clinical meaningfulness (i.e., 10% to 15% at any time point) in favour of pembrolizumab plus SOC. However, the lower bounds of the corresponding 95% confidence intervals were compatible with little-to-no clinically important difference. Accordingly, the Grading of Recommendations, Assessment, Development, and Evaluations (GRADE) assessment of the evidence suggested with moderate certainty that adding pembrolizumab to SOC likely results in a clinically important increase in the probability of survival plus SOC at 12, 18, and 36 months compared with placebo.

• pERC noted that the KEYNOTE-811 trial enrolled patients regardless of their tumour PD-L1 expression status. However, the sponsor’s decision to limit the funding request for pembrolizumab plus SOC to patients who have PD-L1-positive disease was based on a subgroup analysis. Although the subgroup analyses based on PD-L1 expression status were prespecified, they were absent from the statistical testing hierarchy. pERC agreed that, while it presented a risk of inflated type I error (i.e., falsely rejecting the null hypothesis), the subgroup of patients with PD-L1 positive tumours represented approximately 85% of the full study population. pERC noted that the results from the full study population (N = 698) and those from the PD-L1 positive subgroup (N = 594) were consistent and that the clinical benefit observed in the full study population appeared to be driven by the PD-L1 positive subgroup.

• pERC noted that both HER2 and PD-L1 testing are required to implement a reimbursement recommendation for pembrolizumab in the patient population under review. According to the documentation submitted by the sponsor, HER2 testing is funded across all jurisdictions in Canada. It is used for all advanced or metastatic gastric or GEJ tumours across all testing laboratories. However, pERC discussed that some jurisdictions may currently not have validated PD-L1 testing in place and suggested that those jurisdictions may need to consider operationalizing and funding PD-L1 CPS testing to identify patients eligible for pembrolizumab treatment.

• The pharmacoeconomic analysis, which primarily informs pERC’s economic rationale for the recommendation, considered pembrolizumab as a fixed dose, per the product monograph. pERC discussed the results of the scenario analysis conducted by CADTH, where pembrolizumab was assumed to be administered using a weight-based dose. In this analysis, the ICER decreased to $297,169 per QALY gained; a price reduction of 85% would be required for pembrolizumab to achieve an ICER of $50,000 per QALY gained in such a scenario. pERC noted that this analysis is associated with additional uncertainty given that the treatment was not evaluated as a weight-based dose, as this may influence efficacy and adverse events (AEs).

Background

Gastric cancer is a growth of abnormal cells that starts in the stomach. In 2023, an estimated 4,100 Canadians were projected to be diagnosed with gastric cancer. Gastric cancers are generally classified into 2 topographical subsites. Cardia gastric cancers include the upper part of the stomach adjoining the esophagus. Noncardia gastric cancer occurs in the more distal regions of the stomach. GEJ cancer develops in the area where the esophagus meets the gastric cardia. The risk for developing gastric and GEJ cancer increases with age, is greatest after 50 years of age and occurs more frequently among men than women. Approximately 90% of noncardia cancers are attributable to Helicobacter pylori infection. Early-stage gastric and GEJ cancer are potentially curable. However, most patients present with symptoms that are usually nonspecific. As a result, early diagnosis of gastric and GEJ cancers is challenging. Instead, most patients have advanced-stage III or stage IV disease at the time of diagnosis when curative treatments are not possible. Patients with unresectable advanced or metastatic disease typically experience high symptom burden, impaired quality of life (QoL), and frequent bouts of anxiety and depression. The 5-year survival rate for patients diagnosed with gastric and GEJ cancer living in Canada is 29%, reflecting that most patients are diagnosed with an advanced-stage disease that is associated with poor prognosis. Among those with metastatic gastric or GEJ cancer, the 5-year survival rate is 6.6%.

Approximately 90% to 95% of gastric and GEJ cancers are histologically classified as adenocarcinoma. Gastric cancers may contain oncogenic driver mutations that lead to uncontrolled cell growth and proliferation. The most common driver mutation is HER2. HER2 is overexpressed or amplified in 25% to 32% of patients with GEJ and between 9.5% and 18% of patients with gastric cancers. HER2 overexpression in patients with gastric cancer is associated with poor outcomes and more aggressive disease. In clinical practice, laboratory tests for HER2 status and PD-L1 expression are done on a biopsy sample taken from the primary tumour or metastases. HER2 status can be determined via immunohistochemistry (IHC), which measures the amount of HER2 protein in the cancer cells, or via fluorescence in situ hybridization (FISH), which examines the number of copies of the HER2 gene in the cancer cells. PD-L1 expression was determined using a semiquantitative approach through IHC.

In patients with HER2+ disease, the addition of trastuzumab to the standard first-line platinum-fluoropyrimidine doublet is recommended for all patients based on the phase III ToGA trial, which demonstrated improvements in response rates, PFS, and OS with trastuzumab compared with chemotherapy alone. This regimen is supported by the National Institute for Health and Care Excellence (NICE), the European Society for Medical Oncology (ESMO), the National Comprehensive Cancer Network, Cancer Care Ontario (CCO), and Alberta Health Services. In October 2023, ESMO recommended adding pembrolizumab to the trastuzumab plus chemotherapy SOC for patients with positive PD-L1 expression defined by a CPS or 1 or more based on the results of the KEYNOTE-811 clinical trial.

Pembrolizumab is a high-affinity antibody against programmed cell death protein 1 (PD-1), which exerts dual ligand blockade of the PD-1 pathway, including PD-L1 and programmed cell death-ligand 2 (PD-L2), on antigen-presenting or tumour cells. Pembrolizumab reactivates tumour-specific cytotoxic T lymphocytes in the tumour microenvironment by inhibiting the PD-1 receptors from blinding to its ligands. Pembrolizumab received a notice of compliance on February 6, 2024, through the standard review pathway. The Health Canada indication for pembrolizumab, in combination with trastuzumab, fluoropyrimidine- and platinum-containing chemotherapy, is for the first-line treatment of adult patients with locally advanced unresectable or metastatic HER2 positive gastric or GEJ adenocarcinoma, whose tumours express PD-L1 (CPS ≥ 1) as determined by a validated test.

Pembrolizumab has been approved by Health Canada, in combination with trastuzumab, fluoropyrimidine- and platinum-containing chemotherapy, for adult patients with locally advanced unresectable or metastatic HER2 positive gastric or GEJ adenocarcinoma, whose tumour express PD-L1 (CPS ≥ 1) as determined by a validated test. Pembrolizumab is available as a solution for infusion, and the dosage recommended in the product monograph is 200 mg every 3 weeks or 400 mg every 6 weeks by IV infusion.

Sources of Information Used by the Committee

To make its recommendation, the committee considered the following information:

• a review of 1 phase III randomized controlled trial (RCT) clinical study in adult patients with HER2-positive advanced gastric or GEJ adenocarcinoma

• patient perspectives gathered by patient group, My Gut Feeling – Stomach Cancer Foundation of Canada

• input from public drug plans and cancer agencies that participate in the CADTH review process

• 2 clinical specialists with expertise in diagnosing and treating patients with gastric or GEJ cancers

• input from 2 clinician groups, including the Canadian Gastrointestinal Oncology Evidence Network (CGOEN) and Ontario Health – Cancer Care Ontario Gastrointestinal Drug Advisory Committee (OH-CCO GI DAC)

• a review of the pharmacoeconomic model and report submitted by the sponsor.

Stakeholder Perspectives

Patient Input

Patient group input was submitted by 1 patient advocacy group, My Gut Feeling – Stomach Cancer Foundation of Canada, and included input collected from an international online survey conducted between November 10 and November 24, 2023. The survey included responses from 40 patients (77.5%) and caregivers (22.5%). Of those who responded, 72.5% were from Canada, and 15.5% were HER2 positive. All patients who responded to the survey experienced at least 1 symptom before diagnosis, with the most common being weight loss (57.5%), reflux (55%), change in appetite (50%), pain (47.5%), nausea/vomiting (37.5%), and difficulty swallowing (25%). Most patients (95%) reported that their cancer diagnosis had a significant impact on their QoL, physical and mental health, ability to eat and work, finances, social life, identity, and personal image. Psychosocial impacts such as anxiety, depression, sleep loss, feeling crippled, anticipatory grief, and loss of control were cited by 1 patient. Caregivers and family members who responded to the survey also reported being impacted by the cancer diagnosis, which included feeling hopeless (especially with metastatic disease), stress from the impact of chemotherapy-induced side effects causing stress on other family members, and changes to family dynamics that require counselling for children. Other disease- or treatment-related concerns reported by both the patients and caregivers included loss of fertility, feeling isolated, financial difficulty, as well as financial and geographical barriers to accessing treatment, health care providers, and information. All patients who completed the survey experienced at least 1 side effect. The most commonly reported treatment-related side effects included fatigue (87.5%), appetite changes (77.5%), alopecia and taste changes (75% each), weight loss and neuropathy (70% each). Approximately 16% of patients reported discontinuing treatment due to an adverse event resulting in hospitalization. Patients and caregivers who completed the survey indicated that the following outcomes were important in considering new treatments: improved survival, remission, shrinking of cancer, improved symptoms, treatment tolerability, and improved QoL. Patients and caregivers also added that equitable access, convenience of administration (e.g., oral versus IV, less frequent travel to hospital, shorter chair time to receive treatment), and more options for choosing based on their values and preferences were important. Finally, survey responders from Canada emphasized that biomarker testing should be accessible across all centres and provinces at the onset of their disease.

Clinician Input

The information in this section summarizes input provided by the patient and clinician groups who responded to CADTH’s call for input and from clinical experts consulted by CADTH for this review.

Input From Clinical Experts Consulted by CADTH

The clinical experts consulted by CADTH for this review emphasized that locally advanced and metastatic HER2-positive gastroesophageal cancer is a disease associated with a considerable unmet need. The clinical experts advised that although treatment with trastuzumab combined with chemotherapy is available for locally advanced metastatic HER2-positive gastroesophageal cancer, OS outcomes remain unacceptably poor. Both clinical experts suggested that as per the KEYNOTE-811 clinical trial, pembrolizumab would be added to the current SOC first-line therapy (trastuzumab combined with platinum doublet chemotherapy) for patients with locally advanced and metastatic HER2 positive gastroesophageal cancer. This combination – pembrolizumab combined with trastuzumab and platinum doublet chemotherapy – would represent a new first-line SOC treatment for this patient population. Although patients were eligible to enrol in the KEYNOTE-811 trial regardless of PD-L1 status as measured by CPS, the prespecified subgroup analysis showed that the benefit of adding pembrolizumab to SOC was attributable to the subgroup of patients with PD-L1 CPS of 1 or more (85% of the study population). A clear benefit was not observed in the subgroup of patients with PD-L1 CPS of less than 1, which included a small number of patients. The clinical experts consulted by CADTH opined that the addition of pembrolizumab to first-line treatment for locally advanced and metastatic HER2-positive gastroesophageal cancer should be limited to patients with a PD-L1 CPS of 1 or more. As suggested by the clinical experts, CPS testing should be performed using a validated test. The clinical experts listed 3 factors, in descending order of clinical importance, used to determine response to treatment: patient-reported symptoms and side effects; cross-sectional imaging via CT scans or MRI; and tumour markers such as carcinoembryonic antigen (CEA) and carbohydrate antigen 19 to 9 (Ca 19 to 9). The clinical experts emphasized; however, that the only truly clinically meaningful end points across all oncology types are OS and QoL. The clinical experts added that all other end points (e.g., response rate, CEA response, PFS) should be considered surrogates and are of little relevance if they do not predict improved OS or QoL. The clinical experts suggest that a clinician should assess patients after every 2 to 3 cycles of treatment. The clinical experts indicated that patients should undergo CT scans every 2 to 3 months and that tumour markers should be assessed at least once every 4 weeks. The clinical experts suggested that the decision to discontinue treatment with pembrolizumab should be based on patient-reported symptoms, side effects and well-being, in combination with an assessment of treatment response and either radiologic or clinical disease progression. The clinical experts added that treatment with pembrolizumab should be discontinued in the event of life-threatening immune-related AEs by clinical practice guidelines. Clinical experts suggested that pembrolizumab should only be prescribed by or under the supervision of a specialist in medical oncology with expertise in the management of immunotherapy side effects. The clinical experts noted that immunotherapy and trastuzumab are currently delivered as SOC in all oncology centres and may be safely administered in all centres approved for oncology care.

Clinician Group Input

Clinician group input was submitted by 2 clinician groups – CGOEN and OH-CCO GI DAC. Input provided by the CGOEN and the OH-CCO GI DAC collated insights from 8 and 2 clinicians, respectively. The clinical groups noted that there are currently limited treatment options for patients with HER2-positive gastric or GEJ cancers, with poor outcomes. The clinician from CGOEN indicated that the treatment of HER2-positive gastric cancer has not improved in more than a decade and that immunotherapy is currently only available for patients who are HER2-negative. Based on the OH-CCO GI DAC input, prolonging OS is the main treatment goal for this patient population. According to input from the CGOEN, patients best suited for treatment with pembrolizumab are those with a PD-L1 CPS of 1 or more, as determined by a validated test. Based on input from the CGOEN, response to treatment should be based on routine imaging (during timed intervals for objective assessment), as well as patient preference, tolerability, and QoL. Both clinician groups suggested that patients should be evaluated regularly for clinical response and toxicity per current treatment standards. Both clinician groups agreed that the decision to continue or discontinue treatment with pembrolizumab should be based on patient preference, side effects (including life-threatening immune-related AEs), radiologic or clinical disease progression or treatment response, and patient-reported symptoms and well-being. While input from the CGOEN suggested that said pembrolizumab should be administered in oncology centres, the clinical experts consulted by CADTH noted that pembrolizumab could be safely administered in a hospital or an outpatient clinic.

Drug Program Input

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

Table 2: Responses to Questions From the Drug Programs

5-FU = Fluorouracil; CAPOX = Capecitabine + oxaliplatin; CPS = combined positive score; ECOG = Eastern Cooperative Oncology Group; FISH = Fluorescence in situ hybridization; FOLFOX = Fluorouracil + leucovorin + oxaliplatin; GEJ = Gastroesophageal junction; HER2 = Human epidermal growth factor receptor 2; IHC = Immunohistochemistry; pERC = pan-Canadian Oncology Drug Review expert review committee; PD-L1 = Programmed cell death-ligand 1; PS = performance status.

Clinical Evidence

Systematic Review

Description of Studies

One study was included in the sponsor-conducted systematic review: KEYNOTE-811.

KEYNOTE-811 is an ongoing multicentre (92 sites across 19 countries), placebo-controlled, randomized (1:1 ratio), double-blind, phase III study evaluating the efficacy and safety of adding pembrolizumab (200 mg every 3 weeks) to SOC therapy with trastuzumab and platinum-fluoropyrimidine doublet chemotherapy as first-line therapy for HER2-positive advanced gastric or GEJ cancer in adult patients. Patients were randomly allocated to receive either pembrolizumab (full study population n = 350; PD-L1 CPS ≥ 1 subgroup n = 298) or placebo (full study population n = 348; PD-L1 CPS ≥ 1 subgroup n = 296) each in combination with SOC therapy (trastuzumab in combination with cisplatin/5-FU [FP] or capecitabine/oxaliplatin [CAPOX]). Randomization was stratified by geographic region (Australia, Europe, Israel, and North America versus Asia versus rest of the world), investigator’s choice of chemotherapy regimen (FP versus CAPOX), and PD-L1 expression at baseline (CPS ≥ 1 versus < 1). HER2 status and PD-L1 expression were determined by FDA-approved assays, and were conducted at a central laboratory. KEYNOTE-811 assessed PFS per RECIST 1.1 assessed by blinded independent central review (BICR) and OS as dual primary efficacy end points, in which the study success is claimed if statistically significant analysis results are demonstrated for at least 1 of the 2 primary end points. Secondary end points included overall response rate and duration of response per RECIST 1.1, and harms. Exploratory end points included the following HRQoL measures: the European Organization for Research and Treatment of Cancer Quality of Life Questionnaire-Core 30 (EORTC QLQ-C30), European Organization for Research and Treatment of Cancer Quality of Life Questionnaire-STO22 (EORTC QLQSTO22), and the EQ-5D-5L.

KEYNOTE-811 met the protocol-defined criterion of success (1-sided alpha-level testing 0.0013) at the second interim analysis (IA2; data cut-off date May 25, 2022) in which the stratified HR for PFS was 0.72 (95% CI, 0.60 to 0.87; P = 0.0002) in favour of pembrolizumab plus SOC (median, 10.0 months; 95% CI, 8.6 to 11.7 months) versus placebo plus SOC (median, 8.1 months; 95% CI, 7.0 to 8.5 months). Prespecified subgroup analysis noted that the treatment effect of pembrolizumab plus SOC on PFS compared to placebo plus SOC was attributable to the PD-L1 CPS of 1 or more than 1 subgroups, who made up 85.1% of the total population. A clear benefit was not observed in the subgroup of patients with PD-L1 CPS of less than 1, which included relatively few patients. Among patients with PD-L1 CPS of 1 or more (hereafter referred to as the subgroup of patients who have PD-L1 positive disease), PFS was statistically more prolonged in the pembrolizumab plus SOC group than in placebo plus SOC group (HR, 0.70; 95% CI, 0.58 to 0.85). Based on analyses conducted at IA2, the sponsor proposed that the indication population be limited to patients who had the subgroup of patients who have PD-L1 positive disease. Accordingly, this CADTH review of KEYNOTE-811 will present data from the total study population and the subgroup of patients with PD-L1 positive disease as per the Health Canada indication.

The mean age of all patients enrolled in KEYNOTE-811 was 60.4 years (SD = 11.8 years) and 61.7 years (SD = 10.8) in the pembrolizumab plus SOC and placebo plus SOC group, respectively. PD-L1 positive disease was documented in 85.1% of patients in both the pembrolizumab plus SOC and placebo plus SOC groups. Among patients in KEYNOTE-811 who were PD-L1 positive, the mean age of patients randomized to the pembrolizumab plus SOC and placebo plus SOC group were 60.6 (SD = not reported [NR]) and 61.4 (SD = NR) years, respectively. In terms of disease characteristics of the study participants with PD-L1 positive disease, 32.6% of patients in the pembrolizumab plus SOC group presented with adenocarcinoma of the GEJ and 67.4% presented with adenocarcinoma of the stomach; in the placebo plus SOC group, 33.4% and 66.6% of patients presented with adenocarcinoma of the GEJ and stomach, respectively.

Efficacy Results

Overall Survival

Full Study Population

The median duration of follow-up in the total study population at IA2 (data cut-off May 25, 2022) was 16.1 months (range, 0.6 to 41.6 months) and 14.8 months (range, 0.3 to 41.2 months) in the pembrolizumab plus SOC and placebo plus SOC group, respectively. The median duration of follow-up at the time of the third interim analysis (IA3; data cut-off March 29, 2023) was NR for the total study population.

In KEYNOTE-811, the proportion of observed death at the time of IA3 (March 29, 2023) was 70.0% and 73.6% in the pembrolizumab plus SOC and placebo plus SOC groups, respectively. The median OS was 20.0 months (95% CI, 17.8 to 22.1 months) in the pembrolizumab plus SOC group and 16.8 months (95% CI, 15.0 to 18.7 months) in the placebo plus SOC group. The stratified HR for OS was 0.84 (95% CI, 0.70 to 1.01; P = 0.0292) following treatment with pembrolizumab plus SOC versus placebo plus SOC. Risk differences in OS in the total study population following treatment with pembrolizumab plus SOC when compared to placebo plus SOC at months 12, 18 and 36 were |||| |||| ||| |||| || ||||||| |||| |||| ||| ||| || ||||||| ||| |||| |||| ||| |||| || ||||||| respectively.

PD-L1 Positive Subgroup

The median duration of follow-up in the PD-L1 positive subgroup at IA2 was 17.0 months (range, 0.6 to 41.6) and 13.9 months (range, 0.3 to 41.2 months) in the pembrolizumab plus SOC and placebo plus SOC group, respectively. The median duration of follow-up at the time of IA3 was NR for the PD-L1 positive subgroup.

Among patients in the PD-L1 positive subgroup, the proportion of observed death at the time of IA3 (March 29, 2023) was 68.5% and 73.6% in the pembrolizumab plus SOC and placebo plus SOC groups, respectively. The median OS was 20.0 months (95% CI, 17.9 to 22.7 months) in the pembrolizumab plus SOC group and 15.7 months (95% CI, 13.5 to 18.5 months) in the placebo plus SOC group. The HR for OS was 0.81 (95% CI, 0.67 to 0.98; P = 0.0142) in favour of treatment with pembrolizumab plus SOC versus placebo plus SOC. Risk difference in OS in the PD-L1 positive subgroup following treatment with pembrolizumab plus SOC when compared to placebo plus SOC at months 12, 18 and 36 were |||| |||| ||| ||| || ||||||| ||||| |||| || ||||||| ||| |||| |||| ||| |||| || ||||||| respectively.

Progression-Free Survival

Full Study Population

In KEYNOTE-811, the protocol-defined criterion of success was met at IA2, with 80% of the total events expected for the analysis (information fraction) having accrued (data cut-off date May 25, 2022). The stratified HR for PFS was 0.72 (95% CI, 0.60 to 0.87; P = 0.0002; 1-sided superiority boundary was P = 0.0013) in favour of pembrolizumab plus SOC. The stratified HR for PFS based on BICR assessment from sensitivity analyses 1, 2 and 3 using alternative censoring rules were 0.74 (95% CI, 0.62 to 0.88; P = 0.0003), 0.74 (95% CI, 0.62 to 0.87; P = 0.0001) and 0.73 (95% CI, 0.61 to 0.87; P = 0.0003), respectively.

Disease progression or death on or before the IA3 data cut-off date (March 29, 2023) was observed in 72.3% of patients in the pembrolizumab plus SOC group and 75.0% of patients in the placebo plus SOC group. The median PFS in the pembrolizumab plus SOC and placebo plus SOC group was 10.0 months (95% CI, 8.6 to 12.2 months) and 8.1 months (95% CI, 7.1 to 8.6 months), respectively. The stratified HR for disease progression or death was 0.73 (95% CI, 0.61 to 0.87, P = 0.0002) in favour of pembrolizumab plus SOC versus placebo plus SOC. Risk differences in PFS in the total study population following treatment with pembrolizumab plus SOC when compared to placebo plus SOC at months 12, 18 and 36 were ||||| |||| ||| ||| || ||||||| |||| |||| ||| ||| || ||||||| ||| |||| |||| ||| ||| || ||||||| respectively.

PD-L1 Positive Subgroup

Among patients in the PD-L1 positive subgroup, disease progression or death on or before the data cut-off date (March 29, 2023) was observed in 72.8% of patients in the pembrolizumab plus SOC group and 76.0% of patients in the placebo plus SOC group. The median PFS in the pembrolizumab plus SOC and placebo plus SOC group was 10.9 months (95% CI, 8.5 to 12.5 months) and 7.3 months (95% CI, 6.8 to 8.5 months), respectively. The HR for disease progression or death was 0.71 (95% CI, 0.59 to 0.86, P = 0.0002) in favour of pembrolizumab plus SOC versus placebo plus SOC. Risk differences in PFS in the PD-L1 positive subgroup following treatment with pembrolizumab plus SOC when compared to placebo plus SOC at months 12, 18, and 36 were ||||| |||| ||| ||| || ||||||| |||| |||| ||| ||| || |||||| ||| |||| |||| ||| ||| || ||||||| respectively.

Health-Related Quality of Life

EORTC QLQ-C30

EORTC QLQ-C30 is a cancer-specific HRQoL tool consisting of 30 items to assess 5 functional dimensions (physical function, role function, emotional function, cognitive function, and social function), 3 symptoms items (fatigue, nausea or vomiting, and pain), 5 single-item measures assessing additional symptoms commonly experienced by patients with cancer (dyspnea, loss of appetite, insomnia, constipation and diarrhea) and 1 scale assessing global health status and global QoL. Based on input from the clinical experts consulted by CADTH, global health, physical functioning, and appetite loss were scale items most relevant to patients with gastroesophageal cancers. Scores for each scale and item ranged from 0 to 100, with higher scores indicative of greater QoL, greater physical functioning or a greater degree of symptoms. Improvement and deterioration were defined as changes of 10 or more points in the relevant direction.

Full Study Population

In KEYNOTE-811, analysis of the EORTC QLQ-C30 in the total study population was based on IA2 (data cut-off date May 25, 2022). Overall, baseline EORTC QLQ-C30 was completed by 320 (92.8%) patients in the pembrolizumab plus SOC group and 339 (99.7%) patients in the placebo plus SOC group. By week 24, 231 (67.0%) of the available 265 (76.8%) patients in the pembrolizumab plus SOC group completed the questionnaire for a compliance rate of 87.2%. In the placebo plus SOC group, 190 (55.9%) of the available 235 (69.1%) patients completed the questionnaire for a compliance rate of 80.9%.

In the full study population, the between-group difference in the least square mean change from baseline to week 24 for global health status was ||||| |||||| |||| ||| ||||| || |||| |||||||||||||| following treatment with pembrolizumab plus SOC versus placebo plus SOC. Improvement in global health status was reported in 31.6% and 31.8% of patients in the pembrolizumab plus SOC and standard care group, respectively. The between-group difference in improvement was ||||| |||| ||| ||||| || ||||| following treatment with pembrolizumab plus SOC versus placebo plus SOC. Improvement or stability in global health status was reported in 71.9% and 71.5% of patients in the pembrolizumab plus SOC and placebo plus SOC groups, respectively. The between-group difference in improvement or stability was |||| |||| ||| ||||| || ||||| following treatment with pembrolizumab plus SOC compared to placebo plus SOC. The stratified HR for time to deterioration on the global health status scale at 12 months was 1.14 (95% CI, 0.84 to 1.55; P = 0.3951) for pembrolizumab plus SOC relative to placebo plus SOC.

For physical function, the between-group difference in the least square change from baseline to week 24 was ||||| |||||| |||| ||| ||||| || |||| |||||||||||||| following treatment with pembrolizumab plus SOC versus placebo plus SOC. Improvement in physical function was reported in 14.8% and 15.9% of patients in the pembrolizumab plus SOC and standard care group, respectively. The between-group difference in improvement was ||||| |||| ||| ||||| || ||||| following treatment with pembrolizumab plus SOC versus placebo plus SOC. Improvement or stability in physical function was reported in 73.0% and 72.4% of patients in the pembrolizumab plus SOC and placebo plus SOC groups, respectively. The between-group difference in improvement or stability was |||| |||| ||| |||| || ||||| following treatment with pembrolizumab plus SOC compared to placebo plus SOC. The stratified HR for time to deterioration on the physical function scale at 12 months was 1.05 (95% CI, 0.78 to 1.47; P = 0.7663) for pembrolizumab plus SOC relative to placebo plus SOC.

For the single-item appetite loss, the between-group difference in least square change from baseline to week 24 was |||| |||| ||| ||||| || |||| ||||||||. Improvement in appetite loss was reported in 32.5% and 26.6% of patients in the pembrolizumab plus SOC and standard care group, respectively. The between-group difference in improvement was |||| |||| ||| ||||| || |||||| following treatment with pembrolizumab plus SOC versus placebo plus SOC. Improvement or stability in appetite loss was reported in 77.4% and 72.6% of patients in the pembrolizumab plus SOC and placebo plus SOC groups, respectively. The between-group difference in improvement or stability was |||| |||| ||| |||| || |||||| following treatment with pembrolizumab plus SOC compared to placebo plus SOC. The stratified HR for time to deterioration on the single-item appetite loss at 12 months was 1.18 (95% CI, 0.87 to 1.60; P = 0.2898) for pembrolizumab plus SOC relative to placebo plus SOC.

PD-L1 Positive Subgroup

Among patients in the PD-L1 positive subgroup, analysis of the EORTC QLQ-C30 was based on IA3 (data cut-off date March 29, 2023). Baseline EORTC QLQ-C30 was completed by 272 (93.5%) patients in the pembrolizumab plus SOC group and 274 (95.8%) patients in the SOC group. The number of patients available to complete the measure diminished over time. By week 24, 223 (76.6%) patients were available in the pembrolizumab plus SOC group; of those, 196 (67.0%) patients completed the questionnaire for a compliance rate of 87.4%. In the placebo plus SOC group, 151 (52.8%) of the available 192 (67.1%) patients completed the questionnaire for a compliance rate of 78.6%.

In the PD-L1 positive subgroup, the between-group difference in least square mean change from baseline to week 24 for global health status was ||||| |||| ||| ||||| || |||| |||||||| following treatment with pembrolizumab plus SOC versus placebo plus SOC. Improvement in global health status was reported in 31.6% and 32.5% of patients in the pembrolizumab plus SOC and standard care group, respectively. The between-group difference in improvement was ||||| |||| ||| ||||| || ||| ||||||||| following treatment with pembrolizumab plus SOC versus placebo plus SOC. Improvement or stability in global health status was reported in 71.5% and 71.0% of patients in the pembrolizumab plus SOC and placebo plus SOC groups, respectively. The between-group difference in improvement or stability was |||| |||| ||| ||||| || || ||||||||| following treatment with pembrolizumab plus SOC compared to placebo plus SOC. The HR for time to deterioration on the global health status scale at 12 months was 1.16 (95% CI, 0.83 to 1.61; P = 0.3756) for pembrolizumab plus SOC relative to placebo plus SOC.

For physical functioning, the between-group difference in least square change from baseline to week 24 was ||||| |||| ||| ||||| || ||||| |||||||||||||| following treatment with pembrolizumab plus SOC versus placebo plus SOC. Improvement in physical functioning was reported in 15.1% and 17.5% of patients in the pembrolizumab plus SOC and standard care group, respectively. The between-group difference in improvement was ||||| |||| ||| ||||| || ||| ||||||||| following treatment with pembrolizumab plus SOC versus placebo plus SOC. Improvement or stability in physical functioning was reported in 74.9% and 71.7% of patients in the pembrolizumab plus SOC and placebo plus SOC groups, respectively. The between-group difference in improvement or stability was |||| |||| ||| ||||| || ||| |||||||||| following treatment with pembrolizumab plus SOC compared to placebo plus SOC. The HR for time to deterioration on the physical functioning scale at 12 months was 0.99 (95% CI, 0.72 to 1.38; P = 0.9615) for pembrolizumab plus SOC relative to placebo plus SOC.

For the single-item appetite loss, the between-group difference in least square change from baseline to week 24 was NR. Improvement in appetite loss was reported in 32.6% and 28.3% of patients in the pembrolizumab plus SOC and standard care group, respectively. The between-group difference in improvement was |||| |||| ||| ||||| || |||| ||||||||| following treatment with pembrolizumab plus SOC versus placebo plus SOC. Improvement or stability in appetite loss was reported in 78.0% and 72.4% of patients in the pembrolizumab plus SOC and placebo plus SOC groups, respectively. The between-group difference in improvement or stability was |||| |||| ||| ||||| || |||| ||||||||| following treatment with pembrolizumab plus SOC compared to placebo plus SOC. The HR for time to deterioration on the single-item appetite loss at 12 months was 1.23 (95% CI, 0.88 to 1.70; P = 0.2344) for pembrolizumab plus SOC relative to placebo plus SOC.

EORTC QLQ-ST022

EORTC QLQ-ST022 is an HRQoL measure specific to gastric cancer, which consists of 22 items to assess symptoms of dysphagia (4 items), pain or discomfort (3 items), upper GI symptoms (3 items), eating restrictions (5 items), emotional (3 items), dry mouth, hair loss and body image (1 item each). Scores for each symptom scale range from 0 to 100, with higher scores indicative of worsening symptoms. Improvement and deterioration were defined as a decrease or increase of 10 or more points. Results from the EORTC QLQ-ST022 were included in the clinical report as supportive analyses.

Full Study Population

In KEYNOTE-811, analysis of the EORTC QLQ-ST022 was based on IA2 (data cut-off date May 25, 2022). Overall, baseline EORTC QLQ-ST022 was completed by 319 (92.5%) patients in the pembrolizumab plus SOC group and 320 (94.1%) patients in the placebo plus SOC group. The number of patients available to complete the measure diminished over time. By week 24, 229 (66.4%) of the available 265 (76.8%) patients in the pembrolizumab plus SOC group completed the questionnaire for a compliance rate of 86.4%. In the placebo plus SOC group, 190 (55.9%) of the available 235 (69.1%) patients completed the questionnaire for a compliance rate of 80.9%.

In the full study population, the between-group difference in least square mean change from baseline to week 24 on the pain symptom scale of the EORTC QLQ-ST022 was ||||| |||||| |||| ||| ||||| || |||| ||||| ||||||||| following treatment with pembrolizumab plus SOC versus placebo plus SOC. Improvement in pain symptoms was reported in 40.0% and 32.1% of patients in the pembrolizumab plus SOC and standard care group, respectively. The between-group difference in improvement was |||| |||| ||| |||| || ||| |||||||||| favouring treatment with pembrolizumab plus SOC versus placebo plus SOC. Improvement or stability in pain was reported in 82.0% and 78.2% of patients in the pembrolizumab plus SOC and placebo plus SOC groups, respectively. The between-group difference in improvement or stability was |||| |||| ||| |||| || |||| |||||||| following treatment with pembrolizumab plus SOC compared to placebo plus SOC. Deterioration on the pain symptom scale was recorded in 11.3% and 10.6% of patients in the pembrolizumab plus SOC and placebo plus SOC groups, respectively. The stratified HR for time to deterioration on the pain symptom scale at 12 months was 0.99 (95% CI, 0.62 to 1.58; P = 0.9681) for pembrolizumab plus SOC relative to placebo plus SOC.

PD-L1 Positive Subgroup

Among patients in the PD-L1 positive subgroup, analysis of the EORTC QLQ-ST022 was based on IA3 (data cut-off date March 29, 2023). Baseline EORT QLQ-ST022 was completed by 271 (93.1%) patients in the pembrolizumab plus SOC group and 273 (95.5%) patients in the placebo plus SOC group. The number of patients available to complete the measure diminished over time. By week 24, 193 (66.3%) of the available 223 (76.6%) patients in the pembrolizumab plus SOC group completed the questionnaire for a compliance rate of 86.5%. In the placebo plus SOC group, 152 (79.2%) of the available 192 (67.1%) patients completed the questionnaire for a compliance rate of 79.2%.

In the PD-L1 positive subgroup, the between-group difference in least square mean change from baseline to week 24 for the pain symptom scale of the EORTC QLQ-ST022 was NR. Improvement in pain symptoms was reported in 40.2% and 32.9% of patients in the pembrolizumab plus SOC and standard care group, respectively. The between-group difference in improvement was |||| |||| ||| ||||| || ||||| |||||||| following treatment with pembrolizumab plus SOC versus placebo plus SOC. Improvement or stability in pain was reported in 83.2% and 78.3% of patients in the pembrolizumab plus SOC and placebo plus SOC groups, respectively. The between-group difference in improvement or stability was |||| |||| ||| |||| || ||||| |||||||| following treatment with pembrolizumab plus SOC compared to placebo plus SOC. Deterioration on the pain symptom scale was recorded in 11.4% and 10.6% of patients in the pembrolizumab plus SOC and placebo plus SOC groups, respectively. The HR for time to deterioration on the pain symptom scale at 12 months was 1.00 (95% CI, 0.60 to 1.66; P = 0.9943) for pembrolizumab plus SOC relative to placebo plus SOC.

Harms Results

Adverse Events

Full Study Population

In KEYNOTE-811, at least 1 AE was reported by 99.4% and 100% of patients in the pembrolizumab plus SOC group and the placebo plus SOC group, respectively. Among patients randomized to receive pembrolizumab plus SOC, the 5 most commonly reported AEs were diarrhea (52.6%), nausea (48.3%), anemia (45.4%), vomiting (33.1%), and decreased appetite (32.3%). In the placebo plus SOC group, the 5 most commonly reported AEs were nausea (48.3%), diarrhea (47.1%), anemia (46.2%), decreased appetite (32.4%) and vomiting (28.6%).

In the total study population, AEs classified as Grade 3 or higher were reported in 71.7% of patients in the pembrolizumab plus SOC group and 65.9% of patients in the placebo plus SOC group. The most common AEs that were classified as Grade 3 or higher (reported in more than 5% of patients) in the pembrolizumab plus SOC group were anemia (12.6%), diarrhea (9.7%), decreased neutrophil count (8.3%), neutropenia (6.6%), decreased platelet count (6.3%) and hypokalemia (5.7%). The most common AEs that were classified as Grade 3 or higher (reported in more than 5% of patients) in the placebo plus SOC group were anemia (10.1%), decreased neutrophil count (8.7%), diarrhea (8.4%), decreased platelet count (6.9%), hypokalemia (5.8%), nausea (5.5%), and neutropenia (5.2%).

PD-L1 Positive Subgroup

Among patients in the PD-L1 positive subgroup, at least 1 AE was reported by 99.3% and 100% of patients in the pembrolizumab plus SOC group and the placebo plus SOC group, respectively. Among patients who were PD-L1 positive and randomized to receive pembrolizumab plus SOC, the 5 most commonly reported AEs were diarrhea (53.7%), nausea (50.7%), anemia (46.3%), vomiting (35.2%), and decreased appetite (33.2%). In the placebo plus SOC group, the 5 most commonly reported AEs were nausea (48.5%), diarrhea (46.8.1%), anemia (46.8%), vomiting (30.5%) and decreased appetite (30.2%).

In the PD-L1 positive subgroup, AEs classified as Grade 3 or higher were reported in 73.8% of patients in the pembrolizumab plus SOC group and 65.8% in the placebo plus SOC group. The most common AEs that were classified as Grade 3 or higher (reported in more than 5% of patients) in the pembrolizumab plus SOC group were anemia (12.8%), diarrhea (10.7%), decreased neutrophil count (8.4%), neutropenia (7.7%), decreased platelet count (7.4%) and hypokalemia (6.0%). The most common AEs that were classified as Grade 3 or higher (reported in more than 5% of patients) in the placebo plus SOC group were anemia (10.2%), decreased neutrophil count (9.2%), diarrhea (8.5%), decreased platelet count (5.8%), and nausea (5.8%)

Serious Adverse Events

Serious adverse events were AEs resulting in death, or those that were life-threatening, required inpatient hospitalization or prolonged existing hospitalization, resulted in persistent or significant disability and/or incapacity, congenital anomaly and/or birth death or other important medical events.

Full Study Population

In the full study population, at least 1 serious AE was reported in 46.0% of patients in the pembrolizumab plus SOC group and in the group of patients who received placebo plus SOC. In the pembrolizumab plus SOC group, the following SAEs were reported in more than 2% of patients: pneumonia (5.1%), diarrhea (4.9%), and pulmonary embolism (2.9%). In the placebo plus SOC group, the following SAEs were reported in more than 2% of patients: diarrhea (4.6%) and vomiting (2.6%).

PD-L1 Positive Subgroup

In the PD-L1 positive subgroup, at least 1 serious AE was reported in 48.0% and 47.8% of patients in each the pembrolizumab plus SOC group and placebo plus SOC group, respectively. Details of incident SAEs were NR by the sponsor.

Withdrawal of Treatment Due to Adverse Events

Full Study Population

In the total study population, treatment with any of the study drugs was stopped in 41.4% and 38.4% of patients in the pembrolizumab plus SOC and the SOC group, respectively. In the pembrolizumab plus SOC group, pembrolizumab, trastuzumab and any chemotherapy were discontinued due to AEs in 13.1%, 13.1%, and 38.9% of patients, respectively. In the placebo plus SOC group, placebo, trastuzumab and chemotherapy were discontinued in 10.7%, 9.2% and 38.2% of patients, respectively. Overall, 6.3% of patients in the pembrolizumab plus SOC group and 6.9% of patients in the SOC group discontinued all drugs in the regimen.

PD-L1 Positive Subgroup

In the PD-L1 positive subgroup, treatment with any of the study drugs was stopped in 42.6% and 36.6% of patients in the pembrolizumab plus SOC and the SOC group, respectively. In the pembrolizumab plus SOC group, pembrolizumab, trastuzumab and any chemotherapy were discontinued due to AEs in 14.1%, 14.1%, and 40.3% of patients, respectively. In the placebo plus SOC group, placebo, trastuzumab and chemotherapy were discontinued in 11.5%, 10.5% and 36.3% of patients, respectively. Overall, 6.7% of patients in the pembrolizumab plus SOC group and 7.8% of patients in the placebo plus SOC group discontinued all drugs in the regimen.

Mortality

Full Study Population

In the full study population, death due to AEs was documented in 6.6% of patients who received pembrolizumab plus SOC group and in 6.1% of patients who received placebo plus SOC.

PD-L1 Positive Subgroup

In the PD-L1 positive subgroup, death due to AEs was documented in 6.7% of patients who received pembrolizumab plus SOC group and in 6.8% of patients who received placebo plus SOC.

Notable Harms

Immune-mediated AEs were of interest to the CADTH clinical review team.

Full Study Population

In the total study population, at least 1 immune-mediated AE was documented in ||||| ||| ||||| of patients who received pembrolizumab plus SOC and placebo plus SOC, respectively. Grade 3 or worse immune-mediated AEs were reported in |||| of patients in the pembrolizumab plus SOC group and in |||| of patients in the placebo plus SOC group.

PD-L1 Positive Subgroup

In the total study population, at least 1 immune-mediated AE was documented in ||||| ||| ||||| of patients who received pembrolizumab plus SOC and placebo plus SOC, respectively. Grade 3 or worse, immune-mediated AEs were reported in |||| of patients in the pembrolizumab plus SOC group and in |||| of patients in the placebo plus SOC group.

Critical Appraisal

KEYNOTE-811 is a randomized, placebo-controlled, parallel-group, multicentre, double-blinded phase III study. Patients were randomized centrally using interactive response technology, typically adequate for concealing allocation until treatment assignment. The stratification factors for randomization appeared appropriate, as they addressed important prognostic factors identified by the clinical experts consulted for this review, and the baseline characteristics between the treatment groups were generally well-balanced. Since PD-L1 status (CPS ≥ 1 versus CPS < 1) was a stratification factor, the review team assumed that the randomization and prognostic balance hold in this subgroup of interest. In both the total study population and PDL1 positive subgroup, between-group imbalances were noted in the concomitant use of loperamide and unspecified herbal and traditional medicine. However, according to the clinical experts consulted by CADTH for this review, the use of loperamide or unspecified herbal and traditional medicine is not likely to lead to any meaningful impact on treatment response. In the PD-L1 positive subgroup, a greater proportion of patients who received placebo plus SOC received subsequent therapy relative to the pembrolizumab plus SOC group. Given that the reasons for treatment discontinuation were primarily disease progression and AE (which were similar in proportion between the intervention groups), the risk of unblinding driving the use of subsequent therapies appeared to be low.

The dual primary outcomes in the KEYNOTE-811 trial were PFS and OS. An appropriate analysis set (intention-to-treat) was used to measure the effect of treatment on OS and PFS. To minimize the risk of measurement bias, the study investigators were masked to patients’ responses to treatment, and tumour response was confirmed by radiologic evidence and based on BICR as per RECIST 1.1 objective criteria. Sensitivity analysis of PFS demonstrated consistency between the BICR and investigator’s assessment of tumour response, suggesting that the procedures employed to minimize bias associated with knowledge of group assignment were adequate. OS is considered an objective outcome, and it is not prone to bias due to knowledge of group assignment. Risk of bias to due to missing outcome data for OS and PFS appeared to be low as losses to follow-up for reasons other than death were low, and sensitivity analyses with different censoring rules for PFS in the overall population were consistent. KEYNOTE-811 assessed HRQoL – outcomes deemed important by patients and clinicians – as exploratory outcomes. The double-blind nature of the trial minimized the risk of bias in the measurement of the subjective items on the EORTC QLQ-30 and EROTC QLQ-ST022. However, comparative efficacy conclusions based on HRQoL outcomes are limited due to the diminishing number of patients available to complete the questionnaires. The results of HRQoL are at risk of attrition bias. Finally, since the completion rates were not balanced between the groups, there is also a risk that attrition bias may favour 1 of the treatment groups over the other. The extent and direction of the bias, however, cannot be determined since it is not clear if those patients who completed the questionnaires were systematically different from those who did not.

Analysis of efficacy results followed a defined statistical plan and employed appropriate censoring criteria. There was adequate control for multiplicity (type I error) across the dual efficacy end points of PFS and OS and interim analyses in the total study population using a hierarchical testing procedure. Both PFS and OS were modelled using a proportional hazards assumption. Although the proportional hazards assumption underlying the HRs for OS and PFS was not tested, the curves appeared to be relatively parallel based on visual inspection. The choice to limit treatment with pembrolizumab plus SOC to patients who have PD-L1 positive disease was based on subgroup analysis. Although the subgroup analyses were prespecified, they were absent from the statistical testing hierarchy. While this presents a risk of type I error (i.e., falsely excluding the null), the subgroup of patients determined to be PD-L1 positive represented approximately 85% of the total study population. The results observed in the entire study population appeared to be driven by the PD-L1 positive subgroup; qualitatively, the total population and PD-L1 positive subgroup results were similar. Finally, results were based on interim analyses, which may have overestimated the treatment effect estimates. However, given the relatively large sample size and number of events with a 75% information fraction, the effect estimate and confidence are not likely to be highly unstable. Although reassuring, overestimation of the treatment effect cannot be excluded entirely.

GRADE Summary of Findings and Certainty of the Evidence

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 and PFS at months 12, 18 and 36

• HRQoL measured by the EORTC QLQ-C30 (global QoL, physical functioning and appetite loss) at week 24

• notable harms, including immune-mediated AEs and grade 3 or worse immune-mediated AEs.

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 CADTH’s expert committee deliberations. A final certainty rating was determined as outlined by the GRADE Working Group.

Following the GRADE approach, evidence from RCTs started as high-certainty evidence. It could be rated down for concerns related to study limitations (which refer 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 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. The presence or absence of a clinically important OS effect was determined based on a threshold informed by the clinical experts consulted for the purpose of this review. The presence or absence of a clinically important effect on HRQoL was evaluated based on MID estimates identified in the literature. For all other outcomes, the presence or absence of a clinically important effect was based on the non-null effect.

Results of GRADE Assessment

Table 3 presents the GRADE summary of findings for pembrolizumab in combination with SOC versus placebo in combination with SOC.

Table 3: Summary of Findings of Pembrolizumab in Combination With Standard of Care Versus Saline Placebo in Combination With Standard of Care for Patients With HER2 Positive Advanced Gastric or Gastroesophageal Junction Adenocarcinoma in the PD-L1 CPS ≥ 1 Subgroup

BICR = blinded independent central review; CI = confidence interval; CPS = combined positive score; EORTC QLQ-C30 = European Organization for Research and Treatment of Cancer Quality of Life Questionnaire-Core 30; HR = hazard ratio; IA = investigator assessed; LS = least squares; OS = overall survival; MID = minimally important difference; NR = not reported; PD-L1 = programmed death-ligand 1; PFS = progression-free survival; SD = standard deviation; SOC = standard of care.

Notes: 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. Results based on interim analysis 3 (data cut-off date of March 29, 2023).

^aThe sponsor requested between-group differences to aid in interpretation, and they were not part of the sponsor's analysis plan.

^bThe median follow-up time at IA3 for both the full study population and the subgroup of patients with PD-L1 positive disease was NR.

^cRated down 1 level for serious imprecision. The 95% CI is compatible with little-to-no difference and a clinically important benefit (exceeding the 5 to 10% threshold suggested by the clinical experts consulted by CADTH).

^dThe clinical experts consulted by CADTH indicated a lack of clarity about a threshold of clinical importance; therefore, the null was used. Although the certainty of evidence was not rated down for serious indirectness, there were concerns about the clinical importance of PFS.

^eRated down 2 levels for very serious study limitations because of the risk of bias due to missing data, as results were available to less than 60% of patients by week 24.

^fThe clinical experts consulted for this review indicated a lack of clarity about a threshold of clinical importance, therefore the null was employed.

^gRated down 1 level for serious imprecision. The clinical experts consulted by CADTH indicated a lack of clarity about a threshold of clinical importance, so the null was employed. No threshold was crossed, but a small number of events contributed to the estimated treatment effect.

Source: Clinical Study Report for KEYNOTE-811, Statistical Report KN811 IA3, PRO Report. Details included in the table are from the sponsor’s Summary of Clinical Evidence.

Long-Term Extension Studies

No long-term extension studies were included in this submission.

Indirect Comparisons

No indirect treatment comparisons were included in this submission. The sponsor conducted a feasibility assessment of estimating the comparative efficacy and safety of pembrolizumab combined with SOC therapy (trastuzumab in combination with FP or CAPOX) versus other fluoropyrimidine- and platinum-containing chemotherapies used in combination with trastuzumab in Canada, mainly FOLFOX (5-FU plus leucovorin and oxaliplatin) and capecitabine plus cisplatin. A systematic literature review informed the availability of relevant studies to perform an indirect comparison. This review located 1 trial (ToGA) within which some patients in 1 arm received capecitabine plus cisplatin and trastuzumab. However, an indirect comparison was impossible because the analysis pooled this arm with another (5-FU plus cisplatin and trastuzumab). Therefore, an indirect comparison was not deemed possible. A review of the feasibility appraisal by CADTH agrees with this conclusion.

Studies Addressing Gaps in the Evidence From the Systematic Review

No studies addressing gaps in the evidence from the systematic review were included in this submission.

Economic Evidence

Cost and Cost-Effectiveness

Table 4: Summary of Economic Evaluation

CAPOX = capecitabine and oxaliplatin; CISPFU = 5-fluorouracil and cisplatin; GEJ = gastroesophageal junction; HER2 = human epidermal growth factor receptor 2; ICER = incremental cost-effectiveness ratio; KM = Kaplan-Meier; LY = life-year; PFS = progression-free survival; RDI = relative dose intensity; OS = overall survival; PSM = partitioned survival model; QALY = quality-adjusted life-year; WTP = willingness to pay.

^aThe sponsor assumed that chemotherapy would comprise CAPOX plus CISPFU.

Budget Impact

CADTH identified the following key limitations with the sponsor’s analysis: Use of RDI to estimate actual drug costs is inappropriate; The proportion of patients eligible to receive pembrolizumab is uncertain; The attribution of market share to “clinical trials” is inappropriate; The market uptake of pembrolizumab may be underestimated; The use of fixed dosing for pembrolizumab is not aligned with the public drug plans’ implementation strategy; The distribution of doublet chemotherapy regimens does not align with clinical practice in Canada.

The CADTH reanalysis included assuming 100% RDI for all drugs, removing market share attributed to clinical trials, and adopting a distribution of chemotherapies aligned with Canadian clinical practice.

Based on the CADTH base case, the 3-year budget impact is expected to be $38,095,911 (year 1: $1,927,523; year 2: $13,060,487; year 3: $23,107,901) should the public drug plans reimburse pembrolizumab for use in combination with trastuzumab and chemotherapy for the first-line treatment of locally advanced unresectable or metastatic HER2-positive, PD-L1 positive gastric or GEJ adenocarcinoma.

pERC Information

Members of the Committee

Dr. Maureen Trudeau (Chair), Mr. Daryl Bell, Dr. Philip Blanchette, Dr. Kelvin Chan, Dr. Matthew Cheung, Dr. Michael Crump, Dr. Jennifer Fishman, Mr. Terry Hawrysh, Dr. Yoo-Joung Ko, Dr. Christian Kollmannsberger, Dr. Catherine Moltzan, Ms. Amy Peasgood, Dr. Anca Prica, Dr. Adam Raymakers, Dr. Patricia Tang, Dr. Marianne Taylor, and Dr. W. Dominika Wranik

Meeting date: May 8, 2024

Regrets: One expert committee member did not attend.

Conflicts of interest: None