CADTH Reimbursement Review

Atezolizumab (Tecentriq)

Sponsor: Hoffmann-La Roche Ltd.

Therapeutic area: Extensive-stage small cell lung cancer

This multi-part report includes:

Clinical Review

Pharmacoeconomic Review

Stakeholder Input

Clinical Review

Executive Summary

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

Introduction

Lung cancer is the most commonly diagnosed cancer and the leading cause of cancer-related death in Canada.¹ In 2021, an estimated 29,600 Canadians were diagnosed with lung cancer, representing approximately 13% of all new cancer cases, and an estimated 21,000 Canadians died from lung cancer.¹ The main known risk factor for lung cancer is tobacco smoking, including exposure to second-hand smoke. Non–small cell lung cancer (NSCLC) is the predominant subtype of lung cancer, accounting for approximately 85% of cases. Small cell lung cancer (SCLC) accounts for about 15% of cases and is distinguished from NSCLC by its rapid growth, early development of metastatic disease, and initial responsiveness to platinum-based chemotherapy.² SCLC is classified into 2 stages. The first, limited-stage disease, is confined to 1 hemithorax, with no extrathoracic metastases, except for ipsilateral supraclavicular lymph nodes (provided they can be included in the same radiation port as the tumour), and primary tumour and regional nodes that can be adequately encompassed in a radiation port.^3,4 The second, extensive-stage disease, is defined as disease that cannot be classified as limited, including malignant pleural or pericardial effusions, contralateral hilar or supraclavicular lymph nodes, and hematogenous metastases.^3,4 Approximately 2-thirds of patients with SCLC have extensive-stage disease at diagnosis, which is associated with particularly poor prognosis.⁵ Extensive-stage SCLC (ES-SCLC) has a median survival of 7 months to 10 months and a 1-year overall survival (OS) rate of 40% (with treatment).^1,6

Despite the considerable response rates observed with first-line chemotherapy regimens, response is not durable, and most patients with ES-SCLC relapse within 1 year of treatment completion.⁷ Subsequent treatment options for patients with relapsed disease are few, due to the limited efficacy of chemotherapy and other regimens in later lines and the low performance status of many patients with relapsed disease. In addition, brain metastases are common in SCLC, with about 10% of patients presenting with brain metastases at the time of diagnosis; an additional 40% to 50% of patients subsequently develop brain metastases, which further contributes to poor prognosis.^8,9 In the first-line setting, the most important goals of treatment are prolonging survival and improving quality of life (QoL) (reduction in symptom severity, ability to maintain independence in daily activities, and so forth). Patients with ES-SCLC have a significant need for better first-line therapies with more durable response to prolong survival beyond 2 years and preserve QoL.

Until recently, the standard first-line treatment of patients with ES-SCLC was a platinum agent (cisplatin or carboplatin) and etoposide chemotherapy.¹⁰ Despite a median survival limited to approximately 10 months, there has been no considerable improvement in OS in more than 20 years. Recently, immune checkpoint inhibitors (ICIs) added to platinum and etoposide chemotherapy have demonstrated benefit in this setting. Two ICIs, durvalumab and atezolizumab, are approved in Canada, in combination with etoposide and either carboplatin or cisplatin for the first-line treatment of patients with ES-SCLC. However, neither is currently publicly funded. Durvalumab received a CADTH recommendation to reimburse in July 2021, but the Health Technology Assessment process is not yet complete, and price negotiations are ongoing with the pan-Canadian Pharmaceutical Alliance. A reimbursement request for atezolizumab in combination with carboplatin and etoposide for the treatment of ES-SCLC was previously submitted to CADTH and did not receive a recommendation to reimburse. This current CADTH Reimbursement Review for atezolizumab in combination with a platinum-based chemotherapy and etoposide is for of a resubmission filed by the sponsors for the aforementioned indication.

Atezolizumab (Tecentriq) is a humanized monoclonal anti–programmed death ligand 1 that inhibits programmed death ligand1 (PD-L1) engagement with programmed cell death protein 1 (PD-1) and B7.1. Health Canada has approved atezolizumab in combination with carboplatin and etoposide for the first-line treatment of adult patients with ES-SCLC. The sponsor’s funding request differs from the approved Health Canada indication: “for the first-line treatment of patients with extensive-stage small cell lung cancer (ES-SCLC) in combination with a platinum-based chemotherapy and etoposide. Maintenance Tecentriq should be continued until loss of clinical benefit or unacceptable toxicity.”

The objective of this review was to evaluate the efficacy and safety of atezolizumab in combination with carboplatin and etoposide for the first-line treatment of patients with ES-SCLC.

Stakeholder Perspectives

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

Patient Input

The patient and caregiver input received for this review was collected by Lung Cancer Canada (LCC) from interviews with patients with SCLC and their caregivers, gathered from December 2021 to February 2022, as well as information from previous LCC submissions. Six respondents with SCLC had experience with atezolizumab (in combination with chemotherapy or as a single treatment), 4 of whom had extensive-stage disease. Five patients had access to atezolizumab through clinical trial and 1 through a compassionate access program. Four of these respondents resided in Ontario, 1 resided in British Columbia, and 1 resided in Quebec. Respondents indicated that a diagnosis of SCLC and the subsequent treatment had a major impact on the lives of patients and their family members. They reported that they expect the following key outcomes from any new drug or treatment: relief from disease symptoms, manageability of side effects, improved QoL, ability to maintain independence and functionality, greater access across jurisdictions, disease stability, longer periods of remission, and prolonged survival. Patients with SCLC have had a very high unmet need, as there had been no new treatment options for SCLC in 30 years until 2021, when durvalumab was approved for treatment of ES-SCLC. Six respondents who had received or were continuing to receive atezolizumab indicated that this drug had promising and durable treatment results with tolerable side effects. They also mentioned that atezolizumab had helped them regain independence, functionality, and livelihood, which reduced the burden on their caregivers and loved ones.

Clinician Input

Input From Clinical Experts Consulted by CADTH

The clinical experts consulted by CADTH noted that ES-SCLC has a relatively short median OS (10 months). Although patients typically have an initial response to treatment, most patients relapse within 6 months with poor prognosis. Given most patients’ poor performance status and poor response to subsequent therapies, first-line treatment options that increase survival are highly desired. The combination of immunotherapy and chemotherapy is widely accepted as the new standard of care for the management of ES-SCLC. The addition of durvalumab or atezolizumab to a platinum agent and etoposide would be the most appropriate initial therapy for ES-SCLC. The clinical experts consulted also indicated that there is no specific subgroup of patients best suited for treatment with atezolizumab plus carboplatin and etoposide, and all patients with ES-SCLC should be treated with combination immunotherapy and chemotherapy in the first-line setting, irrespective of symptoms, as ES-SCLC is an aggressive disease and requires prompt treatment. Response to treatment is typically assessed every 3 cycles while on chemotherapy, using radiographic imaging with a CT scan, and every 3 months thereafter.

Clinician Group Input

Clinician input was received from the Ontario Health (Cancer Care Ontario) Drug Advisory Committee and from LCC. The clinician groups noted that patients with ES-SCLC have a high unmet need for more effective therapies since most patients progress in a short period of time despite a high response rate to initial therapy. Atezolizumab would be used as initial systemic therapy in patients with ES-SCLC in combination with 4 cycles of platinum and etoposide, followed by maintenance atezolizumab until disease progression. Atezolizumab will be an alternative option to durvalumab (if durvalumab is indeed added to the provincial or territorial public formularies across Canada following negotiations with the pan-Canadian Pharmaceutical Alliance) in the first-line treatment of patients with ES-SCLC. It would fit into the current treatment paradigm only as an agent to be started concurrently with first-line platinum and etoposide chemotherapy, with the intention of continuing until disease progression, intolerance, or a patient’s choice to discontinue therapy. Patients with symptomatic brain metastases would need to receive treatment for their brain metastases before starting systemic therapy. The clinician groups believed that no specific subgroups of patients are more likely to benefit from the addition of atezolizumab to platinum-based chemotherapy and etoposide; therefore, they felt the treatment should be considered for any patient with ES-SCLC and an Eastern Cooperative Oncology Group Performance Status (ECOG PS) of 2 or better.

Drug Program Input

Input was obtained from the drug programs that participate in the CADTH Reimbursement Review process. The following were identified as key factors that could impact the implementation of a CADTH recommendation for atezolizumab:

• considerations for initiation of therapy

• considerations for prescription of therapy

• considerations for discontinuation of therapy

• generalizability of trial populations to the broader populations in the jurisdictions

• system and economic issues.

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

Clinical Evidence

Pivotal Studies and Protocol-Selected Studies

Description of Studies

IMpower133 is a randomized, multi-centre, double-blind, placebo-controlled, phase III study designed to evaluate the efficacy and safety of treatment with atezolizumab plus carboplatin and etoposide compared with treatment with placebo plus carboplatin and etoposide in patients with chemotherapy-naive ES-SCLC.¹¹ The trial was conducted in 106 sites across 21 countries (none in Canada). The co-primary end points were investigator-assessed progression-free survival (PFS) and OS. The key secondary end points were investigator-assessed objective response rate (ORR) and investigator-assessed duration of response (DOR). Patient-reported outcomes (PROs) included health-related QoL (HRQoL). The clinical cut-off date for the primary analysis (primary PFS analysis and interim OS analysis) was April 24, 2018. The clinical cut-off date for the updated analysis (final analysis of OS) was January 24, 2019. Overall, the mean age of the patients was 63.7 years (standard deviation [SD] = 8.9); 64.8% were male, and 79.9% were White. Patients had to have an ECOG PS of 0 or 1, and approximately 64% of the patients in both treatment arms had an ECOG PS of 1. Of the 526 patients screened, 403 were randomized: 201 patients to the atezolizumab arm and 202 patients to the placebo arm. The median duration of follow-up was 13.9 months at the data cut-off date of April 24, 2018 (PFS analysis; interim OS analysis), and 22.9 months at the data cut-off date of January 24, 2019 (final OS analysis).

Efficacy Results

Progression-Free Survival

At the data cut-off date for PFS analysis (April 24, 2018), the median PFS was 5.2 months (95% confidence interval [CI], 4.4 to 5.6) in the atezolizumab arm and 4.3 months (95% CI, 4.2 to 4.5) in the placebo arm. The stratified hazard ratio (HR) for disease progression or death was 0.77 (95% CI, 0.62 to 0.96; P = 0.0170).

Overall Survival

At the time of the OS interim analysis (data cut-off date: April 24, 2018), patients had a median survival follow-up time of 13.9 months. The median OS was 12.3 months (95% CI, 10.8 to 15.9) in the atezolizumab arm and 10.3 months (95% CI, 9.3 to 11.3) in the placebo arm. The stratified HR for death was 0.70 (95% CI, 0.54 to 0.91; P = 0.007).

At the final OS analysis (data cut-off date: January 24, 2019), the median survival follow-up time was 22.9 months. The median OS was 12.3 months (95% CI, 10.8 to 15.8) in the atezolizumab arm and 10.3 months (95% CI, 9.3 to 11.3) in the placebo arm. The stratified HR for death was 0.75 (95% CI, 0.60 to 0.95; P = 0.015). The 2-year event-free rates were 22.0% in the atezolizumab arm and 16.8% in the placebo arm.

Objective Response Rate

The investigator-assessed, confirmed ORR was 60.2% in the atezolizumab arm and 64.4% in the placebo arm; 2.5% and 1.0% of patients in the atezolizumab and placebo arms, respectively, had a complete response (CR). At the updated analysis, the confirmed investigator-assessed, ORR was 60.2% (95% CI, 53.1 to 67.0) in the atezolizumab arm and 64.4% (95% CI, 57.3 to 71.0) in the placebo arm; 3.5% and 1.0% of patients in the atezolizumab and placebo arms, respectively, had a CR.

Duration of Response

The median DOR (confirmed) was 4.2 months (range = 1.4 to 24.3) in the atezolizumab arm and 3.9 months (range = 2.0 to 24.2) in the placebo arm. At data cut-off (April 24, 2018), 14.9% of patients in the atezolizumab arm and 5.4% of patients in the placebo arm had ongoing response. At the updated analysis, the median DOR was 4.2 months (95% CI, 4.1 to 4.5) in the atezolizumab arm and 3.9 months (95% CI, 3.1 to 4.2) in the placebo arm.

Harms Results

The majority of patients in both treatment arms — 100% in the atezolizumab arm and 96.4% in the placebo arm — experienced at least 1 adverse event (AE) of any grade. In the atezolizumab arm, the most common AE of any grade by preferred term experienced by at least 10% of patients were anemia (43.4%), nausea (37.9%), and neutropenia (37.4%). In the chemotherapy arm, the most common AEs of any grade by preferred term experienced by at least 10% of patients were anemia (35.2%), neutropenia (35.2%), and alopecia (34.7%).

Grade 3 or 4 AEs occurred in 67.7% of patients in the atezolizumab arm and 63.3% of patients in the placebo arm. The most common grade 3 or 4 AEs reported in at least 5% of patients in the atezolizumab and placebo arms were neutropenia (22.7% versus 25.0%), decreased neutrophil count (15.7% versus 16.8%), anemia (15.7% versus 13.3%), thrombocytopenia (10.1% versus 8.7%), and hyponatremia (4.5% versus 6.6%).

In the atezolizumab arm, 38.9% of patients had at least 1 serious AE (SAE). In the placebo arm, 35.2% of patients experienced at least 1 SAE. The most common SAEs experienced by at least 1% of patients in either the atezolizumab or the chemotherapy arm were pneumonia (4.5% versus 3.6%), neutropenia (3.5% versus 4.1%), febrile neutropenia (2.5% versus 4.6%), and thrombocytopenia (2.5% versus 2.0%).

Withdrawal from any study treatment due to AEs was reported for 12.1% of patients in the atezolizumab arm and 3.1% of patients in the chemotherapy arm. In the atezolizumab arm, 11.6% of patients experienced AEs leading to discontinuation of atezolizumab, and in the placebo arm 2.6% of patients had AEs leading to discontinuation of placebo. The main reasons for permanently discontinuing atezolizumab in 21 patients in the atezolizumab arm were infusion-related reactions and gastrointestinal disorders.

Grade 5 fatal AEs occurred in 4 patients (2.0%) in the atezolizumab arm and included pneumonia, respiratory failure, death, and neutropenia. Grade 5 fatal AEs occurred in 11 patients (5.6%) in the placebo arm and included pneumonia, pulmonary sepsis, sepsis, septic shock, acute respiratory failure, hemoptysis, cardiopulmonary failure, pericardial effusion, and general physical health deterioration. The only grade 5 AE (by preferred term) that occurred in more than 1 patient was pneumonia (1 patient in the atezolizumab arm and 3 patients in the placebo arm). Of the grade 5 events, 3 events in each arm were considered related to at least 1 component of study treatment. In the atezolizumab arm, a grade 5 death was considered related to all study treatment; there was also 1 case of grade 5 pneumonia and 1 case of grade 5 neutropenia that were both considered related to both carboplatin and etoposide. In the placebo arm, a grade 5 septic shock was considered related to all study treatment, a grade 5 pneumonia was considered related to placebo, and a grade 5 cardiopulmonary failure was considered related to carboplatin.

Immune-related AEs were reported for 41.4% of patients in the atezolizumab arm and 24.5% of patients in the placebo arm. Rash (both treatment arms) and hypothyroidism (atezolizumab arm) were the most common (≥ 10% incidence) and most differentially reported (≥ 5% difference between treatment arms) immune-related AEs during treatment. Immune-related infusion-related reaction events were experienced by 5.6% of patients (n = 11) in the atezolizumab arm and 5.1% of patients (n = 10) in the placebo arm. The majority of these events were grade 1 or 2 (atezolizumab arm: n = 7 [3.5%]; placebo arm: n = 9 [4.6%]). Four patients (2.0%) in the atezolizumab arm and 1 patient (0.5%) in the placebo arm had grade 3 or 4 infusion-related reactions.

Table 2: Summary of Key Results From the IMpower133 Study

AE = adverse event; CE = carboplatin and etoposide; CI = confidence interval; CR = complete response; HR = hazard ratio; ORR = overall response rate; OS = overall survival; PFS = progression-free survival; PR = partial response; SAE = serious adverse event.

^aStratified by sex (male vs. female) and Eastern Cooperative Oncology Group Performance Status (0 vs. 1).

^bLog-rank test.

^cClopper-Pearson interval.

^dWald with continuity correction.

Source: Clinical Study Report for IMpower133 study¹²; Update Clinical Study Report for IMpower133 study.¹³

Critical Appraisal

The baseline demographic and disease characteristics were roughly balanced between the 2 treatment arms. Response outcomes (ORR and DOR) were assessed by investigators per Response Evaluation Criteria in Solid Tumours (RECIST) Version 1.1. Although the trial was double blinded and the investigators were blinded to treatment assignment, risk of bias cannot be ruled out. For example, nearly half the patients in the atezolizumab arm experienced immune-related AEs or other events. These events may have made the investigator aware of the patient’s treatment assignment. Therefore, for all investigator-assessed outcomes there may be a degree of subjectivity that could have biased the results. In addition, although the proportion of patients receiving concomitant and supportive care for symptom control was largely similar in the 2 treatment arms, which may have led to comparable PROs, including QoL outcomes, as observed in the trial, this may not mean that the 2 trial regimens truly have comparable safety and impact on QoL. Interim and final analyses were planned a priori and adequately described. The interim analysis applied the Lan-DeMets alpha spending function with the O’Brien-Fleming stopping boundary, which is deemed conservative in controlling type I error when claiming a treatment effect based on interim analysis. The updated final analysis results of OS were consistent with the interim analysis results.

The patient population in the IMpower133 study generally reflects patients in Canadian clinical practice in this setting. However, some patient groups were not represented, including those with an ECOG PS of 2 and patients with active untreated metastases. The proportion of patients with brain metastases (9%) was lower than that observed in clinical practice (10% to 20%), but this is likely due to the specific inclusion requirements for these patients (e.g., only supratentorial and cerebellar metastases, and no ongoing requirement for corticosteroids as therapy for central nervous system disease). Due to the small number of patients in some subgroups, including brain metastases at baseline, subgroup analyses failed to demonstrate similar effects in patients with brain metastases as in patients free of brain metastases. The comparator in the IMpower133 trial (carboplatin and etoposide) is relevant to the Canadian context as platinum (carboplatin or cisplatin) and etoposide chemotherapy is the current standard of care.

Indirect Comparisons

Description of Studies

One sponsor-submitted indirect treatment comparison (ITC)¹⁴ and 9 published ITCs retrieved from literature were summarized and appraised for this CADTH review.

The sponsor-submitted ITC provided estimates of PFS, OS, ORR, and incidence of SAEs for atezolizumab plus carboplatin and etoposide and for competing interventions, platinum doublet therapies, and immunotherapies used for the first-line treatment of ES-SCLC. The results of comparisons between atezolizumab plus carboplatin and etoposide versus etoposide and carboplatin, etoposide and cisplatin, and durvalumab plus carboplatin (or cisplatin) and etoposide were considered relevant for the purpose of this CADTH review.

The sponsor’s base-case analysis for each outcome included adjusted or stratified HRs reported across the trials in the relevant evidence network. Additional scenario analyses were conducted to investigate the choice of platinum agent for the analyses of PFS and OS and to explore the effect on OS of 1 study with outlier ECOG PS data.

Efficacy Results

This section will focus on the findings of the sponsor-submitted network meta-analysis (NMA).¹⁴

Progression-Free Survival

The results of the base-case analysis showed that atezolizumab plus carboplatin and etoposide was associated with longer PFS than carboplatin (or cisplatin) and etoposide (HR = 0.77; 95% credible interval [CrI], 0.62 to 0.96). Similar findings were observed in the scenario analyses that included unadjusted or unstratified HRs (scenario 1: HR = 0.76; 95% CrI, 0.62 to 0.94) or that considered etoposide and carboplatin as distinct nodes (scenario 2: HR = 0.77; 95% CrI, 0.62 to 0.95).

The results of the base-case analysis for the comparison of atezolizumab plus carboplatin and etoposide versus durvalumab plus carboplatin (or cisplatin) and etoposide showed no statistically significant difference in PFS based on the CrI which included the null, and the point estimate that was close to the null value (HR = 0.97; 95% CrI, 0.73 to 1.28). Similar findings were observed in the scenario analyses that included unadjusted or unstratified HRs (scenario 1: HR = 0.95; 95% CrI, 0.72 to 1.26).

Overall Survival

The results of the base-case analysis suggested that atezolizumab plus carboplatin and etoposide may be associated with improvement in OS, when compared with etoposide and carboplatin (or cisplatin) (HR = 0.75; 95% CrI, 0.60 to 0.95). Similar findings were obtained in the other 3 scenario analyses, which included unadjusted or unstratified HRs (scenario 1: HR = 0.71; 95% CrI, 0.55 to 0.92), investigated the robustness of the results to the exclusion of a study with outlier ECOG PS data (i.e., the Hermes 2008 study) (scenario 2: HR = 0.75; 95% CrI, 0.60 to 0.95), or considered etoposide and cisplatin as distinct nodes (scenario 3: HR = 0.75; 95% CrI, 0.60 to 0.95).

The results of the base-case analysis for the comparison of atezolizumab plus carboplatin and etoposide versus durvalumab plus carboplatin (or cisplatin) and etoposide showed no statistically significant difference in OS based on the CrI which included the null, and the point estimate that was close to the null value (HR = 1.01; 95% CrI, 0.75 to 1.36). Similar findings were obtained in the other 3 scenario analyses, which included unadjusted or unstratified HRs (scenario 1: HR = 0.95; 95% CrI, 0.69 to 1.31), investigated the robustness of the results to exclusion of the Hermes 2008 study (scenario 2: HR = 1.01; 95% CrI, 0.75 to 1.35), or considered etoposide and cisplatin as distinct nodes (scenario 3: HR = 0.95; 95% CrI, 0.69 to 1.31).

Objective Response Rate

The comparison of atezolizumab plus carboplatin and etoposide versus durvalumab plus carboplatin (or cisplatin) and etoposide showed that atezolizumab plus etoposide and carboplatin was associated with lower odds of ORR (odds ratio [OR] = 0.54; 95% CrI, 0.32 to 0.94). The OR for the comparison of atezolizumab plus carboplatin and etoposide versus etoposide and carboplatin or cisplatin was estimated to be 0.84 (95% CrI, 0.56 to 1.25), and the OR for the comparison of atezolizumab plus carboplatin and etoposide versus etoposide and cisplatin was 0.70 (95% CrI, 0.37 to 1.35).

Harms Results

Serious Adverse Events

Two studies were used to inform the evidence network for SAEs. The OR observed in the comparison of atezolizumab plus carboplatin and etoposide versus durvalumab plus carboplatin (or cisplatin) and etoposide was estimated to be 1.37 (95% CrI, 0.79 to 2.37), and in the comparison of atezolizumab plus carboplatin and etoposide versus etoposide and cisplatin, the OR was 1.12 (95% CrI, 0.74 to 1.70). No statistically significant difference was observed based on the CrIs which included the null value, and the point estimates that were close to the null value (i.e., OR = 1).

Critical Appraisal

The sponsor’s systematic review methods for identifying and assessing studies included in the network were considered appropriate for identifying relevant studies. The PICO (population, interventions, comparisons, outcomes) criteria were pre-specified, and articles were reviewed by 2 independent reviewers while a second analyst extracted data. All relevant comparators identified in the CADTH review protocol that were considered relevant to the Canadian practice context were presented in the sponsor’s NMA. Outcomes presented in the trials included in the network analysis were considered relevant and clinically meaningful by the clinician experts consulted during the CADTH review. The population studied in all 8 trials included in the NMA was considered relevant for the reimbursement request. Most studies included untreated patients with ES-SCLC. One study (Skarlos 1994) recruited a different population in the trial but had a subgroup of patients with ES-SCLC. Information from the subgroup analysis was used to inform the network. Quality assessments were conducted using the validated 7-criteria checklist provided by the National Institute for Health and Care Excellence (NICE) single technology appraisal user guide.

A generalized linear regression model with a binomial likelihood, logit link model was used; this model was considered appropriate for the types of outcomes assessed in the network. The sponsor explored both fixed-effect (FE) and random-effect (RE) models in its base-case scenarios, and results from the FE model were presented. The sponsor provided a justified rationale for using the FE model over the RE model based on the model fit criteria, including a judgment on the similarities of the studies included in terms of effect modifiers.

The transitivity assumption was assessed by evaluating potential effect modifiers. There was considerable heterogeneity across trials, particularly in terms of ECOG PS. The Hermes 2008 trial enrolled less than 53% of patients with an ECOG PS of 0 or 1 in both treatment arms versus 100% in the CASPIAN, ECOG-ACRIN EA5161, IMpower133, and KEYNOTE-604 trials. In the Okamoto 2007 trial, patients with an ECOG PS of 0 to 2 were included if they were 70 years or older and those with an ECOG PS of 3 were included if they were younger than 70 years. There was inconsistency in the reporting of the number and type of metastatic sites across the trials. Heterogeneity in the use of subsequent anticancer therapy administered in the second line and higher to patients recruited in the studies was identified as a potential source of bias affecting OS assessment (non-protocol second- and third-line treatment was reported in the Hermes 2008, Okamoto 2007, and Schmittel 2011 trials); this may also affect the generalizability of the findings of the NMA to the Canadian setting. Variability was also observed in the dosing of etoposide plus carboplatin or cisplatin across the trials: 3 studies — CASPIAN, Schmittel 2011, and Skarlos 1994 — randomized patients to combination chemotherapy regimens for up to 4 to 6 cycles, whereas in the IMpower133 trial (including ECOG-ACRIN EA5161, KEYNOTE-604, Hermes 2008, and Okamoto 2007) patients were dosed with the comparator for up to 4 cycles. This may have impacted the findings of the ITC.

According to the sponsor’s ITC report, a meta-regression analysis to investigate inter-trial heterogeneity was not possible as there were insufficient studies (i.e., due to the presence of several single study connections between interventions). Scenario analyses related to certain characteristics of interest were included in the sponsor’s NMA report to address heterogeneity across the trials included in the network (e.g., removal of the Hermes 2008 trial, which was an outlier as it had the smallest proportion of patients with ECOG PS < 2 from the OS base-case analysis). According to the clinical expert consulted, ECOG PS and metastatic sites (liver and brain) were the most significant effect modifiers in the treatment of ES-SCLC patients. The sponsor acknowledged that additional scenario or subgroup analyses were feasible for PFS and OS; however, because relevant subgroup data are not currently available from the trials of the evidence networks investigating immunotherapies (i.e., the CASPIAN, ECOG-ACRIN EA5161, and KEYNOTE-6040 ongoing trials), not all possible subgroup analyses were included in the sponsor’s report. Therefore, the NMA results should be interpreted with caution due to limitations that may arise from between-study differences in some covariates and lack of sufficient evidence to minimize heterogeneity and inconsistency (e.g., by performing meta-regression analysis).

Other Relevant Evidence

No other relevant evidence was identified.

Conclusions

Based on clinical data from the IMpower133 study, atezolizumab in combination with carboplatin and etoposide demonstrated a statistically significant benefit compared to placebo in combination with carboplatin and etoposide in the first-line treatment of patients with ES-SCLC. The updated OS analysis, with a median of 22.9 months of follow-up, showed results consistent with those reported at the interim OS analysis, which suggests maintained clinical benefit for atezolizumab in combination with carboplatin and etoposide. Although the net gain of about 1 month in median PFS and 2 months in median OS observed with the addition of atezolizumab to carboplatin and etoposide is modest, it was considered by the clinical experts consulted by CADTH to be clinically meaningful in this setting where patients experience rapid tumour growth and fast clinical deterioration and have poor prognosis. The toxicity profile of atezolizumab was consistent with its immune-mediated mechanism of action, with no new safety concerns. Based on the results of the sponsor-submitted ITC, atezolizumab appears to demonstrate, in terms of improving PFS and OS, comparable benefit to durvalumab, the only other immunotherapy agent approved for the first-line treatment of ES-SCLC in Canada (but not currently funded by the drug plans in Canada). However, no firm conclusions could be drawn due to the small number of studies per comparison, leading to lower precision in effect estimates.

Introduction

Disease Background

Lung cancer is the most commonly diagnosed cancer and the leading cause of cancer-related death in Canada.¹ In 2021, an estimated 29,600 Canadians were diagnosed with lung cancer, representing approximately 13% of all new cancer cases, and 21,000 Canadians died from lung cancer, representing 25% of all cancer deaths in 2021.¹ It is estimated that 1 in 15 Canadians will develop lung cancer during their lifetime; 1 in 18 men and 1 in 20 women will die from it. The overall 5-year net survival for lung cancer from 2015 to 2017 was estimated to be 19% for men and 26% for women.¹ The incidence of lung cancer begins to rise at 40 years of age and peaks between 65 and 84 years of age.¹⁵ The main known risk factors for lung cancer include tobacco smoking (including exposure to second-hand smoke) and exposure to asbestos, arsenic, radon, non-tobacco-related polycyclic aromatic hydrocarbons, and air pollution. The most common symptoms of lung cancer are cough, dyspnea, hemoptysis, chest pain, and systemic symptoms such as fatigue and weight loss. The diagnostic evaluation entails imaging that can include chest X-ray, CT scan, PET, MRI and bone scans, and tissue biopsy for histologic confirmation.

Lung cancer is histologically divided into NSCLC, which accounts for approximately 85% of cases, and SCLC which accounts for about 12% to 15% of cases, with more than 4,000 cases diagnosed annually across Canada.^2,16 Small cell lung cancer has pathological, clinical, and molecular characteristics that are distinct from those of NSCLC. SCLC is a high-grade neuro-endocrine carcinoma arising predominantly in current or former smokers.⁹ It is marked by an exceptionally high proliferative rate and early development of widespread metastases.^9,17 SCLC is primarily classified into limited-stage and extensive-stage disease. The Veterans Administration Lung Study Group 2-stage system has been routinely used for the clinical staging of SCLC since the late 1950s. Limited-stage disease is defined as a disease confined to 1 hemithorax (although local extension may be present); no extrathoracic metastases, except for ipsilateral supraclavicular lymph nodes, provided they can be included in the same radiation port as the tumour; and primary tumour and regional nodes that can be adequately encompassed in a radiation port.^3,4 Extensive-stage disease is defined as disease that cannot be classified as limited, including malignant pleural or pericardial effusions, contralateral hilar or supraclavicular lymph nodes, and hematogenous metastases.^3,4 More recently, the International Association for the Study of Lung Cancer has promoted the use of the tumour-node-metastasis staging classification to provide better anatomic discrimination for the measurement of outcomes, prognostic information, and more precise lymph node staging.⁹ However, patient selection in clinical trials has so far predominantly relied on the Veterans Administration Lung Study Group classification system. Approximately 2-thirds of patients with SCLC have extensive-stage disease at diagnosis, which is associated with particularly poor prognosis.⁵ Extensive-stage SCLC has a median survival of 7 months to 10 months (with treatment) and a 1-year OS rate of 40%.^1,6 Survival beyond 2 years is generally no more than 15%, and the 5-year survival rate is less than 7%.¹⁸ In addition to extensive-stage disease, poor prognostic factors in SCLC include older age, male sex, impaired performance status, weight loss, elevated lactate dehydrogenase, and higher total gross tumour volume.¹⁹

As SCLC is a chemosensitive tumour, rapid responses with symptomatic improvement are often observed with platinum doublet chemotherapy. Evidence has shown that chemotherapy dramatically prolongs survival compared to best supportive care; 60% to 80% of patients with ES-SCLC respond to first-line chemotherapy.²⁰ However, despite the remarkable response rates observed with first-line chemotherapy regimens, response is not durable. Most patients with ES-SCLC develop chemotherapy-resistant disease and relapse within 1 year of treatment completion.⁷ Prognosis continues to remain poor for those with relapsed disease, with a median survival of 5 months to 6 months.²¹ Subsequent therapy options for patients with relapsed disease are few due to limited efficacy of chemotherapy and other regimens in later lines and the low performance status of many patients with relapsed disease. In addition, brain metastases are common in SCLC, with about 10% of patients presenting with brain metastases at the time of diagnosis and an additional 40% to 50% of patients subsequently developing brain metastases, which further contributes to poor prognosis.^8,9 Although prophylactic cranial irradiation (PCI) has been shown to improve systemic control of disease in some patients with ES-SCLC, it is associated with significant impairment in neurocognitive functioning, and the overall evidence for the effectiveness of this approach in improving OS in all patients with ES-SCLC with central nervous system metastases remains limited.^22-24 In the first-line setting, the most important goals of treatment are prolonging survival and improving QoL (reducing symptom severity, maintaining independence in daily activities, and so forth).

Standards of Therapy

The standard first-line treatment for patients with newly diagnosed ES-SCLC consists of a platinum agent (cisplatin or carboplatin) with etoposide. The treatment landscape for SCLC had remained virtually unchanged for the past 3 decades. Recently, multiple phase III randomized controlled trials (RCTs) have demonstrated the benefits of adding an ICI to first-line chemotherapy in patients with newly diagnosed ES-SCLC.^11,25,26 The addition of either of 2 anti-PD-L1 monoclonal antibodies, durvalumab or atezolizumab, to standard platinum-etoposide, with continuation of immunotherapy as maintenance, improved PFS and OS. In Canada, 2 ICIs — durvalumab and atezolizumab — are approved, in combination with etoposide and either carboplatin or cisplatin for the first-line treatment of patients with ES-SCLC. However, neither is currently publicly funded. Durvalumab received a CADTH recommendation to reimburse in July 2021, but the Health Technology Assessment process is not yet complete, and price negotiations are ongoing with the pan-Canadian Pharmaceutical Alliance. Durvalumab can currently be obtained only through the sponsor’s compassionate access program in some provinces. A reimbursement request for atezolizumab in combination with carboplatin and etoposide for the treatment of ES-SCLC was previously submitted to CADTH and did not receive a recommendation to reimburse. The current CADTH Reimbursement Review has been conducted for a resubmission filed by the sponsors for the aforementioned indication.

Drug

Atezolizumab (Tecentriq) is an engineered humanized immunoglobulin monoclonal antibody targeting PD-L1 and provides a dual blockade of interactions between PD-L1 and its receptors PD-1 and B7.1, restoring tumour-specific T-cell immunity.

In the US, the FDA has approved atezolizumab in combination with carboplatin and etoposide for the first-line treatment of adult patients with ES-SCLC. Atezolizumab has been approved for the same indication by the European Medicines Agency’s Committee for Medicinal Products for Human Use.

Health Canada has approved atezolizumab in combination with carboplatin and etoposide for the first-line treatment of adult patients with ES-SCLC. The sponsor’s reimbursement request differs from the approved Health Canada indication. The requested reimbursement is for the first-line treatment of patients with ES-SCLC in combination with a platinum-based chemotherapy and etoposide. Maintenance Tecentriq should be continued until loss of clinical benefit or unacceptable toxicity.

Atezolizumab for injection is supplied as a concentrate for solution for infusion in 60 mg/mL, 1,200 mg/20 mL, and 840 mg/14mL single-use vials. During the induction phase, the recommended dose of atezolizumab is 1,200 mg administered by IV infusion followed by carboplatin, and then etoposide administered by IV infusion on day 1. Etoposide is administered by IV infusion on days 2 and 3. This regimen is administered every 3 weeks for 4 cycles. The induction phase is followed by a maintenance phase without chemotherapy in which 1,200 mg of atezolizumab is administered by IV infusion every 3 weeks. Patients are treated with atezolizumab until loss of clinical benefit or unacceptable toxicity.²⁷

Stakeholder Perspectives

Patient Group Input

This section was prepared by CADTH staff based on the input provided by patient groups. Please refer to the Stakeholder Input for the full patient group input submitted to CADTH.

The patient and caregiver input received for this review was collected by LCC, a registered national charitable organization focused on lung cancer education, patient support, research, and advocacy. The input was sourced from interviews with patients with SCLC and caregiver testimonies gathered from December 2021 to February 2022, as well as information from previous LCC submissions. Six respondents with SCLC had experience with atezolizumab (in combination with chemotherapy or as a single treatment), 4 of whom had extensive-stage disease. Five patients had access to atezolizumab through clinical trial and 1 through a compassionate access program. Four of these respondents resided in Ontario, 1 in British Columbia, and 1 in Quebec.

Respondents indicated that a diagnosis of SCLC and the subsequent treatment had a major impact on the lives of patients and their family members. Several respondents reported that the diagnosis of SCLC was devastating to them and their family. Respondents indicated a varying range of SCLC symptoms that affected their daily activities. The daily activities that were most commonly impacted included the ability work, drive, travel, participate in activities they enjoy, and spend time with family and friends. Caregivers might need to retire or take time off work to provide care. Respondents reported that they expect the following key outcomes from any new drug or treatment: relief from disease symptoms, manageability of side effects, improved QoL, ability to maintain independence and functionality, greater access across jurisdictions, disease stability, longer periods of remission, and prolonged survival. According to the patient input received, the SCLC patient population has had a significant unmet need, as there had been no new treatment options for SCLC in 30 years, until 2021, when durvalumab was approved for treatment of ES-SCLC. Six respondents who had received or were continuing to receive atezolizumab indicated that this drug had had promising and durable treatment results with tolerable side effects. They also mentioned that atezolizumab had helped them regain independence, functionality, and livelihood, which reduced the burden on their caregivers and loved ones.

Clinician Input

Input From Clinical Experts Consulted by CADTH

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

Unmet Needs

The clinical experts consulted by CADTH noted that ES-SCLC is considered incurable, with a relatively short median OS (10 months). Although patients typically have initial response, the majority of patients will experience a relapse within 6 months. At relapse, outcomes are poor due both to patient-related factors, including poor performance status, and treatment-related factors such as poor response to subsequent therapies and significant toxicity. Therefore, first-line treatment options that increase expected survival are highly desired.

Place in Therapy

Atezolizumab is a monoclonal antibody against the PD-1/PD-L1 pathway, which would be used in combination with carboplatin and etoposide. This drug would be used in the first-line setting and would change the current treatment paradigm for SCLC. The clinical experts noted that there are numerous examples in other cancers, including NSCLC, demonstrating that the combination of chemotherapy and immunotherapy results in longer PFS and OS than with standard chemotherapy alone. The combination of immunotherapy and chemotherapy is widely accepted as the new standard of care for the management of ES-SCLC. The addition of durvalumab, or atezolizumab, to a platinum agent and etoposide would be the most appropriate initial therapy for ES-SCLC.

Patient Population

The clinical experts believed that no specific subgroup of patients is best suited for treatment with atezolizumab plus carboplatin and etoposide. They noted that data on various subgroups (e.g., age, sex, brain metastasis, PD-L1, and tumour mutational burden) were collected in the pivotal trials of atezolizumab and durvalumab (i.e., IMpower133 and CASPIAN), but none of the clinical characteristics studied as subgroups were found to be predictive or prognostic in this setting. In the absence of any contraindications (e.g., active or uncontrolled autoimmune conditions, or paraneoplastic autoimmune conditions requiring systemic therapy), atezolizumab should be offered to all patients with ES-SCLC (too extensive to be treated safely with curative intent radiotherapy). Patients should be treated irrespective of symptoms, as ES-SCLC is an aggressive disease and treatment needs to be started on an urgent basis. The clinical experts indicated that the diagnosis of ES-SCLC is easily made by oncologists involved in the treatment of SCLC. Diagnosis is routinely made by pathologists, and staging investigations are standardized to determine if limited- or extensive-stage disease exists. There is no population of pre-symptomatic patients who would be observed and not treated.

Assessing Response to Treatment

The clinical experts consulted by CADTH noted that there are currently no clinical features or biomarkers known to be predictive of response to treatment in patients with ES-SCLC. Thus, all patients should be treated with combination immunotherapy and chemotherapy in the first-line setting. Response to treatment is typically assessed every 3 cycles while on chemotherapy, using radiographic imaging with a CT scan, and every 3 months thereafter. It is challenging at an individual patient level to measure if survival is improved. Important outcomes to measure for individual patients are response rate, DOR, improvement in symptoms, and QoL.

Discontinuing Treatment

The decision to discontinue treatment with atezolizumab may be made after occurrence of disease progression or development of significant toxicity (i.e., grade 3 or higher or persistent grade 2 toxicity that is impacting function). The clinical experts also noted that it is important to recognize that some patients may have progression according to RECIST but might be benefiting from treatment. In these cases, it may be appropriate to continue treatment until it is clear that the treatment is failing.

Prescribing Conditions

Atezolizumab would typically be prescribed by medical oncologists. In some regions, pulmonologists who treat thoracic malignancies may also manage patients with ES-SCLC. No companion testing is required.

Clinician Group Input

This section was prepared by CADTH staff based on the input provided by clinician groups. Please refer to the Stakeholder Input section for the full clinical group input submitted to CADTH.

Clinician input was received from the Ontario Health (Cancer Care Ontario) Drug Advisory Committee, which provides evidence-based clinical and health system guidance on drug-related issues in support of Cancer Care Ontario’s mandate, including the Provincial Drug Reimbursement Programs and the Systemic Treatment Program, and from LCC, a national charity with the objective of increasing awareness about lung cancer, supporting research and advocating access to treatments for patients with lung cancer. The clinician groups noted that the most important goal of treatment for ES-SCLC is improved OS. Moreover, since ES-SCLC is a cancer with a high propensity to spread to the brain, a systemic therapy with significant activity in the brain would be important to avoid brain irradiation and preserve functioning and QoL. Patients with ES-SCLC have a high unmet need for more effective therapies since most patients progress in a short period of time despite a high response rate to initial therapy. Atezolizumab would be used as initial systemic therapy in patients with ES-SCLC in combination with 4 cycles of platinum and etoposide, followed by maintenance atezolizumab until disease progression. The clinician groups noted that atezolizumab would be considered as an alternative option to durvalumab for the first-line treatment of patients with ES-SCLC. It would fit into the current treatment paradigm only as an agent to be started concurrently with first-line platinum and etoposide chemotherapy, with the intention of continuing until disease progression, intolerance, or a patient’s choice to discontinue therapy. Patients with symptomatic brain metastases should have treatment for their brain metastases before starting systemic therapy. The addition of atezolizumab to platinum and etoposide will not have any downstream impact on other treatment options. Progression after atezolizumab therapy would be treated with additional systemic chemotherapy or other regimens, including cyclophosphamide, Adriamycin, and vincristine, as per the current paradigm. The clinician groups also indicated that no specific subgroups of patients are more likely to benefit from the addition of atezolizumab; therefore, the treatment should be considered for any patient with ES-SCLC and an ECOG PS of 2 or better.

Drug Program Input

Input was obtained from the drug programs that participate in the CADTH Reimbursement Review process. The following were identified as key factors that could impact the implementation of a CADTH recommendation for atezolizumab.

The implementation questions and corresponding responses from the clinical experts consulted by CADTH are summarized in Table 3.

Table 3: Summary of Drug Plan Input and Clinical Expert Response

ECOG PS = Eastern Cooperative Oncology Group Performance Status; ES-SCLC = extensive-stage small cell lung cancer; PAG = provincial advisory group; pERC = pan-Canadian Oncology Drug Review Expert Review Committee; RECIST = Response Evaluation Criteria in Solid Tumours.

Clinical Evidence

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

Systematic Review: Pivotal and Protocol-Selected Studies

Objectives

To evaluate the efficacy and safety of atezolizumab in combination with carboplatin and etoposide for the first-line treatment of patients with ES-SCLC.

Methods

Studies selected for inclusion in the systematic review included pivotal studies provided in the sponsor’s submission to CADTH and Health Canada, as well as those meeting the selection criteria presented in Table 4. Outcomes included in the CADTH review protocol reflect outcomes considered to be important to patients, clinicians, and drug plans.

Table 4: Inclusion Criteria for the Systematic Review

DOR = duration of response; ECOG PS = Eastern Cooperative Oncology Group Performance Status; HRQoL = health-related quality of life; ORR = objective response rate; OS = overall survival; PFS = progression-free survival; RCT = randomized controlled trials; RECIST = Response Evaluation Criteria in Solid Tumours.

^aCurrent standard of care.

^bNot publicly funded. Available through compassionate access only.

^cThese outcomes were identified as being of particular importance to patients in the input received by CADTH from patient groups.

The literature search for clinical studies was performed by an information specialist using a peer-reviewed search strategy according to the PRESS (Peer Review of Electronic Search Strategies) checklist.²⁸

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

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

The initial search was completed on February 22, 2022. Regular alerts updated the search until the meeting of the CADTH pan-Canadian Oncology Drug Review Expert Committee on July 13, 2022.

Grey literature (literature that is not commercially published) was identified by searching relevant websites from the Grey Matters: A Practical Tool For Searching Health-Related Grey Literature checklist.²⁹ Included in this search were the websites of regulatory agencies (the FDA and the European Medicines Agency). Google was used to search for additional internet-based materials. See Appendix 1 for more information on the grey literature search strategy.

These searches were supplemented through the review of bibliographies of key papers and through contacts with appropriate experts. In addition, the sponsor was contacted for information regarding unpublished studies.

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

A focused literature search for NMAs dealing with atezolizumab or SCLC was run in MEDLINE All (1946–) on February 18, 2022. No search limits were applied.

Findings From the Literature

A total of 934 studies were identified from the literature for inclusion in the systematic review (Figure 1). The included study is summarized in Table 5. A list of excluded studies is presented in Appendix 2.

Figure 1: Flow Diagram for Inclusion and Exclusion of Studies

Table 5: Details of the IMpower133 Study

AE = adverse event; AUC = area under the curve; CNS = central nervous system; DOR: duration of response; 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; EORTC QLQ-LC13 = European Organisation for Research and Treatment of Cancer Quality of Life Questionnaire Lung Cancer 13; ES-SCLC = extensive-stage small cell lung cancer; ORR = objective response rate; OS = overall survival; PFS = progression-free survival; RCT = randomized controlled trial; RECIST 1.1 = Response Evaluation Criteria in Solid Tumours Version 1.1; SCLC = small cell lung cancer.

Source: Clinical Study Report for IMpower133 study.¹²

Description of the IMpower133 Study

IMpower133 is a randomized, multi-centre, double-blind, placebo-controlled phase III study designed to evaluate the efficacy and safety of treatment with atezolizumab plus carboplatin and etoposide compared with treatment with placebo plus carboplatin and etoposide in patients with chemotherapy-naive ES-SCLC. The study included a phase I safety run-in period to establish tolerability of the study treatment. The trial was conducted in 106 sites across 21 countries (none in Canada). The clinical cut-off date for the primary analysis (primary PFS analysis and interim OS analysis) was April 24, 2018. The clinical cut-off date for the updated analysis (final analysis of OS) was January 24, 2019. The trial was funded by Hoffmann-La Roche Ltd.

Randomization and treatment allocation: A total of 403 eligible patients were randomized (1:1) to receive either atezolizumab plus carboplatin and etoposide or placebo plus carboplatin and etoposide. Randomization was performed with the use of a permuted-block randomization method — the interactive voice or web response system IxRS — and was stratified according to sex, ECOG PS (0 versus 1), and presence of brain metastases (yes versus no). After written informed consent had been obtained, all screening procedures and assessments had been completed, and eligibility had been established, the study site obtained each patient’s identification number and treatment assignment from the IxRS for eligible patients.

Blinding: IMpower133 was a double-blind study. The sponsor, the study site personnel including the investigators, and the patients were blinded to treatment assignment.

Study phases: The induction phase of the study consisted of 4 cycles of atezolizumab or placebo plus chemotherapy, with each cycle being 21 days in duration. Following the induction phase, patients continued maintenance therapy with either atezolizumab or placebo (21-day cycles). The patients received their first dose of the study drug on the day of randomization if possible. If not possible, the first dose occurred within 5 days after randomization (Figure 2).

Figure 2: IMpower133 Study Schema

ECOG PS = Eastern Cooperative Oncology Group Performance Status; RECIST v1.1 = Response Evaluation Criteria in Solid Tumours Version 1.1; SCLC = small cell lung cancer.

Source: Update Clinical Study Report for IMpower133 study.¹³

Protocol Amendments

Several changes were made to the original study protocol throughout the duration of the study. The main amendments (from most to least recent) were as follows. On August 29, 2017, modifications were made to the statistical analysis plan and the timing for the efficacy analyses for PFS and OS in the global study. The OS event-patient ratio for the interim OS analysis was increased from 45% to 55%; for the final OS analysis, the ratio was reduced from 74% to 70%. The second OS interim analysis, at the time when 258 OS events had occurred, was also removed. Consequently, 280 deaths were required for the final OS analysis, estimated to be achieved at approximately first patient randomized plus 31 months, compared to 298 OS events at 37 months under a 74% event-patient ratio. These changes were implemented to be consistent with other studies in the atezolizumab first-line lung cancer program. The multiplicity strategy was adjusted from splitting alpha to a group sequential Holm procedure so that alpha spent on PFS could be recycled to OS when PFS was significant, and vice versa, to most efficiently use alpha and maximize power. On August 25, 2016, the requirement for a tumour response assessment at the treatment discontinuation visit was removed. On June 8, 2016, it was added that in the case of an early termination of the study, patients who were deriving clinical benefit from treatment with atezolizumab would be permitted to continue treatment with atezolizumab at the discretion of the investigator.

Populations

Inclusion and Exclusion Criteria

Eligible patients were adults with histologically or cytologically confirmed ES-SCLC (defined according to the Veterans Administration Lung Study Group staging system), measurable ES-SCLC according to RECIST 1.1, and an ECOG PS of 0 or 1 who had not received previous systemic treatment for ES-SCLC. Patients with treated asymptomatic central nervous system metastases were eligible. Patients with a history of autoimmune disease and previous treatment with CD137 agonists or immune checkpoint blockade therapies were ineligible.

Baseline Characteristics

Between June 6, 2016, and May 31, 2017, a total of 403 patients who met the eligibility criteria were enrolled at 106 sites in 21 countries and were randomly assigned to receive chemotherapy with either atezolizumab (n = 201) or placebo (n = 202). In the intention-to-treat (ITT) population, most patients (79.9%) were White and most (64.8%) were male. More than half the patients (53.8%) were 65 years or younger, with a median (range) age of 64.0 (26 to 90) years. Almost 2-thirds of patients (65.3%) had an ECOG PS of 1. The majority of patients (97.0%) were either past or current smokers. Overall, 8.7% of patients had brain metastases and 37.0% of patients had liver metastases at enrolment. Most patients (93.1%) enrolled in the study were initially diagnosed with ES-SCLC; 6.7% were initially diagnosed with limited-stage SCLC before progressing to ES-SCLC. The median time from diagnosis of ES-SCLC to study enrolment was 0.7 months. Twenty-two patients in each treatment arm received PCI. The baseline demographic and disease characteristics were well balanced between the 2 treatment arms (Table 6).

Table 6: Summary of Baseline Characteristics

CE = carboplatin and etoposide; ECOG = Eastern Cooperative Oncology Group; ES-SCLC = extensive-stage small cell lung cancer; ITT = intention to treat; SD = standard deviation.

Source: Clinical Study Report for IMpower133 study.¹²

Interventions

Induction phase

Atezolizumab Plus Carboplatin and Etoposide

For patients randomly assigned to the atezolizumab plus carboplatin and etoposide arm, treatment in the induction phase consisted of 4 21-day cycles of carboplatin (area under the curve of 5 mg/mL/min administered intravenously on day 1 of each cycle) and etoposide (100 mg/m² of body surface area administered intravenously on days 1 through 3 of each cycle), plus atezolizumab (at a dose of 1,200 mg, administered on day 1 of each cycle). Atezolizumab was provided as a sterile liquid in a single-use, 20 mL glass vial. The vial was designed to deliver 20 mL (1,200 mg) of atezolizumab solution but could contain more than the stated volume to enable delivery of the entire 20 mL.

Placebo Plus Carboplatin and Etoposide

For patients randomly assigned to the placebo plus carboplatin and etoposide arm, treatment in the induction phase consisted of 4 21-day cycles of carboplatin (area under the curve of 5 mg/mL/min administered intravenously on day 1 of each cycle) and etoposide (100 mg/m² of body surface area administered intravenously on days 1 through 3 of each cycle), plus placebo (administered on day 1 of each cycle). Carboplatin and etoposide were used in the commercially available formulations.

Maintenance phase

In both treatment arms, the induction phase was followed by a maintenance phase during which patients received either atezolizumab (1,200 mg) or placebo according to their previous random assignment, until the occurrence of unacceptable toxic effects or disease progression according to RECIST 1.1 criteria. During the maintenance phase, PCI was permitted per local standard of care. Thoracic radiation with curative intent or the intent to eliminate residual disease was not permitted. Palliative thoracic radiation was allowed.

Concomitant Treatment

Concomitant treatment included any medication (e.g., prescription drugs, over-the-counter drugs, or homeopathic remedies and nutritional supplements) used by a patient from 7 days before screening until the treatment discontinuation visit. Pre-medication with antihistamines could be administered for any atezolizumab or placebo infusions after cycle 1.

Patient care with supportive therapies was managed as clinically indicated per local standards. Patients who experienced infusion-associated symptoms could be treated symptomatically with acetaminophen, ibuprofen, diphenhydramine, and/or famotidine or another H2 receptor antagonist per standard practice. Serious infusion-associated events — manifested by dyspnea, hypotension, wheezing, bronchospasm, tachycardia, reduced oxygen saturation, or respiratory distress — could be managed with supportive therapies (e.g., supplemental oxygen and Beta2-adrenergic agonists) as clinically indicated^. Systemic corticosteroids and tumour necrosis factor alpha inhibitors could be administered at the discretion of the treating physician for the treatment of specific AEs when associated with atezolizumab therapy.

Study Treatment Discontinuation

Patients discontinued study treatment if they experienced symptomatic deterioration attributed to the following: disease progression as determined by the investigator after integrated assessment of radiographic data, biopsy results if available, and the patient’s clinical status; intolerable toxic effects related to atezolizumab (including immune-mediated AEs determined by the investigator to be unacceptable given the individual patient’s potential response to therapy and severity of the event); or intolerable toxic effects related to other components of the study treatment. If 1 component of study treatment was discontinued permanently because of tolerability concerns, the patient was allowed to continue with other components of study treatment until disease progression if agreed by the investigator and patient. In addition, radiographic progressive disease per RECIST 1.1, use of another non-protocol-specified anticancer therapy, or presence of any medical condition that could jeopardize the patient’s safety by continued treatment were grounds for discontinuation of study treatment.

Patients could be considered for treatment beyond radiographic progression per RECIST 1.1 at the discretion of the investigator and after discussion with the patient if the following criteria were met: evidence of clinical benefit as assessed by the investigator; no decline in ECOG PS that could be attributed to disease progression; and absence of tumour progression at critical anatomic sites (e.g., leptomeningeal disease) that could not be managed by protocol-allowed medical interventions. Patients were required to provide informed consent to acknowledge deferring other treatment options in favour of continuing study treatment at the time of initial progression. Patients were monitored clinically and with a follow-up scan after 6 weeks, or sooner if symptomatic deterioration occurred. Treatment was discontinued for unacceptable toxic effects or clinical deterioration due to disease progression.

Dose Modification or Interruption

No dose reductions for atezolizumab or placebo were permitted. Patients could temporarily suspend treatment with atezolizumab or placebo for up to 105 days beyond the last dose if they experienced an AE that required a dose to be withheld. If atezolizumab or placebo was withheld because of AEs for more than 105 days beyond the last dose, the patient was discontinued from atezolizumab or placebo treatment. If a patient had to be tapered off steroids used to treat AEs, atezolizumab could be withheld for additional time beyond 105 days from the last dose until steroids were discontinued or reduced to a prednisone dose (or dose equivalent) less than or equal to 10 mg/day. The acceptable length of interruption depended on agreement between the investigator and the medical monitor.

Dose modifications for carboplatin and etoposide were permitted for toxicity according to the prescribing information and local standard of care. Once reduced, the dose could not be increased back to 100%. Treatment with carboplatin or etoposide was recommended to be discontinued if a patient experienced any hematologic or non-hematologic grade 3 or 4 toxicity after 2 dose reductions or a treatment delay of more than 63 days due to toxicities.

Outcomes

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

Efficacy

The primary end points were PFS and OS. PFS was defined as the time from randomization to the first occurrence of disease progression, as determined by the investigator using RECIST 1.1 criteria, or death from any cause, whichever occurred first. OS was defined as the time from randomization to death from any cause.

Key secondary end points included ORR and DOR. ORR was defined as CR or partial response (PR) as determined by the investigator according to RECIST 1.1. Confirmation of responses was not required per protocol, but confirmed response rates were reported. DOR was defined as the time interval from first occurrence of a documented objective response to the time of disease progression, as determined by the investigator using RECIST 1.1, or death from any cause, whichever came first.

Patients underwent tumour assessments at baseline and every 6 weeks (± 7 days) for 48 weeks following day 1 of cycle 1, regardless of treatment dose delays. After completion of the week 48 tumour assessment, tumour assessments were required every 9 weeks (± 7 days), regardless of treatment dose delays. Patients underwent tumour assessments until radiographic disease progression per RECIST 1.1, withdrawal of consent, study termination by the sponsor, or death, whichever occurred first. Patients who continued treatment beyond radiographic disease progression per RECIST 1.1 continued to undergo tumour assessments every 6 weeks (± 7 days), or sooner if symptomatic deterioration occurred. For these patients, tumour assessments continued every 6 weeks (± 7 days), regardless of time in the study, until study treatment was discontinued. Patients who discontinued treatment for reasons other than radiographic disease progression per RECIST 1.1 (e.g., toxicity or symptomatic deterioration) continued scheduled tumour assessments at the same frequency as would have been followed if the patient had remained on study treatment (i.e., every 6 weeks [± 7 days] for 48 weeks following cycle 1, day 1, and every 9 weeks [± 7 days] thereafter, regardless of treatment dose delays) until radiographic disease progression per RECIST 1.1, withdrawal of consent, study termination by the sponsor, or death, whichever occurred first, regardless of whether patients started a new anticancer therapy.

Patient-reported outcomes were evaluated as secondary and exploratory end points and measured using the European Organisation for Research and Treatment of Cancer Quality of Life Questionnaire Core 30 (EORTC QLQ-C30) and the European Organisation for Research and Treatment of Cancer Quality of Life Questionnaire Lung Cancer 13 (EORTC QLQ-LC13).

The EORTC QLQ-C30 instrument is a generic questionnaire consisting of 30 items developed to assess the symptoms and functioning of cancer patients. The instrument includes 5 functional scales, 4 symptom scales, 1 global health status (GHS) scale, and 1 financial impact score. Most items are scored 1 (“not at all”) to 4 (“very much”), except for the items contributing to the GHS scale, which are scored 1 (“very poor”) to 7 (“excellent”). The recall period for each question is “during the past week.” An outcome variable consisting of a score from 0 to 100 is derived for each of the symptom scales and items, according to the EORTC QLQ-C30 instructions. Higher scores on symptoms indicate a worse health state. Higher scores on the GHS and functioning scales indicate better health status or function.

The EORTC QLQ-LC13 is a questionnaire measuring lung cancer symptoms and side effects from conventional chemo- and radiotherapy. It comprises 13 questions assessing lung cancer symptoms (cough, hemoptysis, dyspnea, and site-specific pain), treatment-related side effects (sore mouth, dysphagia, peripheral neuropathy, and alopecia), and pain medication. Except for a multi-item scale for dyspnea, all are single items. An outcome variable consisting of a score from 0 to 100 is derived for each of the symptom scales and items, according to the EORTC QLQ-LC13 instructions. Higher scores on the symptom scales indicate greater symptom burden and therefore a worse health state.

The validity and reliability of both instruments have been established in populations of patients with lung cancer. However, there is a dearth of information with respect to their responsiveness in this setting. A detailed discussion and critical appraisal of the EORTC QLQ-30 and EORTC QLQ-LC13 is provided in Appendix 3.

The EORTC QLQ-C30 and EORTC QLQ-LC13 assessments were completed on day 1 of each 21-day treatment cycle at scheduled study visits during treatment, and at 3 months and 6 months after treatment discontinuation. The instruments, translated into the local language, were completed by patients on an electronic PRO device before administration of study treatment and before any other study assessments.

Time to deterioration in patient-reported lung cancer symptoms was defined as time from randomization to deterioration (10-point change) on each of the EORTC QLQ-C30 and EORTC QLQ-LC13 symptom subscales maintained for 2 assessments or reported at 1 assessment followed by death from any cause within 3 weeks.

Safety

Patient safety was assessed based on reported AEs, clinical laboratory data, vital signs, electrocardiogram, and physical examination. The incidence, nature, and severity of AEs were graded according to the National Cancer Institute Common Terminology Criteria for Adverse Events Version 4.0 and coded using the standard Medical Dictionary for Regulatory Activities version 21.0 terms. AEs of special interest were immune-related AEs, defined based on the mechanism of action of atezolizumab, organized by medical concepts. Patients were assessed for AEs before each dose, and dosing occurred only if the clinical assessment and local laboratory test values were acceptable. The investigators determined whether AEs were related to the trial regimen. AEs were recorded during the study and for up to 30 days after the last dose of the study (90 days for SAEs and AEs of special interest) or until the initiation of new systemic anticancer therapy after the last dose of study treatment, whichever occurred first. After that period, any SAEs or AEs of special interest were reported if they were considered related to prior exposure to study treatment by the investigators. Causality of the AEs was assessed by the investigators.

Statistical Analysis

Sample Size Calculations

The sample size of the trial was determined by the analysis of OS. To detect an improvement of HR equal to 0.68 in OS using a log-rank test, approximately 306 deaths in the ITT population were required to achieve a 91% power at a 2-sided significance level of 0.045. One OS interim analysis was performed when 238 OS events in the ITT population had occurred (data cut-off date: April 24, 2018), which was estimated to occur 25 months after the first patient was randomized, with a 2-sided alpha level of 0.0193 (stopping boundary), computed on the basis of the Lan-DeMets alpha spending function approximating the O’Brien-Fleming boundary.

The OS interim analysis was planned for when approximately 240 OS events in the ITT population had been observed. The pre-specified primary analysis of PFS was planned to be conducted at the time of the OS interim analysis and was estimated to occur when approximately 295 PFS events in the ITT population had occurred, which was expected at approximately 25 months after the first patient was randomized. This provided a 99% power to detect an improvement of HR equal to 0.55 in PFS at a 2-sided significance level of 0.005. The study planned to randomize 400 patients during the global enrolment phase.

The calculation of sample size and estimates of the analysis timelines were based on the following assumptions: PFS and OS are exponentially distributed; the median duration of PFS in the control arm is 6 months; the median duration of OS in the control arm is 10 months; the interim and final analyses of OS use the Lan-DeMets alpha spending function to approximate the O’Brien-Fleming boundary; and there is a dropout rate of 5% over 12 months for PFS and OS.

Analyses of Outcomes

For the primary analyses of PFS and OS, the Kaplan–Meier method was used to estimate median PFS and OS, and the Brookmeyer-Crowley methodology and log-log transformation for normal approximation were used to construct the 95% CIs for the median PFS and median OS in each treatment arm. For the PFS analysis, patients who did not experience disease progression or death at the time of analysis were censored at the time of the last tumour assessment. For the analysis of OS, data for patients who were alive were censored at the time of the last contact. For both PFS and OS analyses, patients with no post-baseline tumour assessment were censored at the date of randomization plus 1 day.

Treatment comparisons were based on the stratified log-rank test (stratified according to sex and ECOG PS score [0 versus 1]). The stratification factor that contained the level with the smallest size was dropped from the stratified analysis if at least 1 stratum had fewer than 10 events. This was pre-specified in the statistical analysis plan. As a result, the stratification factor of presence or absence of brain metastases was removed from the stratified analysis as it contained the level with the lowest number of patients. The HR and corresponding 95% CIs were estimated with the use of a stratified Cox regression model with the same stratification factors that were used in the stratified log-rank test.

To control the overall 2-sided type I error rate at 0.05, a group sequential weighted Holm procedure was used, for which the 2-sided significance levels of 0.005 and 0.045 were allocated to the primary comparisons for PFS and OS, respectively. The test that was significant could pass its alpha level to the test that was not statistically significant at the original allocated alpha level. If PFS in the ITT population was statistically significant at the 2-sided alpha level of 0.005, OS in the ITT population was tested at a 2-sided alpha level of 0.045. Similarly, if OS in the ITT population was statistically significant at the 2-sided alpha level of 0.045, PFS in the ITT population was tested at a 2-sided alpha level of 0.005.

The ORR was estimated using the Clopper-Pearson method for 95% CI of response rates. The 95% CI for the difference in ORRs between the 2 treatment arms was estimated using the normal approximation to the binomial distribution method.

The DOR was assessed for patients who had an objective response as determined by the investigator using RECIST 1.1. Patients whose disease has not progressed and who had not died at the time of analysis were censored at the time of last tumour assessment date. If no tumour assessments were performed after the date of the first occurrence of a CR or PR, DOR was censored at the date of the first occurrence of a CR or PR plus 1 day. Since DOR was based on a nonrandomized subset of patients (i.e., patients who achieved an objective response), formal hypothesis testing was not performed for this end point. Comparisons between treatment arms were made for descriptive purposes. A similar approach to the primary analyses of PFS was used for the DOR analysis.

HRQoL was pre-specified as a secondary and exploratory end point. Descriptive analysis included time to deterioration in lung cancer symptoms, change from baseline summaries, and cumulative distribution function curves of within-patient change from baseline.

All safety end points were reported using descriptive statistics.

Summaries of AEs by treatment group included the following:

• The number and percentage of patients with at least 1 AE by standard of care and preferred term

• AEs by severity (National Cancer Institute Common Terminology Criteria for Adverse Events Version 4.0), presented by standard of care and preferred term

• AEs leading to treatment discontinuation

• AEs related to study drug

• AEs of special interest

  • Immune-related AEs (defined based on the mechanism of action of atezolizumab)

• Infusion-related reactions

Deaths and SAEs were listed and summarized by treatment group. Events of National Cancer Institute Common Terminology Criteria for Adverse Events grade 3 and grade 4 severity were summarized by treatment group. Treatment-emergent AEs leading to permanent treatment discontinuation were listed and summarized by treatment group.

Subgroup Analyses

HRs for OS and PFS were evaluated by pre-planned subgroups according to ECOG PS and presence of brain metastases at baseline, which were subgroups of interest identified in the CADTH systematic review protocol. Between-group treatment effects, with a nominal 95% CI for these end points, were estimated within each category. There was no multiplicity control. As such, all subgroup analyses are exploratory in nature.

Sensitivity Analyses

The potential impact of missing scheduled tumour assessments on the primary analysis of PFS as determined by the investigator was assessed using a PFS event imputation rule: If a patient missed 2 or more assessments scheduled immediately before the date of the PFS event, the patient would be counted as having progressed on the date of the first of these missed assessments, and patients with a PFS event who missed 2 or more scheduled assessments immediately before the PFS event would be censored at the last tumour assessment before the missed visits. The imputation rule would be applied to patients in both treatment arms. Statistical analyses were similar to those used in the primary analysis of PFS. Analyses were also conducted to assess the impact of non-prior anticancer therapy on PFS for patients who switched to other treatment before a PFS event. The Kaplan–Meier method was used to estimate median PFS; 95% CIs for the median were computed using the method of Brookmeyer-Crowley, and HRs were estimated by Cox regression.

Interim Analyses

One interim efficacy analysis of OS was planned for when approximately 240 OS events had been observed. The interim analysis of OS (data cut-off date: April 24, 2018) was conducted when 238 of the planned 240 OS events had been observed.

The primary analysis of PFS was conducted at the time of the interim analysis of OS, and the exact timing of the analysis depended on when 240 OS events in the ITT population had occurred. No interim analysis of PFS was planned.

Updated Analyses

The final OS analysis was planned to be conducted when approximately 306 OS events in the ITT population had been observed, expected to occur at approximately 36 months after the first patient. The updated analysis of OS (data cut-off date: January 24, 2019) was conducted when 302 of the planned 306 OS events had occurred.

Analysis Populations

The ITT population (N = 403) was defined as all the randomized patients, regardless of whether the patient received the assigned treatment. The primary end points were assessed in the ITT population.

The safety population included all treated patients (N = 394), defined as patients who received any amount of any component of study treatment. For the safety analyses, patients who received any amount of atezolizumab were analyzed as part of the atezolizumab arm, even if atezolizumab was given in error.

Results

Patient Disposition

A total of 526 patients were screened; 123 patients failed screening based on information collected in the IxRS system. The most common reasons for screen failure were active or untreated central nervous system metastases (n = 25), withdrawal by patient (n = 13), and lack of evidence of histologically or cytologically confirmed ES-SCLC per the Veterans Administration Lung Study Group staging system (n = 10 patients). Of the 403 patients who were randomly assigned to receive atezolizumab (n = 201) or placebo (n = 202) plus carboplatin and etoposide, 9 did not receive any study treatment (4 patients in the atezolizumab arm and 5 patients in the placebo arm). As of the first clinical cut-off date (April 24, 2018), all 9 untreated patients had discontinued the study due to withdrawal by patient (n = 4), death (n = 4), or physician decision (n = 1). Treatment allocation was unblinded for 4 patients (2 patients in each arm) for safety reasons and for 6 patients (2 patients in the atezolizumab arm and 4 patients in the placebo arm) for other reasons, such as to inform subsequent treatment decisions after disease progression. These were cases of individual patient unblinding that occurred at the site level; the study sponsor continued to remain blinded to the treatment assignment. Patients who were unblinded were included in the analysis populations.

At the time of the primary clinical data cut-off, 124 patients (61.7%) in the atezolizumab arm and 142 (70.3%) in the placebo arm had discontinued the study. Withdrawal of consent was more common in the atezolizumab arm (9.0%) than the placebo arm (4.5%). At the time of the data cut-off date for the updated analysis, 161 patients (80.1%) in the atezolizumab arm and 172 (85.1%) in the placebo arm had discontinued the study. Overall, 14 patients (3.5%) were still on study treatment, and 56 (13.9%) were in follow-up (Table 7).

Treatment Beyond Progressive Disease

Forty-nine patients (24.4%) in the atezolizumab arm were treated beyond investigator-assessed disease progression per RECIST 1.1. The median duration of atezolizumab treatment following investigator-assessed disease progression was 0.7 months (range = 0 to 16). In the atezolizumab arm, 7 of 49 patients (14.3%) treated with atezolizumab beyond progressive disease were still receiving treatment at the time of the first clinical data cut-off date. In total, 85.7% of patients were withdrawn from atezolizumab treatment, mainly due to subsequent progressive disease (79.6%), symptomatic deterioration (4.1%), or withdrawal by patient (2.0%).

Table 7: Patient Disposition

^aData cut-off date: April 24, 2018.

^bData cut-off date: January 24, 2019.

^cOne patient randomized to the placebo arm received atezolizumab and was therefore counted in the atezolizumab arm in the safety population.

Source: Clinical Study Report for IMpower133 study¹²; Liu et al. (2021).³⁰

Protocol Deviations

During the course of the study, 222 protocol deviations were reported for 118 patients in the atezolizumab arm and 104 patients in the placebo arm. Overall, 153 patients (38.0%) had at least 1 protocol deviation (39.3% and 36.6% in the atezolizumab and placebo arms, respectively). Major protocol deviations were described as procedural (e.g., error with stratification, omission of safety labs required by protocol, tumour assessment significantly out of window, or omission of tumour assessment); related to inclusion criteria (e.g., inclusion or exclusion tests not done or out of window, ineligible history or current SCLC stage, or inclusion lab values outside limit); or related to medication (e.g., significant deviation from planned study drug dose or received incorrect study drug or wrong dose) in 32.0% (atezolizumab: 32.3%; placebo: 31.7%), 6.0% (atezolizumab: 7.5%; placebo: 4.5%), and 4.0% (atezolizumab: 3.5%; placebo: 4.5%) of patients.

Exposure to Study Treatments

The safety population included 198 patients who received at least 1 dose of atezolizumab and 196 patients who received placebo. The median duration of treatment with atezolizumab was 4.7 months (range = 0 to 21). Patients received a median of 7 doses of atezolizumab (range = 1 to 30) and 6 doses of placebo. The median number of doses of chemotherapy was the same in the 2 treatment arms (carboplatin: 4 doses; etoposide: 12 doses). Median exposure to carboplatin and etoposide was 2.3 months in the atezolizumab arm and 2.2 months in the placebo arm. The median dose intensity of atezolizumab or placebo was 95% in both arms, while the mean dose intensity of carboplatin and etoposide, respectively, was 92.3% and 89.4% in the atezolizumab arm and 93.3% and 90.3% in the placebo arm. Most patients in the atezolizumab and placebo arms (80% and 90%, respectively) completed the planned 4 cycles of induction treatment (Table 8).

Table 8: Summary of Treatment Exposure — Safety Population

CE = carboplatin and etoposide; SD = standard deviation.

^aDose intensity is the number of doses actually received divided by the expected number of doses.

Source: Clinical Study Report for IMpower133 study.¹²

Prior and Concomitant Treatments

At the time of study enrolment, 20 patients (5.0%) had received non-anthracycline chemotherapy: 8 patients (4.0%) in the atezolizumab arm and 12 patients (5.9%) in the placebo arm. All patients who received prior anticancer therapy had non-anthracycline chemotherapy, with the most common treatment being cisplatin or carboplatin plus etoposide and concurrent radiation. Overall, 14.4% of patients had undergone prior surgery for SCLC, with “other” surgeries (atezolizumab: 13.4%; placebo: 11.4%), metastasectomy (atezolizumab: 1.5%; placebo: 0.5%), lobectomy (atezolizumab: 1%; placebo: 0%), and thoracotomy (atezolizumab: 0.5%; placebo: 0.5%) being the most commonly reported procedures. Prior radiotherapy was reported for 12.4% and 13.9% of patients in the atezolizumab and placebo arms, respectively. A total of 13.2% of patients had received prior radiotherapy for SCLC, including patients in the extensive-stage setting (atezolizumab: 7.0% and placebo: 7.9%, including patients with prior radiotherapy to the brain [atezolizumab: 6.5%; placebo: 6.4%]), in the limited-stage setting (atezolizumab: 5.0%; placebo: 5.4%), and in other settings (atezolizumab: 1.0%; placebo: 1.5%) (Table 9).

At the time of study enrolment and before the randomization date, 90.6% of patients were receiving at least 1 concomitant medication (atezolizumab: 89.6%; placebo: 91.6%). The most commonly used classes of drugs (≥ 25% of patients in any arm) were opioid analgesics (atezolizumab: 29.4%; placebo: 33.2%), statins (atezolizumab: 25.4%; placebo: 27.7%), beta-adrenoceptor blocking agents (atezolizumab: 28.9%; placebo: 22.3%), bronchodilators and anti-asthmatics (atezolizumab: 21.9%; placebo: 28.7%), and proton pump inhibitors (atezolizumab: 21.4%; placebo: 26.7%).

Nearly all patients (97.0%) received at least 1 concomitant medication initiated on or after the randomization date (atezolizumab: 97.5%; placebo: 96.5%). The most commonly used classes of drugs (≥ 25% of patients in any arm) were 5-HT3 antagonists (atezolizumab: 77.6%; placebo: 73.8%), steroids (atezolizumab: 65.7%; placebo: 62.9%), colony-stimulating factors (atezolizumab: 41.8%; placebo: 42.1%), supplements (atezolizumab: 41.3%; placebo: 36.1%), antiemetics not elsewhere classified (atezolizumab: 39.3%; placebo: 35.1%), opioid analgesics (atezolizumab: 33.3%; placebo: 33.7%), nonsteroidal anti-inflammatories (atezolizumab: 24.9%; placebo: 27.2%), laxatives and stool softeners (atezolizumab: 25.4%; placebo: 24.8%), and antihistamines (atezolizumab: 26.4%; placebo: 19.8%).

There were no reported concomitant cancer-related surgeries during the study up to the clinical data cut-off date. No immunosuppressive agents, excluding corticosteroids, started on the day of or after atezolizumab or placebo treatment were reported during the study up to the clinical data cut-off date.

Overall, 7.9% of patients received on-study cancer radiotherapy, with bone (atezolizumab: 2.0%; placebo: 4.0%), brain (excluding PCI) (atezolizumab: 2.0%; placebo: 2.5%), and lung (atezolizumab: 1.5%; placebo: 2.0%) being the most common sites irradiated.

Overall, 10.9% of patients in each treatment arm received PCI during the maintenance phase, For patients who did not receive PCI, the reasons reported were as follows: not considered standard of care (atezolizumab: 52.2%; placebo: 46.5%), progressive disease (atezolizumab: 12.4%; placebo 18.3%), “other” reasons (atezolizumab: 13.4%; placebo: 16.8%), patient preference (atezolizumab: 7.0%; placebo: 3.0%), and clinical deterioration or comorbidities (atezolizumab: 4.0%; placebo: 3.0%).

Table 9: Prior Anticancer Treatments — ITT Population

CE = carboplatin and etoposide; ITT = intention to treat.

Source: Clinical Study Report for IMpower133 study.¹²

Subsequent Anticancer Treatments

In the ITT population, 51.7% of patients in the atezolizumab arm and 57.4% of patients in the placebo arm received at least 1 non-protocol or follow-up non-protocol anticancer therapy. The most common treatment received in the atezolizumab and placebo arms was non-anthracycline chemotherapy: 40.3% and 43.6%, respectively. In the atezolizumab and placebo arms, 3.0% and 7.4% of patients, respectively, received subsequent immunotherapy (Table 10).

Table 10: Treatments — ITT Population

CE = carboplatin and etoposide; ITT = intention to treat.

Note: Multiple cases within a specific line of therapy and regimen for a patient were counted once for the frequency of line of therapy or regimen name.

A patient was counted more than once if they received more than 1 therapy type under each line and regimen.

Source: Clinical Study Report for IMpower133 study.¹²

Efficacy

Progression-Free Survival

As of the data cut-off date (April 24, 2018), 171 patients (85.1%) in the atezolizumab arm and 189 patients (93.6%) in the placebo arm had disease progression or died. The median PFS was 5.2 (95% CI, 4.4 to 5.6) months in the atezolizumab arm and 4.3 (95% CI, 4.2 to 4.5) months in the placebo arm. The stratified HR for disease progression or death was 0.77 (95% CI, 0.62 to 0.96; P = 0.0170) (Table 11 and Figure 3).

The 6-month event-free rate in the atezolizumab and placebo arms was 30.86 (95% CI, 24.26 to 37.45) versus 22.39 (95% CI, 16.56 to 28.22). The 12-month event-free rate was 12.62 (95% CI, 7.85 to 17.40) in the atezolizumab arm and 5.35 (95% CI, 2.14 to 8.56) in the placebo arm (Table 11).

In the 2 subgroups of interest identified for this review (ECOG PS and brain metastases at baseline), while the overall PFS benefit with atezolizumab observed in the overall population was evident in patients with an ECOG PS of 1 and those without brain metastases, for those with an ECOG PS of 0 and those with brain metastases at baseline, the benefit of adding atezolizumab to carboplatin and etoposide in terms of PFS was uncertain (Table 12). The number of patients in these subgroups was low (ECOG PS of 0: n = 128; brain metastases at baseline: n = 33).

Sensitivity Analysis of PFS

The sensitivity analysis of PFS censored for missing scheduled tumour assessments was consistent with the primary analysis. The median PFS was 5.1 (95% CI, 4.3 to 5.6) months in the atezolizumab arm and 4.3 (95% CI, 4.2 to 4.5) months in the placebo arm. The stratified HR for disease progression or death was 0.77 (95% CI, 0.63 to 0.96; P = 0.0209).

The sensitivity analysis of PFS censoring at non-prior anticancer therapy showed similar results to those of the main analysis. The median PFS was 5.2 (95% CI, 4.4 to 5.6) months in the atezolizumab arm and 4.3 (95% CI, 4.2 to 4.5) months in the placebo arm. The stratified HR for disease progression or death was 0.78 (95% CI, 0.63 to 0.96; P = 0.0212).

Table 11: PFS — ITT Population

CE = carboplatin and etoposide; CI = confidence interval; HR = hazard ratio; ITT = intention to treat; PFS = progression-free survival.

^aSince null hypothesis for overall survival was rejected at an overall 2-sided significance level of 0.045, PFS was tested at 2-sided type I error of 0.05.

Source: Clinical Study Report for IMpower133 study.¹²

Figure 3: Kaplan–Meier Curves for PFS — ITT Population

Atezo = atezolizumab; CE = carboplatin and etoposide; CI = confidence interval; ITT = intention to treat; PBO = placebo.

Source: Clinical Study Report for IMpower133 study.¹²

Table 12: PFS by Subgroup at Baseline — ITT Population

CE = carboplatin and etoposide; CI = confidence interval; ECOG PS = Eastern Cooperative Oncology Group Performance Status; HR = hazard ratio; ITT = intention to treat; PFS = progression-free survival.

Note: Medians were estimated using the Kaplan–Meier method. Hazard ratios relative to placebo and the associated CIs were estimated using unstratified Cox regression.

Source: Clinical Study Report for IMpower133 study.¹²

Overall Survival

At the time of the OS interim analysis (data cut-off date: April 24, 2018), patients had a median survival follow-up time of 13.9 months. A total of 104 patients (51.7%) in the atezolizumab arm and 134 patients (66.3%) in the placebo arm had died. The median OS was 12.3 months (95% CI, 10.8 to 15.9) in the atezolizumab arm and 10.3 months (95% CI, 9.3 to 11.3) in the placebo arm. The stratified HR for death was 0.70 (95% CI, 0.54 to 0.91; P = 0.007). The 1-year OS rate was 51.7% in the atezolizumab arm and 38.2% in the placebo arm (Table 13 and Figure 4).

Table 13: OS — ITT Population

CE = carboplatin and etoposide; CI = confidence interval; HR = hazard ratio; ITT = intention to treat; NE = not estimable; OS = overall survival.

^aInterim analysis: OS was tested at 2-sided alpha of 0.0193 (with 238 observed OS events at the data cut-off date of April 24, 2018) to control the overall 2-sided type I error for OS at 0.045 by Lan-DeMets function approximating O’Brien-Fleming boundary.

^bDescriptive purposes only.

Source: Update Clinical Study Report for IMpower133 study.¹³

Updated Analysis

At the final OS analysis (data cut-off date: January 24, 2019), median survival follow-up time was 22.9 months. The median duration of survival follow-up was 23.1 months (range = 0.0 to 29.5) in the atezolizumab arm and 22.6 months (range = 0.0 to 30.7) in the placebo arm.

The median OS was 12.3 months (95% CI, 10.8 to 15.8) in the atezolizumab arm and 10.3 months (95% CI, 9.3 to 11.3) in the placebo arm. The stratified HR for death was 0.75 (95% CI, 0.60 to 0.95; P = 0.015). Based on the landmark analysis, in the atezolizumab versus placebo arms, 34.0% versus 21.0% of patients were alive at 18 months and 22.0% versus 16.8% of patients were alive at 24 months (Table 13 and Table 5).

Of the 2 subgroups of interest identified for this review (ECOG PS and brain metastases at baseline), the OS benefit was consistent for patients with an ECOG PS of 0 and an ECOG PS of 1. With respect to brain metastases, an OS benefit was not observed in patients with brain metastases at baseline but was observed in patients without brain metastases at baseline (Table 14).

Objective Response Rate

The confirmed ORR was 60.2% in the atezolizumab arm and 64.4% in the placebo arm; 2.5% and 1.0% of patients in the atezolizumab and placebo arms, respectively, had a CR (Table 15).

At the updated analysis (data cut-off date: January 24, 2019), the confirmed ORR was 60.2% (95% CI, 53.1 to 67.0) in the atezolizumab arm and 64.4% (95% CI, 57.3 to 71.0) in the placebo arm; 3.5% and 1.0% of patients in the atezolizumab and placebo arms, respectively, had a CR.

Figure 4: Kaplan-Meier Curves for OS — ITT Population

Atezo = atezolizumab; CE = carboplatin and etoposide; CI = confidence interval; ITT = intention to treat; PBO = placebo.

Note: Analyses were stratified by sex (male versus female) and Eastern Cooperative Oncology Group Performance Status (0 versus 1).

Source: Clinical Study Report for IMpower133 study.¹²

Table 14: OS by Subgroup at Baseline — ITT Population

CE = carboplatin and etoposide; CI = confidence interval; ECOG PS = Eastern Cooperative Oncology Group Performance Status; HR = hazard ratio; ITT = intention to treat; OS = overall survival.

Note: Medians were estimated using the Kaplan–Meier method. Hazard ratios relative to placebo and the associated CIs were estimated using unstratified Cox regression.

Source: Clinical Study Report for IMpower133 study¹²; Updated Clinical Study Report for IMpower133 study.¹³

Table 15: Objective Response Rate — ITT Population

CE = carboplatin and etoposide; CI = confidence interval; ITT = intention to treat.

Note: Objective response rate with confirmed response was assessed by investigator per Response Evaluation Criteria in Solid Tumours Version 1.1.

^aClopper-Pearson interval.

^bWald with continuity correction.

^cCochran-Mantel-Haenszel test.

^dStratified by sex (male vs. female) and Eastern Cooperative Oncology Group Performance Status (0 vs. 1).

Source: Clinical Study Report for IMpower133 study.¹²

Duration of Response

The median DOR (confirmed) was 4.2 months (range = 1.4 to 19.5) in the atezolizumab arm and 3.9 months (range = 2.0 to 16.1) in the placebo arm. At data cut-off (April 24, 2018), 14.9% of patients in the atezolizumab arm and 5.4% of patients in the placebo arm had ongoing response. Median time to event (progression) was 4.2 months (95% CI, 4.1 to 4.5) in the atezolizumab arm and 3.9 months (3.1 to 4.2) in the placebo arm (Table 16).

At the updated analysis, the median DOR was 4.2 months (range = 1.4 to 24.3) in the atezolizumab arm and 3.9 months (range = 2.0 to 24.2) in the placebo arm. As of the data cut-off date (January 24, 2019), 9.1% of patients in the atezolizumab arm and 2.3% of patients in the placebo arm had ongoing response.

Figure 5: Kaplan-Meier Curves for OS, Updated Analysis — ITT Population

Atezo = atezolizumab; CE = carboplatin and etoposide; CI = confidence interval; ITT = intention to treat; PBO = placebo.

Note: Analyses were stratified by sex (male versus female) and Eastern Cooperative Oncology Group Performance Status (0 versus 1).

Source: Update Clinical Study Report for IMpower133 study.¹³

Table 16: Duration of Response — ITT Population Patients With Confirmed Response Assessed by Investigator per RECIST 1.1

CE = carboplatin and etoposide; CI = confidence interval; HR = hazard ratio; ITT = intention to treat; RECIST 1.1 = Response Evaluation Criteria in Solid Tumours Version 1.1.

^aKaplan–Meier estimates. The 95% CIs for medians were computed using the Brookmeyer-Crowley method.

^bEstimated by Cox regression.

Source: Clinical Study Report for IMpower133 study.¹²

Time to Next Treatment

Time to next treatment was not an end point in the IMpower133 trial.

Health-Related Quality of Life

At baseline, 175 patients (87%) in the atezolizumab arm and 179 patients (89%) in the placebo arm completed the EORTC QLQ-C30, and 176 (88%) and 168 (83%), respectively, completed the EORTC QLQ-LC13. Completion rates remained above 80% up to week 36 in the atezolizumab arm and up to week 24 in the placebo arm. At week 54, 34 (8%) of the 403 randomized patients remained on study treatment and were eligible to complete PRO assessment.

Patients in the IMpower133 study generally reported worse disease-related symptoms (cough, chest pain, dyspnea, arm/shoulder pain, or pain in other parts) at baseline than are reported in the normative scores of patients with SCLC.³¹ Changes from baseline in treatment-related symptoms, including diarrhea, dysphagia, sore mouth, peripheral neuropathy, nausea/vomiting, and insomnia, were generally similar between treatment arms at most visits through week 54.³¹

Mean change from baseline in function scores (physical, cognitive, emotional, social, and role) were similar in both treatment arms through week 54, with a general trend for improvement of function (physical, emotional, and social) or maintenance of function (role and cognitive) (Figure 6).

Harms

Adverse Events

The majority of patients in both treatment arms — 100% in the atezolizumab arm and 96.4% in the placebo arm — experienced at least 1 AE of any grade (Table 17). In the atezolizumab arm, the most common AEs of any grade by preferred term experienced by at least 10% of patients were anemia (43.4%), nausea (37.9%), and neutropenia (37.4%). In the placebo arm, the most common AEs of any grade by preferred term experienced by at least 10% of patients were anemia (35.2%), neutropenia (35.2%), and alopecia (34.7%) (Table 18).

Grade 3 or 4 AEs, regardless of attribution, occurred in 67.7% of patients in the atezolizumab arm and 63.3% of patients in the placebo arm. The most common grade 3 or 4 AEs reported in at least 5% of patients in any treatment arm were neutropenia (22.7% versus 25.0% in the atezolizumab and placebo arms, respectively), decreased neutrophil count (15.7% versus 16.8%), anemia (15.7% versus 13.3%), thrombocytopenia (10.1% versus 8.7%), and hyponatremia (4.5% versus 6.6%).

Serious Adverse Events

In the atezolizumab arm, 37.4% of patients had at least 1 SAE. In the placebo arm, 34.7% of patients experienced at least 1 SAE (38.9% and 35.2% as of Jan 24, 2019 cut-off date). The most common SAEs experienced by at least 1% of patients in either the atezolizumab or placebo arm were pneumonia (4.5% versus 3.6%), neutropenia (3.5% versus 4.1%), febrile neutropenia (2.5% versus 4.6%), and thrombocytopenia (2.5% versus 2.0%) (Table 19).

Figure 6: Change From Baseline Through Week 54 in Function and in Health-Related Quality of Life

CI = confidence interval; CP = carboplatin; ET = etoposide; mOS = mean overall survival; mPFS = mean progression-free survival.

Source: Clinical Summary for IMpower133 study.¹²

Withdrawals Due to AEs

Withdrawal from any study treatment due to AEs was reported for 12.1% of patients in the atezolizumab arm and 3.1% of patients in the placebo arm. In the atezolizumab arm, 11.6% of patients experienced AEs leading to discontinuation of atezolizumab, and in the placebo arm, 2.6% of patients had AEs leading to discontinuation of placebo. The main reasons for permanently discontinuing atezolizumab in 21 patients in the atezolizumab arm were infusion-related reactions and gastrointestinal disorders.

In the atezolizumab arm, 70.2% of patients had AEs resulting in dose modification or interruption. In the placebo arm, 60.7% of patients had dose modification or interruption to their treatment due to AEs. Neutropenia was the most common AE leading to dose modification or interruption of any study treatment in both treatment arms. The most frequent AEs leading to dose modification or interruption of any study treatment reported in at least 2% patients were neutropenia, anemia, leukopenia, fatigue, pyrexia, and increased alanine aminotransferase incidence in the atezolizumab arm, whereas those more common in the placebo arm were decreased neutrophil, platelet, and white blood cell counts.

Table 17: Summary of AEs — Safety Population

AE = adverse event; CE = carboplatin and etoposide; SAE = serious adverse event.

Note: Clinical data cut-off date: January 24, 2019.

^aAn event consistent with an immune-mediated mechanism of action requiring treatment with systemic corticosteroids or hormone replacement therapy.

^bIncidence of treatment-related AEs and AEs leading to withdrawal from any treatment are for any treatment component.

Source: Clinical Summary for IMpower133 study¹²; Liu et al. (2021)³⁰; Reck et al. (2019).³²

Table 18: AEs by Preferred Term Occurring in at Least 10% of Patients in Either Treatment Arm — Safety Population

AE = adverse event; CE = carboplatin and etoposide; MedDRA = Medical Dictionary for Regulatory Activities.

Source: Clinical Study Report for IMpower133 study.¹²

Table 19: Serious Adverse Events by Preferred Term Occurring in at Least 1% of Patients in Either Treatment Arm — Safety Population

CE = carboplatin and etoposide; MedDRA = Medical Dictionary for Regulatory Activities; SAE = serious adverse event.

Note: Grade 5 adverse events due to progressive disease are excluded.

Source: Clinical Study Report for IMpower133 study.¹²

Death

As of the primary data cut-off date, 103 deaths had occurred in the atezolizumab arm (52.0% of patients), and 130 deaths had occurred in the placebo arm (66.3% of patients). The most common cause of death in both arms was progressive disease, which accounted for 87.4% of deaths in the atezolizumab arm and 88.5% of deaths in the placebo arm (Table 20).

Grade 5 AEs occurred in 4 patients (2.0%) in the atezolizumab arm, and included pneumonia, respiratory failure, death, and neutropenia. Grade 5 fatal AEs occurred in 11 patients (5.6%) in the placebo arm and included pneumonia, pulmonary sepsis, sepsis, septic shock, acute respiratory failure, hemoptysis, cardiopulmonary failure, pericardial effusion, and general physical health deterioration. The only grade 5 AE (by preferred term) that occurred in more than 1 patient was pneumonia (1 patient in the atezolizumab arm and 3 patients in the placebo arm).

Table 20: All Deaths and Primary Causes of Death — Safety Population

CE = carboplatin and etoposide.

Note: “Other” primary causes of death include unrelated adverse events outside of reporting window.

Source: Clinical Study Report for IMpower133 study.¹²

Of the grade 5 events, 3 events in each arm were considered related to at least 1 component of study treatment. In the atezolizumab arm, 1 grade 5 death was considered related to all study treatment; there was also 1 case of grade 5 pneumonia and 1 case of grade 5 neutropenia that were both considered related to both carboplatin and etoposide. In the placebo arm, 1 case of grade 5 septic shock was considered related to all study treatment, 1 case of grade 5 pneumonia was considered related to placebo, and 1 case of grade 5 cardiopulmonary failure was considered related to carboplatin.

Notable Harms

Immune-Related AEs

Immune-related AEs were reported for 41.4% of patients in the atezolizumab arm and 24.5% of patients in the placebo arm (Table 21). Rash (both treatment arms) and hypothyroidism (atezolizumab arm) were the most common (≥ 10% incidence) and most differentially reported (≥ 5% difference between treatment arms) immune-related AEs during treatment. More patients in the atezolizumab arm experienced immune-related rash than in the placebo arm (20.2% versus 10.7%). The majority of rash AEs were grade 1 or 2 in severity; 2.0% of patients in the atezolizumab arm and no patients in the placebo arm experienced a grade 3 or 4 rash AE. One patient in each arm experienced an immune-relate rash that was experienced as serious. One patient in the atezolizumab arm had erythema, which led to atezolizumab withdrawal. Rash maculo-papular led to treatment modification or interruption in 2 patients (1.0%) in the atezolizumab arm and no patients in the placebo arm. Immune-related rash that required systemic corticosteroid treatment occurred in 2 patients (1.0%) in the atezolizumab arm and no patients in the placebo arm.

Immune-related hypothyroidism was reported for 12.6% of patients in the atezolizumab arm and 0.5% of patients in the placebo arm. The proportion of patients with immune-related hyperthyroidism in the atezolizumab and placebo arms was 5.6% and 2.6%, respectively. All hyperthyroidism AEs were grade 1 or 2 in severity, and none led to study treatment withdrawal. Hyperthyroidism led to treatment modification or interruption in 1 patient in each arm (0.5%).

Table 21: Adverse Events of Special Interest

AE = adverse event; CE = carboplatin and etoposide.

Note: Clinical data cut-off date: January 24, 2019.

^aAn event consistent with an immune-mediated mechanism of action, not taking into account whether treatment was given for the event.

Source: Reck et al. (2019).³²

Infusion-Related Reactions

Immune-related infusion-related reaction events were experienced by 5.6% of patients (n = 11) in the atezolizumab arm and 5.1% of patients (n = 10) in the placebo arm. The majority of these events were grade 1 or 2 (atezolizumab arm: n = 7 [3.5%]; placebo arm: n = 9 [4.6%]). Four patients (2.0%) in the atezolizumab arm and 1 patient (0.5%) in the placebo arm had grade 3 or 4 infusion-related reactions. One patient (0.5%) in the atezolizumab arm and 2 patients (1.0%) in the placebo arm experienced infusion-related reactions that were reported as serious. Five patients (2.5%) in the atezolizumab arm (versus none in the placebo arm) had an infusion-related reaction that led to withdrawal of any study treatment. Infusion-related reactions led to treatment modification or interruption in 7 patients (3.5%) in the atezolizumab arm and 6 patients (3.1%) in the placebo arm. An infusion-related reaction that required systemic corticosteroid treatment occurred in 5 patients (2.5%) in the atezolizumab arm and 3 patients (1.5%) in the placebo arm.

Critical Appraisal

Internal Validity

The baseline demographic and disease characteristics across treatment arms were roughly balanced between the 2 treatment arms. Protocol deviations were reported in approximately 30% to 40% of patients across the 2 treatment arms. Although the overall proportion of protocol deviations was comparable between the 2 treatment arms, the impact on outcome assessments remains. Concomitant anticancer therapies, including all second-, third-, and fourth-line therapies, were higher (roughly 12% in total) in the placebo arm than in the atezolizumab arm, which may have led to biased estimates of treatment effect against atezolizumab.

Response outcomes (ORR and DOR) were assessed by investigators per RECIST 1.1. While the trial was double blinded and the investigators were blinded to treatment assignment, risk of bias cannot be ruled out. For example, nearly half the patients in the atezolizumab arm experienced immune-related AEs or other events. These events may have made the investigator aware of the patient’s treatment assignment. Therefore, for all investigator-assessed outcomes a certain degree of subjectivity may exist, which may have biased the results. In addition, although the proportion of patients receiving concomitant and supportive care for symptom control was largely similar in the 2 treatment arms, which may have led to comparable PROs including QoL outcomes as observed in the trial, this may not mean that the 2 trial regimens truly have comparable safety and impact on QoL.

Interim and final analyses were planned a priori and adequately described. The interim analysis applied Lan-DeMets alpha spending function with the O’Brien-Fleming stopping boundary, which is deemed conservative in controlling type I error when claiming a treatment effect based on interim analysis. The updated final analysis results of OS were consistent with the interim analysis results.

Treatment of atezolizumab continued until disease progression per RECIST 1.1, treatment discontinuation or interruption due to AEs, and early withdrawals; immune-related AEs were particularly disproportional between treatment arms. Patients could be considered for treatment beyond radiographic disease progression if they had evidence of clinical benefit. During the maintenance phase, concomitant treatments including PCI and palliative thoracic radiation were permitted per local standard of care and approved indications. While treatment beyond progression would not affect PFS, it might impact OS estimates. The clinical experts consulted by CADTH noted that the concept of treatment beyond progression for select patients who show evidence of ongoing benefit with treatment has broad acceptance among oncologists in clinical practice; however, the IMpower133 trial was not designed to evaluate the effect of treatment beyond progression.

External Validity

The patient population in the IMpower133 study generally reflects patients in Canadian clinical practice in this setting. However, some patient groups were not represented, including those with an ECOG PS of 2 and patients with active untreated metastases. The clinical experts consulted by CADTH for this review indicated that patients with an ECOG PS of 2 should not be excluded from treatment with combination immunotherapy and chemotherapy, including atezolizumab plus carboplatin and etoposide. The exclusion of these patients from the IMpower133 trial limits the generalizability of clinical evidence with respect to the efficacy and safety of atezolizumab in these groups of patients. The proportion of patients with brain metastases (9%) was lower than that observed in clinical practice (10% to 20%), but this is likely due to the specific inclusion requirements for these patients (e.g., only supratentorial and cerebellar metastases, and no ongoing requirement for corticosteroids as therapy for central nervous system disease). Due to the small number of patients in some subgroups, including brain metastases at baseline, subgroup analyses failed to demonstrate similar effects in patients with brain metastases as in patients free of brain metastases. In addition, the study recruited patients mainly from Asia and Europe; no patients from Canada were recruited. However, the clinical experts consulted by CADTH noted that the lower-than-expected proportion of trial patients with brain metastases and the lack of representation of Canadian patients does not reduce the generalizability of the results to Canadian clinical practice.

The comparator in the IMpower133 trial (carboplatin and etoposide) is relevant to the Canadian context as platinum (carboplatin or cisplatin) and etoposide chemotherapy is the current standard of care. The clinical experts consulted by CADTH noted that in Canadian clinical practice, approximately 70% to 75% of patients with ES-SCLC receive carboplatin and etoposide. Although cisplatin is generally considered more effective, carboplatin is frequently substituted for cisplatin in clinical practice as it reduces the risk of toxicity, including nausea and vomiting, neuropathy, neutropenia, and infection. The clinical experts indicated that although cisplatin was not included in the IMpower133 trial, atezolizumab may be added to cisplatin and etoposide, as well as to carboplatin and etoposide. Thoracic radiation, a treatment used in some patients with ES-SCLC in clinical practice, was not allowed in the trial. The clinical experts noted that while up to 1-quarter of patients with ES-SCLC might have thoracic radiation, including for palliative reasons, the evidence to support the addition of thoracic radiation to systemic therapy in SCLC is inconclusive.³³

Indirect Evidence

Objectives and Methods for the Summary of Indirect Evidence

The IMpower133 study compared atezolizumab plus carboplatin and etoposide against placebo plus carboplatin and etoposide in patients with ES-SCLC. No direct comparative evidence was submitted that compared atezolizumab plus carboplatin and etoposide with other current standard of care treatments for ES-SCLC in Canada. The clinical experts consulted by CADTH identified carboplatin plus etoposide or cisplatin plus etoposide as standard of care therapies currently funded in Canada. In addition, the clinical experts noted that durvalumab plus etoposide and carboplatin (or cisplatin) was also a suitable comparator, although the combination may not currently be available through public drug plans in all Canadian jurisdictions. The objective of this section is to summarize and critically appraise available indirect evidence comparing atezolizumab plus carboplatin and etoposide against other relevant treatments for ES-SCLC.

In the absence of a head-to-head RCT comparing atezolizumab plus carboplatin and etoposide with a standard of care regimen, the sponsor submitted an ITC in the form of an NMA, which provides comparative evidence of the efficacy of atezolizumab plus carboplatin and etoposide against several comparators.

A supplemental literature search was conducted by CADTH to retrieve published ITCs for atezolizumab. A focused literature search for NMAs including atezolizumab as a treatment option for SCLC was run in MEDLINE All (1946–) on February 18, 2022. No search limits were applied. Nine ITCs were retrieved from the CADTH literature search.

Description of the Sponsor-Submitted NMA

Objectives

The primary objective of the sponsor-submitted NMA was to compare atezolizumab in combination with etoposide plus a platinum-based chemotherapy for the first-line treatment of ES-SCLC against relevant platinum doublet therapies and immunotherapies used in clinical practice.¹⁴

Study Selection Methods

The sponsor conducted a systematic review to select studies based on pre-specified PICO criteria, as outlined in Table 22. The systematic review including a feasibility assessment, ||| ||||||||| ||||||||| || |||| || ||||| ||| ||||||| || ||| ||||. The systematic review was restricted to phase II, III, and IV RCTs with active or placebo or best supportive care as controls, with no restriction on blinding, that were conducted in adult patients (≥ 18 years) with histologically or cytologically confirmed ES-SCLC and no prior systemic treatment for ES-SCLC. The sponsor considered different comparators in the systematic review, as presented in Table 22. The primary focus was to retrieve English-language publications or non–English-language publications with an English abstract. A grey literature search was conducted in congress proceedings and other available grey literature sites. Citations were screened by 2 independent analysts, and a second analyst extracted data. Risk of bias assessments of the studies included were conducted based on the 7-criteria checklist by the NICE single technology appraisal user guide.

Table 22: Study Selection Criteria and Methods for Sponsor-Submitted NMA

ES-SCLC = extensive-stage small cell lung cancer; HRQoL = health-related quality of life; NMA = network meta-analysis; RCT = randomized controlled trial.

^aThe immunotherapies were not considered interventions of interest for the original search. However, relevant search terms were included in the systematic review update search strategy (without a related date restriction).

Source: Sponsor-submitted NMA.¹⁴

The sponsor-submitted systematic review and feasibility assessment¹⁴ included a best-case evidence network (irrespective of inter-trial heterogeneity or reported outcomes) that considered platinum-based combination chemotherapy regimens and immunotherapy treatments. As the comparator arms of the 3 immunotherapy trials (the CASPIAN, ECOG-ACRIN EA5161, and KEYNOTE-604 trials) included etoposide administered in combination with cisplatin or carboplatin, a grouped node was required for inclusion of these trials in the network, irrespective of which chemotherapy was administered. The Skarlos 1994 and Okamoto 2007 trials, which compared etoposide plus carboplatin against etoposide plus cisplatin, were also kept in the best-case scenario network. In the feasibility assessment, outcome-specific evidence networks were explored for all primary and secondary outcomes of the IMpower133 trial. Connected networks were obtained for OS, PFS, ORR, and incidence of SAEs. However, an NMA was found not to be feasible for DOR, time in response, time to deterioration, and incidence of grade 3 to 5 AEs or treatment-related SAEs.

ITC Analysis Methods

Table 23 presents the ITC methods used in the sponsor-submitted NMA.¹⁴

Table 23: Sponsor-Submitted ITC Analysis Methods

FE = fixed effect; GLM = generalized linear model; ITC = indirect treatment comparison; MCMC = Markov chain Monte Carlo; NMA = network meta-analysis; ORR = objective response rate; OS = overall survival; PFS = progression-free survival; RE = random effect; SAE = serious adverse event.

Source: Sponsor-submitted NMA.¹⁴

The sponsor-submitted NMA methodology used a Bayesian approach. Both FE and RE models were explored in the base-case scenarios to fit the model. All findings presented by the sponsor were based on the FE model, which assumes homogeneity of the underlying true treatment effects. The sponsor noted that the FE model gave a better fit to the data in terms of residual deviance (closer to the number of data points) in all base-case analyses than the RE model did. The sponsor highlighted that the results of the RE model were associated with high levels of uncertainty due to insufficient power in the analyses to accurately estimate between-study SD and that the relative treatment effects of treatment comparisons were not aligned with the trial-level estimates. Vague priors with a mean of 1 and variance of 100,² as recommended by Dias (2011), were used to estimate the treatment effect sizes. The RE model constructed stated that uniformly distributed prior distributions, which ran between 0 and 5 as recommended by Dias 2011, should be used.

The sponsor included 1 3-armed trial (CASPIAN) in the network analyses of OS and PFS. Adjustments were made to account for the potential likelihood of correlation between treatments in trial-level data. According to the sponsor, the variance of the log-hazard for the baseline treatments in the 3-armed trial was unknown, and approximations were made based on the variances of the differences using methods described by Woods (2010).

Markov chain Monte Carlo simulations with 2 chains starting from different initial values of selected unknown parameters were used to construct the models in estimating the relative treatment effects. The binomial likelihood, logit link model was considered appropriate for the NMA by the sponsor. Model convergence for each model constructed was assessed by analyzing history and density plots and Brooks-Gelman-Rubin diagnostic plots. In addition, autocorrelation plots were assessed to detect the presence of autocorrelation in the chains. Following the model convergence, inferences were made from data obtained by sampling 20,000 iterations.

Heterogeneity was assessed across the trials included in the sponsor-submitted network by conducting subgroup analyses for selected outcomes. There were insufficient studies in the sponsor’s network to conduct a meta-regression. Scenario analyses were conducted for OS (with the removal of the Hermes 2008 trial from the base-case analysis) as the sponsor considered some studies as potential outliers due to differences in ECOG PS in the patients recruited.

The transitivity assumption was assessed by evaluating the effect of potential treatment effect modifiers across studies that may impact the ITC results. All trials reported a median age of patients ranging from 60 years to 74 years, with an overall median of 65 years across the trials. The Okamoto 2007 trial was conducted in an elderly, high-risk population and included patients with ages ranging from 55 years to 86 years (92% of patients were ≥ 70 years). In total, 54% of patients enrolled in the IMpower133 trial were younger than 65 years. Most patients recruited across the trials included in the sponsor’s network were male (mean = 68%; range = 44% to 91%). Baseline data on ethnicity were only available in the IMpower133 and CASPIAN trials, and the majority of patients were White or Asian. Variability was observed across and within trial arms for gender. Across the trials included in the network, the proportion of patients with an ECOG PS of 0 to 1 ranged from 52% to 100%, with a mean of 85%. In 3 studies that permitted the enrolment of patients with ECOG PS greater than 1 (Okamoto 2007, Hermes 2008, and Skarlos 1994), the majority of patients were classified as having an ECOG PS of 2, with only a small proportion reported as having an ECOG PS of 3. The sponsor observed variability in ECOG PS across and within trials investigating non-immunotherapy regimens. There were inconsistencies in the reporting of the number and type of metastatic sites across trials in the network, and the most common were brain and liver metastases. IMpower133 included a smaller proportion of patients with brain or liver metastases than the other trials in the network. According to the clinical experts consulted, it is not anticipated that smoking status will be a contributing effect modifier given that smoking is an established risk factor associated with SCLC.

Intervention nodes were connected to the network diagram through etoposide plus cisplatin or carboplatin. The sponsor included a grouped node within the network for 3 trials (CASPIAN, ECOG-ACRIN EAS161, and KEYNOTE-604) that required an assumption of equivalence of carboplatin and cisplatin.

Sensitivity analyses were conducted for the outcomes investigated. Analyses using the unadjusted HRs were conducted by the sponsor given that both adjusted (or stratified) and unadjusted (or unstratified) HRs were reported for some trials. The removal of Okamoto 2007 was suggested as a scenario for sensitivity analysis as the trial was considered a potential outlier.

The risk of bias was generally considered low or unclear across the studies for the remaining elements of bias assessment. The studies appeared sufficiently homogenous to combine in analyses.

Treatments

Indirect comparisons for atezolizumab plus carboplatin and etoposide were performed for the following interventions: irinotecan plus carboplatin, etoposide plus carboplatin, etoposide plus cisplatin, durvalumab plus etoposide and carboplatin (or cisplatin), durvalumab plus tremelimumab plus etoposide and carboplatin (or cisplatin), and nivolumab plus etoposide and carboplatin (or cisplatin).

Only comparisons between atezolizumab plus carboplatin and etoposide versus etoposide plus carboplatin, etoposide plus cisplatin, durvalumab plus etoposide and carboplatin (or cisplatin) were considered for this CADTH review.

End Points

The sponsor created several outcome-specific networks for OS, PFS, ORR, and SAEs.

Results of Sponsor-Submitted NMA

Summary of Included Studies

Ninety-one publications were identified by the sponsor, and a feasibility assessment was conducted to determine a best-case evidence network. Eight trials (IMpower133,¹¹ Hermes 2008,³⁴ Schmittel 2011,³⁵ CASPIAN,²⁵ KEYNOTE-604,²⁶ Skarlos 1994,³⁶ Okamoto 2007,³⁷ and ECOG-ACRIN EA5161³⁸) formed a connected network. There were notable differences observed in the treatment pathway between IMpower133 and the “non-immunotherapy” trials of the best-case network (Skarlos 1994, Okamoto 2007, Hermes 2008, and Schmittel 2011). The CASPIAN, KEYNOTE-604, and ECOG-ACRIN EA5161 trials had a similar design to the IMpower133 trial in that randomized induction therapy was followed by maintenance therapy with placebo or immunotherapy.

Etoposide plus carboplatin or cisplatin was consistent as a comparator across the trials included in the network. Dosing regimens differed across some trials.

In 3 trials (CASPIAN, Schmittel 2011, and Skarlos 1994), dosing of etoposide plus carboplatin or cisplatin was performed for up to 6 cycles, and in 5 trials (including the ECOG-ACRIN EA5161 and KEYNOTE-604 trials, Hermes 2008, and Okamoto 2007) dosing of the comparator was for up to 4 cycles. The primary outcome was stated for all trials except Skarlos 1994. Table 24 summarizes the assessment of homogeneity within the sponsor-submitted NMA.

A summary of study design and patient baseline characteristics of the included trials is presented in Appendix 4. Of the 8 studies included in the sponsor’s network, 7 included patients 18 years and older with ES-SCLC. In Skarlos 1994, the authors recruited previously untreated SCLC patients, and a subgroup with ES-SCLC was available, which provided data for the analyses. All trials reported a median age of patients ranging from 60 years to 74 years, with an overall median of 65 years across the trials. The proportion of patients with an ECOG PS of 0 to 1 ranged from 52% to 100%, with a mean of 85% across the 8 trials included.

Table 24: Assessment of Homogeneity for Sponsor-Submitted NMA

AE = adverse event; ECOG PS = Eastern Cooperative Oncology Group Performance Status; ES-SCLC = extensive-stage small cell lung cancer; NMA = network meta-analysis; ORR = objective response rate; OS = overall survival; PFS = progression-free survival.

Source: Sponsor-submitted NMA.¹⁴

Results

Only results of the FE model are presented in this section.

Progression-Free Survival

The sponsor conducted 3 scenario analyses: a base-case scenario (which included adjusted or stratified HRs reported across the trials); scenario 1 (which included the unadjusted or unstratified HRs reported); and scenario 2 (which consisted of an analysis where etoposide plus cisplatin, and etoposide plus carboplatin, were considered as distinct nodes).

Of the 8 trials selected, PFS data were available in 5 (CASPIAN, ECOG-ACRIN EA5161, KEYNOTE-604, Schmittel 2011, and IMpower133); these data were used to inform the base-case and scenario 1 network structure. Figure 7 presents the overall network diagram for the base case and scenario 1.

In the base-case and scenario analyses, atezolizumab plus carboplatin and etoposide was associated with longer PFS than carboplatin or cisplatin plus etoposide, and the comparisons were considered to be statistically meaningful as the 95% CrIs excluded the null value (refer to Table 25).

In the base-case and scenario analyses of the comparison of PFS for atezolizumab plus carboplatin and etoposide against durvalumab plus carboplatin (or cisplatin) and etoposide, no statistically significant difference was observed based on the CrI intervals, which included the null value, and the point estimates were close to null.

Figure 7: Redacted

Confidential figure redacted at the request of the sponsor.

Source: Sponsor-submitted network meta-analysis.¹⁴

Table 25: Progression-Free Survival NMA Results for the Comparison of Atezolizumab Plus Carboplatin and Etoposide Versus Relevant Comparators — FE Model

CrI = credible interval; FE = fixed effect; HR = hazard ratio; NA = not applicable; NMA = network meta-analysis; PFS = progression-free survival.

Note: The 95% CrIs for atezolizumab plus carboplatin and etoposide against carboplatin (or cisplatin) and etoposide plus durvalumab included the null value.

^aBase case: ||| |||||||| |||||||| |||| ||| ||||| ||| |||| || |||||||||| |||| ||| |||| |||||||

^bScenario 1: ||| |||||||| |||||||| |||| ||| ||||| ||| |||| || |||||||||| |||| ||| |||| |||||||

^cScenario 2: ||| |||||||| |||||||| |||| ||| ||||| ||| |||| || |||||||||| |||| ||||| |||| ||||||

Source: Sponsor-submitted NMA.¹⁴

Overall Survival

Seven studies (the CASPIAN, ECOG-ACRIN EA5161, KEYNOTE-604, and Impower133 trials; Hermes 2008; Okamoto 2007; and Schmittel 2011) were used to inform the network for OS. Four scenario analyses were conducted: a base case (adjusted or stratified HRs), scenario 1 (unadjusted or unstratified HRs), scenario 2 (Hermes 2008 trial excluded), and scenario 3 (etoposide plus cisplatin, and etoposide plus carboplatin, considered as distinct nodes). The evidence network for the base-case scenario is presented in Figure 8, and Table 26 presents the summary of the NMA results (HRs with their corresponding 95% CrIs) for the comparison of atezolizumab plus etoposide and carboplatin versus comparators from each scenario.

In all scenarios assessed, atezolizumab plus carboplatin and etoposide was favoured in terms of OS when compared with etoposide plus carboplatin or cisplatin (|||| ||||| ||| |||| |||| ||||| |||| || ||||); |||||||| || ||||||| |||| |||| |||| || ||||| and the 95% CrIs excluded the null value (refer to Table 26). The HRs presented in the base-case and scenario analyses of atezolizumab plus carboplatin and etoposide against durvalumab plus carboplatin (or cisplatin) and etoposide (|||| |||| ||||||||| ||||||| |||| |||| |||| || ||||) showed no statistically significant difference between the 2 combinations based on the CrI which included the null value, and the point estimates were close to the null.

Figure 8: Redacted

Confidential figure redacted at the request of the sponsor.

Source: Sponsor-submitted network meta-analysis.¹⁴

Table 26: Overall Survival NMA Results for the Comparison of Atezolizumab Plus Carboplatin and Etoposide Versus Relevant Comparators – FE Model

CrI = credible interval; FE = fixed effect; HR = hazard ratio; NMA = network meta-analysis.

Note: The 95% CrIs for comparison of atezolizumab plus carboplatin and cisplatin against etoposide plus cisplatin and etoposide and carboplatin (or cisplatin) plus durvalumab include the null value.

^aScenario 1: ||| |||||||| |||||||| |||| ||| ||||| ||| |||| || |||||||||| |||| ||||| |||| |||||||

^bScenario 2: ||| |||||||| |||||||| |||| ||| ||||| ||| |||| || |||||||||| |||| ||||| |||| ||||||

^cScenario 3: ||| |||||||| |||||||| |||| ||| ||||| ||| |||| || |||||||||| |||| ||||| |||| ||||||

Source: Sponsor-submitted NMA.¹⁴

Objective Response Rate

Eight studies (the CASPIAN, ECOG-ACRIN EA5161, KEYNOTE-604, Okamoto 2007, Schmittel 2011, Hermes 2008, Skarlos 1994, and IMpower133 trials) were used to inform the network for ORR. The residual deviance from the model was 14.88, in comparison with 15 data points. Figure 9 presents the evidence network for ORR. The results of the analysis of ORR are presented in Table 27.

The OR reported for the comparison of atezolizumab plus carboplatin and etoposide against etoposide plus carboplatin or cisplatin was |||| |||| |||| |||| || ||||, and the OR for the comparison of the atezolizumab combination against etoposide plus cisplatin was |||| |||| |||| |||| || ||||. The comparison of atezolizumab plus carboplatin and etoposide against durvalumab plus carboplatin (or cisplatin) and etoposide was associated with a lower ORR (|||||||| ||| |||| |||| || ||||), which indicated that the atezolizumab combination may be associated with a statistically meaningful improvement in ORR as the 95% CrI did not include the null value, although the CrI was wide.

Figure 9: Redacted

Confidential figure redacted at the request of the sponsor.

Source: Sponsor-submitted network meta-analysis.¹⁴

Table 27: Objective Response Rate NMA Results for the Comparison of Atezolizumab Plus Carboplatin and Etoposide Versus Relevant Comparators — FE Model

CrI = credible interval; FE = fixed effect; NMA = network meta-analysis; OR = odds ratio.

Note: The ORs (95% CrIs) are presented for treatment A (row) vs. treatment B (column).

Source: Sponsor-submitted NMA.¹⁴

Serious Adverse Events

Data from 2 studies (CASPIAN and IMpower133) were used to populate the SAE network.

Figure 10 provides the evidence network for SAEs, and Table 28 provides the ORs and corresponding 95% CrIs for the analysis. The OR was estimated to be |||| |||| |||| |||| || ||||| for the comparison between atezolizumab plus carboplatin and etoposide against durvalumab plus carboplatin (or cisplatin) and etoposide, and |||| |||| |||| |||| || ||||| for the comparison of the atezolizumab combination against etoposide plus cisplatin. The CrIs for both comparisons included the null value.

Figure 10: Redacted

Confidential figure redacted at the request of the sponsor.

Source: Sponsor-submitted network meta-analysis.¹⁴

Table 28: Serious Adverse Event NMA Results for the Comparison of Atezolizumab Plus Carboplatin and Etoposide Versus Relevant Comparators — FE Model

CrI = credible interval; FE = fixed effect; NMA = network meta-analysis; OR = odds ratio.

Note: The ORs (95% CrIs) are presented for treatment A (row) vs. treatment B (column).

Source: Sponsor-submitted NMA.¹⁴

Critical Appraisal of Sponsor-Submitted NMA

The sponsor’s systematic review methods for identifying and assessing studies included in the network were considered appropriate for identifying relevant studies. The PICO criteria were pre-specified, and articles were reviewed by 2 independent reviewers, while a second analyst extracted data. All relevant comparators identified in the CADTH review protocol that were considered relevant to the Canadian practice context were presented in the sponsor’s NMA. The outcomes presented in the trials included in the network analysis were considered relevant and clinically meaningful by the clinician experts consulted during the CADTH review. The population studied in all 8 trials was considered relevant for the reimbursement request. Most studies included untreated patients with ES-SCLC. One study (Skarlos 1994) recruited a different population in the trial but had a subgroup of patients with ES-SCLC. Information from the subgroup analysis was used to inform the network. Quality assessments were conducted using the validated 7-criteria checklist provided by the NICE single technology appraisal user guide.

A generalized linear regression model with a binomial likelihood, logit link model was used; the model was considered appropriate for the types of outcomes assessed in the network. The sponsor explored both FE and RE models in its base-case scenarios, and results from the FE model were presented. The sponsor provided a justified rationale for using the FE model over the RE model based on the model fit criteria, including a judgment on the similarities of the studies included in terms of effect modifiers.

The transitivity assumption was assessed by evaluating potential effect modifiers. There was considerable heterogeneity across trials, particularly in terms of ECOG PS. The Hermes 2008 trial enrolled less than 53% of patients with an ECOG PS of 0 or 1 in both treatment arms versus 100% in the CASPIAN, ECOG-ACRIN EA5161, IMpower133, and KEYNOTE-604 trials. In the Okamoto 2007 trial, in those aged 70 years and older, patients with an ECOG PS of 0 to 2 were included, whereas those with an ECOG PS of 3 were included if they were younger than 70 years of age. There was inconsistency in the reporting of the number and type of metastatic sites across the trials. Heterogeneity in the use of subsequent anticancer therapy administered to patients in the second line and higher recruited in the studies was identified as a potential source of bias affecting OS assessment (non-protocol second- and third-line treatment was reported in the Hermes 2008, Okamoto 2007, and Schmittel 2011 trials) and as potentially affecting the generalizability of the findings of the NMA to the Canadian setting. Variability was also observed in the dosing of etoposide plus carboplatin or cisplatin across the trials (3 studies — CASPIAN, Schmittel 2011, and Skarlos 1994 — randomized patients to combination chemotherapy regimens for up to 4 to 6 cycles, while in the IMpower133 trial [including ECOG-ACRIN EA5161, KEYNOTE-604, Hermes 2008, and Okamoto 2007], patients were dosed with the comparator for up to 4 cycles), and this may impact the findings of the ITC.

According to the sponsor’s ITC report, a meta-regression analysis was not possible to investigate inter-trial heterogeneity due to insufficient studies (i.e., the presence of several single study connections between interventions). Scenario analyses related to certain characteristics of interest were included in the sponsor’s NMA report to address heterogeneity across the trials included in the network (e.g., removal of the Hermes 2008 trial, which was an outlier due to having the smallest proportion of patients with ECOG PS < 2, from the OS base-case analysis). According to the clinical expert consulted, ECOG PS and metastatic sites (liver and brain) were the most significant effect modifiers in the treatment of ES-SCLC patients. The sponsor acknowledged that additional scenario or subgroup analyses were feasible for PFS and OS; however, because relevant subgroup data are not currently available from the trials of the evidence networks investigating immunotherapies (i.e., CASPIAN, ECOG-ACRIN EA5161, and KEYNOTE-6040 ongoing trials), not all possible subgroup analyses were included in the sponsor’s report. Therefore, the NMA results should be interpreted with caution due to limitations that may arise from between-study differences in some covariates and lack of sufficient evidence to minimize heterogeneity and inconsistency (e.g., a meta-regression analysis).

Description of Published Indirect Comparisons

Nine ITCs were retrieved from literature that assessed other treatment options in patients ES-SCLC.

Objectives

The objectives of the 9 ITCs retrieved from the CADTH literature search are outlined in Table 34 and Table 35.

Study Selection Methods

Table 34 and Table 35 in Appendix 3 present the study selection methods applied by the published ITCs to select studies for their network. All 9 studies included untreated patients with ES-SCLC, although 1 study (Gristina et al. [2021]) focused on patients with advanced or unresectable ES-SCLC with stage IVA or IVB disease. Chemotherapy (platinum-based agents) plus etoposide was the common comparator identified across the published ITCs. Eight studies in total presented data on OS, PFS, and AEs as the outcomes investigated. One study (Chen et al. [2017]) investigated AEs as the main outcome.

All 9 ITCs performed literature searches in electronic databases, some of which included PubMed and other grey literature websites. Most studies limited their search to English-language publications, and 3 studies reportedly had no language restrictions. Most studies had their literature search conducted independently by 2 reviewers and further validated by a third author. Most studies (8 in total) conducted risk of bias analysis across the studies included in the network using the Cochrane risk of bias tool; 1 study used the CONSORT checklist to assess bias.

Methods of Published Indirect Comparisons

ITC Analysis Methods

Table 36 and Table 37 in Appendix 4 provide a summary of the ITC methods used in the published ITCs.

Of the 9 published ITCs assessed in the CADTH review, 4 (Ando et al. [2021]³⁹; Kang et al. [2021]⁴⁰; Gristina et al. [2021]⁴¹; Zhou et al. [2020]⁴²) used Bayesian methods to conduct the NMA, 3 (Chen et al. [2021]⁴³; Chen HL et al. [2020]⁴⁴; Wang et al. [2020]⁴⁵) reported a frequentist approach, and the rest did not specify clearly the NMA methods used in the ITC. None of the studies reported the priors used in conducting the NMA. Only 1 study (Kang et al. [2021]⁴⁰) reported information on the assessment of model fit. Consistency assessments were reported for some studies, most of which used the I² statistics to assess heterogeneity between the trials included in the network. In other studies, consistency could not be assessed due to the absence of closed loops. Only 1 study (Ando et al. [2021]³⁹) reported how convergence was assessed in the NMA. Three studies (Chen HL et al. [2020]⁴⁴; Ando et al. [2021]³⁹; Gristina et al. [2020]⁴¹) included subgroup analyses. In 4 studies (Chen et al. [2020]⁴⁶; Zhou et al. [2020]⁴²; Chen et al. [2017]⁴⁷; Kang et al. [2021]⁴⁰), the authors reported reasons why subgroup analyses could not be conducted. Only 3 studies (Ando et al. [2021]³⁹; Chen et al. [2017]⁴⁷; Kang et al. [2021]⁴⁰) had conducted sensitivity analyses.

Results of Published Indirect Comparisons

Summary of Included Studies

The KEYNOTE-604, CASPIAN, IMpower133, and EA5161 trials were common studies included in the network structures across the 9 published ITCs assessed (refer to Table 34 and Table 35). These studies were also included in the sponsor-submitted NMA. In the ITC by Chen et al. (2020), the network geometry included 46 trials compared to the other ITCs retrieved and the sponsor-submitted NMA. The lowest number of studies included in the ITCs was observed in Wang et al. (2020) (4 studies). Given that most ITCs included studies that were also in the sponsor’s ITC, the baseline characteristics of those studies are also presented in Appendix 4.

Efficacy Outcomes

Progression-Free Survival

In Ando et al. (2021),³⁹ the authors concluded that PFS was statistically different between the atezolizumab plus etoposide and platinum (HR = 0.71; 95% CrI, 0.53 to 0.93) and the durvalumab platinum-based chemotherapy plus etoposide (HR = 0.72; 95% CrI, 0.55 to 0.91) groups, respectively, compared to platinum plus irinotecan, as the CrIs for these comparisons excluded the null value.

The ITC by Chen et al. (2021)⁴³ observed no statistically significant differences in PFS in the indirect comparison between atezolizumab and durvalumab (HR = 0.96; 95% CI, 0.72 to 1.29).

In the ITC by Gristina et al. (2021),⁴¹ the pooled results did not suggest a difference in terms of PFS (HR = 1.21; 95% CI, 0.98 to 1.49) between the anti-PD-1 class of therapies and chemotherapy.

The ITC by Chen et al. (2020)⁴⁶ suggested that atezolizumab plus carboplatin and etoposide may be associated with an improvement in PFS compared with the common comparator carboplatin plus etoposide, considering the CIs that did not include the null value (HR = 0.77; 95% CI, 0.60 to 0.99).

Similarly, the ITC by Chen HL et al. (2020)⁴⁴ showed an improvement in PFS with atezolizumab plus chemotherapy (HR = 0.77; 95% CI, 0.61 to 0.96) when compared with chemotherapy alone. No superior effects were observed in this ITC for the pairwise comparisons between the different ICIs.

In Wang et al. (2020),⁴⁵ the pooled HR estimates showed that immunotherapy (including ipilimumab, atezolizumab, and durvalumab) plus chemotherapy significantly improved PFS (HR = 0.81; 95% CI, 0.74 to 0.88;) when compared against placebo plus chemotherapy.

In the ITC by Zhou et al. (2020),⁴² the addition of PD-L1 inhibitors (durvalumab and atezolizumab) to etoposide plus platinum chemotherapy resulted in an HR of 1.29 (95% CI, 0.96 to 1.75) for PFS, compared with etoposide plus platinum chemotherapy alone.

Overall Survival

In the ITC by Ando et al. (2021),³⁹ the OS in the groups treated with atezolizumab plus etoposide and platinum-based chemotherapy, durvalumab plus etoposide and platinum-based chemotherapy, or pembrolizumab plus etoposide and platinum-based chemotherapy was found to be significantly higher than that in groups treated with etoposide plus platinum, with HRs of 0.71 (95% CrI, 0.54 to 0.91), 0.73 (95% CrI, 0.59 to 0.91), and 0.81 (95% CrI, 0.65 to 0.99), respectively. The OS of groups treated with atezolizumab plus etoposide and platinum-based chemotherapy or durvalumab plus etoposide and platinum-based chemotherapy was significantly higher than that of groups treated with platinum-amrubicin, with HRs of 0.73 (95% CrI, 0.52 to 0.99) and 0.757 (95% CrI, 0.56 to 0.99), respectively. No significant differences in OS were observed between each pair of 3 ICIs plus etoposide and platinum.

In the ITC by Chen et al. (2021),⁴³ no significant difference in OS was observed in the indirect comparison between atezolizumab and durvalumab (HR = 0.93; 95% CI, 0.67 to 1.30).

The ITC by Gristina et al. (2021)⁴¹ reported an HR of 1.22 (95% CI, 0.97 to 1.53) for treatments in the anti-PD-1 class against chemotherapy.

In the ITC by Chen et al. (2020),⁴⁶ findings in the comparison groups of atezolizumab plus platinum-based chemotherapy and etoposide (HR = 0.70; 95% CI, 0.52 to 0.94) versus platinum and etoposide plus durvalumab (HR = 0.73; 95% CI, 0.57 to 0.94) suggested improved OS for the atezolizumab combination.

The ITC by Chen HL et al. (2020)⁴⁴ suggested an improvement in OS in the atezolizumab plus chemotherapy group (HR = 0.70; 95% CI, 0.54 to 0.91) compared with chemotherapy alone.

In the ITC by Wang et al. (2020),⁴⁵ the pooled HR estimates suggested that immunotherapy (including ipilimumab, atezolizumab, and durvalumab) plus chemotherapy may improve the survival outcomes in terms of OS (HR = 0.84; 95% CI, 0.75 to 0.93) in comparison with chemotherapy plus placebo.

In the ITC by Zhou et al. (2020),⁴² the HR for OS was reported to be 1.40 (95% CI, 1.09 to 1.80) with the addition of PD-L1 inhibitors (durvalumab and atezolizumab) to etoposide-platinum chemotherapy compared with etoposide-platinum chemotherapy alone.

Objective Response Rate

In the ITC by Chen et al. (2021),⁴³ the OR observed for ORR in the comparison of durvalumab versus atezolizumab (monotherapy) was 0.79 (95% CI, 0.64 to 0.98).

In the ITC by Chen HL et al. (2020),⁴⁴ the authors observed no significant difference between any comparable ICIs except for durvalumab, which produced a noticeable benefit over atezolizumab.

In Wang et al. (2020),⁴⁵ the pooled risk ratios for ORR in direct comparisons were 1.04 (95% CI, 0.94 to 1.16; P = 0.452) for immunotherapy (including ipilimumab, atezolizumab, and durvalumab) plus chemotherapy against placebo plus chemotherapy.

In the ITC by Zhou et al. (2020),⁴² no significant difference was observed in ORR by the authors between the etoposide-platinum chemotherapy with PD-L1 inhibitors (including atezolizumab and durvalumab), etoposide plus cisplatin, or carboplatin plus a PD-L1 inhibitor (OR = 0.86; 95% CI, 0.48 to 1.58), compared with etoposide plus cisplatin or carboplatin alone.

Safety Outcomes: AEs

In the ITC by Ando et al. (2021),³⁹ the incidence of at least grade 3 AEs was significantly higher with ICIs plus etoposide and platinum than with irinotecan and platinum. There were no significant differences observed in grade 3 AEs between each pair of 3 ICIs plus etoposide and platinum.

In the ITC by Chen et al. (2021),⁴³ the authors reported an OR of 0.22 (95% CI, 0.1 to 0.5) in the comparison between durvalumab versus atezolizumab (OR = 0.22; 95% CI, 0.10 to 0.50).

In the ITC by Gristina et al. (2021),⁴¹ no statistically meaningful differences were observed in AEs between the treatments included in the network compared with chemotherapy alone (risk ratio [RR] = 2.27; 95% CI, 1.02 to 5.1).

In the ITC by Chen HL et al. (2020),⁴⁴ no significant differences were observed in the risk of grade 3 or 4 AEs for etoposide and platinum plus ICIs (nivolumab, atezolizumab, or durvalumab) compared with etoposide plus platinum alone.

In the ITC by Wang et al. (2020),⁴⁵ the pooled relative risks of the direct comparisons in the network were 1.03 (95% CI, 0.98 to 1.08) for any grade AEs and 0.97 (95% CI, 0.89 to 1.05) for at least grade 3 AEs in the immunotherapy (including ipilimumab, atezolizumab, and durvalumab) plus chemotherapy groups against placebo plus chemotherapy.

In the ITC by Zhou et al. (2020),⁴² the addition of PD-L1 inhibitors to etoposide-platinum chemotherapy was reported to be associated with no meaningfully different toxic effects in general (OR = 1.14; 95% CI, 0.36 to 2.31) when compared with etoposide plus cisplatin or carboplatin, considering the wide CI for OR that includes the null value.

Critical Appraisal of Published Indirect Comparisons

The NMA methodology reported in the published ITCs did not include sufficient details to assess the credibility and validity of the ITCs. However, the CADTH review team attempted to identify commonalities and differences between the identified ITCs and the sponsor-submitted NMA. A summary is provided below.

In the ITC by Ando et al. (2021),³⁹ variability was reported in the age and performance status of patients recruited across trials, although upon assessment of risk of bias, the authors concluded that heterogeneity had little impact on the final conclusions, and no significant inconsistency was detected in the global inconsistency test conducted.

The trials included in the ITC by Chen et al. (2021)⁴³ were identical to the sponsor-submitted ITC. The authors reported a low risk of bias in their quality assessments in allocation concealment, random sequence generation, blinding of outcome assessments, and blinding of participants and personnel; however, they reported potential heterogeneity in the diversity of the racial population given that most of the studies included in the network recruited non-Asian populations.

The ITC by Gristina et al. (2021)⁴¹ included 4 of the 6 studies also included in the sponsor-submitted NMA. The authors conducted a risk of bias analysis using the CONSORT checklist, and the authors reported average quality results for the trials. The authors identified some issues with the trials in Gristina et al. (2021)⁴¹ that were related to the likelihood of performance bias and detection bias owing to the open-label nature of the trials.

The ITC by Chen et al. (2020)⁴⁶ reported baseline characteristics for all 46 trials included in the network diagram. Six of the 46 trials also featured in the sponsor-submitted NMA. Forty of the 46 trials (87%) compared platinum plus etoposide with other regimens. The authors reported that comparisons between carboplatin and cisplatin were not robust because of the sparse network. In addition, most of the RCTs included only patients with ECOG PS 0 to 1 or 0 to 2.

In the ITC by Chen HL et al. (2020),⁴⁴ 4 of the 6 studies that were included in the network geometry were also included in the sponsor-submitted NMA. Baseline characteristics were presented, and the authors also performed quality assessments and reported heterogeneity in the RCTs included in the network diagram in terms of regimen for chemotherapy and criteria for treatment response or progression.

The ITC by Wang et al. (2020)⁴⁵ also presented the baseline characteristics of patients included in the 4 RCTs in the network, 2 of them also presented in the sponsor’s NMA. The quality assessments conducted by the authors showed that all studies achieved randomization and had low-to-moderate risk of bias. The network geometry was considered sparse, and the authors concluded that more data were needed to supplement the findings.

The ITC by Zhou et al. (2020)⁴² included 14 studies, of which 4 featured in the sponsor’s NMA. The authors did not present the baseline characteristics, and the quality assessments showed low risk of detection and reporting bias. All trials that were included implemented blinding of patients and personnel; 8 studies were reported to have low risk of attrition bias. Subgroup analyses were not reported in the ITC, and some of the patients included in the RCTs had received second-line and later therapies. Adverse events data could not be investigated.

In the ITC by Chen et al. (2017),⁴⁷ 9 studies were included, of which 3 were similar to the RCTs in the sponsor’s NMA. Baseline characteristics were reported for the study, and the quality assessments showed low risk of attrition or reporting bias given that all studies included required the blinding of participants and personnel. The authors reported no heterogeneity in the RCTs following their assessment. The authors reported variability in the follow-up durations, pre-medication dosage, and racial differences. In addition, genetic variations between the comparisons could not be adjusted. The authors also reported a limited sample size, which would affect the precision of the results, and inconsistency could not be evaluated.

The ITC by Kang et al. (2021)⁴⁰ included 5 trials in the network, of which 4 were identical to those included in the sponsor’s NMA. Risk of bias was assessed, and a key limitation identified was the absence of subgroup analyses.

Summary

Overall, heterogeneity was identified in the sponsor-submitted NMA, and comparisons had imprecise estimates (atezolizumab plus carboplatin and etoposide against durvalumab plus carboplatin (or cisplatin) and etoposide for PFS, OS, and ORR). The PFS and OS comparisons of atezolizumab plus carboplatin and etoposide against durvalumab plus carboplatin (or cisplatin) and etoposide showed no statistically significant differences based on the wide CrIs which included the null value, and the point estimates that were close to the null. Therefore, it is likely that the treatment effects for both regimens are comparable. However, no firm conclusion could be drawn due to the small number of studies per comparison, leading to a sparce network and lower precision of effect estimates.

The safety data were considered uncertain given that only 2 trials informed the network, and the estimates were imprecise (the 95% CrIs obtained were wide and included the null value).

The published ITCs had too little information related to the networks constructed to make definitive conclusions on the findings reported. The limitations identified were mainly related to the NMA methodology used, heterogeneity across the studies included in the networks, and heterogeneity in the baseline characteristics of patients within the included trials. Further, the findings on OS and PFS reported in most of the ITCs had imprecise estimates (wide 95% CIs which included the null). The CADTH review team considered that the findings from the ITCs published did not provide sufficient evidence to inform the comparison of the efficacy and safety of atezolizumab plus carboplatin and etoposide against relevant comparators.

Other Relevant Evidence

The sponsor submitted additional exploratory analyses from the IMpower133 trial, in conference abstract format, which included an analysis of long-term survivors (defined as patients who lived for at least 18 months after randomization). Of the 373 patients included in this analysis, more long-term survivors were treated with atezolizumab plus carboplatin (or cisplatin) and etoposide (33.5%) than with carboplatin or cisplatin plus etoposide (20.4%).⁴⁸ An exploratory analysis that assessed the benefit of atezolizumab plus carboplatin and etoposide compared to carboplatin and etoposide alone in patients who reached the maintenance phase of the trial (i.e., patients who received at least the first dose of maintenance therapy regardless of the number of chemotherapy cycles received) showed that a similar proportion of patients received maintenance treatment in the 2 treatment arms (77% in the atezolizumab arm, and 81% in the placebo arm). The analysis suggested an OS and PFS benefit in the maintenance population in patients receiving atezolizumab plus carboplatin and etoposide versus carboplatin and etoposide alone (OS: HR = 0.59; 95% CI, 0.43 to 0.81; PFS: HR = 0.64; 95% CI, 0.50 to 0.82). The authors concluded that induction treatment with atezolizumab plus carboplatin and etoposide as well as maintenance with atezolizumab both appeared to contribute to the OS benefit observed in the IMpower133 trial.⁴⁹ Another post hoc exploratory analysis showed that time to intra-cranial progression was delayed with atezolizumab plus carboplatin and etoposide (median time to intra-cranial progression [95% CI] = 20.2 months [11.0 to NE], when compared to carboplatin and etoposide alone (10.5 months [8.7 to 17.3]; HR = 0.66; 95% CI, 0.44 to 1.00).⁵⁰

Discussion

Summary of Available Evidence

The evidence base for this review consists of 1 RCT, 1 ITC submitted by the sponsor, and 9 published ITCs. IMpower133 is a randomized, multi-centre, double-blind, placebo-controlled phase III study designed to evaluate the efficacy and safety of treatment with atezolizumab plus carboplatin and etoposide compared with treatment with placebo plus carboplatin and etoposide in patients with chemotherapy-naive ES-SCLC. The trial was conducted in 106 sites across 21 countries (none in Canada). The co-primary end points were investigator-assessed PFS and OS. The key secondary end points were investigator-assessed ORR and investigator-assessed DOR. Patient-reported outcomes included HRQoL. Overall, the mean age was 63.7 years (SD = 8.9); 64.8% of patients were male and 79.9% were White. Patients had to have an ECOG PS of 0 or 1, and approximately 64% of the patients in both treatment arms had an ECOG PS of 1. Of the 526 patients screened, 403 were randomized, 201 patients to the atezolizumab arm and 202 patients to the placebo arm.

The sponsor-submitted ITC compared atezolizumab in combination with etoposide and platinum-based chemotherapy for the first-line treatment of ES-SCLC with relevant platinum doublet therapies and immunotherapies used in clinical practice. Only comparisons of atezolizumab plus carboplatin and etoposide versus etoposide plus carboplatin, etoposide plus cisplatin, and durvalumab plus etoposide and carboplatin (or cisplatin) were considered for the CADTH review. However, uncertainty remains around the findings of the sponsor-submitted ITC due to heterogeneity between the included studies (e.g., patient characteristics across studies) and wide CrIs around the reported point estimates of comparative treatment effect. Few inferences can be made from the published ITCs due to important limitations related to the NMA methodology used and heterogeneity across the studies included in the networks and in the baseline characteristics of patients within the trials.

Interpretation of Results

Efficacy

In the IMpower133 study, the addition of atezolizumab to carboplatin and etoposide chemotherapy resulted in longer PFS and OS than was observed with placebo plus carboplatin and etoposide in the first-line treatment of patients with ES-SCLC. At the first data cut-off date (April 24, 2018), 233 death events (59%) and 360 PFS events (89%) had occurred and both co-primary end points of the study had been met. The OS showed a statistically significant difference between the 2 arms in favour of atezolizumab (median OS = 12.3 months) over placebo (median OS = 10.3 months), as indicated by a stratified HR of 0.701 (95% CI, 0.54 to 0.91; P = 0.0069). OS results from the placebo arm of the trial are in line with data from most published studies of platinum plus etoposide. The final OS analysis after a median follow-up of 22.9 months (data cut-off January 24, 2019; 302 out of 403 OS events = 75%) was consistent with the interim OS analysis. The median OS in both treatment arms was unchanged (12.3 months in the atezolizumab arm and 10.3 months in the placebo arm; HR = 0.76; 95% CI, 0.60 to 0.95; P = 0.0154). At 24 months, the survival rate was approximately 5% higher in the atezolizumab arm than in the placebo arm (22.0% versus 16.8%). This difference in survival rate at 2 years was noted as clinically important by the clinical experts consulted by CADTH. Subgroup analyses did not identify a particular group of patients with considerably higher or lower benefit from atezolizumab plus carboplatin and etoposide. Progression-free survival results were also in favour of atezolizumab, with a median investigator-assessed PFS of 5.2 months, versus 4.3 months in the placebo arm (HR = 0.772; 95% CI, 0.62 to 0.96; P = 0.0170). The clinical experts consulted by CADTH noted that although the net benefits of 2 months in OS and about 1 month in PFS are not striking, in the context of ES-SCLC, which is an aggressive disease with poor prognosis, these marginally improved OS and PFS effects are clinically meaningful.

The benefit of the addition of atezolizumab to carboplatin and etoposide was not substantially supported by secondary end points. The confirmed ORR was numerically higher in the placebo arm (64.4%, compared to 60.2% in the atezolizumab arm). However, this difference between treatment arms may reflect the initial high response rate to chemotherapy that is characteristic of patients with ES-SCLC. The median DOR was similar in both treatment arms (4.2 months versus 3.9 months in the atezolizumab and placebo arms, respectively). The clinical experts consulted indicated that durability of survival benefit as measured by survival rates is key in the setting of ES-SCLC, where most patients do not live longer than 8 months to 10 months with chemotherapy alone, and ORR does not capture this aspect of treatment effect. Landmark OS rates were numerically higher in the atezolizumab arm than in the placebo arm at 12 months (51.9% versus 39.0%), 18 months (34.0% versus 21.0%), and 24 months (22.0% versus 16.8%) after randomization. At 12 months, OS rates were approximately 13% higher, and at 24 months they were 5.2% higher in the atezolizumab arm than in the placebo arm. However, these results should be treated with caution as landmark analyses do not take censoring into account.

In the CASPIAN study evaluating the efficacy and safety of durvalumab in combination with carboplatin (or cisplatin) and etoposide for the first-line treatment of patients with ES-SCLC, the median OS was 13.0 months (95% CI, 11.5 to 14.8) in the durvalumab plus platinum-etoposide arm versus 10.3 months (95% CI, 9.3 to 11.2) in the platinum-etoposide arm (HR = 0.73; 95% CI, 0.59 to 0.91).²⁵ As of the final analysis, the median OS was 12.9 months (95% CI, 11.3 to 14.7) in the durvalumab plus carboplatin (or cisplatin) and etoposide arm, compared to 10.5 months (95% CI, 9.3 to 11.2) in the platinum-etoposide arm (HR = 0.78; 95% CI, 0.64 to 0.94). The landmark analysis of OS at 12, 18, and 24 months in the CASPIAN study showed similar event-free rates as observed in the IMpower133 trial (52.8%, 32%, and 22.9%, respectively). The effect of durvalumab on OS appears to be comparable to the OS benefit observed with the addition of atezolizumab to carboplatin and etoposide.⁵¹ There is some suggestion from the ITC included in this review that atezolizumab plus carboplatin and etoposide may be similar to durvalumab plus carboplatin (or cisplatin) and etoposide in terms of OS benefit based on the HR; however, the 95% CrIs observed were imprecise.

Selected patients were allowed to continue treatment with atezolizumab monotherapy beyond radiographic progression. In the setting of ES-SCLC, given patients’ poor prognosis and the limited efficacy of later-line treatments, achieving clinically significant benefit from therapy in the first-line setting is important. In addition, there is a potential for pseudoprogression or tumour-immune infiltration. The clinical experts consulted by CADTH agreed that conventional response criteria may not adequately assess the activity of immunotherapy agents and that progressive disease may not necessarily imply therapeutic failure. While the rationale for implementing treatment beyond progression in the IMpower133 trial is acceptable and is endorsed by the clinical experts consulted, current data from the IMpower133 trial are insufficient to establish a clear benefit of treatment beyond progression.

An important aspect of treatment desired by both clinicians and patient groups consulted by CADTH was improvement in patients’ cancer symptoms and HRQoL. The overall HRQoL improved in both treatment arms. However, overall, there was no significant difference in any domain of HRQoL between the 2 treatment arms, suggesting that the addition of atezolizumab to chemotherapy neither improves nor impedes HRQoL.

Harms

Atezolizumab exhibited an acceptable toxicity profile, as expected from an anti-PD-1 checkpoint inhibitor. The incidence and severity of immune-related AEs were also reasonable, and AEs were consistent with atezolizumab’s immune-mediated mechanism of action. Adverse events were observed in almost all patients in the trial. Although the proportion of patients with grade 3 or 4 AEs was high, it was comparable between the 2 treatment arms (atezolizumab arm: 67.7%; placebo arm: 63.3%). The incidence of SAEs was also similar in the 2 treatment arms (atezolizumab arm: 38.9%; placebo arm: 35.2%). However, grade 5 fatal AEs were more common in the placebo arm than in the atezolizumab arm (5.6% versus 2.0%). More patients with AEs required treatment withdrawal in the atezolizumab arm (11.6%) than in the placebo arm (2.6%).

The most common AEs of any grade that occurred in the trial were anemia (39%), neutropenia (36%), alopecia (36%), nausea (35%), constipation (28%), and fatigue (26%), which is consistent with what is expected from carboplatin and etoposide chemotherapy, the backbone treatment in both treatment arms. Grade 3 or 4 AEs that occurred in the trial were in general also related to myelotoxicity and thus more likely related to carboplatin and etoposide backbone treatment. The incidence of grade 3 or 4 neutropenia was comparable in both arms (atezolizumab arm: 22.7%; placebo arm: 25.0%). The majority of SAEs were also related to myelotoxicity and were observed in a similar proportion of patients in the 2 treatment arms. The proportion of patients with febrile neutropenia was higher in the placebo arm (4.6%, versus 2.5% in the atezolizumab arm).

The proportion of patients with immune-related AEs was considerably higher in the atezolizumab arm than in the placebo arm (41.4% versus 24.5%). Approximately a quarter of the patients from each treatment arm required systemic corticosteroids. Of the immune-related AEs, the most frequent was rash, followed by thyroid disorders and hepatitis. The clinical experts consulted by CADTH indicated that the safety profile of carboplatin and etoposide in the trial is consistent with the known safety profile of these drugs in clinical practice. The incidence and type of immune-related AEs observed in the atezolizumab arm are also largely consistent with what is expected in clinical practice. Although the addition of atezolizumab to carboplatin and etoposide appears to increase the incidence of grade 3 or 4 AEs and SAEs, leading to a higher proportion of patients requiring treatment modification or withdrawal from treatment, the majority of AEs occurring with atezolizumab and other ICIs in this setting are amenable to timely detection and manageable with treatment.

The results of the ITC relating to the safety of atezolizumab with respect to other comparable regimens are highly uncertain as only 2 trials informed the network, and the estimates were imprecise. There is no evidence of differences in the safety profiles of atezolizumab plus carboplatin and etoposide versus durvalumab plus carboplatin (or cisplatin) and etoposide.

Conclusions

Based on clinical data from the IMpower133 study, atezolizumab in combination with carboplatin and etoposide demonstrated a statistically significant benefit compared to placebo in combination with carboplatin and etoposide in the first-line treatment of patients with ES-SCLC. The updated OS analysis with a median of 22.9 months of follow-up showed consistent results with those reported at the interim OS analysis and suggests maintained clinical benefit from atezolizumab in combination with carboplatin and etoposide. Although the net gain of about 1 month in median PFS and 2 months in median OS observed with the addition of atezolizumab to carboplatin and etoposide is modest, it was considered by the clinical experts consulted by CADTH to be clinically meaningful in this setting, where patients experience rapid tumour growth and fast clinical deterioration and have poor prognosis. The toxicity profile of atezolizumab was consistent with its immune-mediated mechanism of action, with no new safety concerns. In terms of improving OS, atezolizumab appears to demonstrate comparable benefit to durvalumab, the only other immunotherapy agent approved (but not currently funded by the drug plans) in Canada for this indication, and can be considered an alternative treatment in combination with carboplatin and etoposide chemotherapy for the first-line treatment of ES-SCLC in Canada.

References

1.Canadian Cancer Society. Lung and bronchus cancer statistics. 2021; https://cancer.ca/en/cancer-information/cancer-types/lung/statistics. Accessed 2022 Apr 1.

2.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

3.Kalemkerian GP. Staging and imaging of small cell lung cancer. Cancer Imaging. 2012;11:253-258. PubMed

4.Kalemkerian GP, Gadgeel SM. Modern staging of small cell lung cancer. J Natl Compr Canc Netw. 2013;11(1):99-104. PubMed

5.Nicholson AG, Chansky K, Crowley J, et al. The International Association for the Study of Lung Cancer lung cancer staging project: Proposals for the revision of the clinical and pathologic staging of small cell lung cancer in the forthcoming eighth edition of the TNM classification for lung cancer. J Thorac Oncol. 2016;11(3):300-311.

6.Socinski MA, Smit EF, Lorigan P, et al. Phase III study of pemetrexed plus carboplatin compared with etoposide plus carboplatin in chemotherapy-naive patients with extensive-stage small-cell lung cancer. J Clin Oncol. 2009;27(28):4787-4792. PubMed

7.Small cell lung cancer, version 2.2022. (NCCN Clinical Practice Guidelines in Oncology). Plymouth Meeting (PA): National Comprehensive Cancer Network; 2021: https://www.nccn.org/guidelines/guidelines-detail?category=1&id=1462 (free registration required). Accessed 2022 Apr 1.

8.Auperin A, Arriagada R, Pignon JP, et al. Prophylactic cranial irradiation for patients with small-cell lung cancer in complete remission. Prophylactic Cranial Irradiation Overview Collaborative Group. N Engl J Med. 1999;341(7):476-484. PubMed

9.Rudin CM, Brambilla E, Faivre-Finn C, Sage J. Small-cell lung cancer. Nat Rev Dis Primers. 2021;7(1):3. PubMed

10.Ganti AKP, Loo BW, Bassetti M, et al. Small cell lung cancer, version 2.2022, NCCN Clinical Practice Guidelines in Oncology. J Natl Compr Canc Netw. 2021;19(12):1441-1464. https://jnccn.org/view/journals/jnccn/19/12/article-p1441.xml. Accessed 2022 Jun 10. PubMed

11.Horn L, Mansfield AS, Szczesna A, et al. First-line atezolizumab plus chemotherapy in extensive-stage small-cell lung cancer. N Engl J Med. 2018;379(23):2220-2229. PubMed

12.Clinical Study Report: Primary CSR Study GO30081 (IMpower133). A Phase I/III, randomized, double-blind, placebo-controlled study of carboplatin plus etoposide with or without atezolizumab (anti-PD-L1 antibody) in patients with untreated extensive-stage small cell lung cancer. Report No.1084268 [internal sponsor's report]. Mississauga (ON): F. Hoffmann-La Roche Ltd; 2018 Sep.

13.Clinical Study Report: Update CSR Study GO30081 (IMpower133). A Phase I/III, randomized, double-blind, placebo-controlled study of carboplatin plus etoposide with or without atezolizumab(anti-PD-L1 antibody) in patients with untreated extensive-stage small cell lung cancer. Report No.1093769 [internal sponsor's report]. Mississauga (ON): F. Hoffmann-La Roche Ltd; 2019 Jun.

14.Center for Drug Evaluation and Research. Cross-discipline team leader review. Tecentriq (atezolizumab), injection for intravenous administration. Company: Genentech, Inc. Application No.:761034. Approval date: 01/12/2016 (FDA approval package). Silver Spring (MD): U.S. Food and Drug Administration (FDA); 2016 Apr 8: https://www.accessdata.fda.gov/drugsatfda_docs/nda/2016/761034Orig1s000CrossR.pdf. Accessed 2022 Feb 8.

15.Duma N, Santana-Davila R, Molina JR. Non-small cell lung cancer: Epidemiology, screening, diagnosis, and treatment. Mayo Clin Proc. 2019;94(8):1623-1640. PubMed

16.Canadian Cancer Statistics Advisory Committee. Canadian cancer statistics 2018. Toronto: Canadian Cancer Society; 2018: https://cdn.cancer.ca/-/media/files/research/cancer-statistics/2018-statistics/canadian-cancer-statistics-2018-en.pdf. Accessed 2022 Jun 10.

17.Karachaliou N, Pilotto S, Lazzari C, Bria E, de Marinis F, Rosell R. Cellular and molecular biology of small cell lung cancer: an overview. Transl Lung Cancer Res. 2016;5(1):2-15. PubMed

18.Bernhardt EB, Jalal SI. Small cell lung cancer. Cancer Treat Res. 2016;170:301-322. PubMed

19.Foster NR, Mandrekar SJ, Schild SE, et al. Prognostic factors differ by tumor stage for small cell lung cancer: a pooled analysis of North Central Cancer Treatment Group trials. Cancer. 2009;115(12):2721-2731. PubMed

20.Agra Y, Pelayo M, Sacristan M, Sacristan A, Serra C, Bonfill X. Chemotherapy versus best supportive care for extensive small cell lung cancer. Cochrane Database Syst Rev. 2003(4):CD001990. PubMed

21.Qin A, Kalemkerian GP. Cisplatin, etoposide, and irinotecan for relapsed small-cell lung cancer. Transl Cancer Res. 2016;5(Suppl 6):S1142-S1144. PubMed

22.Edelman MJ. Prophylactic cranial irradiation for small-cell lung cancer: time for a reassessment. Am Soc Clin Oncol Educ Book. 2020;40:24-28. PubMed

23.Slotman BJ, Mauer ME, Bottomley A, et al. Prophylactic cranial irradiation in extensive disease small-cell lung cancer: short-term health-related quality of life and patient reported symptoms: results of an international Phase III randomized controlled trial by the EORTC Radiation Oncology and Lung Cancer Groups. J Clin Oncol. 2009;27(1):78-84. PubMed

24.Yin X, Yan D, Qiu M, Huang L, Yan SX. Prophylactic cranial irradiation in small cell lung cancer: a systematic review and meta-analysis. BMC Cancer. 2019;19(1):95. PubMed

25.Paz-Ares L, Dvorkin M, Chen Y, et al. Durvalumab plus platinum-etoposide versus platinum-etoposide in first-line treatment of extensive-stage small-cell lung cancer (CASPIAN): a randomised, controlled, open-label, phase 3 trial. Lancet. 2019;394(10212):1929-1939. PubMed

26.Rudin CM, Awad MM, Navarro A, et al. Pembrolizumab or Placebo Plus Etoposide and Platinum as First-Line Therapy for Extensive-Stage Small-Cell Lung Cancer: Randomized, Double-Blind, Phase III KEYNOTE-604 Study. J Clin Oncol. 2020;38(21):2369-2379. PubMed

27.Tecentriq (atezolizumab): 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 Ltd; 2022 Jan 14.

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

29.Grey matters: a tool for searching health-related grey literature. Ottawa (ON): CADTH; 2022: https://greymatters.cadth.ca/. Accessed 2022 Jun 10.

30.Liu SV, Reck M, Mansfield AS, et al. Updated overall survival and PD-L1 subgroup analysis of patients with extensive-stage small-cell lung cancer treated with atezolizumab, carboplatin, and etoposide (IMpower133). J Clin Oncol. 2021;39(6):619-630. PubMed

31.Mansfield AS, Kazarnowicz A, Karaseva N, et al. Safety and patient-reported outcomes of atezolizumab, carboplatin, and etoposide in extensive-stage small-cell lung cancer (IMpower133): a randomized phase I/III trial. Ann Oncol. 2020;31(2):310-317. PubMed

32.Reck M, Liu S, Mansfield AS, et al. IMPower 133: updated overall survival (OS) analysis of first-line (1L) atezolizumab + carboplatin + etoposide in extensive-stage SCLC (ES-SCLC) [conference paper presented at: ESMO 2019 Congress; 2019 Sep 27 - Oct 1; Barcelona, Spain]. Ann Oncol. 2019;30(suppl 5):v710-v717. https://oncologypro.esmo.org/meeting-resources/esmo-2019-congress/IMpower133-updated-overall-survival-OS-analysis-of-first-line-1L-atezolizumab-atezo-carboplatin-etoposide-in-extensive-stage-SCLC-ES-SCLC. Accessed 2022 Jun 10.

33.Ellis PM, Swaminath A, Pond GR. Patterns of relapse in small cell lung cancer: competing risks of thoracic versus CNS relapse. Curr Oncol. 2021;28(4):2778-2788. PubMed

34.Hermes A, Bergman B, Bremnes R, et al. Irinotecan plus carboplatin versus oral etoposide plus carboplatin in extensive small-cell lung cancer: a randomized phase III trial. J Clin Oncol. 2008;26(26):4261-4267. PubMed

35.Schmittel A, Sebastian M, Fischer von Weikersthal L, et al. A German multicenter, randomized phase III trial comparing irinotecan-carboplatin with etoposide-carboplatin as first-line therapy for extensive-disease small-cell lung cancer. Ann Oncol. 2011;22(8):1798-1804. PubMed

36.Skarlos DV, Samantas E, Kosmidis P, et al. Randomized comparison of etoposide-cisplatin vs. etoposide-carboplatin and irradiation in small-cell lung cancer. A Hellenic Co-operative Oncology Group study. Ann Oncol. 1994;5(7):601-607. PubMed

37.Okamoto H, Watanabe K, Kunikane H, et al. Randomised phase III trial of carboplatin plus etoposide vs split doses of cisplatin plus etoposide in elderly or poor-risk patients with extensive disease small-cell lung cancer: JCOG 9702. Br J Cancer. 2007;97(2):162-169. PubMed

38.Leal T, Wang Y, Dowlati A, et al. Randomized phase II clinical trial of cisplatin/carboplatin and etoposide (CE) alone or in combination with nivolumab as frontline therapy for extensive-stage small cell lung cancer (ES-SCLC): ECOG-ACRIN EA5161. J Clin Oncol. 2020;38(15_suppl):9000-9000.

39.Ando K, Manabe R, Kishino Y, et al. Comparative efficacy and safety of immunotherapeutic regimens with PD-1/PD-L1 inhibitors for previously untreated extensive-stage small cell lung cancer: a systematic review and network meta-analysis. Curr Oncol. 2021;28(2):1094-1113. PubMed

40.Kang S, Wang X, Zhang Y, Zhang B, Shang F, Guo W. First-Line Treatments for Extensive-Stage Small-Cell Lung Cancer With Immune Checkpoint Inhibitors Plus Chemotherapy: A Network Meta-Analysis and Cost-Effectiveness Analysis. Front Oncol. 2021;11:740091. PubMed

41.Gristina V, Galvano A, Castellana L, et al. Is there any room for PD-1 inhibitors in combination with platinum-based chemotherapy as frontline treatment of extensive-stage small cell lung cancer? A systematic review and meta-analysis with indirect comparisons among subgroups and landmark survival analyses. Ther Adv Med Oncol. 2021;13:17588359211018018. PubMed

42.Zhou T, Zhang Z, Luo F, et al. Comparison of First-Line Treatments for Patients With Extensive-Stage Small Cell Lung Cancer: A Systematic Review and Network Meta-analysis. JAMA netw. 2020;3(10):e2015748.

43.Chen J, Wang J, Xu H. Comparison of atezolizumab, durvalumab, pembrolizumab, and nivolumab as first-line treatment in patients with extensive-stage small cell lung cancer: A systematic review and network meta-analysis. Medicine. 2021;100(15):e25180. PubMed

44.Chen HL, Tu YK, Chang HM, et al. Systematic review and network meta-analysis of immune checkpoint inhibitors in combination with chemotherapy as a first-line therapy for extensive-stage small cell carcinoma. Cancers (Basel). 2020;12(12):03.

45.Wang BC, Xiao BY, Li PC, et al. Efficacy and safety of first-line immunotherapy in combination with chemotherapy for patients with extensive-stage small cell lung cancer: A systematic review and network meta-analysis. J Oncol. 2020;2020:2368164. PubMed

46.Chen H, Horita N, Ito K, et al. Systematic review of first-line chemotherapy for chemo-naive extensive-stage small-cell lung cancer: network meta-analysis. Ther Adv Med Oncol. 2020;12:1758835920965841. PubMed

47.Chen Y, Chen L, Zhong D. Comparing the adverse effects of platinum in combination with etoposide or irinotecan in previously untreated small-cell lung cancer patients with extensive disease: A network meta-analyses. Thorac Cancer. 2017;8(3):170-180. PubMed

48.Liu S, Horn L, Mok T, et al. IMpower133: characterization of long-term survivors treated first line with chemotherapy +/− atezolizumab in extensive-stage small cell lung cancer [conference abstract presented at ESMO Virtual Congress 2020; 2020 Sep 19 - Oct 18; Switzerland]. Ann Oncol. 2020;31(suppl 4):S974-S987. https://oncologypro.esmo.org/meeting-resources/esmo-virtual-congress-2020/impower133-characterisation-of-long-term-survivors-treated-first-line-with-chemotherapy-atezolizumab-in-extensive-stage-small-cell-lung-cancer. Accessed 2022 Jun 10.

49.Reck M, Horn L, Mok T, et al. IMpower133: exploratory analysis of maintenance therapy in patients with extensive-stage small cell lung cancer [conference paper]. Proceedings of the IASLC 2020 World Conference on Lung Cancer; 2020 Aug 9-12; Singapore. 2020; Abstract: https://library.iaslc.org/virtual-library-search?product_id=20&author=Reck&category=Small+Cell+Lung+Cancer%2FNET; Slides: https://dialogoroche.com.ar/content/dam/roche-dialogo/global-assets/downloadable-assets/oncology/roche-wclc-2020/presentaciones/WCLC-2020-presentation-reck-IMpower133-exploratory-analysis-of-maintenance-therapy-in-patients-with-extensive-stage-small-cell-lung-cancer.pdf. Accessed 2022 Jun 10.

50.Higgins K, Curran W, Liu S, Yu W, Brockman M, Johnson A. Patterns of disease progression after carboplatin/etoposide + atezolizumab in extensive-stage small cell lung cancer (ES-SCLC) [conference abstract at ASTRO’s 62nd Annual Meeting; 2020 Oct 23-29]. Int J Radiat Oncol Biol Phys. 2020;108(5):1398. PubMed

51.Paz-Ares L, Chen Y, Reinmuth N, et al. Durvalumab, with or without tremelimumab, plus platinum-etoposide in first-line treatment of extensive-stage small-cell lung cancer: 3-year overall survival update from CASPIAN. ESMO Open. 2022;7(2):100408. PubMed

52.Nishio M, Sugawara S, Atagi S, et al. Subgroup analysis of Japanese patients in a phase III study of atezolizumab in extensive-stage small-cell lung cancer (IMpower133). Clin Lung Cancer. 2019;20(6):469-476.e461. PubMed

53.Nishio M, Saito H, Goto K, et al. IMpower132: Atezolizumab plus platinum-based chemotherapy vs chemotherapy for advanced NSCLC in Japanese patients. Cancer Sci. 2021;112(4):1534-1544. PubMed

54.Agrawal S, Kumar A, Srivastava V, Mishra BN. Cloning, expression, activity and folding studies of serine hydroxymethyltransferase: a target enzyme for cancer chemotherapy. J Mol Microbiol Biotechnol. 2003;6(2):67-75. PubMed

55.Aaronson NK, Ahmedzai S, Bergman B, et al. The European Organization 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

56.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

57.Osoba D, Aaronson N, Zee B, Sprangers M, te Velde A. Modification of the EORTC QLQ-C30 (version 2.0) based on content validity and reliability testing in large samples of patients with cancer. The Study Group on Quality of Life of the EORTC and the Symptom Control and Quality of Life Committees of the NCI of Canada Clinical Trials Group. Qual Life Res. 1997;6(2):103-108. PubMed

58.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

59.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

60.Bedard G, Zeng L, Zhang L, et al. Minimal important differences in the EORTC QLQ-C30 in patients with advanced cancer. Asia Pac J Clin Oncol. 2014;10(2):109-117. PubMed

61.Bjordal K, de Graeff A, Fayers PM, et al. A 12 country field study of the EORTC QLQ-C30 (version 3.0) and the head and neck cancer specific module (EORTC QLQ-H&N35) in head and neck patients. EORTC Quality of Life Group. Eur J Cancer. 2000;36(14):1796-1807. PubMed

62.Zhang Z, Li M, Wang H, Agrawal S, Zhang R. Antisense therapy targeting MDM2 oncogene in prostate cancer: Effects on proliferation, apoptosis, multiple gene expression, and chemotherapy. Proc Natl Acad Sci U S A. 2003;100(20):11636-11641. PubMed

63.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

Appendix 1: Literature Search Strategy

Note that this appendix has not been copy-edited.

Clinical Literature Search

Overview

Interface: Ovid

Databases:

• MEDLINE All (1946-present)

• Embase (1974-present)

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

Date of search: February 22, 2022.

Alerts: Biweekly search updates until project completion.

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

Limits:

• Conference abstracts: excluded

Table 29: Syntax Guide

Multi-Database Strategy

1. (atezolizumab* or Tecentriq* or tecntriq or MPDL3280A or MPDL-3280A or RG7446 or RG-7446 or L01FF05 or 52CMI0WC3Y).ti,ab,kf,ot,hw,rn,nm.

2. Small Cell Lung Carcinoma/

3. (SCLC or SCLCs or ESSCLC? or ESCLC? or CSCLC?).ti,ab,kf.

4. ((lung* or bronchial or pulmonary) adj4 ((small cell* or smallcell* or microcellular*) adj3 (cancer* or carcinoma* or neoplas* or tumour* or tumor* or malignan*))).ti,ab,kf.

5. (oat cell adj (cancer* or carcinoma* or neoplas* or tumour* or tumor* or malignan*)).ti,ab,kf.

6. or/2-5

7. 1 and 6

8. (IMpower133 or NCT02763579).ti,ab,kf. use medall

9. 7 or 8

10. use medall

11. *atezolizumab/ or (atezolizumab* or Tecentriq* or tecntriq or MPDL3280A or MPDL-3280A or RG7446 or RG-7446 or L01FF05).ti,ab,kf,dq.

12. small cell lung cancer/

13. (SCLC or SCLCs or ESSCLC? or ESCLC? or CSCLC?).ti,ab,kf,dq.

14. ((lung* or bronchial or pulmonary) adj4 ((small cell* or smallcell* or microcellular*) adj3 (cancer* or carcinoma* or neoplas* or tumour* or tumor* or malignan*))).ti,ab,kf,dq.

15. (oat cell adj (cancer* or carcinoma* or neoplas* or tumour* or tumor* or malignan*)).ti,ab,kf,dq.

16. or/12-15

17. 11 and 16

18. (IMpower133 or NCT02763579).ti,ab,kf,dq.

19. 17 or 18

20. use oemezd

21. not (conference abstract or conference review).pt.

22. 10 or 21

23. remove duplicates from 22

Clinical Trials Registries

ClinicalTrials.gov

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

• Search – (atezolizumab OR Tecentriq OR MPDL3280A OR MPDL-3280A OR RG7446 OR RG-7446 OR L01FF05 OR 52CMI0WC3Y) and Disease: (small cell OR smallcell OR SCLC OR ESSCLC OR ESCLC) AND (lung OR pulmonary OR bronchial)

WHO ICTRP

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

• Search - (atezolizumab OR Tecentriq OR MPDL3280A OR MPDL-3280A OR RG7446 OR RG-7446 OR L01FF05 OR 52CMI0WC3Y) AND (small cell OR smallcell OR SCLC OR ESCLC OR ESSCLC) NOT NCT*

Health Canada’s Clinical Trials Database

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

• Search terms – atezolizumab, Tecentriq, small cell lung cancer

EU Clinical Trials Register

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

• Search - (atezolizumab OR Tecentriq OR MPDL3280A OR MPDL-3280A OR RG7446 OR RG-7446 OR L01FF05 OR 52CMI0WC3Y) AND (small cell OR smallcell OR SCLC OR ESSCLC OR ESCLC) AND (lung OR pulmonary OR bronchial) NOT NCT*

Grey Literature

Search dates: February 8 to 15, 2022.

Keywords: Tecentriq, atezolizumab; small cell lung cancer.

Limits: no limits by date or language.

Updated: Search updated before the completion of stakeholder feedback period

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

• Health Technology Assessment Agencies

• Health Economics

• Clinical Practice Guidelines

• Drug and Device Regulatory Approvals

• Advisories and Warnings

• Drug Class Reviews

• Databases (free)

• Internet Search

The complete search archive of sites consulted for this report is available on request.

Appendix 2: Excluded Studies

Note that this appendix has not been copy-edited.

Table 30: Excluded Studies

Appendix 3: Description and Appraisal of Outcome Measures

Note that this appendix has not been copy-edited.

Aim

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

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

QLQ LC13 Lung Cancer Supplement

The outcomes are evaluated in the included study and their properties are shown in : Summary of outcome measures and their measurement properties.

Findings

Table 31: Summary of Outcome Measures and Their Measurement Properties

MID = minimal important difference.

European Organization for Research and Treatment of Cancer, 30 Item Core Quality of Life Questionnaire (EORTC QLQ C-30) and Lung Cancer Supplement (LC13)

The EORTC QLQ C-30 (version 3) is a generic, patient self-administered questionnaire for evaluating the QoL of patients with cancer participating in clinical trials.^54,55 This questionnaire is intended to be complemented by tumour-specific questionnaire modules or supplements, such as the 1 for patients with lung cancer, or QLQ-LC13.^54,56

The EORTC QLQ-C30 consists of 30 items that are scored to create 5 multi-item functional scales, 3 multi-item symptom scales, 6 single-item symptom scales, and a global QoL scale (Table 32).⁵⁴ Version 3.0 of the questionnaire is the most current version and has been in use since December of 1997.⁶¹ It is intended for use in the adult population only.⁵⁴ The QLQ-C30 uses a 1-week recall period to assess function and symptoms.⁵⁴ Twenty-eight questions are scored on a 4–point Likert with anchors at 1 (“not at all”) and 4 (“very much”). The 2 questions that make up the global QoL scale are scored on a 7–point Likert scale with anchors at 1 (“very poor”) and 7 (“excellent”).⁶² Raw scores for each scale are computed as the average of the items that contribute to a particular scale. Each raw scale score is converted to a standardized score that ranges from 0 to 100 using a linear transformation, with a higher score reflecting better function on the function scales, and better HRQoL, while a higher score in the symptom scale means higher burden of symptoms and therefore a worse health state.^58,62

The EORTC QLQ-LC13 is a questionnaire measuring lung cancer symptoms and side effects from conventional chemo- and radiotherapy. It comprises 13 questions of lung cancer-associated symptoms (e.g., coughing, dyspnea and pain) and side effects from conventional chemo- and radiotherapy (e.g., hair loss, neuropathy, sore mouth and dysphagia).⁵⁶ All questions are scored on a 4–point Likert scale with anchors at 1 (“not at all”) and 4 (“very much”). The scoring approach for the QLQ-LC13 is identical in principle to that for the symptom scales/single items of the QLQ-C30. An outcome variable consists of a score from 0 to 100, with a higher score reflecting greater symptom burden and therefore a worse health state.⁶²

Table 32: EORTC QLQ-C30 Scales

Psychometric Properties

The validity and reliability of the QLQ-C30 has been assessed in several studies among patients with cancer, including lung cancer.^55-58 In a validation study of the QLQ-C30, patients with lung cancer (n = 160) and a heterogenous group of other cancers (n = 375) completed the questionnaire at baseline and at day 8 after chemotherapy.⁵⁷ Item-domain correlations were determined for the entire patient group at baseline and at day 8 after chemotherapy. At baseline, all items were strongly correlated within their own domain than with any other domains (r = −0.65 to 0.95), except for item 5 (whether the responders needed help with eating, dressing, washing, or using the toilet) and the physical function domain (r = −0.3). On day 8 after chemotherapy, the item-domain for item 5 and the physical function domain was higher (r = 0.49), indicating that item 5 was more relevant in the week after chemotherapy than before chemotherapy.⁵⁷ Similarly, items asking about vomiting showed a higher correlation with domains for nausea/vomiting on day 8 after chemotherapy (r = 0.89) than before chemotherapy (r = 0.74). The questionnaire also demonstrated adequate internal consistency for most domains at baseline and at day 8 after chemotherapy (Cronbach Alpha > 0.70).⁵⁷ There were observed strong inter-domain correlations between physical function and role function (r > 0.6), and between physical function and fatigue (r = –0.62). In the same study, the mean scores for physical, role and social functions, as well as for global QoL decreased significantly with decreasing ECOG Performance Status.⁵⁷

In a study consisting of 112 patients with advanced lung cancer or pleural mesothelioma, criterion and concurrent validity of the QLQ-C30 was supported for most of the functioning and symptom scales, and global QoL.⁵⁸ When used as a continuous score, the 6-minute walking score was found to be strongly correlated with the QLQ-C30 physical functioning, moderately to strongly correlated with fatigue, role functioning and global QoL (r > 0.4), and poorly correlated with dyspnea (r = 0.21). In the same study, the Hospital anxiety and depression (HADS) anxiety scale was found to be strongly correlated with the QLQ-30 emotional functioning (r = –0.75), and moderately correlated with global QoL (r = –0.47). The HADS depression scale has been found to have a moderate to strong correlation with all functioning scales, fatigue and appetite loss (r = 0.48 to 0.55).⁵⁸ In terms of reliability, both QLQ-C30 and QLQ-LC13 demonstrated an adequate internal consistency for most multi-item scales (Cronbach Alpha > 0.70), except for the cognitive functioning scale (Cronbach Alpha = 0.57).⁵⁸ The findings of another study demonstrate that the QLQ-30 and QLQ-LC13 items and scales pertaining to symptoms of lung cancer discriminate clearly between patient subgroups on the basis of ECOG Performance Status and disease stage, while items pertaining to treatment toxicity were found to discriminate between subgroups of patients receiving chemotherapy versus radiotherapy.⁵⁶

Minimally Important Difference

Change in the EORTC QLQ-C30 may be interpreted in terms of small, moderate or large changes in HRQoL.⁵⁹ A study of patients with SCLC and breast cancer estimated a clinically relevant change in score on any of the QLQ-C30 scales using an anchor-based approach.⁵⁹ The MID estimates in patients who reported “a little” change (for better or worse) on the subjective significance questionnaire (SSQ) had corresponding changes on a function or symptom scale of the QLQ-C30 of approximately 5 to 10 points. Participants who reported a “moderate” change in the SSQ had corresponding changes in the QLQ-C30 of about 10 to 20 points, and those who reported “very much” change had corresponding changes of more than 20 points.⁵⁹

Another study estimated the MID for the QLQ-C30 among 369 patients with advanced cancer (including lung cancer), using anchor and distribution-based methods for improvement and deterioration. Using 2 patient-based anchors (overall health and overall QoL), MID estimates of the QLQ-30 ranged from a meaningful change for improvement of 9.1 units (cognitive functioning) to 23.5 units (pain), and a meaningful change for deterioration ranging from 7.2 units (physical functioning) to 13.5 units (role functioning). Distribution-based estimates were closest to 0.5 SD.⁶⁰

In another study, WHO performance status and weight change were used as clinical anchors to determine estimates of MIDs for the QLQ-C30 scales in 812 patients with NSCLC.⁶³ The MID estimates for improvement (based on the performance status and weight gain) were physical functioning (9, 5); role functioning (14, 7); social functioning (5, 7); GHS (9, 4); fatigue (14,5); and pain (16, 2). The respective MID estimates for deterioration were physical (4, 6); role (5, 5); social (7, 9); GHS (4, 4); fatigue (6, 11); and pain (3, 7), based on the performance status and weight, respectively.⁶³

Limitations

The reliability of social items such as social functioning and the social worker’s ratings of social support and activity still need more study and are yet to be proven. The difference in methodology and timing of the studies assessing these values can imply uncertainty in the body of evidence for the properties of the measurement scales.

Table 34: Study Selection Criteria and Methods for Published ITCs

AE = adverse event; anti-PD-L1 = anti- programmed cell death-1 ligand 1; DCR = disease control rate; DOR = duration of response; ES-SCLC = extensive-stage small cell lung cancer; ORR = objective response rate; OS = overall survival; PFS = progression-free survival; RCT = randomized controlled trial; RoB = risk of bias.

Table 35: Study Selection Criteria and Methods for Published ITCs

AE = adverse event; anti-PD-L1 = anti- programmed cell death-1 ligand 1; DCR = disease control rate; ES-SCLC = extensive-stage small cell lung cancer; ORR = objective response rate; OS = overall survival; PFS = progression-free survival; RCT = randomized controlled trial; RoB = risk of bias.

Table 36: ITC Analysis Methods of the Published ITCs

AE = adverse event; CI = confidence interval; CrI = credible interval; DCR = disease control rate; DORR = duration of objective response rate; DR = discontinuation rate; ECOG PS = Eastern Cooperative Oncology Group Performance Status; HR = hazard ratio; ITC = indirect treatment comparison; NR = not reported; ORR = objective response rate; OS = overall survival; PFS = progression-free survival.

Table 37: ITC Analysis Methods of the Published ITCs

AE = adverse event; CI = confidence interval; CrI = credible interval; DCR = disease control rate; ECOG PS = Eastern Cooperative Oncology Group Performance Status; HR = hazard ratio; ITC = indirect treatment comparison; NR = not reported; ORR = objective response rate; OS = overall survival; PFS = progression-free survival; RR = risk ratio; TRAE = treatment-related adverse event.

Pharmacoeconomic Review

Executive Summary

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

Table 1: Submitted for Review