CADTH Reimbursement Review

Lenvatinib and Pembrolizumab (Lenvima and Keytruda)

Sponsor: Eisai Limited

Therapeutic area: Advanced or metastatic renal cell carcinoma

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.

Table 1: Submitted for Review

NOC = Notice of Compliance; RCC = renal cell carcinoma; TBD = to be determined.

Source: Lenvatinib product monograph.¹

Introduction

Renal cell carcinoma (RCC) is the most common form of kidney cancer, accounting for more than 85% of all cases around the world.² RCCs are further classified into different subtypes based on histology (clear cell, papillary, chromophobe, clear cell papillary, collecting duct, medullary, and unclassified). The clear cell component is the most prevalent form of RCC and represents more than 70% of all RCC cases in practice.^3,4 More than 33% of cases identified at initial diagnosis have metastatic disease⁵ due to the fact that most patients experience few or no symptoms at earlier stages, which restricts the number of cases identified with early disease.⁶ Common symptoms are blood in urine, dull pain around the flank region that does not go away, fullness in the upper abdomen or a lump in this area, fever, appetite loss, nausea, vomiting, constipation, weakness, fatigue, anemia, polycythemia, and unexplained weight loss.^3,4,6 Projected estimates in Canada in 2021 show that kidney and renal pelvis cancers were the seventh most diagnosed cancers in men (5,200 new cases; 2.8% of disease-related deaths) and the 12th most diagnosed cancers in women (2,600 new cases; 1.7% of disease-related deaths). The predicted 5-year age-standardized survival rate was 73% for both sexes. Established risk factors include smoking, hypertension, obesity, medications (over-the-counter pain killers, phenacetin-containing compounds, and diuretics), family history of RCC, and genetic conditions (von Hippel-Lindau disease) or hereditary papillary RCC.^3,4,6

Treatment selection in practice is based on prognostic risk models, particularly the International Metastatic Renal Cell Carcinoma Database Consortium (IMDC) risk group classification (favourable, intermediate, and poor).⁷ For patients in the favourable risk group, preferred therapies outlined by the Kidney Cancer Research Network of Canada (KCRNC) practice guideline include pembrolizumab (PEM) plus axitinib (AXI), and nivolumab (NIVO) plus cabozantinib. Other options include sunitinib (SUN) and pazopanib (PAZO). For patients in the intermediate- or poor risk groups, the preferred options include ipilimumab (IPI) plus NIVO, AXI-PEM, and NIVO plus cabozantinib. Other available options for patients in the intermediate and poor risk groups include SUN, PAZO, and cabozantinib (cabozantinib received market approval from Health Canada on October 6, 2021, as a first-line treatment option for patients with advanced RCC who are in the intermediate or poor IMDC risk group).

Lenvatinib (LEN) is a multiple-receptor tyrosine kinase (RTK) inhibitor that selectively inhibits kinase activities of vascular endothelial growth factor receptor 1 (VEGFR1), FMS-like tyrosine kinase 1 [FLT1], VEGFR2 [KDR], and VEGFR3 [FLT4]), in addition to other proangiogenic and oncogenic pathway–related RTKs. PEM is a high-affinity antibody against programmed cell death 1 protein (PD-1), which exerts dual ligand blockade of the PD-1 pathway, including programmed cell death 1 ligand 1 (PD-L1) and programmed cell death1 ligand 2 (PD-L2), on antigen-presenting or tumour cells. PEM is a solution for IV infusion available in a 100 mg per 4 mL vial. The Health Canada–approved dose is 20 mg (two 10 mg capsules) of LEN orally once daily in combination with PEM 200 mg administered as an IV infusion over 30 minutes every 3 weeks, or 400 mg every 6 weeks. This is continued until unacceptable toxicity or disease progression or for up to 24 months, or until administration of thirty-five 200 mg doses or eighteen 400 mg doses, whichever is longer. After completing combination therapy, LEN may be administered as a single drug until disease progression or unacceptable toxicity.

The objective of this report was to perform a systematic review of the beneficial and harmful effects of LEN combined with PEM for the treatment of adult patients with advanced or metastatic RCC who have not received prior systemic therapy for metastatic RCC.

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

Two patient groups, CanCertainty and Kidney Cancer Canada (KCC), provided input for this submission. The full patient group input is included in the stakeholder section of this review.

The CanCertainty group expressed concerns related to inconsistent provincial coverage for oncology treatment regimens containing orally administered drugs and the resulting financial burden on vulnerable patients.

The KCC group included 2 online surveys of patients with kidney cancer and caregivers conducted in 2018 (the KCC survey) and the 2020 International Kidney Cancer Coalition (IKCC) survey of 241 Canadian respondents (47% with no evidence of disease, 6% with local disease, and 35% with advanced or metastatic disease) and 1 patient telephone interview conducted on November 26, 2021. In the IKCC survey, patients reported that having no access to up-to-date treatment or equipment is 1 of the top barriers to treatment. The side effects of kidney cancer therapies that were reported most often in the KCC survey included fatigue or lack of energy, diarrhea, loss of appetite, hand-foot syndrome, skin problems (including itching and rash), nausea or vomiting, pain, shortness of breath, and bleeding. Approximately one-quarter of respondents indicated the treatment was difficult to tolerate. Patients highlighted that improvement to their physical condition, such as tumour response and symptom control (breathing and pain), quality of life (QoL) improvement, and the chance for long-term disease control, are highly important considerations when deciding to take a new therapy. One clinical trial participant who was interviewed about their experience with LEN and PEM for metastatic RCC described the treatment as effective, very tolerable, and with manageable side effects (e.g., total body rash [managed with prednisone], nausea, fatigue, reduced appetite), and a reasonable QoL.

Clinician Input

Input From the Clinical Experts Consulted by CADTH

The clinical experts consulted during this CADTH review considered prolonged overall survival (OS), progression-free survival (PFS), reduction in metastatic lesions (objective response rate [ORR]), and improved QoL as the most important treatment goals. The experts noted that not all patients respond to treatments and some patients become resistant to therapy in the long run.

The experts considered ORR, PFS, and OS clinically meaningful to patients with metastatic RCC. According to the experts, a clinically meaningful response to treatment will be associated with a reduction in the size of metastatic disease by CT, reduction in pain from local metastases, and generally improved well-being of the patient. The clinical experts stated that CT imaging, history, and physical examination are commonly used in practice to assess patient response to therapy and assessments are conducted every 2 to 3 months. The clinical experts highlighted disease progression or serious autoimmune side effects related to PEM as deciding factors for treatment discontinuation. The clinical experts consulted thought that LEN-PEM will offer an additional therapy to patients with metastatic RCC in the first-line setting and patients in all IMDC risk groups will benefit from LEN-PEM.

One clinical expert highlighted that the significant benefit of the LEN-PEM treatment versus AXI-PEM is the much lower liver toxicity associated with LEN, noting that the incidence of liver toxicity with AXI-PEM is between 22% and 29%.⁸ In the opinion of the experts, differentiating liver toxicity in practice following the use of AXI instead of immunotherapy is challenging, and is often responsible for prolonged breaks from all therapy. As highlighted by 1 expert, the toxicity may be lower with LEN-PEM in terms of hepatotoxicity; however, the full toxicity profile of the combinations will only be evident in their use outside of the clinical trial setting.

Clinician Group Input

This section was prepared by CADTH staff based on the input provided by patient groups. The full clinician group input is included in the stakeholder section of this review.

Two clinician groups provided input for this CADTH review. The Ontario Health (Cancer Care Ontario) (OH-CCO) Genitourinary Drug Advisory Committee is a group that provides timely evidence-based clinical and health system guidance on drug-related issues in support of OH-CCO’s mandate, including the provincial drug reimbursement programs and the Systemic Treatment Program. The KCRNC is a virtual and inclusive national network of researchers committed to the facilitation of kidney cancer research to enhance the knowledge of kidney cancer and its treatment.

Both clinician groups highlighted improved OS and PFS, reduction in tumour size (measured as ORR), and improved QoL as treatment goals. Both clinician groups identified treatment options that were consistent with those listed by KCRNC practice guidelines for kidney cancer management. Both clinician groups identified poor response and resistance to treatment as issues faced by patients and clinicians with current treatment options. Both clinician groups anticipated that LEN-PEM will be an effective first-line option for patients with advanced RCC. Both groups considered the PFS and ORR findings from the CLEAR trial clinically significant.

Drug Program Input

The drug plans anticipate that LEN-PEM will change the comparator drug status and shift subsequent line therapies in the Canadian setting. The drug plans anticipate dose modifications in practice. The drug plans noted that LEN is available as 4 mg and 10 mg capsules, with packaging flexibility for dispensing for different treatment durations. The drug plans highlighted a potential for drug wastage for any previously dispensed supply of LEN if dose reductions are required in prescription fills (e.g., mid-cycle), as the drug cannot be re-dispensed. The implementation questions and corresponding responses from the clinical experts consulted by CADTH are summarized in Table 4.

Clinical Evidence

Pivotal Studies and Protocol-Selected Studies

Description of Studies

The CLEAR trial is an ongoing multi-centre, randomized, parallel-arm, open-label, phase III study with a primary objective to compare the efficacy and safety of LEN in combination with either everolimus or PEM versus SUN as first-line treatment in adult patients with advanced RCC. The study enrolled patients who were 18 years and older with a histologically or cytologically confirmed diagnosis of RCC with a clear cell component and documented evidence of advanced disease. Patients were also required to have at least 1 measurable target lesion assessed using Response Evaluation Criteria in Solid Tumours Version 1.1 (RECIST 1.1) criteria; adequate liver, bone marrow, blood coagulation, and renal function; a Karnofsky Performance Status (KPS) score of 70 or greater; and an adequately controlled blood pressure with or without antihypertensive medications.

The primary outcome investigated in the CLEAR trial was PFS measured by independent imaging review (IIR) using the RECIST 1.1 criteria. Secondary and exploratory outcomes included OS, ORR, health-related quality of life (HRQoL) (from 3 questionnaires: the Functional Assessment of Cancer Therapy Kidney Symptom Index – Disease Related Symptoms [FKSI-DRS], the European Organisation for Research and Treatment of Cancer Quality of Life Questionnaire Core 30 [EORTC QLQ-C30], and the EuroQol 5-Dimensions 3-Levels questionnaire [EQ-5D-3L] with the associated EuroQol Visual Analogue Scale), safety and tolerability, duration of response (DOR), and disease control rate (DCR).⁹

Patients were randomized into 3 study arms (the LEN-PEM, LEN plus everolimus, and SUN arms) in a 1:1:1 ratio based on 2 stratification factors: geographic region and the Memorial Sloan Kettering Cancer Center (MSKCC) prognostic risk groups. This CADTH review focuses on the comparison between LEN-PEM and SUN as per the sponsor’s reimbursement request and the Health Canada indication. There were more than 200 participating sites across North America (including 6 sites in Canada), Europe, Asia, and Australia.⁹ Patients received either 20 mg of LEN orally once daily plus 200 mg of PEM administered intravenously every 3 weeks, or 50 mg of SUN taken orally once daily for 4 weeks followed by 2 weeks off treatment, until the investigator discontinued treatment for the patient, the patient withdrew consent, or the patient moved into the follow-up phase.

By the third interim analysis data cut-off (August 28, 2020), a total of 1,417 patients had been screened, of which 1,069 were randomized to receive a study treatment in 1 of the 3 study arms. In total, 355 patients were randomized into the LEN-PEM arm and 357 in the SUN arm. The median age of patients enrolled in CLEAR was 62 years, more males were enrolled compared with females, and the majority of patients were White or Asian. Baseline characteristics were equally distributed among the 2 study arms except for age; more patients randomized into the SUN arm were younger than 65 years compared with the LEN-PEM arm (Table 7). More patients discontinued treatment in the SUN arm (76.5%) compared with the LEN-PEM arm (59.2%), and more patients in the SUN arm (57.7%) received subsequent systemic anti-cancer medication during survival follow-up compared with the LEN-PEM arm (33%).⁹

Efficacy Results

Table 2 provides a summary of findings for the outcomes of interest identified in the CADTH protocol.

Table 2: Summary of Key Results From Pivotal and Protocol-Selected Studies (August 28, Data Cut-Off)

AE = adverse event; CI = confidence interval; CR = complete response; DCR = disease control rate; DOR = duration of response; EORTC = European Organisation for Research and Treatment of Cancer; IxRS = interactive voice and web response system; LEN = lenvatinib; MSKCC = Memorial Sloan Kettering Cancer Center; NE = not estimable; OS = overall survival; PEM = pembrolizumab; PFS = progression-free survival; PR = partial response; SUN = sunitinib.

Note: Results are from the August 28, 2020, data cut-off unless specified otherwise.

^aHazard ratio is based on a Cox proportional hazard model including treatment group as a factor. The Efron method is used for ties.

^bStratified by geographic region (region 1 = Western Europe and North America; region 2 = rest of the world) and MSKCC prognostic groups (favourable, intermediate, and poor risk) in IxRS.

^cQuartiles are estimated by Kaplan-Meier method, and the 95% CIs are estimated with a generalized Brookmeyer and Crowley method.

^dEstimates for progression-free survival follow-up time are calculated in the same way as the Kaplan-Meier estimate of PFS but with the meaning of “censor” and “event” status indicator reversed.

^eThe 95% CI is constructed using the method of normal approximation.

^fOdds ratio and nominal P value are calculated using the Cochran-Mantel-Haenszel method, stratified by IxRS stratification factors.

^gEstimates for survival follow-up time are calculated in the same way as the Kaplan-Meier estimate of OS but with the meaning of “censor” and “event” status indicator reversed.

^hQuartiles are estimated by Kaplan-Meier method, and the 95% CIs are estimated with a generalized Brookmeyer and Crowley method.

ⁱEach patient may be counted in multiple categories.

^jInclude Medical Dictionary for Regulatory Activities preferred terms “neoplasm progression,” “malignant neoplasm.”

^kLEN or SUN. Dose reduction is not applicable for PEM.

^lDose modification includes dose reduction or drug interruption.

Source: Clinical Study Report.⁹

Progression-Free Survival

By the third interim data cut-off (August 28, 2020), a total of 365 PFS events had occurred and the median PFS was 23.9 months (95% confidence interval [CI], 20.8 to 27.7) in the LEN-PEM arm and 9.2 months (95% CI, 6.0 to 11.0) in the SUN arm. The hazard ratio (HR) obtained between the LEN-PEM arm versus the SUN arm was 0.39 (95% CI, 0.32 to 0.49; P < 0.0001). The median estimated PFS follow-up was 22.3 months (95% CI, 21.1 to 25.6) in the LEN-PEM arm and 16.6 months (95% CI, 13.1 to 18.5) in the SUN arm.⁹

The PFS in the subgroups of interest (risk groups according to the IMDC prognostic model) was as follows:

• Favourable risk group: The median estimated PFS was 28.1 months in the LEN-PEM arm and 12.9 months in the SUN arm. The HR between the LEN-PEM arm versus the SUN arm was 0.41 (95% CI, 0.28 to 0.62).

• Intermediate risk group: The median estimated PFS in the LEN-PEM arm was 22.1 months and 7.1 months in the SUN group. The HR obtained between the LEN-PEM arm and the SUN arm was 0.39 (95% CI, 0.29 to 0.52).

• Poor risk group: The median estimated PFS in the LEN-PEM arm was 22.1 months and 4 months in the SUN arm. The HR between the LEN-PEM arm versus the SUN arm was 0.28 (95% CI, 0.13 to 0.60).

Objective Response Rate

The ORR estimated by IIR in the LEN-PEM arm at the August 28, 2020, data cut-off was 71% (95% CI, 66.3 to 75.7). In total, 16.1% of patients receiving LEN-PEM had a confirmed complete response (CR) and 54.9% had a confirmed partial response (PR). In the SUN arm, the estimated ORR was 36.1% (95% CI, 31.2 to 41.1). In total, 4.2% of patients receiving SUN had a confirmed CR and 31.9% had a confirmed PR. The estimated odds ratio (OR) in the LEN-PEM arm versus the SUN arm was 4.35 (95% CI, 3.16 to 5.97) in favour of LEN-PEM.⁹

Overall Survival

The median OS by IIR was not estimable in either treatment arm at the August 28, 2020, data cut-off (interim analysis 3), and at the subsequent follow-up analysis performed on March 31, 2021. The HR estimated between the LEN-PEM arm versus the SUN arm was 0.66 (95% CI, 0.49 to 0.88; P = 0.0049).

The median duration of follow-up at the August 28, 2020, data cut-off was 26.7 months (95% CI, 25.9 to 27.4) in the LEN-PEM arm and 26.3 months (95% CI, 25.4 to 27.2) in the SUN arm. At the March 31, 2021, data cut-off, median OS was not estimable. The HR estimated between the LEN-PEM arm and the SUN arm was 0.72 (95% CI, 0.55 to 0.93). The median duration of follow-up was 33.7 months (95% CI, 32.8 to 34.4) in the LEN-PEM arm and 33.4 months (95% CI, 32.5 to 34.1) in the SUN arm.⁹

Duration of Response

By the August 28, 2020, data cut-off, the median DOR observed in patients with a response was 25.8 months (95% CI, 22.1 to 27.9) in the LEN-PEM arm and 14.6 months (95% CI, 9.4 to 16.7) in the SUN arm.⁹

Health-Related Quality of Life

HRQoL assessments between the LEN-PEM arm and the SUN arm for the EORTC QLQ-C30 questionnaire were as follows:

• The overall least squares mean difference assessments after 46 weeks of treatment for physical function was 3.01 (95% CI, 0.48 to 5.54).

• For the symptom scales, the least squares mean differences were −2.8 for fatigue (95% CI, −5.52 to −0.08), −2.79 for dyspnea (95% CI, −5.53 to −0.25), and −2.19 for constipation (95% CI, −4.19 to −0.18).

Time to First Deterioration Assessments

EORTC QLQ-C30 questionnaire: In physical functioning, the median time to first deterioration (TTD) in weeks in the LEN-PEM arm was 15.29 (95% CI, 12.29 to 21.43), while in the SUN arm, median TTD was 12.71 (95% CI, 9.29 to 18.14; nominal log-rank difference P = 0.03). The median TTD obtained in the dyspnea subscale was 39.29 (95% CI, 24.43 to 51) in the LEN-PEM arm and 21.14 (95% CI, 15.43 to 32.71) in the SUN arm (nominal log-rank difference P value = 0.02). In the appetite loss subscale, the median TTD in the LEN-PEM arm was 18.29 (95% CI, 15.14 to 21.71), while in the SUN arm, the median TTD was 9.14 (95% CI, 6.29 to 15.14). The nominal P value of the log-rank test was 0.03.

EQ-5D-3L Visual Analogue Scale: The median TTD in weeks obtained in the Visual Analogue Scale was 9.43 (95% CI, 6.43 to 12.29) in the LEN-PEM arm and, in the SUN arm, the median TTD was 9.14 (95% CI, 6.29 to 12.0). A nominal P value of 0.04 was obtained in the log-rank difference.⁹

Time Until Definitive Deterioration

FKSI-DRS total score: In the LEN-PEM arm, the median time until definitive deterioration (TUDD) in weeks was 134.14 (95% CI, 120 to not estimable), while in the SUN arm, the TUDD in weeks was 117.43 (95% CI, 90.14 to 131.29). The nominal P value obtained was less than 0.01.

EORTC QLQ-C30: The median TUDD in the global health status/QoL scale in weeks in the LEN-PEM arm was 114.29 (95% CI, 102.14 to 153.29), while in the SUN arm, the median TUDD in weeks was 75.14 (95% CI, 57.29 to 105.14). The nominal P value obtained was less than 0.0001.

In the physical function domain of the EORTC, the median TUDD in weeks in the LEN-PEM arm was 134.14 (95% CI, 109.14 to not estimable), while in the SUN arm, the median TUDD in weeks was 78.14 (95% CI, 63.14 to 111.0). The nominal P value obtained from the log-rank difference was less than 0.0001.

EQ-5D-3L Visual Analogue Scale: The median TUDD in weeks obtained in the LEN-PEM arm was 124.86 (95% CI, 94.71 to 134.57), while in the SUN arm, the median TUDD in weeks was 74.86 (95% CI, 54.14 to 94.0). The nominal P value obtained was less than 0.01.⁹

Disease Control Rate

By the August 28, 2020, data cut-off, the DCR observed in the LEN-PEM arm was 90.1%, while in the SUN arm, the DCR was 74.2%.⁹

Time to Treatment Discontinuation

This outcome was not investigated in the CLEAR trial.

Harms

Overall, the proportions of patients reporting at least 1 adverse event (AE) were comparable in both study arms (99.7% in the LEN-PEM arm and 98.5% in the SUN arm) in the CLEAR study by the August 28, 2020, data cut-off. Diarrhea, hypertension, hypothyroidism, decreased appetite, fatigue, nausea, and stomatitis were the most common AEs reported in the LEN-PEM arm, and diarrhea, hypertension, stomatitis, palmar-plantar erythrodysesthesia (PPE) syndrome, fatigue, nausea, and decreased appetite were the most commonly reported for SUN.

Serious AEs (SAEs) were reported in 50.6% of patients in the LEN-PEM arm compared with 33.2% in the SUN arm. There were more AEs leading to drug discontinuations (37.2% versus 14.4%), dose reductions (68.8% versus 50.3%), drug interruptions (78.4% versus 53.8%), and dose modifications (87.5% versus 70.3%) in the LEN-PEM arm compared with the SUN arm, respectively. Overall, more deaths were reported in the SUN arm (29.1%) compared with the LEN-PEM arm (22.2%).

The following notable harms were reported in the LEN-PEM arm and SUN arm. The notable harms observed in the LEN-PEM arm versus the SUN arm, respectively, were: hypertension (56.3% versus 42.6%), hypothyroidism (56.8% versus 32.1%), hepatotoxicity (27.3% versus 24.1%), proteinuria (29.5% versus 12.6%), hemorrhage (27.3% versus 26.5%), PPE syndrome (29.5% versus 37.9%), renal events (22.2% versus 17.6%), QT prolongation (6.5% versus 3.8%), arterial thromboembolic events (5.4% versus 2.1%), gastrointestinal perforation (1.4% versus 0.9%), hypocalcemia (1.4% versus 2.6%), cardiac dysfunction (2.6% versus 2.1%), fistula formation (0.6% versus 0.6%), and posterior reversible encephalopathy syndrome (0.6% versus 0.3%)SUN.⁹

Critical Appraisal

The CLEAR trial is a randomized, parallel-arm study. The randomization scheme implemented minimized the risk of bias owing to unknown confounders, including known and unknown prognostic factors. Baseline and demographic characteristics were balanced across the 2 study arms of interest for this review (except for age), suggesting that randomization was successful. The open-label design was the key limitation of the CLEAR trial because it increases the risk of assessment and reporting bias, especially for subjective outcomes such as HRQoL and safety. The primary outcome (PFS) and secondary outcomes (ORR, DOR, and DCR) were assessed by an IIR team using the RECIST 1.1 criteria, thus minimizing assessment bias. The time-to-event outcomes (OS, PFS) and other secondary outcomes (ORR, DOR, DCR, HRQoL, and safety) investigated in the trial were considered clinically meaningful by the clinical experts and reflective of outcomes assessed in clinical practice. The magnitude of the effect of LEN-PEM on HRQoL is uncertain because of the potential bias in reporting and attrition (the questionnaire completion rate went below 50% at cycle 26 for LEN-PEM and cycle 12 for SUN). The concomitant medications permitted (including subsequent anti-cancer therapies permitted in the follow-up phase) were also considered appropriate by the clinical experts and reflective of treatments used in Canadian practice. Several interim analyses and subgroup analyses were pre-specified in the protocol before the third interim data cut-off (August 28, 2020). The final OS analysis will take place after approximately 304 deaths are observed in the LEN-PEM arm and the SUN arm. Adjustments were made to account for alpha spending during the interim analysis. Multiplicity adjustments were implemented adequately for the analysis of PFS, OS, and ORR, and sensitivity analyses were also conducted for PFS. The findings from the sensitivity analyses were consistent with the primary intention-to-treat (ITT) analyses. No multiplicity adjustments were made during the analysis of DOR, DCR, HRQoL, and defined subgroups; thus, the findings were considered exploratory. The study was considered adequately powered to detect changes in PFS between the LEN-PEM arm versus the SUN arm. The threshold margin defined by the sponsor for PFS (including the OS), and the ORR was considered clinically significant by the clinical experts consulted.

The clinical experts consulted considered the baseline characteristics and the findings of the CLEAR trial generalizable to adult patients with untreated advanced or metastatic RCC with a clear cell component in the Canadian setting. The dosage of LEN and PEM used in the trial aligns with the Health Canada indication. SUN was considered an appropriate comparator. The experts noted that treatment options such as AXI-PEM were not available in practice for patients at the time of the trial initiation; at that time, SUN was the standard-of-care option for untreated RCC patients with advanced or metastatic disease in Canada. According to the clinical experts consulted, patients with brain metastases who had received prior treatment for brain metastasis can benefit and are eligible to receive treatment, except in cases of uncontrolled disease. Patients recruited in the CLEAR trial had better access to disease assessments and follow-up procedures compared with patients in real-world practice. The frequency of disease assessments and follow-up procedures in the CLEAR trial were considered appropriate by the clinical experts.

Indirect Comparisons

Description of Studies

One network meta-analysis (NMA) submitted by the sponsor and 4 published indirect treatment comparisons (ITCs) identified in the literature were summarized for this review. The objectives of the sponsor-submitted NMA and published ITCs were to assess the comparative clinical efficacy and/or safety of LEN-PEM compared with other first-line treatments for advanced RCC based on evidence from randomized controlled trials (RCTs).

The network informing the NMA submitted by the sponsor was composed of 24 phase II and phase III RCTs. The trials included adults with advanced or metastatic RCC who received first-line systemic treatments for advanced or metastatic RCC administered alone or in combination, best supportive care, or placebo. The studies enrolled patients between 1992 and 2019 and the study sample sizes ranged from 101 patients to 1,110 patients. A total of 18 studies reported on the timing of response assessments, which varied across studies from every 6 weeks to every 12 weeks. Among the 24 trials, the median age of the study populations ranged from 55 years to 68 years. Patients were described by risk category using the MSKCC criteria (16 studies), IMDC criteria (5 studies), or both (2 studies). Where baseline risk was reported (in all but 1 study), 23.5% to 81% of patients in each treatment group were categorized as intermediate risk. In most of the studies included in the network (21 studies), the majority of patients had either a Karnofsky score of at least 70 or an Eastern Cooperative Oncology Group (ECOG) score of 0 or 1 (less than 13% of patients included in 4 studies had an ECOG score of 2 and 80% to 83% of patients included in 1 study had a Karnofsky score of 70 or less). In all studies that reported information regarding histology (21 studies), the most common histological RCC subtype was clear cell, with at least 78% of patients possessing clear cell or predominantly clear cell histology location.

The studies included in the published ITCs were also included in the sponsor-submitted NMA. The methodology used for the published ITCs lacked important details, which hindered the ability to appropriately interpret the reported results. Further, individual estimates of treatment effects for the indirect comparisons of LEN-PEM versus other combination therapies were not reported for any outcomes. The NMA submitted by the sponsor was the most comprehensive assessment of indirect evidence among these studies and, as such, it is the focus of the following summary. The published ITCs were considered supportive of the sponsor-submitted NMA.

Efficacy Results

The summary of results herein focuses on comparisons between LEN-PEM versus the comparators included in the CADTH systematic review protocol (AXI-PEM, NIVO-IPI, PAZO, and SUN). For each of these comparators, the evidence was based on a single (different) RCT.

Progression-Free Survival

The base-case analysis of PFS (FDA censoring) used a random-effects (RE) model and included 18 comparators from 21 RCTs. The reported HR for LEN-PEM compared with the following comparators was 0.44 (95% credible interval [CrI], 0.23 to 0.82) versus NIVO-IPI, 0.57 (95% CrI, 0.31 to 1.08) versus AXI-PEM, and 0.38 (0.21 to 0.67) versus PAZO. The author indicated that the point estimates of the fixed-effects (FE) model were similar to the RE model, although the CrIs were narrower ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||. For PFS, based on an RE model, LEN-PEM showed benefit compared with NIVO-IPI and compared with PAZO. The RE model did not show a difference for the comparison with AXI-PEM, whereas the results for the FE model favoured LEN-PEM.

Overall Response Rate

The base-case analysis of ORR used an FE model and included 13 comparators from 14 RCTs. The OR for LEN-PEM was 3.24 (95% CrI, 2.18 to 4.85) compared with NIVO-IPI, 1.86 (95% CrI, 1.23 to 2.84) compared with AXI-PEM, and 3.00 (95% CrI, 2.02 to 4.47) compared with PAZO. The author reported that the CrIs were larger in the RE model, which only impacted the comparison of AXI-PEM ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||. Similar to the results for PFS, the results of the analysis of ORR based on an FE model showed a benefit of LEN-PEM when compared with other treatments.

Overall Survival

The base-case analysis of OS was performed using an FE model only and included 13 comparators from 12 RCTs. The HR for comparisons of LEN-PEM was 1.04 (95% CrI, 0.77 to 1.42) when compared with NIVO-IPI, 0.99 (95% CrI, 0.71 to 1.37) compared with AXI-PEM, and 0.78 (95% CrI, 0.58 to 1.06) compared with PAZO. These results suggest that the analysis of OS did not show a difference for LEN-PEM compared with other treatments.

Harms Results

All-Cause Grade 3 or Greater AEs

||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||

Treatment Discontinuation Due to AEs

|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||

Critical Appraisal

The methodology used for the study selection in the systematic literature review was pre-specified and used an appropriate set of criteria in terms of the study characteristics for a systematic review, the databases searched, the data extraction process, and the quality assessment. The literature review was comprehensive and was expected to have captured the relevant studies of interest. Despite an inclusive literature search, most of the connections within the network were limited to 1 study. Comparisons of interest (due to their relevance in the Canadian treatment setting) within the network were limited to indirect estimates only and based on 1 open-label RCT; therefore, inconsistency could not be assessed in these connections. Based on a qualitative review of the populations of the included studies, there were some concerns regarding potential bias due to effect modifiers. This included some differences between study populations in terms of number of metastases, prior nephrectomy, presence of sarcomatoid features, and distribution of patients by risk status that may warrant further review. This remains a source of uncertainty in the network. The quality of the included studies was assessed using Cochrane risk-of-bias tool 2.0, but information about the results of the quality assessment of the individual studies was not reported. Additionally, information about study withdrawal or dropouts was not reported, therefore limiting the ability to evaluate the internal validity of included studies.

The clinical experts consulted by CADTH indicated that the sponsor-submitted NMA considered all relevant comparators in the Canadian context. Information about the dosing of the treatments included in all of the trials that were included in the network was limited, with details regarding relative dose intensity, compliance, or missed dosing either not reported or poorly reported. The efficacy and safety outcomes included in the NMA were clinically relevant, but HRQoL was not included, which was a limitation of the sponsor-submitted NMA. Some of the patient characteristics were inconsistently reported across trials and, in particular, details about race and ethnicity, PD-L1 status, and cancer staging were infrequently reported. In general, heterogeneity that was identified as a limitation was not adjusted for, but some sensitivity and subgroup analyses were performed. Subgroup analyses were limited by sample size (patients in the poor- and favourable risk subgroups represented a small proportion of patients in the overall population). Overall, the interpretation of the results for the subgroup analyses of the NMA is limited.

Differences in time point assessments and actual treatment duration were also acknowledged as a limitation of the NMA, as was the impact of a lack of data maturity on efficacy assessments. A sensitivity analysis was conducted where trials with a follow-up period of less than 12 months were excluded; however, no adjustments were made for the variation in follow-up duration in studies where the duration was greater than 12 months. For reference, in the CLEAR trial,¹⁰ the analysis of OS was based on data with a median follow-up of approximately 33 to 34 months and the analysis of PFS was based on a median follow-up of 26 to 27 months. The results for OS and PFS were based on a median follow-up of 43 months in the KEYNOTE-426 trial,¹¹ and a minimum of 48 months in CheckMate-214.¹² The impact of the heterogeneity in the follow-up duration on these outcomes is unknown.

The sponsor-submitted ITC included justification of model selection (FE versus RE) based on an assessment of model fit or a lower deviance information criterion (DIC), although reported differences were very small. Assessments of heterogeneity based on I² and inconsistency were also considered, although most connections were formed by a single RCT and there were few closed loops. The RE model used an informative before stabilize estimates of between-study variance. The prior was based on plausible values, and sensitivity analyses were conducted. There was uncertainty in the results, with wide CrIs. This is likely due to the sparsity of the network. The results for the ORR had very wide CrIs and the results for OS and all-cause AEs of grade 3 or higher included CrIs that crossed 1 and included values suggesting a strong treatment effect, limiting interpretation of these results. The analysis of treatment discontinuation due to AEs was also associated with a lack of precision and uncertainty from wide CrIs that crossed 1 while including values suggesting a strong treatment effect, although the FE model improved precision.

Conclusions

One pivotal study and 5 ITCs provided evidence for the CADTH systematic review. This review focused on the comparison between LEN-PEM versus SUN investigated in the CLEAR trial as per the sponsor’s reimbursement request and the Health Canada indication. No other evidence directly comparing LEN-PEM with other standard therapies for advanced or metastatic RCC was identified. In CLEAR, the median PFS estimated by IIR at the final interim analysis for PFS (August 28, 2020) was 23.9 months in patients receiving LEN-PEM compared with 9.2 months in patients receiving SUN. The HR estimated for PFS between LEN-PEM against SUN was considered statistically and clinically significant. The median OS was not estimable in both study arms at the data cut-off for interim analysis 3 and at the follow-up analysis data cut-off of March 31, 2021. However, the HR estimated between LEN-PEM against SUN was considered statistically significant. The ORR estimated in the LEN-PEM arm was also considered statistically significant. The HRQoL assessments were considered exploratory due to the lack of multiplicity adjustments in the analysis and the potential for reporting bias. The findings of the CLEAR trial were considered by the clinical experts consulted during the review to be meaningful for patients with advanced or metastatic RCC and were aligned with outcomes of importance to patients. In the opinion of the clinical experts, clinical judgment is required to evaluate LEN-PEM’s clinical benefit and management of AEs in practice. The experts anticipate that the treatment-related adverse events (TRAEs) resulting from the use of LEN-PEM will be managed in practice using similar strategies already in place for other treatment options (frequent AE monitoring and dose adjustments, reductions, and modifications are anticipated for this treatment). The open-label design was a key limitation of the CLEAR trial, and the OS data are interim. The study was randomized and adjustments for multiplicity were conducted for key outcomes (PFS, OS, and ORR), which minimized bias in the study. The clinical experts considered the baseline characteristics and the findings from the CLEAR trial generalizable to patients in Canada diagnosed with advanced or metastatic RCC in the first-line setting with at least a clear cell component.

No direct evidence was available to assess the relative efficacy of LEN-PEM versus other current standard-of-care therapies. Indirect evidence of LEN-PEM for first-line treatment of patients with advanced or metastatic RCC was available based on 5 ITCs: 1 NMA submitted by the sponsor and 4 ITCs identified in published literature. The sponsor-submitted NMA of LEN-PEM compared with other available therapies showed benefit for LEN-PEM for PFS and ORR but not for OS, compared with other therapies. Sources of uncertainty identified during the review included heterogeneity in the RCTs, a sparse network, and a lack of data maturity (shorter follow-up duration) for the CLEAR trial. The sponsor-submitted NMA results of the analysis of treatment discontinuation due to AEs ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||, although these results were limited by a lack of precision in addition to a number of assumptions made about the outcome that cause uncertainty in the results. Findings of OS, PFS, and ORR obtained from 4 additional published ITCs assessed in this review were consistent with the results of the sponsor-submitted NMA. However, the methodology used for the analyses lacked important details, which hindered the ability to appropriately interpret the reported results.

Introduction

Disease Background

RCC is the most common form of kidney cancer, accounting for more than 85% of all cases across the world.² RCCs are further classified into different subtypes based on histology (clear cell, papillary, chromophobe, clear cell papillary, collecting duct, medullary, and unclassified). The clear cell component is the most prevalent form of RCC and represents more than 70% of all RCC cases in practice.^3,4 More than 33% of cases identified at initial diagnosis include metastatic disease⁵ due to the fact that most patients experience few or no symptoms at earlier stages, which restricts the number of cases of early disease identified.⁶ Common symptoms are blood in urine, dull pain around the flank region that does not go away, fullness in the upper abdomen or a lump in this area, fever, appetite loss, nausea, vomiting, constipation, weakness, fatigue, anemia, polycythemia, and unexplained weight loss.^3,4,6 In metastatic disease, patients may experience additional symptoms such as bone pain, adenopathy and pulmonary symptoms, anemia, or varicocele.² Disease staging at diagnosis predicts prognosis and facilitates treatment choice in real-world settings.^2,3 CT scans, MRIs, X-rays, and bone scans are common diagnostic methods used for identifying and characterizing tumours and assessing disease progression in patients in both real-world and clinical trial settings.^2,3

Projected estimates in Canada in 2021 showed that kidney and renal pelvis cancers were the seventh most diagnosed cancers in males (accounting for 5,200 new cases and 2.8% of disease-related deaths) and the 12th most diagnosed in females (2,600 new cases and 1.7% of disease-related deaths). The predicted 5-year age-standardized survival rate was 73% for both sexes. The predicted net survival was higher (92%) for patients aged 15 to 44 years compared with patients 85 years and older (33%).¹³ The 5-year survival rate is said to be highly dependent on key factors such as tumour stage, grade, and local extent of tumour; the presence of regional nodal metastasis and evidence of metastatic disease at presentation are key determinants. The 5-year relative survival for localized (stage I) kidney and renal pelvis cancer was 92.7% and falls to 13.9% for patients with metastatic disease.¹⁴ Established risk factors include smoking, hypertension, obesity, medications (over-the-counter pain killers, phenacetin-containing compounds, and diuretics), family history of RCC, and genetic conditions (von Hippel-Lindau disease) or hereditary papillary RCC.^3,4,6

The KCRNC practice guideline recommends the use of prognostic models, particularly the IMDC model for managing patients with advanced or metastatic RCC and for treatment selection in the first-line setting.^2,7 The IMDC model relies on 6 clinical parameters to characterize patients into 3 risk groups (favourable, intermediate, and poor) which are stratified based on hemoglobin less than the lower limit of normal; serum-corrected calcium greater than the upper limit of normal (ULN); KPS of less than 80%, time from initial diagnosis to initiation of therapy of less than 1 year, absolute neutrophil count greater than ULN, and platelets greater than ULN. Patients are classified into the favourable risk group if they possess none of these 6 factors, the intermediate risk group if they have 1 or 2 factors, and the poor risk group if they have 3 to 6 adverse factors. Clinical decision-making for individualized therapy relies on a patient’s risk category.^2,7

Standards of Therapy

Treatment options for RCC are based on the IMDC risk group classification, which is supported by available evidence from clinical trials and real-world data.⁷ The KCRNC guidelines outline the following options for untreated RCC patients with advanced disease.

For patients who fall under the favourable risk category, the “preferred” options outlined include AXI-PEM and NIVO plus cabozantinib. “Other” options include SUN and PAZO. Of note, therapies classified as “other” have studies that demonstrated PFS and not necessarily OS survival, and those labelled as “preferred” have studies that demonstrated improvement in OS.

For patients in the intermediate or poor risk category, the preferred options outlined include NIVO-IPI, AXI-PEM, and NIVO plus cabozantinib. Other available options for this risk group include SUN, PAZO, and cabozantinib. Of note, cabozantinib received market approval from Health Canada on October 6, 2021, as a first-line treatment option for patients with advanced RCC who fall under the intermediate or poor IMDC risk group.

Treatment options identified by the clinical experts and clinician groups consulted during this CADTH review were consistent with those outlined in the 2021 KCRNC consensus guidelines.

Drug

LEN is a multiple-RTK inhibitor that selectively inhibits kinase activities of VEGFR1 (FLT1), VEGFR2 (KDR), and VEGFR3 (FLT4), in addition to other proangiogenic and oncogenic pathway–related RTKs. LEN also inhibits other kinases that have been implicated in pathogenic angiogenesis, tumour growth, and cancer progression in addition to their normal cellular functions, including fibroblast growth factor (FGF) receptors (FGFR) 1, 2, 3, and 4; platelet-derived growth factor receptor alpha (PDGFRA), RTK (KIT proto-oncogene, RTK [KIT]), and rearranged during transfection (RET). LEN also exhibits antiproliferative activity in hepatocellular carcinoma cell lines dependent on activated FGFR signalling with concurrent inhibition of FGFR substrate 2 alpha phosphorylation. In syngeneic mouse tumour models, LEN treatment decreased the tumour-associated macrophage population and increased activated cytotoxic T-cell populations, leading to antitumour activity. The antitumour activity of the combination of LEN and an anti–PD-1 monoclonal antibody was greater than that of monotherapy.⁹

PD-1 is an immune-checkpoint receptor that limits the activity of T lymphocytes in peripheral tissues. The PD-1 pathway is an immune-control checkpoint that may be engaged by tumour cells to inhibit active T-cell immune surveillance. PEM is a high-affinity antibody against PD-1, which exerts dual ligand blockade of the PD-1 pathway, including PD-L1 and PD-L2, on antigen-presenting or tumour cells. By inhibiting the PD-1 receptor from binding to its ligands, PEM reactivates tumour-specific cytotoxic T lymphocytes in the tumour microenvironment.¹⁵

LEN underwent a standard review by Health Canada and received a Notice of Compliance on May 5, 2022, for the following indication: in combination with PEM, for the treatment of adult patients with advanced (not amenable to curative surgery or radiation) or metastatic RCC with no prior systemic therapy for metastatic RCC. The Health Canada–approved dose is 20 mg (two 10 mg capsules) of LEN taken orally once daily in combination with PEM 200 mg administered as an IV infusion over 30 minutes every 3 weeks or 400 mg every 6 weeks. This is continued until unacceptable toxicity or disease progression or for up to 24 months, or until administration of thirty-five 200 mg doses or eighteen 400 mg doses, whichever is longer. After completing combination therapy, LEN may be administered as a single drug until disease progression or unacceptable toxicity.

Table 3: Key Characteristics of Lenvatinib, Pembrolizumab, and Sunitinib

FGFR = fibroblast growth factor receptor; FLT = FMS-like tyrosine kinase; KDR = kinase insert domain receptor; KIT = KIT proto-oncogene, receptor tyrosine kinase; PD-1 = programmed cell death 1 protein; PD-L1 = programmed cell death 1 ligand 1; PD-L2 = programmed cell death 1 ligand 2; PDGF = platelet-derived growth factor; PDGFRA = platelet-derived growth factor receptor alpha; PDGFRB = platelet-derived growth factor receptor beta; RCC = renal cell carcinoma; RET = rearranged during transfection; RTK = receptor tyrosine kinase; VEGF = vascular endothelial growth factor; VEGFR1, VEGFR2, VEGFR3 = vascular endothelial growth factor receptor 1, 2, 3.

^aHealth Canada–approved indication.

Source: Product monographs for Keytruda,¹⁵ lenvatinib,¹ and sunitinib.¹⁶

Stakeholder Perspectives

Patient Group Input

This section was prepared by CADTH staff based on the input provided by patient groups. The full patient group input is included in the Stakeholder Input section of this review.

Two patient groups, CanCertainty and KCC, provided input for this submission. The CanCertainty Coalition is the united voice of more than 30 Canadian patient groups, cancer health charities, and caregiver organizations from across the country joining together with oncologists and cancer care professionals to significantly improve the affordability and accessibility of cancer treatment. The KCC group is a national community of patients, caregivers, and health professionals who work to provide every Canadian touched by kidney cancer with support, education, and advocacy for their care pathways and treatment options.

The CanCertainty group expressed concerns related to inconsistent provincial coverage for oncology treatment regimens containing orally administered drugs and the resulting financial burden on vulnerable patients.

The KCC group included 2 online surveys of patients with kidney cancer and caregivers conducted in 2018 (the KCC survey) and the 2020 IKCC survey (comprising 241 Canadian respondents: 47% with no evidence of disease, 6% with local disease, and 35% with advanced or metastatic disease) and 1 patient telephone interview conducted on November 26, 2021. In the IKCC survey, patients reported that having no access to up-to-date treatment or equipment is 1 of the top barriers to treatment. The side effects of kidney cancer therapies that were reported most often in the KCC survey include fatigue or lack of energy, diarrhea, loss of appetite, hand-foot syndrome, skin problems (including itching and rash), nausea or vomiting, pain, shortness of breath, and bleeding. Approximately one-quarter of respondents indicated the treatment was difficult to tolerate. Patients highlighted that improvement to their physical condition, such as tumour response and symptom control (breathing and pain), QoL improvement, and the chance for long-term disease control are highly important considerations before deciding to take a new therapy. One clinical trial participant who was interviewed about their experience with LEN and PEM for metastatic RCC described the treatment as effective, very tolerable, and with manageable side effects (e.g., total body rash [managed with prednisone], nausea, fatigue, reduced appetite), and a reasonable QoL.

Clinician Input

Input From the 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 RCC.

Unmet Needs

Prolonged OS, PFS, reduction in the size of metastatic lesions, and improved QoL were considered by the clinical experts to be the most important treatment goals for patients with advanced metastatic RCC. In the opinion of the clinical experts, not all patients respond to treatment in current practice and those who respond often become resistant to therapy in the long run.

Place in Therapy

According to the clinical experts consulted, LEN-PEM will provide an additional first-line treatment option, in addition to AXI-PEM, NIVO-IPI, SUN, and PAZO, for the treatment of patients with advanced or metastatic RCC in Canada. The clinical experts highlighted that since the CLEAR trial was designed to investigate LEN-PEM against SUN in the first-line setting, the therapy would only be appropriate for patients in the first-line setting.

Patient Population

In the opinion of the clinical experts consulted, patients in all IMDC prognostic risk groups will benefit from the LEN-PEM therapy. The experts cited that the diagnosis of metastatic RCC is based on imaging and histology confirmation, neither of which is challenging to perform, and both are associated with a low probability of misdiagnosis. The clinical experts noted that patients with metastasis often present with asymptomatic disease; therefore, it is very unlikely that a treating oncologist will wait for symptom presentation before initiating treatment in patients. The experts noted there are no clear contraindications to LEN-PEM except in patients with pre-existing autoimmune disorders; these patients may have a higher risk of developing AEs with any form of immunotherapy and thus are less suited to receive LEN-PEM in practice. The experts also noted that although some patients may do better on treatment with NIVO-IPI, owing to the absence of biomarkers, predicting which patients will be least suitable for treatment or who will likely exhibit a response to LEN-PEM is uncertain.

Assessing Response to Treatment

The clinical experts highlighted CT imaging, history, and physical examination as common diagnostic methods used to assess response to treatment in real-world practice. The experts considered the outcomes of ORR, PFS, and OS to be clinically meaningful for patients. As cited by the experts, a favourable outcome following treatment is 1 that is associated with a reduction in the size of metastatic disease (assessed by CT), a reduction in pain from local metastases, and a general improvement in patient well-being. The clinical experts stated that treatment response is assessed every 2 to 3 months in real-world settings.

Discontinuing Treatment

According to the clinical experts, disease progression or serious autoimmune AEs related to PEM will be considered when deciding treatment discontinuation in patients. The experts further noted that SAEs from LEN are rare and can be managed with dose reduction.

Prescribing Conditions

The clinical experts consulted during the review thought that it will be appropriate for patients to be generally seen and treated by a medical oncologist experienced in using both tyrosine kinase inhibitors and immunotherapy, as considerable clinical judgment is required in evaluating clinical benefit and managing toxicities. The experts noted there is a small number of oncology urologists in Canada who administer systemic therapy and have the required experience and expertise to administer this therapy.

Additional Considerations

One clinical expert highlighted that the significant benefit of the LEN-PEM treatment over AXI-PEM is the much lower liver toxicity associated with LEN, noting that the incidence of liver toxicity with AXI-PEM is between 22% to 29%.⁸ In the opinion of the experts, differentiating liver toxicity in practice following the use of AXI instead of immunotherapy is challenging and it is often responsible for prolonged breaks from all therapies. As highlighted by the other expert, the toxicity may be lower with LEN-PEM in terms of hepatotoxicity; however, the full toxicity profile of the combinations will only be evident in their use outside of the clinical trial setting.

Clinician Group Input

This section was prepared by CADTH staff based on the input provided by clinician groups. The full clinician group input is included in the stakeholder input section of this review.

Two clinician groups, the OH-CCO Genitourinary Drug Advisory Committee and the KCRNC, provided input for this CADTH review. The OH-CCO’s Drug Advisory Committee provides timely evidence-based clinical and health system guidance on drug-related issues in support of OH-CCO’s mandate, including the provincial drug reimbursement programs and the Systemic Treatment Program. The KCRNC is a virtual and inclusive national network of researchers committed to the facilitation of kidney cancer research to enhance the knowledge of kidney cancer and its treatment.

The OH-CCO and the KCRNC clinician groups both highlighted improved OS and PFS, reduction in tumour size (measured as ORR), and improved QoL as treatment goals. Both clinician groups identified treatment options that were consistent with the Canadian guidelines for kidney cancer management. Both clinician groups identified poor response and resistance to treatment as issues with current treatment options. The OH-CCO group added that patients with advanced RCC are not routinely cured by current therapies and, as such, have developed resistance to treatment over time, causing patients to die of the disease. Treatment options for refractory disease were identified as an important unmet need by the clinician groups. Both clinician groups anticipated that LEN-PEM will be an effective first-line option for patients with advanced RCC across all IMDC risk groups. The clinician groups consulted considered the ORR (71%) and CR rate (16%) obtained in the LEN-PEM arm to be the highest compared with the other immunotherapy and tyrosine kinase inhibitor combinations available in practice. The clinician groups also considered the PFS (23.9 months) in the LEN-PEM arm the longest PFS and clinically significant.

According to the OH-CCO group, a patient’s IMDC group classification, suitability for immunotherapy, tumour burden, and preference will be assessed when deciding treatment discontinuation. The OH-CCO clinician group stated that it will be uncommon to recommend other systemic therapies before starting LEN-PEM. Both clinician groups consulted considered improved or stable clinical status, stable disease, reduction in pain from local metastases, and shrinkage (reduction in the size) of the disease based on radiographic imaging (i.e., CT scan) to be clinically meaningful. Both groups noted that, consistent with other funded treatment options, patients will be assessed for a response based on history, physical examination, and radiographic imaging (most commonly CT scans, usually every 2 to 3 months). Both clinician groups highlighted that disease progression, SAEs from PEM (such as high-grade immune-related AEs) or high-grade AEs from LEN (despite dose reduction or schedule change), will be considered when deciding treatment discontinuation.

Additional Considerations

The KCRNC group highlighted that a significant benefit of LEN-PEM versus the AXI-PEM combination is the much lower probability of liver toxicity with LEN. The KCRNC group cited that the incidence of liver toxicity with AXI-PEM is 22% to 29%.⁸ As underlined by the clinician group, liver toxicity is often responsible for prolonged breaks from all therapy and, in practice, it is challenging to differentiate liver toxicity resulting from AXI versus immunotherapy.

Drug Program Input

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

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

AXI = axitinib; CNS = central nervous system; IPI = ipilimumab; LEN = lenvatinib; PAZO = pazopanib; PEM = pembrolizumab; mRCC = metastatic renal cell carcinoma; MSKCC = Memorial Sloan Kettering Cancer Center; pERC = CADTH pan-Canadian Oncology Drug Review Expert Review Committee; RCC = renal cell carcinoma; SUN = sunitinib.

Clinical Evidence

The clinical evidence included in the review of LEN-PEM is presented in 3 sections. The first section, the systematic review, includes pivotal studies provided in the sponsor’s submission to CADTH and Health Canada, as well as those studies that were selected according to an a priori protocol. The second section includes indirect evidence from the sponsor and indirect evidence selected from the literature that met the selection criteria specified in the review. The third section includes 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 LEN 20 mg, taken orally once daily, in combination with PEM (200 mg IV administered once every 3 weeks), for the treatment of adult patients with advanced or metastatic RCC with no prior systemic therapy for metastatic RCC.

Methods

Studies selected for inclusion in the systematic review included pivotal studies provided in the sponsor’s submission to CADTH and Health Canada, as well as those meeting the selection criteria presented in Table 5. Outcomes included in the CADTH review protocol reflect outcomes considered to be important to patients, clinicians, and drug plans.

Of note, the systematic review protocol presented in Table 5 was established before the granting of a Notice of Compliance from Health Canada.

Table 5: Inclusion Criteria for the Systematic Review

IMDC = International Metastatic Renal Cell Carcinoma Database Consortium; LEN = lenvatinib; NIVO-IPI = nivolumab plus ipilimumab; PAZO = pazopanib; PEM = pembrolizumab; PEM-AXI = pembrolizumab plus axitinib; RCC = renal cell carcinoma; RCT = randomized controlled trial; SUN = sunitinib.

^aFor poor and intermediate risk groups.

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

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.

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 tool.¹⁷

Published literature was identified by searching the following bibliographic databases: MEDLINE All (1946‒) through Ovid and Embase (1974‒) through Ovid. All Ovid searches were run simultaneously as a multi-file search. 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 lenvatinib (Lenvima) and pembrolizumab (Keytruda). Clinical trials registries were searched: the US National Institutes of Health’s clinicaltrials.gov, the WHO’s International Clinical Trials Registry Platform (ICTRP) search portal, Health Canada’s Clinical Trials Database, and the European Union Clinical Trials Register.

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

The initial search was completed on December 9, 2021. Regular alerts updated the search until the meeting of the CADTH pan-Canadian Oncology Drug Review Expert Committee (pERC) on April 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 tool.¹⁸ Included in this search were the websites of regulatory agencies (FDA and European Medicines Agency [EMA]). 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 by reviewing bibliographies of key papers and through contacts with appropriate experts. In addition, the manufacturer of the drug was contacted for information regarding unpublished studies.

Findings From the Literature

Two reports of a single study were identified from the literature for inclusion in the systematic review (Figure 1). The included studies are summarized in Table 6.

Figure 1: Flow Diagram for Inclusion and Exclusion of Studies

Table 6: Details of the CLEAR Study

AE = adverse event; CNS = central nervous system; CR = complete response; DCR = disease control rate; DOR = duration of response; ECOG PS = Eastern Cooperative Oncology Group Performance Score; HRQoL = health-related quality of life; IA = interim analysis; IIR = independent imaging review; LEN = lenvatinib; ORR = objective response rate; OS = overall survival; PD = pharmacodynamic; PEM = pembrolizumab; PFS = progression-free survival; PFS2 = PFS with next line of therapy; PK = pharmacokinetics; RCC = renal cell carcinoma; RECIST 1.1 = Response Evaluation Criteria in Solid Tumours Version 1.1; SUN = sunitinib; TBD = to be determined.

Note: Other reports were included in the review: FDA¹⁹ and European Medicines Agency²⁰ reports and the proposed Canadian product monograph.¹

Description of Studies

CLEAR is an ongoing multi-centre, randomized, open-label, parallel-arm, phase III study with a primary objective to compare the efficacy and safety of LEN in combination with either everolimus or PEM versus SUN as a first-line treatment for adult patients with advanced RCC. Patients enrolled at baseline were 18 years and older with a histologically or cytologically confirmed diagnosis of RCC with a clear cell component, documented evidence of advanced disease, and at least 1 measurable target lesion according to RECIST 1.1 criteria. Patients were also eligible if they had adequate liver, bone marrow, blood coagulation, and renal function; a KPS score of 70 or greater; and adequately controlled blood pressure with or without antihypertensive medications.⁹ The primary outcome (PFS), was assessed by an IIR and tumour response assessments were based on RECIST 1.1 guidelines. Other secondary and exploratory outcomes of interest for the CADTH review included OS, ORR, HRQoL, safety and tolerability, DOR, and DCR. Stratified randomization using an interactive voice and web response system (IxRS) was implemented across participating trial sites. Patients were stratified based on 2 predefined factors:

• geographic region: region 1 (Western Europe and North America) and region 2 (rest of the world)

• MSKCC prognostic groups: favourable, intermediate, and poor risk

In total, 1,417 patients were screened at baseline and 1,069 were randomized in a 1:1:1 ratio to receive 1 of 3 therapies (LEN-PEM, LEN plus everolimus, or SUN) in planned parallel treatment arms, as presented in Figure 2.

The study was first initiated on October 13, 2016, and is currently ongoing in more than 200 centres across North America (6 sites in Canada), Europe, Asia, and Australia. CLEAR is an open-label study with patients and investigators aware of the treatments being administered. Blinding of the sponsor’s personnel (Eisai clinical, biostatistics, and imaging core laboratory personnel) was achieved, and data integrity was maintained using a data integrity protection plan and an operational and communication plan. The study was conducted in 3 phases: a pre-randomization phase (with a screening and a baseline period), a randomization phase (with treatment and follow-up period), and an extension phase, which are further described subsequently.⁹ Figure 2 summarizes the study design of the CLEAR trial.

This CADTH review focuses on comparisons between arm B (LEN-PEM) and arm C (SUN) as outlined in the sponsor’s reimbursement request and the Health Canada indication.⁹

Figure 2: CLEAR Study Design

R = randomization.

^a Extension phase includes a treatment and follow-up period. all patients still on treatment at the end of the randomization phase will enter the extension phase and continue to receive the same study treatment they received in the randomization phase.

^b Lenvatinib 18 mg plus everolimus 5 mg given orally once daily.

^c Lenvatinib 20 mg once daily plus pembrolizumab 200 mg intravenously every 3 weeks.

^d Sunitinib 50 mg once daily on a schedule of 4 weeks on treatment followed by 2 weeks off.

Source: Clinical Summary Report.⁹

Study Phases

Pre-Randomization Phase

This phase lasted no longer than 28 days. It included a screening period (which occurred between day −28 and day −3) that allowed investigators to obtain patient consent and establish patient eligibility. The baseline period was also included in this phase and allowed confirmation of patient eligibility and establishment of baseline disease characteristics.⁹

Randomization

The randomization phase began when the first patient was randomized to any treatment arm in the study and ended at the data cut-off for the third planned interim analysis (final PFS analysis) on August 28, 2020. The randomization phase included a treatment period and a follow-up period.⁹

Treatment Period

This period began for each patient after randomization and ended with the completion of the off-treatment visit, which occurred within 30 days after the final dose of study treatment. Study treatment was administered in a 21-day cycle and these cycles were counted continuously irrespective of any dose interruptions.⁹

Follow-Up Period

This phase began the day after a patient had completed their off-treatment visit. This period continued if the patient was alive, unless the patient withdrew consent, was lost to follow-up, or the sponsor terminated the study. Patients were followed every 12 weeks (± 1 week) for PFS with the next line of therapy (PFS2), survival, and all subsequent anti-cancer and local standard-of-care treatments received from the investigator. Patients who had discontinued the study treatment before disease progression underwent tumour assessments every 8 weeks and a bone scan every 24 weeks until disease progression was documented and confirmed by IIR, or a new anti-cancer therapy was initiated, unless the patient withdrew consent or was lost to follow-up.⁹

Extension Phase

The extension phase consisted of patients still receiving study treatments (in any of the arms) or those who had transitioned into the follow-up phase by the interim data cut-off. The extension phase was divided into a treatment and follow-up period. In the treatment phase, patients still receiving a study drug at the interim analysis 3 cut-off continued treatment in the extension phase, while the follow-up phase was made up of patients whose disease had progressed during the randomization phase and patients who had discontinued any of the study drugs. Patients remained in the follow-up phase if they were alive, unless they withdrew consent, were lost to follow-up, or the study was terminated by the sponsor.⁹

Populations

Inclusion and Exclusion Criteria

The patients enrolled were 18 years or older and had a histologically or cytologically confirmed diagnosis of RCC with a clear cell component with documented evidence of advanced disease. They were required to have at least 1 measurable target lesion according to RECIST 1.1 guidelines; adequate liver, bone marrow, blood coagulation, and renal function; a KPS score of 70 or greater; and adequately controlled blood pressure with or without antihypertensive medications.

Patients who had received prior systemic cancer therapy for RCC, had a history of a significant cardiac impairment within the past 12 months, had a history of or current non-infectious pneumonitis that required steroid treatment, had a history of organ allograft, or had tested positive for HIV, hepatitis B, or hepatitis C were excluded from the CLEAR trial. Those with central nervous system (CNS) metastases were eligible if they had received local therapy (e.g., whole brain radiation therapy, surgery, or radiosurgery) and had discontinued the use of corticosteroids for at least 4 weeks before the initiation of study treatment.⁹

Baseline Characteristics

At the third interim analysis data cut-off (August 28, 2020), the median age of the patients randomized into the CLEAR study was 62 years; more males were enrolled compared with females and a majority of patients were White or of Asian descent. There were more patients with a KPS score of 80 or greater in the 2 arms of interest compared with patients with a KPS score of less than 80. Baseline characteristics were balanced across the 2 study arms with the exception of age (more patients randomized in the SUN arm were younger than 65 years compared with the LEN-PEM arm). Table 7 presents the baseline summary in the ITT population.

Table 7: Summary of Baseline Demographic and Disease Characteristics (ITT Population)

IMDC = International Metastatic Renal Cell Carcinoma Database Consortium; ITT = intention to treat; LEN = lenvatinib; MSKCC = Memorial Sloan Kettering Cancer Center; PD-L1 = programmed cell death 1 ligand 1; PEM = pembrolizumab; RCC = renal cell carcinoma; SD = standard deviation; SUN = sunitinib.

Note: Percentages may not total 100 because of rounding. One patient in the LEN-PEM group had carcinoma without a clear cell component. Data cut-off date: August 28, 2020.

^aPatients may be represented in more than 1 category.

^bKarnofsky Performance Status scores range from 0 to 100, with lower scores indicating greater disability. Scores were missing for 2 patients: 1 in the LEN plus everolimus group and 1 in the SUN group.

^cAn MSKCC score of 0 indicates favourable risk, a score of 1 or 2 indicates intermediate risk, and a score of 3 or higher indicates poor risk.

^dAn IMDC score of 0 indicates favourable risk, a score of 1 or 2 indicates intermediate risk, and a score of 3 to 6 indicates poor risk.

^ePD-L1 expression was assessed with the PD-L1 immunohistochemistry 22C3 pharmDx assay (Agilent Technologies) and reported as the combined positive score, defined as the number of PD-L1–staining cells (tumour cells, lymphocytes, and macrophages) divided by the total number of viable tumour cells, multiplied by 100.

^fKidneys were not included in the number of metastatic organs or sites. The only tumour location in the kidney applied to 3 patients (0.8%) in the LEN plus everolimus group, 4 patients (1.1%) in the LEN-PEM group, and 3 patients (0.8%) in the SUN group.

^gFour common sites of metastasis are shown. Patients may have had metastasis at more than 1 site.

Source: Clinical Study Report⁹ and Motzer et al. (2021).¹⁰

Interventions

Patients were randomized to receive treatments in 1 of 3 study arms:

• LEN plus everolimus (arm A), consisting of LEN 18 mg orally once daily plus everolimus 5 mg orally once daily in each 21-day cycle

• LEN-PEM (arm B), consisting of LEN 20 mg orally once daily plus PEM 200 mg IV every 3 weeks during each 21-day cycle

• SUN (arm C), consisting of SUN 50 mg orally once daily taken for 4 weeks on treatment followed by 2 weeks off.

In the LEN-PEM treatment arm, the starting dose of LEN was 20 mg per day. Planned LEN dose reductions were designed to occur in succession based on the patient’s previous dose level (14 mg, 10 mg, and 8 mg per day and so forth). The starting dose of PEM was 200 mg per administration. PEM was withheld, interrupted, or discontinued in the event of a drug-related toxicity and severe or life-threatening AEs. In the SUN treatment arm, the starting dose of SUN was 50 mg per day based on 4 weeks on and 2 weeks off treatment schedule. SUN dose reductions were allowed in succession based on the patient’s previous dose level (37.5 mg per day or 25 mg per day on a schedule of 4 weeks on and 2 weeks off treatment). Once a dose of a study drug had been reduced, it was not increased at a later date unless the dose had been mistakenly decreased; in that situation, the sponsor’s approval was required to increase the dose.⁹

Treatment Discontinuation Criteria

Patients received the study drugs during the treatment period until disease progression confirmed by independent review, loss of clinical benefit based on investigator assessment (upon confirmation by IIR and imaging core laboratory), development of unacceptable toxicity, patient request, withdrawal of consent, completion of 35 treatments (approximately 2 years) with PEM, or study termination by the sponsor.

Treatment discontinuation during the treatment period was allowed for patients who had:

• attained a confirmed CR

• been treated for at least 8 cycles (at least 24 weeks) with PEM

• received at least 2 treatments with PEM beyond the date when the initial CR was declared.

In the presence of clinical benefit, patients in the LEN-PEM arm who had discontinued PEM were allowed to continue treatment with LEN alone unless any of the other discontinuation criteria applied. Patients were also permitted to continue receiving PEM alone after discontinuing LEN. Patients were permitted to continue the study treatment beyond RECIST 1.1–defined disease progression as long as the treating investigator considered that the patient was tolerating the study drug and had clinical benefit. Patients who discontinued treatment owing to disease progression or loss of clinical benefit could receive alternative treatment at the investigator’s discretion.⁹

Concomitant Medication

Concomitant medications were permitted and properly documented in the CLEAR trial (Table 14). Patients were prohibited from using other anti-cancer therapies during the treatment phase of the trial. Prohibited therapies included chemotherapy, tyrosine kinase inhibitors, antitumour interventions, cancer immunotherapy, and radiotherapy (except for palliative radiotherapy for up to 2 painful pre-existing, non-target bone metastases). Patients were also discouraged from receiving other concurrent investigational drugs, live vaccines, and systemic glucocorticoids (except for the purpose of modulating AE symptoms of immunologic etiology) during the screening and treatment phase. Physiologic doses of corticosteroids (up to 10 mg per day) were permitted.⁹

Outcomes

A list of the efficacy end points identified in the CADTH review protocol that were assessed in the CLEAR trial and included in this review is provided in Table 8. These outcomes are standard outcomes approved by regulatory agencies (Health Canada, FDA,²¹ and EMA²²) for oncology trials. They were also considered clinically meaningful to patients by the clinical experts and clinician groups consulted during the review.

Table 8: Summary of Outcomes of Interest Identified in the CADTH Review Protocol

CR = complete response; DCR = disease control rate; DOR = duration of objective response; HRQoL = health-related quality of life; IIR = independent imaging review; ORR = objective response rate; OS = overall survival; QLQ-C30 = Quality of Life Questionnaire Core 30; PFS = progression-free survival; PR = partial response; RECIST 1.1 = Response Evaluation Criteria in Solid Tumours Version 1.1; SAE = serious adverse event; TEAE = treatment-emergent adverse event.

Source: Clinical Study Report.⁹

Efficacy Measurement for Primary and Secondary Outcomes

Tumour assessments (consisting of CT scans of the chest and CT or MRI scans of the abdomen, pelvis, and other known or suspected sites of disease) were performed at pre-randomization, every 8 weeks from the date of randomization in treatment cycles in the randomization phase, and as per investigator decision in the extension phase. For each patient, the same imaging modality and image-acquisition protocol was used consistently across all time points. Bone scans were performed within 6 weeks before randomization, and those performed within a target of 1 week but no more than 2 weeks following a CR were assessed by the investigator. Brain scans were performed at screening and as clinically indicated thereafter, and within a target of 1 week but no more than 2 weeks following achievement of a CR. For patients with a history of protocol-eligible treated brain metastases, a brain scan was required at all tumour assessment time points.⁹

All patients were permitted to continue treatment beyond initial RECIST 1.1–defined progression as long as the investigator believed that the patient was still receiving clinical benefit and was tolerating the study drug treatment. Clinical benefit was defined as:

• absence of signs and symptoms of disease progression (including laboratory results)

• no decline in performance status

• absence of rapid progression of disease

• absence of progressive tumour(s) at critical sites requiring urgent intervention (e.g., spinal cord compression).

Patients who discontinued study treatment without disease progression in the randomization phase were allowed to continue to undergo tumour assessments every 8 weeks and bone scans every 24 weeks in the follow-up period until disease progression was documented or another anti-cancer therapy was initiated.

Patients who discontinued study treatment without tumour progression in the extension phase had tumour assessments performed as clinically indicated following the prevailing local standard of care, at the investigator’s discretion. Copies of tumour assessment scans were no longer required to be sent to the imaging core laboratory and an independent review was not conducted in the extension phase.⁹

Health-Related Quality of Life

HRQoL measures were assessed using the generic cancer HRQoL instrument (EORTC QLQ-C30), the RCC-specific HRQoL instrument (FKSI-DRS), and the generic HRQoL instrument, EQ 5D-3L. A detailed discussion and critical appraisal of these outcomes is available in Appendix 2.

Functional Assessment of Cancer Therapy Kidney Symptom Index – Disease Related Symptoms

The FKSI-DRS is a kidney cancer–specific, patient-reported instrument that evaluates disease-related symptoms.²³ The questionnaire consists of 9 questions that assess the symptoms of kidney cancer deemed by patients and clinicians to be the most important to monitor (lack of energy, pain, weight loss, bone pain, fatigue, shortness of breath, cough, fever, blood in urine) when treating advanced kidney cancer.²³ Evidence of convergent validity and discriminative validity, internal reliability, and adequate responsiveness to change has been demonstrated in patients with RCC. The minimal important difference (MID) estimated using different anchors ranged from 0.62 to 3 points.

EORTC Quality of Life Questionnaire Core 30

The EORTC QLQ-C30 is 1 of the most commonly used patient-reported outcome measures in oncology clinical trials.²⁴ It is a cancer-specific, multi-dimensional measure of HRQoL that is designed to assess the change in HRQoL in clinical trial participants in response to treatments.²⁵ The QLQ-C30 consists of 30 questions that are scored to create 5 multi-item function scales (physical, role, cognitive, emotional, and social), 3 multi-item symptom scales (fatigue, pain, nausea and vomiting), 6 single-item symptom scales (dyspnea, insomnia, appetite loss, constipation, diarrhea, and financial impact) and a 2-item global health status/QoL scale.²⁶ The validity, reliability, and responsiveness of the EORTC QLQ-C30 has not been evaluated in patients with advanced RCC.

EQ-5D-3L

The EQ-5D-3L is a generic, utility-based measure of HRQoL comprising 2 components: the EuroQol descriptive system and the EuroQol VAS.²⁷ For the EuroQol descriptive system, respondents are assigned a 5-digit descriptive health state based on their report on 5 health status dimensions that day (mobility, self-care, usual activities, pain/discomfort, and anxiety/depression).²⁷ Three response options are available for each dimension to reflect the 3 possible levels of functioning (level 1 = no problems, level 2 = some problems, level 3 = extreme problems). An index score is then calculated by applying a population-specific (e.g., UK, US) utility function to the health state vector. A score of 0 represents the health state “dead” and 1.0 reflects “perfect health.” A negative score represents the health state that society considers to be “worse than dead.”²⁷ The validity, reliability, and responsiveness of the EQ-5D-3L has not been evaluated in patients with advanced RCC.

Patients enrolled in the CLEAR trial completed questionnaires at baseline, cycle 2 day 1, cycle 3 day 1, and cycle 4 day 1 to day 1 of the last cycle, and when they were off treatment. Change of QoL from baseline to different cycle dates was assessed using a mixed-model analysis. The least squares mean change from baseline with the 95% CI was calculated.

• TTD was defined as the number of weeks between randomization and the first deterioration event. A deterioration event for any particular HRQoL outcome was defined as a detrimental change in score, relative to baseline, that exceeded the MID for that score. TTD was calculated and compared using the Kaplan-Meier method stratified log-rank tests. TUDD was defined as the number of weeks between randomization and the earliest deterioration event with no subsequent recovery above the deterioration threshold or no subsequent HRQoL assessment data.

  • Thresholds defined in the study:

    • FKSI-DRS: Decrease of 3 points or greater

    • EORTC QLQ-C30 functional and the global health status/QoL score: Decrease of 10 points or greater

    • EORTC QLQ-C30 symptom scores: Increase of 10 points or greater

    • EQ-5D-3L index: Decrease of 0.08 points or greater; EQ-5D-3L Visual Analogue Scale: decrease of 7 points or greater.

No adjustments for multiplicity were performed during the analysis. All P values obtained were 2-sided and the CIs were nominal and descriptive. All randomized patients with available HRQoL data who had received at least 1 dose of the study treatment were included in the analysis unless otherwise specified.

Harms Outcomes

All safety analyses were performed on the safety analysis set. Safety data presented by treatment groups were summarized on an “as treated” basis using descriptive statistics. Safety variables assessed included treatment-emergent adverse events (TEAEs), clinical laboratory parameters, vital signs, 12-lead electrocardiogram results, and echocardiogram results, including left ventricular ejection fraction. Categorical variables were summarized by number and percentage. Continuous variables were summarized using descriptive statistics. Due to the longer period of follow-up at interim analysis 3 (August 28, 2020, data cut-off and final PFS analysis) compared with interim analysis 2 (November 15, 2019, data cut-off), safety analyses presented in this report are based on the August 28, 2020, cut-off.⁹

Statistical Analysis

Sample Size and Power Calculation

The sample size for the study was estimated based on the primary end point, PFS, and it was estimated that approximately 1,050 patients would be randomized into the 3 arms in a 1:1:1 ratio and stratified based on geographic regions (Western Europe and North America versus Other) and MSKCC prognostic groups (favourable, intermediate, and poor risk).⁹

The same treatment effect was assumed for the 2 planned primary comparisons between the 3 study arms (LEN plus everolimus [arm A] and LEN-PEM [arm B], each compared with SUN alone [arm C]). Based on the assumption that a median PFS of 12.3 months would be obtained in the SUN arm, and an HR of 0.714 estimated between the LEN plus everolimus versus SUN and lenvatinib plus PEM versus SUN arms, the sponsor presumed that this would correspond to a 40% improvement (4.9 months) in median PFS from 12.3 months to 17.2 months in both the LEN plus everolimus versus SUN and LEN-PEM versus SUN comparisons. A yearly loss of PFS event rate of 22% was assumed.

For the 2 PFS comparisons (1 for each test arm), an alpha of 0.0499 (2-sided) was split, as initial allocations, into an alpha of 0.045 for the comparison between LEN-PEM and SUN, and an alpha of 0.0049 for the comparison between LEN plus everolimus versus SUN.⁹

The study was designed to achieve 90% power at an alpha of 0.045 to detect a statistically significant difference in PFS between LEN-PEM and SUN. A total of 388 PFS events were expected to occur over both the LEN-PEM and the SUN arms for the final PFS analysis and the same number (388) for the final PFS analysis comparison between LEN plus everolimus versus SUN (the same number of events were expected to occur over both arms).

The power to detect a statistically significant difference in PFS between LEN plus everolimus versus SUN was approximately 70% at the initial assigned alpha of 0.0049. It was expected to be at least 90% after alpha reallocation when the hypothesis tests of PFS and OS in the comparison of LEN-PEM and SUN were considered statistically significant.

In the calculation of power for the PFS analysis, the sponsor assumed that 1 interim analysis of PFS would be performed at the 80% information fraction and a Lan-DeMets spending function with an O’Brien-Fleming boundary would be used between the interim and final analysis of PFS. The analysis was planned for approximately 4 months after the last patient was randomized when approximately 310 PFS events should have occurred in the LEN-PEM arm and the SUN arm.

For the OS analysis, a total of 304 deaths in each comparison (456 death events among the 3 arms) were expected in the final OS analysis. For the OS testing, it was assumed that when the corresponding PFS testing was considered statistically significant at the initial assigned alpha, the study would provide 80% power to detect a statistically significant difference at an alpha level of 0.045 for the comparison between LEN-PEM and SUN, and 50% power at an alpha level of 0.0049 for the comparison between LEN plus everolimus versus SUN. The following assumptions were made for the OS power calculations:

• The HR assumed was 0.70 (a median OS of 54.1 months in the LEN plus everolimus arm or LEN-PEM arm, and 37.9 months in the SUN arm)

• interim analyses occurred when the information fraction for death events was approximately 45%, 60%, and 80%

• a Lan-DeMets spending function with Pocock boundary was used in the sponsor’s analysis to control alpha levels

• the yearly rate for loss to follow-up was assumed to be 3%

Based on the planned sample size and given assumptions, the final analysis of OS was expected to occur approximately 69 months after the first patient was randomly assigned to treatment.

The ORR was obtained by assuming an ORR of 32% in arm C and 48% in the LEN plus everolimus arm or the LEN-PEM arm; the study was expected to provide at least a 95% power to detect a difference with alpha reallocation when testing for PFS and OS was positive for each comparison of LEN-PEM versus SUN, and LEN plus everolimus versus SUN.⁹

Analyses, Multiple-Testing Procedure, and Alpha Spending

Multiplicity testing adjustments were made using the overall familywise error rate for the primary outcome (PFS) and 2 secondary outcomes (OS and ORR). The Maurer and Bretz approach presented in Figure 3 was implemented for the primary end point and key secondary end points (OS and ORR). Multiplicity adjustments were not made for other secondary end points or exploratory analyses. Subgroup analyses were considered exploratory outcomes and, thus, were not accounted for multiplicity (no adjustments made). An alpha of 0.0001 was subtracted from the total alpha of 0.05 to account for the interim analysis of ORR from the LEN-PEM arm. Figure 3 shows the initial alpha allocation (the remaining alpha was 0.0499) for each hypothesis and the graphical approach for multiple analyses of PFS, OS, and ORR.⁹

Figure 3: Graphical Approach to Control Familywise Error Rate for Testing Primary and Key Secondary End Points

α = alpha; EP = end point; ORR = objective response rate; OS = overall survival; PFS = progression-free survival.

Note: Hypothesis 1 (H₁): The PFS of the lenvatinib plus pembrolizumab arm is superior to that of the sunitinib arm. Hypothesis 2 (H₂): The PFS of the lenvatinib + everolimus arm is superior to that of the sunitinib arm. Hypothesis 3 (H₃): The OS of the lenvatinib plus pembrolizumab arm is superior to that of the sunitinib arm. Hypothesis 4 (H₄): The OS of the lenvatinib plus everolimus arm is superior to that of the sunitinib arm. Hypothesis 5 (H₅): The ORR of the lenvatinib plus pembrolizumab arm is superior to that of the sunitinib arm. Hypothesis 6 (H₆): The ORR of the lenvatinib plus everolimus arm is superior to that of the sunitinib arm.

Source: Clinical Study Report.⁹

Adjustments for Covariates

Cox proportional hazard models and log-rank tests were obtained for PFS and OS, and the Cochrane-Mantel-Haenszel test was used to determine the ORR, stratified by region (Western Europe and North America, rest of world), and MSKCC risk group (favourable, intermediate, and poor risk). Analyses of PFS included other covariates of interest based on the subgroups.

Planned Analyses

There were 4 planned interim analyses and 1 planned final study analysis, which are summarized in Table 9. These interim efficacy analyses were conducted by an independent statistical group with no other responsibilities for the study. The safety monitoring was conducted by an independent data management committee (DMC) that had exclusive access to data with treatment information.⁹ The frequency of the safety reviews was defined in the DMC charter. The safety monitoring and interim analyses for PFS and OS were planned to assess whether to stop the trial based on the review of the safety and efficacy findings with treatment information.

Table 9: Summary of Interim and Final Efficacy Analyses

DOR = duration of response; IF = information fraction; OS = overall survival; ORR = objective response rate; progression-free survival; TBD = to be determined.

^aThe P value for hypothesis testing of ORR will be based on the ORR data at analysis number 2.

Source: Clinical Study Report.⁹

Primary Outcome Analysis

PFS was assessed by IIR using the RECIST 1.1 guidelines. The Lan-DeMets spending function with an O’Brien-Fleming boundary was used to determine the nominal alpha level for each PFS comparison at the first interim analysis of PFS (corresponds to interim analysis 2) (nominal alpha for hypothesis 1 [H₁] = 0.0216 and H₂ = 0.0014) and final analysis of PFS at interim analysis 3 (nominal alpha = 0.0386 for H₁ and 0.0046 for H₂). For each comparison, statistical significance could be claimed based on either the interim analysis or final analysis for PFS at the specified alpha levels (Table 9).⁹ Statistical significance was observed at interim analysis 2 and was consistent with PFS findings at interim analysis 3.

The PFS curve in each treatment group was estimated using the Kaplan-Meier method and the difference in PFS for each of the 2 primary comparisons was tested by stratified log-rank test. The tests were stratified by geographic region and MSKCC prognostic groups. The HR (for LEN plus everolimus relative to SUN and for LEN-PEM relative to SUN) and the corresponding 95% CIs were estimated using the Cox regression model and the Efron method for ties, stratified by the factors used for stratified randomization. The median PFS, and the PFS rates at various time points, were calculated using the Kaplan-Meier product-limit estimates for each treatment arm and presented with 2-sided 95% CIs.

The final analysis of PFS was performed when approximately 388 PFS events assessed by the IIR were observed for each comparison. A graphical approach was used to control the familywise error rate at a 2-sided alpha of 0.0499 for multiple comparisons, including PFS, OS, and ORR comparisons of LEN-PEM versus SUN and LEN plus everolimus versus SUN. For each comparison, a statistical significance was claimed based on either the interim or final analysis of PFS at specified alpha levels.

PFS censoring rules and the definition of progression date were based on FDA guidance²¹ and EMA guidelines.²² The ITT dataset was used to determine the primary outcome, and the secondary efficacy outcomes analyses were based on the per-protocol dataset. Censoring rules for PFS are captured in Table 10.

Table 10: Censoring Rules for Derivation of Progression-Free Survival

CR = complete response; PD = progressive disease; PFS = progression-free survival; PR = partial response; SD = stable disease.

^aAdequate tumour assessment is a radiologic assessment of CR, PR, SD, non-CR or non-PD, or PD as determined by investigators at regular intervals as defined in the protocol. Any tumour assessments after new anti-cancer treatment starts will be removed in the definition of PFS.

^bMore than 1 missed visit or adequate tumour assessment is defined as a duration between the last adequate tumour assessment and PD or death of longer than 16 weeks plus 10 days (tumour assessment window) minus 1 day, which is 121 days for patients on every 8-week tumour assessment schedule in this study.

Source: Clinical Study Report.⁹

Sensitivity Analyses

Sensitivity analyses were performed using an unstratified log-rank test for the comparisons of the PFS of LEN plus everolimus versus SUN alone, and LEN-PEM versus SUN alone, as well as using the unstratified Cox proportional hazards model with the Efron method used for ties, including treatment arms as a single covariate for the estimation of the HR.

The additional sensitivity analyses performed included the following:

• The actual reported date of progression by IIR or death was used to define PFS, regardless of missing assessments or use of new anti-cancer therapy (as per the EMA guidance document).²²

• The radiologic assessment data (as assessed by the investigator) and death were used to define PFS.

• The different derivation rule was used when a patient had more than 1 consecutive missed visit or tumour assessment and death or progression followed immediately after more than 1 missed visit or tumour assessment (i.e., if a patient missed 2 or more tumour assessments right before disease progression or death), the patient was censored on the date of the last adequate tumour assessment before disease progression or death. If a patient was censored by both this criterion and the anti-cancer treatment criterion, the earliest censoring date was used.

Subgroup Analyses

HRs and 2-sided 95% CIs were derived for comparing PFS, OS, and ORR (by the IIR and investigator) in the study arms (the LEN plus everolimus arm versus SUN, or the LEN-PEM arm versus SUN). Forest plots with the median PFS and corresponding 95% CIs were constructed for each predefined subgroup analysis. Similar summary statistics and plots were determined for the OS. The OR and a 2-sided 95% CI for comparing ORR as assessed by IIR was also summarized and presented in forest plots. The following subgroups were investigated in the CLEAR trial:

• age group (< 65 years, ≥ 65 years)

• sex (male, female)

• race (White, Asian, all others)

• geographic region (Western Europe and North America, rest of world) per IxRS

• MSKCC risk group (favourable, intermediate, poor) per IxRS

• IMDC risk group (favourable, intermediate, poor)

• number of metastatic sites per IIR (0, 1, 2, ≥ 3)

• KPS score group (100 to 90, 80 to 70)

• baseline bone metastasis (yes, no)

• baseline liver metastasis (yes, no)

• baseline lung metastasis (yes, no)

• PD-L1 status (combined positive score ≥ 1, < 1, or not available)

• prior nephrectomy (yes, no)

• histologic clear component featuring sarcomatoid (yes, no).

This CADTH review identified 3 subgroups of interest in the protocol that were based on the IMDC prognostic model (favourable, intermediate, and poor risk).

Secondary Outcome Analysis

Table 11 presents the statistical analyses of the key efficacy outcomes investigated in the CLEAR trial.⁹

Table 11: Statistical Analysis of Efficacy End Points in the CLEAR Trial

CI = confidence interval; DOR = duration of response; HRQoL = health-related quality of life; KM = Kaplan-Meier; NR = not reported; ORR = objective response rate; OS = overall survival; PFS = progression-free survival.

Source: Clinical Study Report.⁹

Overall Survival

The OS between treatment arms was determined using the stratified log-rank test with geographic region and MSKCC prognostic groups as strata. The HR and its 95% CI comparing LEN plus everolimus versus SUN and LEN-PEM versus SUN was estimated using the stratified Cox proportional hazards model and the Efron method for ties, stratified by geographic region and MSKCC prognostic groups. Median OS and the OS rate at various time points with 2-sided 95% CIs were calculated using the Kaplan-Meier product-limit estimates for each treatment arm, and Kaplan-Meier estimates of OS were plotted over time. The Lan-DeMets spending function with Pocock boundary was used to control alpha levels in the interim and final analyses of OS. In the first 2 interim analyses, the OS was determined based on approximately 45% and 60% of the information fractions on OS events (Table 9). All events related to death were included irrespective of whether patients were on therapy or had discontinued. Patients still alive at data cut-off were censored, including those who were discontinued because they were lost to follow-up or withdrew consent (censored at their last date known to be alive).

Objective Response Rate

The ORR (CR plus PR) was estimated based on tumour response according to IIR assessment using the RECIST 1.1 guidelines. It was calculated with an exact 95% CI using the Clopper and Pearson method. Differences between treatment arms (the LEN plus everolimus arm versus SUN, and the LEN-PEM arm versus SUN) were tested using the Cochran-Mantel-Haenszel test, stratified by geographic region and MSKCC prognostic groups. A 2-sided 95% CI for the OR and the difference in ORR were calculated. A P value was calculated for the hypothesis test based on the ORR data at the time of the PFS analysis.

Duration of Response

The median DOR among responders for each treatment arm was determined along with corresponding 2-sided 95% CIs. Censoring rules for DOR were the same as those for PFS.

Disease Control Rate

DCR was calculated with exact 95% CIs using the Clopper and Pearson method. The differences and ORs of the previously described rates (OS, ORR, DOR) between treatment arms and corresponding 2-sided 95% CIs were calculated. These analyses were performed based on both IIR and investigator assessments. The null hypothesis of no difference in DCR and clinical benefit rate comparing the LEN plus everolimus arm versus the SUN and LEN-PEM arm versus SUN alone was tested using the Cochran-Mantel-Haenszel test stratified by geographic region and MSKCC prognostic groups.⁹

Handling of Missing Data, Dropouts, and Outliers

AEs with incomplete start dates were considered treatment-emergent if the:

• day and month are missing and the year is equal to or after the year of the first dose date

• day is missing and the year is after the year of the first dose

• day is missing and the year is equal to the year of the first dose date and the month is equal to or after the month of the first dose date

• year is missing

• complete date is missing.

Medications were considered concomitant if the:

• day and month are missing and the year is equal to or after the year of the first dose date

• day is missing and the year is after the year of the first dose

• day is missing and the year is equal to the year of the first dose date and the month is equal to or after the month of the first dose date

• year is missing

• complete date is missing.

For incomplete dates involving efficacy and other safety data, a conservative imputation was calculated as needed.⁹

Analysis Populations

• Full analysis set: used for all efficacy analyses (ITT population) and consisted of all randomized patients regardless of the treatment received.

• Per-protocol set: used for all secondary analyses for efficacy end points. It was composed of patients who had received at least 1 dose of the study drug, had no major protocol deviations, and had a baseline and at least 1 post-baseline tumour assessment

• Safety analysis set: consisted of patients who had received at least 1 dose of any study drug and the analysis was based on the as-treated principle.

• HRQoL analysis set: all patients who had any HRQoL data and received at least 1 dose of the study treatment.

Protocol Amendments

The original protocol (v.1.0) was approved on June 22, 2016.²⁸ There were 7 protocol amendments made by the August 28, 2020, data cut-off date.

• Amendment 1 (September 16, 2016): This included secondary end points and updates to the exclusion criteria. The proportion of patients who discontinued treatment due to toxicity and time to treatment failure due to toxicity were added as a new secondary end point. The characterization of the population pharmacokinetics of PEM was added as an exploratory objective.

• Amendment 2 (February 03, 2017): This included updates to the follow-up period and the inclusion of secondary end points. The assessment of PFS by investigator was added as a secondary objective. A pregnancy assessment was added to the follow-up period.

• Amendment 3 (January 30, 2018): This made revisions to the exclusion criteria and added dose modifications. Exclusion criterion 15 was updated (the cardiovascular impairment window was extended from 6 months to 12 months). Revisions were made to the management of notable outcomes proteinuria, hypertension, and hemorrhage.

• Amendment 4 (June 30, 2018): This updated the planned patient enrolment number and made revisions to the exclusion criteria. The planned enrolment was increased to 1,050 patients (approximately 350 patients per arm) to address slow enrolment in the first 12 months and the high loss of PFS events and to provide adequate power for intergroup comparisons of OS. Two interim analyses were added. For the primary analysis of PFS, the alpha was decreased to 0.0499 for all comparisons due to the addition of an interim analysis for which an alpha of 0.0001 was allocated. For the multiplicity adjustment, the P value thresholds for the primary analysis of PFS were changed because of the addition of an interim analysis.

• Amendment 5 (December 19, 2018): This updated the interim analysis outcomes (ORR and DOR) and clarified that the results may be considered for an early submission to regulatory agencies in regions outside of EMA jurisdiction.

• Amendment 6 (September 10, 2019): The protocol for the interim analysis of OS and the multiplicity strategy was updated. The sponsor added the interim analysis of PFS, OS, and ORR.

• Amendment 7 (August 6, 2020): The sponsor removed an exploratory objective to assess PFS using immune-related RECIST criteria in patients treated with LEN-PEM.

All protocol amendments were submitted to the appropriate health authorities and institutional review boards or independent ethics committees for information and approval in accordance with local requirements.⁹

Changes to the Planned Analyses

All changes to the planned analyses outlined in the original protocol were documented in the final statistical analysis plan (version 3.0) (August 14, 2020). The 95% CI for the ORR and DCR was calculated for each treatment arm using the normal approximation method as opposed to the Clopper and Pearson method because, in the sponsor’s opinion, the sample size was large enough to use normal approximation.

Post Hoc Analyses

To evaluate the impact of subsequent anti-cancer medication received post-treatment (during the follow-up phase) on the treatment effect of LEN-PEM versus SUN on OS, Kaplan-Meier plots of LEN-PEM versus SUN were constructed in patients treated with and without subsequent anti-cancer medication.⁹

Results

Patient Disposition

Enrolment in the CLEAR study was completed on July 24, 2019. By the third interim data cut-off (August 28, 2020), a total of 1,417 patients had been screened and 1,069 randomized to receive a study treatment in any 1 of the 3 study arms. Of the total who were screened out, 268 patients (18.9%) failed to meet the inclusion criteria or met exclusion criteria, 34 (2.4%) withdrew consent, 10 (0.7%) had AEs, 2 (0.1%) were lost to follow-up, and 34 (2.4%) failed for reasons categorized as “other.” Only the findings in study arm B (LEN-PEM) and arm C (SUN) are reported in this CADTH review.

In total, 142 patients (40%) in the LEN-PEM arm, and 67 patients (18.8%) in the SUN arm were still receiving treatment at the data cut-off for interim analysis 3 (August 28, 2020). The total number of patients who discontinued treatment was higher in the SUN arm (n = 273) compared with the LEN-PEM arm (n = 210). Table 12 presents the patient disposition in the LEN-PEM and SUN arms in the CLEAR trial.⁹

Table 12: Patient Disposition

LEN = lenvatinib; PEM = pembrolizumab; NA = not applicable; SUN = sunitinib.

Note: Percentages are based on the total number of patients in the full analysis set within the relevant treatment group. Data cut-off date: August 28, 2020.

^aAs reported in the Patient Disposition section of the electronic case report form.

^b“Treatment ongoing” is based on data available in the database at the time of data cut-off. Patients receiving SUN or at least 1 study drug in combination therapy are deemed to have treatment ongoing in absence of an off-treatment visit, or with a treatment ongoing at data cut-off in the patient disposition (randomization phase) section of the electronic case report form.

^cTreatment discontinuation includes patients who discontinued SUN or both study drugs in combination therapy.

^dDiscontinued from study refers to patients who were no longer followed up for survival as of the cut-off date.

Source: Clinical Study Report.⁹

Major Protocol Deviations

Major protocol deviations were as follows:

• Exclusion criteria (2 patients: 1 in the LEN-PEM arm): One patient was enrolled with active CNS metastasis and 1 patient had significant cardiovascular impairment.

• Inclusion criteria: 1 patient was enrolled without histological confirmation of RCC with a clear cell component (in the LEN-PEM arm).

• 1 patient in the LEN-PEM arm had other prohibited concomitant medications or procedures.

• Prohibited concomitant nondrug therapy: 10 patients received a prohibited anti-cancer procedure (tumour resection or radiation therapy: n = 3 in the LEN-PEM arm; n = 4 in the SUN arm) during the study, leading to tumour assessments that were not evaluable and the censoring of PFS events by IIR.

• Tumour assessment: 5 patients (n = 2 in the LEN-PEM arm; n = 2 in the SUN arm) missed more than 1 consecutive tumour assessment scan, leading to censoring of PFS events by IIR.⁹

Exposure to Study Treatments

Treatment duration was defined as the duration between the start date of the first study drug and the end date of the last study drug. Exposure to treatment was obtained from the safety analysis set. The median duration on treatment observed by data cut-off (August 28, 2020) was 17 months in the LEN-PEM arm and 7.84 months in the SUN arm. Exposure to treatments in the LEN-PEM arm was 2.5 times longer than patient exposure in the SUN arm. Table 13 presents treatment exposure to study drugs for the LEN-PEM arm (arm B) in comparison with SUN (arm C) in the CLEAR trial.⁹

Table 13: Exposure to Study Treatments

LEN = lenvatinib; PEM = pembrolizumab; NA = not applicable; SD = standard deviation; SUN = sunitinib.

Note: Percentages are based on the total number of patients in the safety analysis set within the relevant treatment group. Data cut-off date: August 28, 2020.

^aDuration of treatment (months) = (date of last dose of study drug minus date of the first dose of the study drug plus 1) divided by 30.4375. Overall duration of treatment is defined as the duration between the earliest first dose start date of either medication or the latest last dose end date of either medication.

Source: Clinical Study Report.⁹

Concomitant Medications

Overall, 96.4% of patients received at least 1 concomitant medication during the study and this rate was comparable in the 2 arms of interest (Table 14). The most frequently administered prior medications were in the Anatomical Therapeutic Chemical (ATC) pharmacologic subclasses of plain lipid-modifying drugs (25.9%) and antithrombotic drugs (22.8%).⁹ Table 14 lists the concomitant medications reported in at least 30% of patients enrolled in either arm in the CLEAR trial.

Table 14: Concomitant Medications Reported in Patients in Any Treatment Arm by Pharmacologic Subclass — Full Analysis Set

ATC = Anatomical Therapeutic Chemical; LEN = lenvatinib; PEM = pembrolizumab; SUN = sunitinib.

Note: Percentages are based on the total number of patients in the full analysis set within the relevant treatment group. Concomitant medications include medications that either: started before the first dose of the study drug and were continuing at the time of the first dose of study drug, or started on or after the date of the first dose of the study drug up to 30 days after the patient’s last dose. Patients with 2 or more medications within an ATC level (or drug name) are counted only once within that ATC level (or drug name). Medications were coded using the WHO Drug Dictionary version WHODDMAR20B3G. Data cut-off date: August 28, 2020.

Source: Clinical Study Report.⁹

Anti-Cancer Medications Received by Patients During the Follow-Up Phase

The use of post-treatment anti-cancer medications was permitted at the survival follow-up phase after study treatment had been discontinued in patients. The proportion of patients receiving subsequent anti-cancer medications was higher in the SUN arm (57.7%) compared with the LEN-PEM arm (33.0%). The most common were anti-VEGF therapies (n = 328; 30.7%) and PD-1 and PD-L1 checkpoint inhibitors (n = 309; 28.9%). Use of PD-1 and PD-L1 checkpoint inhibitors was greatest in the SUN arm (43.1%) when compared with the LEN-PEM arm (8.2%).²⁸ Table 15 presents a summary of anti-cancer treatments received by patients in the LEN-PEM arm versus the SUN arm during the survival follow-up phase of CLEAR.⁹

Table 15: Anti-Cancer Medications Approved During Survival Follow-Up at Interim Analysis 3 — Full Analysis Set

CTLA-4 = cytotoxic T-lymphocyte-associated protein 4; LEN = lenvatinib; MTOR = mammalian target of rapamycin; PD-1 = programmed cell death 1 protein; PD-L1 = programmed cell death 1 ligand 1; PEM = pembrolizumab; SD = standard deviation; SUN = sunitinib; VEGF = Visual Analogue Scale.

Note: Percentages are based on the total number of patients in the full analysis set within the relevant treatment group. Patients with 2 or more anti-cancer medications may be counted in multiple categories. Medications were coded using WHO Drug Dictionary version WHODDMAR20B3G. Data cut-off date: August 28, 2020.

^aMapping/coding is based on verbatim = XmAb20717, which is a bi-specific antibody for PD-1 and CTLA-4.

Source: Clinical Study Report.⁹

Efficacy

Only those efficacy outcomes and analyses of subgroups identified in the review protocol are reported subsequently. The findings presented were obtained at interim analysis 3 (August 28, 2020, data cut-off).

Progression-Free Survival (By Independent Imaging Review)

By the August 28, 2020, data cut-off (which also corresponds to the third and final interim analysis of PFS), a total of 365 PFS events had occurred in the LEN-PEM and SUN arms. The median PFS estimated by IIR was 23.9 months (20.8 to 27.7) in the LEN-PEM arm and 9.2 months (6.0 to 11.0) in the SUN arm. The HR between the LEN-PEM versus SUN arms was 0.39 (95% CI, 0.32 to 0.49; P < 0.0001).⁹ The sponsor reviewed PFS data at interim analysis 2 (November 15, 2019, data cut-off) retrospectively, and statistical significance was observed. Findings at interim analysis 2 were consistent with the estimates obtained at the data cut-off for interim analysis 3. Table 16 provides a summary of PFS by IIR per RECIST 1.1 criteria in the LEN-PEM arm versus the SUN arm. Figure 4 presents the Kaplan-Meier plot of PFS at the data cut-off for interim analysis 3.

Table 16: Progression-Free Survival at Interim Analysis 3 — Full Analysis Set

CI = confidence interval; IxRS = interactive voice and web response system; LEN = lenvatinib; MSKCC = Memorial Sloan Kettering Cancer Center; PEM = pembrolizumab; SUN = sunitinib.

Note: Percentages are based on the total number of patients in the full analysis set within the relevant treatment group. Data cut-off date: August 28, 2020.

^aQuartiles are estimated by Kaplan-Meier method, and the 95% CIs are estimated with a generalized Brookmeyer and Crowley method.

^bEstimates for progression-free survival follow-up time are calculated in the same way as the Kaplan-Meier estimate of progression-free survival but with the meaning of “censor” and “event” status indicator reversed.

^cHazard ratio is based on a Cox proportional hazard model including treatment group as a factor, Efron method is used for ties.

^dStratified by geographic region (region 1: Western Europe and North America; region 2: rest of the world) and MSKCC prognostic groups (favourable, intermediate, and poor risk) in IxRS.

^eProgression-free survival rate and 95% CIs are calculated using Kaplan-Meier product-limit method and Greenwood formula.

Source: Clinical Study Report.⁹

PFS Subgroup Analyses

Risk Groups Based on the IMDC Prognostic Model: The CADTH review protocol identified 3 subgroups — the IMDC prognostic model risk groups (favourable, intermediate, and poor) — for the systematic review. Results are presented for the LEN-PEM versus SUN comparison.

• Favourable risk group: By the August 28, 2020, data cut-off, a total of 43 patients out of 110 in the LEN-PEM arm had events, and the median PFS was 28.1 months. In the SUN arm, a total of 67 patients out of 124 had events, and the estimated median PFS was 12.9 months. The HR between the LEN-PEM arm and the SUN arm in the favourable risk group was 0.41 (95% CI, 0.28 to 0.62).

• Intermediate risk group: By the August 28, 2020, data cut-off, a total of 97 patients out of 210 had events in the LEN-PEM arm and the estimated median PFS was 22.1 months. In the SUN arm, 110 of 192 patients had events and the estimated median PFS was 7.1 months. The HR obtained between the LEN-PEM arm and the SUN arm was 0.39 (95% CI, 0.29 to 0.52).

  Figure 4: Kaplan-Meier Plot of Progression-Free Survival at Interim Analysis 3 — Full Analysis Set

  

  + = censored observations; CI = confidence interval; HR = hazard ratio; IxRS = interactive voice and web response system. LEN+PEMBRO = lenvatinib + pembrolizumab; mo = months; SUN = sunitinib; RECIST = Response Evaluation Criteria in Solid Tumours; PFS = progression-free survival; vs = versus.

  Note: Median is estimated by Kaplan-Meier method, and the 95% CIs are estimated with a generalized Brookmeyer and Crowley method. HR is estimated from a Cox proportional hazard model including treatment group as a factor and stratified by IxRS stratification factors; the Efron method is used for ties. P value is calculated using log-rank test stratified by IxRS stratification factors. Data cut-off date: August 28, 2020.

  Source: Clinical Study Report.⁹

• Poor risk group: By the August 28, 2020, data cut-off, a total of 18 patients out of 32 had events in the LEN-PEM arm at a median PFS of 22.1 months. In the SUN arm, 26 of the 37 patients in this group had events, with an estimated median PFS of 4.0 months. The HR obtained between LEN-PEM and SUN was 0.28 (95% CI, 0.13 to 0.60).

The subgroup analysis of PFS indicated that LEN-PEM showed a benefit over SUN in the 3 risk groups. All subgroup analyses were considered exploratory because no adjustments for multiplicity were made.⁹

PFS Sensitivity Analysis

By the August 28, 2020, data cut-off, the unstratified HR obtained from the unstratified Cox regression model was 0.42 (95% CI, 0.34 to 0.52; P < 0.0001) for the LEN-PEM versus SUN comparison. The sensitivity analysis was consistent with the findings observed in the primary analysis of PFS and showed that the benefit of using LEN-PEM was maintained over SUN. Three additional sensitivity analyses for PFS were conducted and showed results that were consistent with the primary analysis.⁹

Objective Response Rate

By the August 28, 2020, data cut-off, the ORR estimated by IIR in the LEN-PEM arm was 71.0% (95% CI, 66.3 to 75.7). In total, 16.1% of patients receiving LEN-PEM had a confirmed CR and 54.9% had a confirmed PR. In the SUN arm, the ORR estimated was 36.1% (95% CI, 31.2 to 41.1). In total, 4.2% of patients receiving SUN had a confirmed CR and 31.9% of patients had a confirmed PR. The difference between the 2 treatment arms was 34.9% (95% CI, 28.0 to 41.7). The estimated OR between the LEN-PEM arm and the SUN arm was 4.35 (95% CI, 3.16 to 5.97; P < 0.0001) in favour of LEN-PEM.⁹ Table 17 presents a summary of the ORR estimated by IIR in the LEN-PEM arm versus the SUN arm.

Table 17: Objective Response Rate at Interim Analysis 3 — Full Analysis Set

CI = confidence interval, CR = complete response; IxRS = interactive voice and web response system; LEN = lenvatinib; PEM = pembrolizumab; PD = progressive disease; PR = partial response; SD = standard deviation; SUN = sunitinib.

Note: Percentages are based on the total number of patients in the full analysis set within the relevant treatment group. Stable disease must be ≥ 7 weeks after randomization. Durable stable disease must be ≥ 23 weeks after randomization. Time to first objective response (months) = (date of first objective response minus date of randomization plus 1) multiplied by 12 divided by 365.25, for patients with best overall response of CR or PR. It is censored for patients without a best overall response of CR or PR. Data cut-off date: August 28, 2020.

^a95% CI is constructed using the method of normal approximation.

^bOdds ratio and P value are calculated using the Cochran-Mantel-Haenszel method, stratified by IxRS stratification factors.

Source: Clinical Study Report.⁹

Overall Survival

By the August 28, 2020, data cut-off (which also corresponds to the second interim analysis for OS), the median OS was not estimable. An HR of 0.66 (95% CI, 0.49 to 0.88; P = 0.0049) was estimated based on IIR interpretation, representing a 34% reduction in the risk of death in the LEN-PEM arm compared with the SUN arm at any particular time point. The median duration of OS follow-up was 26.7 months (95% CI, 25.9 to 27.4) for LEN-PEM and 26.3 months (95% CI, 25.4 to 27.2) for SUN.⁹ Table 18 presents a summary of OS findings by IIR estimated for LEN-PEM and SUN. Figure 5 presents the Kaplan-Meier plot of the OS at the data cut-off for interim analysis 3.⁹

The sponsor considered that the use of post-treatment anti-cancer medication by patients in the survival follow-up phase may have confounded OS estimates at the August 28, 2020, data cut-off. A post hoc sensitivity analysis was conducted to assess the impact on OS (Appendix 3).

Table 18: Overall Survival at Interim Analysis 3 — Full Analysis Set

CI = confidence interval; IxRS = interactive voice and web response system; LEN = lenvatinib; MSKCC = Memorial Sloan Kettering Cancer Center; NE = not estimable; PEM = pembrolizumab; SUN = sunitinib.

Note: Percentages are based on the total number of patients in the full analysis set within the relevant treatment group. Data cut-off date: August 28, 2020.

^aQuartiles are estimated by Kaplan-Meier method, and the 95% CIs are estimated with a generalized Brookmeyer and Crowley method.

^bHazard ratio is based on a Cox proportional hazard model including treatment group as a factor; the Efron method is used for ties.

^cStratified by geographic region (region 1: Western Europe and North America or region 2: rest of the world) and MSKCC prognostic groups (favourable, intermediate, and poor risk) in IxRS.

^dEstimates for survival follow-up time are calculated in the same way as the Kaplan-Meier estimate of overall survival but with the meaning of “censor” and “event” status indicator reversed.

^eOverall survival rate and 95% CIs are calculated using the Kaplan-Meier product-limit method and Greenwood formula.

Source: Clinical Study Report.⁹

Figure 5: Kaplan-Meier Plot of Overall Survival at Interim Analysis 3 — Full Analysis Set

+ = censored observations; CI = confidence interval; HR = hazard ratio; IxRS = interactive voice and web response system; LEN+PEMBRO = lenvatinib + pembrolizumab; mo = months; NE = not evaluable; NR = not reached; OS = overall survival; SUN = sunitinib; vs = versus.

Note: Median was estimated by Kaplan-Meier method, and the 95% CIs were estimated using a generalized Brookmeyer and Crowley method. HR was estimated from a Cox proportional hazard model including treatment group as a factor and stratified by IxRS stratification factors; the Efron method was used for ties. Data cut-off date: August 28, 2020.

Source: Clinical Study Report.⁹

The sponsor conducted a follow-up analysis for OS 7 months after the August 28, 2020, data cut-off (i.e., March 31, 2021) (Appendix 3). The findings were consistent with the results obtained at interim analysis 3.

Duration of Response

By the August 28, 2020, data cut-off, the median DOR estimated by IIR in patients with a response was 25.8 months (95% CI, 22.1 to 27.9) in the LEN-PEM arm and 14.6 months (95% CI, 9.4 to 16.7) in the SUN arm.

Table 19 presents a summary of the DOR by IIR in the LEN-PEM arm versus the SUN arm.⁹

Table 19: Duration of Response at Interim Analysis 3 — Full Analysis Set

CI = confidence interval; LEN = lenvatinib; PEM = pembrolizumab; SUN = sunitinib

Note: Duration of objective response (months) = (date of progressive disease, death, or censor date minus date of first objective response plus 1) multiplied by (12 divided by 365.25) for patients with an objective response. Data cut-off date: August 28, 2020.

^aQuartiles were estimated by Kaplan-Meier method; 95% confidence intervals were estimated using a generalized Brookmeyer and Crowley method.

Source: Clinical Study Report.⁹

Health-Related Quality of Life

The completion and compliance rates observed for patients were greater than 90% at baseline across groups. Rates of questionnaire completion declined below 50% at cycle 26 for LEN-PEM and at cycle 12 for SUN because of treatment discontinuation. The adherence rates were greater than 80% up until cycle 51 across the treatment arms, while the adherence rate at the off-treatment visit was greater than 78% across the study arms.⁹

Overall Least Squares Mean Difference

Patients enrolled in the LEN-PEM arm had better maintenance of HRQoL and less severe symptoms compared with those who received SUN. The overall mean difference estimated favoured LEN-PEM for the EORTC QLQ-C30 physical functioning scale as well as for symptoms of fatigue, dyspnea, and constipation. Figure 6 presents the overall least squares mean difference estimated for the LEN-PEM arm versus SUN.

Figure 6: Overall Least Squares Mean Difference — LEN-PEM Versus SUN

CI = confidence interval; EORTC QLQ-C30 = European Organisation for Research and Treatment of Cancer Quality of Life Questionnaire Core 30; FKSI-DRS = Functional Assessment of Cancer Therapy Kidney Symptom Index - Disease Related Symptoms; GHS/QoL = global health status/quality of life; LEN = lenvatinib; LS = least squares; PEMBRO = pembrolizumab; SUN = sunitinib.

Note: Data cut-off date is August 28, 2020.

Source: Sponsor’s submission report.⁹

Time to First Deterioration

At the August 28, 2020, data cut-off, the longitudinal analysis of change from baseline for the different QoL scales was estimated after a mean follow-up time of 46 weeks.

In the EORTC QLQ-C30, physical functioning, dyspnea, and appetite loss showed improvement in patients receiving LEN-PEM (Figure 7 and Figure 8). The TTD assessed showed improvement in patients receiving LEN-PEM for every scale except cognitive functioning and financial difficulties. Figure 8 summarizes TTD in the LEN-PEM arm versus the SUN arm.

EORTC QLQ-C30 Questionnaire

In physical functioning, the median TTD in the LEN-PEM arm was 15.29 (95% CI, 12.29 to 21.43), while in the SUN arm, median TTD in weeks was 12.71 (95% CI, 9.29 to 18.14; nominal log-rank difference P value = 0.03).

The median TTD obtained in the dyspnea subscale was 39.29 (95% CI, 24.43 to 51) in the LEN-PEM arm and 21.14 (95% CI, 15.43 to 32.71) in the SUN arm (nominal log-rank difference P value = 0.02).

In the appetite loss subscale, the median TTD in weeks in the LEN-PEM arm was 18.29 (95% CI, 15.14 to 21.71), while in the SUN arm, the median TTD in weeks was 9.14 (95% CI, 6.29 to 15.14). The nominal P value of the log-rank test was 0.03.

EQ-5D-3L VAS

The median TTD in weeks obtained in the Visual Analogue Scale was 9.43 (95% CI, 6.43 to 12.29) in the LEN-PEM arm and, in the SUN arm, the median TTD in weeks was 9.14 (95% CI, 6.29 to 12.0). Nominal P value of 0.04 was obtained in the log-rank difference.

Figure 7: Time to First Deterioration Forest Plot — LEN-PEM Versus SUN

CI = confidence interval; EORTC QLQ-C30 = European Organisation for Research and Treatment of Cancer Quality of Life Questionnaire Core 30; FKSI-DRS = Functional Assessment of Cancer Therapy Kidney Symptom Index – Disease Related Symptoms; GHS/QoL = global health status/quality of life; L + P/S = lenvatinib + pembrolizumab versus sunitinib; LEN = lenvatinib; PEMBRO = pembrolizumab; pt = patient; SUN = sunitinib.

Note: Data cut-off date is August 28, 2020.

Source: Sponsor’s submission report.⁹

Figure 8: Time to First Deterioration Survival Analyses — LEN-PEM Versus SUN

CI = confidence interval; EORTC QLQ-C30 = European Organisation for Research and Treatment of Cancer Quality of Life Questionnaire Core 30; FKSI-DRS = Functional Assessment of Cancer Therapy Kidney Symptom Index – Disease Related Symptoms; LEN = lenvatinib; PEMBRO = pembrolizumab; SUN = sunitinib; TTD = time to first deterioration.

Note: Data cut-off date was August 28, 2020.

Source: Sponsor’s submission report.⁹

Time Until Definitive Deterioration

Prolonged TUDD in physical functioning, role functioning, social functioning, fatigue, insomnia, dyspnea, nausea and vomiting, pain, appetite loss, and diarrhea was observed in patients receiving LEN-PEM compared with those receiving SUN (Figure 9 and Figure 10).⁹

FKSI-DRS Total Score

In the LEN-PEM arm, the median TUDD in weeks was 134.14 (95% CI, 120 to not estimable), while in the SUN arm, the TUDD in weeks was 117.43 (95% CI, 90.14 to 131.29). The nominal P value obtained was less than 0.01 (Figure 9 and Figure 10).

EORTC Quality of Life Questionnaire Core 30

The median TUDD in weeks in the global health status/QoL scale in the LEN-PEM arm was 114.29 (95% CI, 102.14 to 153.29) while, in the SUN arm, the median TUDD in weeks was 75.14 (95% CI, 57.29 to 105.14). The nominal P value obtained was less than 0.0001.

Figure 9: Time Until Definitive Deterioration Forest Plot — LEN-PEM Versus SUN

CI = confidence interval; EORTC = European Organisation for Research and Treatment of Cancer; FKSI-DRS = Functional Assessment of Cancer Therapy Kidney Symptom Index - Disease Related Symptoms; GHS/QoL = global health status/quality of life; L + P/S = lenvatinib + pembrolizumab versus sunitinib; LEN = lenvatinib; PEMBRO = pembrolizumab; pt = patient; SUN = sunitinib; TUDD = time until definitive deterioration.

Data cut-off date: August 28, 2020.

Source: Sponsor’s submission report.⁹

In the physical function domain of the EORTC, the median TUDD in weeks in the LEN-PEM arm was 134.14 (95% CI, 109.14 to not estimable), while in the SUN arm, the median TUDD in weeks was 78.14 (95% 63.14 to 111.0). The nominal P value obtained from the log-rank difference was less than 0.0001.

EQ-5D-3L Visual Analogue Scale

The median TUDD in weeks obtained in the LEN-PEM arm was 124.86 (95% CI, 94.71 to 134.57) while, in the SUN arm, the median TUDD in weeks was 74.86 (95% CI, 54.14 to 94.0). The nominal P value obtained was less than 0.01 (Figure 9 and Figure 10).

Figure 10: Time Until Definitive Deterioration — LEN-PEM Versus SUN

CI = confidence interval; EORTC QLQ-C30 = European Organisation for Research and Treatment of Cancer Quality of Life Questionnaire Core 30; FKSI-DRS = Functional Assessment of Cancer Therapy Kidney Symptom Index – Disease Related Symptoms; GHS/QoL = global health status/quality of life; HRQoL = health-related quality of life; LEN = lenvatinib; PEMBRO = pembrolizumab; NE = not estimable; SUN = sunitinib; TUDD = time until definitive deterioration; vs = versus.

Data cut-off date: August 28, 2020.

Source: Sponsor’s submission report.⁹

Disease Control Rate

By the August 28, 2020, data cut-off, the DCR observed in the LEN-PEM arm was 90.1%, while in the SUN arm, the DCR was 74.2%. The clinical benefit rate was 84.2% in the LEN-PEM arm and 59.4% in the SUN arm.⁹ Table 20 summarizes findings of the DCR estimated by IIR in the LEN-PEM arm versus the SUN arm at interim analysis 3.

Table 20: Disease Control Rate at Interim Analysis 3 — Full Analysis Set

CI = confidence interval; CR = complete response; IxRS = interactive voice and web response system; LEN = lenvatinib; PEM = pembrolizumab; PR = partial response; SUN = sunitinib.

Note: Percentages are based on the total number of patients in the full analysis set within the relevant treatment group. Data cut-off date: August 28, 2020.

^a95% CI is constructed using the method of normal approximation.

^bOdds ratio and nominal P value are calculated using the Cochran-Mantel-Haenszel method, stratified by IxRS stratification factors.

Source: Clinical Study Report.⁹

Harms

Only those harms identified in the CADTH review protocol are reported subsequently. The safety evaluation focuses on the data from the LEN-PEM and SUN arms obtained at the August 28, 2020, data cut-off.⁹

Adverse Events

Overall, 99.7% of patients enrolled in the LEN-PEM arm and 98.5% in the SUN arm reported at least 1 TEAE by the August 28, 2020, data cut-off. AEs of grade 3 or higher occurred in 82.4% of patients in the LEN-PEM arm and 71.8% in the SUN arm. Table 21 presents an overview of TEAEs in the LEN-PEM arm and the SUN arm.⁹

By the August 28, 2020, data cut-off, the most frequently reported AEs occurring in more than 30% of patients receiving treatment in the LEN-PEM arm were diarrhea, hypertension, hypothyroidism, decreased appetite, fatigue, nausea, and stomatitis. In the SUN arm, diarrhea, hypertension, stomatitis, PPE syndrome, fatigue, nausea, and decreased appetite were the most frequently reported events. Table 22 summarizes the AEs occurring in at least 10% of patients receiving LEN-PEM or SUN in the CLEAR trial.⁹

Table 21: Summary of AEs at Interim Analysis 3 — Safety Analysis Set

AE = adverse event; CTCAE = Common Terminology Criteria for Adverse Events; LEN = lenvatinib; MedDRA = Medical Dictionary for Regulatory Activities; NA = not applicable; PEM = pembrolizumab; SUN = sunitinib; TEAE = treatment-emergent adverse event.

Note: Percentages are based on the total number of patients in the safety analysis set within the relevant treatment group. MedDRA preferred terms “neoplasm progression,” “malignant neoplasm progression,” and “disease progression,” which are unrelated to the study drug are excluded. For each row category, patients with 2 or more AEs in that category were counted only once. For serious TEAEs, the follow-up window is 120 days after the last dose date. AEs were graded using CTCAE version 4.03. Data cut-off date: August 28, 2020.

^aEach patient can be counted in multiple categories.

^bLEN or PEM (or SUN). Dose reduction is not applicable for PEM.

^cRegardless of action taken for PEM.

^dRegardless of action taken for lenvatinib.

^eDue to the same AE.

^fDose modification includes dose reduction or drug interruption.

Source: Clinical Study Report.⁹

Table 22: Adverse Events Occurring in at Least 10% of Patients on LEN-PEM and SUN in the CLEAR Trial at Interim Analysis 3 — Safety Analysis Set

AE = adverse event; LEN = lenvatinib; MedDRA = Medical Dictionary for Regulatory Activities; PEM = pembrolizumab; SUN = sunitinib; TEAE = treatment-emergent adverse event.

Note: Percentages are based on the total number of patients in the safety analysis set within the relevant treatment group. AEs categorized under the MedDRA preferred terms “neoplasm progression,” “malignant neoplasm progression,” and “disease progression” that were unrelated to the study drug were excluded. Patients with 2 or more TEAEs reported under the same preferred term were counted only once. MedDRA-categorized AEs were included if the incidence rate was 5% or higher in any treatment group. For the LEN-PEM group, AEs categorized by preferred term were sorted in decreasing order of incidence. If the incidence rates of AEs in 2 or more categories of terms were identical, the preferred terms were sorted alphabetically. AE terms were coded using MedDRA version 23.0. Data cut-off: August 28, 2020.

Source: Clinical Study Report.⁹

Serious Adverse Events

The SAEs of grade 3 and higher reported in at least 1% of patients at the August 28, 2020, cut-off are presented in Table 23. SAEs of grade 3 and higher were more frequently reported in the LEN-PEM arm compared with the SUN arm (82.4% of patients receiving LEN-PEM versus 71.8% receiving SUN) by the August 28, 2020, data cut-off. The most reported SAEs (occurring in ≥ 2% of patients in either arm) were diarrhea, vomiting, pneumonitis, acute kidney injury, hypertension, pneumonia, dyspnea, adrenal insufficiency, and pyrexia.⁹ Table 24 summarizes the SAEs reported in patients receiving either study treatment.

Table 23: Grade 3 or Higher TEAEs Occurring in at Least 1% of Patients in Any Treatment Arm — Safety Analysis Set

LEN = lenvatinib; MedDRA = Medical Dictionary for Regulatory Activities; PEM = pembrolizumab; SUN = sunitinib; TEAE = treatment-emergent adverse event.

Note: Percentages are based on the total number of patients in the safety analysis set within the relevant treatment group. TEAEs categorized under the MedDRA preferred terms “neoplasm progression,” “malignant neoplasm progression,” and “disease progression” that were unrelated to the study drug were excluded. The TEAEs in the lenvatinib and everolimus group are listed in the table by decreasing order of frequency. Patients with 2 or more TEAEs reported under the same preferred term were counted only once. Adverse event terms were coded using MedDRA version 23.0 and graded using Common Terminology Criteria for Adverse Events version 4.03. Data cut-off date: August 28, 2020.

Source: Clinical Study Report.⁹

Table 24: Serious Adverse Events of Any Grade Occurring in Patients in Any Treatment Arm — Safety Analysis Set

LEN = lenvatinib; MedDRA = Medical Dictionary for Regulatory Activities; PEM = pembrolizumab; SUN = sunitinib; TEAE = treatment-emergent adverse event.

Note: Percentages are based on the total number of patients in the safety analysis set within the relevant treatment group. AEs categorized under the MedDRA preferred terms “neoplasm progression,” “malignant neoplasm progression,” and “disease progression” that are unrelated to the study drug were excluded. Treatment-related TEAEs include those considered by the investigator to be related to the study drug and those with a missing causality. The treatment-related TEAEs in the LEN plus everolimus group are listed in the table by decreasing order of frequency. Patients with 2 or more TEAEs reported under the same preferred term were counted only once. Adverse events were coded using MedDRA version 23.0. Data cut-off: August 28, 2020.

Source: Clinical Study Report.⁹

Mortality

By the August 28, 2020, data cut-off, all death-related events, other than “malignant neoplasm progression,” were included in the frequency counts for fatal AEs.

In total, 78 deaths (22.2%) were reported in the LEN-PEM arm and 99 deaths (29.1%) were reported in the SUN arm. Deaths that occurred during the survival follow-up phase were not reported as AEs. During that phase, there were 51 deaths (14.5%) reported in the LEN-PEM arm and 76 deaths (22.2%) in the SUN arm; in both study arms, most deaths were attributed to disease progression.⁹ Table 25 summarizes the number of deaths in the LEN-PEM arm versus the SUN arm.

Table 25: Summary of Deaths at Interim Analysis 3 — Safety Analysis Set

LEN = lenvatinib; PD = progressive disease; PEM = pembrolizumab; SUN = sunitinib; TEAE = treatment-emergent adverse event.

Note: Percentages are based on the total number of patients in the safety analysis set within the relevant treatment group. Data cut-off date: August 28, 2020.

^aIncludes TEAEs categorized under Medical Dictionary for Regulatory Activities preferred terms “neoplasm progression,” “malignant neoplasm progression,” and “disease progression” that were not related to the study drug.

^bIncludes 4 patients (3 on LEN plus everolimus and 1 on SUN) who took new anti-cancer therapy and died more than 30 days after but within 120 days after the last dose of the study drug.

^cDeaths occurred during the survival follow-up period.

Source: Clinical Study Report.⁹

Withdrawals, Dose Reduction, or Interruptions Due to AEs

Discontinuation Due to AEs

Overall, 13.4% of patients receiving LEN-PEM and 14.4% receiving SUN discontinued all study drugs due to TEAEs. In total, 37.2% of patients in the LEN-PEM arm discontinued a study drug due to AEs, most of which included pneumonitis, diarrhea, rash, acute myocardial infarction, proteinuria, myocardial infarction, acute kidney injury, and renal failure, which occurred in at least 1% of patients in either arm.⁹

In total, 25.6% of patients receiving LEN-PEM discontinued LEN due to AEs. AEs commonly associated with LEN discontinuation (in > 1% of patients) were proteinuria (1.7%), diarrhea (1.4%), myocardial infarction (1.1%), and acute myocardial infarction (1.1%). A total of 28.7% of patients discontinued PEM treatment in the LEN-PEM arm. The AEs commonly associated with discontinuation of PEM (in > 1% of patients) were pneumonitis (2.8%), rash (1.7%), diarrhea (1.1%), and alanine aminotransferase increased (1.1%).⁹

In total, 14.4% of patients discontinued SUN. The most common AEs associated with discontinuation of SUN were nausea (0.9%), asthenia (0.9%), fatigue (0.9%), acute kidney injury (0.9%), and metastases to CNS (0.9%).⁹

Dose Reduction or Interruptions Due to AEs

More patients (68.8%) receiving LEN-PEM experienced AEs leading to dose reduction compared with those receiving SUN (50.3%). The most common AEs associated with the reduction of LEN were diarrhea (15.9%), hypertension (11.6%), proteinuria (10.2%), PPE (8.8%), decreased appetite (7.7%), and nausea (5.1%).

In total, 73% of patients enrolled in the LEN-PEM arm experienced AEs leading to interruption of LEN compared with the patients receiving SUN (53.8%). The most common AEs leading to interruption of LEN were diarrhea (17.6%), hypertension (8.2%), proteinuria (7.7%), asthenia (6.3%), lipase increased (5.4%), and fatigue (5.1%). AEs leading to dose interruption of PEM were reported in 55.1% of patients in the LEN-PEM arm. The AEs that commonly led to dose interruption of PEM were diarrhea (10.2%), lipase increased (5.1%), asthenia (4.5), alanine aminotransferase increased (3.4%), amylase increased (4%), aspartate aminotransferase increased (2.6%), and fatigue (2.6%).⁹ The most common AEs leading to dose modifications in patients receiving SUN were PPE (12.6%), hypertension (9.7%), platelet count decreased (10%), diarrhea (8.2%), fatigue (8.2%), thrombocytopenia (7.1%), stomatitis (6.2%), and neutropenia (6.2%). Dose reduction of SUN was commonly associated with platelet decrease (8.8%), diarrhea (5%), hypertension (4.7%), asthenia (3.8%), and fatigue (6.8%).

Notable Harms

The notable harms identified in both the LEN-PEM arm and the SUN arm in the CLEAR trial included hypertension, hypothyroidism, hepatotoxicity, proteinuria, hemorrhage, renal events, PPE events, QT prolongation, gastrointestinal perforation, arterial thromboembolic events, hypocalcemia, cardiac dysfunction, arterial thromboembolic events, posterior reversible encephalopathy syndrome, and fistula formation. Table 26 presents a summary of the proportion of patients reporting these AEs in the LEN-PEM arm versus the SUN arm. A higher proportion of reports was observed in the LEN-PEM arm for hypertension, hypothyroidism, hepatotoxicity, proteinuria, and QT prolongation compared with the SUN arm.⁹

Table 26: Notable Harms Common in the LEN-PEM and SUN Arm at Interim Analysis 3 — Safety Analysis Set

ALT = alanine aminotransferase; AST = aspartate aminotransferase; LEN = lenvatinib; NR = not reported; PEM = pembrolizumab; QT = heart rate QT interval; SUN = sunitinib.

Note: Data cut-off date was August 28, 2020.

Source: Clinical Study Report.⁹

Critical Appraisal

Internal Validity

The CLEAR trial is a randomized, parallel-arm design. Treatment allocation was performed centrally using an IxRS, allowing patients to be randomly assigned to any 1 of the predefined study arms in a 1:1:1 ratio. The study was stratified based on 2 factors: geographic region and the MSKCC prognostic risk groups. The baseline and demographic characteristics were balanced within the 2 study arms of interest, other than a slight imbalance in age between the LEN-PEM and SUN arms (i.e., there were more patients younger than 65 years old in the SUN arm and more patients older than 65 years in the LEN-PEM arm), suggesting that randomization was implemented successfully.

The CLEAR trial is an open-label trial that is open to patients and investigators. The pre-specified dosage of lenvatinib is different for each study arm, and the treatment administration schedules between everolimus, PEM, and SUN are also different. The open-label design introduces a potential bias in the assessment of PFS, ORR, DOR, and DCR, and a potential reporting bias of subjective outcomes such as HRQoL and safety. To minimize the risk of differential measurement error, the sponsor performed tumour assessments using RECIST 1.1 criteria and radiographic scans were assessed by an IIR team. The RECIST guideline was considered appropriate by the CADTH reviewers, given that it has been used in many oncology trials and in real-world settings by regulatory agencies, including Health Canada, to assess tumour changes.²⁹ Tumour scans and assessments were also performed by an approved independent core laboratory.

There was a low risk of bias owing to the use of non-protocol interventions. Patients were permitted to receive anti-cancer therapy only in the follow-up phase (after study treatments had been discontinued). More patients in the SUN arm (57.7%) received post-treatment anti-cancer therapies compared with the LEN-PEM arm (33%), which may have introduced bias in the assessment of OS. A post hoc analysis of survival conducted by the sponsor at the August 28, 2020, data cut-off suggested that anti-cancer medications received post-treatment might have biased the OS estimates in the opposite direction (did not favour the LEN-PEM treatment). CADTH considered the post hoc analysis appropriate to assess the impact of subsequent anti-cancer therapy on OS.

CADTH found there was a low risk of selective reporting bias in the CLEAR study, as the sponsor analyzed data in accordance with the pre-specified statistical plan, which was finalized before the unblinded data were available for the analyses. The sponsor reported findings for all pre-specified analyses outlined in the study protocol. All safety data occurring in the treatment arms were also reported.

The PFS and OS investigated during the CLEAR trial were considered appropriate outcomes for patients with advanced and metastatic RCC, according to the clinical experts consulted. The censoring rules for PFS were pre-specified based on guidance documents,^21,22 and sample size and power calculations were based on PFS. The Cox proportional hazard model, which relies on the assumption of proportional hazards in both treatment groups, was used. A violation of the proportional hazard assumption may lead to bias in estimates in the regression model. It was not clearly stated in the sponsor’s statistical analysis plan whether the proportional hazard assumption was assessed formally and whether it was violated or not. It is also unclear whether adjustments were made to meet the proportional hazard assumption in the study. The visual assessment of the PFS survival curves did not suggest any violation of the proportional hazard assumption. Survival curves were estimated using the Kaplan-Meier model. The median OS was not estimated in either treatment arm at the 2 data cut-offs (August 28, 2020, and March 31, 2021). CADTH considered the benefit of LEN-PEM in improving OS in patients with advanced RCC to be uncertain, owing to data immaturity in the trial.

HRQoL was assessed as a secondary outcome in the CLEAR trial using the FKSI-DRS, EORTC QLQ-C30, and EQ-5D-3L questionnaires. There is a potential for reporting bias owing to the open-label nature of the trial. The FKSI-DRS questionnaire has been validated in patients with RCC with evidence of reliability and responsiveness. Although the EORTC QLQ-C30 and EQ-5D-3L have been widely used in oncology trials in different cancer populations, these questionnaires have not been validated in patients with advanced or metastatic RCC. The clinical experts consulted highlighted that all 3 scales were appropriate for assessing patient-reported outcomes in RCC patients. In the opinion of the clinical experts, these scales are most sensitive to changes in symptoms related to RCC but are not sensitive to detecting changes in symptoms due to TRAEs. HRQoL outcomes had missing data at later time points, as evidenced by the declined rates of questionnaire completion: below 50% at cycle 26 for LEN-PEM and at cycle 12 for the SUN arm owing to treatment discontinuation, thus impacting the interpretability of trends over time and raising the potential for biased results from patients who remained in the trial. The sponsor did not provide details on how missing data were handled during the analyses. It is unclear whether the methods used for handling missing data introduced bias in the findings presented. The benefit of LEN-PEM in improving HRQoL over time was considered uncertain by the CADTH reviewers, owing to the potential of reporting and attrition bias.

The sponsor assumed that a median PFS of 12.3 months in the SUN arm and an HR of 0.714 in the comparisons between the study arms (LEN-PEM against SUN and LEN plus everolimus against SUN) and SUN in their hypothesis tests. According to the sponsor, this corresponds to a 40% improvement (4.9 months) in median PFS (from 12.3 months to 17.2 months) in both study arms (LEN plus everolimus and LEN-PEM) against the SUN arm. The sponsor also provided threshold margins for the ORR and OS outcomes. The sponsor did not provide a rationale for using the margins for PFS, OS, and ORR in the statistical plan. The clinical experts consulted during the review considered the threshold margins for PFS, ORR, and OS to be clinically meaningful.

The PFS, OS, and ORR estimates obtained at the August 28, 2020, cut-off (including the OS follow-up analysis of March 31, 2021) were considered robust because multiplicity adjustments were performed during analysis. The type I error rate was adjusted using the overall familywise error rate approach, which was considered appropriate by the CADTH reviewers. Sensitivity analyses and adjustments of covariates were conducted for PFS, and the findings were consistent with the primary analysis in the ITT set. Multiplicity adjustments were not conducted for other secondary and exploratory end points (DOR, DCR, and HRQoL), including the analysis of subgroups.

All subgroup analyses were pre-specified in the protocol. No statistical tests of interaction were conducted for subgroups to test whether treatment effects differed among subgroups and other exploratory outcomes. The sample size for the subgroup analyses (IMDC prognostic risk groups) was small and possibly underpowered to detect a statistically meaningful difference between the study arms. Multiplicity adjustments were not made in these analyses; therefore, the P values reported were considered nominal. The magnitude of efficacy for subgroups was considered uncertain.

All interim analyses conducted were planned a priori, and stopping rules were pre-specified in the protocol. Alpha levels were properly accounted for in each of the analyses conducted. The approaches used to preserve alpha and the power in the interim analyses were considered appropriate by the CADTH reviewers.

More patients discontinued study treatment in the SUN arm (76.5%) compared with the LEN-PEM arm (59.5%). Patients were not allowed to cross over from 1 arm to the other in the CLEAR trial, which reduced bias. Treatment discontinuations had minimal impact on the analysis of PFS and OS because missing values and dropouts were handled using conservative approaches (patients who discontinued study treatments were censored for the PFS analysis but remained in the follow-up phase and were assessed for OS unless they withdrew consent). The ITT and per-protocol sets were used to assess the primary and secondary outcomes to account for missing data and unbalanced treatment discontinuations between study arms. Consistency in the results between both analysis sets was assessed by the sponsor and no major differences were reported between both sets. The CADTH reviewers considered that appropriate methods were used to control attrition for the analysis of the primary and key secondary outcomes.

Reduced drug exposure due to nonadherence and reduced dose intensity could potentially underestimate AEs and efficacy, leading to an overestimation of dose intensity.³⁰ CADTH considered that there was a low risk of bias owing to nonadherence to protocol interventions in the CLEAR trial because few patients were reported as having major protocol deviations and these numbers were equally distributed across the 2 study arms of interest. All protocol deviations led to a censoring event for PFS by IIR. According to the sponsor’s dose intensity calculations presented, adherence bias was not considered a potential issue by the CADTH reviewers, given that all dose adjustments for study drugs were pre-specified in the protocol and recorded for each study drug. CADTH noted that patients in the LEN-PEM arm had longer exposure to treatment (mean = 17.29 months) compared with patients receiving SUN (mean = 11.33 months), which may have influenced the reporting of AEs in both study arms.

External Validity

Table 27 summarizes the generalizability of the evidence from the CLEAR trial. The population requested for the reimbursement aligns with the Health Canada indication. The clinical experts consulted during the CADTH review agreed that the inclusion and exclusion criteria of the CLEAR trial were appropriate for the RCC population under consideration. The experts also noted that the baseline characteristics of patients in the CLEAR trial are generalizable to patients in the Canadian setting.

LEN 20 mg, taken orally once daily, was administered with PEM 200 mg IV every 3 weeks following a 21-day cycle. The dosage aligns with the Health Canada indication. In the CLEAR trial, patients were also allowed to discontinue either 1 of the study drugs in the treatment phase. Dose adjustments or modifications were allowed in the trial and the methods were outlined in the protocol. According to the clinical experts consulted, dose adjustments or modifications are anticipated in a clinical practice setting to manage AEs while maintaining drug benefit. The clinical experts also noted that the discontinuation of 1 treatment while the patient continues another is frequently observed in practice for drug combinations.

SUN is an approved treatment option for untreated patients with metastatic RCC in Canada. It was considered an appropriate comparator because it was available as a standard-of-care option for RCC patients in the first-line setting when the CLEAR trial was initiated. As noted by the clinical experts, AXI-PEM and other standard-of-care treatments were not available when the CLEAR trial was initiated.

Table 27: Assessment of Generalizability of Evidence for LEN-PEM

AE = adverse event; AXI = axitinib; CNS = central nervous system; DCR = disease control rate; DOR = duration of response; HRQoL = health-related quality of life; ITT = intention to treat; KPS = Karnofsky Performance Status; LEN = lenvatinib; ORR = objective response rate; OS = overall survival; PEM = pembrolizumab; PFS = progression-free survival; PFS2 = PFS with next line of therapy; SUN = sunitinib.

Indirect Evidence

Objectives and Methods for the Summary of Indirect Evidence

The aim of this section was to appraise the indirect evidence used to inform the pharmacoeconomic model and to identify indirect comparisons that fill gaps in the evidence from the systematic review. Although direct evidence is available on the efficacy and safety of LEN-PEM versus SUN, comparative efficacy studies versus other first-line treatments for advanced or metastatic RCC of interest were not identified in the systematic literature search.

A focused literature search for NMAs dealing with at least 1 of the following search terms, “renal carcinoma” or “lenvatinib” in combination with “pembrolizumab,” was run in MEDLINE All (1946–) on December 9, 2021. No limits were applied.

Four potentially relevant ITCs were identified in the literature in addition to the sponsor-submitted NMA. This section will appraise the sponsor-submitted NMA, and the following section will appraise published ITCs.

Description of Sponsor-Submitted NMA

The sponsor submitted an ITC that evaluated the efficacy and safety of LEN-PEM in patients with advanced RCC.

Objectives

The objective of the sponsor-submitted ITC was to assess the comparative clinical efficacy and safety of LEN-PEM compared with other first-line treatments currently approved, recommended, or under development in advanced RCC, based on evidence from RCTs.

Study Selection Methods

The selection of studies used to inform the NMA was based on a systematic literature review, described in Table 28. The review included phase II and III RCTs in adults with advanced or metastatic RCC without prior lines of systemic therapy who received first-line systemic treatments for advanced or metastatic RCC administered alone or in combination (Table 28), best supportive care, or placebo, and reported on 1 of the efficacy outcomes (PFS, OS, ORR, PD, DOR, time to next treatment) or safety outcomes of interest (patients with discontinuations, treatment discontinuation due to AEs, total all-cause grade 3 or higher AEs, total grade 3 or higher TRAEs, duration on intervention, and subsequent treatments). Studies were limited to English-language reports and excluded trials based on the exclusion criteria described in Table 28. The literature search included multiple electronic databases (up to June 2021) and other sources such as grey literature, bibliography review, and client results. Reports were screened independently by 2 researchers, with disagreements resolved by a third reviewer. Data were extracted by 1 researcher and verified by another. The quality assessment of all RCTs was conducted using the Cochrane risk-of-bias assessment tool (version 2.0).

Table 28: Study Selection Criteria and Methods for Sponsor-Submitted NMA

AE = adverse event; BSC = best supportive care; CENTRAL = Cochrane Central Register of Controlled Trials; DOR = duration of response; NMA = network meta-analysis; ORR = overall response rate; OS = overall survival; PD = progressive disease; PFS = progression-free survival; RCC = renal cell carcinoma; RCT = randomized controlled trial; SLR = systematic literature review; TRAE = treatment-related adverse event.

^aDosing was not specified in the study selection protocol.

^bInterventions include: Axitinib, avelumab plus axitinib, atezolizumab plus bevacizumab, bevacizumab, bevacizumab plus erlotinib, bevacizumab plus everolimus, bevacizumab plus interferon, bevacizumab plus low-dose interferon, cabozantinib, erlotinib, everolimus, high-dose interleukin-2, lenvatinib plus everolimus, nivolumab, nivolumab plus cabozantinib, nivolumab plus ipilimumab, pazopanib, pembrolizumab plus axitinib, sorafenib, sunitinib, temsirolimus, and tivozanib.

^cGrey literature, bibliography review, client results.

Source: Sponsor-submitted indirect treatment comparison (Evidera).³¹

Methods of Sponsor-Submitted NMA

ITC Analysis Methods

A feasibility assessment was conducted to determine whether a connected network comparing the treatments of interest using the outcomes of interest could be constructed, and if any differences in the characteristics of the included RCTs, including common comparators and network connections, patient characteristics (histology, cancer stage, risk score, other patient characteristics), or trial outcomes (outcome assessment, survival analysis, subgroup definitions, and follow-up duration) may be modifiers.

A Bayesian NMA using an FE or RE model was conducted due to the presence of multiple studies per comparison to account for potential substantial heterogeneity of the network. Model fit was assessed by comparing the DIC and the posterior mean of the residual deviance for the FE and RE models. A substantially better-fitting model was identified by DICs that are lower by more than 5 points. An FE model was run for every scenario and chosen as the basis for the interpretation of results when only single-study comparisons existed within the network; RE models were considered based on the goodness-of-fit assessment (DIC greater than 3 to 5 points), heterogeneity (I² > 50% for any direct comparison), or statistically significant inconsistency between results from direct evidence versus indirect evidence within closed loops in a network. Of note, the author reported that the RE model was considered the default in networks with at least 2 instances of comparisons of at least 2 studies, given that the FE model assumption of homogeneity is unrealistic.

All Bayesian NMAs used a non-informative prior (N) of N(0 to 10,000) for the baseline effects and treatment effects. The RE NMAs used somewhat informative priors with a uniform distribution (U) for the standard deviations, specifically, U(0 to 0.4) for PFS and OS, and U(0 to 1) for binary outcomes. These were chosen to allow for a moderate to large variation, but to discourage the estimation of extreme values. Sensitivity analyses explored the influence of these priors. All Bayesian analyses used Markov chain Monte Carlo simulations with a 100,000 iteration burn-in and 3 sets of 100,000 iterations (posterior samples) for parameter estimation. Model convergence was assessed through inspection of ratios of Monte Carlo error to standard deviations of the posteriors. Where necessary, convergence was confirmed using plots of 3 chains. Issues related to convergence were handled by increasing the run-in and/or examination of other factors, such as choice of prior and starting values.

Inconsistency was assessed by comparing the results of the Bucher ITCs with that of direct evidence. Of note, comparators of interest to this review were not involved in a closed loop within the network. Heterogeneity for each comparison was assessed by conducting classical pairwise meta-analyses and calculating I². Where statistical heterogeneity was identified, study and patient characteristics were further reviewed to determine whether there were any clinically relevant effect modifiers. Potential effect modifiers were reviewed and handled by either study exclusion, sensitivity analysis, or subgroup analysis. Sensitivity analyses were planned that assessed the difference across potential effect modifiers, such as the differences between the enrolment of patient populations with 100% clear cell histology and differences in follow-up durations (based on a cut-off of less than 12 months or more than 12 months). Additionally, sensitivity analyses that considered the equivalence of SUN, PAZO, and tivozanib (treated as 1 node) were also performed. The planned subgroup analyses were to assess outcomes by risk subgroup. More specifically, sensitivity analysis 3 (risk subgroups: MSKCC = IMDC) and sensitivity analysis 11 (risk subgroups: IMDC = MSKCC) evaluated the assumption that the IMDC and MSKCC risk score definitions were equivalent. If both the IMDC and MSKCC definitions were available for a single trial, MSKCC was prioritized in sensitivity analysis 3 and IMDC was prioritized in sensitivity analysis 11. Two subgroup analyses, where only trials reporting IMDC definitions were included (sensitivity analysis 4) or only trials reporting MSKCC definitions were included (sensitivity analysis 5), were also performed.

For base-case analyses, separate nodes were used for interventions and comparators within the networks and only treatment and associated standard dosing were inputted into the model. Differences in time point assessments and actual treatment duration were acknowledged as a limitation; no studies were excluded, and sensitivity analyses were conducted that excluded 2 studies with a follow-up duration of less than 12 months. The excluded studies did not involve comparators of interest for this review.

Regarding the handling of missing data, 1 study digitized a Kaplan-Meier curve for PFS to estimate the HR (and standard errors of the log HR), which was not provided in the publication. No other studies were missing data related to OS or PFS. The ITT denominator was used for response outcomes and no calculations were applied to exclude unknown or unevaluable responses if included in the ITT denominator.

Table 29: Sponsor-Submitted NMA Analysis Methods

AE = adverse event; DIC = deviance information criterion; EMA = European Medicines Agency; FE = fixed effects; IFN = interferon; IMDC = International Metastatic Renal Cell Carcinoma Database Consortium; IRC = independent review committee; ITC = indirect treatment comparison; MSKCC = Memorial Sloan Kettering Cancer Center; NMA = network meta-analysis; OS = overall survival; PFS = progression-free survival; PD-L1 = programmed cell death 1 ligand 1; RE = random effects; RECIST 1.1 = Response Evaluation Criteria in Solid Tumours version 1.1; SD = standard deviation.

^aPFS, according to FDA censoring criteria, included progression date assigned to the earliest date when any RECIST 1.1–defined disease progression is observed without missing more than 1 adequate radiologic assessment (FDA, 2018).

^bPFS, according to EMA censoring criteria, included using the actual date of progression reported by an independent imaging review or death to define PFS, regardless of missing assessments, or use of a new anti-cancer therapy (EMA, 2019).

Source: Sponsor-submitted ITC (Evidera).³¹

Results of Sponsor-Submitted NMA

Summary of Included Studies

Twenty-four trials were included in the network informing the NMA. Most of the studies were phase II or phase III open-label RCTs (19 studies), 2 were phase III double-blind RCTs, and 3 were phase II or III studies that did not report blinding. The studies enrolled patients between 1992 and 2019 and the study sample sizes ranged from 101 to 1,110 patients. A total of 18 studies reported on the timing of response assessments, which varied across studies from every 6 weeks to every 12 weeks. One study (Negrier [1998]) was excluded from the base-case analysis due to differences in the response criteria. The sponsor indicated that PFS was consistently defined as the time to disease progression or death assessed by investigator or IRC review. The follow-up duration varied between trials, with multiple outcome assessment times according to different data cut-offs. The sponsor indicated that most studies reported outcomes at follow-up durations of 20 to 25 months, with the exception of a few outliers, and acknowledged this as a limitation.

A summary of the patient characteristics of included trials is presented in Appendix 4 (Table 39). Among the 24 trials, the median age of the study populations ranged from 55 years to 68 years, and more than half of patients were male (range, 56.5% to 83.5%). From the 8 studies that reported information about ethnicity, the majority of patients were White (69% to 97%), followed by Asian (1% to 25%) or Black (0% to 3%). Trials included in the network described patients by risk category using the MSKCC criteria (16 studies), IMDC criteria (5 studies), or both (2 studies). Where baseline risk was reported (in all but 1 study), 23.5% to 81% of patients in each treatment group were classified as intermediate risk. Three studies recruited only patients who were classified as intermediate or poor risk (Choueiri [2017]; Hudes [2007]; Zurita [2018]). Performance status was reported using the Karnofsky scale (11 studies) and the ECOG scale (12 studies) or not reported (1 study). In most of the studies included in the network (21 studies), the majority of patients had either a Karnofsky score of at least 70 or an ECOG score of 0 or 1 (less than 13% of patients included in 4 studies had an ECOG score of 2, and 80% to 83% of patients in 1 study had a Karnofsky score of 70 or less). In all studies that reported information regarding histology (21 studies), the most common histological RCC subtype was clear cell, with at least 78% of patients possessing clear cell or predominantly clear cell histology location. All included studies enrolled patients with metastatic or advanced metastatic cancer. Ten studies specified the American Joint Committee on Cancer (AJCC) cancer stage; at least 95% of the patients enrolled in 8 studies had stage IV cancer while, in the remaining 2 studies (including CLEAR), 40% to 55% had stage IV cancer. In 11 studies, between 56.5% and 89% of patients had at least 2 metastases; the lungs were the most common site of metastasis, representing more than 50% of the metastatic sites in the 17 studies that reported metastasis location. Lastly, 6 studies reported PD-L1 expression status at baseline and the proportion of patients with less than 1% PD-L1–positive cells ranged from 29% to 77%.

Table 30: Assessment of Homogeneity for Sponsor-Submitted NMA

IFN = interferon; IMDC = International Metastatic Renal Cell Carcinoma Database Consortium; MSKCC = Memorial Sloan Kettering Cancer Center; NMA = network meta-analysis; PD-L1 = programmed cell death 1 ligand 1; PFS = progression-free survival; q.6.w. = every 6 weeks; q.8.w. = every 8 weeks; q.12.w. = every 12 weeks; RECIST = Response Evaluation Criteria in Solid Tumours; RCC = renal cell carcinoma; RCT = randomized controlled trial.

Source: Sponsor-submitted indirect treatment comparison (Evidera).³¹

Results

A global network diagram for the NMA is presented in Figure 11. The network included 22 comparators from 24 RCTs (N = 12,577, based on the sum of the reported study sample sizes, excluding 1 study that did not report sample size.)

Figure 11: Network Diagram for Sponsor-Submitted NMA

HD IL-2 = high-dose interleukin-2; IFN = interferon.

Note: Because the CBOSUN, Global ARCC, and TemPa trials enrolled only patients with an intermediate or poor risk profile, those studies were not included in the base-case analyses but were included in the risk subgroup analyses. Only a treatment-naive subgroup of patients from the TIVO-1 trial was included.

Source: Sponsor-submitted indirect treatment comparison.³¹

Progression-Free Survival

PFS was evaluated using FDA censoring criteria and EMA censoring in the CLEAR trial and assessed as such in the sponsor-submitted NMA (Figure 12 and Figure 13). The CLEAR trial was the only included study that reported PFS assessments by FDA or EMA censoring criteria; therefore, only data from the CLEAR trial differs between the 2 analyses.

The base-case analysis of PFS (FDA censoring) included 18 comparators from 21 RCTs. This included comparisons of LEN-PEM, AXI-PEM, NIVO-IPI, and PAZO versus SUN, each informed by evidence from 1 trial (CLEAR, KEYNOTE-426, CheckMate-214, and COMPARZ, respectively). Of note, the data informing the PFS results from the KEYNOTE-426³² and CheckMate-214³³ trials are based on a 42-month follow-up and 4-year follow-up, respectively. The author reported that an RE model was selected because the DIC was lower (42.6 versus 42.9), although they noted the difference was within the 5 points and, therefore, was not considered meaningful. The results of the base-case analysis of PFS (FDA censoring) are presented in Figure 12. The reported HR for LEN-PEM was 0.44 (95% CrI, 0.23 to 0.82) compared with NIVO-IPI, 0.57 (95% CrI, 0.31 to 1.08) compared with AXI-PEM, and 0.38 (95% CrI, 0.21 to 0.67) compared with PAZO. The author indicated that the point estimates of the FE model were similar to the RE model, although the CrIs were narrower ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||.

Figure 12: PFS (FDA Censoring) Results — LEN-PEM Versus Other Treatments (Base Case, RE)

ATE = atezolizumab; AVE = avelumab; AXI = axitinib; BEV = bevacizumab; CAB = cabozantinib; CrI = credible interval; EVE = everolimus; IFN = interferon; IL = interleukin; IPI = ipilimumab; LEN = lenvatinib; NIV = nivolumab; PAZ = pazopanib; PEM = pembrolizumab; PFS = progression-free survival; prob = probability; RE = random effects; SOR = sorafenib; SUN = sunitinib; TIV = tivozanib.

Source: Sponsor-submitted indirect treatment comparison.³¹

The results of the base-case analysis of PFS (EMA censoring) were similar to the results of the base-case analysis using FDA censoring. An RE model was selected for the analysis of PFS (EMA censoring) based on the same rationale for PFS (FDA censoring). The base-case analysis included 19 comparators from 21 RCTs. |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||.

Analysis of PFS (FDA censoring) by risk status was considered feasible under the assumption that the MSKCC and IMDC risk status definitions were equivalent. This was analyzed using 2 sensitivity analyses: the first (sensitivity analysis 3) prioritized MSKCC assessments from trials that provided MSKCC and IMDC results, and the second (sensitivity analysis 11) prioritized IMDC assessments from trials that provided IMDC and MSKCC results. The results for relevant comparisons are presented by risk subgroup in Table 31. The analysis of PFS by risk subgroup was also performed based on the data using the EMA censoring rules. The results were similar to the results presented in Table 31 for PFS by risk subgroup (FDA censoring), with the exception of the comparison with AXI-PEM performed in the poor risk subgroup. |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||

Table 31: Redacted

Note: Table and title redacted as per sponsor’s request.

Overall Response Rate

The base-case analysis of ORR included 13 comparators from 14 RCTs, including comparisons of LEN-PEM, AXI-PEM, NIVO-IPI, and PAZO versus SUN (each informed by evidence from 1 trial). The author reported that an FE model was selected based on a DIC that was lower for the FE model (DIC = 55.5) than the RE model (DIC = 55.8). A summary of the results is presented in Figure 13. The OR for LEN-PEM was 3.24 (95% CrI, 2.18 to 4.85) compared with NIVO-IPI, 1.86 (95% CrI, 1.23 to 2.84) compared with AXI-PEM, and 3.00 (95% CrI, 2.02 to 4.47) compared with PAZO. The author reported that the CrIs were larger in the RE model, which only impacted the comparison with AXI-PEM |||||||||||||||||||||||||||||||||||||||||||||||||||.

Overall Survival

The base-case analysis of OS included 13 comparators from 12 RCTs, including comparisons of LEN-PEM, AXI-PEM, NIVO-IPI, and PAZO versus SUN (each informed by evidence from 1 trial). Of note, the data informing the OS results of the CLEAR trial are based on follow-up data from a conference abstract and the results for the KEYNOTE-426³² and CheckMate-214³³ trials are based on a 42-month follow-up and 4-year follow-up, respectively. The author reported that FE models were used for all analyses of OS due to sparse networks with only 1 trial per connection; RE models were not run for this outcome. The sponsor also reported that the OS network was sparser than the PFS network due to the exclusion of the 6 trials with planned or sequential crossover designs. These trials included patients who had crossed over to receive a second treatment as part of the planned study design; therefore, the OS results were ineligible for inclusion in the NMA, as they pertain to second-line treatment.³¹ The HR for comparisons of LEN-PEM with NIVO-IPI was 1.04 (95% CrI, 0.77 to 1.42); for AXI-PEM, the HR was 0.99 (95% CrI, 0.71 to 1.37), and for PAZO, the HR was 0.78 (95% CrI, 0.58 to 1.06).

Figure 13: ORR Results — LEN-PEM Versus Other Treatments (Base Case, Fixed Effects)

ATE = atezolizumab; AVE = avelumab; AXI = axitinib; BEV = bevacizumab; CAB = cabozantinib; CrI = credible interval; EVE = everolimus; IPI = ipilimumab; LEN = lenvatinib; NIV = nivolumab; ORR = objective response rate; PAZ = pazopanib; PEM = pembrolizumab prob = probability; PEM = pembrolizumab; SOR = sorafenib; SUN = sunitinib.

Source: Sponsor-submitted indirect treatment comparison.³¹

Figure 14: OS Results — Lenvatinib Plus Pembrolizumab Versus Other Treatments (Base Case, Fixed Effects)

ATE = atezolizumab; AVE = avelumab; AXI = axitinib; BEV = bevacizumab; CAB = cabozantinib; CrI = credible interval; EVE = everolimus; IFN = interferon; IL = interleukin; IPI = ipilimumab; LEN = lenvatinib; NIV = nivolumab; PAZ = pazopanib; PEM = pembrolizumab prob = probability; SUN = sunitinib.

Source: Sponsor-submitted indirect treatment comparison.³¹

Similar to the analysis of PFS, the analysis of OS by risk status was considered feasible under the assumption that the MSKCC and IMDC risk status definitions were equivalent and analyzed using sensitivity analysis 3 and sensitivity analysis 11, as previously described. The results for relevant comparisons are presented by risk subgroup in Table 32.

Table 32: Redacted

Note: Table and title redacted as per sponsor’s request.

All-Cause Grade 3 and Higher AEs

The base-case analysis of all-cause grade 3 and higher AEs included evidence from 13 trials that involved 12 comparators, including LEN-PEM, AXI-PEM, and PAZO. Each of these interventions was compared with SUN and informed by evidence from 1 trial. Both an FE and RE model were run, and the FE model was selected based on a lower DIC (46.5 versus 47.4), no comparisons that included more than 2 studies, no statistical heterogeneity, and no evidence of inconsistency within the closed loop of the network. |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||.

Figure 15: Redacted

Note: Figure and title redacted as per sponsor’s request.

Treatment Discontinuation Due to AE

The base-case analysis included 14 comparators from 19 RCTs, including comparisons of the following interventions versus SUN: LEN-PEM, AXI-PEM, NIVO-IPI, and PAZO (each informed by evidence from 1 trial). An RE model was selected for analysis based on a lower DIC (68.3 versus 69.2). |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||

Figure 16: Redacted

Note: Figure and title redacted as per sponsor’s request.

Critical Appraisal of Sponsor-Submitted NMA

The methodology used for the study selection in the systematic literature review was pre-specified and used an appropriate set of criteria in terms of the study characteristics for a systematic review, databases searched, data extraction process, and quality assessment. The literature review was comprehensive and was expected to have captured the relevant studies of interest. Based on the systematic literature review, a total of 34 RCTs had eligible populations, comparators, and outcomes of interest for the NMA; however, 10 were excluded from the NMA because they did not form a connected network, did not report a clear cell histology subgroup, did not include interventions of interest, or because of the study design (such as the inclusion of an open-label run-in phase or having varying drug regimens for different patient subgroups). The rationale for the exclusion of these RCTs was reasonable. Despite an inclusive literature search, most of the connections within the network were limited to 1 study. Comparisons of interest (due to their relevance in the Canadian treatment setting) within the network were limited to indirect estimates only and based on 1 open-label RCT; therefore, inconsistency could not be assessed in these connections. Where appropriate, inconsistency was assessed by comparing the results of the Bucher ITCs with that of direct evidence. Areas of inconsistency within the network were explored through sensitivity analyses. Based on a qualitative review of the populations of the included studies, there were some concerns regarding potential bias due to effect modifiers. This included some differences between study populations in terms of the number of metastases, prior nephrectomy, presence of sarcomatoid features, and distribution of patients by risk status that may warrant further review. This remains a source of uncertainty in the network. The quality of included studies was assessed using the Cochrane risk-of-bias assessment tool (2.0), but information about the results of the quality assessment of individual studies was not reported. Additionally, information about study withdrawal or dropouts was not reported, therefore limiting the ability to evaluate the internal validity of the included studies.

The clinical experts consulted by CADTH indicated that the sponsor-submitted NMA considered all relevant comparators in the Canadian context. Specifically, NIVO-IPI, AXI-PEM, and PAZO were identified as the most relevant comparators against LEN-PEM and were therefore included in the evidence informing the NMA. Information about the dosing of treatments included in all of the trials in the network was limited, with details regarding relative dose intensity, compliance, or missed dosing either not reported or poorly reported. Both efficacy (OS, PFS, ORR, CR) and safety (all-cause grade 3 or higher AEs, treatment-related grade 3 or higher AEs, and treatment discontinuation due to AEs) outcomes were assessed in the NMA. All outcomes were reported here except for CR and analyses of treatment-related grade 3 or higher AEs. HRQoL was not included as an outcome in the sponsor-submitted NMA.

As part of the feasibility assessment, the sponsor identified the following parameters as potential effect modifiers a priori: histology, cancer stage, risk score, other patient characteristics, outcome assessment, survival analysis, subgroup definitions, and follow-up duration. The sponsor reported a summary of related findings, assumptions, and recommendations, and whether or not the exclusion of a study, or a sensitivity or subgroup analysis, was required. Some of the patient characteristics were reported inconsistently across trials and, in particular, details about race and ethnicity, PD-L1 status, and cancer staging were reported infrequently. In general, heterogeneity that was identified as a limitation was not adjusted for, but some sensitivity and subgroup analyses were performed. Sensitivity analyses that explored the equivalence of certain comparators (SUN, PAZO, and tivozanib), differences in clear cell histology (included studies reporting on patients with 100% clear cell histology) and in PD-L1 positivity (included studies reporting on subgroups with a PD-L1 expression of at least 1%), were consistent with the base-case analyses. Subgroup analyses were also performed by risk group and risk group definition (IMDC and MSKCC risk scoring). Stratification by risk status at randomization was not reported in the NMA, but a review of the published CLEAR,¹⁰ KEYNOTE-426,¹¹ and CheckMate-214¹² trials indicated that patients were stratified by risk status at randomization. Patients enrolled in COMPARZ³⁴ were not stratified by risk status. Additionally, patients in the poor risk and favourable risk subgroups represented a small proportion of patients in the overall population and, therefore, a small sample size in the trials. Some of the results of the subgroup analyses by risk status were inconsistent with the base-case analyses, particularly among the favourable- and poor risk groups, which are subject to the limitations that have been described. Overall, the interpretation of the results for the subgroup analyses of the NMA is limited.

The sponsor also noted some heterogeneity in terms of race (notably, the proportion of Asian patients ranged from 1% to 25% in trials) and the number of metastases (for example, the proportion of patients with at least 2 metastases ranged from 57% to 89%) among the included trials and reported this as a limitation. Additionally, the proportion of patients with sarcomatoid features was lower in the CLEAR trial (6% to 8%) than in the KEYNOTE-426 trial (18% based on patients with data available) as was the proportion of patients with a history of nephrectomy (74% to 77% for CLEAR, 80% to 83% in KEYNOTE-426 and CheckMate-214). This may suggest the patient population enrolled in CLEAR had less severe disease than in other trials of interest (also, the proportion of patients with poor risk status was lower in CLEAR than in KEYNOTE-426 and CheckMate-214). Differences in time point assessments and actual treatment duration were also acknowledged as a limitation of the NMA, as was the impact of a lack of data maturity on efficacy assessments. A sensitivity analysis was conducted where trials with a follow-up period of less than 12 months were excluded; however, no adjustments were made for the variation in follow-up duration in studies where the duration was greater than 12 months. For reference, in the CLEAR trial,¹⁰ the analysis of OS was based on data with a median follow-up of approximately 33 to 34 months, and the analysis of PFS was based on a median follow-up of 26 to 27 months. The results for OS and PFS were based on a median follow-up of 43 months in the KEYNOTE-426 trial¹¹ and a minimum of 48 months in CheckMate-214.¹² The impact of the heterogeneity in the follow-up duration on these outcomes is unknown.

The sponsor-submitted ITC included justification of model selection (FE versus RE) based on an assessment of model fit or a lower DIC, although reported differences were very small. Assessments of heterogeneity based on I² and inconsistency were also considered, although most connections were formed by a single RCT and there were few closed loops. The RE model used an informative before stabilize estimates of between-study variance. The prior was based on plausible values, and sensitivity analyses were conducted. There was uncertainty in the results, with wide CrIs. This is likely due to the sparsity of the network. The results for the ORR had very wide CrIs and the results for OS and all-cause grade 3 or higher AEs included CrIs that crossed 1 and included values suggesting a strong treatment effect, thus limiting the interpretation of these results. The analysis of treatment discontinuation due to AEs was also associated with a lack of precision and uncertainty due to wide CrIs that crossed 1 while including values suggesting a strong treatment effect, although the FE model improved precision.

Description of Published Indirect Comparisons

Four ITCs that evaluated the efficacy and safety of LEN-PEM in patients with advanced or metastatic RCC were identified in the literature and summarized for this review.^35-38

Objectives

• The objective of the ITC published by Quhal et al. (2021a)³⁵ was to perform indirect comparisons of efficacy and safety of first-line immune-checkpoint inhibitor (ICI)-based combination therapies for metastatic RCC.

• The objective of the ITC published by Quhal et al. (2021b)³⁶ was to compare the safety profiles of systemic ICI-based combination therapies that were evaluated in the first-line setting for the management of patients with advanced or metastatic RCC.

• The objective of the ITC published by Cattrini et al. (2021)³⁷ was to address the lack of head-to-head comparisons and the uncertainty of the benefit from immunotherapy-based combinations in all of the IMDC subgroups.

• The objective of the ITC published by Nocera et al. (2021)³⁸ was to provide formal comparisons among immune-oncology combinations in terms of OS, PFS, ORR, and TRAEs.

Study Selection Methods

A summary of the study selection methods for each of the published ITCs is presented in Table 33. All of the published ITCs included RCTs that studied patients with advanced or metastatic clear cell RCC who received first-line treatment with an ICI-based combination. The comparators were SUN or standard of care. Three of the ITCs^35,37,38 included outcomes related to efficacy (such as OS and PFS) and safety (TRAEs and treatment discontinuation), and 1 ITC³⁶ included only studies that reported safety outcomes (TRAEs, treatment discontinuation, and treatment-related mortality).

All of the published ITCs conducted their literature search using PubMed. Other databases used included Web of Science, Scopus, Embase, and Cochrane Library. The bibliographies of included studies and/or relevant conference abstracts were also searched in 2 of the ITCs.^35,38 Where reported, studies were selected by 2 independent reviewers based on title and abstract screening with disagreements resolved by consensus. Details about study selection were not provided for 1 of the published ITCs.³⁶ Two of the published ITCs reported details of the data extraction process either performed or verified in duplicate by 2 independent reviewers.^35,37 Details of the data extraction process were not provided for the other 2 published ITCs.^36,38 In 3 of the 4 published ITCs, the Cochrane risk-of-bias tool was used to assess study quality and, in 1 ITC, study quality was assessed using the Jadad score.

Table 33: Study Selection Criteria and Methods for Published ITCs

AE = adverse event; ASCO = American Society of Clinical Oncology; BSC = best supportive care; CENTRAL = Cochrane Central Register of Controlled Trial; CRR = complete response rate; ICI = immune-checkpoint inhibitor; IMDC = International Metastatic Renal Cell Carcinoma Database Consortium; ITC = indirect treatment comparison; NA = not applicable; NR = not reported; ORR = overall response rate; OS = overall survival; PD-L1 = programmed cell death 1 ligand 1; PFS = progression-free survival; RCC = renal cell carcinoma; RCT = randomized controlled trial; TRAE = treatment-related adverse event.

^aExcluded study designs: observational studies, reviews, letters, editorials, replies from authors, case reports.

^bExcluded study designs: observational studies, review articles, commentaries, editorials, and articles without peer review.

Source: Quhal et al. (2021a),³⁵ Quhal et al. (2021b),³⁶ Cattrini et al. (2021),³⁷ Nocera et al. (2021).³⁸

Methods of Published Indirect Comparisons

ITC Analysis Methods

All of the published ITCs reported only limited details about the ITC analysis methods used (Table 34). All of the published ITCs reported that an NMA was conducted using both RE and FE models. The 2 ITCs published by Quhal et al. (2021) did not specify whether a Bayesian or frequentist approach was used; however, they did note whether contrast-based analyses or arms-based analyses were used for specific outcomes (described in Table 34). The ITC published by Cattrini et al. (2021) reported the use of a Bayesian approach and the ITC published by Nocera et al. (2021) reported the use of a frequentist approach. No information about priors or convergence was reported. Inconsistency was not assessed, as the networks did not include any closed loops (all comparisons were made to SUN). The only study that reported information about the assessment of model fit was the Cattrini et al. (2021) ITC, which used heterogeneity assessed by I² to determine model fit. Additional information about the assessment of model fit was not provided.

Table 34: ITC Analysis Methods

AE = adverse event; CR = complete response; CRR = complete response rate; FE = fixed effects; IMDC = International Metastatic Renal Cell Carcinoma Database Consortium; ITC = indirect treatment comparison; NA = not applicable; NMA = network meta-analysis; NR = not reported; ORR = objective response rate; OS = overall survival; PD-L1 = programmed cell death 1 ligand 1; PFS = progression-free survival; RE = random effects; SD = stable disease; TRAE = treatment-related adverse event.

Source: Quhal et al. (2021a),³⁵ Quhal et al., (2021b),³⁶ Cattrini et al. (2021),³⁷ Nocera et al. (2021).³⁸

The 4 published ITCs estimated the relative ranking of alternative treatments by outcome using a P-score or a surface under the cumulative ranking (SUCRA) analysis. All outcomes that were reported in the published ITCs were pre-specified, which include:

• Quhal et al. (2021a): OS, PFS, CRR, and ORR

• Quhal et al. (2021b): Treatment-related morality, treatment discontinuation due to AEs, grade 3 and higher TRAEs

• Cattrini et al. (2021): OS, OS by subgroups (IMDC risk status, PD-L1 expression status) and grade 3 and higher AEs

• Nocera et al. (2021): OS, PFS, CR, any response (CR + PR) and no progression rate (CR + PR + stable disease), and grade 3+ TRAEs

As noted, Cattrini (2021) evaluated OS by subgroup, which was indicated as a secondary outcome of the NMA. The authors reported that if data were not available for the risk subgroups, the estimate for the poor and intermediate subgroup was obtained by pooling the HRs and 95% CIs (or performing a meta-analysis) of the estimates from the poor and intermediate subgroups. Information about the subgroup analyses in the Quhal et al. (2021a) ITC was limited. None of the ITCs reported sensitivity analyses and no steps were taken to address potential sources of heterogeneity.

Results of Published Indirect Comparisons

Summary of Included Studies

A summary of the studies included in the published ITCs is provided in Table 35. A total of 6 phase III open-label RCTs were included in the Quhal (2021a), Quhal (2021 b), and Cattrini (2021) ITCs. In the Nocera (2021) ITC, 4 of the 6 ITCs were included (all except IMmotion 151 and JAVELIN Renal 101). All of the trials included in the published ITCs were also included in the sponsor-submitted NMA described in the previous section. The following interventions were included in the 4 common RCTs: LEN-PEM, AXI-PEM, NIVO-IPI, and NIVO plus cabozantinib. The additional 2 RCTs included the 3 published ITCs evaluated atezolizumab plus bevacizumab and avelumab plus AXI. All RCTs included SUN as the comparator.

Table 35: Studies Included in the Published ITCs

AE = adverse event; ITC = indirect treatment comparison.

Source: Quhal et al. (2021a),³⁵ Quhal et al. (2021b),³⁶ Cattrini et al. (2021),³⁷ Nocera et al. (2021).³⁸

The main study characteristics were poorly reported in all of the published ITCs except Cattrini et al. (2021), which the following information is derived from. Among the 6 included RCTs, the number of patients enrolled ranged from 651 to 1,096 and the median follow-up for OS ranged from 18.1 months to 55.0 months. The primary end point included PFS in all studies, in addition to OS in 4 of the RCTs and ORR in 1 of the RCTs. Using the IMDC risk status, the proportion of patients with a favourable risk status ranged from 18.8% to 32.9%, intermediate ranged from 56.5% to 64.1%, and poor risk status ranged from 9.8% to 19.8%.

Results

A network diagram was reported for the Quhal et al. (2021a) and Nocera et al. (2021) ITCs, as presented in Figure 17 and Figure 18, respectively. The network described for Quhal et al. (2021a) should apply to the remaining 2 ITCs (Quhal et al. [2021b]; Cattrini et al. [2021]). Each connection was informed by 1 RCT and all comparisons between interventions were based on indirect evidence only.

Figure 17: Network Plots for Overall Survival and Progression-Free Survival

Source: Quhal et al. (2021a).³⁵

Figure 18: Network Diagram for Overall Survival and Progression-Free Survival

Source: Nocera et al. (2021).³⁸

Efficacy Outcomes

In all of the published ITCs, outcomes were assessed by ranking the comparisons made between the interventions and SUN. Individual estimates of treatment effects for the indirect comparisons involving LEN-PEM versus other combination therapies were not reported for any outcomes. Due to the lack of robustness associated with the networks and the lack of estimates of the treatment effects of interest, a detailed description of the results of these analyses has not been presented. Briefly, in the 3 ITCs^35,37,38 that analyzed OS, all treatments demonstrated benefit in OS compared with SUN. None of the ITCs reported LEN-PEM as the preferred treatment based on the ranking results for OS. In the 2 ITCs^35,38 that analyzed PFS, all comparisons demonstrated benefit in PFS compared with SUN except for NIVO-IPI, and LEN-PEM was the preferred treatment based on the ranking results for PFS. In the 2 ITCs^35,38 that analyzed ORR, all comparisons demonstrated benefit compared with SUN except for NIVO-IPI, which yielded different results in the ITCs. Both ITCs reported that LEN-PEM was the preferred treatment in terms of ORR based on the ranking results.

Safety Outcomes (Treatment Discontinuation Due to AEs and Grade 3 and Higher AEs)

The ITC published by Quhal et al. (2021b)³⁶ analyzed treatment discontinuation due to AEs, and LEN-PEM was reported as being associated with the highest likelihood of discontinuation relative to other treatments based on ranking results, including AXI-PEM and NIVO-IPI. The ITC published by Cattrini et al. (2021)³⁷ analyzed grade 3 and higher AEs based on a ranking of comparisons to SUN. LEN-PEM was not reported as being the preferred treatment based on grade 3 and higher AEs.

Critical Appraisal of Published Indirect Comparisons

The systematic literature review informing the published ITCs was limited in terms of the databases searched; however, key trials of relevance to the review of LEN-PEM were identified and included in each of the 4 published ITCs. The included trials were based on studies of acceptable quality as assessed by the Cochrane risk-of-bias tool (version 2.0 or not reported) or the Jadad score, although the risk of bias was driven predominantly by the open-label study designs of the included RCTs. Each of the publications acknowledged heterogeneity among the included trials resulting from variation in patient characteristics and follow-up duration. Additionally, the small number of studies informing indirect comparisons was a weakness of the networks.

The methodology used for the analyses reported in the published ITCs lacked important details, which hindered the ability to appropriately interpret the reported results. All of the published ITCs evaluated outcomes based on comparisons between the interventions and SUN and the subsequent ranking of comparisons to determine the preferred treatment. Individual estimates of treatment effects for the indirect comparisons of LEN-PEM versus other combination therapies were not reported for any outcomes. Due to issues related to the methodology and robustness of the networks, this method of analysis does not adequately provide evidence of the relative efficacy and safety of LEN-PEM compared with the other treatments assessed in the network.

Discussion

Summary of Available Evidence

The CADTH systematic review included evidence from 1 pivotal study (CLEAR trial) and 5 ITCs (1 sponsor-submitted ITC and 4 ITCs identified from published literature). Additional input from patient groups, clinician groups, drug plans, and the clinical experts consulted was considered during the review.

The CLEAR trial is an ongoing multi-centre, randomized, parallel-arm, open-label, phase III study comparing the efficacy and safety of LEN in combination with either everolimus or PEM versus SUN as a first-line treatment option in adult patients with advanced RCC. The study enrolled patients who were 18 years and older with a histologically or cytologically confirmed clear cell component and documented evidence of advanced disease. Patients were randomized in a 1:1:1 ratio into 3 treatment arms (LEN-PEM, LEN plus everolimus, and SUN) based on 2 stratification factors: geographic region and the MSKCC prognostic risk groups. The primary outcome of the CLEAR trial, PFS, was measured by IIR using RECIST v.1.1 criteria. Other secondary and exploratory outcomes investigated included OS, ORR, HRQoL, safety and tolerability, DOR, DCR, and so forth.²⁸

Patients in the CLEAR trial received either 20 mg of LEN orally once daily plus 200 mg of PEM administered intravenously every 3 weeks, or 50 mg of SUN orally once daily for 4 weeks followed by a 2-week off-treatment period until the investigator discontinued treatment, patient withdrew consent, or patient moved into the follow-up phase.²⁸ The median age of patients randomized into the CLEAR study was 62 years; more males were enrolled compared with females and the majority of patients were White or of Asian descent. There were more patients with a KPS score of 80 or greater in the LEN-PEM and SUN arms compared with patients with a KPS score of less than 80. Baseline characteristics were balanced across the 2 study arms with the exception of age (more patients randomized in the SUN arm were younger than 65 years compared with those in the LEN-PEM arm).

The sponsor’s submission included an NMA to support the evidence generated in the CLEAR trial. The NMA presented comparative evidence between LEN-PEM and other approved therapies, other than SUN, and was appraised for this review. Four additional published ITCs that evaluated the efficacy and safety of LEN-PEM in patients with advanced or metastatic RCC were summarized and appraised for this review.

Interpretation of Results

Efficacy

Both the clinical experts and clinician groups consulted during the CADTH review highlighted improved OS and PFS, reduction in tumour size (ORR), and improved QoL as treatment goals for untreated patients with metastatic or advanced RCC. The patient groups highlighted the need for more treatment options that allow oncologists to set up individualized treatment plans for patients. Improved OS, reduction in disease progression, and control drug resistance were some of the goals cited by the patient groups consulted.

The CADTH review protocol identified PFS, ORR, OS, DOR, HRQoL, DCR, time to treatment discontinuation, and improvement in 3 subgroups (the IMDC prognostic risk groups: poor, intermediate, and favourable), as important outcomes for patients. These outcomes were pre-specified in the CLEAR trial before the data cut-off for interim analysis 3 (August 28, 2020) and the type I error rate was adequately controlled for PFS, OS, and ORR. Sensitivity analyses were conducted for PFS and the findings obtained were consistent with the primary PFS findings. Analyses of other outcomes such as HRQoL, DOR, and DCR were not adjusted for multiplicity during the analysis.

The study was powered to detect a 40% difference in PFS between the 2 treatment arms. The primary outcome was statistically significant at the data cut-off for the final PFS analysis (interim analysis 3). The median PFS estimated by IIR was 23.9 months (95% CI, 20.8 to 27.7) in the LEN-PEM arm while, in the SUN arm, the median PFS estimated was 9.2 months (95% CI, 6.0 to 11.0). The HR between the LEN-PEM and SUN arms was 0.39 (95% CI, 0.32 to 0.49; P < 0.0001), corresponding to a 61% reduction in the risk of disease progression or death at any given time for LEN-PEM compared with SUN. The clinical experts consulted considered the tumour response in the LEN-PEM arm to be large compared with the SUN arm. The experts acknowledged that the findings were clinically meaningful and important to patients in the Canadian practice setting. The PFS findings were consistent across the IMDC prognostic groups (poor, intermediate, and favourable), suggesting that LEN-PEM treatment was beneficial to patients in all 3 groups; this finding was consistent with the primary analysis. The clinical experts considered these PFS estimates to be clinically meaningful.

The ORR estimated by IIR was 71% (95% CI, 66.3 to 75.7) in the LEN-PEM arm, with 16.1% of patients achieving a CR and 54.9% achieving a PR. In the SUN arm, the estimated ORR was 36.1% (95% CI, 31.2 to 41.1) with 4.2% of patients achieving a CR and 31.9% achieving a PR. The findings were considered statistically significant. The clinical experts consulted considered the tumour response in the LEN-PEM arm to be considerably large compared with the SUN arm and clinically meaningful. The experts noted the tumour response observed in the LEN-PEM arm was larger than responses observed from the use of the other treatment options available in practice. The DCR rate (CR plus PR plus stable disease) estimated by IIR in the LEN-PEM arm was 90.1% (95% CI, 87.0 to 93.2) and 74.2% (95% CI, 69.7 to 78.8) in the SUN arm. The OR estimated between LEN-PEM and SUN was 4.35 (95% CI, 3.16 to 5.97). The experts mentioned that the DCR observed in patients receiving SUN in practice was lower compared with the estimates observed in the CLEAR trial. The OR obtained between LEN-PEM and SUN was 3.26 (95% CI, 2.13 to 5.0). The median DOR estimated in the LEN-PEM arm was 25.8 months (95% CI, 22.1 to 27.9) and, in the SUN arm, it was 14.6 months (95% CI, 9.4 to 16.7).⁹

The median OS was not estimable at either the August 28, 2020, data cut-off or the follow-up data cut-off (March 31, 2021). The HR estimated between the LEN-PEM arm and the SUN arm was 0.66 (95% CI, 0.49 to 0.88; P = 0.0049), representing a 34% reduction in the risk of death in the LEN-PEM arm compared with the SUN arm at the August 28, 2020, data cut-off, and an HR of 0.72 (95% CI, 0.55 to 0.93) at the follow-up data cut-off (March 31, 2021), representing a 28% reduction in risk of death in the LEN-PEM arm versus the SUN arm.⁹ The HR findings were considered statistically significant. The final OS result is planned to be obtained at another interim analysis, once a pre-specified number of events have accrued.

The clinical experts consulted highlighted that patients with advanced or metastatic RCC with a clear cell component will benefit from the treatment even if they have additional clinical features (other histologies in addition to the clear cell component). The clinical experts acknowledged that patients with KPS scores of less than 70 will not likely benefit from the treatment because of their poor performance status.

The clinician and patient groups consulted during the CADTH review highlighted HRQoL as an important treatment goal. In the CLEAR trial, HRQoL was assessed using 3 questionnaires. The FKSI-DRS has been validated in RCC populations, with MIDs identified in the literature that range from 0.62 to 3 points based on calculations using different anchors. Although the EORTC QLQ-C30 and EQ-5D-3L instruments have been validated in other cancers, no studies evaluating the validity, reliability, and responsiveness of these instruments in patients with advanced or metastatic RCC were identified in published literature. The clinical experts highlighted that these questionnaires were appropriate for the RCC patient population. In their opinion, these scales are most sensitive to symptom-related changes due to RCC (if the therapy is active with associated improvements in symptoms) but not as sensitive to changes in symptoms owing to TRAEs. The rates of questionnaire completion declined below 50% at cycle 26 for LEN-PEM and at cycle 12 for the SUN arm owing to treatment discontinuation. The sponsor did not provide information on how missing data were handled during the analyses. It is therefore uncertain whether the methods used to address missing data were appropriate and free of bias. Further, no adjustments for multiplicity were made in the analyses; thus, the results were considered descriptive. It is uncertain whether LEN-PEM improves HRQoL in patients with advanced RCC, given the limitations identified in the analyses.

There was no direct evidence available to assess the relative efficacy of LEN-PEM versus other current standard-of-care therapies. Indirect evidence of LEN-PEM for first-line treatment of patients with advanced or metastatic RCC was available based on 5 ITCs: 1 NMA submitted by the sponsor and 4 ITCs identified in the published literature.^35-38

The sponsor-submitted NMA was designed to evaluate the efficacy of LEN-PEM compared with other available therapies. The relative efficacy was assessed as OS, PFS, ORR, and safety outcomes (HRQoL was not included as an outcome) using a Bayesian NMA approach. All comparators included in the NMA were considered relevant by the clinical experts consulted; NIVO-IPI, AXI-PEM, and PAZO were identified as the most relevant. For PFS, based on an RE model, LEN-PEM showed benefit compared with NIVO-IPI and PAZO. The RE model did not show a difference for the comparison with AXI-PEM, whereas the results for the FE model favoured LEN-PEM. Similarly, the results of the analysis of ORR based on an FE model showed a benefit with LEN-PEM when compared with other treatments. The analysis of OS did not show a difference for LEN-PEM compared with other treatments. The demonstration of a benefit with LEN-PEM in terms of PFS but not in OS was not consistent with the expectations of the clinical experts consulted by CADTH. Sources of heterogeneity between the included RCTs, a sparse network, and a lack of data maturity (shorter follow-up duration) for the CLEAR trial are potential sources of uncertainty that limit the analysis.

The 4 published ITCs included a subset of studies that were included in the sponsor-submitted NMA. Limited details about the methodology used for the published ITCs were provided, preventing an adequate interpretation of the results. Estimates of indirect treatment effects between LEN-PEM and other combinations were not reported. Despite the limitations, the results that were available, including an assessment of efficacy based on the rankings for OS, PFS, and ORR, as well as safety outcomes, were consistent with the results observed in the sponsor-submitted NMA.

Harms

Overall, the proportion of patients reporting at least 1 AE was comparable in both study arms (99.7% in the LEN-PEM arm and 98.5% in the SUN arm). Diarrhea, hypertension, hypothyroidism, decreased appetite, fatigue, nausea, and stomatitis were the most common AEs reported in patients receiving LEN-PEM while, in the SUN arm, diarrhea, hypertension, stomatitis, PPE syndrome, fatigue, nausea, and decreased appetite were the most common AEs.

SAEs were reported in 50.6% of patients receiving LEN-PEM compared with 33.2% of patients receiving SUN. There were more AEs leading to drug discontinuations (37.2% versus 14.4%), dose reductions (68.8% versus 50.3%), drug interruptions (78.4% versus 53.8%), and dose modifications (87.5% versus 70.3%) in the LEN-PEM arm compared with the SUN arm, respectively. Overall, more deaths were reported in the SUN arm (29.1%) compared with the LEN-PEM arm (22.2%). Of note, patients receiving LEN-PEM had longer exposures to the study treatment (mean = 17.29 months) compared with patients receiving SUN (mean = 11.33 months), which may have influenced the reporting of AEs.⁹

The most common notable harms reported in the 2 arms were hypertension, hypothyroidism, hepatotoxicity, proteinuria, hemorrhage, PPE, renal events, QT prolongation, arterial thromboembolic events, gastrointestinal perforation, hypocalcemia, cardiac dysfunction, fistula formation, and posterior reversible encephalopathy syndrome. More AEs associated with hypertension, hypothyroidism, hepatotoxicity, proteinuria, and QT prolongation were reported in patients receiving LEN-PEM compared with patients receiving SUN.

The clinical experts indicated that the safety profiles of LEN-PEM and SUN in the CLEAR trial were similar and are consistent with the safety profiles of current treatment options in practice. The experts considered the safety profile of LEN-PEM to be acceptable and manageable in practice. The experts noted similarities in the notable harms reported in both treatment arms and indicated that these were consistent with the AEs reported for other treatment options used in practice. As noted by the clinical experts, patients receiving current treatment options require frequent monitoring and dose adjustments (reductions, modifications, or withdrawal of a drug) to manage TRAEs (e.g., hypertensive, hepatotoxic, and immune-related events) related to treatments with PEM combinations. The clinical experts anticipated implementing intensive strategies similar to those currently used in practice to manage AEs from LEN-PEM.

The indirect evidence available for the comparison of LEN-PEM versus the other treatments for advanced or metastatic RCC that were previously described also included an analysis of safety outcomes. ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||. The 4 published ITCs were subject to substantial uncertainty, but the results of the assessment of safety were consistent with the results observed in the sponsor-submitted NMA.

Conclusions

One pivotal study and 5 ITCs provided evidence for the CADTH systematic review. This review focused on the comparison between LEN-PEM and SUN investigated in the CLEAR trial as per the sponsor’s reimbursement request and the Health Canada indication. No other evidence directly comparing LEN-PEM with other standard therapies for advanced or metastatic RCC was identified. In CLEAR, the median PFS estimated by IIR at the final interim analysis for PFS (August 28, 2020) was 23.9 months in patients receiving LEN-PEM compared with 9.2 months in patients receiving SUN. The HR estimated for PFS between LEN-PEM against SUN was considered statistically and clinically significant. The median OS was not estimable in either study arm at the data cut-off for interim analysis 3 (August 28, 2020) or at the data cut-off for the follow-up analysis (March 31, 2021). However, the HR estimated between LEN-PEM against SUN was considered statistically significant. The ORR estimated in the LEN-PEM arm was also considered statistically significant. The HRQoL assessments were considered exploratory due to the lack of multiplicity adjustments in the analysis and the potential for reporting bias. The findings of the CLEAR trial were considered by the clinical experts consulted during the review to be meaningful for patients with advanced or metastatic RCC and were aligned with outcomes of importance to patients. In the opinion of the clinical experts, clinical judgment is required to evaluate the clinical benefit of LEN-PEM and the management in practice of AE . The experts anticipate that the TRAEs resulting from the use of LEN-PEM will be managed in practice using strategies similar to those already in place for other treatment options (frequent AE monitoring and dose adjustments, reductions, and modifications are anticipated for the treatment). The open-label design was a key limitation of the CLEAR trial, and the OS data are interim. The study was randomized and adjustments for multiplicity were conducted for key outcomes (PFS, OS, and ORR), which minimized bias in the study. The clinical experts considered the baseline characteristics and the findings from the CLEAR trial to be generalizable to patients diagnosed in the first-line setting in Canada with advanced or metastatic RCC with at least a clear cell component.

No direct evidence was available to assess the relative efficacy of LEN-PEM versus other current standard-of-care therapies. Indirect evidence of LEN-PEM for the first-line treatment of patients with advanced or metastatic RCC was available based on 5 ITCs: 1 NMA submitted by the sponsor and 4 ITCs identified in published literature. The sponsor-submitted NMA of LEN-PEM compared with other available therapies showed benefit for LEN-PEM for PFS and ORR but not for OS compared with other therapies. Sources of uncertainty identified during the review included heterogeneity in the RCTs, sparse network, and lack of data maturity (shorter follow-up duration) for the CLEAR trial. The sponsor-submitted NMA results of the analysis of treatment discontinuation due to AEs ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||, although these results were limited by a lack of precision in addition to a number of assumptions made about the outcome that cause uncertainty in the results. The OS, PFS, and ORR findings obtained from the 4 additional published ITCs assessed in this review were consistent with the results of the sponsor-submitted NMA. However, the methodology used for the analyses lacked important details, which hindered the ability to appropriately interpret the reported results.

References

1.Lenvima (lenvatinib capsules), 4 mg and 10 mg lenvatinib (as lenvatinib mesylate) [product monograph]. Mississauga (ON): Eisai Limited; 2021 May 3.

2.Kidney cancer, version 4.2022. (NCCN Clinical Practice Guidelines in Oncology). Plymouth Meeting (PA): National Comprehensive Cancer Network; 2021 Dec 21: https://www.nccn.org/professionals/physician_gls/pdf/kidney.pdf (free registration required). Accessed 2022 Jan 23.

3.National Cancer Institute. Renal Cell Cancer Treatment (PDQ®) – Patient Version 2021; https://www.cancer.gov/types/kidney/patient/kidney-treatment-pdq#section/all. Accessed 2022 Jan 23.

4.National Comprehensive Cancer Network (NCCN). Kidney cancer - guidelines for patients. 2022; https://www.nccn.org/patients/guidelines/content/PDF/kidney-patient.pdf. Accessed 2022 Jan 23.

5.Flanigan RC, Campbell SC, Clark JI, Picken MM. Metastatic renal cell carcinoma. Curr Treat Options Oncol. 2003;4(5):385-390. PubMed

6.BC Cancer Agency. Kidney cancer. 2022; http://www.bccancer.bc.ca/health-info/types-of-cancer/urinary/kidney. Accessed 2022 Jan 23.

7.Canil C, Kapoor A, Basappa NS, et al. Management of advanced kidney cancer: Kidney Cancer Research Network of Canada (KCRNC) consensus update 2021. Can Urol Assoc J. 2021;15(4):84-97. PubMed

8.Rini BI, Atkins MB, Plimack ER, et al. Characterization and management of treatment-emergent hepatic toxicity in patients with advanced renal cell carcinoma receiving first-line pembrolizumab plus axitinib. Results from the KEYNOTE-426 trial. Eur Urol Oncol. 2021. PubMed

9.Drug Reimbursement Review sponsor submission: Lenvima (lenvatinib) in combination with pembrolizumab (Keytruda), 4 mg and 10 mg lenvatinib mesylate capsules [internal sponsor's package]. Mississauga (ON): Eisai Limited; 2021 Nov.

10.Motzer R, Alekseev B, Rha SY, et al. Lenvatinib plus pembrolizumab or everolimus for advanced renal cell carcinoma. N Engl J Med. 2021;384(14):1289-1300. PubMed

11.Rini BI, Plimack ER, Stus V, et al. Pembrolizumab plus axitinib versus sunitinib for advanced renal-cell carcinoma. N Engl J Med. 2019;380(12):1116-1127. PubMed

12.Motzer RJ, Tannir NM, McDermott DF, et al. Nivolumab plus ipilimumab versus sunitinib in advanced renal-cell carcinoma. N Engl J Med. 2018;378(14):1277-1290. PubMed