CADTH Reimbursement Review

Tebentafusp (Kimmtrak)

Sponsor: Medison Pharma Canada Inc.

Therapeutic area: Unresectable or metastatic uveal melanoma

This multi-part report includes:

Clinical Review

Pharmacoeconomic Review

Stakeholder Input

Clinical Review

Executive Summary

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

Introduction

Uveal melanoma (UM) is a rare subset of melanoma that arises from the uveal tract (which consists of the choroid, ciliary body, and iris) in the eye.¹ It is estimated that there were 3.75 new cases of UM per million in Canada each year from 1992 to 2010, with an increase of 0.074 new cases per million individuals annually.² At diagnosis, most patients have localized disease and approximately 50% of patients are symptomatic (vision loss or disturbances).¹ About 50% of patients will progress to metastatic disease,³ with metastasis most commonly in the liver (93%).¹ Survival of patients with metastatic uveal melanoma (mUM) is not favourable; the estimated 1-year survival rate is 43% to 52% in the first-line setting, and about 37% in the second-line setting or later.^4,5 The leading cause of death is parenchymal liver failure.⁶

Patient with a limited number of metastatic lesions in the liver alone (oligometastatic disease) may benefit from local therapies. Treatment modalities include radiofrequency ablation, stereotactic radiation, and surgical resection. In contrast, patients with mUM who have a larger number of metastatic lesions and/or disease external to the liver are usually prescribed systemic therapies, according to the clinical expert consulted by CADTH. The most commonly prescribed systemic therapies in the first-line setting are immunotherapies (off-label), administered as monotherapies (i.e., nivolumab or pembrolizumab alone) or in combination (i.e., ipilimumab plus nivolumab). Ipilimumab monotherapy is generally given in later-line settings, according to the clinical expert. Systemic chemotherapies have a limited role in the treatment of mUM in Canada because of toxicity and a low response rate, the clinical expert explained. Given that there are poor survival benefits and a poor tumour response to all available systemic treatments, the standard of care for mUM is enrolment in a clinical trial, if available.^7,8 Of note, funding for systemic therapies for mUM is usually assessed on a case-by-case basis by the jurisdictions.

Tebentafusp is a glycoprotein (gp)100 peptide-human leukocyte antigen (HLA)-A*02:01 directed T-cell receptor CD3 bispecific T-cell engager, which has been shown, in vitro, to activate polyclonal T-cells to release inflammatory cytokines and cytolytic proteins, resulting in the direct lysis of UM tumour cells. It is the first treatment approved by Health Canada specifically for the treatment of HLA-A*02:01-positive adults with unresectable UM or mUM. The recommended dose of tebentafusp, administered by IV infusion, is 20 mcg on day 1, 30 mcg on day 8, 68 mcg on day 15, and 68 mcg once every week thereafter. Per the Health Canada product monograph, it is recommended that patients be treated until unacceptable toxicity or disease progression occur.⁹ The first 3 infusions should be administered in an appropriate health care setting, where the patient can be monitored during the infusion and for at least 16 hours after to mitigate the risk of cytokine release syndrome (CRS).⁹ Only patients with a positive HLA-A*02:01 genotype status are eligible for treatment with tebentafusp.⁹

The objective of this report was to perform a systematic review of the beneficial and harmful effects of tebentafusp (100 mcg/0.5 mL) solution administered by IV infusion for the treatment of HLA-A*02:01-positive adults with unresectable UM or mUM.

Stakeholder Perspectives

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

Patient Input

Two patient groups provided input for this reimbursement review: Melanoma Canada (MC) and Save Your Skin Foundation (SYSF). MC gathered information from patients with UM and their caregivers (n = 19) using an online survey. SYSF collected responses from patients with ocular melanoma (n = 38) using patient interviews, patient roundtables, and online surveys. Most of the SYSF respondents (n = 33) resided in Canada. The majority of respondents to the MC and SYSF requests for input were diagnosed with early-stage or primary disease.

MC respondents mentioned that the diagnosis of UM affects their day-to-day life and quality of life (QoL), and reported that the most common issues are loss of vision, vision impairment, fear or anxiety, depression, and fatigue. SYSF respondents mentioned that their balance is affected, which causes huge physical and psychological deterioration.

Five MC respondents and 10 SYSF respondents indicated that they had experience with tebentafusp through clinical trials or a Health Canada program for compassionate access. According to the submissions by MC and SYSF, 3 patients (2 MC respondents and 1 SYSF respondents) who had access to tebentafusp indicated that the drug was effective in slowing down disease progression, and another 2 patients could not comment on the effectiveness because it was too early for them to tell. The side effects frequently reported by patients from both groups were skin rash, fever, fatigue, cognitive impairment, gastrointestinal issues, nausea, muscle, joint pain, and headaches. Most patients described the side effects as short-term, tolerable, and manageable, and reported that, in their experience, the benefits of the treatments outweighed the negative side effects. Only 1 patient from SYSF said the side effects were not manageable. Patients from both groups reported that they had to travel long distances to access the drug and had financial difficulties.

Patients from both groups expressed their desire for an effective treatment that can preserve vision and provide a good QoL and longer survival. In addition, patients interview by MC indicated that they would like to know of improvements in earlier diagnosis and the detection of metastasis.

Clinician Input

Input From Clinical Expert Consulted by CADTH

The clinical expert highlighted that mUM is an aggressive disease with poor survival and no currently available therapies that predictably improve outcomes. The clinical expert expressed that a shift in the current treatment paradigm is anticipated, in which tebentafusp is used in the first-line setting or later, and is supported by clinical trial and anecdotal evidence.

The clinical expert indicated that only HLA-A*02:01-positive patients with mUM are expected to benefit from tebentafusp, owing to its unique mechanism of action. However, the clinical expert was unable to identify which HLA-A*02:01-positive patients are most likely to benefit from tebentafusp, noting that there is currently no good clinical or biologic predictor of response to tebentafusp. The clinical expert indicated that patients with a poor Eastern Cooperative Oncology Group Performance Status [ECOG PS], (3 or above) are generally not eligible for treatment in clinical practice as they are unlikely to benefit.

The clinical expert strongly suggested that patients be allowed to continue treatment as long as they continue to derive clinical benefits from tebentafusp, noting that there is some evidence from a posthoc analysis of Study 202 showing that patients with radiographic progression on tebentafusp can continue to benefit from treatment beyond progression.¹⁰ The clinical expert highlighted that, given the complexity of the clinical considerations involved, treatment response and the decision to discontinue treatment should be left to the discretion of the attending oncologist, based on assessments of history and physical examinations (every 3 to 4 weeks), laboratory tests (weekly), and imaging (every 12 to 16 weeks) findings. According to the clinical expert, treatment discontinuation is generally considered in clinical practice when a patient develops intolerable toxicities or clear evidence of significant progression, which is indicated by a decline in ECOG PS, an increase in pain, rising lactate dehydrogenase (LDH) levels, and marked radiographic progression.

The clinical expert recommended that tebentafusp initially be prescribed only by specialist physicians experienced in the use of tebentafusp and familiar with the management of CRS, noting that the risk of CRS is highest with the first 3 or 4 doses and substantially lower with subsequent doses. The clinical expert noted that once a pattern of use is established and the risk of CRS is absent, subsequent infusions can be performed in a community clinic setting.

Clinician Group Input

CADTH received input from 1 clinician group — the Ontario Health (Cancer Care Ontario) Skin Cancer Drug Advisory Committee — based on responses from 6 clinicians. The Skin Cancer Drug Advisory Committee provides timely evidence-based clinical and health-system guidance on drug-related issues in support of Cancer Care Ontario’s mandate, including the Provincial Drug Reimbursement Programs and the Systemic Treatment Program. No major contrary views from those provided by the clinical expert consulted by CADTH for this review were presented.

The clinician group stated that UM, which is distinct from cutaneous melanoma, is a disease that has a low tumour mutational burden and a low 1-year survival rate (50%) in the metastatic setting. None of the current systemic treatments for UM have been proven to have an overall survival (OS) benefit, which is a major unmet need in patients with mUM. The clinician group noted that tebentafusp has demonstrated an improvement in OS in clinical trials. The clinician group noted that the treatment goal is to improve OS and QoL. The clinician group explained that tebentafusp would be the first-line treatment of choice for HLA-A*02:01-positive patients with mUM, although they did not comment on the use of tebentafusp in second-line settings or later. The clinician group stated that HLA-A*02:01-positive patients with mUM who do not meet the exclusion criteria of Study 202 would be suitable for tebentafusp. They noted that ongoing ECOG PS, tumour size, and Response Evaluation Criteria in Solid Tumours (RECIST) progression requirements are the clinical outcomes used to determine whether a patient is responding. They indicated that treatment response is considered clinically meaningful when tumour size is reduced by at least 20% and performance status improves, but added that treatment may be continued in some patients with radiographic progression (new and/or increasing target lesions) if clinical benefits are observed. The clinician group noted that toxicity or symptomatic disease progression without clinical benefit would be considered when the decision to discontinue tebentafusp is made. The clinician group indicated that oncologists with experience in the inpatient management of side effects related to tebentafusp are required because of the toxicities that commonly occur after the first 3 doses.

Drug Program Input

The drug programs expressed an interest in understanding treatment eligibility for patients who received prior systemic treatments in the advanced or metastatic setting, who have central nervous systemic (CNS) metastases, or who have an ECOG PS of 2 or greater, as well as identifying appropriate discontinuation criteria.

The clinical expert consulted by CADTH indicated that the use of tebentafusp in patients who have received prior therapies in the advanced or metastatic setting is appropriate, and noted that there is a high unmet need for effective systemic treatments in the first-line setting and beyond and that the use of tebentafusp in these settings aligns with clinical trial and anecdotal evidence. The clinical expert indicated that patients with CNS metastases should be eligible for treatment if CNS metastases have been controlled with radiation or surgery. The clinical expert noted that in clinical practice, patients with an ECOG PS of 3 or worse are not eligible for systemic treatment because they are unlikely to benefit from treatment.

In terms of discontinuation criteria, clinical trial evidence has shown that patients may continue to derive clinical benefits from tebentafusp despite initial radiographic progression, the clinical expert explained. In general, treatment discontinuation may be warranted in patients with clear evidence of significant progression, suggested by a decline in ECOG PS, an increase in pain, rising LDH levels, and marked progression on imaging, the clinical expert noted. However, given that the clinical considerations involved are complex, the decision to discontinue treatment should be left to the discretion of attending oncologist, the clinical expert explained.

Clinical Evidence

Pivotal Studies and Protocol-Selected Studies

Description of Study

Study 202 met the inclusion criteria for the CADTH Systematic Review.¹⁰ Study 202 was a phase III, open-label, randomized, active-controlled study that aimed to compare the efficacy and safety of tebentafusp with investigator’s choice of treatment (pembrolizumab, ipilimumab, or dacarbazine) in HLA-A*02:01-positive adults with mUM who had not received prior therapy in the metastatic setting (N = 378). Patients were randomized in a 2:1 ratio to receive IV infusions of tebentafusp (20 mcg on day 1, 30 mcg on day 8, 68 mcg on day 15, and weekly thereafter) or IV infusions of dacarbazine (1,000 mg/m²), ipilimumab (3 mg/kg), or pembrolizumab (2 mg/kg per dose [up to 200 mg] or a 200 mg flat dose) every 3 weeks. Treatments were continued until disease progression per RECIST 1.1, unacceptable toxicity, or completion of a maximum of 4 doses of ipilimumab. Patients receiving tebentafusp, pembrolizumab, or ipilimumab were permitted to continue treatment after initial disease progression if prespecified criteria were met that indicated clinical benefit and tolerance of the study drugs.

Study 202 aimed to establish the superiority of tebentafusp to investigator’s choice of therapy through the coprimary end points of OS in the rash analysis set (RAS), which consisted of patients receiving tebentafusp who developed a rash in the first week after treatment and all patients in the investigator’s-choice arm, and OS in the intention-to-treat (ITT) analysis set. The key secondary end points were progression-free survival (PFS) and best overall response (BOR) (evaluated statistically as objective response rate [ORR]). Other secondary outcomes included health-related quality of life (HRQoL) (5-Level EQ-5D [EQ-5D-5L] and European Organisation for Research and Treatment of Cancer Quality of Life Questionnaire Core 30 [EORTC QLQ-C30]), disease control rate (DCR), duration of response (DOR), and safety, all of which were measured without control for multiplicity. The first interim OS analysis and the final PFS analysis were performed with a data cut-off date of October 13, 2020. An informal updated OS analysis and the primary ORR analysis were performed with a data cut-off date of August 12, 2021, to fulfill regulatory requirements in Europe.

The baseline patient characteristics were balanced between treatment arms. Overall, the mean age of patients was 62.1 (standard deviation [SD] = 11.6) years. Approximately half of the patients were female. The majority of patients were white, had an ECOG PS of 0, and had liver metastases. Most patients had not undergone surgery for metastatic disease (91.3%). A small proportion of patients (3.7%) had received prior antineoplastic systemic treatments (in any setting), and 40.2% of patients had received prior local radiotherapy. In the investigator’s-choice arm, the majority of patients were assigned pembrolizumab (81.7%), whereas 12.7% received ipilimumab and 5.6% received dacarbazine.

Efficacy Results

The key efficacy results from Study 202 are summarized in Table 2.

Overall Survival

In the first interim OS analysis (median duration of follow-up of 14.1 months), the median OS in the RAS was 27.4 (95% confidence interval [CI], 20.2 to not reported [NR]) months in the tebentafusp arm and 16.0 (95% CI, 9.7 to 18.4) months in the investigator’s-choice arm, with a hazard ratio [HR] of 0.38 (95% CI, 0.25 to 0.56) and a P value of < 0.0001 in favour of tebentafusp. The median OS in the ITT analysis set was 21.7 (95% CI, 18.6 to 28.6) months in the tebentafusp arm and 16.0 (95% CI, 9.7 to 18.4) months in the investigator’s-choice arm, with an HR of 0.51 (95% CI, 0.37 to 0.71) and a P value of < 0.0001 in favour of tebentafusp. The results of the informal updated OS analysis were consistent with those of the first interim analysis.

Progression-Free Survival

In the final PFS analysis (median follow-up duration of 11.4 months), the median PFS in the ITT analysis set was 3.3 (95% CI, 3.0 to 5.0) months in the tebentafusp arm and 2.9 (95% CI, 2.8 to 3.0) months in the investigator’s-choice arm, with an HR of 0.73 (95% CI, 0.58 to 0.94) and a P value of 0.0139 in favour of tebentafusp.

Health-Related Quality of Life

As of the data cut-off date of October 13, 2020, the change from baseline in EORTC QLQ-C30 score was generally stable and similar between treatment arms at most time points for almost all domains. With respect to the fatigue scale (higher scores indicate worse symptoms), the difference in least square (LS) mean change in score from baseline at end of treatment between tebentafusp versus investigator’s choice was −9.259. This analysis was not adjusted for multiplicity.

The baseline mean EQ visual analogue scale (VAS) (higher scores indicate better HRQoL) was 81.0 (SD = 16.4) in the tebentafusp arm and 80.4 (SD = 18.3) in the investigator’s-choice arm. The mean change from baseline at end of treatment was –10.1 (SD = 22.53) in the tebentafusp arm and –11.7 (SD = 21.40) in the investigator’s-choice arm. The difference between arms was not tested statistically.

Duration of Response

As of the data cut-off date of August 12, 2021, 26 (10.3%) patients receiving tebentafusp and 6 (4.76%) patients receiving investigator’s choice of therapy had achieved a complete response (CR) or partial response (PR). Among these patients, the median DOR was 9.9 (95% CI, 5.6 to 22.1) months in the tebentafusp arm and 9.7 (95% CI, 2.7 to NR) months in the investigator’s-choice arm. The difference between arms was not tested statistically.

Objective Response Rate

In the primary ORR analysis (data cut-off date of August 12, 2021), the ORR in the ITT analysis set was 10.3% (95% CI, 6.9% to 14.8%) in the tebentafusp arm and 4.8% (95% CI, 1.8% to 10.1%) in the investigator’s-choice arm, corresponding to an odds ratio (OR) of 2.26 (95% CI, 0.91 to 5.61) and a P value of 0.684. Because of imprecision, the results for ORR are inconclusive.

Disease Control Rate

As of the data cut-off date of August 12, 2021, the DCR in the ITT analysis set was 45.6% (95% CI, 39.4% to 52.0%) in the tebentafusp arm and 27.0% (95% CI, 19.5% to 35.6%) in the investigator’s-choice arm, corresponding to an OR of 2.34 (95% CI, 1.45 to 3.76). This analysis was not adjusted for multiplicity.

Table 2: Summary of Key Efficacy Results From Study 202

CI = confidence interval; EQ VAS = EQ visual analogue scale; HRQoL = health-related quality of life; ITT = intention to treat; NR = not reported; OS = overall survival; PFS = progression-free survival; RAS = rash analysis set; SD = standard deviation.

^aN = 252 refers to the intention-to-treat analysis set sample size; N = 149 refers to the RAS sample size.

^bBased on Kaplan-Meier estimate.

^cHR was estimated using a Cox proportional hazards model stratified by LDH status.

^dP value based on a log-rank test of the Kaplan-Meier curve stratified by LDH status.

^eAccording to the sponsor, the OS analysis at the data cut-off date of August 12, 2021, was informal; as such, the P value should be considered exploratory.

^fBased on the patient population that had achieved a CR or PR (at the data cut-off date of October 13, 2020, there were 23 patients in the tebentafusp arm and 6 in the investigator’s-choice arm; at the data cut-off date of August 12, 2021, there were 26 patients in the tebentafusp arm and 6 in the investigator’s-choice arm).

^gDuration of response is defined as the time from the date of first documentation of partial response or better to the date of first documentation of progressive disease or death due to any cause, whichever comes first.

^hOdds ratio was calculated using Cochran-Mantel-Haenszel test stratified by LDH status.

Source: Study 202 Clinical Study Report.¹¹

Harms Results

The key harms outcomes from Study 202 are summarized in Table 3.

As of the data cut-off date of October 13, 2020, treatment-emergent adverse events (TEAE) were reported in all patients in the tebentafusp arm and in 94.6% of patients in the investigator’s-choice arm. The most common TEAEs (tebentafusp arm versus investigator’s-choice arm) were pyrexia (76.0% versus 7.2%), pruritus (69.0% versus 23.4%), rash (55.1% versus 16.2%), and fatigue (51.0% versus 35.1%). Serious TEAE were reported in 28.2% of patients in the tebentafusp arm and 23.1% of patients in the investigator’s-choice arm. The most common serious TEAE in the tebentafusp arm was CRS (9.8% versus 0.0% in the investigator’s-choice arm). The proportion of patients who discontinued treatment due to TEAEs was 3.3% in the tebentafusp arm and 6.6% in the investigator’s-choice arm. There were 84 deaths (34.3%) and 57 deaths (51.4%) reported in the tebentafusp arm and the investigator’s-choice arm, respectively. The majority of deaths in both arms was attributed to disease progression.

The proportion of patients who reported CRS (or presentations related to CRS) and dermatological was notably higher in the tebentafusp arm than in the investigator’s-choice arm. The most common notable harms of any grade (30% or greater) were pyrexia, pruritus, rash, fatigue, nausea, chills, hypotension, dry skin, headache, and maculopapular rash, all of which were mostly grade 1 or 2.

Table 3: Summary of Key Harms Results From Study 202 (SAS)

NA = not applicable; NR = not reported; SAS = safety analysis set; TEAE = treatment-emergent adverse event.

^aTEAE was defined as an AE that had an onset date on or after the date of the first dose or pretreatment TEAEs that increase in severity on or after the date of the first dose up to an including 90 days after the date of the last dose.

^bCRS was reported by the investigator based on Lee et al. (2014)¹² grading criteria.

Source: Study 202 Clinical Study Report.¹¹

Critical Appraisal

The overall design of Study 202 was appropriate for the objectives of the study. There was no particular concern with the methods of randomization or allocation concealment. The open-label design was considered reasonable by the clinical expert., There is potential for reporting bias on tumour response (ORR, DCR, BOR, DOR) and subjective harms outcomes because these outcomes were based on the investigator’s assessment, although the extent and direction of bias are unclear. The statistical analyses were generally appropriate, with proper processes in place to preserve power in the interim and final OS analyses and to account for multiplicity for the coprimary end points and key secondary end points using a hierarchal approach. DCR, DOR, and HRQoL outcomes, however, were not adjusted for multiplicity and were considered exploratory because of the increased risk of a type I error. It should be noted that the OS analyses were interim, and interim analyses are typically associated with a risk of overestimation of the treatment effects in favour of the experimental intervention (i.e., tebentafusp).¹³ Considering that the OS analyses were based on a relatively small number of events, the OS results are prone to imprecision. The OS analysis in the RAS should also be interpreted with caution because of the risk of confounding, resulting from the absence of randomization in the comparison, although the direction of bias could not be determined. There is also uncertainty in the HRQoL outcomes related to potential reporting and attrition bias and because the instruments used (EORTC QLQ-C30 and EQ-5D-5L) had not been validated in patients with mUM.

In terms of generalizability, a limitation to note is that the studies included patients who had received no prior therapies in the metastatic setting; therefore, the applicability of trial results to patients who had received prior therapies in the metastatic setting is unclear. The treatments included in the comparator arm account for a small proportion of systemic treatments prescribed for mUM in Canada, which increases uncertainty about the generalizability of study results. The clinical expert consulted by CADTH noted that the impact on generalizability is likely to be small because the efficacy of immunotherapies is considered similar by clinicians. However, this opinion was based on a small retrospective cohort study and is associated with some uncertainties. The OS benefits of tebentafusp were considered clinically meaningful by the clinical expert, who noted, however, that PFS and tumour response outcomes have limited clinical relevance because tumour response is poorly correlated with OS in patients with mUM receiving systemic therapy in general. The clinical relevance of HRQoL outcomes was also uncertain because the instruments used are not routinely administered in clinical practice, although they are able to capture some of the most common HRQoL concerns (e.g., anxiety, depression, fatigue) reported by patients. With respect to safety, specifically CRS, the clinical expert expected the study findings to be generalizable to clinical practice, provided that tebentafusp is administered in appropriate treatment settings, as specified in the product monograph, but noted that CRS generally occurs after the first 3 to 4 infusions and is manageable if proper supportive care is provided.

Indirect Comparisons

Description of Study

One indirect treatment comparison (ITC) was submitted by the sponsor and included in this review.¹⁴ No additional ITCs were identified in the literature. The sponsor performed an unanchored matching-adjusted indirect comparison (MAIC) to estimate OS and PFS with tebentafusp compared with ipilimumab plus nivolumab in patients with mUM who had received no prior therapy in the metastatic setting, based on data from the GEM-1402 comparator trial,¹⁵ which was identified in a systematic literature review (SLR), and the Study 202 index trial. Of note, the GEM-1402 trial enrolled patients with mUM regardless of HLA status and Study 202 enrolled HLA-A*02:01-positive patients with mUM. The sponsor submitted a supplementary analysis of the prognostic value of HLA-A*02:01 for OS to support the MAIC analyses.¹⁶

Efficacy Results

The ITC included 237 patients in the tebentafusp arm from Study 202, and the effective sample size (ESS) of the ipilimumab plus nivolumab arm from the GEM-1402 trial was 115.9. The MAIC analysis between tebentafusp and ipilimumab plus nivolumab showed results in favour of tebentafusp with respect to both OS (HR, 0.507; 95% CI, 0.324 to 0.793) and PFS (HR, 0.647; 95% CI, 0.445 to 0.941).

Of the 80 patients included in the supplementary HLA status analysis, 43 were HLA-A*02:01-positive and 37 were HLA-A*02:01-negative. Median OS was 45.9 months (range, NR) in HLA-A*02:01-positive patients and 45.2 months (range, NR) in HLA-A*02:01-negative patients, with an HR of 0.82 (95% CI, 0.36 to 1.88).

Harms Results

The ITC did not assess safety outcomes.

Critical Appraisal

A key limitation of the ITC was that the selection criteria for the SLR were not determined a priori, increasing the risk of selection bias for comparator trials included in the MAIC analysis. In addition, patient population heterogeneity, specifically with respect to HLA-A*02:01 status, could be a potential source of confounding, given that the comparator trial included patients regardless of HLA-A*02:01 status. The sponsor-submitted supplementary HLA analysis did not allow the CADTH review team to rule out the possible confounding effect of HLA status on OS, given that the study was based on a small observational cohort and results are subject to imprecision (a wide 95% CI for HR with respect to OS). Further, time since primary diagnosis, a covariate identified in the multivariate analysis, was excluded from the adjustment because it was not reported in the comparator trial, which may contribute to the uncertainty of the results. There is also concern about a loss of precision in the results, given the significant reduction in the ESS. Overall, the direction of bias of the limitations noted could not be determined, and the results of the MAIC should be interpreted with caution because of the potential biases. There is also an evidence gap, in that HRQoL and harms outcomes, as well as the efficacy of tebentafusp in treatment-experienced patients, were not addressed in this analysis.

Other Relevant Evidence

This section includes 2 additional studies from the sponsor’s submission to CADTH that were considered to address important gaps in the evidence included in the Systematic Review. The first was a phase I/II, multicentre, open-label study (Study 102) that analyzed the efficacy and safety of tebentafusp in HLA-A*02:01-positive patients with mUM who had received 1 or 2 prior lines of therapy in the metastatic or advanced setting.¹⁷ The second study was observational and compared patients receiving tebentafusp from Study 202 with patients receiving ipilimumab plus nivolumab from the GEM-1402 trial in a first-line mUM setting after propensity score weighting.¹⁸ The analysis was based on the same studies that informed the sponsor-submitted MAIC and was submitted to CADTH after the sponsor had obtained individual patient data (IPD) from the GEM-1402 trial. The study aimed to address limitations of the MAIC analysis related to the use of aggregate data from the GEM-1402 trial.

Study 102

The Study 102 was a phase I/II study that analyzed the efficacy and safety of tebentafusp in HLA-A*02:01-positive patients with mUM who had received 1 or 2 prior lines of therapy in the advanced or metastatic setting.¹⁷ The primary end point for the phase I part was the incidence (number) of dose-limiting toxicities, whereas the primary end point for the phase II, single-arm, dose-expansion part was ORR. The secondary end points for the phase II part were OS, PFS, DOR, DCR, and BOR.

Efficacy Results

In the phase II expansion cohort (data cut-off date of March 20, 2020), the ORR of 4.7% (95% CI, 1.8% to 10.0%) was based on 6 of the 127 patients who received tebentafusp and achieved a PR; no patients achieved a CR. After a median follow-up of 19.6 (95% CI, 16.0 to 22.2) months, median OS was 16.8 (95% CI, 12.9 to 21.3) months. Median PFS was 2.8 (95% CI, 2.0 to 3.7) months, according to RECIST 1.1 assessment by independent central review (ICR). Median DOR (CR or PR) was 8.7 (95% CI, 5.6 to 24.5) months, according to assessment of RECIST 1.1 by ICR. DCR (CR, PR, or stable disease) was 22.8% (95% CI, 15.7% to 31.2%) at 24 weeks or beyond. The most frequently observed BOR was disease progression (47.2%), followed by stable disease (44.9%) and PR (4.7%).

Harms Results

As in Study 202, all patients in Study 102 experienced AEs of any grade. In addition, the phase II expansion part of Study 102 and Study 202 had comparable rates of TEAEs of grade 3 or higher (59.1% versus 54.3%). In the phase II expansion part of Study 102, 9 patients (7.1%) experienced TEAEs leading to tebentafusp discontinuation, a proportion comparable to that of Study 202 (6.6%). The most frequently reported any-grade TEAEs were pyrexia (81.1%), pruritus (68.5%), nausea (67.7%), chills (66.1%), and hypotension (41.7%). These TEAEs were also observed in the tebentafusp arm in Study 202. In the phase II expansion part of Study 102, 86 serious TEAEs were reported in 42 (33.1%) patients, a proportion comparable to that of Study 202 (28.2%). Serious TEAEs were CRS, reported in 4 (3.1%) patients, and sepsis, alanine transaminase increase, rash maculopapular, and hypotension, reported in 3 (2.4%) patients each. No deaths related to TEAEs or the study drug were observed.

Critical Appraisal

The noncomparative design of Study 102, with no statistical testing, is the key limitation. The lack of direct comparative data means there is uncertainty regarding the magnitude of effects obtained for the efficacy outcomes. Although the clinical expert consulted noted that the efficacy outcomes of tebentafusp in Study 102 were clinically meaningful, demonstrated the activity of the drug, and were compatible with the Study 202 phase III trial, CADTH review team noted that in the absence of a comparative arm, the findings from the efficacy and safety analysis are uncertain, as the single-arm design does not allow for conclusions to be drawn about the comparative efficacy of tebentafusp or for symptoms of underlying mUM disease to be differentiated from treatment-related AEs.

Propensity Score Analysis (IPTW Approach)

The sponsor-submitted observational study compared patients in the Study 202 tebentafusp arm with patients in the GEM-1402 trial who received ipilimumab plus nivolumab in a first-line metastatic setting.¹⁸ The study was not randomized, and propensity score weighting with the inverse probability of treatment weighting (IPTW) approach was used in an attempt to adjust for confounding.

Efficacy

The ESS of the GEM-1402 trial was 34.4 after weighting, compared to a sample size of 45.0 before weighting. The patient characteristics were generally balanced between the tebentafusp and ipilimumab plus nivolumab cohorts after weighting.

In the primary analysis, the median OS of the tebentafusp cohort was 21.7 months (SD = NR), and the weighted median OS of the ipilimumab plus nivolumab cohort was 12.6 months (SD = NR). The HR between tebentafusp and ipilimumab plus nivolumab with respect to OS was 0.430 (95% CI, 0.287 to 0.643), in favour of tebentafusp.

Harms Results

The study did not assess safety outcomes.

Critical Appraisal

The IPTW approach improved on the MAIC by leveraging IPD from the GEM-1402 trial; however, many of the limitations of the MAIC analysis also apply to the current analysis. Specifically, the lack of a priori selection criteria for the SLR is a potential source of selection bias for comparator studies. It is also unclear whether all known or unknown confounding factors have been adequately adjusted for. Heterogeneity in HLA status was noted in the comparator groups and considered in the supplementary analysis submitted by the sponsor that assessed the impact of HLA status on OS, the CADTH review was unable to confidently rule out confounding effects related to the difference in HLA status between the cohorts. Last, outcomes that are of interest to stakeholders, such as HRQoL, harms, and the efficacy of tebentafusp in patients who received prior systemic therapy in the metastatic setting, were not addressed in this analysis.

Conclusions

Tebentafusp demonstrated a statistically significant and clinically meaningful improvement in OS compared with investigator’s choice of therapy (pembrolizumab, ipilimumab, dacarbazine) in HLA-A*02:01-positive adults with mUM who had received no prior therapy in the metastatic setting, based on the phase III pivotal study. Tebentafusp led to a statistically significant improvement in PFS but not in ORR, although these outcomes are of limited clinical relevance in the treatment of mUM, according to the clinical expert consulted by CADTH. Analyses of other secondary end points showed results in favour of tebentafusp (DCR) or results that were comparable between treatment arms (DOR, HRQoL), but no conclusion can be drawn because of the lack of statistical testing or control for multiplicity. The comparator treatments account for a small proportion of systemic treatments prescribed for mUM in Canada, which increases uncertainty about the generalizability of the study results. Results of a phase II, single-arm study suggested clinical activity of tebentafusp in patients who had received prior therapy in the metastatic setting, but they are subject to limitations related to the noncomparative study design. No definitive conclusions can be drawn from the sponsor-submitted MAIC or propensity score analysis about the relative efficacy of tebentafusp compared with ipilimumab plus nivolumab because of significant limitations of the analyses. CRS and dermatological AEs were common with tebentafusp but were generally grade 1 or 2, transient, and manageable with supportive care.

Introduction

Disease Background

UM is a rare subset of melanoma that accounts for 2.9% of all melanoma cases in Canada.² UM refers to melanomas that arise from the uveal tract (choroid, ciliary body, and iris) in the eye. It is the most common primary intraocular malignancy in adults, representing 95% of all ocular melanomas.¹ It is estimated that 3.75 new cases of UM developed per million people each year in Canada from 1992 to 2010, with an increase of 0.074 new cases per million individuals annually.²

UM is most commonly diagnosed in Caucasians [as cited] (97.8%), and the mean age at diagnosis is approximately 58 years.^19,20 Light eye and skin colour, skin and iris nevi (dark spots), UV light exposure, and melanocytosis (hyperpigmentation) of the eye or skin near the eye are known risk factors for UM.¹

Visual symptoms, such as flashes, floaters, and vision loss, are present in approximately 50% of patients with UM; other patients are often asymptomatic and identified during routine eye examination.¹ Most patients have localized disease at diagnosis, and about 50% of patients will eventually progress to metastatic disease,³ with metastases most commonly involving the liver (93%), lung (24%), bone (16%), and skin or subcutaneous tissues (11%).¹ The survival of patients with mUM is unfavourable; it has an estimated 1-year survival rate of 43% to 52% in the first-line treatment setting, and about 37% in the second-line setting or beyond.^4,5 Parenchymal liver failure is the leading cause of death in patients with mUM.⁶

The diagnosis of UM involves clinical evaluations of the history and physical status of the patient and an eye examination using imaging techniques such as biomicroscopy, dilated fundus exam, ultrasound, optical coherence tomography, and fluorescein angiography.^7,21 Metastatic disease is diagnosed with biopsy (if clinically appropriate), and imaging is used for baseline staging and the evaluation of specific signs and symptoms.⁷

Standards of Therapy

The primary tumour in the eye may be treated with brachytherapy, external electron beam therapy, or enucleation (removal of the eye), according to the clinical expert. The clinical expert noted that the goal is cessation of growth and slow shrinkage of the tumour, which denotes a long-term local clinical response. Brachytherapy controls the local tumour in 90% of cases; in the remaining 10%, enucleation is used as salvage therapy, according to the clinical expert. OS is identical with brachytherapy and enucleation, so the former is generally preferred, the clinical expert noted, and enucleation is reserved for large or invasive tumours that cannot be treated with brachytherapy.

Treatments for mUM can be broadly divided into local therapies and systemic therapies.¹

Local therapies, which include surgical resection, radiofrequency ablation, stereotactic radiation, and infusion of radioactive beads or chemotherapy into the liver through the hepatic arteries, are generally used in patients with limited or oligometastatic disease with small number of lesions in the liver, according to the clinical expert consulted by CADTH.

Systemic therapies are usually prescribed for patients with a larger number of metastatic lesions and/or disease external to the liver. Immunotherapies (off-label) are commonly prescribed systemic treatments for mUM, either as monotherapies (i.e., nivolumab, or pembrolizumab alone) or in combination (i.e., ipilimumab plus nivolumab) in the first-line setting. According to the clinical expert, ipilimumab monotherapy is generally used in later-line settings in Canada and systemic chemotherapies are rarely prescribed because of the poor response rate and poor toxicity tolerance. The clinical expert noted that UM tumours typically have a low ORR (approximately 10% to 20%) to immunotherapies because of their low tumour mutation burden; as such, the standard of care for mUM is enrolment in a clinical trial when available and clinically appropriate.^7,8 Funding for systemic therapies for mUM is usually assessed on a case-by-case basis by the jurisdictions.

In patients with mUM and multiple metastases restricted to the liver, hepatic artery infusion may be considered. Therapeutic drugs include chemotherapy or radiolabelled microbeads.²²

Drug

Tebentafusp is a bispecific T-cell engager comprising 2 domains: a T-cell-receptor-targeting domain that binds to gp100 peptides presented by HLA-A*02:01 on UM tumour cell surface; and an anticluster of differentiation 3 (CD3) effector domain that binds and activates T-cells.²³ Tebentafusp forms an immune synapse between UM cells and T-cells, thereby redirecting and activating T-cells to release inflammatory cytokines and cytolytic proteins that cause lysis of UM tumour cells.²³ HLA-A*02:01 expression is required in patients receiving tebentafusp, as the T-cell-receptor-targeting domain is specific to the HLA-A*02:01-gp100_{280 to 288} complex.²³

This is the first CADTH review for tebentafusp. Tebentafusp was granted a Health Canada Notice of Compliance for the indication of treatment of HLA-A*02:01-positive adults with unresectable UM or mUM on June 7, 2022. A positive HLA-A*02:01 genotype status is required for patients to be eligible for treatment with tebentafusp. The recommended dose of tebentafusp, per the Health Canada product monograph, is 20 mcg on day 1, 30 mcg on day 8, 68 mcg on day 15, and 68 mcg once every week thereafter.⁹ Patients should be treated until unacceptable toxicity or disease progression occur.⁹ The first 3 infusions of tebentafusp should be administered in an appropriate health care setting by IV infusion over 15 to 20 minutes.⁹ Patients should be monitored during the infusion and for at least 16 hours after the infusion is complete.⁹

The sponsor is seeking reimbursement of tebentafusp per the indication.

The key characteristics of tebentafusp and the comparator drugs are summarized in Table 4.

Table 4: Key Characteristics of Tebentafusp, Pembrolizumab, Nivolumab, and Ipilimumab

AE = adverse event; CTLA-4 = cytotoxic T-lymphocyte-associated antigen 4; ImmTAC = immune mobilizing monoclonal T-cell receptors against cancer; mAb = monoclonal antibody; mUM = metastatic uveal melanoma; PD-1 = programmed cell death 1 protein; UM = uveal melanoma.

^aOff-label treatment. The Health Canada indication is for the treatment of unresectable or metastatic melanoma. Patients with UM were excluded from the pivotal trials that led to approval of the indication.

^bHealth Canada–approved indication.

^cIpilimumab monotherapy is used in second- or later-line settings in the treatment of mUM in Canada.

^dOff-label dosing. Dosing shown is approved for the treatment of melanoma (cutaneous).

Sources: Tebentafusp product monograph,⁹ Keytruda product monograph,²⁴ Opdivo product monograph,²⁵ Yervoy product monograph.²⁶

Stakeholder Perspectives

Patient Group Input

This section was prepared by CADTH staff based on the input provided by patient groups. The original patient group submissions can be found on CADTH’s website.

Two patient groups provided input for this review: MC and SYSF. MC (formerly the Melanoma Network of Canada) is an organization based in Canada that advocates on behalf of patients with melanoma, coordinates educational and prevention strategies, and assists in funding for melanoma research. SYSF is a national patient-led not-for-profit group committed to the fight against nonmelanoma skin cancers, melanoma, and ocular melanoma through nationwide education, advocacy, and awareness initiatives.

MC gathered information from an online survey. A total of 19 patients with UM and 15 caregivers responded to the survey. There were 9 patient respondents from Ontario, 2 from Nova Scotia, 2 from British Columbia, 3 from Alberta, 1 from Prince Edward Island, 1 from Newfoundland, and 1 from outside Canada. The majority of respondents had early-stage disease or did not know their staging. SYSF obtained information and feedback from patient interviews, virtual patient roundtables, and online surveys. Input was gathered from 38 patients with ocular melanoma, the majority of whom were diagnosed with primary ocular melanoma (28 patients). There were 9 respondents from British Columbia, 9 from Ontario, 8 from Alberta, 3 from Saskatchewan, 2 from Newfoundland, 1 from Manitoba, 1 from Quebec, 2 from outside Canada, and 3 did not identify their province of residence.

Respondents from MC mentioned that the diagnosis of UM affected their day-to-day life and QoL. Loss of vision or vision impairment (77.78%) was reported to be the major issue affecting their daily life and QoL, followed by fear or anxiety (61.11%), depression (38.89%) and fatigue (27.78%). In addition, respondents from SYSF mentioned that their balance was affected, which caused serious physical and psychological deterioration. The respondents also indicated that they had very few symptoms leading up to their diagnosis and were concerned about the risk of metastasis.

Respondents from MC and SYSF indicated that they had experience with a variety of treatments, such as surgery, immunotherapies, radiation, and brachytherapy. According to the MC respondents, the frequently reported issues with current treatments were loss of vision or vision impairment (82.35%), fatigue (35.29%), disrupted sleep (29.41%), post-traumatic stress (23.53%), negative impact to family or social life (23.53%), pain (17.65%), and headaches (17.65%).

There were 5 patients from MC and 10 patients from SYSF who indicated that they had experience with tebentafusp through clinical trials or a Health Canada program for compassionate access. According to the MC and SYSF submissions, 3 patients (2 from MC and 1 from SYSF) who had experience with tebentafusp indicated that the drug had shown effectiveness in slowing disease progression, and another 2 patients could not comment on the effectiveness as it was too early for them to tell. Patients from both groups reported a variety of side effects: skin rash, fever, fatigue, cognitive impairment, stomach-related issues, nausea, muscle and joint pain, and headaches. Most tebentafusp-experienced patients described the side effects as short-term, tolerable, and manageable, and reported that the benefits of treatments outweighed the negative side effects. Only 1 patient from SYSF reported that the side effects were not manageable. In addition, patients from both groups reported that they had to travel long distances to get access to the drug and had financial difficulties.

Patients from both groups expressed their desire for an effective treatment that can preserve vision and provide a good QoL and longer survival. In addition, patients from MC indicated that they would like to see improvements in earlier diagnosis and detection of metastasis.

Clinician Input

Input From Clinical Expert Consulted by CADTH

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

Unmet Needs

The clinical expert consulted by CADTH noted that mUM is an aggressive disease associated with poor survival and that available systemic treatments for mUM have not demonstrated improvement in OS, PFS, or ORR. Therefore, the clinical expert noted that the development of novel therapies for mUM is a priority. The clinical expert also noted that UM tumours in general have a low ORR to immune checkpoint inhibitors (approximately 10% to 20%) because of the low tumour mutation burden, although the ORR is slightly higher in atypical UM. This predicts a poor response to immunotherapy, per the clinical expert. Chemotherapies are associated with a low ORR (about 5% to 15%) and considerable toxicity, according to the clinical expert.

Place in Therapy

The clinical expert indicated that the mechanism of action of tebentafusp is unique among current therapies for UM, in that the drug exploits the biology of the UM cell. The clinical expert explained that patients must have an HLA-A*02:01-positive status for tebentafusp to be effective, owing to the unique mechanism of action, which leaves only 45% of all patients with mUM (mainly of European origin) eligible for tebentafusp.

The clinical expert considered tebentafusp to be a major advance in the therapy of UM, and noted that although it is not curative, it will likely fulfill the unmet need for an effective treatment that improves survival. A shift in the current treatment paradigm is anticipated by the clinical expert. The clinical expert noted that in melanoma clinics where tebentafusp can be accessed through compassionate drug programs, tebentafusp is currently the preferred first-line systemic treatment for mUM. The clinical expert stated that tebentafusp should be used as a first-line treatment, if possible, but that its use in the second-line setting and beyond is also rational, particularly in patients for whom access to compassionate supplies of tebentafusp was not possible in the past and who therefore had to be treated with other therapies.

Patient Population

The clinical expert indicated that only HLA-A*02:01-positive patients with mUM are expected to benefit from tebentafusp. The HLA subtype can be determined with a routine HLA genotype testing, which is available in larger treatment centres.

The clinical expert indicated that there is currently no good clinical or biologic predictor of response to tebentafusp. The clinical expert noted that patients with low-volume disease tend to respond better to tebentafusp than patients with large metastases, although this was based on anecdotal evidence. In addition, patients with a better performance status are also more likely to benefit from treatment in general in clinical practice, according to the clinical expert.

The clinical expert indicated that patients with a poor performance status are least suitable for treatment with tebentafusp, noting that patients with an ECOG PS of 3 or above are generally not eligible for treatment in clinical practice.

Assessing Response to Treatment

According to the clinical expert, OS is considered the most important and clinically meaningful outcome in patients with mUM, whereas PFS and ORR have limited relevance, as survival does not corelate with ORR in patients with mUM receiving tebentafusp. The clinical expert noted that a posthoc analysis of Study 202 showed that patients with progression on tebentafusp could still benefit from treatment beyond progression.¹⁰ The clinical expert, therefore, concluded that patients should be allowed to continue treatment as long as they derive clinical benefits from tebentafusp, and treatment response should be left to the discretion of the attending oncologist based on assessments of history and physical examinations (every 3 to 4 weeks), laboratory tests (weekly), and imaging (every 12 to 16 weeks) findings.

Discontinuing Treatment

The clinical expert highlighted that the decision to discontinue treatment should be left to the discretion of the attending oncologist, given the complexity of the considerations involved. In clinical practice, treatment discontinuation is generally considered in the presence of intolerable toxicities or when there is clear evidence of significant progression, as indicated by a decline in performance status, an increase in pain, rising LDH levels, and marked radiographic progression, according to the clinical expert.

Prescribing Conditions

The clinical expert emphasized that the first 3 or 4 infusions of tebentafusp should be given in hospital because of the risk of CRS, which usually presents with rash, fever or chills, hypotension, and, in rare cases, compromise of pulmonary, renal, or hepatic function. After cycle 4, the risk of CRS is substantially lower, and most patients have no side effects from tebentafusp thereafter.

The clinical expert recommended that initially tebentafusp should only be prescribed by a specialist physician experienced in the use of tebentafusp and familiar with the management of CRS. Once a pattern of use is established and the risk of CRS is absent, subsequent infusions can be performed in a community clinic setting.

Clinician Group Input

This section was prepared by CADTH staff based on the input provided by clinician groups. The original clinician group submissions can be found on CADTH’s website.

CADTH received input from 1 clinician group, Ontario Health (Cancer Care Ontario) Skin Cancer Drug Advisory Committee, based on responses from 6 clinicians. The Skin Cancer Drug Advisory Committee provides timely evidence-based clinical and health-system guidance on drug-related issues in support of Cancer Care Ontario’s mandate, including the Provincial Drug Reimbursement Programs and the Systemic Treatment Program. No major views in contrast to those provided by the clinical expert consulted by CADTH for this review were presented.

The clinician group stated that UM is a disease that has a low tumour mutational burden and a low 1-year survival rate (50%) in the metastatic setting, which is distinct from cutaneous melanoma. None of the current systemic treatments for UM have been proven to have an OS benefit, which is a major unmet need in patients with mUM. The clinician group noted that tebentafusp has demonstrated an improvement in OS in clinical trials. The clinician group noted that the treatment goal is to improve OS and QoL. The clinician group explained that tebentafusp would be the first-line treatment of choice for HLA-A*02:01-positive patients with mUM, although did not comment on the use of tebentafusp in second-line settings or beyond. The clinician group stated that HLA-A*02:01-positive patients with mUM who do not meet the exclusion criteria of Study 202¹⁰ would be suitable for tebentafusp. They noted that ongoing performance status, tumour size, and RECIST progression requirements are the clinical outcomes used to determine whether a patient is responding to treatment. They indicated that treatment response is considered clinically meaningful in the presence of a reduction in tumour size of at least 20% and improved performance status, and noted that treatment can be continued in some patients with radiographic progression (new and/or increasing target lesions) if clinical benefits are observed. Toxicity and symptomatic disease progression without clinical benefit would be considered when deciding whether to discontinue tebentafusp, according to the clinician group. They indicated that oncologists with experience in the inpatient management of side effects related to tebentafusp are required because of the known toxicities that commonly occur after the first 3 doses.

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 expert consulted by CADTH are summarized in Table 5.

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

CNS = central nervous system; CRS = cytokine release syndrome; ECOG PS = Eastern Cooperative Oncology Group Performance Status; HLA = human leukocyte antigen; LDH = lactate dehydrogenase; mUM = metastatic uveal melanoma; OS = overall survival; pERC = CADTH pan-Canadian Oncology Drug Review Expert Review Committee; RECIST = Response Evaluation Criteria in Solid Tumours; UM = uveal melanoma.

Clinical Evidence

The clinical evidence included in the review of tebentafusp 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 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 perform a systematic review of the beneficial and harmful effects of tebentafusp (100 mcg/0.5 mL) IV infusion for the treatment of HLA-A*02:01-positive adults with unresectable UM or mUM.

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 6. Outcomes included in the CADTH review protocol reflect outcomes considered to be important to patients, clinicians, and drug plans.

Table 6: Inclusion Criteria for the Systematic Review

AE = adverse event; CNS = central nervous system; HLA = human leukocyte antigen; HRQoL = health-related quality of life; RCT = randomized controlled trial; SAE = serious adverse event; UM = uveal melanoma; WDAE = withdrawal due to adverse event.

^aOff-label treatment. The Health Canada indication is for the treatment of unresectable or metastatic melanoma. Patients with UM were excluded from the pivotal trials that led to approval of the indication.

^bIpilimumab monotherapy is used in a second-line setting or beyond for mUM in Canada.

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

Published literature was identified by searching the following bibliographic databases: MEDLINE All (1946–) via Ovid and Embase (1974–) via Ovid. All Ovid searches were run simultaneously as a multifile search. Duplicates were removed using Ovid deduplication for multifile 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 concept was tebentafusp. The following clinical trials registries were searched: the US National Institutes of Health’s clinicaltrials.gov, 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. Refer to Appendix 1 for the detailed search strategies.

The initial search was completed on May 18, 2022. Regular alerts updated the search until the meeting of the CADTH pan-Canadian Oncology Drug Review Expert Committee (pERC) on November 9, 2022.

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

In addition, the manufacturer of the drug was contacted for information regarding unpublished studies.

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

Findings From the Literature

One study was identified from the literature for inclusion in the Systematic Review (Figure 1).^10,11,30 The included study is summarized in Table 7. A list of excluded studies is presented in Appendix 2.

Figure 1: Flow Diagram for Inclusion and Exclusion of Studies

Table 7: Details of Included Study

AE = adverse event; CNS = central nervous system; ECOG PS = Eastern Cooperative Oncology Group Performance Status; EORTC QLQ-C30 = European Organisation for Research and Treatment of Cancer Quality of Life Questionnaire Core 30; HLA = human leukocyte antigen; mUM = metastatic uveal melanoma; NR = not reported; SAE = serious adverse event.

Note: One additional report from the FDA was included.³⁰

^aHistologically or cytologically confirmed mUV.

^bPrior surgical resection of oligometastatic disease was allowed. Prior neoadjuvant or adjuvant therapy was allowed, provided it was administered in the curative setting in patients with localized disease. Patients must not have been re-treated with an investigator’s-choice therapy that was administered as adjuvant or neoadjuvant treatment. Additionally, patients who received nivolumab as prior adjuvant or neoadjuvant treatment should not have received pembrolizumab as investigator’s choice of therapy.

^cPatients with brain metastases are eligible if lesions have been treated with localized therapy and there is no evidence of progression for at least 4 weeks before the study.

Source: Study 202 Clinical Study Report.¹¹

Description of Studies

One study (Study 202) met the inclusion criteria for the CADTH Systematic Review. Study 202 was a phase III, open-label, randomized, active-controlled study that aimed to compare the efficacy and safety of tebentafusp with investigator’s choice (pembrolizumab, ipilimumab, or dacarbazine) in HLA-A*02:01-positive adults with mUM who had not received prior therapy in the metastatic setting (N = 378). The study was initiated on October 4, 2017, and is currently ongoing at 58 cites (2 in Canada) in 14 countries.

Study 202 consisted of 4 phases: prescreening, screening, treatment, and follow-up. Patients who were identified as HLA-A*02:01-positive in the prescreening phase entered a screening phase of up to 28 days to assess study eligibility. Upon entering the treatment phase, eligible patients were randomized to either tebentafusp or investigator’s choice (pembrolizumab, ipilimumab, or dacarbazine) in a 2:1 ratio. Randomization was stratified by baseline LDH level (higher than the upper limit of normal [ULN] range, defined as 250 units/L, or lower than or equal to the ULN) using the Interactive Response Technology system. Patients received study treatments until completion of 4 planned doses (ipilimumab only), disease progression, or intolerable toxicity, and they were subsequently monitored in the follow-up phase for safety, survival, and disease progression (in patients who discontinued treatment for reasons other than progressive disease).

This report presents the primary analysis of Study 202 as of the data cut-off date of October 13, 2020, which corresponded to the first interim analysis of OS and the final analysis of PFS. An informal updated analysis of OS and the primary analysis of ORR (data cut-off date of August 12, 2021), which were conducted to fulfill regulatory requirements in Europe, are also presented.

Populations

Inclusion and Exclusion Criteria

The key inclusion and exclusion criteria of Study 202 are summarized in Table 7. The study enrolled HLA-A*02:01-positive adults, aged 18 years and older, with mUM, an ECOG PS of 0 or 1, and no prior systemic therapy (chemotherapy, immunotherapy, or targeted therapy) or prior liver-directed therapy (chemotherapy, radiotherapy, or embolization) in the metastatic or advanced setting. Patients with prior surgical resection of oligometastatic disease were allowed. Patients who received prior neoadjuvant or adjuvant therapy for localized disease were allowed if the therapy was administered in the curative setting. Patients who had symptomatic or untreated CNS metastases or who were receiving systemic immunosuppressive medications were excluded.

Baseline Characteristics

Baseline patient demographics, disease characteristics, and prior cancer treatments were, overall, balanced between treatment arms, as shown in Table 8. Approximately half of the patients were female and most were white. Overall, the mean age of the study population was 62.1 (SD = 11.6) years. The majority of patients had an ECOG PS of 0 (73.3%) and liver metastasis, with melanoma initially located in the choroid (75.7%) and no metastasis at diagnosis (92.3%). One-third of patients had an elevated baseline LDH. Most patients had received no prior surgery for metastatic disease (91.3%). A small proportion of patients (3.7%) had received prior antineoplastic systemic treatments (in any setting) and 40.2% of patients had received prior local radiotherapy (in any setting).

Table 8: Summary of Baseline Characteristics (ITT Population)

ECOG PS = Eastern Cooperative Oncology Group Performance Status; ITT = intention to treat; LDH = lactate dehydrogenase; SD = standard deviation; ULN = upper limit of normal.

Note: Data cut-off date of October 13, 2020.

^aSunitinib or sunitinib malate.

Source: Study 202 Clinical Study Report.¹¹

Interventions

Patients in the tebentafusp arm received 20 mcg on day 1, 30 mcg on day 8, and 68 mcg on day 15 and weekly thereafter by IV infusion over 15 minutes.

Patients in the investigator’s-choice arm received 1 of the following 3 treatments selected by the investigator(s) before randomization:

• dacarbazine 1,000 mg/m² by IV infusion on day 1 of each 21-day cycle

• ipilimumab 3 mg/kg by IV infusion over 90 minutes on day 1 of each 21-day cycle for a maximum of 4 doses

• pembrolizumab at either 2 mg/kg per dose (up to 200 mg) or a flat dose of 200 mg (where approved locally) by IV infusion over 30 minutes on day 1 of each 21-day cycle.

No blinding of interventions occurred. All treatments were continued until radiographic progression per RECIST 1.1 (except for patients receiving ipilimumab, for which a maximum of 4 doses were given), or unacceptable toxicity. Exceptions were applied to tebentafusp, pembrolizumab, and ipilimumab, which could be continued after initial disease progression when all of the following prespecified criteria were met:

• no signs or symptoms indicating clinically significant progressive disease

• no decline in ECOG PS

• no impending threat to vital organs or critical anatomic sites requiring urgent alternative medical intervention or no cases in which continuation of the study therapy would have prevented the institution of such an intervention

• the absence of any of the study treatment discontinuation criteria (i.e., completion of 4 doses of ipilimumab; initiation of alternative anticancer therapy; unacceptable toxicity; patient met 1 or more of the exclusion criteria or failed to meet all of the inclusion criteria and continuation of the study treatment would pose a safety risk; consent withdrawal; or pregnancy).

In patients who received treatment beyond initial progression, treatments were discontinued in the presence of further progression, which is defined as any 1 of the following:

• an additional increase in tumour burden of at least 20% (sum of the diameters of both target and new measurable lesions) accompanied by an absolute increase of at least 5 mm

• unequivocal progression of nontarget lesions

• new nonmeasurable lesions.

In the tebentafusp arm, inpatient safety monitoring was required for at least 16 hours after each of the first 3 doses; outpatient administration was permitted for subsequent doses unless the patient experienced hypotension of grade 2 or higher after the third dose. In the investigator’s-choice arm, no extended postinfusion monitoring was required.

Patients who experienced an infusion reaction to prior doses were given premedication to manage potential infusion reactions (e.g., acetaminophen, antihistamine). Concomitant medications and therapies that were necessary for the supportive care and safety of such patients were generally allowed (e.g., antidiarrheals, antiemetics, electrolyte supplementation), and palliative radiotherapy, hematopoietic colony-stimulating growth factors, anticoagulants, antihypertensives, bisphosphonates, and denosumab were allowed at the discretion of the investigator(s).

Anticancer therapy (other than study treatments) and most monoclonal antibody therapies were prohibited from the study. Immunosuppressive medications were prohibited, with the exception of systemic corticosteroids, which were permitted for the management of infusion reactions, immune-mediated toxicities, and other toxicities (e.g., hypotension, CRS).

Outcomes

A list of efficacy end points identified in the CADTH review protocol that were assessed in the clinical trials included in this review is provided in Table 9. These end points are subsequently summarized. A detailed discussion and critical appraisal of the outcome measures is provided in Appendix 4.

Table 9: Summary of Outcomes of Interest Identified in the CADTH Review Protocol

AE = adverse event; HRQoL = health-related quality of life; SAE = serious adverse event; WDAE = withdrawal due to adverse event.

^aThe sponsor considered the definition of best overall response and objective response rate to be synonymous. Objective response rate was the end point that was evaluated statistically.

The coprimary end points of Study 202 were OS in the ITT population and OS in the RAS (i.e., patients who developed rash in the week after tebentafusp treatment). OS was defined as the time between the date of randomization and the date of death from any cause in an individual patient. OS was followed continuously while the patients received study treatment and every 3 months during the follow-up phase. The inclusion of the RAS in the OS analysis was informed by phase II trial data that suggested rash (an on-target but off-tumour AE) might be associated with a clinical benefit,¹⁷ per the sponsor.

The key secondary end points of Study 202 were PFS and BOR (evaluated statistically as ORR). Other secondary efficacy end points included HRQoL measured with EORTC QLQ-C-30 and EQ-5D-5L scores, DOR, and DCR. Tumour response end points were based on investigator assessments, according to RECIST 1.1, conducted every 3 months during the treatment phase and the follow-up phase. HRQoL end points were based on assessments conducted every other cycle until cycle 5 and then every fourth cycle during the treatment period, and conducted every 3 months during the follow-up period. Secondary end points were defined as follows:

• PFS was defined as the time from randomization to the date of first documented progression (based on RECIST 1.1) or death from any cause, whichever occurred first.

• HRQoL was measured with the EORTC QLQ-C-30 and EQ-5D-5L instruments. Detailed information on the scoring and validity of the EORTC QLQ-C30 and EQ-5D-5L instruments is provided in Appendix 4.

  • EORTC QLQ-C30 is a patient-reported outcome instrument designed to assess the HRQoL of cancer patients by the evaluation of functional, symptom, global health status/QoL domains.³¹ Outcome scores are computed using a linear transformation of the raw score such that scores range from 0 to 100. A higher score represents a higher (better) level of functioning and a higher (worse) level of symptoms. A higher score for global health status/QoL represents a better level of global health status. The validity, reliability, responsiveness, and minimally important difference (MID) estimate of the EORTC QLQ-C30 instrument have not been evaluated in patients with mUM. An MID estimate of 10 points was identified in 2 studies of various cancer types^32,33 and was a commonly used threshold for the assessment of a clinically meaningful change in HRQoL, although other MIDs have been reported by some studies^33,34 (refer to Table 40 in Appendix 4 for more details).

  • EQ-5D-5L is a generic patient-reported outcome instrument that measures overall health status in patients by generating an index-based summary score on 5 health dimensions (mobility, self-care, daily activities, pain, and anxiety) based upon societal preference weights.³⁵ The EQ-5D-5L also uses a vertical VAS for patients to rate their health, with 0 indicating the best health you can imagine and 100 indicating the worst health you can imagine. The validity, reliability, responsiveness, and MID estimates of the EQ-5D-5L instrument have not been evaluated in patients with mUM. In a study of patients with various cancer types, MID estimates of 7 to 12 have been reported for EQ VAS.³⁶

• DOR was calculated only for patients who achieved a CR or PR and was defined as the time from the date of first documented response (based on RECIST 1.1) to the date of first progression or death.

• ORR was defined as the number of randomized patients with at least 1 visit response of CR or PR (based on RECIST 1.1) divided by the number of randomized patients and presented as a percentage for each treatment arm in the ITT population.

• DCR was defined as the proportion of patients with a BOR of CR or PR or stable disease (based on RECIST 1.1) at 24 weeks or more after randomization to the study drug and before any progressive disease event.

• BOR was defined as the best response (CR, PR, stable disease, progressive disease, or not evaluable, based on RECIST 1.1) from randomization until progression, death, or the start of new anticancer therapy.

Harms

The harms outcomes were measured with TEAEs, death, safety laboratory parameters, vital signs, electrocardiogram results, physical examination, and extent of exposure. The outcomes were based on safety assessments conducted once every cycle in the investigator’s-choice arm, and 3 times every cycle in the tebentafusp arm (except in cycle 1, where 6 assessments were conducted) during the treatment phase and safety follow-up phase.

Statistical Analysis

A summary of the statistical analyses of efficacy outcomes in Study 202 is shown in Table 11.

Sample Size and Power Calculation

The sample-size calculation was based on the primary end point of OS in the ITT population. Assuming 33 months of nonuniform recruitment and a 10% dropout rate, the study required a sample size of 369 patients in the ITT population, to be randomized in a 2:1 ratio to the tebentafusp arm and the investigator’s-choice arm, to achieve at least 250 death events to detect an HR of 0.645 with respect to OS at a 2-sided alpha of 0.045 with a power of 89% using a log-rank test.

In the RAS, 164 death events were needed to detect an HR of 0.531 with respect to OS at a 2-sided alpha of 0.005 with a power of 89% using a log-rank test, assuming 50% of patients in the tebentafusp arm developed a rash in the first week after treatment.

Interim and Final Analyses

A group sequential approach was used to account for 2 formal interim analyses and 1 final analysis. The first and second interim analyses were planned to occur after 150 and 200 death events, respectively, in the ITT population; and the final analysis will be conducted after 250 death events have occurred in the ITT population. The interim analyses were reviewed by an independent data monitoring committee. The primary end point of OS was tested at an overall alpha of 0.05 (2-sided), with 10% allocated to the RAS analysis (i.e., alpha_RAS = 0.005) and 90% allocated to the ITT analysis (i.e., alpha_ITT = 0.045), based on the O’Brien-Fleming boundary as implemented using the Lan-DeMets approach. The stopping boundaries for OS are outlined in Table 10. If the stopping boundaries were crossed at the first interim analysis, then the alpha in RAS was carried over to the ITT analysis, resulting in an alpha_ITT of 0.05. PFS and BOR were not tested on an interim basis. A final PFS analysis was planned to occur after 274 progression events, and a BOR (evaluated statistically as ORR) analysis was planned to occur after all randomized patients had been followed for approximately 9 months (i.e., 3 planned assessments).

Table 10: Stopping Boundaries for OS in the Interim and Final Analyses

ITT = intention to treat; OS = overall survival; RAS = rash analysis set.

Source: Study 202 Statistical Analysis Plan.³⁷

Multiplicity Adjustment

Hypothesis testing was formally performed on the coprimary end points of OS in the RAS analysis and OS in the ITT analysis, and the key secondary end points of PFS and BOR (evaluated statistically as ORR) in the ITT analysis. The end points were tested in a sequential order using the graphical approach, as outlined in Figure 2, to control for multiplicity. In Figure 2, the arrows specify the alpha transfer path. Once the null hypothesis for an end point was rejected (i.e., deemed statistically significant) at its assigned significance level, its significance level was transferred to subsequent end point(s), following the arrow. The fraction indicated at each node denotes the fraction of alpha allocated to the end point, and the number on the arrows denote the weights for transferring significance level. OS in the RAS was tested first, and if the test was statistically significant, the significance was transferred to test OS in the ITT analysis set. If OS in both RAS and ITT analyses were statistically significant, the significance was transferred to test PFS in the ITT analysis set, and so on.

Figure 2: Multiplicity Strategy Involving Sequential Testing of OS, PFS, and BOR

BOR = best overall response; H1 = null hypothesis for OS in the RAS; H2 = null hypothesis for OS in the ITT analysis set; H3 = null hypothesis for PFS in ITT; H4 = null hypothesis for BOR (evaluated statistically as ORR) in ITT; ITT = intention to treat; ORR = objective response rate; PFS = progression-free survival; RAS = rash analysis set.

Notes: The fraction (or whole number) at the bottom of each node represents the allocated weight for alpha in the testing sequence.

This figure was adapted in consultation with the sponsor to correct errors identified in the allocation fractions.

Source: Study 202 Statistical Analysis Plan.³⁷

Statistical Analysis for Coprimary Efficacy End Points

In the primary analyses of the coprimary end points of OS in the RAS and ITT set, the difference in OS between the tebentafusp arm and the investigator’s-choice arm were tested using a 2-sided log-rank test stratified by LDH level (LDH > 250 units/L versus LDH ≤ 250 units/L). The HR and the 95% CI were based on a Cox proportional hazards model stratified by LDH level using the Efron approach for handling ties. OS was analyzed using the Kaplan-Meier (KM) method. The median follow-up time for OS and the 95% CI were presented based on the reverse KM method. OS rate every 6 months was derived from KM estimates.

A sensitivity analysis of OS, which aimed to examine the impact of misstratification, was planned based on strata determined by the central LDH results if discrepancies of more than 10% were observed in LDH results compared to strata recorded at the time of randomization in the Interactive Response Technology used in the primary analysis.

Exploratory subgroup analyses were conducted with respect to OS. The ECOG PS subgroup (ECOG PS of 0 versus 1) was relevant to this review. The HR and 95% CI between treatment arms were estimated using the Cox proportional hazards model using Efron’s approach for handling ties, adjusted for treatment, subgroup, and treatment-by-subgroup interaction term. No adjustment for type I error was involved in the subgroup analysis.

No imputation was performed to account for missing OS data in general (except for partial dates).

Statistical Analysis for Secondary Efficacy End Points

The secondary efficacy analyses were performed in the ITT population and were based on tumour response per RECIST 1.1 as assessed by the investigator. PFS was analyzed using the same methodology as was used in the primary analysis of OS. For ORR and DCR, the difference between treatment arms was analyzed using the Cochran-Mantel-Haenszel test stratified by LDH level, and OR, along with 95% CI, were reported. Descriptive data were provided for DOR based on KM estimates. The median duration of follow-up for DOR and the 95% CI were presented based on the reverse KM method.

Descriptive statistics were used to report EQ-5D-5L and EORTC QLQ-C30 outcomes. For each of the EORTC QLQ-C30 domain, the LS mean change from baseline in scale score was estimated in each treatment arm using the mixed model repeated measures approach with treatment group (binary), time point (categorical), LDH status (binary), and baseline score (continuous) as covariates, and the difference between treatment groups was reported.

Two sensitivity analyses for PFS were conducted to assess:

• evaluation-time bias (when scans are not performed as scheduled in the protocol), in which the midpoint between the time of progression and the previous evaluable tumour assessment were analyzed using a stratified log-rank test stratified by LDH level

• attrition bias, in which the PFS analysis was repeated with actual PFS event times (as opposed to censored times) for patients who progressed or died in the absence of progression immediately after 2 or more nonevaluable or missed tumour assessments.

Exploratory subgroup analyses were conducted with respect to PFS using the same methodology as was used in subgroup analyses of OS, outlined previously. The ECOG PS subgroup (ECOG PS of 0 versus 1) was relevant to this review.

No imputation was performed to account for missing data in any secondary efficacy end points in general (except for partial dates).

Safety Analyses

The safety analyses were performed on the safety analysis set and were summarized using descriptive statistics.

Table 11: Statistical Analysis of Efficacy End Points

CI = confidence interval; CMH = Cochran-Mantel-Haenszel; EORTC QLQ-C30 = European Organisation for Research and Treatment of Cancer Quality of Life Questionnaire Core 30; EQ-5D-5L = 5-Level EQ-5L; IRT = Interactive Response Technology; ITT = intention to treat; KM = Kaplan-Meier; LDH = lactate dehydrogenase; LS = least squares; MMRM = mixed model repeated measures; NA = not applicable; RAS = rash analysis set.

Source: Study 202 Statistical Analysis Plan.³⁷

Analysis Populations

• The ITT analysis set included all patients randomized in the study and was analyzed by treatment assignment. The ITT set was used for all demographic characteristics, baseline characteristics, and efficacy analyses.

• The RAS included all patients randomized to tebentafusp who developed a rash in the first week after treatment and all patients randomized to the investigator’s-choice arm. It was prespecified that the results (demographic characteristics, baseline characteristics, efficacy, and safety) of the RAS analysis would only be reported if the RAS analysis crossed the stopping boundaries for OS but the ITT analysis did not.

• The safety analysis set included all patients who received at least 1 full or partial dose of tebentafusp or investigator’s choice of therapy and was analyzed by the initial treatment received. The safety set was used for safety analyses of the study.

• The pharmacokinetic analysis set was not reviewed in this report because it was of limited relevance.

Results

Patient Disposition

A summary of patient disposition is shown in Table 12. Of the 447 patients screened, 378 were randomized (252 to the tebentafusp arm and 126 to the investigator’s-choice arm). In the investigator’s-choice arm, the number of patients who received pembrolizumab, ipilimumab, and dacarbazine were 103 (81.7%), 16 (12.7%), and 7 (5.6%), respectively. As of the data cut-off date, the proportion of patients who discontinued the study was numerically higher in the investigator’s-choice arm (54.8%) than in the tebentafusp arm (38.1%), mostly due to death (50.0% versus 34.5%). The proportion of patients who discontinued the study drug was numerically higher in the investigator’s-choice arm (79.4%) than in the tebentafusp arm (68.3%), with the primary reason being disease progression (61.9% versus 61.1%).

Table 12: Patient Disposition

ITT = intention to treat.

Note: Data cut-off date of October 13, 2020.

^aThe number of patients receiving pembrolizumab (n = 103), ipilimumab (n = 16), or dacarbazine (n = 7).

^bTen patients received a maximum of 4 doses of ipilimumab, per protocol.

Source: Study 202 Clinical Study Report.¹¹

Protocol Deviation

A numerically higher proportion of patients reported at least 1 type of important protocol deviation in the tebentafusp arm (39.3%) than in the investigator’s-choice arm (28.6%). The most common protocol deviation in the tebentafusp arm and investigator’s-choice arm was related to trial procedures (22.6% versus 11.1%). The investigators attributed the imbalance in protocol deviations to the more frequent dosing and the intensive schedule of visits in the tebentafusp arm, and stated that the protocol deviations were not expected to affect interpretation of the results.

Exposure to Study Treatments

Exposure to study treatments is summarized in Table 13. The median duration of treatment was 163.0 (range, 1 to 1,016) days in the tebentafusp arm and 65.0 (range, 1 to 658) days in the investigator’s-choice arm. The median relative dose intensity was 100.0% in the tebentafusp and investigator’s-choice arms (range, 95% to 100% and 100% to 100%, respectively).

Table 13: Study Drug Exposure (Safety Analysis Set)

SD = standard deviation.

Note: Data cut-off date of October 13, 2020.

^aDate of last study drug administration – date of first study drug administration + 1.

^b(Total actual dose received intensity / total planned dose intensity) * 100.

Source: Study 202 Clinical Study Report.¹¹

Treatment Beyond Initial Progression

The proportion of patients who continued the assigned study treatment beyond initial disease progression was 43.3% in the tebentafusp arm and 14.3% in the investigator’s-choice arm, whereas the median duration of treatment beyond initial progression was 1.9 (range, 0.1 to 28.8) months and 1.1 (range, 0.10 to 13.0) months, respectively.

Concomitant Medications

A summary of common concomitant medications used during the study is shown in Table 15. The most frequently used concomitant medications (tebentafusp arm versus investigator’s-choice arm) were analgesics (89.7% versus 41.3%), systemic antihistamine (86.1% versus 18.3%), blood substitutes and perfusion solutions (56.7% versus 14.3%), topical corticosteroids (54.0% versus 14.3%), and anti-inflammatory and antirheumatic products (47.2% versus 29.4%). The proportion of patients who received these medications was notably higher in the tebentafusp arm than in the investigator’s-choice arm, including psycholeptics (32.9% versus 20.6%).

Subsequent Anticancer Treatments

A summary of anticancer treatments used after discontinuation of the study treatment is shown in Table 16. The most common subsequent anticancer treatment was systemic therapy, reported in 43.3% of patients in the tebentafusp arm and 43.7% of patients in the investigator’s-choice arm. The median duration of subsequent systemic therapy was 110 (range, 70 to 653) days in the tebentafusp arm and 70 (range, 1 to 721) days in the investigator’s-choice arm.

Table 14: Treatment Beyond Initial Progression (ITT Analysis Set)

ITT = intention to treat; SD = standard deviation.

^aDuration of treatment beyond RECIST 1.1 progression is the time from RECIST 1.1 progression to the date of last dosing recorded at the time of the data cut-off date.

Source: Study 202 Clinical Study Report.¹¹

Table 15: Common Concomitant Medications (in > 30% of Patients) (ITT Analysis Set)

ITT = intention to treat.

Note: Data cut-off date of October 13, 2020.

Source: Study 202 Clinical Study Report.¹¹

Table 16: Subsequent Anticancer Treatments (ITT Analysis Set)

ITT = intention to treat.

Note: Data cut-off date of October 13, 2020.

Source: Study 202 Clinical Study Report.¹¹

Efficacy

Only efficacy outcomes and analyses of subgroups identified in the review protocol are reported here. Refer to Appendix 3 for detailed efficacy data.

Survival Outcomes

Overall Survival

The coprimary end points of OS in the ITT analysis set and the RAS are summarized Table 17 and Table 18, respectively. A KM plot of OS in the ITT analysis set is shown in Figure 3.

Table 17: OS Outcomes (ITT Analysis Set)

CI = confidence interval; ITT = intention to treat; OS = overall survival.

^aBased on Kaplan-Meier estimate.

^bHazard ratio was estimated using a Cox proportional hazards model stratified by LDH status.

^cP value based on log-rank test of the Kaplan-Meier curve stratified by LDH status.

^dAccording to the sponsor, the OS analysis at the data cut-off date of August 12, 2021, was informal; as such, the P value should be considered exploratory.

Source: Study 202 Clinical Study Report.¹¹

Table 18: OS Outcomes (RAS Analysis Set)

CI = confidence interval; NR = not reported; OS = overall survival; RAS = rash analysis set.

Note: The RAS included all patients randomized to tebentafusp who developed a rash in the first week after treatment and all patients randomized to the investigator’s-choice arm.

^aBased on Kaplan-Meier estimate.

^bHazard ratio was estimated using a Cox proportional hazards model stratified by LDH status.

^cP value based on log-rank test of the Kaplan-Meier curve stratified by LDH status.

Source: Study 202 Clinical Study Report.¹¹

The primary analysis of OS (data cut-off date of October 13, 2020) in the ITT analysis set was based on a median duration of follow-up of 14.1 (95% CI,12.7 to 15.6) months when 150 death events occurred (87 events [34.5%] in the tebentafusp arm; 63 events [50.0%] in the investigator’s-choice arm). The KM estimate for the median OS in the ITT analysis set was 21.7 (95% CI, 18.6 to 28.6) months in the tebentafusp arm and 16.0 (95% CI, 9.7 to 18.4) months in the investigator’s-choice arm, with an HR of 0.51 (95% CI, 0.37 to 0.71; P < 0.0001) in favour of tebentafusp. Results of the updated analysis (data cut-off date of August 12, 2021) were consistent with those of the primary analysis, and showed that the KM estimate for the median OS in the ITT analysis set was 21.7 (95% CI, 19.1 to 26.0) months in the tebentafusp arm and 16.7 (95% CI, 11.8 to 19.3) months in the investigator’s-choice arm, with an HR of 0.58 (95% CI, 0.44 to 0.77).

The primary analysis of OS (data cut-off date of October 13, 2020) in the RAS, which included all patients randomized to tebentafusp who developed a rash in the first week after treatment and all patients randomized to the investigator’s-choice arm, was based on a median duration of follow-up of 15.2 (95% CI, 13.3 to 17.0) months when 106 death events occurred (43 patients [28.9%] in the tebentafusp arm; 63 patients [50.0%] in the investigator’s-choice arm). The KM estimate for the median OS in the RAS was 27.4 (95% CI, 20.2 to NR) months in the tebentafusp arm and 16.0 (95% CI, 9.7 to 18.4) months in the investigator’s-choice arm, with an HR of 0.38 (95% CI, 0.25 to 0.56; P < 0.0001) in favour of tebentafusp.

Figure 3: KM Plot of OS (ITT) — First Interim Analysis (Data Cut-Off Date of October 13, 2020)

CI = confidence interval; HR = hazard ratio; KM = Kaplan-Meier; IMCgp100 = tebentafusp; ITT = intention to treat; OS = overall survival.

Source: Study 202 Clinical Study Report.¹¹

The prespecified sensitivity analysis of OS in the ITT analysis set (data cut-off date of October 13, 2020) did not occur because the predefined criteria for conducting such an analysis (if discrepancies were more than 10% between the LDH stratum used at randomization and the central LDH stratum) was not met.

In an exploratory subgroup analysis based on ECOG PS (data cut-off date of October 13, 2020), the HR between the tebentafusp arm and the investigator’s-choice arm with respect to OS in the ITT analysis set was 0.48 (95% CI, 0.33 to 0.72) in the ECOG PS of 0 subgroup and 0.72 (95% CI, 0.39 to 1.36) in the ECOG PS of 1 subgroup (Table 35 in Appendix 3).

Progression-Free Survival

PFS was a key secondary end point, and the final PFS analysis was available at the data cut-off date of October 13, 2020 (Table 19). A KM plot of PFS is presented in Figure 4.

The final PFS analysis was conducted based on a median follow-up duration of 11.4 (95% CI, 11.1 to 16.6) months and a total of 295 PFS events (198 events [78.6%] in the tebentafusp arm; 97 events [77.0%] in the investigator’s-choice arm). The KM estimate for the median PFS was 3.3 (95% CI, 3.0 to 5.0) months in the tebentafusp arm and 2.9 (95% CI, 2.8 to 3.0) months in the investigator’s-choice arm, with an HR of 0.73 (95% CI, 0.58 to 0.94; P = 0.0139) in favour of tebentafusp.

Table 19: PFS Outcomes (ITT)

CI = confidence interval; ITT = intention to treat; NR = not reported; PFS = progression-free survival.

^aBased on Kaplan-Meier estimate.

^bHazard ratio was estimated using a Cox proportional hazards model stratified by LDH status.

^cP value based on log-rank test of the Kaplan-Meier curve stratified by LDH status.

Source: Study 202 Clinical Study Report.¹¹

Figure 4: KM Plot of PFS (ITT) — Final Analysis (Data Cut-Off Date of October 13, 2020)

CI = confidence interval; HR = hazard ratio; IMCgp100 = tebentafusp; ITT = intention to treat.

Source: Study 202 Clinical Study Report.¹¹

The results of 2 sensitivity analyses of PFS that assessed evaluation-time bias and attrition bias were consistent with those of the primary analysis (Table 36 in Appendix 3).

In an exploratory subgroup analysis of PFS by ECOG PS, the HR between the tebentafusp arm and the investigator’s-choice arm was 0.68 (95% CI, 0.51 to 0.91) in the ECOG PS of 0 subgroup and 0.86 (95% CI, 0.53 to 1.41) in the ECOG PS of 1 subgroup (Table 37 in Appendix 3).

Health-Related Quality of Life

EORTC QLQ-C30

EORTC QLQ-C30 score was a secondary end point. Compliance rates were above 50% through cycle 29 in the tebentafusp arm, whereas in the investigator’s-choice arm, compliance rates declined to less than 50% after cycle 17.

As of the data cut-off date of October 13, 2020, the mean global health status/QoL scale score at baseline was 76.1 (SD = 20.2) in the tebentafusp arm (n = 173) and 74.9 (SD = 20.4) in the investigator’s-choice arm (n = 65). At the end of treatment, the LS mean change in score from baseline was –10.7 (SD = NR) in the tebentafusp arm (n = 76) and –11.0 (SD = NR) in the investigator’s-choice arm (n = 34). The LS mean difference between treatment groups was 0.334 (SD = NR) (Figure 5).

The mean fatigue score at baseline was 20.9 (SD = 20.5) in the tebentafusp arm (n = 173) and 22.9 (SD = 24.6) in the investigator’s-choice arm (n = 65). The LS mean change in score from baseline trended toward deterioration in both arms, and was 10.9 (SD = NR) in the tebentafusp arm (n = 76) and 20.1 (SD = NR) in the investigator’s-choice (n = 34) arm by the end of treatment, with a LS mean difference between groups of –9.259 (SD = NR) (Figure 6).

Analyses of all other EORTC QLQ-30 function and symptom scores were generally consistent with those of the global health status/QoL scale, where the change in score from baseline was stable and similar between treatment arms at most time points.

EQ-5D-5L

EQ-5D-5L score was a secondary outcome. Compliance rates were above 50% through cycle 29 in the tebentafusp arm, whereas in the investigator’s arm, compliance rates declined to below 50% at most time points after cycle 17.

As summarized in Table 20, the mean baseline EQ VAS score was 81.0 (SD = 16.4) in the tebentafusp arm (n = 194) and 80.4 (SD = 18.3) in the investigator’s-choice arm (n = 78). As of the data cut-off date of October 13, 2020, the mean change in EQ VAS score from baseline to end of treatment was –10.1 (SD = 22.53) in the tebentafusp arm (n = 95) and –11.7 (SD = 21.40) in the investigator’s-choice arm (n = 45).

Results of the EQ-5D-5L descriptive system are summarized in Table 38 in Appendix 3. As of the data cut-off date of October 13, 2020, the proportion of patients reporting “no problems” in EQ-5D-5L domains at end of treatment in the tebentafusp arm versus the investigator’s-choice arm, respectively, was 74.8% versus 71.2% for mobility, 89.2% versus 91.5% for self-care, 63.1% versus 61.0% for usual activities, 49.5% versus 40.7% for pain or discomfort, and 55.6% versus 42.4% for anxiety or depression.

Tumour Response Outcomes

Duration of Response

The DOR (secondary end point) in patients who achieved a CR or PR is summarized in Table 21. As of the data cut-off date of August 12, 2021, the median DOR was 9.9 (95% CI, 5.6 to 22.1) months in the tebentafusp arm (n = 26) and 9.7 (95% CI, 2.7 to NR) months in the investigator’s-choice arm (n = 6).

ORR, DCR, BOR

ORR, DCR, and BOR are summarized in Table 22. As of the data cut-off date of August 12, 2021, the prespecified time point for the ORR analysis (i.e., all randomized patients had been followed for at least 3 assessments) was met. The ORR was 10.3% (95% CI, 6.9% to 14.8%) in the tebentafusp arm and 4.8% (95% CI, 1.8% to 10.1%) in the investigator’s-choice arm, corresponding to an OR of 2.26 (95% CI, 0.91 to 5.61; P = 0.684). The DCR was 45.6% (95% CI, 39.4% to 52.0%) in the tebentafusp arm and 27.0% (95% CI, 19.5% to 35.6%) in the investigator’s-choice arm, corresponding to an OR of 2.34 (95% CI, 1.45 to 3.76). In both the tebentafusp and investigator’s-choice arms, the most frequently observed BOR was disease progression (52.0% versus 64.3%), followed by stable disease (35.3% versus 22.2%), PR (9.9% versus 4.8%), and CR (0.4% versus 0%).

Figure 5: LS Mean Change From Baseline in Mean Global Health Status/QoL Scale Score Over Time (ITT Analysis Set) — Data Cut-Off Date of October 13, 2020

C3D1 = cycle 3 day 1; C5D1 = cycle 5 day 1; C9D1 = cycle 9 day 1; EOT = end of treatment; IMCgp100 = tebentafusp; ITT = intention to treat; LS = least squares; QoL = quality of life.

Note: The LS mean change from baseline in scale score was computed using the mixed model repeated measures approach, with treatment group (binary), time point (categorical), LDH status (binary), and baseline score (continuous) as covariates.

Source: Study 202 Clinical Study Report.¹¹

Figure 6: LS Mean Change From Baseline in Mean Fatigue Scale Score Over Time (ITT Analysis Set) — Data Cut-Off Date of October 13, 2020

C3D1 = cycle 3 day 1; C5D1 = cycle 5 day 1; C9D1 = cycle 9 day 1; EOT = end of treatment; IMCgp100 = tebentafusp; ITT = intention to treat; LS = least squares.

Note: The LS mean change from baseline in scale score was computed using the mixed model repeated measures approach, with treatment group (binary), time point (categorical), LDH status (binary), and baseline score (continuous) as covariates.

Source: Study 202 Clinical Study Report.¹¹

Table 20: EQ VAS Results (ITT Analysis Set)

C3D1 = cycle 3 day 1; C9D1 = cycle 9 day 1; C17D1 = cycle 17 day 1; EOT = end of treatment; EQ VAS = EQ visual analogue scale; ITT = intention to treat; SD = standard deviation.

Source: Study 202 Clinical Study Report.¹¹

Table 21: DOR (All Patients Who Achieved a CR or PR in the ITT Analysis Set)

CI = confidence interval; CR = complete response; DOR = duration of response; ITT = intention to treat, NR = not reported; PR = partial response.

Note: Only patients who achieved a CR or PR were included.

^aDOR is defined as the time from the date of first documentation of PR or better to the date of first documentation of progressive disease or death from any cause, whichever comes first.

^bCalculated using the reverse Kaplan-Meier method.

Source: Study 202 Clinical Study Report.¹¹

Table 22: ORR, DCR, and BOR (ITT)

BOR = best overall response; CI = confidence interval; DCR = disease control rate; ITT = intention to treat; NR = not reported; ORR = objective response rate.

^a95% CIs were calculated using the exact Clopper-Pearson method.

^bOdds ratio was calculated using the Cochran-Mantel-Haenszel test, stratified by LDH status.

^cNot controlled for multiplicity.

^dStable disease was classified based on assessments made at least 77 days after the date of randomization.

Source: Study 202 Clinical Study Report.¹¹

Posthoc OS Analysis in Patients With a BOR of Progressive Disease

A posthoc analysis of landmark OS on day 100 was conducted in patients who had a BOR of disease progression to understand the role of treatment beyond progression. The proportion of patients who were alive and had a best response of progressive disease by day 100 was 41.7% in the tebentafusp arm and 42.1% in the investigator’s-choice arm. In these patients, the HR of OS between the tebentafusp arm and the investigator’s-choice arm was 0.43 (95% CI, 0.27 to 0.68). A postlandmark KM plot of OS in patients with a BOR of progressive disease is shown in Figure 10 in Appendix 3.

Harms

Only harms identified in the review protocol are reported here. Refer to Table 23 for detailed harms data as of the data cut-off date of October 13, 2020.

Adverse Events

All patients in the tebentafusp arm and 94.6% of patients in the investigator’s-choice arm reported at least 1 TEAE. TEAEs of grade 3 or higher occurred in 54.3% of the tebentafusp arm and 36.0% of the investigator’s-choice arm. The most common TEAEs of any grade were (tebentafusp versus investigator’s choice) pyrexia (76% versus 7.2%), pruritus (69.0% versus 23.4%), rash (55.1% versus 16.2%), and fatigue (51.0% versus 35.1%).

Serious Adverse Events

Serious TEAEs were reported in 28.2% of patients in the tebentafusp arm and 23.1% of patients in the investigator’s-choice arm. The most common serious TEAE was CRS, reported in 9.8% of patients in the tebentafusp arm and no patients in the investigator’s-choice arm.

Withdrawal Due to Adverse Events

The proportion of patients who discontinued treatment due to TEAEs was 3.3% in the tebentafusp arm and 6.6% the investigator’s-choice arm. No specific AE was identified to account for the majority of treatment discontinuations.

Mortality

Eighty-four deaths (34.3%) and 57 deaths (51.4%) were reported in the tebentafusp arm and the investigator’s-choice arm, respectively. The majority of deaths in the tebentafusp and investigator’s-choice arms were attributed to disease progression (32.7% versus 46.8%). No specific AE was identified to account for the majority of deaths in either arm.

Notable Harms

The proportion of patients who reported CRS (or presentations related to CRS) and dermatological AEs was notably higher in the tebentafusp arm than in the investigator’s-choice arm, with the most common notable harms (30% or greater) being pyrexia, pruritus, rash, fatigue, nausea, chills, hypotension, dry skin, headache, and maculopapular rash. The most commonly reported notable harms of grade 3 or higher in the tebentafusp arm were rash (9.4%), maculopapular rash (8.6%), fatigue (5.3%), and pruritus (4.5%).

Table 23: Summary of Harms (Safety Analysis Set)

AE = adverse event; ALT = alanine transaminase; AST = aspartate transaminase; TEAE = treatment-emergent adverse event.

Note: Data cut-off date of October 13, 2020.

^aTEAE was defined as an AE that had an onset date on or after the date of first dose or pre-treatment TEAEs that increase in severity on or after the date of first dose up to and including 90 days after the date of last dose of the study drug.

^bFrequency ≥ 20%.

^cCRS was reported by the investigator based on Lee et al. (2014)¹² grading criteria.

^dFrequency ≥ 5%.

^eFrequency ≥ 2%.

Source: Study 202 Clinical Study Report.¹¹

Critical Appraisal

Internal Validity

Study 202 was a phase III, open-label, randomized controlled trial (RCT). The methods of randomization, which involved stratification by LDH level and the use of an Interactive Response Technology system for randomized assignment, were considered appropriate. There was generally no notable imbalance in baseline patient characteristics between treatment groups, which suggests that randomization was likely successful. A notably higher proportion of patients received concomitant analgesics, systemic antihistamines, blood substitutes and perfusion solutions, topical corticosteroids, and anti-inflammatory and antirheumatic products in the tebentafusp arm than the investigator’s-choice arm. The clinical expert noted this to be reasonable, as these medications are routinely used for the management of CRS and rash associated with tebentafusp and were not expected to confound the study results.

The CADTH review team and the clinical expert considered the open-label design to be reasonable, given that the distinct dosing regimens and toxicity profiles of the interventions could likely allow investigators and patients to make inferences about treatment assignment regardless of blinding. It is unlikely that the reporting of OS was influenced by knowledge of treatment assignment, considering the objective nature of this outcome; however, subjective outcomes reported by patients, such as HRQoL and AEs, might be subject to reporting bias. In addition, tumour response outcomes (PFS, ORR, BOR, DCR, DOR) were based on investigator assessment; no independent blinding committee was involved in the assessment process to help mitigate the biases associated with the open-label study design. Therefore, there is a risk of bias associated with the measurement of these outcomes, although the extent and direction of bias are unclear.

The primary analysis presented constitutes the first interim analysis of OS and the final analysis of PFS, which were conducted as planned a priori based on the number of OS or PFS events that occurred. Stopping rules for interim analyses of OS were prespecified in the protocol and alpha levels were properly accounted for to preserve power. A graphical approach involving sequential testing of the coprimary end point of OS in the RAS and ITT populations, and key secondary end points (PFS and BOR [evaluated statistically as ORR]), was appropriately used to account for multiplicity. No formal hypothesis testing was carried out for other secondary outcomes (DCR, DOR, HRQoL), so these results are considered exploratory. There is an increased risk of false-positive conclusions for outcomes that were not controlled for multiplicity. No conclusion could be drawn regarding the subgroup effects of ECOG PS on OS, considering the subgroup analysis was exploratory in nature, and there was a lack of sample-size consideration and control for multiplicity.

With respect to the Cox proportional hazards model used for OS and PFS analyses, no assessment of the proportional hazards assumption was reported; therefore, it is unknown whether there were any violations of the assumption or whether adjustments were involved to address any potential violations. The censoring rules and the method of handling missing data used were, overall, considered appropriate.

There are 2 limitations to note with respect to the OS analyses. First, there is typically a risk of overestimating the treatment effect in favour of the experimental intervention (i.e., tebentafusp) in an interim analysis.¹³ Considering that the interim OS analyses were based on a relatively small number of events, the OS results are prone to imprecision. Second, the results of the OS analysis of the RAS and the posthoc exploratory landmark OS analysis in patients with a BOR of progressive disease should be interpreted with caution, given that no randomization was involved in the comparison. The results were subject to confounding by the potential differences in prognostic factors and treatment-effect modifiers between the treatment arms; however, the direction of bias is inconclusive.

It should also be noted that 16 patients in the investigator’s-choice arm were allowed to cross over to the tebentafusp arm between the data cut-off dates of October 13, 2020, and August 12, 2021. No adjustments were made to account for the crossover in the informal updated analysis of OS or the primary BOR (evaluated statistically as ORR) analysis (data cut-off date of August 12, 2021), according to the sponsor. The treatment crossover might have biased the findings in favour of the investigator’s-choice arm, rather than the tebentafusp arm, and therefore is not expected to have had a major impact on interpretation of OS findings in this study, specifically with respect to the direction of OS benefits.

The EORTC QLQ-C30 and EQ-5D-5L scales used for HRQoL assessment in this study are commonly used in oncology trials, although the validity, reliability, and responsiveness of these instruments have not been studied in patients with UM. There are also uncertainties about the HRQoL outcomes because of the large amount of missing data as a result of treatment discontinuation and poor questionnaire compliance rates (above 50% in the tebentafusp arm throughout the study but declined to less than 50% in the investigator’s-choice arm at most time points after cycle 17). There is a risk that reporting bias results from patients who remained in the trial affected the interpretability of HRQoL trends over time. Also, no data imputation was performed to account for the missing data.

External Validity

The clinical expert noted that the inclusion and exclusion criteria of Study 202 were generally reflective of the eligibility criteria in clinical practice for systemic treatments; however, patients with prior therapy in the metastatic setting were not captured in the study. This was a limitation, and the clinical expert noted that tebentafusp should not only be offered in the first-line setting, but also in the second-line setting and beyond, based on clinical trial and anecdotal evidence. The clinical expert also indicated that clinical practice is usually less restrictive than clinical trials with respect to ECOG PS and comorbidities. The baseline characteristics of enrolled patients were generally similar to the Canadian patient population, per the clinical expert.

The dosing of tebentafusp in Study 202 was consistent with the recommended dosing in the product monograph. Of note, although the product monograph recommended continuation of treatment until disease progression or unacceptable toxicity, the clinical expert explained that continuation of tebentafusp beyond initial progression until no clinical benefits are observed is more appropriate in clinical practice. The dosing approach of tebentafusp in Study 202 was consistent with the approach used in clinical practice and, therefore, would not undermine the external validity of the results.

In the investigator’s-choice arm, patients were assigned to pembrolizumab (81.7%), ipilimumab (12.7%), or chemotherapy (5.6%). According to the clinical expert, ipilimumab and chemotherapy are not the standard of care for patients with mUM who are naive to systemic treatment. Further, the clinical expert indicated that pembrolizumab only accounts for approximately 10% of systemic treatments prescribed in the first-line setting. The clinical expert indicated that ipilimumab plus nivolumab is currently used in most patients with mUM receiving systemic treatments. However, this regimen was not included in the comparator arm. The clinical expert commented that the potential impact on generalizability is likely to be small, noting that a retrospective study showed no significant difference in response rates between single-drug and combination immunotherapy regimens,³⁸ although the study results might be subject to bias, owing to the small sample size and retrospective design.

OS is an important clinical outcome, according to clinician and patient input. The clinical expert noted that given the data available, the OS benefits of tebentafusp were considered clinically meaningful, although a longer follow-up (2 to 3 more years) would help provide further confidence in the OS benefits of tebentafusp. PFS and tumour response outcomes were noted to be less clinically important by the clinical expert, as tumour response is poorly correlated with OS in patients with mUM who receive systemic therapy in general. Further, the PFS benefit of tebentafusp was small in Study 202, with a median extension in PFS of 0.4 months compared with investigator’s choice. Improvement in HRQoL was highly valued by patients with mUM. The clinical expert noted that the EORTC QLQ-C-30 and EQ-5D-5L scales are not routinely used in clinical practice, so the generalizability of HRQoL findings in Study 202 could not be ascertained. It is worth noting that these HRQoL instruments capture some of the most common HRQoL impacts reported by patients, including fear, anxiety, depression, fatigue, and decline in physical and psychological functioning; symptoms such as loss of vision and balance were not assessed. Prevention of vision loss was noted to be an important outcome by the patient group; however, this outcome was not assessed in the study. The clinical expert noted that prevention of vision loss is not an expected outcome of tebentafusp.

The product monograph recommends that the first 3 doses of tebentafusp be administered in a health care setting that would allow a patient to be monitored during the infusion and for 16 hours after the infusion is complete, whereas subsequent doses can be administered in an ambulatory setting in the absence grade 2 or worse hypotension, which reflects the treatment settings in Study 202. The clinical expert indicated that the safety findings of tebentafusp are generalizable to clinical practice, provided tebentafusp is administered in appropriate treatment settings, as specified in the product monograph, and noted that, based on clinical experience, CRS generally occurs after the first 3 or 4 infusions and is generally manageable if proper supportive care is provided.

Indirect Evidence

Objectives and Methods for the Summary of Indirect Evidence

Because of a lack of studies directly comparing tebentafusp with systemic treatments other than pembrolizumab, ipilimumab, and dacarbazine monotherapies for mUM, a review of indirect evidence was undertaken.

A focused literature search for ITCs dealing with UM was run in MEDLINE All (1946–) on May 18, 2022. No limits were applied to the search. Three published ITCs were identified, none of which met the selection criteria in the review protocol (Table 6). One sponsor-submitted ITC was included in this review. The sponsor also submitted a supplementary analysis of the prognostic value of HLA-A*02:01 status for OS to support the MAIC analyses.

Description of Indirect Comparisons

The sponsor submitted 1 report that included 3 MAIC analyses of survival results in patients with previously untreated mUM who received ipilimumab plus nivolumab versus tebentafusp, versus pembrolizumab, and versus the investigator’s-choice arm in Study 202. The comparison between tebentafusp and ipilimumab plus nivolumab is of interest for this review and will be addressed in this report.

Methods of Sponsor-Submitted ITC

Objectives

The primary objective of the sponsor-submitted ITC was to compare OS between tebentafusp and ipilimumab plus nivolumab in patients with previously untreated mUM.

Study Selection Methods

To identify the clinical evidence for tebentafusp and relevant comparator interventions, an SLR was conducted using 7 electronic databases and 2 trial registries, along with handsearches of review articles, reference lists of included full-text publications, and free-text keyword searches in internet search engines. The selection criteria of the SLR are summarized in Table 24. Of note, the SLR that was originally designed to fulfill regulatory requirements in Europe and the inclusion criteria were wider than intended for the ITC, specifically the cutaneous melanoma population, and outcomes such as ORR, DCR, AEs, and QoL were not of interest for the sponsor-submitted ITC. It was inferred that the selection criteria could have been refined to align with the objectives of the analysis, although no details were stated in the sponsor-submitted ITC report.

Table 24: Study Selection Criteria and Methods for the Sponsor-Submitted Systematic Review for ITC

CDSR = Cochrane Database of Systematic Reviews; CENTRAL = Cochrane Central Register of Controlled Trials; CM = cutaneous melanoma; DARE = Database of Abstracts of Reviews of Effects; HTAD = Health Technology Assessment Database; ITC = indirect treatment comparison; RCT = randomized controlled trial; UM = uveal melanoma; WHO ICTRP = WHO International Clinical Trials Registry.

^aThe SLR protocol was initially designed to fulfill regulatory requirements and included patients with advanced or metastatic CM and outcomes including overall response rate, disease control rate, adverse events, and quality of life. These elements, however, were not of interest for the sponsor-submitted ITC.

Source: Sponsor-submitted SLR for ITC.³⁹

The criteria included studies published before March 2020 and the search was updated September 2021. Article screening was performed independently by 2 reviewers at 2 stages (title and abstracts, and then full papers), and a third reviewer helped resolve disagreements. Reasons for exclusion were documented. Data extraction was performed by 1 reviewer and quality checked by a second reviewer.

Quality assessment of the selected studies was conducted by 1 reviewer to assess the risk of bias using the Cochrane RoB 2.0 tool for RCTs, and Cochrane ROBINS-I tool for single-arm trials

ITC Analysis Methods

A summary of the analysis methods for the MAIC is shown in Table 25. An unanchored MAIC approach was selected for the indirect comparison between tebentafusp and ipilimumab plus nivolumab, owing to the lack of a common comparator. The covariates used for match adjustment were informed by a multivariate analysis of OS in Study 202. Adaptations were applied to the covariates to account for the limitations of available data and to allow matching to be implemented. Note that time since primary diagnosis was an identified covariate but could not be used in matching as it was not reported in the GEM-1402 trial. IPD from Study 202 were then weighted to match baseline characteristics in the comparator trial (GEM-1402) using weights estimated with the method of moments described in the Signorovitch et al. (2012)^40,41 studies, and weights were examined for extreme values. The ESS was calculated using the methods described in the Pillippo et al. (2016)⁴² study. The covariate characteristics of patients before and after weighting were presented. A weighted Cox proportional hazards model was fit, with the MAIC weights applied to produce the efficacy end points.

Results of an unadjusted indirect comparison (UAIC) and MAIC were presented.

The sponsor noted that potential issues with ESS calculations and modelling could arise because of the notably smaller proportion of patients with extrahepatic disease only in Study 202 than in the GEM-1402 trial (3.6% versus 21.2%); as such, 2 exploratory sensitivity analyses were conducted to assess the impact of using alternative definitions for the disease location covariate for matching.

Note that OS was the prespecified outcome of interest in the MAIC analyses, and the PFS analysis was added subsequent to observing the results for OS.

Table 25: ITC Analysis Methods

ALP = alkaline phosphatase; ECOG PS = Eastern Cooperative Oncology Group Performance Status; ITC = indirect treatment comparison; LDH = lactate dehydrogenase; MAIC = matching-adjusted indirect comparison.

Source: Sponsor-submitted ITC Technical Report.¹⁴

Results of Sponsor-Submitted ITC

Summary of Included Studies

After removal of duplicates, 1,531 citations were screened by title and abstract; of these, 336 citations were reviewed by full-text screening and eventually 60 studies met the selection criteria. Two published studies of ipilimumab plus nivolumab were identified —Pelster et al. (2021)⁴³ and GEM-1402¹⁵ — both of which were single-arm studies of patients with mUM. Study 202¹⁰ (data cut-off date of October 13, 2020) was the index trial, which contained IPD for the tebentafusp group.

The GEM-1402 trial was selected by the sponsor as the comparator study for the MAIC because the trial had a larger sample size and reported more key covariates for matching than the Pelster et al. (2021) study. In addition, the sponsor noted that the GEM-1402 trial was a multicentre study that contained treatment-naive patients only, similar to Study 202, whereas the Pelster et al. (2021) study was a single-centre study in which only 57% of patients were treatment-naive. The characteristics of Study 202 and the GEM-1402 trial are summarized in Table 26. The risk of bias (per ROBINS-I appraisal by a single reviewer) in the GEM-1402 trial was noted to be low in the sponsor’s assessment.

Results

Patient characteristics from Study 202 (unadjusted and match-adjusted) and the GEM-1402 trial are summarized in Table 27. The ESS for tebentafusp after match adjustment was 115.93 (48.9% of the original sample size from Study 202).

Table 26: Summary of Included Studies in the MAIC

ECOG PS = Eastern Cooperative Oncology Group Performance Status; HLA = human leukocyte antigen; IPD = individual patient data; mUM = metastatic uveal melanoma; OS = overall survival; PFS = progression-free survival; RCT = randomized controlled trial; RECIST = Response Evaluation Criteria in Solid Tumours; UM = uveal melanoma.

Sources: Study 202 Clinical Study Report,¹¹ Piulats et al. (2020).¹⁵

Table 27: Summary of Model Covariates in Unadjusted Indirect Comparison and MAIC

ALP = alkaline phosphatase; ECOG PS = Eastern Cooperative Oncology Group Performance Status; ESS = effective sample size; LDH = lactate dehydrogenase.

Note: Study 202 data cut-off date: October 13, 2020; GEM-1402 data cut-off date: July 9, 2019).

^aMissing values for any variables were excluded from the analysis.

^bMissing data in the GEM-1402 trial were excluded from the calculation of proportions.

Source: Sponsor-submitted Indirect Treatment Comparison Technical Report.¹⁴

UAIC and MAIC (primary analysis) of OS and PFS between tebentafusp and ipilimumab plus nivolumab are summarized in Table 28. The KM curves of OS and PFS from MAIC analyses are shown in Figure 7 and Figure 8, respectively.

The MAIC analysis between tebentafusp and ipilimumab plus nivolumab showed results in favour of tebentafusp with respect to OS (HR, 0.507; 95% CI, 0.324 to 0.793) and PFS (HR, 0.647; 95% CI, 0.445 to 0.941). Results of the sensitivity analyses, which assessed the impact of using alternative definitions for the disease location covariate for matching, were consistent with the primary analyses (Table 39).

Table 28: OS and PFS Results in Unadjusted Indirect Comparison and MAIC (Primary Analysis)

CI = confidence interval; ESS = effective sample size; OS = overall survival; PFS = progression-free survival; SE = standard error.

Source: Sponsor-submitted Indirect Treatment Comparison Technical Report.¹⁴

Sponsor-Submitted Supplementary HLA Analysis

The sponsor submitted a supplementary analysis (N = 80) assessing the comparative OS in patients with mUM who were HLA-A*02:01-positive (n = 43) versus patients with mUM who were HLA-A*02:01-negative (n = 37).¹⁶ The OS data were sourced from a clinical dataset published in the Liu et al. (2018)⁴⁴ study. With respect to statistical analysis, the sponsor stated that “Survival analysis was carried out using R package survminer v0.4.9 and Cox log-rank test was used to assess differences between the survival curves. Univariate Cox proportional hazards methods (R package survival version 3.2-11) were used to model the prognostic importance of potential predictors of survival. Tests for correlation (2-sided) were conducted using R stats package 4.1.” The median OS was 45.9 months (range, NR) in HLA-A*02:01-positive patients and 45.2 months (range, NR) in HLA-A*02:01-negative patients, with an HR of 0.82 (95% CI, 0.36 to 1.88). The OS curves of HLA-A*02:01-positive and HLA-A*02:01-negative patients with mUM are shown in Figure 12 in Appendix 3.

Critical Appraisal of Sponsor-Submitted ITC

The MAIC was conducted based on studies identified by an SLR. The selection criteria of the SLR were broad and, as a result, many studies that were not of interest for the MAIC were identified in the search. It appeared that the selection process was refined to select studies that align with the MAIC study objectives, but there is a lack of transparency on how the process was carried out. There is potential for selection bias because the criteria for selecting the comparator trial was not decided a priori. The search approach was comprehensive, involving multiple electronic databases, clinical trial registries, and a supplementary manual search. The study selection and data extraction processes were considered appropriate, involving at least 2 independent reviewers. Reasons for study exclusion were reported. A quality assessment was performed for all included trials. Because the risk of bias appraisals were performed by a single reviewer without verification or consensus, there is an increased risk of error and bias in the appraisals. The comparator trial was noted to have a low risk of bias, based on the assessment by the sponsor.

Figure 7: KM Plot of OS From MAIC (Primary Analysis)

IMCgp100 = tebentafusp; Ipi = ipilimumab; KM = Kaplan-Meier; MAIC = matching-adjusted indirect comparison; Nivo = nivolumab; OS = overall survival.

Source: Sponsor-submitted Indirect Treatment Comparison Technical Report.¹⁴

An important limitation of the MAIC is that the patient population in the index trial (i.e., HLA-A*02:01-positive patients with mUM) was narrower than the comparator trial (patients with mUM regardless of HLA status). It is not possible to account for such difference in the adjustment process because only aggregate data were available for the comparator trial. It is unclear what proportion of patients in the comparator trial was HLA-A*02:01-positive. The sponsor submitted a supplementary analysis exploring the impact of HLA-A*02:01 status on disease prognosis in patients with mUM; however, CADTH was unable to draw any definitive conclusions, given that the study was based on a small observational cohort and results are subject to imprecision, considering the wide 95% CI. It is also not clear if the methodologies of the analysis were robust from the brief description of statistical analysis provided. As such, results of the MAIC should be interpreted with caution, owing to potential confounding by the heterogeneity in patient population between the trials. All other inclusion and exclusion criteria, as well as the median duration of follow-up, were generally similar in the trials. There was inconsistency in the definition of PFS and OS between the trials, although the impact on study findings is unlikely to be substantial.

Figure 8: KM Plot of PFS From MAIC (Primary Analysis)

IMCgp100 = tebentafusp; Ipi = ipilimumab; KM = Kaplan-Meier; MAIC = matching-adjusted indirect comparison; Nivo = nivolumab; PFS = progression-free survival.

Source: Sponsor-submitted Indirect Treatment Comparison Technical Report.¹⁴

The choice to conduct an unanchored MAIC was appropriately justified by the lack of a common comparator across the included trials. The prognostic and treatment-effect modifying factors that were adjusted for in the weighting process were identified from a multivariate analysis of OS results in Study 202. The methods of the multivariate analysis were not available for assessment by the CADTH review team; as such, there is uncertainty about the appropriateness of the identified factors and the direction of potential bias. Another point to note is that time since primary diagnosis, a covariate identified in the multivariate analysis, was excluded from adjustment because it was not reported in the comparator trial.

Unadjusted and match-adjusted baseline covariates were reported. Rescaling of extreme weights was involved after an examination of weight distribution, and the distribution of rescaled weights was reported. There was a notable reduction in ESS, by approximately 51.1%, after the weighting process, which suggest poor population overlap. A significant reduction in sample size can contribute to imprecision, increasing uncertainty of the results. A similar issue was noted with the sensitivity analyses, which further limits the ability to draw any conclusion from the MAIC. With respect to the UAIC, the result should be considered exploratory because naive direct comparisons are subject to significant confounding and bias, owing to the lack of control for systematic differences between trials.⁴⁵ It should also be noted that the PFS results should be interpreted with caution because the posthoc analysis was performed after results of the OS analysis were available and was subject to potential bias.

With respect to external validity, ipilimumab plus nivolumab was a relevant comparator as it accounts for most first-line systemic treatment prescribed for mUM in Canada, according to the clinical expert. A limitation to note is that the analysis was restricted to patients who were treatment-naive in the metastatic setting. The clinical expert consulted by CADTH noted that use of tebentafusp in patients with mUM who had received prior systemic therapy was anticipated; thus, the findings of the MAIC may have limited generalizability to this specific patient population. Further, the MAIC assessed OS and PFS end points only, and other efficacy outcomes that were of interest to the stakeholders, such as HRQoL and harms, were not investigated.

Other Relevant Evidence

This section includes 2 additional studies from the sponsor’s submission to CADTH that were considered to address important gaps in the evidence included in the Systematic Review.

The first study (Study 102) was a phase I/II, multicentre, open-label study that analyzed the efficacy and safety of tebentafusp in HLA-A*02:01-positive patients with mUM who had received 1 or 2 prior lines of therapy in the metastatic or advanced setting. Although the study did not meet the Systematic Review inclusion criteria, as it is not a phase III study, the CADTH review team summarized the study design and results of Study 102 to provide supporting evidence for tebentafusp in patients with mUM who had received prior therapies.

The second study was an observational study that compared patients receiving tebentafusp from Study 202 with patients receiving ipilimumab plus from the GEM-1402 trial after propensity score weighting. The analysis was based on the same studies that informed the sponsor-submitted MAIC and was submitted to CADTH after the sponsor had obtained IPD from the GEM-1402 trial. The study aimed to address the limitations of the MAIC analysis related to the use of aggregate data from the GEM-1402 trial.

Study 102

The characteristics of Study 102 are presented in Table 29. Study 102 was designed as a phase I/II study of tebentafusp administered on a weekly basis with an intrapatient escalation dosing regimen. The intrapatient escalation regimen consisted of fixed low doses of tebentafusp at cycle 1 day 1 (20 mcg) and cycle 1 day 8 (30 mcg), followed by an escalated cohort specified weekly dose administered at cycle 1 day 15 and beyond. According to this regimen, all patients in the study received 2 weekly doses of tebentafusp at a dose level below the identified recommended phase II dose of 50 mcg, and then dose escalation commenced at the third weekly dose (cycle 1 day 15), with the goal of achieve a long-term dosing regimen at a dose higher than that identified for the recommended phase II dose. Doses of 54 mcg, 64 mcg, 73 mcg, 68 mcg were evaluated in the phase I dose-escalation cohorts. The dose-escalation regimen identified the intrapatient escalation regimen of 68 mcg was selected for the phase II dose-expansion cohort. The overview of the study designs for Study 102 is shown in Figure 9. The primary end point for phase I was the incidence (number) of dose-limiting toxicity, whereas ORR was the primary end point for the phase II, single-arm, dose-expansion cohort. The key secondary end points for phase II — OS, PFS, DOR, DCR, and BOR — were detailed in this report, according the prespecified CADTH systematic review protocol. Patients were eligible for the phase II expansion component if they were HLA-A*02:01-positive, had an ECOG PS of 0 or 1, had previously received systemic treatment for mUM, and had measurable disease according to RECIST 1.1. The exclusion criteria were generally consistent with those used in Study 202.

Study 102 is ongoing, with an estimated completion date of March 20, 2024. The evidence summarized in this section is from a planned interim analysis with a data cut-off date of March 20, 2020, and contains results of the phase II expansion cohort only. Results of the phase I study were not of interest to this review, owing to the dose-finding nature and the small sample size.

Table 29: Details of Study 102

CNS = central nervous system; DCR = disease control rate; DOR = duration of response; ECOG PS = Eastern Cooperative Oncology Group Performance Status; HLA = human leukocyte antigen; mUM = metastatic uveal melanoma; ORR = objective response rate; OS = overall survival; PFS = progression-free survival; RECIST = Response Evaluation Criteria in Solid Tumours.

^aThe date on which the last participant in a clinical study was examined or received an intervention or treatment to collect final data for the primary outcome measures, secondary outcome measures, and adverse events (that is, the last participant's final visit).

^bUp to 38 months.

^cORR is defined as the percentage of participants with measurable disease with at least 1 visit response of CR or PR that is confirmed at least 4 weeks later, as defined in RECIST 1.1 and assessed by an investigator. The denominator in the calculation of ORR is the number of participants in the full analysis set with measurable disease at baseline.

^dUp to 49 months.

^ePFS is defined as the time (in months) from first dose of study drug until the date of disease progression or death (from any cause in the absence of disease progression), as assessed by RECIST 1.1 by the investigator for phase I and by the ICR for phase II.

^f24 weeks.

Sources: Study 102 Clinical Study Report;²⁷ Study 102 Registration.⁴⁶

Figure 9: Study Design Schematic for Study 102

Source: Study 102 Clinical Study Report.²⁷

Baseline Characteristics

In the phase II expansion cohort, the majority of patients were white (99.2%) and approximately half were female (49.6%), with a mean age of 61.0 (SD = 10.93) years. There was a higher proportion of patients who were younger than 65 years (63.0%) than who were 65 years or older (37.0%). Most patients (70.1%) had an ECOG PS of 0 at baseline. A total of 58.3% of patients had baseline LDH greater than the ULN, 80.3% of patients had a baseline absolute lymphocyte count of at least 1.0 × 10⁹/L, and 70.9% of patients had a baseline alkaline phosphatase level no higher than the ULN. Of the 127 patients in the phase II dose-expansion cohort, at initial diagnosis, most patients (32.3%) had stage II disease, followed by stage III (22.0%), stage IV (16.5%), and stage I (8.7%). The median time from primary diagnosis to metastatic disease was 2.96 years and the median time from primary diagnosis to study entry was 4.42 years. Approximately two-thirds of patients had received 1 prior anticancer regimen and 33.9% of patients had received at least 2 prior therapies. The most common prior treatments were radiotherapy (75.6%), immunotherapy (73.2%), and liver-directed therapy (44.9%). At study entry, per the investigator’s assessment, 117 patients had at least 1 measurable metastatic lesion in the liver; the largest lesion in 43 (33.9%) patients was no larger than 3 cm in diameter and in 69 (54.3%) patients was at least 3 cm. The baseline characteristics of patients in Study 102, phase II are summarized Table 30.

Table 30: Baseline Characteristics of Patients in Study 102, Phase II (Full Analysis Set)

ALC = absolute lymphocyte count; ALP = alkaline phosphatase; ECOG PS = Eastern Cooperative Oncology Group Performance Status; ICR = independent central review; LDH = lactate dehydrogenase; PD-1 = programmed cell death 1 protein; PD-L1 = programmed cell death 1 ligand 1; SD = standard deviation; ULN = upper limit of normal.

^aLiver metastases measurements based on investigator assessment.

^bLiver metastases measurements based on ICR.

^cMedically reviewed therapy types based on the coded term, reported term, electronic Case Report Form therapy class, and reason for therapy. Includes therapies for primary and metastatic disease only. Also includes liver-directed chemotherapeutic drugs.

^dLiver-directed therapies include the following subcategories: ablation, bland embolization, chemoembolization, immunoembolization, perfusion, radiation, and radioembolization.

Source: Study 102 Clinical Study Report.²⁷

Statistical Analysis

Study 102 data were summarized using descriptive statistics. Categorical data were presented as frequencies and percentages. For continuous data, mean, SD, 95% CI (where applicable), median, minimum, and maximum were presented. The primary end point of the phase II dose expansion, ORR, was estimated according to the number and proportion of patients with an objective response (BOR of CR or PR), based on ICR assessment according to RECIST 1.1. The associated 2-sided 95% CI for the true ORR using the exact Clopper-Pearson method was also presented. This was based on the full analysis set, which included all patients assigned to treatment. The key secondary end points of the phase II dose expansion identified by the CADTH review protocol — OS, PFS, DOR, DCR, and BOR — were derived and reported based on ICR data.

Results

Patient Disposition

A summary of patient disposition is shown in Table 31. A total of 148 patients were screened and 127 patients were enrolled in the phase II dose-expansion part of the study. Of the 127 (100.0%) patients who received the study drug, 106 (83.5%) discontinued the study drug and 21 (16.5%) remained on treatment. The primary reason for treatment discontinuation was disease progression (70.1%). As of the data cut-off date of March 20, 2020, 53 (41.7%) patients remained on study and 74 (58.3%) patients were reported to have ended the study. Death was the primary cause of study discontinuation (54.3%).

Efficacy

The summary of results in this section focuses on the primary and key secondary outcomes of Study 102, phase II (OS, PFS, DOR, DCR, and BOR) to assess the effects of tebentafusp on patients with mUM who had received 1 or 2 prior lines of therapy (shown in Table 32).

For ORR, 6 of 127 patients achieved a PR (4.7%; 95% CI, 1.8% to 10.0%); no patients achieved a CR. After a median follow-up of 19.6 (95% CI, 16.0 to 22.2) months, the median OS was 16.8 (95% CI, 12.9 to 21.3) months. The median PFS was 2.8 (95% CI, 2.0 to 3.7) months, per the RECIST 1.1 assessment by ICR. The median DOR (CR or PR) per the RECIST 1.1 assessment by ICR, was 8.7 (95% CI, 5.6 to 24.5) months. The DCR (CR, PR, or stable disease) was 22.8% (95% CI, 15.7% to 31.2%) at 24 weeks and beyond. Of the 127 patients who received the study drug, the most frequently observed BOR was disease progression (47.2%), followed by stable disease (44.9%), and PR (4.7%).

Table 31: Patient Disposition in Study 102, Phase II (Full Analysis Set)

Note: Data cut-off date of March 20, 2020.

Source: Study 102 Clinical Study Report.²⁷

Table 32: Summary of Key Outcomes in Study 102, Phase II (Full Analysis Set)

CI = confidence interval; DCR = disease control rate; DOR = duration of response; ORR = objective response rate; OS = overall survival; PFS = progression-free survival.

Note: Data cut-off date of March 20, 2020.

^aORR is defined as the percentage of participants with measurable disease with at least 1 visit response of CR or PR that is confirmed at least 4 weeks later, as defined in RECIST 1.1 and assessed by ICR. The denominator in the calculation of the ORR is the number of participants in the full analysis set with measurable disease at baseline.

^b95% CIs were calculated using the exact Clopper-Pearson method.

^cAssessed by RECIST 1.1 by ICR.

^dDOR is defined as the time (in months) from the date of first documented objective response (CR or PR) until the date of documented disease progression or death from cause in the absence of disease progression, as assessed by ICR.

^eParticipants with a BOR of CR or PR or stable disease recorded at least 24 weeks (± 1 week) after commencement of study drug and before any progressive disease event, as assessed by ICR.

^fResponse confirmation at least 4 weeks after initial evaluation, as defined in RECIST 1.1 and assessed by an ICR.

Source: Study 102 Clinical Study Report.²⁷

Harms

As in Study 202, all patients in Study 102 experienced an AE of any grad. In addition, the Study 102 phase II expansion and Study 202 had comparable rates of grade 3 or higher TEAEs (59.1% versus 54.3%). In the Study 102 phase II expansion, 9 patients (7.1%) experienced TEAEs leading to tebentafusp discontinuation, a proportion comparable to that of Study 202 (6.6%). The most frequently reported TEAEs of any grade were pyrexia (81.1%), pruritus (68.5%), nausea (67.7%), chills (66.1%), and hypotension (41.7%). These TEAEs were also observed in the tebentafusp arm of Study 202. A total of 86 serious TEAEs were reported in 42 (33.1%) patients, a proportion comparable to that of Study 202 (28.2%). The serious TEAEs were CRS in 4 (3.1%) patients and sepsis, elevated alanine transaminase, rash maculopapular, and hypotension in 3 (2.4%) patients each. No deaths related to TEAEs were observed in Study 102.

Critical Appraisal

Study 102 was designed as a phase I/II, open-label, single-arm study to determine the optimal drug dosage(s) and to generate hypotheses for the phase III trial (Study 202). The primary end point of ORR and the key secondary end points, such as PFS and DOR, were assessed by ICR based on RECIST 1.1. The use of ICR assessments reduced potential bias in the estimation of treatment effects related to the investigator or patient knowing the intervention received, which could affect the subjectively assessed outcomes, such as response and subjective AEs.

However, the noncomparative design of the Study 102, with no statistical testing, is the key limitation. The lack of a randomized control group means there is uncertainty regarding the magnitude of effects obtained for the efficacy outcomes and their relationship to the study drug. Although the clinical expert consulted explained that the efficacy outcomes of tebentafusp in Study 102 were clinically meaningful, demonstrated the activity of the drug, and were comparable with the Study 202 phase III trial, the CADTH review team notes that in the absence of a comparative arm, the findings obtained from the efficacy and safety analysis are uncertain, as the single-arm design does not allow for conclusions to be drawn about the comparative efficacy for tebentafusp or the differentiation of symptoms of the underlying mUM disease from treatment-related AEs.

Study 102 included patients who had received prior therapy in the metastatic setting, which is supported by the clinical expert’s comments, based on clinical trial and anecdotal evidence, that tebentafusp should not only be offered in first-line setting, but also in the second-setting and beyond. The clinical expert also indicated that clinical practice is usually less restrictive with respect to ECOG PS and comorbidities. The baseline characteristics of Study 102 were suggestive of an overrepresentation of patients with a higher functioning status (ECOG PS of 0) and may compromise the representativeness of the study sample to the general population of adults with mUM.

Propensity Score Analysis (IPTW Approach)

This observational study compared patients in the tebentafusp arm of Study 202 to patients who received ipilimumab plus nivolumab in the GEM-1402 trial.¹⁸ The study was not randomized, and propensity score weighting was used in an attempt to adjust for confounding.

Methods

Objectives

The primary objective of this analysis was to compare OS between tebentafusp and ipilimumab plus nivolumab in patients with previously untreated mUM.

Patient Selection Criteria

Key selection criteria of the 2 cohorts included in the study are summarized in Table 26. As mentioned in the MAIC section, although both studies enrolled patients with unresectable UM or mUM who had received no prior therapy in the metastatic setting, Study 202 enrolled HLA-A*02:01-positive patients only, whereas the GEM-1402 trial enrolled patients regardless of HLA status. All other selection criteria were generally consistent between the studies.

Propensity Score Development

The patient characteristics that were used to balance the cohorts for effect estimation included time from diagnosis to metastasis and the same covariates used in the MAIC analysis, including median age, sex, baseline LDH (normal versus elevated), baseline alkaline phosphatase (normal versus elevated), disease location (hepatic only, extrahepatic only, hepatic and extrahepatic), and ECOG PS, all of which were informed by a multivariate analysis of OS in Study 202. These covariates were used as predictors in a logistic regression model in which tebentafusp treatment was the binary outcome used to estimate the probability of receiving tebentafusp for each patient in the GEM-1402 trial.

Statistical Methods

The propensity score analysis weighted patients using the IPTW approach to estimate treatment effects. In the primary analysis, the average treatment effect of the treated (ATT) method was used, in which all patients who received tebentafusp were assigned a weight of 1, and patients who received ipilimumab plus nivolumab received a weight calculated as the propensity score of treatment divided by 1 minus the propensity score of treatment. The distribution of weights was assessed visually, using box plots for outliners.

Two additional analyses that used different methods to calculate weights were conducted. One analysis used the average treatment effect of the control weights, in which the roles of tebentafusp and ipilimumab plus nivolumab were reversed in the previously described propensity score equation, whereas the second analysis used the average treatment effect of the combined population weights, in which patients who received tebentafusp were assigned a weight of 1 divided by the propensity score, and patients who received ipilimumab plus nivolumab received a weight of 1 divided by 1 minus the propensity score.

After weighting, the median OS and OS curves were estimated using the KM method. The HR and 95% CI were calculated using an IPTW Cox regression model. Variance was calculated using robust sandwich estimation. A sensitivity analysis using a multivariate Cox regression model was conducted.

Patient with missing data for at least 1 relevant covariate were excluded from the primary analysis. A sensitivity analysis using multiple imputation of missing baseline covariate values was performed.

Results

Based on visual inspection of the boxplots, there were no clear weight outliers and no extreme weights. Patient characteristics before and after propensity score weighting are summarized in Table 33. The ESS of the GEM-1402 trial was 34.4 after weighting, compared to a sample size of 45 before weighting. The patient characteristics were generally balanced in the tebentafusp and ipilimumab plus nivolumab cohorts after weighting.

In the primary analysis, median OS of the tebentafusp cohort was 21.7 months (SD = NR), and the weighted median OS of the ipilimumab plus nivolumab cohort was 12.6 months (SD = NR). The HR between tebentafusp and ipilimumab plus nivolumab with respect to OS was 0.430 (95% CI, 0.287 to 0.643) in favour of tebentafusp. A KM plot of OS for the tebentafusp cohort and the weighted ipilimumab plus nivolumab cohort in the primary analysis is shown in Figure 10. Results of the sensitivity analyses were consistent with those of the primary analysis.

Table 33: Patient Characteristics Before and After Weighting (Propensity Score Analysis [IPTW])

ALP = alkaline phosphatase; ECOG PS = Eastern Cooperative Oncology Group Performance Status; ESS = effective sample size; IPTW = inverse probability of treatment weights; LDH = lactate dehydrogenase; SD = standard deviation.

Note: Study 202 data cut-off date: October 13, 2020; GEM-1402 data cut-off date: July 9, 2019.

Source: Sponsor-submitted Propensity Score Analysis Statistical Report.⁴⁷

Critical Appraisal

The propensity score analysis (IPTW) and the MAIC analysis conducted by the sponsor share some common features and, as such, much of the appraisal of the MAIC analysis with respect to study selection, covariate identification, and external validity also applies to the current analysis.

As previously noted, the GEM-1402 trial was identified from an SLR. There is potential for selection bias in the SLR, owing to the lack of a priori selection criteria for the comparator trial. Other methods of the SLR were generally appropriate, with proper literature search, study selection, and data extraction methods in place. The GEM-1402 trial was noted to have a low risk of bias based on a quality assessment conducted by the sponsor; however, a single reviewer was involved in the process with no verification, which increases the risk of error and bias in appraisal.

The sponsor’s justification for an IPTW approach over other propensity score approaches (matching, stratification, and covariate adjustment) was that IPTW is better suited for analyzing a small sample size than matching approaches, and for analyzing time-to-event end points. The sponsor noted that for time-to-event analysis, there is potential bias in KM estimates and marginal HRs with other propensity score approaches.

Figure 10: KM Plot of OS After Weighting (Primary Propensity Score Analysis [IPTW])

IMCgp100 = tebentafusp; Ipi = ipilimumab; IPTW = inverse probability of treatment weighting; KM = Kaplan-Meier; Nivo = nivolumab; OS = overall survival.

Source: Sponsor-submitted Propensity Score Analysis Statistical Report.⁴⁷

The covariates used for propensity score calculation were informed by a multivariate analysis of OS in Study 202 conducted by the sponsor. Because the methods of the multivariate analysis were not available for assessment by the CADTH review team, there is uncertainty as to whether the identified factors were appropriate or comprehensive. Further, Austin (2014)⁴⁸ suggested that the identification of potentially prognostically important covariates should be based on clinical expertise or an SLR of available literature, rather than statistical testing in the study sample. It should also be noted that there is heterogeneity in the HLA status of patients in Study 202 (HLA-A*02:01-positive patients with mUM) and the GEM-1402 trial (patients with mUM regardless of HLA status). It remains unclear to the CADTH review team after a review of the sponsor-submitted supplementary analysis that assessed the impact of HLA status on OS whether HLA status is or is not a possible confounder of OS, given that the analysis was based on a small observational cohort and the results are subject to imprecision. Overall, it is uncertain if all known and unknown confounding factors have been adequately accounted for in the weighting process, which is a key assumption of a propensity score model.⁴⁸

There is a reasonable overlap in the distribution of propensity scores between the treatment groups. The weights were calculated with the IPTW approach using the ATT method in the primary analysis. Additional analyses using alternative methods (ATT stabilized, average treatment effect, and average treatment effect of the control) were conducted, which showed results consistent with the primary analysis. The distribution of generated weights was examined, and no extreme outlying weight was identified. Baseline covariates before and after weighting were reported and were similar between the comparison groups. Some patients were excluded from the primary analysis because of missing data for baseline covariates. The sponsor conducted a sensitivity analysis using multiple imputation, which suggested results similar to the primary analysis. Of note, no formal hypothesis testing was conducted, which renders the trial findings exploratory.

It should also be noted that the analysis was informed with interim OS data from Study 202, which adds uncertainty to the results; there is typically a risk of overestimation of the treatment effects in an interim analysis.¹³

The generalizability of study results is expected to be similar to that of Study 202. The analysis was restricted to patients with mUM who were treatment-naive in the metastatic setting. The findings may have limited generalizability to patients who had received prior systemic therapy, which is a relevant patient population, according to the clinical expert, because tebentafusp will likely be used in previously treated patients. In addition, OS was the only assessed outcome; HRQoL and harms outcomes that were of interest to the stakeholders were not assessed.

Discussion

Summary of Available Evidence

This report summarizes the evidence on tebentafusp based on 1 phase III RCT, 1 ITC, 1 phase I/II study, and 1 propensity score (IPTW-based) analysis.

Study 202, a phase III, open-label RCT, was included in the CADTH Systematic Review. The study aimed to demonstrate the superiority of tebentafusp to investigator’s choice of therapy (pembrolizumab, ipilimumab, or dacarbazine) in HLA-A*02:01-positive adults with mUM who had received no prior therapy in the metastatic setting (N = 378) through the coprimary end points of OS in the RAS and ITT analysis set, and the key secondary end points of PFS and BOR (evaluated statistically as ORR) in the ITT analysis set. In the primary analysis (data cut-off date of October 13, 2020), the mean age of the study population was 62.1 (SD = 11.6) years. Approximately half of the patients were female. The majority of patients were white, had an ECOG PS of 0, and had liver metastases at baseline. Most patients had undergone no prior surgery for metastatic disease (91.3%) and had received no prior antineoplastic systemic treatments (in any setting) (96.3%). Prior radiotherapy was reported in 40.2% of patients.

One sponsor-submitted ITC was summarized and critically appraised. Given the lack of direct comparative evidence between tebentafusp and ipilimumab plus nivolumab, the most commonly prescribed immunotherapy for mUM in Canada, the sponsor performed an unanchored MAIC to estimate the comparative OS and PFS benefits between the interventions in patients with mUM who had received no prior therapy in the metastatic setting. The analysis was informed by an SLR that identified the open-label, single-arm GEM-1402 trial of ipilimumab plus nivolumab as the comparator trial.

Study 102, a phase I/II, open-label study, was also summarized and appraised as supporting evidence. The phase II, single-arm expansion cohort is the only evidence to date for tebentafusp in HLA-A*02:01-positive patients with mUM who had received 1 or 2 prior therapies (chemotherapy, immunotherapy, or targeted therapy) in the metastatic setting (N = 127 in the phase II expansion cohort). The primary end point in the phase II expansion cohort was ORR, and the secondary end points included OS, PFS, DOR, DCR, and safety. At baseline, the mean age of patients was 60.4 (SD = 10.93) years. The majority of patients were white, had an ECOG PS of 0, liver metastasis, and an LDH level that was above the ULN. The majority of patients (66.1%) had received 1 prior anticancer therapy, whereas others had received 2 or more prior anticancer therapies. Of note, 83.5% of patients received prior systemic therapies (73.2% received immunotherapy, 11.0% received chemotherapy, and 8.7% received targeted therapy. Other common prior therapies included radiotherapy (75.6%) and liver-directed therapy (44.9%).

The report also includes an observation study that used propensity score weighting (IPTW approach) to compare OS between tebentafusp and ipilimumab plus nivolumab based on IPD from the same studies that informed the sponsor-submitted MAIC. The study was submitted to CADTH after the sponsor had obtained IPD from the GEM-1402 trial. The study aimed to address the limitations of the MAIC analysis related to the use of aggregate data from the GEM-1402 trial.

Interpretation of Results

Efficacy

In the pivotal trial, Study 202, the coprimary end point of OS in the RAS and OS in the ITT analysis set were met in the first interim OS analysis (data cut-off date of October 13, 2020), supporting the superiority of tebentafusp to investigator’s choice of therapy in the first-line treatment setting for mUM with respect to OS, the most clinically relevant outcome noted by patients and clinicians. The clinical expert consulted by CADTH considered the magnitude of the OS benefits of tebentafusp in the ITT analysis to be clinically meaningful, but noted that a longer follow-up (2 to 3 more years) would help provide further confidence in the OS benefits of tebentafusp. OS analysis in the RAS, which aimed to assess the comparative OS benefits of tebentafusp with investigator’s choice therapy in patients who developed a rash in the first week after treatment with tebentafusp, however, should be interpreted with caution because of the potential for confounding related to the lack of randomization in the comparison, although the extent and direction of any bias cannot be determined. No conclusion could be drawn regarding the subgroup effects of ECOG PS on OS or from the posthoc landmark OS analysis among patients with a BOR of progressive disease, owing to the exploratory nature, lack of control for multiplicity, and the risk of confounding in the absence of randomization in the comparisons.

The key secondary end point of PFS was also met, in favour of tebentafusp. No statistically significant difference was noted in the ORR analysis between treatment arms; however, considering the wide 95% CI of the OR, the results are subject to serious imprecision, rendering the comparative effect of tebentafusp versus investigator’s choice therapy on ORR inconclusive. In addition, tumour response was assessed by investigators using RECIST 1.1 and, considering the open-label design, PFS and ORR, along with other secondary tumour response outcomes (DOR, DCR) might be subject to reporting bias; however, the extent and the direction of the bias is unclear. The clinical expert noted that PFS and ORR outcomes were of limited clinical relevance in the treatment of mUM because there is generally a poor correlation between tumour response and survival in patients with mUM receiving systemic treatments. The DOR in patients who achieved a CR or PR were numerically similar between the treatment arms; however, the difference between treatment arms was not statistically tested, which precludes the drawing of conclusions regarding the effect of tebentafusp on DOR. The DCR analysis suggested results in favour of tebentafusp; however, these results should be interpreted with caution because of the lack of control for multiplicity and, thus, an increased risk of a type I error.

HRQoL, a secondary end point in the study, was noted to be an important outcome of interest by patients and clinicians. The EORTC QLQ-C30 analysis showed that the change in score from baseline at most time points for all domains was stable and numerically similar between treatment arms, based on the commonly used MID estimate of 10 points for cancer patients. Of note, the sponsor considered a LS mean difference of fatigue score between treatment arms of –9.259 points by end of treatment to be statistically significant and clinically significant in the Clinical Study Report, although the MID used was not reported. Given the lack of control for type I error in HRQoL outcomes, the EORTC QLQ-C30 analyses should be considered exploratory. For the same reason, results of the EQ VAS analysis, which suggested a numerically similar change in score from baseline at all time points between treatment arms, should be considered exploratory. It should be noted that there is a significant amount of missing data in HRQoL outcomes, owing to deaths and poor questionnaire completion rates, increasing uncertainty of the results because of potential attrition bias. Also, considering that these instruments have not been validated in patients with mUM and that they are not routinely administered to patients with mUM in clinical practice, per the clinical expert, it is impossible to draw any conclusion about the effect of tebentafusp on HRQoL from available data.

Further, the generalizability of study results is uncertain because the treatments included in the comparator arm account for a small proportion of systemic therapies for mUM prescribed in Canada. Of note, 2 commonly prescribed immunotherapy regimens, ipilimumab plus nivolumab and nivolumab monotherapy, were not included in the comparator arm. The impact on generalizability was likely to be small because, as the clinical expert noted, immunotherapy regimens are similar with respect to tumour response rates; however, this opinion was formed based on a small retrospective cohort. The lack of direct evidence comparing tebentafusp with ipilimumab plus nivolumab and with nivolumab represents an evidence gap in the treatment of mUM.

In an effort to address the evidence gap, the sponsor conducted an MAIC, which suggested that tebentafusp demonstrated OS and PFS benefits over ipilimumab plus nivolumab in patients with mUM who had received no prior therapies in the metastatic setting; however, the results are highly uncertain considering the limitations of the analysis. Specifically, there was a risk of selection bias in the SLR in the absence of a priori selection criteria. The comparator trial included patients with mUM regardless of HLA-A*02:01 status. The heterogeneity in study populations between trials could have potentially confounded the study results. Further, time since primary diagnosis, a covariate identified in the multivariate analysis, was excluded from adjustment because it was not reported in the comparator trial, which may contribute to the uncertainty of the results. There is also concern with a loss of precision in the results, given a significant reduction in the ESS. Overall, the direction of bias of the aforementioned limitations could not be determined, and the results of the MAIC should be interpreted with caution. There is also an evidence gap in that HRQoL and harms outcomes, as well as the efficacy of tebentafusp in treatment-experienced patients, were not addressed in this analysis. The IPTW analysis improved upon the MAIC by using IPD from the comparator trial; however, many of the limitations of the MAIC analysis (risk of selection bias in the SLR, confounding adjustment, and lack of assessment of HRQoL, harms, and efficacy in a treatment-experience population) remained unaddressed. As such, the CADTH review team considered the results of this analysis to be uncertain.

The phase II, single-arm expansion cohort of Study 102 provided noncomparative evidence of tebentafusp treatment in HLA-A*02:01-positive patients with mUM, the majority (83.5%) of whom had received prior systemic therapies. The clinical expert stated that the study demonstrated activity of tebentafusp in the study population and that the results were compatible with the pivotal study, but noted that it is a challenge to draw any definitive conclusions, given the phase II, single-arm study design. The CADTH review team agreed that the noncomparative study design was a key limitation of the study.

Harms

In the pivotal study, tebentafusp is associated with a notably higher frequency of CRS (or presentations related to CRS) and dermatological AEs than investigator’s choice of therapy, with the most common AEs (> 50%) being pyrexia, pruritus, rash, and fatigue in the tebentafusp arm. However, it should be noted that most of the AEs were grade 1 or 2. In addition, the proportion of patients who discontinued treatment due to a TEAE (3.3%) and the proportion of patients who died due to an AE (0.8%) were both low in the tebentafusp arm. The harms of tebentafusp in patients with mUM who had received prior therapies in phase II of Study 102 were consistent with the harms observed in the pivotal study.

Although CRS is a common AE of tebentafusp, in the clinical experience of the clinical expert consulted by CADTH, it is usually transient, and the risk is substantially lower after 3 or 4 infusions. In addition, CRS is generally manageable with monitoring and supportive care (e.g., saline infusion, antihistamine, acetaminophen) in an inpatient setting for initial doses, and admission to the intensive care unit is rarely required, according to the clinical expert. This is consistent with the perspective of patients included in the patient group input, who reported that the side effects of tebentafusp are generally short-term, tolerable, and manageable. The clinical expert added that the recommended dosing of tebentafusp, which involves titration of initial doses, allows patients to build up tolerance to the treatment, and dose adjustments can be implemented to address CRS during the titration process.

Conclusions

Tebentafusp demonstrated a statistically significant and clinically meaningful improvement in OS compared with investigator’s choice of therapy (pembrolizumab, ipilimumab, or dacarbazine) in HLA-A*02:01-positive adults with mUM who had received no prior therapy in the metastatic setting, based on the phase III pivotal study. Tebentafusp resulted in a statistically significant improvement in PFS but not in ORR, although these outcomes are of limited clinical relevance in the treatment of mUM, per the clinical expert. Analyses of other secondary end points showed results in favour of tebentafusp (DCR) or comparable results between treatment arms (DOR, HRQoL), but no conclusions can be drawn because of the lack of statistical testing or control for multiplicity. The comparator treatments account for a small proportion of systemic treatments prescribed for mUM in Canada, which increases uncertainty in the generalizability of the study results. A phase II, single-arm study suggested clinical activity of tebentafusp in patients who had received prior therapy in the metastatic setting, but the results are subject to limitations arising from the noncomparative study design. No definitive conclusion can be drawn from the sponsor-submitted MAIC and propensity score analyses on the relative efficacy of tebentafusp versus ipilimumab plus nivolumab because of significant limitations with the analysis. CRS and dermatological AEs were common with tebentafusp, but were generally grade 1 or 2, transient, and manageable with supportive care.

References

1.Carvajal RD, Harbour JW. Management of metastatic uveal melanoma. In: Post TW, ed. UpToDate. Waltham (MA): UpToDate; 2021: http://www.uptodate.com. Accessed 2022 Jun 2.

2.Ghazawi FM, Darwich R, Le M, et al. Uveal melanoma incidence trends in Canada: a national comprehensive population-based study. Br J Ophthalmol. 2019;103(12):1872-1876. PubMed

3.Damato BE, Dukes J, Goodall H, Carvajal RD. Tebentafusp: T Cell Redirection for the Treatment of Metastatic Uveal Melanoma. Cancers (Basel). 2019;11(7):11. PubMed

4.Khoja L, Atenafu EG, Suciu S, et al. Meta-analysis in metastatic uveal melanoma to determine progression free and overall survival benchmarks: an international rare cancers initiative (IRCI) ocular melanoma study. Ann Oncol. 2019;30(8):1370-1380. PubMed

5.Rantala ES, Hernberg M, Kivela TT. Overall survival after treatment for metastatic uveal melanoma: a systematic review and meta-analysis. Melanoma Res. 2019;29(6):561-568. PubMed

6.Nathan P, Cohen V, Coupland S, et al. Uveal Melanoma UK National Guidelines. Eur J Cancer. 2015;51(16):2404-2412. PubMed

7.NCCN Guidelines. Melanoma: uveal, version 2.2022. Plymouth Meeting (PA): National Comprehensive Cancer Network (NCCN); 2022: https://www.nccn.org/. Accessed 2022 Jun 2.

8.Guideline Resource Unit, Cancer Care Alberta. Uveal Melanoma. (Clinical practice guideline CU-015 – version 2). Edmonton (AB): Alberta Health Services; 2021: https://www.albertahealthservices.ca/assets/info/hp/cancer/if-hp-cancer-guide-cu015-uveal-melanoma.pdf. Accessed 2022 Jun 2.

9.Kimmtrak (tebentafusp): 100 mcg / 0.5 mL solution for intravenous infusion [product monograph]. Dublin (IE): Immunocore Ireland Limited; 2022 Jun 7.

10.Nathan P, Hassel JC, Rutkowski P, et al. Overall Survival Benefit with Tebentafusp in Metastatic Uveal Melanoma. N Engl J Med. 2021;385(13):1196-1206. PubMed

11.Clinical Study Report: IMCgp100-202. A phase II randomized, open-label, multi-center study of the safety and efficacy of IMCgp100 compared with investigator’s choice in HLA-A*0201 positive patients with previously untreated advanced uveal melanoma [internal sponsor's report]. Oxfordshire (GB): Immunocore Ltd; 2021 Apr 13.

12.Lee DW, Gardner R, Porter DL, et al. Current concepts in the diagnosis and management of cytokine release syndrome. Blood. 2014;124(2):188-195. PubMed

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

14.Match-adjusted indirect comparison of tebentafusp vs ipilimumab+nivolumab in patients with previously untreated uveal melanoma statistical report [internal sponsor's report]. In: Drug Reimbursement Review sponsor submission: Kimmtrak (tebentafusp), 100mcg/0.5 mL for intraveous infusion. Dublin (IE): Immunocore; 2021 Dec 7.

15.Piulats JM, Espinosa E, de la Cruz Merino L, et al. Nivolumab Plus Ipilimumab for Treatment-Naive Metastatic Uveal Melanoma: An Open-Label, Multicenter, Phase II Trial by the Spanish Multidisciplinary Melanoma Group (GEM-1402). J Clin Oncol. 2021;39(6):586-598. PubMed

16.TCGA UVM overall survival and metastasis free survival. Patients stratified by possession of HLA*A02:01 allele [internal sponsor's report]. In: Drug Reimbursement Review sponsor submission: Kimmtrak (tebentafusp), 100 mcg / 0.5 mL solution for intravenous infusion, 100mcg/0.5 mL for Intraveous infusion. Clonshaugh (Dublin): Imunocore; 2022 Jul 28.

17.Sacco JJ, Carvajal R, Butler MO, Shoushtari AN, Hassel JC. 64MO - A phase (ph) II, multi-center study of the safety and efficacy of tebentafusp (tebe) (IMCgp100) in patients (pts) with metastatic uveal melanoma (mUM). Ann Oncol. 2020;31:S1442-S1443.

18.Propensity score analysis comparing IMCgp100 or pembrolizumab (from study IMCgp100-202) to ipilimumab+nivolumab (from GEM-1402 study) in previously untreated uveal melanoma patients [internal sponsor's report]. In: Drug Reimbursement Review sponsor submission: Kimmtrak (tebentafusp), 100 mcg / 0.5 mL solution for intravenous infusion, 100mcg/0.5 mL for intraveous infusion. Dublin (IE): Immunocore; 2022 Jun 16.

19.Krantz BA, Dave N, Komatsubara KM, Marr BP, Carvajal RD. Uveal melanoma: epidemiology, etiology, and treatment of primary disease. Clin Ophthalmol. 2017;11:279-289. PubMed

20.Rodriguez A, Duenas-Gonzalez A, Delgado-Pelayo S. Clinical presentation and management of uveal melanoma. Mol Clin Oncol. 2016;5(6):675-677. PubMed

21.Harbour JW, Shih HA. Initial management of uveal and conjunctival melanomas. In: Post TW, ed. UpToDate. Waltham (MA): UpToDate; 2022: http://www.uptodate.com. Accessed 2022 May 10.

22.Sato T. Locoregional management of hepatic metastasis from primary uveal melanoma. Semin Oncol. 2010;37(2):127-138. PubMed

23.Drug Reimbursement Review sponsor submission: Kimmtrak (tebentafusp), 100 mcg / 0.5 mL solution for intravenous infusion [internal sponsor's package]. Toronto (ON): Medison Pharma Canada Inc.; 2022 Apr 20.

24.Keytruda (pembrolizumab): solution for infusion 100 mg/4 mL vial [product monograph]. Kirkland (QC): Merck Canada Inc; 2022 May 4. Accessed 2022 Jun 2.

25.Opdivo (nivoluman): intravenous solution (10mg/mL) 40 and 100 mg single-use vials [product monograph]. Montreal (QC): Bristol-Myers Squibb Canada Co; 2022 Feb 28. Accessed 2022 Jun 2.

26.Yervoy (ipilimumab): intravenous infusion, 5 mg/mL in 10 mL and 40 mL vials [product monograph]. Montreal (QC): Bristol-Myers Squibb Canada Co; 2022 March 3. Accessed 2022 Jun 2.

27.Clinical Study Report: IMCgp100-102. A phase 1/2, open-label, multi-center study of the safety and efficacy of IMCgp100 using the intra-patient escalation dosing regimen in patients with advanced uveal melanoma [internal sponsor's report]. Oxfordshire (GB): Immunocore Ltd; 2021 Feb 8.

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

29.Grey matters: a practical tool for searching health-related grey literature. Ottawa (ON): CADTH; 2019: https://www.cadth.ca/grey-matters. Accessed 2022 Jul 07.

30.Center for Drug Evaluation Research. Multidiscipline review(s). Kimmtrak (tebentafusp) intravenous infusion. Company: Immunocore, LLC. Application No.: 761228 . Approval date: 01/25/2022 (FDA approval package). Silver Spring (MD): U.S. Food and Drug Administration (FDA); 2022: https://www.accessdata.fda.gov/drugsatfda_docs/nda/2022/761228Orig1s000MultidisciplineR.pdf. Accessed 2022 Jun 23.

31.Fayers PM, Aaronson NK, Bjordal K, et al. The EORTC QLC-C30 scoring manual (3rd Edition). Brussels (BE): European Organisation for Research and Treatment of Cancer (EORTC); 2001: https://www.eortc.org/app/uploads/sites/2/2018/02/SCmanual.pdf. Accessed 2021 Apr 23.

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

33.Snyder CF, Blackford AL, Sussman J, et al. Identifying changes in scores on the EORTC-QLQ-C30 representing a change in patients' supportive care needs. Qual Life Res. 2015;24(5):1207-1216. PubMed

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

35.Herdman M, Gudex C, Lloyd A, et al. Development and preliminary testing of the new five-level version of EQ-5D (EQ-5D-5L). Qual Life Res. 2011;20(10):1727-1736. PubMed

36.Pickard AS, Neary MP, Cella D. Estimation of minimally important differences in EQ-5D utility and VAS scores in cancer. Health Qual Life Outcomes. 2007;5:70. PubMed

37.Statistical analysis plan: a phase II randomized, open-label, multi-center study of the safety and efficacy of IMCgp100 compared with investigator choice in HLA-A*0201 positive patients with previously untreated advanced uveal melanoma [internal sponsor's report] In: Drug Reimbursement Review sponsor submission: Kimmtrak (tebentafusp), 100 mcg / 0.5 mL solution for intravenous infusion [internal sponsor's package]. Toronto (ON): Medison Pharma Canada Inc. 2020 Oct 28.

38.Kelly D, Rose AAN, Muniz TP, et al. Development of a Metastatic Uveal Melanoma Prognostic Score (MUMPS) for Use in Patients Receiving Immune Checkpoint Inhibitors. Cancers (Basel). 2021;13(14). PubMed

39.Tebentafusp for the treatment of HLA-A*02:01-positive metastatic uveal melanoma - clinical systematic literature review [internal sponsor's report]. In: Drug Reimbursement Review sponsor submission: Kimmtrak (tebentafusp), 100mcg/0.5 mL for intraveous infusion. Abingdon (GB): Immunocore; 2021 Oct 19.

40.Signorovitch JE, Sikirica V, Erder MH, et al. Matching-adjusted indirect comparisons: a new tool for timely comparative effectiveness research. Value Health. 2012;15(6):940-947. PubMed

41.Signorovitch JE, Wu EQ, Yu AP, et al. Comparative effectiveness without head-to-head trials: a method for matching-adjusted indirect comparisons applied to psoriasis treatment with adalimumab or etanercept. Pharmacoeconomics. 2010;28(10):935-945. PubMed

42.Phillippo D, Ades T, Dias S, Palmer S, Abrams K, Welton N. NICE DSU Technical Support Document 18: Methods for population-adjusted indirect comparisons in submissions to NICE. Vol 18. Sheffield (UK): Decision Support Unit, ScHARR, University of Sheffield; 2016: https://research-information.bris.ac.uk/ws/portalfiles/portal/94868463/Population_adjustment_TSD_FINAL.pdf. Accessed 2022 Jun 24.

43.Pelster MS, Gruschkus SK, Bassett R, et al. Nivolumab and Ipilimumab in Metastatic Uveal Melanoma: Results From a Single-Arm Phase II Study. J Clin Oncol. 2021;39(6):599-607. PubMed

44.Liu J, Lichtenberg T, Hoadley KA, et al. An Integrated TCGA Pan-Cancer Clinical Data Resource to Drive High-Quality Survival Outcome Analytics. Cell. 2018;173(2):400-416 e411. PubMed

45.Kim H, Gurrin L, Ademi Z, Liew D. Overview of methods for comparing the efficacies of drugs in the absence of head-to-head clinical trial data. Br J Clin Pharmacol. 2014;77(1):116-121. PubMed

46.Immunocore. NCT02570308: A phase I/II open-label, multi-center study of the safety and efficacy of IMCgp100 using the intra-patient escalation dosing regimen in patients with advanced uveal melanoma. ClinicalTrials.gov. Bethesda (MD): U.S. National Library of Medicine; 2021: https://clinicaltrials.gov/ct2/show/study/NCT02570308. Accessed 2022 Jul 4.

47.Propensity score analysis comparing IMCgp100 (from study IMCgp100-202) to ipilimumab+nivolumab (from GEM-1402 study) in previously untreated uveal melanoma patients statistical report [internal sponsor's report]. In: Drug Reimbursement Review sponsor submission: Kimmtrak (tebentafusp), 100 mcg / 0.5 mL solution for intravenous infusion, 100mcg/0.5 mL for intraveous infusion. Dublin (IE): Immunocore; 2022 Jun 16.

48.Austin PC. The use of propensity score methods with survival or time-to-event outcomes: reporting measures of effect similar to those used in randomized experiments. Stat Med. 2014;33(7):1242-1258. PubMed

49.Luo N, Fones CS, Lim SE, Xie F, Thumboo J, Li SC. The European Organization for Research and Treatment of Cancer Quality of Life Questionnaire (EORTC QLQ-c30): validation of English version in Singapore. Qual Life Res. 2005;14(4):1181-1186. PubMed

50.Davda J, Kibet H, Achieng E, Atundo L, Komen T. Assessing the acceptability, reliability, and validity of the EORTC Quality of Life Questionnaire (QLQ-C30) in Kenyan cancer patients: a cross-sectional study. J Patient Rep Outcomes. 2021;5(1):4. PubMed

51.Richardson J, Iezzi A, Khan MA, Chen G, Maxwell A. Measuring the Sensitivity and Construct Validity of 6 Utility Instruments in 7 Disease Areas. Med Decis Making. 2016;36(2):147-159. PubMed

52.McClure NS, Sayah FA, Xie F, Luo N, Johnson JA. Instrument-Defined Estimates of the Minimally Important Difference for EQ-5D-5L Index Scores. Value Health. 2017;20(4):644-650. PubMed

53.Giesinger JM, Kieffer JM, Fayers PM, et al. Replication and validation of higher order models demonstrated that a summary score for the EORTC QLQ-C30 is robust. J Clin Epidemiol. 2016;69:79-88. PubMed

54.Herdman M, Kerr C, Pavesi M, et al. Testing the validity and responsiveness of a new cancer-specific health utility measure (FACT-8D) in relapsed/refractory mantle cell lymphoma, and comparison to EQ-5D-5L. J Patient Rep Outcomes. 2020;4(1):22. PubMed

Appendix 1: Literature Search Strategy

Note that this appendix has not been copy-edited.

Clinical Literature Search

Overview

Interface: Ovid

Databases:

• MEDLINE All (1946-present)

• Embase (1974-present)

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

Date of search: May 18, 2022

Alerts: Bi-weekly search updates until project completion

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

Limits:

• Humans

• Conference abstracts: excluded

Table 34: Syntax Guide

Multidatabase Strategy

1. (Kimmtrak* or tebentafusp* or imcgp100* or imcgp 100* or immtac gp* or WHO10664 or WHO 10664 or N658GY6L3E).ti,ab,kf,ot,hw,nm,rn.

2. 1 use medall

3. *tebentafusp/

4. (Kimmtrak* or tebentafusp* or imcgp100* or imcgp 100* or immtac gp* or WHO10664 or WHO 10664).ti,ab,kf,dq.

5. 3 or 4

6. 5 use oemezd

7. (conference review or conference abstract).pt.

8. 6 not 7

9. 2 or 8

10. remove duplicates from 9

Clinical Trials Registries

ClinicalTrials.gov

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

Search terms -- Kimmtrak, tebentafusp]

WHO ICTRP

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

[Search terms -- Kimmtrak, tebentafusp]

Health Canada’s Clinical Trials Database

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

[Search terms -- Kimmtrak, tebentafusp]

EU Clinical Trials Register

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

[Search terms -- Kimmtrak, tebentafusp]

Grey Literature

Search dates: May 1-15, 2022

Keywords: Kimmtrak, tebentafusp, melanoma

Limits: none

Updated: Search updated prior to the meeting of CADTH pan-Canadian Oncology Drug Review Expert Committee (pERC).

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

• Health Technology Assessment Agencies

• Health Economics

• Clinical Practice Guidelines

• Drug and Device Regulatory Approvals

• Advisories and Warnings

• Drug Class Reviews

• Clinical Trials Registries

• Databases (free)

• Health Statistics

• Internet Search

• Open Access Journals.

Appendix 2: Excluded Studies

Note that this appendix has not been copy-edited.

Table 35: Excluded Studies

Appendix 3: Detailed Outcome Data

Note that this appendix has not been copy-edited.

Table 36: Subgroup Analysis of OS by ECOG PS (ITT) - Primary Analysis (Data Cut-Off Date of October 13, 2020)

CI = confidence interval; ECOG = Eastern Cooperative Oncology Group; ITT = intention to treat.

^aBased on Kaplan-Meier estimate.

^bEstimated using the Cox proportional hazards model with Efron’s approach for handling ties adjusted for treatment, subgroup, and treatment-by-subgroup interaction term.

^cP value < 0.1 suggests there is an interaction between treatment and the subgroup. The p value was not adjusted for multiplicity.

Source: Study 202 Clinical Study Report¹¹

Table 37: Sensitivity Analyses of PFS Outcomes (ITT) — Final Analysis (Data Cut-Off Date of October 13, 2020)

CI = confidence interval; ITT = intention to treat; PFS = progression-free survival.

^aBased on Kaplan-Meier estimate.

^bHazard ratio was estimated using a Cox proportional hazards model stratified by LDH status.

^cP value based on log-rank test of the Kaplan-Meier curve stratified by LDH status. P value was not adjusted for multiplicity.

Source: Study 202 Clinical Study Report.¹¹

Table 38: Subgroup Analysis of PFS by ECOG PS (ITT Analysis Set) — Final Analysis (Data Cut-Off Date of October 13, 2020)

CI = confidence interval; ECOG = Eastern Cooperative Oncology Group; ITT = intention to treat.

^aBased on Kaplan-Meier estimate.

^bEstimated using the Cox proportional hazards model with Efron’s approach for handling ties adjusted for treatment, subgroup, and treatment-by-subgroup interaction term.

Source: Study 202 Clinical Study Report.¹¹

Table 39: EQ-5D-5L Descriptive System Results (ITT) — Data Cut-Off Date of October 13, 2020