Reimbursement Review

Erdafitinib (Balversa)

Sponsor: Janssen Inc.

Therapeutic area: Locally advanced unresectable or metastatic urothelial carcinoma

This multi-part report includes:

Clinical Review

TPA Review

Pharmacoeconomic Review

Clinical Review

Abbreviations

Executive Summary

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

Table 1: Background Information of Application Submitted for Review

FGFR = fibroblast growth factor receptor; NOC = Notice of Compliance.

Introduction

Urothelial carcinoma (UC) is a malignant transformation of urothelial cells, primarily affecting the bladder, and accounts for 90% to 95% of bladder cancer cases. In 2023, Canada saw an estimated 13,400 new bladder cancer cases, with 2,600 deaths expected annually. The most common symptom of bladder cancer is hematuria, although other symptoms, such as pain during urination, abdominal pain, and fatigue, may also occur. The cancer predominantly affects individuals older than 50 years, and there is a higher incidence in males and in those with risk factors like smoking, chemical exposure, or a family history of bladder cancer. Diagnosis is typically made through cystoscopy and biopsy; most patients are diagnosed with non–muscle-invasive bladder cancer (NMIBC), although a significant portion may progress to muscle-invasive bladder cancer (MIBC) or metastatic disease. Fibroblast growth factor receptors (FGFRs) play a crucial role in UC, and abnormalities in the FGFR3 gene are linked to up to 42% of UC cases.

Treatment for locally advanced unresectable or metastatic urothelial carcinoma (la/mUC) focuses on slowing disease progression, extending life, and improving quality of life. The standard first-line treatment is cisplatin-based chemotherapy, although platinum-based chemotherapy is generally preferred. For patients who are ineligible for platinum-based chemotherapy, immune checkpoint inhibitors (ICIs), like pembrolizumab and avelumab, are offered as an alternative first-line option and are commonly used as second-line treatment after progression on chemotherapy. Newer therapies, such as enfortumab vedotin (Padcev) and erdafitinib (Balversa), are also part of the treatment landscape, particularly for patients with specific genetic alterations. Despite these therapies, la/mUC remains largely incurable, with a poor prognosis.

Erdafitinib, a targeted therapy, inhibits FGFRs and is used in cases of la/mUC with FGFR3 alterations. It has been approved by Health Canada for the for the treatment of adult patients with locally advanced unresectable or metastatic UC, harbouring susceptible FGFR3 genetic alterations, who have disease progression during or following at least 1 line of prior therapy, including within 12 months of neoadjuvant or adjuvant therapy. Erdafitinib should not be used for the treatment of patients who are eligible for and have not received prior PD-1 or PD-L1 inhibitor therapy. Treatment with erdafitinib should be initiated following confirmation of a susceptible FGFR genetic alteration using a validated test.

The objective of this report is to review and critically appraise the evidence submitted by the sponsor on the beneficial and harmful effects of erdafitinib tablets — 3 mg, 4 mg, and 5 mg — administered orally for the treatment of adult patients with la/mUC harbouring susceptible FGFR3 genetic alterations.

Patient, Clinician, and Drug Program Perspectives

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

Patient Input

One patient group, Bladder Cancer Canada (BCC), provided input for this CDA-AMC review. BCC is a registered national charity that serves those facing a bladder cancer diagnosis. The information provided by BCC was collected with an online survey, which asked about the impact of FGFR3 metastatic UC on the lives of patients, the effect of current treatments, and patient experience with erdafitinib. The online survey was completed by 4 people (identified by BCC as patients A, B, C, and D), all of whom were from Canada and all of whom had la/mUC with an FGFR3 mutation. Of the 2 survey respondents (patients A and B) who had treatment experience with erdafitinib, 1 agreed to participate in a telephone interview to discuss their survey responses.

The most common cancer symptoms reported by respondents were fatigue, insomnia, neuropathy, and decreased mobility. Survey responses suggested that the cancer symptoms of participants were not adequately managed with current therapies. Fatigue, neuropathy, and hair loss were the most commonly reported side effects of treatments. One patient-reported difficulty in accessing treatment was related to travel time. Regarding respondent willingness to tolerate new side effects from drugs that can control disease progression or improve overall survival (OS), the average score was 6.25 on a scale of 1 (will not tolerate side effects) to 10 (will tolerate significant side effects), with scores ranging from 4 to 10.

Of the respondents who were treated with erdafitinib, patient A had completed a full course of treatment and patient B had been receiving erdafitinib for 6 weeks. When these patients were asked to rate how their lives had changed with erdafitinib in terms of certain categories (metastatic cancer symptoms, drug side effects, maintaining quality of life, controlling disease progression, and preventing recurrence), compared to other therapies they had received, scores generally suggested that neither respondent experienced a major difference with erdafitinib. Reported side effects from erdafitinib were dysgeusia (patient A) and dry mouth, nausea, and leg pain (patient B). Tolerability of the side effects of erdafitinib, on a scale of 1 (completely tolerable) to 10 (completely intolerable), was rated as 9 by patient A and 3 by patient B. Both patients reported that taking erdafitinib orally made their treatment easier and both indicated that they would recommend erdafitinib to other patients with bladder cancer.

Clinician Input

Input From Clinical Experts Consulted by CDA-AMC

Clinical experts consulted by CDA-AMC for the review of erdafitinib in the treatment of la/mUC identified several key considerations. For unmet needs, one of the most significant challenges is determining the optimal sequencing of erdafitinib and enfortumab vedotin for patients with FGFR alterations. Although enfortumab vedotin is currently the standard treatment after progression on platinum-based chemotherapy and ICIs, it may not be suitable for all patients because of its associated toxicities. Additionally, the anticipated approval of enfortumab vedotin used in combination with pembrolizumab as a first-line therapy raises concerns about the lack of subsequent treatment options, highlighting another unmet need.

Erdafitinib is expected to be used as monotherapy after platinum-based chemotherapy and ICI therapy, particularly in patients with FGFR alterations, as supported by the pivotal THOR trial. However, clinical experts noted that the ideal sequencing of enfortumab vedotin and erdafitinib remains uncertain, which could impact treatment decisions.

The target population for erdafitinib includes patients with tumours harbouring FGFR alterations, necessitating genetic testing to identify these alterations. Experts agreed that la/mUC is not a rare disease and that the patient population studied in clinical trials is consistent with the intended target population for this therapy.

In terms of treatment assessment, the experts recommended monitoring clinical symptoms and using imaging studies to evaluate disease progression. Treatment discontinuation should be considered in cases of disease progression, significant toxicity, or intolerability, with decisions made on a case-by-case basis, following the criteria used in clinical trials. The experts also emphasized that the management of diagnosis, treatment, and ongoing patient monitoring should be handled by oncology specialists, including in outpatient settings, to ensure the best possible care for patients.

Clinician Group Input

Seven clinicians from the Ontario Health (Cancer Care Ontario) Genitourinary Cancer Drug Advisory Committee (GU DAC) provided input for this review. Ontario Health (Cancer Care Ontario) Drug Advisory Committees provide timely evidence-based clinical and health system guidance on drug-related issues, including the Provincial Drug Reimbursement Programs and the Systemic Treatment Program, in support of Cancer Care Ontario’s mandate.

In communicating current treatments for la/mUC, the GU DAC noted that patients who have previously received an ICI, chemotherapy, or the combination of both are eligible for treatment with enfortumab vedotin, and the goal of treatment is to improve OS. The GU DAC expressed an unmet need for a treatment for patients with genomic alterations, noting that erdafitinib is effective for the FGFR genetic alteration and would be the first targeted therapy identified for this patient population based on molecular testing. Regarding place in therapy, the GU DAC indicated that patients who would be eligible and best suited for treatment with erdafitinib are those with FGFR mutations and/or alterations who have previously received or who have a contraindication to ICI therapy. The GU DAC stated that erdafitinib would be administered in the outpatient setting under the supervision of a medical oncologist. Patient response to treatment is assessed in clinical practice using conventional imaging (CT scan of the chest, abdomen, and/or pelvis), at the physician’s discretion. Discontinuation of treatment with erdafitinib would be considered when the patient experiences unacceptable toxicity or clinically significant disease progression.

Drug Program Input

The drug programs provided input and posed several questions regarding the implementation of erdafitinib in patients with la/mUC harbouring FGFR alterations. These questions aim to clarify the drug's place in therapy, the appropriate patient population, and considerations for its use.

In response to the question about relevant comparators from the THOR clinical trial, the clinical experts agreed that a platinum-doublet chemotherapy regimen serves as an appropriate comparator for erdafitinib in fit patients with FGFR alterations who experience disease progression on first-line enfortumab vedotin with pembrolizumab (if recommended and funded). Furthermore, the drug programs also noted that single-drug chemotherapy with docetaxel or paclitaxel would likely be the next line of therapy in patients who experience disease progression on first-line enfortumab vedotin with pembrolizumab, validating the choice of comparators. This aligns with the standard treatment protocols currently in use for la/mUC, again validating the choice of comparators in the trial.

When asked about the initiation of erdafitinib therapy, particularly in patients with different histologic subtypes of UC, the experts noted that erdafitinib should be considered specifically for patients who have been treated previously with platinum-based chemotherapy and an ICI and have shown disease progression. This underscores the importance of considering FGFR alterations as a critical factor when determining the suitability of erdafitinib for a patient. The experts recommended starting erdafitinib after other standard treatments, and emphasized its role as a subsequent line of therapy, particularly in patients with FGFR alterations.

Clinical Evidence

Systematic Review

Description of Studies

One pivotal study was included in this submission. The THOR study was a global, phase III, randomized controlled trial (RCT) designed to evaluate the efficacy and safety of erdafitinib compared to standard chemotherapy in patients with la/mUC who had specific FGFR3 and/or FGFR2 genetic alterations and had progressed after treatment with anti-PD-1 or anti-PD-L1 checkpoint inhibitors. The trial enrolled 266 patients, who were randomly assigned in a 1:1 ratio to receive either erdafitinib (n = 136) or the investigator’s choice of chemotherapy (docetaxel or vinflunine; n = 130). The primary outcomes of interest were OS and progression-free survival (PFS), with secondary outcomes including objective response rate (ORR) and safety profiles. The pivotal trial had 2 cohorts; the population of interest for this submission is cohort 1 (i.e., those previously treated with an anti-PD-L1 drug).

The baseline characteristics of the study population were well balanced between the 2 treatment arms. The median age of patients was approximately 67 years, and a majority were male. Most patients had an Eastern Cooperative Oncology Group Performance Status Performance Status (ECOG PS) of 0 or 1, and all had previously received at least 1 line of therapy, including anti-PD-1 or anti-PD-L1 drugs. The genetic profile of the patients, defined by FGFR3 and/or FGFR2 alterations, was consistent across both groups, ensuring comparability when evaluating the treatment effects of erdafitinib and chemotherapy.

Of note, the Independent Data Monitoring Committee (IDMC) recommended that cohort 1 be stopped early because of the superiority of erdafitinib treatment over chemotherapy.

Efficacy Results

Overall Survival

The primary end point of the THOR trial was OS, which was deemed critical for decision-making by clinical experts consulted by CDA-AMC. The median OS was 12.06 months (95% confidence interval [CI], 10.28 to 16.36 months) in the erdafitinib group and 7.79 months (95% CI, 6.54 to 11.07 months) in the chemotherapy group. The estimated 6-month survival rates were 85% for erdafitinib and 66% for chemotherapy, a difference of █████ ████ ███ ████ █████, whereas the 24-month rates were 26% and 20%, respectively, with a difference of ████ ████ ███ █████ █████. The OS analysis demonstrated that erdafitinib significantly reduced the time to death compared to chemotherapy (hazard ratio [HR] = 0.64; P = 0.005). Subgroup analyses were not designed to establish treatment effects within specific subgroups.

Progression-Free Survival

PFS was a secondary end point in the THOR study. Median PFS was 5.55 months (95% CI, 4.40 to 5.65 months) for erdafitinib and 2.73 months (95% CI, 1.81 to 3.68 months) for chemotherapy. The estimated 6-month PFS rates were higher in the erdafitinib group (37%) than in the chemotherapy group (27%), a difference (absolute effect) of ████ ████ ███ ███ █████, with both groups showing low PFS rates by 24 months. These absolute effects in PFS at 6 months and 24 months presented imprecise CIs (i.e., no evidence of a meaningful difference between erdafitinib and chemotherapy). However, the overall relative effects, measured with a Cox proportional hazards (PH) regression model analysis, showed an improvement in PFS for erdafitinib against chemotherapy, with a HR of 0.58 (95% CI, 0.44 to 0.78; P = 0.0002).

Objective Response Rate

The ORR, defined as the proportion of patients achieving a complete response (CR) or a partial response (PR), was significantly higher in the erdafitinib group, with 45.6% of patients responding, than in the chemotherapy group, with 11.5% of patients responding. An absolute difference of ███ ████ ███ █████ █████. The risk ratio (RR) of achieving an objective response was almost 4 times higher with erdafitinib as with chemotherapy (RR = 3.94; 95% CI, 2.37 to 6.57).

Duration of Response

The median duration of response (DoR) was 4.86 months (95% CI, 3.84 to 7.46 months) for erdafitinib and 5.55 months (95% CI, 2.14 to 6.01 months) for chemotherapy. The results suggest that although erdafitinib is more effective in inducing responses, the duration of these responses may be comparable to that of chemotherapy. For instance, the 6-month probability of remaining in response was 42.0% in the erdafitinib group (95% CI, 0.29% to 0.55%) and 32.0% in the chemotherapy group (95% CI, 10.0% to 57.0%), a difference of █████ ████ ███ ██████ █████. The overall HR was 0.85 (95% CI, 0.43 to 1.66) for erdafitinib compared to chemotherapy. However, the DoR analysis was based on a small number of patients, especially in the chemotherapy group, which can lead to imprecision in the estimates.

Patient-Reported Outcomes (Health-Related Quality of Life)

Patient-reported outcomes (PROs) were assessed to evaluate the impact of treatments on health-related quality of life (HRQoL). Baseline HRQoL scores were similar between in the 2 treatment groups, and compliance with HRQoL assessments remained high through the early treatment cycles but declined in later cycles because of disease progression and death. Across all PRO measures — including the Functional Assessment of Cancer Therapy–Bladder (FACT-Bl), the EQ-5D-5L, and the Patient Global Impression of Severity (PGI-S) — there were no significant differences in HRQoL between erdafitinib and chemotherapy. Both treatment groups maintained general HRQoL and overall health status throughout the study, suggesting that although erdafitinib improves survival outcomes, it does not lead to a substantial difference in PROs compared to chemotherapy.

Harms Results

The THOR study revealed that both erdafitinib and chemotherapy were associated with a high incidence of adverse events (AEs), with nearly all patients experiencing at least 1 AE. For instance, at least 1 AE was reported in 133 of 135 patients (98.5%) in the erdafitinib group and 109 of 112 patients (97.3%) in the chemotherapy group. The most common AEs (i.e., reported by at least 10% of patients) included hyperphosphatemia (80.0%), diarrhea (62.2%), and stomatitis (48.1%) in the erdafitinib group, and anemia (32.1%), constipation (27.7%), and asthenia (25.0%). in the chemotherapy group

At least 1 serious adverse event (SAE) was reported in 56 patients (41.5%) in the erdafitinib treatment group and 47 patients (41.9%) in the chemotherapy group. The most frequently reported SAEs (by at least > 2% of patients) were urinary tract infection (4.4%) and hematuria (3.7%) in the erdafitinib group, and febrile neutropenia (6.3%) and febrile bone marrow aplasia (3.6%) in the chemotherapy group.

Among AEs of special interest, central serous retinopathy (CSR) was a notable AE specific to erdafitinib, occurring in 23 (17.0%) patients (compared to 0 patients in the chemotherapy group), which necessitated regular ophthalmologic monitoring because of the potential for vision impairment. Hyperphosphatemia, another event of interest, occurred in 108 of 135 patients (80.0%) in the erdafitinib group and 0 patients in the chemotherapy group. Last, nail and skin disorders were deemed worthy of attention, with nail disorders reported in 90 patients (66.7%) in the erdafitinib group and in 6 patients (5.4%) in the chemotherapy group. Similarly, skin disorders were reported in 74 patients (54.8%) and 14 patients (12.5%), respectively.

The rates of treatment discontinuation due to AEs were slightly lower in the erdafitinib group (14.1%) than in the chemotherapy group (17.9%).

Critical Appraisal

The THOR study was a well-designed RCT comparing erdafitinib to chemotherapy in patients with la/mUC harbouring FGFR alterations who have previously been treated with anti-PD-1 or anti-PD-L1 therapies. The randomization process was properly conducted, although some imbalances, such as the difference in the number of patients declining chemotherapy, were noted. Despite these minor issues, most baseline characteristics were balanced. The open-label design could potentially introduce bias, particularly in subjective measures. The exploratory subgroup analyses were not predefined, raising concerns about potential type I errors, although these analyses were not the primary focus of the study and there was no evidence of imbalance among subgroups (i.e., evidence suggesting subgroup effects).

In terms of external validity, the THOR study was conducted in 23 countries, which enhances its generalizability, although the underrepresentation of certain demographic groups, particularly Black patients, could limit its applicability in multicultural settings like Canada. According to the clinical experts consulted for this review, the median participant age aligns with the typical age range for patients with UC in Canada, supporting the relevance of the findings. The focus on patients with FGFR alterations highlights the importance the availability of molecular testing for the generalizability of the results. The chemotherapy options used in the study are consistent with those available in Canada, further supporting the study’s external validity. Overall, the clinical experts believe that the study's findings are applicable to most patients with la/mUC in Canada who meet the specified criteria.

GRADE Summary of Findings and Certainty of the Evidence

For pivotal studies and RCTs identified in the sponsor’s systematic review, the Grading of Recommendations Assessment, Development and Evaluation (GRADE) approach was used to assess the certainty of the evidence for outcomes considered most relevant to CDA-AMC expert committee deliberations, and a final certainty rating was determined as outlined by the GRADE Working Group.

Table 2: Summary of Findings for Erdafitinib Versus Chemotherapy for Patients With la/mUC Harbouring an FGFR Alteration

AE = adverse event; CI = confidence interval; CR = complete response; FACT-Bl = Functional Assessment of Cancer Therapy–Bladder; HRQoL = health-related quality of life; la/mUC = locally advanced unresectable or metastatic urothelial carcinoma; PGI-S = Patient Global Impression of Severity; PR = partial response; RCT = randomized controlled trial; RR = risk ratio; SAE = serious adverse event; VAS = visual analogue scale.

^aThe 95% CI excludes the null and a conservative threshold of 20 patients per 1,000 treated; the sample size or optimal information size (in the THOR study calculated for OS) was not reached (N = 280) before the study was stopped early for benefit. Therefore, the estimate was rated down 1 level for imprecision.

^bThe 95% CI includes the null but also a conservative threshold of 20 patients per 1,000 treated (for benefit or harm). Furthermore, the sample size was considered small for a conservative estimate of an optimal information size of 280. Therefore, the imprecision domain was rated down 2 levels.

^cNo statistical tests were performed. However, the number of patients in these analyses tend to decrease over the assessment period, decreasing the sample size. No evidence of difference was detected, with wide CIs in the estimates at the end of treatment for both arms. Therefore, 2 levels were rated down for imprecision.

^dNo statistical tests were performed. The difference in effect between the 2 groups was considered very large and certain based on input from clinical experts, so it was not rated down for imprecision.

Source: Sponsor’s Summary of Clinical Evidence.

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

The GRADE assessments are presented in Table 2. The selection of outcomes for the GRADE approach was based on the sponsor’s Summary of Clinical Evidence, consultation with clinical experts, and input received from patient group, clinician groups, and public drug plans. The following list of outcomes was finalized in consultation with expert committee members: OS, PFS, ORR, DoR, HRQoL, and harms.

Long-Term Extension Studies

No long-term extension studies were available for this submission.

Indirect Comparisons

Description of Studies

The sponsor submitted an indirect treatment comparison (ITC) designed to assess the efficacy and safety of erdafitinib relative to enfortumab vedotin in patients with UC who had progressed after 1 or 2 prior treatments, including at least 1 anti-PD-L1 drug. Because of the lack of direct head-to-head evidence between erdafitinib and enfortumab vedotin, an anchored matching-adjusted indirect comparison (MAIC) was performed. This approach allowed for indirect comparisons while adjusting for differences in measured baseline characteristics between the trials. To estimate the relative efficacy and safety of these treatments, the authors used individual patient data from 1 pivotal study (THOR) that assessed erdafitinib and aggregate data from 1 available study (EV-301 trial) that evaluated enfortumab vedotin. A Bayesian approach was used, but the Bucher method was used to complete the sensitivity analysis.

The MAIC produced estimates comparing erdafitinib to enfortumab vedotin in the EV-301 trial population, incorporating additional eligibility criteria that excluded patients with no prior exposure to platinum-based chemotherapy, patients with an ECOG PS of 2, and those who had received more than 1 prior chemotherapy. The population-matching process reduced the effective sample size (ESS) from 197 to 126, a 36% reduction, which was deemed adequate to support comparisons across all efficacy and safety outcomes.

Efficacy Results

In the base-case analyses, the effect estimates for OS, PFS, ORR, and CR showed wide credible intervals (CrIs), reflecting an inability to determine whether one treatment is superior to the other for any of these end points.

In terms of survival end points, the postadjustment HR for OS with erdafitinib was 0.92 (95% CrI, 0.54 to 1.57) compared to enfortumab vedotin. Although the CrIs widened after the adjustment, this was consistent with the observed reduction in ESS.

Similar effects were observed for PFS, where the postadjustment HR for erdafitinib was 0.93 (95% CrI, 0.55 to 1.56).

For the ORR, the RR was 1.49 (95% CrI, 0.56 to 3.90), and for CR, it was 2.89 (95% CrI, 0.27 to 30.33). Despite these increases, the wide CrIs reflect substantial uncertainty, particularly for CR. Sensitivity analyses, adjusted for each covariate cumulatively, confirmed the consistency of the Bayesian and Bucher estimates across all efficacy outcomes.

When matched patients in the THOR study were compared to the intent-to-treat (ITT) population and to those receiving 1 or 2 prior lines of therapy in the EV-301 study, similar results were observed for OS and PFS.

Harms Results

For harms, the RR for any AE was 1.02 (95% CrI, 0.98 to 1.06). For the remaining AEs, erdafitinib demonstrated a safety profile comparable to that of enfortumab vedotin, with an RR close to 1 in most cases.

Critical Appraisal

Overall, the results from the ITC suggest substantial uncertainty about whether erdafitinib and enfortumab vedotin differ meaningfully in terms of survival, response outcomes, or harms. This uncertainty is primarily attributed to the imprecision observed in the CrIs of the effect estimates.

Conclusions

Evidence from the pivotal THOR study shows that in patients with la/mUC harbouring FGFR3 and/or FGFR2 alterations, erdafitinib improves OS, PFS, and the overall response rate, compared with chemotherapy. These benefits come with specific, although clinically manageable, AEs, such as hyperphosphatemia and CSR. Although the study provides robust evidence of efficacy, there are gaps in the available data, particularly regarding long-term safety and the effects in specific populations, such as those with a poorer performance status or extensive prior chemotherapy exposure. From a clinical perspective, the benefits of erdafitinib align with patient needs and expectations for new treatments that prolong survival.

There is a lack of direct comparative data on the clinical effects and position of erdafitinib against other emerging relevant comparators in the treatment landscape of la/mUC in Canada. The sponsor-submitted ITC, which used a MAIC, assessed the efficacy and safety of erdafitinib relative to enfortumab vedotin using data from the THOR and EV-301 trials. The results showed adjusted HRs for OS and PFS with wide CrIs, indicating substantial imprecision, making it difficult to determine whether erdafitinib or enfortumab vedotin offers a superior outcome. Although a potential increase in ORR and CR rates is suggested with erdafitinib, the wide CrIs again reflect a high level of uncertainty. In terms of safety, risk ratios for AEs also showed wide CrIs, which contribute to the uncertainty and make it difficult to draw definitive conclusions.

Overall, erdafitinib offers a significant clinical benefit over chemotherapy by extending survival and inducing responses for a targeted group of patients with la/mUC, but more certainty is needed to assess the effects erdafitinib against other emerging relevant comparator therapies in the clinical landscape of Canada.

Introduction

The objective of this report is to review and critically appraise the evidence submitted by the sponsor on the beneficial and harmful effects of erdafitinib tablets — 3 mg, 4 mg, and 5 mg — administered orally for the treatment of adult patients with la/mUC harbouring susceptible FGFR3 genetic alterations.

Disease Background

The contents within this section have been informed by materials submitted by the sponsor and clinical expert input. The following has been summarized and validated by the CDA-AMC review team.

UC is the malignant transformation and growth of urothelial cells, which line the renal pelvis, ureters, bladder, and urethra of the urinary tract.^1,2

UC accounts for 90% to 95% of all cases of bladder cancer.^3-5 Bladder cancer is the fifth most common cancer in Canada and is the fourth and ninth most common cancer among men and women, respectively.⁶ In Canada in 2023, it was estimated that 13,400 new cases of bladder cancer would be diagnosed and that 2,600 deaths due to bladder cancer would occur; for 2024, these estimates are expected to be 12,300 and 2,600, respectively.^6,7 The estimated 5-year prevalence is estimated at 134.7 per 100,000 people in Canada.⁸

Microscopic or gross hematuria (blood in the urine) is the most common presenting sign of bladder cancer. Frequent or urgent urination, burning or pain during urination, difficulty urinating, abdominal pain, lower back pain, and pelvic pain may also be present. Other symptoms, such as fatigue, weight loss, anorexia, and bone pain, may develop with advanced or metastatic disease.^9-13 Bladder cancer typically affects people older than 50 years and more commonly occurs in males than females.^6,14 Other risk factors for bladder cancer include white race, personal or family history of bladder cancer, smoking, and environmental and occupational exposures (e.g., chemical carcinogens, arsenic).^4,9,15,16

Bladder cancer is usually diagnosed by a urologist in an outpatient setting. If urinalysis detects blood or abnormal cells in the urine, cystoscopy is performed to visually examine the bladder and urethra for the presence of tumours.^9,17-20 A biopsy, most commonly transurethral resection of bladder tumour, is collected to confirm the diagnosis.^9,17-19 The transurethral resection of bladder tumour involves removal of the tumour and some surrounding muscle to evaluate the spread of the cancer to nearby tissues.¹⁷ These diagnostic procedures are combined with various blood tests and imaging assessments.^9,17

Bladder cancer can be categorized as NMIBC, MIBC, or locally advanced and/or metastatic.^9,21,22 In NMIBC, cancer cells are present in the inner lining of the bladder or connective tissue of the urothelial wall, whereas MIBC is characterized by spread through the connective tissue into the muscle layers of the bladder. Locally advanced bladder cancer involves spread to nearby lymph nodes or other tissues (e.g., prostate, vagina, uterus), and metastatic disease is characterized by spread to the abdomen wall, pelvis, distant lymph nodes, or other areas of the body (e.g., lungs, liver, bones).^22-25 Cancer that cannot be surgically removed is considered unresectable.²³ A clinical expert in Canada consulted by CDA-AMC indicated that, at diagnosis, approximately 70% of patients have NMIBC, 25% have MIBC, and 5% have metastatic disease. Furthermore, 20% of individuals diagnosed with NMIBC go on to develop MIBC and, depending on the stage (e.g., T4, N1), approximately 50% of patients with MIBC will develop advanced and/or metastatic disease recurrence.

Canada-specific survival statistics for locally advanced and metastatic bladder cancer are not available.²⁶ In the US, the 5-year relative survival rate for patients with metastatic disease is 8.8%.²⁷ The clinical experts consulted by CDA-AMC indicated that, based on RCTs, the 5-year OS rate is closer to 10% to 15%.

FGFR1 through FGFR4 comprise a family of transmembrane receptor tyrosine kinases that play a role in proliferation, differentiation, cell migration, and survival.²⁸ Abnormal activity of these FGFRs have been associated with UC, and various gene abnormalities in FGFR3 have been identified in this cancer.^29,30 Erdafitinib binds to and inhibits FGFR1 through FGFR4, resulting in reduced cell viability in cancer cell lines that express activating FGFR genetic alterations; the indication for erdafitinib includes the requirement that la/mUC harbour susceptible FGFR3 genetic alterations.^31,32 According to the literature, mutations in FGFR3 have been found in up to 42% of all UCs, up to 15% of MIBCs, and up to 20% of cases of metastatic disease.³³ A clinical expert in Canada consulted by CDA-AMC estimated the proportion of patients with metastatic bladder cancer with an FGFR3 mutation to be 10% to 15%. This expert also estimated that approximately 40% of patients are currently receiving testing for FGFR alterations; this testing is part of routine care for la/mUC in some settings, although there is a patchwork across Canada with respect to this testing.

Standards of Therapy

The contents within this section have been informed by materials submitted by the sponsor and clinical expert input. The following has been summarized and validated by the CDA-AMC review team.

Health professionals provide tailored treatment plans for patients with la/mUC that follow overarching guidelines and recommendations.³⁴ The goal for patients with la/mUC is to slow disease progression, promote remission, extend life, and improve patient quality of life.^34,35 This is in agreement with the clinical experts consulted by CDA-AMC.

According to the 2019 consensus statement issued by the Canadian Urological Association and the Genitourinary Medical Oncologists of Canada ,³⁶ the 2022 Cancer Care Ontario guideline on muscle-invasive and metastatic bladder cancer,³⁷ and the 2023 Cancer Care Alberta Locally Advanced/Metastatic Bladder Cancer Clinical Practice Guideline,³ therapies for la/mUC include chemotherapy, immunotherapy, and therapies such as enfortumab vedotin and erdafitinib. Locally advanced, unresectable, or metastatic UC has a poor prognosis and is largely incurable. Cisplatin-based chemotherapy is currently the preferred standard of care for patients with newly diagnosed la/mUC. Platinum-based chemotherapy is typically first-line treatment for patients with la/mUC.^3,36,37 Immunotherapy with ICIs, such as pembrolizumab or avelumab, can be offered as a first-line treatment option for patients who are ineligible for platinum-based chemotherapy,³⁷ but is typically offered as second-line treatment for patients who experience disease progression during or after platinum-based chemotherapy, for patients who relapse in the 12 months after receiving neoadjuvant or adjuvant platinum-based chemotherapy for earlier-stage disease, or as maintenance therapy for patients who have stable disease or a response to first-line platinum-based chemotherapy.^3,36,37

An outline summary of the therapies that are currently used in Canada for the treatment of la/mUC is presented in Figure 1, with the proposed place in therapy, as outlined by the sponsor. The information was obtained from Canadian guidelines and the latest CADTH Provisional Funding Algorithm Diagram for Metastatic Urothelial Carcinoma developed in January 2023.³⁸

Figure 1: Sponsor’s Proposed Place in Therapy for Erdafitinib in la/mUC

CDA-AMC = Canada’s Drug Agency; la/mUC = locally advanced unresectable or metastatic urothelial carcinoma; Pt = platinum; UC = urothelial carcinoma.

^aPatients who stopped pembrolizumab treatment after 2 years (35 cycles) for reasons other than disease progression or intolerability are eligible for up to 1 additional year of pembrolizumab upon relapse.

^bIn usual practice, individuals would not receive the platinum-based chemotherapy sequentially in both the neoadjuvant and adjuvant settings.

^cOnly if patients received 4 to 6 cycles of chemotherapy without disease progression.

^dFor patients with muscle-invasive urothelial carcinoma who are at high risk of disease recurrence after undergoing radical resection, treatment with nivolumab should be initiated no more than 120 days after completion of therapy.

^eWith the introduction of enfortumab vedotin, taxanes have largely been displaced and are received after enfortumab vedotin; however, they remain an available therapeutic option after PD-L1 inhibitor therapy.

^fPatients would receive a taxane after enfortumab vedotin and/or erdafitinib (if they have not previously received taxane).

^gThe CADTH metastatic UC Provisional Funding Algorithm is the only Canadian algorithm to include treatment sequencing after nivolumab adjuvant therapy.

Source: Sponsor Clinical Evidence submission.

Drug Under Review

Key characteristics of erdafitinib are summarized in Table 3, along with other treatments available for la/mUC.

Erdafitinib has been approved by Health Canada for the treatment of adult patients with la/mUC harbouring susceptible FGFR3 genetic alterations, who have disease progression during or after at least 1 line of prior therapy, including in the 12 months after neoadjuvant or adjuvant therapy. Erdafitinib should not be used for the treatment of patients who are eligible for but have not received PD-1 or PD-L1 inhibitor therapy. Treatment with erdafitinib should be initiated after confirmation of a susceptible FGFR genetic alteration with a validated test.³¹ Erdafitinib is a pan-FGFR tyrosine kinase inhibitor that binds to and inhibits FGFR1 through FGFR4. Inhibition of FGFR results in decreased FGFR-related signalling and reduced cell viability in cancer cell lines that express activating FGFR genetic alterations.^31,32

Erdafitinib is available as 3 mg, 4 mg, and 5 mg tablets for oral administration. The recommended starting dose of erdafitinib is 8 mg taken orally once daily. The product monograph recommends a dose increase to 9 mg once daily if there is no drug-related toxicity and if serum phosphate concentration is less than 9.0 mg/dL 14 to 21 days after the start of treatment. Dose-modification instructions, based on serum phosphate concentrations and for the management of adverse effects, are detailed in the product monograph. Regular monitoring for elevated serum phosphate concentrations and ocular disorders (with ophthalmological examinations and patient-administered tests) is required. The presence of a susceptible FGFR genetic alteration should be confirmed before initiating treatment with erdafitinib, and erdafitinib should be prescribed and managed by a health care professional with experience in the use of anticancer drug.³¹

The sponsor has requested reimbursement per the Health Canada–approved indication.^39,40 Erdafitinib has not previously been reviewed by CDA-AMC.

Erdafitinib was approved by the FDA in January 2024 for the treatment of “adult patients with locally advanced or metastatic UC with susceptible FGFR3 genetic alterations, as determined by an FDA-approved companion diagnostic test, whose disease has progressed on or after at least one line of prior systemic therapy,” with the limitation that erdafitinib is “not recommended for the treatment of patients who are eligible for and have not received prior PD-1 or PD-L1 inhibitor therapy.”^40,41

On June 27, 2024, the European Medicines Agency Committee for Medicinal Products for Human Use recommended granting marketing authorization for erdafitinib to be used as monotherapy for “the treatment of adult patients with unresectable or metastatic UC harbouring susceptible FGFR3 genetic alterations who have previously received at least one line of therapy containing a PD-1 or PD-L1 inhibitor in the unresectable or metastatic treatment setting.”⁴²

Table 3: Key Characteristics of Erdafitinib, Enfortumab Vedotin, Docetaxel, and Paclitaxel

ADC = antibody-drug conjugate; CSR = central serous retinopathy; FGFR = fibroblast growth factor receptor; MMAE = monomethyl auristatin E; UC = urothelial carcinoma.

^aHealth Canada–approved indication.

^bDepending on the indication, this medication may be used as monotherapy or in combination with other medications.^43,44

Sources: Balversa (erdafitinib) product monograph,³¹ UpToDate erdafitinib drug information,³² Padcev (enfortumab vedotin) product monograph,⁴⁵ docetaxel product monograph,⁴³ paclitaxel product monograph,⁴⁴ National Institutes of Health National Library of Medicine StatPearls – Paclitaxel.⁴⁶

Patient, Clinician, and Drug Program Perspectives

Patient Group Input

This section was prepared by the review team based on the input provided by patient groups.

One patient group, BCC, provided input for this CDA-AMC review. BCC is a registered national charity that serves those facing a bladder cancer diagnosis. The objectives of BCC are to help bladder cancer patients and their support teams address the day-to-day issues of this disease; increase awareness among the public and medical community; and fund research into the diagnosis, treatment, and elimination of bladder cancer.

The information provided by BCC was collected with an online survey posted in May, June, and September of 2024, which asked questions about the impact of FGFR3 metastatic UC on the lives of patients, the effect of current treatments, and patient experience with erdafitinib. Despite the various methods used to reach potential respondents (e.g., BCC mailing list, social media posts, online discussion boards), patients with appropriate experience were difficult to identify. The online survey was completed by 4 people (identified by BCC as patients A, B, C, and D), all of whom were from Canada and had la/mUC with an FGFR3 mutation. The year of diagnosis and number of previously received treatments varied among patients. Patient D is now deceased. Of the 2 respondents (patients A and B) who had treatment experience with erdafitinib, 1 agreed to participate in a telephone interview to discuss their survey responses.

The most common cancer symptoms reported by the 4 respondents were fatigue (75%), as well as insomnia, neuropathy, and decreased mobility (all 50%). Respondents had treatment experience with enfortumab vedotin, gemcitabine, cisplatin, pembrolizumab, carboplatin, docetaxel, and doxorubicin (among others). When asked to rate agreement with the statement “My current therapies are able to manage my cancer symptoms” on a scale of 1 (strongly disagree) to 10 (strongly agree), the average score of 4 (n = 3) suggested that respondents’ symptoms were not adequately managed by current therapies. Fatigue (75%), as well as neuropathy and hair loss (50% each) were the most commonly reported side effects of these treatments. One respondent (patient A) reported difficulty accessing treatment because of travel time. Regarding respondent willingness to tolerate new side effects from drugs that can control disease progression or improve OS on a scale of 1 (will not tolerate side effects) to 10 (will tolerate significant side effects), the 2 patients with erdafitinib experience each responded with a score of 4, whereas the other 2 respondents had scores of 7 and 10; the average score was 6.25.

Patient respondents who were treated with erdafitinib received this treatment after at least 1 line of PD-1 or PD-L1 inhibitor therapy, including in the 12 months after neoadjuvant or adjuvant therapy. Patient A had completed the full course of treatment and patient B had been receiving erdafitinib for 6 weeks. Respondents rated, on a scale of 1 (much worse) to 5 (much better), how their lives had changed on erdafitinib compared to other therapies they had received in terms of categories of metastatic cancer symptoms, drug side effects, maintaining quality of life, controlling disease progression, and preventing recurrence. Scores of 3 from both respondents in every category (except for a score of 4 for maintaining quality of life from patient B) suggested that neither respondent experienced a major difference with erdafitinib, although 1 patient reported that it helped manage their insomnia. Reported side effects from erdafitinib were dysgeusia (patient A) and dry mouth, nausea, and leg pain (patient B). Tolerability of the side effects of erdafitinib, on a scale of 1 (completely tolerable) to 10 (completely intolerable), was rated as 9 by patient A and 3 by patient B, who also commented that the side effects were “better than expected.” Both patients reported that taking erdafitinib orally made their treatment easier, with 1 person commenting, “I like the freedom of it.” Both respondents indicated that they would recommend erdafitinib to other patients with bladder cancer.

Clinician Input

Input From Clinical Experts Consulted by CDA-AMC

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

Unmet Needs

The primary goals in treating la/mUC are to prolong survival and improve or maintain quality of life. The standard treatment includes platinum-based chemotherapy followed by maintenance avelumab for patients who show a response or have stable disease. Alternatives include second-line pembrolizumab. For patients progressing after ICI therapy, enfortumab vedotin is the standard treatment. After enfortumab vedotin, options are limited, with single-drug chemotherapy (e.g., taxanes) being the most common choice. Although enfortumab vedotin is currently the standard after progression on platinum-based chemotherapy and an ICI, it does have toxicities that may not be suitable for all patients. A significant unmet need is determining the optimal sequencing of erdafitinib and enfortumab vedotin, particularly in patients with FGFR alterations. Additionally, with the anticipated approval of enfortumab vedotin combined with pembrolizumab as a first-line therapy, there will be an unmet need for treatment options after enfortumab vedotin plus pembrolizumab.

Place in Therapy

Erdafitinib would be used as monotherapy after platinum-based chemotherapy and an ICI, as supported by the pivotal THOR trial, specifically in patients with FGFR alterations. However, the ideal sequencing of enfortumab vedotin and erdafitinib remains unclear, according to the clinical experts consulted by CDA-AMC.

Patient Population

The target population includes patients whose tumours harbour FGFR alterations, as identified in the pivotal trial. This requires initial testing to detect the genetic alteration. The clinical experts agree that la/mUC is not a rare disease, and that the patient population evaluated in the body of evidence is consistent with the intended target population for this therapy.

Assessing the Response Treatment

The response to treatment should be determined by the response in clinical symptoms to monitor deterioration, as well as by imaging studies that assess the progression of the condition. The timing suggested by experts to assess response (supported by imaging studies) is every 3 months.

Discontinuing Treatment

Clinical experts recommend discontinuing treatment if there is disease progression, significant toxicity, or intolerability, and decisions should be made on a case-by-case basis. These criteria should align with those established in the pivotal trial.

Prescribing Considerations

Diagnosis, treatment prescribing, and ongoing patient monitoring should be managed by clinicians or health care teams with expertise in oncology, including in outpatient settings, as advised by clinical experts.

Clinician Group Input

This section was prepared by the CDA-AMC review team based on the input provided by clinician groups. The full original clinician group input received by CDA-AMC have been included in the Patient, Clinician, and Drug Program Input section of this report.

Seven clinicians from the Ontario Health (Cancer Care Ontario) GU DAC provided input for this review. Ontario Health (Cancer Care Ontario) Drug Advisory Committees provide timely evidence-based clinical and health system guidance on drug-related issues, including the Provincial Drug Reimbursement Programs and the Systemic Treatment Program, in support of Cancer Care Ontario’s mandate.

In communicating current treatments for la/mUC, the GU DAC noted that patients who have previously received an ICI, chemotherapy, or a combination of both are eligible for treatment with enfortumab vedotin, with the goal of improving OS (this goal of treatment was also identified by clinical experts consulted by CDA-AMC). The GU DAC expressed an unmet need for a treatment for patients with genomic alterations, noting that erdafitinib is effective for the FGFR genetic alteration and would be the first targeted therapy identified for this patient population based on molecular testing. Regarding place in therapy, the GU DAC indicated that patients who would be eligible and best suited for treatment with erdafitinib are those with FGFR mutations and/or alterations who have previously received, or have a contraindication to, ICI therapy. A clinical expert consulted by CDA-AMC noted that erdafitinib would be used as monotherapy after chemotherapy and ICI, per the THOR trial. The GU DAC stated that erdafitinib would be administered in the outpatient setting under the supervision of a medical oncologist. Patient response to treatment is assessed in clinical practice using conventional imaging (CT scan of the chest, abdomen, and/or pelvis), at the physician’s discretion. Discontinuation of treatment with erdafitinib would be considered upon unacceptable toxicity or clinically significant disease progression. These treatment aspects (setting, assessment of response, and criteria for discontinuation) are in alignment with those identified by a clinical expert consulted by CDA-AMC.

Drug Program Input

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

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

FGFR = fibroblast growth factor receptor; ICI = immune checkpoint inhibitor; la/mUC = locally advanced unresectable or metastatic urothelial carcinoma; pCODR = pan-Canadian Oncology Drug Review; pCPA = pan-Canadian Pharmaceutical Association; pERC = pan-Canadian Oncology Drug Review Expert Review Committee.

Clinical Evidence

The objective of this CDA-AMC Clinical Review Report is to review and critically appraise the clinical evidence submitted by the sponsor on the beneficial and harmful effects of erdafitinib, 3 mg, 4 mg, or 5 mg, administered orally, in the treatment of patients (≥ 18 years) with la/mUC, FGFR3 alterations, and disease progression during or after at least 1 line of PD-1 or PD-L1 inhibitor therapy, including in the 12 months after neoadjuvant or adjuvant therapy. The focus will be placed on comparing erdafitinib to relevant comparators and identifying gaps in the current evidence.

A summary of the clinical evidence submitted by the sponsor in this review of erdafitinib is presented in 2 sections, with the CDA-AMC critical appraisal of the evidence included at the end of each section. The first section, the Systematic Review, includes pivotal studies and RCTs that were selected in accordance with the sponsor’s systematic review protocol. The CDA-AMC assessment of the certainty of the evidence in this first section, using the GRADE approach, follows the critical appraisal of the evidence. The second section includes indirect evidence from the sponsor.

Included Studies

The following bodies of evidence are included and appraised in this CDA-AMC review document:

• 1 pivotal study (RCT) identified in the systematic review

• 1 ITC (a MAIC).

Systematic Review

The contents of this section have been informed by materials submitted by the sponsor. The following information has been summarized and validated by the CDA-AMC review team.

Description of Studies

One pivotal trial (THOR) was included in the systematic review. The study was a global, phase III clinical trial designed to evaluate the efficacy of erdafitinib compared to standard chemotherapy in patients with metastatic UC who had specific FGFR3 and/or FGFR2 genetic alterations and had progressed after treatment with anti-PD-1 or anti-PD-L1 checkpoint inhibitors. The study involved 266 patients who were randomly assigned to receive either erdafitinib or the investigator’s choice of chemotherapy (docetaxel or vinflunine). The population of interest for this submission is cohort 1 of the THOR study (as opposed to cohort 2, which compared a different population against a group treated with pembrolizumab). Characteristics of cohort 1 are summarized in Table 5. A schematic overview of the study design is presented in Figure 2.

Table 5: Details of the THOR Trial

AE = adverse event; CR = complete response; DoR = duration of response; ECOG PS = Eastern Cooperative Oncology Group Performance Status; FACT-Bl = Functional Assessment of Cancer Therapy–Bladder; FGFR = fibroblast growth factor receptor; HRQoL = health-related quality of life; NMIBC = non–muscle-invasive bladder cancer; ORR = objective response rate; OS = overall survival; PD = progressive disease; PFS = progression-free survival; PGI-S = Patient Global Impression of Severity; PR = partial response; RECIST = Response Evaluation Criteria in Solid Tumours Version 1.1; SAE = serious adverse event.

Sources: 2023 Clinical Study Report for the THOR study, cohort 1;⁴⁷ sponsor’s Summary of Clinical Evidence.

Figure 2: Schematic Overview of the THOR Trial

Chemo = chemotherapy; FU = follow-up; lab = laboratory.

Note: This figure was adapted from the THOR Clinical Study Report to only depict cohort 1.

Source: 2023 Clinical Study Report for the THOR study, cohort 1.⁴⁷

Populations

Inclusion and Exclusion Criteria

Eligible patients from cohort 1 of the THOR trial were adults with a histologic demonstration of transitional cell carcinoma of the urothelium, specifically, metastatic or surgically unresectable urothelial cancer. Patients were required to have an ECOG PS of 0, 1, or 2, documented disease progression before randomization, and specific FGFR genomic alterations identified by a companion diagnostic test. Additionally, eligible patients must have undergone prior treatment with an anti-PD-L1 drug as monotherapy or as a combination therapy; however, patients could have received no more than 2 prior lines of systemic treatment. Prior treatment with an anti-PD-L1 drug could have been given as frontline or maintenance therapy in a neoadjuvant, adjuvant, or metastatic setting.

Patients were ineligible if they had received prior FGFR inhibitor treatment, or had known allergies, hypersensitivity, or intolerance to erdafitinib or its excipients.

Interventions

Erdafitinib was administered orally in tablet form once daily over the course of 21-day cycles until disease progression, intolerable toxicity, withdrawal of consent, or a decision by the investigator to discontinue treatment after disease progression. Treatment with erdafitinib could continue beyond the initial demonstration of disease progression if the patient was clinically stable, at the discretion of the investigator. If a clinically significant increase in tumour burden was identified by tumour imaging at least 4 weeks after the first imaging that indicated progressive disease (PD), erdafitinib was discontinued. Each dose of erdafitinib was taken at home at approximately the same time each day, with or without food. Patients were instructed to avoid consuming grapefruit or Seville oranges due to cytochrome P450 3A4 (CYP450 3A4/5) inhibition. If a dose was missed, it could be taken up to 6 hours after the scheduled time and the patient could return to the normal schedule the next day. If it was more than 6 hours after the missed dose, that dose was skipped and the patient continued treatment at the scheduled time the next day. If vomiting occurred with drug administration, no replacement dose was taken.

All patients randomized to erdafitinib received 8 mg of erdafitinib once daily from day 1 to day 14 of cycle 1, with the option for up-titration to 9 mg, depending on serum phosphate levels on day 14 of cycle 1. If a patient’s serum phosphate level was higher than 9.00 mg/dL (2.91 mmol/L), erdafitinib treatment was withheld. Patients with serum phosphate levels of 7.00 to 8.99 mg/dL (2.25 mmol/L to 2.90 mmol/L) had the option to increase the dose of erdafitinib to 9 mg once daily and concurrently initiate treatment with a phosphate binder. Patients with serum phosphate levels of less than 7.00 mg/dL (< 2.25 mmol/L) had the option to increase the dose of erdafitinib to 9 mg once daily. Dose reductions were allowed, if recommended by a physician in response to an AE. If the AE that resulted in the dose reduction was resolved, the patient was allowed to resume treatment with erdafitinib, with doses re-escalated on a schedule that was the inverse of the dose reduction schedule.

Docetaxel was administered as a 1-hour IV infusion every 3 weeks at a dose of 75 mg/m². Patients continued treatment until disease progression, intolerable toxicity, withdrawal of consent, or a decision by the investigator to discontinue treatment. All patients were premedicated with oral corticosteroids, such as dexamethasone 16 mg per day (e.g., 8 mg twice daily) for 3 days, starting 1 day before docetaxel administration, to reduce the incidence and severity of fluid retention and the severity of hypersensitivity reactions. Docetaxel was not given to patients with bilirubin levels greater than 1 times the upper limit of normal (ULN), or to patients with aspartate transaminase or alanine transaminase levels greater than 1.5 times the ULN and concomitant alkaline phosphatase levels greater than 2.5 times the ULN. Docetaxel could not be given to patients with a neutrophil count of less than1,500 cells/mm³. Treatment with docetaxel was withheld for all grades of neutropenia, thrombocytopenia, and anemia.

Vinflunine was administered as a 20-minute IV infusion once every 3 weeks at a dose of 320 mg/m². Treatment was administered until disease progression, intolerable toxicity, withdrawal of consent, or a decision by the investigator to discontinue treatment. In patients with an ECOG PS of 0 and prior pelvic irradiation or an ECOG PS of at least 1, vinflunine treatment was started at the dose of 280 mg/m². The dose was increased to 320 mg/m² if no hematological toxicity causing treatment delay or dose reduction occurred during the first cycle. Patients who began vinflunine treatment at 320 mg/m² had a dose reduction to 280 mg/m² if an AE occurred. For patients who initiated treatment at 280 mg/m² and experienced an AE, the dose was reduced to 250 mg/m² after the first occurrence and resolution of the AE. If a second AE occurred, treatment was discontinued.

Outcomes

A list of efficacy end points assessed in this Clinical Review Report is provided in Table 6, followed by descriptions of the outcome measures. Summarized end points are based on outcomes included in the sponsor’s Summary of Clinical Evidence, as well as any outcomes identified as important to this review by the clinical experts consulted by CDA-AMC and input from patient and clinician groups and public drug plans. Using the same considerations, the CDA-AMC review team selected end points that were most relevant to CDA-AMC expert committee deliberations and finalized this list of end points in consultation with members of the expert committee. All summarized efficacy end points were assessed using the GRADE approach. Select notable harms outcomes considered important for CDA-AMC expert committee deliberations were also assessed using the GRADE approach.

Table 6: Outcomes Summarized in the THOR Trial Included in the Systematic Review

AE = adverse event; CR = complete response; DoR = duration of response; FACT-Bl = Functional Assessment of Cancer Therapy–Bladder; HRQoL = health-related quality of life; ORR = objective response rate; OS = overall survival; PD = progressive disease; PFS = progression-free survival; PGI-S = Patient Global Impression of Severity; PR = partial response; PRO = patient-reported outcome; SAE = serious adverse event.

^aStatistical testing for these end points was adjusted for multiple comparisons (e.g., hierarchal testing approach, with the following testing order: PFS, ORR, and time to urinary bladder cancer symptom deterioration). Of the PRO end points, only time to urinary bladder cancer symptom deterioration was a key secondary end point and subjected to hierarchical statistical testing. All other PRO end points were nonkey secondary end points.

Sources: 2023 Clinical Study Report for cohort 1 of the THOR trial, sponsor’s Summary of Clinical Evidence.

The outcomes obtained by CDA-AMC were the same as depicted in Table 6. The clinical experts consulted by CDA-AMC agreed on the importance of these outcomes and their evaluation in the body of evidence submitted by the sponsor.

Overall Survival

OS was defined as the time from randomization to the date of death from any cause. OS (in months) was calculated as:

(date of death − date of randomization + 1) / 30.4375

If the patient was alive or the vital status was unknown (for example, if the patient was lost to follow-up or withdrew consent), the patient was censored at the date the patient was last known to be alive. Patients lacking data beyond randomization had their OS censored at the date of randomization. The ITT analysis set was used for OS analyses.

Clinically meaningful survival benefits are being assessed in the cancer landscape in Canada, with 1 study citing an improvement of 2 months or more in median OS, a HR for survival of 0.75 or lower, or both, as the threshold for a clinically meaningful benefit in metastatic colorectal cancer.^48,49 However, no minimal important difference (MID) was identified for OS in la/mUC. The clinical experts consulted by CDA-AMC agreed that patients with la/mUC have, overall, a poor prognosis and will likely consider an improvement in median survival of 2 or 3 months to be meaningful.

Progression-Free Survival

PFS was defined as the duration (in days) from the date of randomization to the date of disease progression (assessed with Response Evaluation Criteria in Solid Tumours Version 1.1 [RECIST 1.1] by the investigator) or relapse from CR or death, whichever is reported first.

RECIST defines PD as at least a 20% increase in the sum of diameters of up to 5 target lesions (maximum 2 lesions per organ), taking as reference the smallest sum on the study and an absolute lesion increase of at least 5 mm or the appearance of new lesions. A CR is the disappearance of all target lesions, and a PR is defined as at least a 30% decrease in the sum of the target lesions. Stable disease is defined as not fitting the criteria for PD or for a PR.⁵⁰

PFS was censored at the date of the last adequate disease assessment for patients who did not have disease progression and were alive, and for patients with unknown disease progression or unknown survival status at the clinical cut-off date. If there is no postbaseline tumour assessment for a patient, PFS was censored on the date of randomization. No threshold or MID for PFS was identified for la/mUC.

Objective Response Rate

ORR was defined as the proportion of patients who achieved a CR or PR, assessed with RECIST 1.1 by the investigator, in the ITT analysis set.

Patients were considered nonresponders if they did not have a CR or PR while on the study, did not have a baseline or postbaseline tumour assessment, or did not have adequate baseline tumour evaluation, or if they died, had PD, dropped out for any reason, or received subsequent therapy before achieving a CR or PR. No threshold or MID for ORR was identified for la/mUC.

Duration of Response

DoR was only measured in responders, and was defined as the duration, in days, from the date of initial documentation of an overall response (CR or PR) to the date of first documented evidence of PD, or relapse for patients who achieved a CR during the study, or death. The censoring employed was similar to that used for PFS. No MID was identified for DoR in la/mUC.

HRQoL or PROs

Functional status, symptoms, and HRQoL were assessed using 3 PRO measures: FACT-Bl, PGI-S, and EQ-5D-5L. The performance values of these tools are presented in Table 7. Time to urinary bladder cancer symptom deterioration was a novel end point generated from the urinary bladder cancer symptoms score from 3 items on the FACT-Bl.

Functional Assessment of Cancer Therapy–Bladder

The FACT-Bl consists of 36 core items, with 5-point Likert response subscales that cover 5 primary domains: physical well-being, social/family well-being, emotional well-being, functional well-being, and bladder cancer subscale. The 5-point Likert scales ranged from 0, meaning not at all to 4, meaning very much. Each subscale consists of multiple questions, and the subscale score is calculated by summing the question scores, multiplying this sum by the number of items in the subscale, then dividing by the number of items answered. Total subscale scores can range from 0 to 28 on the physical well-being, social/family well-being, and functional well-being subscales, 0 to 24 on the emotional well-being scale, and 0 to 48 on the bladder cancer subscale. The trial outcome index, Functional Assessment of Cancer Therapy–General (FACT-G), and total FACT-Bl scores are derived from the subscale scores. The trial outcome index is calculated as the sum of the physical well-being, functional well-being, and bladder cancer symptom scores. Trial outcome index scores can range from 0 to 104. The FACT-G total score is calculated as the sum of the physical well-being, social/family well-being, emotional well-being, and functional well-being scores. FACT-G total scores can range from 0 to 108. The FACT-Bl total score is calculated as the sum of the physical well-being, social/family well-being, emotional well-being, functional well-being, and bladder cancer symptom scores. FACT-Bl total scores can range from 0 to 156. According to the FACT-Bl version 4 scoring guidelines, a higher score indicates a better quality of life. The MID of the FACT-Bl was determined to be 11.26.

Time to urinary bladder cancer symptom deterioration was a novel end point generated from the urinary bladder cancer symptoms score from 3 items on the FACT-Bl (I have trouble controlling my urine, I urinate more frequently than usual, and it burns when I urinate).

Time to deterioration was defined as the first time to worsening from baseline in the urinary bladder cancer symptoms score beyond a meaningful change threshold (−1 point) from the day of randomization. The meaningful change threshold was established using anchor-based analysis, with the PGI-S as the anchor. Patients who experienced disease progression or death were censored from the analyses. The validity, reliability, and responsiveness measures of this tool have not been assessed.

Patient Global Impression of Severity

The PGI-S is a single verbal question regarding disease severity: “Considering all aspects of your bladder cancer symptoms right now, would you say your bladder cancer symptoms are none (0), mild (1), moderate (2), severe (3), or very severe (4)?” The PGI-S was an anchor question that was used to establish the magnitude of MID thresholds for each PRO-derived score, allowing the classification of patients as improved, stable, worse. For patients with UC, specific studies focusing on the MID for PGI-S are absent.

EQ-5D-5L

The EQ-5D-5L is a generic measure of health status. It is a 5-item questionnaire that assesses 5 domains, each scored across 5 levels (no problems, slight problems, moderate problems, severe problems, and extreme problems). The domains include mobility, self-care, usual activities, pain/discomfort, and anxiety/depression, and a visual analogue scale is used to rate a patient’s health today, with anchors ranging from 0 (worst imaginable health state) to 100 (best imaginable health state).⁵¹

The scores for the 5 separate questions are categorical and cannot be analyzed as cardinal numbers. However, the scores for the 5 dimensions were used to compute a single utility score that ranged from 0.0 to 1.0, representing the general health status of the individual. The EQ-5D-5L asks respondents to select their response based on their current health (today), and takes less than 5 minutes to complete. A unique EQ-5D health state is defined by combining 1 level from each of the 5 dimensions, and responses to the 5 items are converted to a weighted health state index (utility score) that ranges from 0 (death) to 1 (perfect health).

A simulation-based study showed that the MID of the EQ-5D-5L index score ranged from 0.037 and 0.069 for algorithms from Canada, China, England, Japan, Spain, and Uruguay, and the Canadian-specific estimate was 0.037 (standard deviation [SD] = 0.001).⁵² Currently, there is no specific mention of an MID for la/mUC.

Table 7: Summary of Patient-Reported Outcome Measures and Their Measurement Properties

BCI = Bladder Cancer Index; FACT-Bl =Functional Assessment of Cancer Therapy–Bladder; HRQoL = health-related quality of life; ICC = intraclass correlation coefficient; MID = minimal important difference; PGI-S = Patient Global Impression of Severity; PRO = patient-reported outcome; UC = urothelial carcinoma.

Statistical Analysis

The statistical analyses of all clinical trial end points are summarized in Table 8.

Overall Survival

The primary efficacy analysis, OS, was based on the ITT analysis set. The survival curves of OS were described using the Kaplan-Meier (KM) method together with the estimated median (in months), with a 95% CI for each treatment group. A stratified log-rank test was used to compare the treatment groups within cohort 1 at an overall alpha level of 0.05. Additionally, the HR for erdafitinib relative to chemotherapy and its associated 95% CI were calculated using a Cox PH model. A stratified analysis for OS using ECOG PS (0 or 1 versus 2), disease distribution (presence versus absence of visceral [lung, liver, or bone] metastases), and region (North America versus Europe versus rest of the world) was prespecified. However, due to the limited number of events by treatment group per strata (i.e., < 10) for OS, the analyses for OS were unstratified.⁵⁸

Additional sensitivity, exploratory, and subgroup analyses were performed for OS:^47,58

• Sensitivity analysis of OS compared the erdafitinib and chemotherapy treatment arms using a log-rank test without adjustments for stratification factors.

• Sensitivity analyses for OS were performed in the safety population (all randomized patients who received at least 1 dose of study drug).

• Exploratory analyses on OS were performed using a selected set of potential prognostic variables (obtained at or before baseline) as covariates in Cox regression models. Potential prognostic factors included age, sex, race, region, baseline ECOG PS, baseline disease distribution, hemoglobin level, and FGFR alteration.

Subgroup analyses were conducted to assess efficacy consistency across patient populations with different demographics or baseline characteristics. Subgroups are outlined in Table 9.

• Subgroup analysis of median OS was performed for patients who received vinflunine and docetaxel.

• Subgroup analysis of median OS was performed for patients who received prior cisplatin-based or carboplatin-based chemotherapy.

The key secondary end points were tested for statistical significance using a hierarchical testing approach, with the following testing order: PFS, ORR, time to urinary bladder cancer symptom deterioration), meaning that if the primary efficacy end point of OS showed a statistically significant improvement in OS for patients in the erdafitinib group compared to the chemotherapy group, the statistical significance would be assessed in the secondary end points.

Table 8: Statistical Analysis of Efficacy End Points for Patients Treated in Cohort 1 of the THOR Trial

CI = confidence interval; DoR = duration of response; ECOG PS = Eastern Cooperative Oncology Group Performance Status; FACT-Bl = Functional Assessment of Cancer Therapy–Bladder; HR = hazard ratio; KM = Kaplan-Meier; MMRM = mixed model for repeated measures; NR = not reported; ORR = objective response rate; OS = overall survival; PFS = progression-free survival; PGI-S = Patient Global Impression of Severity; PH = proportional hazards; PRO = patient-reported outcome; SD = standard deviation.

Sources: 2023 Statistical Analysis Plan for the THOR study,⁵⁸ sponsor’s Summary of Clinical Evidence.

Progression-Free Survival

The ITT analysis set was used for all secondary analyses. PFS was analyzed using the same method as for OS. A stratified analysis for PFS, with the same stratification factors used to analyze OS, was prespecified. However, due to the limited number of events by treatment group per strata (i.e., < 10) for PFS, the analyses for PFS were unstratified.

The following subgroup and sensitivity analyses for PFS were performed:^47,58

• Sensitivity analyses for PFS were performed in the safety population (all randomized patients who received at least 1 dose of study drug).

• Sensitivity analysis of PFS was performed with patients censored at their last disease evaluation if they had progressed or died after missing ≥ 2 consecutive assessments.

• Subgroup analyses were conducted to assess efficacy consistency across patient populations with different demographics or baseline characteristics. Subgroups are outlined in Table 9.

• Subgroup analysis of median PFS was performed for patients who received vinflunine and docetaxel.

• Subgroup analysis of median PFS was performed for patients who received prior cisplatin-based or carboplatin-based chemotherapy.

Objective Response Rate

A stratified Cochran-Mantel-Haenszel test was used to assess treatment differences in ORR at a significance level of 0.05 and to estimate the RR and P values between the treatment groups. The 95% CIs for the RRs were calculated based on Wald statistics and were estimated using ECOG PS (0 or 1 versus 2) as a stratification factor. An analysis that used region (North America versus Europe versus rest of the world), ECOG PS (0 or 1 versus 2), and disease distribution (presence versus absence of visceral [lung, liver, or bone] metastases) as stratification factors was prespecified. However, data were pooled for the stratification factors, per the pooling algorithm specified in the statistical analysis plan. The pooling algorithm requires that the patients be pooled together, and drops stratification factors until there are at least 10 events in each stratum. The stratification factors were dropped in the following order: region, disease distribution, ECOG PS (0 or 1 versus 2).

Subgroup and sensitivity analyses for ORR were performed:⁵⁸

• Subgroup analyses were conducted to assess efficacy consistency across patient populations with different demographics or baseline characteristics. Subgroups are outlined in Table 9.

• Subgroup analysis of median ORR was performed for patients who received vinflunine and docetaxel.

• Subgroup analysis of median ORR was performed for patients who received prior cisplatin-based or carboplatin-based chemotherapy.

Duration of Response

DoR was described using the KM method together with estimated median (in months), with a 95% CI for each treatment group. Additionally, the HR for erdafitinib relative to chemotherapy and its associated 95% CI were calculated using a Cox PH model.⁵⁸

Table 9: Definition of Subgroups in the THOR Trial

CPS = combined positive score; ECOG PS = European Cooperative Oncology Group Performance Status; FGFR = fibroblast growth factor receptor.

^aIf the number of patients from a treatment group (i.e., American Indian or Alaska Native [wording from original source], Asian, Native Hawaiian or other Pacific Islander, and not reported) was less than 10, then those patients were combined into a new subgroup, named “other.”

Source: 2023 THOR Statistical Analysis Plan.⁵⁸

Patient-Reported Outcomes

All PROs were analyzed using the ITT analysis set. To evaluate changes in scores from baseline over time, descriptive statistics (number of observations, mean, SD, minimum, maximum) of scores at baseline and change from baseline at each scheduled visit were reported for FACT-Bl and EQ-5D-5L. For EQ-5D-5L scores, separate summaries were performed at the treatment phase and follow-up phase. For PGI-S, the frequency count and percentage of each score level (1 to 5) over time were provided.

A mixed model for repeated measures analysis was conducted to estimate the change from baseline at each scheduled visit for the 2 treatment arms for FACT-Bl and EQ-5D-5L scores. Changes from baseline were fitted to a mixed-effects model that included patients as a random effect, and baseline value, treatment group, time in month, treatment-by-time interaction, and stratification factors as fixed effects. The stratification factors that were used in the analysis were region (North America versus Europe versus rest of the world), ECOG PS (0 or 1 versus 2), and disease distribution (presence versus absence of visceral [lung, liver, or bone] metastases). Swim plots of the change from baseline, with SD, over time by treatment group were produced.

The median time to deterioration was estimated using a KM method, and the HR and associated 95% CI were estimated using a stratified Cox PH model, with stratification variables identical to those listed for FACT-Bl and EQ-5D-5L scores. A sensitivity analysis for time to urinary bladder cancer symptom deterioration was also conducted, and included disease progression and death as deterioration events.

Safety

Safety data were analyzed using the safety analysis set. The baseline value for safety assessment was defined as the value collected at the time closest to, but before, the administration of the first dose of the study drug. Descriptive statistics or frequency counts and percentages by treatment group were derived for continuous and categorical safety variables, respectively.

Sample Size and Power Calculation

The planned enrolment for cohort 1 was approximately 280 patients (approximately 140 patients to each arm). The data cut-off date for the final analysis was when approximately 208 death events had occurred. Assuming a 53% improvement in median OS for the erdafitinib arm over the chemotherapy arm (an HR of 0.65, under the exponential distribution assumption), the study had at least 85% power to detect an HR of 0.65 at a statistical significance level of 5% (2-sided), with 1 interim analysis for efficacy at an approximately 65% information fraction (approximately 136 deaths) and a final analysis.

Statistical Testing

The type I error was controlled at 5% (2-sided) in cohort 1. All tests were conducted at a 2-sided alpha level of 0.05, and 95% CIs were provided. The familywise type I error was strongly controlled at 5% (2-sided) for the secondary end points. A hierarchical testing approach was used for testing the secondary end points (i.e., a secondary null hypothesis was tested in each cohort if, and only if, the primary null hypothesis and all the secondary null hypotheses that preceded it had been rejected). For the PROs, the level of testing was assumed to be 0.05.

At both the interim and final analyses, the secondary end points were assessed at the same significance levels as specified for testing the primary end point (OS), based on the O'Brien-Fleming alpha-spending function, to protect the overall type I error rate, per the THOR statistical analysis plan. The testing order of these end points was as follows: PFS, ORR, and time to urinary bladder cancer symptom deterioration. If the null hypothesis for any of these end points failed to be rejected at the interim analysis, then any of the subsequent end point(s) were not tested until the final analysis. If the null hypothesis for an end point was rejected at the interim analysis, it remained rejected and was not retested at the final analysis.

The plan was that the interim analysis for cohort 1 would occur after an approximately 65% information fraction (approximately 136 of 208 deaths) occurred. Both superiority and futility were assessed at the interim analysis. The classic O’Brien-Fleming boundaries were used for these assessments. The stopping boundaries were implemented with the Lan-DeMets spending function to control the type I error at the 0.05 significance level overall. The futility assessment was not binding.

The cohort 1 interim analysis occurred after 155 death events had been observed (at approximately 75% information fraction), and the significance level was determined using the O’Brien-Fleming alpha-spending function to control the type I error at the 0.05 significance level (2-sided). The efficacy boundary P value was predetermined to be 0.019 (2-sided). The IDMC recommended that cohort 1 be stopped due to the superiority of erdafitinib treatment over chemotherapy, and that patients be allowed to crossover to erdafitinib treatment. Because cohort 1 was stopped, the interim analysis of cohort 1 is considered the final analysis.

Analysis Populations

The 5 analysis sets employed in this study are listed in Table 10.

Table 10: Analysis Populations of the THOR Trial

ITT = intention-to-treat; OS = overall survival; PFS= progression-free survival; PRO = patient-reported outcome.

Source: 2023 THOR Statistical Analysis Plan.⁵⁸

Results

Patient Disposition

Patient disposition in the cohort 1 ITT analysis set of the THOR trial is summarized in Table 11 and in Figure 3. All these variables were equally distributed between groups, except for the proportion of patients who received the study treatment in the chemotherapy group (86.2%), which was lower than in the erdafitinib group (99.3%).

Table 11: Patient Disposition in Cohort 1 of the THOR Trial (ITT Analysis Set; January 15, 2023, Clinical Cut-Off)

ITT = intention-to-treat.

Notes: Patients were considered to have completed the study if they died during the study or had not been withdrawn from the study by the end of the study.

Three patients did not receive prior anti-PD-L1 therapy but were incorrectly randomized to cohort 1.

^aPercentages are based on patients who discontinued the study treatment.

^bPercentages are based on patients who discontinued the study.

Source: 2023 Clinical Study Report for the THOR study, cohort 1.⁴⁷

Baseline Characteristics

Baseline demographics, disease characteristics, and a summary of prior anticancer therapy of patients from cohort 1 of the THOR trial are summarized in Table 12, Table 13, and Table 14, respectively.

In the THOR trial, imbalances were noted in some variables, such as the difference in number of patients receiving 2 previous lines of chemotherapy and the number declining chemotherapy; however, these were not considered to affect the properness of randomization, and the distribution of baseline characteristics was considered, overall, to be well balanced in the erdafitinib and chemotherapy treatment groups. Although the inclusion criteria allowed for patients with FGFR2 and/or FGFR3 alterations, no enrolled patients had an FGFR2 alteration, which is rare.

Figure 3: Disposition of Patients in Cohort 1 of the THOR Trial

ITT = intention-to-treat.

Source: 2023 Clinical Study Report for the THOR study, cohort 1.⁴⁷

In the cohort 1 ITT analysis set, 78 patients (29.3%) had received 1 prior line of systemic therapy (33.1% of patients in the erdafitinib group and 25.4% of patients in the chemotherapy group) and 187 patients (70.3%) had received 2 prior lines of systemic therapy (66.2% patients in the erdafitinib group and 74.6% patients in the chemotherapy group). One patient (0.4%) had received 3 prior lines of systemic therapy and 3 patients did not receive a prior anti-PD-L1 drug (1 patient randomized to erdafitinib, 2 patients randomized to chemotherapy); they were incorrectly assigned to cohort 1 because of incorrect cohort assignment flow in the Interactive Web Response System at the time of randomization. All 4 patients were included in the final analyses.

Note that in the THOR trial, induction chemotherapy followed by maintenance anti-PD-L1 therapy was considered a single line of therapy. More than half of the patients in both treatment groups (55.9% erdafitinib, 58.5% chemotherapy) received anti-PD-1 or anti-PD-L1 as a single drug in the second-line setting. The most received anti-PD-L1 drugs were pembrolizumab (35.3%), avelumab (22.2%), and atezolizumab (19.5%).

Table 12: Summary of Baseline Characteristics in Cohort 1 of the THOR Trial (ITT Analysis Set; January 15, 2023, Clinical Cut-Off)

ITT = intention-to-treat; SD = standard deviation.

Source: 2023 Clinical Study Report for the THOR study, cohort 1.⁴⁷

Table 13: Baseline Disease Characteristics of Patients in Cohort 1 of the THOR Trial (ITT Analysis Set; January 15, 2023, Clinical Cut-Off)

CPS = combined positive score; ECOG PS = Eastern Cooperative Oncology Group Performance Status; FGFR = fibroblast growth factor receptor; ITT = intention-to-treat; SD = standard deviation.

Source: 2023 Clinical Study Report for the THOR study, cohort 1.⁴⁷

Table 14: Summary of Prior Anticancer Therapy in Cohort 1 of the THOR Trial (ITT Analysis Set; January 15, 2023, Clinical Cut-Off)

ADC = antibody-drug conjugate; ITT = intention-to-treat; gem = gemcitabine; MVAC = methotrexate, vinblastine, doxorubicin (Adriamycin), cisplatin.

Notes: Percentages are based on the number of patients in the analysis set of the corresponding treatment group.

One patient with 3 prior lines of prior systemic therapy is not included in this table.

The + symbol indicates medications that are part of the same line of therapy, which may have been administered sequentially or concurrently. The ; symbol indicates medication(s) administered as a separate line of therapy.

Source: 2023 Clinical Study Report for the THOR study, cohort 1.⁴⁷

Exposure to Study Treatments

Study Treatments

Patient exposure in cohort 1 of the THOR trial for the safety analysis set is summarized in Table 15. The median extent of exposure in cohort 1 was 146.0 days (range, 5 days to 1,162 days) in the erdafitinib group and 43.0 days (range, 1 day to 820 days) in the chemotherapy group. At the time of clinical cut-off for the cohort 1 analysis (January 15, 2023), 29 patients in the erdafitinib group and 10 patients in the chemotherapy group were still on the study treatment.

Table 15: Patient Exposure Characteristics in Cohort 1 of the THOR Trial (Safety Analysis Set; January 15, 2023, Clinical Cut-Off)

NA = not applicable; SD = standard deviation.

Source: 2023 Clinical Study Report for the THOR study, cohort 1.⁴⁷

Concomitant Medications and Cointerventions

No concomitant medications or cointerventions were required during the THOR trial. Concomitant medications were allowed; however, concomitant treatment with investigational drugs, antineoplastic drugs, and hormonal anticancer therapy were prohibited for all patients in cohort 1. Concomitant treatment with strong inhibitors of the CYP3A4 enzymes or live vaccines in the 30 days before the first dose, during the study, and for 3 months after the last dose of study therapy were prohibited for patients in the chemotherapy arm. Concomitant treatment with strong inducers of CYP3A4 enzymes and QT- and/or QTc-prolonging drugs were prohibited for patients in the chemotherapy arm receiving vinflunine.

Permitted concomitant medications for cohort 1 of the THOR trial were as follows:

• Symptomatic treatments: Supportive care, such as antibiotics, analgesics, transfusions, diet, and concomitant medications for the symptomatic treatment of related toxicities (grade 1 to 4) could be administered in accordance with the standard of care at the site, at the treating physician’s discretion, as clinically indicated.

• Prophylactic medications: Appropriate prophylactic antiemetic regimens could be provided if required, in accordance with institutional practice and current European Society of Medical Oncology guidelines.

• Chronic supportive therapies: Ongoing bisphosphonates and denosumab or other supportive therapies were permitted.

• Palliative radiotherapy: Localized radiotherapy for symptomatic control was permitted, but did not include definitive radiation to target lesions.

The most frequently used concomitant medications started during treatment (received by ≥ 30% of patients overall), by anatomic therapeutic chemical class, are summarized in Table 16. After study entry, 130 of 135 patients (96.3%) in the erdafitinib group and 106 of 112 patients (94.6%) in the chemotherapy group received concomitant medication.

Subsequent Treatment

Crossover to the erdafitinib treatment group was permitted after the January 15, 2023, clinical cut-off because of the superiority of erdafitinib treatment over chemotherapy. Crossover results are not reported in the 2023 Clinical Study Report for cohort 1 of the THOR trial and, therefore, are not included in this report, as crossover occurred after the interim efficacy and safety analyses were conducted.

Table 16: Summary of Concomitant Medication Started During Treatment in Cohort 1 of the THOR Trial (Safety Analysis Set; January 15, 2023, Clinical Cut-Off)

ATC = anatomic therapeutic chemical.

Source: 2023 Clinical Study Report for the THOR study, cohort 1.⁴⁷

A summary of subsequent anticancer therapy based on the January 15, 2023; clinical cut-off is reported in Table 17. Chemotherapy was the most common type of subsequent anticancer therapy in both groups.

Table 17: Summary of Subsequent Anticancer Therapy in Cohort 1 of the THOR Trial (ITT Analysis Set; January 15, 2023, Clinical Cut-Off)

FGFR = fibroblast growth factor receptor; ITT = intention-to-treat.

Source: 2023 Clinical Study Report for the THOR study, cohort 1.⁴⁷

Efficacy

The outcomes outlined in this section are presented in the order of importance, determined by the CDA-AMC team from input from clinical experts, clinical groups, and others.

A summary of the primary and secondary efficacy end points for patients treated with erdafitinib in cohort 1 of the THOR trial is provided in Table 18.

Of note, cohort 1 was stopped early; the IDMC recommended that cohort 1 be stopped because of the superiority of erdafitinib treatment over chemotherapy. As such, the interim analysis of cohort 1 is considered the final analysis, and information was extracted from the Clinical Study Report with a January 15, 2023, clinical cut-off.

Table 18: Primary and Secondary Efficacy Outcomes in Cohort 1 of the THOR Trial (ITT Analysis Set; January 15, 2023, Clinical Cut-Off)

CI = confidence interval; CR = complete response; DoR = duration of response; ITT = intention-to-treat; NE = not estimable; ORR = objective response rate; OS = overall survival; PD = progressive disease; PFS = progression-free survival; PR = partial response.

Note: For ORR, a risk ratio greater than 1 indicates that the probability of achieving an objective response (PR or CR) is higher in the erdafitinib arm than in the chemotherapy arm.

^aNumbers represent the proportion (i.e., 1 = 100%), unless otherwise specified. OS, in months, is calculated as: (date of death – date of randomization + 1)/30.4375. If the patient is alive or the vital status is unknown (for example, if the patient is lost to follow-up or withdrew consent), OS is censored at the date the patient was last known to be alive. Patients lacking data beyond randomization have their OS censored at the date of randomization.

^bHazard ratio and 95% CI are estimated using a Cox proportional hazards regression model, with treatment as the only explanatory variable. A hazard ratio of less than 1 indicates longer survival in the erdafitinib arm than in the chemotherapy (vinflunine or docetaxel) arm.

^cP value is 2-sided and is based on a log-rank test.

^dNumbers represent the proportion (i.e., 1 = 100%), unless otherwise specified. PFS, in months, is defined as the time from the date of randomization to the date of disease progression (assessed with RECIST 1.1 by the investigator) or relapse from CR or death, whichever is reported first, divided by 30.4375. PFS is censored at the date of the last adequate disease assessment for patients who do not have disease progression and are alive, as well as for patients with unknown disease progression or unknown survival status as of the clinical cut-off date. Also, if there is no postbaseline tumour assessment for a patient, PFS is censored on the date of randomization. Patients who die or have disease progression after starting subsequent anticancer therapy are censored at the last tumour assessment date.

^eRR, 95% CI, and P value are estimated using Cochran-Mantel-Haenszel test, with ECOG PS (0 or 1 versus 2) as a stratification factor.

^fNumbers represent the proportion (i.e., 1 = 100%), unless otherwise specified. DoR, in months, is defined as the time from the date of initial documentation of CR or PR to the date of disease progression (assessed with RECIST 1.1 by the investigator) or relapse from CR or death, whichever is reported first, divided by 30.4375. DoR is only calculated for the subgroup of patients with an overall response of CR or PR. DoR is censored at the date of the last adequate disease assessment for patients who do not have disease progression and are alive, as well as for patients with unknown disease progression or unknown survival status as of the clinical cut-off date. Patients who die or have disease progression after starting subsequent anticancer therapy will be censored at the last tumour assessment date.

Source: 2023 Clinical Study Report for the THOR study, cohort 1.⁴⁷

Overall Survival

The primary end point of the THOR trial was OS, which was also considered a critical outcome by the clinical experts consulted by CDA-AMC and others involved. This end point was assessed in the ITT analysis set and was measured from the date of randomization to the date of death. If the patient was alive or the vital status was unknown, the patient was censored at the date the patient was last known to be alive.

In an unstratified OS analysis, the median follow-up for OS was 15.90 months (18.04 months for the erdafitinib treatment group and 14.92 months for the chemotherapy treatment group). Median OS was reached in both treatment groups. In the erdafitinib group, the median OS was 12.1 months (95% CI, 10.3 to 16.4 months), with 43.4% of patients censored. The estimated 6-month and 24-month survival rates were 0.85 (95% CI, 0.77 to 0.90) and 0.26 months (95% CI, 0.17 to 0.36 months), respectively. In the chemotherapy group, the median OS was 7.8 months (95% CI, 6.5 to 11.1 months), with 40.0% of patients censored. The estimated 6-month and 24-month survival rates were 0.66 (95% CI, 0.56 to 0.74) and 0.20 (95% CI, 0.11 to 0.31), respectively (Table 18).

Based on a predetermined 2-sided efficacy boundary P value of 0.019, OS was better in patients treated with erdafitinib than with chemotherapy (HR = 0.64; 95% CI, 0.47, 0.88; P value = 0.0050). Based on the HR, treatment with erdafitinib resulted in a 36% reduction in time to death compared to treatment with chemotherapy. The KM curve for OS for patients in the ITT analysis set is provided in Figure 4. The subgroup analysis is presented in Figure 5 and Figure 6. Although subgroup analyses were conducted, the study was not designed to assess the treatment effect within subgroups.

A post hoc sensitivity analysis was performed using the safety analysis set to evaluate the impact of the asymmetry in the number of patients randomized to chemotherapy versus erdafitinib who did not receive treatment. The sensitivity analysis excluded patients who were randomized but did not receive the study treatment (1 patient in the erdafitinib arm and 18 patients in the chemotherapy arm). The results were consistent with the primary analysis (HR = 0.68; 95% CI, 0.49 to 0.94; P value = 0.0199).

Progression-Free Survival

PFS was measured in the ITT analysis set. For patients who did not have disease progression and were alive, as well as for patients with unknown disease progression or unknown survival status as of the clinical cut-off date, PFS was censored at the date of the last adequate disease assessment. If there was no postbaseline tumour assessment for a patient, PFS was censored on the date of randomization. Adequate disease assessment was defined as having sufficient evidence to indicate correctly that progression has or has not occurred.

In an unstratified PFS analysis, median PFS was achieved in both arms. In the erdafitinib arm, median PFS was 5.55 months (95% CI, 4.40 to 5.65 months), with 25.7% of patients censored. The estimated 6-month and 24-month PFS rates were 0.37 (95% CI, 0.28 to 0.46) and 0.05 (95% CI, 0.01 to 0.12), respectively. In the chemotherapy arm, median PFS was 2.73 months (95% CI, 1.81 to 3.68 months), with 30.8% of patients censored. The estimated 6-month and 24-month PFS rates were 0.27 (95% CI, 0.19 to 0.37) and 0.04 (95% CI, 0.00 to 0.13), respectively.

Figure 4: KM Plot for OS in Cohort 1 of the THOR Trial (ITT Analysis Set; January 15, 2023, Clinical Cut-Off)

ITT = intention-to-treat; KM = Kaplan-Meier; OS = overall survival.

Source: 2023 Clinical Study Report for the THOR study, cohort 1.

Figure 5: OS Subgroup Analysis for Cohort 1 of the THOR Trial, Part 1 of 2 (ITT Analysis Set; January 15, 2023, Clinical Cut-Off)

Chemo = chemotherapy; CI = confidence interval; Erda = erdafitinib; EVT = events; ITT = intention-to-treat; NE = not estimable; OS = overall survival.

*Includes cisplatin, carboplatin, and multiple platinum-based chemotherapies.

Source: 2023 Clinical Study Report for the THOR study, cohort 1.⁴⁷

Figure 6: OS Subgroup Analysis for Cohort 1 of the THOR Trial, Part 2 of 2 (ITT Analysis Set; January 15, 2023, Clinical Cut-Off)

Chemo = chemotherapy; CI = confidence interval; CPS = combined positive score; ECOG = Eastern Cooperative Oncology Group; Erda = erdafitinib; EVT = events; FGFR = fibroblast growth factor receptor; ITT = intention-to-treat; OS = overall survival.

Source: 2023 Clinical Study Report for the THOR study, cohort 1.⁴⁷

Based on a 2-sided and multiplicity-adjusted significance level of 0.019, the secondary efficacy end point of PFS showed a statistically significant improvement for patients treated with erdafitinib compared to chemotherapy (HR = 0.58; 95% CI, 0.44 to 0.78; P value = 0.0002). Based on the HR, treatment with erdafitinib resulted in a 42% reduction in time to disease progression or death compared to treatment with chemotherapy. The KM curve for PFS for patients in the ITT analysis set is provided in Figure 7.

Figure 7: KM Plot for PFS in Cohort 1 of the THOR Trial (ITT Analysis Set; January 15, 2023, Clinical Cut-Off)

ITT = intention-to-treat; KM = Kaplan-Meier; PFS = progression-free survival.

Source: 2023 Clinical Study Report for the THOR study, cohort 1.⁴⁷

Two post hoc sensitivity analyses were conducted for PFS. One post hoc sensitivity analysis assessed patients who were censored at their last disease evaluation if they progressed or died after missing at least 2 consecutive assessments. The results of this analysis were consistent with the main PFS analysis. Another post hoc sensitivity analysis, using the safety analysis set, was performed to evaluate the impact of the asymmetry in the number of patients randomized to chemotherapy versus erdafitinib who did not receive treatment. This sensitivity analysis excluded patients who were randomized but did not receive study treatment (1 patient in the erdafitinib arm and 18 patients in the chemotherapy arm). The results were also consistent with the primary analysis (HR = 0.60; 95% CI, 0.45 to 0.81; P value = 0.0006). As with OS, subgroup analyses were conducted but the study was not designed to assess the treatment effect within subgroups, so subgroup effects were not reported.

Objective Response Rate

ORR was measured in the ITT analysis set and defined as the proportion of patients who achieved a CR or PR, as assessed with RECIST 1.1 by the investigator.

ORR was achieved by 62 patients (45.6%) in the erdafitinib treatment arm and 15 patients (11.5%) in the chemotherapy treatment arm (Table 18). Based on a 2-sided and multiplicity-adjusted significance level of 0.019, the difference in ORR was statistically significant between the 2 treatment groups (observed RR, 3.94; 95% CI, 2.37 to 6.57; P < 0.001). Based on the RR, patients treated with erdafitinib are approximately 4 times more likely to achieve an objective response than patients treated with chemotherapy.

The ORR was also examined in prespecified clinically relevant subgroups in the ITT population (FGFR alteration type, PD-L1 status, tumour location, visceral metastases, demographic and baseline characteristics). As with other outcomes, the study was not designed to assess the treatment effect within subgroups, and the number of patients is small. No subgroup effects were apparent.

A post hoc analysis of ORR that compared erdafitinib with docetaxel or vinflunine was consistent (based on HR) with the main ORR analysis comparing erdafitinib with chemotherapy.

Duration of Response

DoR was measured in responders as the duration, in days, from the date of initial documentation of a response to the date of first documented evidence of PD (or relapse for patients who achieved a CR during the study) or death. The censoring was similar to that used for PFS.

The median DoR determined by the investigator was 4.86 months (95% CI, 3.84 to 7.46 months) for the erdafitinib treatment group and 5.55 months (95% CI, 2.14 to 6.01 months) for the chemotherapy treatment group. Patients treated with erdafitinib had a numerically shorter DoR than those treated with chemotherapy (HR = 0.85; 95% CI, 0.43 to 1.66). The DoR in the chemotherapy group was evaluated based on a small number of patients (Table 18).

Of the 17 patients in the erdafitinib group with no disease progression as of the data cut-off date, 16 patients were still on the study treatment and 13 patients continued in the study for more than 2 months. Of the 3 patients in the chemotherapy group with no disease progression as of the data cut-off date, 2 patients were still on the study treatment and 2 patients continued in the study for more than 2 months.

Patient-Reported Outcomes

PROs were measured in the ITT analysis set and evaluated as the change from baseline in HRQoL, assessed with the FACT-Bl, EQ-5D-5L, PGI-S, and time until urinary bladder cancer symptom deterioration (subset of FACT-Bl items). However, the results from the PRO assessments should be interpreted in light of the limitation that the number of patients with available PRO data decreased over time.⁴⁷

Compliance with the FACT-Bl, EQ-5D-5L, and PGI-S assessments was greater than 80% at baseline and for the majority of treatment cycles through cycle 10 for both treatment groups. In later cycles, the number of expected PRO assessments declined largely because of disease progression or death, particularly in the chemotherapy group. Therefore, adherence to measurements was more variable at later cycles due to smaller sample sizes.⁴⁷

Change in Scores From Baseline Over Time for FACT-Bl

Baseline FACT-Bl scores were similar in the erdafitinib and chemotherapy groups. A mixed model for repeated measures was used to assess the change from baseline in FACT-Bl scores at each scheduled visit through cycle 11 in the 2 treatment groups. Results from the FACT-Bl assessment for all domains (Figure 8, Figure 9, Figure 10, Figure 11, Figure 12, Figure 13, Figure 14) and total scores (Figure 15) in cohort 1 showed that the effects on HRQoL measures were similar in the erdafitinib and chemotherapy groups.

Figure 8: Mean Change (± SE) From Baseline on the FACT-Bl Physical Well-Being Domain for Patients in Cohort 1 of the THOR Trial (ITT Analysis Set; January 15, 2023, Clinical Cut-Off)

ECOG PS = Eastern Cooperative Oncology Group Performance Status; FACT-Bl = Functional Assessment of Cancer Therapy–Bladder; HRQoL = health-related quality of life; ITT = intention-to-treat; PRO = patient-reported outcome; SE = standard error.

Notes: The primary model includes treatment, time point (categorical), treatment-by-time interaction, baseline PRO score, region, ECOG PS, and disease distribution. The model includes all scheduled visits through cycle 11.

The FACT-Bl is a 39-item questionnaire, with 5-point Likert scales, covering 5 primary domains: physical well-being, social/family well-being, emotional well-being, functional well-being, and additional concerns for patients with bladder cancer. The higher the score, the better the quality of life.

Patients with available PROs data in the ITT population were used.

Source: 2023 Clinical Study Report for the THOR study, cohort 1.⁴⁷

Figure 9: Mean Change (± SE) From Baseline on the FACT-Bl Social/Family Well-Being Domain for Patients in Cohort 1 of the THOR Trial (ITT Analysis Set; January 15, 2023, Clinical Cut-Off)

ECOG PS = Eastern Cooperative Oncology Group Performance Status; FACT-Bl = Functional Assessment of Cancer Therapy–Bladder; HRQoL = health-related quality of life; ITT = intention-to-treat; PRO = patient-reported outcome; SE = standard error.

Notes: The primary model includes treatment, time point (categorical), treatment-by-time interaction, baseline PRO score, region, ECOG PS, and disease distribution. The model includes all scheduled visits through cycle 11.

The FACT-Bl is a 39-item questionnaire, with 5-point Likert scales, covering 5 primary domains: physical well-being, social/family well-being, emotional well-being, functional well-being, and additional concerns for patients with bladder cancer. The higher the score, the better the quality of life.

Patients with available PROs data in the ITT population were used.

Source: 2023 Clinical Study Report for the THOR study, cohort 1.⁴⁷

Figure 10: Mean Change (± SE) From Baseline on the FACT-Bl Emotional Well-Being Domain for Patients in Cohort 1 of the THOR Trial (ITT Analysis Set; January 15, 2023, Clinical Cut-Off)

ECOG PS = Eastern Cooperative Oncology Group Performance Status; FACT-Bl = Functional Assessment of Cancer Therapy–Bladder; HRQoL = health-related quality of life; ITT = intention-to-treat; PRO = patient-reported outcome; SE = standard error.

Notes: The primary model includes treatment, time point (categorical), treatment-by-time interaction, baseline PRO score, region, ECOG PS, and disease distribution. The model includes all scheduled visits through cycle 11.

The FACT-Bl is a 39-item questionnaire, with 5-point Likert scales, covering 5 primary domains: physical well-being, social/family well-being, emotional well-being, functional well-being, and additional concerns for patients with bladder cancer. The higher the score, the better the quality of life.

Patients with available PROs data in the ITT population were used.

Source: 2023 Clinical Study Report for the THOR study, cohort 1.⁴⁷

Figure 11: Mean Change (± SE) From Baseline on the FACT-Bl Functional Well-Being Domain for Patients in Cohort 1 of the THOR Trial (ITT Analysis Set; January 15, 2023, Clinical Cut-Off)

ECOG PS = Eastern Cooperative Oncology Group Performance Status; FACT-Bl = Functional Assessment of Cancer Therapy–Bladder; HRQoL = health-related quality of life; ITT = intention-to-treat; PRO = patient-reported outcome; SE = standard error.

Notes: The primary model includes treatment, time point (categorical), treatment-by-time interaction, baseline PRO score, region, ECOG PS, and disease distribution. The model includes all scheduled visits through cycle 11.

The FACT-Bl is a 39-item questionnaire, with 5-point Likert scales, covering 5 primary domains: physical well-being, social/family well-being, emotional well-being, functional well-being, and additional concerns for patients with bladder cancer. The higher the score, the better the quality of life.

Patients with available PROs data in the ITT population were used.

Source: 2023 Clinical Study Report for the THOR study, cohort 1.⁴⁷

Figure 12: Mean Change (± SE) From Baseline on the FACT-Bl Bladder Cancer Symptoms Domain for Patients in Cohort 1 of the THOR Trial (ITT Analysis Set; January 15, 2023 Clinical Cut-Off)

ECOG PS = Eastern Cooperative Oncology Group Performance Status; FACT-Bl = Functional Assessment of Cancer Therapy–Bladder; HRQoL = health-related quality of life; ITT = intention-to-treat; PRO = patient-reported outcome; SE = standard error.

Notes: The primary model includes treatment, time point (categorical), treatment-by-time interaction, baseline PRO score, region, ECOG PS, and disease distribution. The model includes all scheduled visits through cycle 11.

The FACT-Bl is a 39-item questionnaire, with 5-point Likert scales, covering 5 primary domains: physical well-being, social/family well-being, emotional well-being, functional well-being, and additional concerns for patients with bladder cancer. The higher the score, the better the quality of life.

Patients with available PROs data in the ITT population were used.

Source: 2023 Clinical Study Report for the THOR study, cohort 1.⁴⁷

Figure 13: Mean Change (± SE) From Baseline on the FACT-Bl Trial Outcome Index^a for Patients in Cohort 1 of the THOR Trial (ITT Analysis Set; January 15, 2023 Clinical Cut-Off)

ECOG PS = Eastern Cooperative Oncology Group Performance Status; FACT-Bl = Functional Assessment of Cancer Therapy–Bladder; HRQoL = health-related quality of life; ITT = intention-to-treat; PRO = patient-reported outcome; SE = standard error.

Notes: The primary model includes treatment, time point (categorical), treatment-by-time interaction, baseline PRO score, region, ECOG PS, and disease distribution. The model includes all scheduled visits through cycle 11.

The FACT-Bl is a 39-item questionnaire, with 5-point Likert scales, covering 5 primary domains: physical well-being, social/family well-being, emotional well-being, functional well-being, and additional concerns for patients with bladder cancer. The higher the score, the better the quality of life.

Patients with available PROs data in the ITT population were used.

^aThe FACT-Bl Trial Outcome Index is calculated as the sum of the physical well-being, functional well-being, and bladder cancer symptom scores.

Source: 2023 Clinical Study Report for the THOR study, cohort 1.⁴⁷

Figure 14: Mean Change (± SE) From Baseline on the FACT-Bl and FACT-G Total Score^a for Patients in Cohort 1 of the THOR Trial (ITT Analysis Set; January 15, 2023 Clinical Cut-Off)

ECOG PS = Eastern Cooperative Oncology Group Performance Status; FACT-Bl = Functional Assessment of Cancer Therapy–Bladder; FACT-G = Functional Assessment of Cancer Therapy–General; HRQoL = health-related quality of life; ITT = intention-to-treat; PRO = patient-reported outcome; SE = standard error.

Notes: The primary model includes treatment, time point (categorical), treatment-by-time interaction, baseline PRO score, region, ECOG PS, and disease distribution. The model includes all scheduled visits through cycle 11.

The FACT-Bl is a 39-item questionnaire, with 5-point Likert scales, covering 5 primary domains: physical well-being, social/family well-being, emotional well-being, functional well-being, and additional concerns for patients with bladder cancer. The higher the score, the better the quality of life.

Patients with available PROs data in the ITT population were used.

^aThe FACT-Bl and FACT-G total score is calculated as the sum of the physical well-being, social/family well-being, emotional well-being, functional well-being scores.

Source: 2023 Clinical Study Report for the THOR study, cohort 1.⁴⁷

Figure 15: Mean Change (± SE) From Baseline on the FACT-Bl Total Score^a for Patients in Cohort 1 of the THOR Trial (ITT Analysis Set; January 15, 2023, Clinical Cut-Off)

ECOG PS = Eastern Cooperative Oncology Group Performance Status; FACT-Bl = Functional Assessment of Cancer Therapy–Bladder; HRQoL = health-related quality of life; ITT = intention-to-treat; PRO = patient-reported outcome; SE = standard error.

Notes: The primary model includes treatment, time point (categorical), treatment-by-time interaction, baseline PRO score, region, ECOG PS, and disease distribution. The model includes all scheduled visits through cycle 11.

The FACT-Bl is a 39-item questionnaire, with 5-point Likert scales, covering 5 primary domains: physical well-being, social/family well-being, emotional well-being, functional well-being, and additional concerns for patients with bladder cancer. The higher the score, the better the quality of life.

Patients with available PROs data in the ITT population were used.

^aThe FACT-Bl total score is calculated as the sum of the physical well-being, social/family well-being, emotional well-being, functional well-being, and bladder cancer symptom scores.

Source: 2023 Clinical Study Report for the THOR study, cohort 1.⁴⁷

Change in Scores From Baseline Over Time for PGI-S

At cycle 1, day 1, 16.7% and 26.2% of patients in the erdafitinib treatment group reported disease severity as none or mild, respectively, whereas in the chemotherapy group, 28.3% and 23.6% of patients reported disease severity as none or mild, respectively. As treatment progressed, the overall proportion of patients who reported disease severity as none or mild decreased in both treatment groups. By the end of treatment, 16.7% and 18.5% of patients in the erdafitinib treatment group, and 20.7% and 29.3% of patients in the chemotherapy group, reported disease severity as none or mild, respectively (Table 19).

Table 19: Proportion of Responses on the PGI-S (All-Treated Analysis Set; January 15, 2023, Clinical Cut-Off)

PGI-S = Patient Global Impression of Severity.

Source: 2023 Clinical Study Report for the THOR study, cohort 1.⁴⁷

Change in Scores From Baseline Over Time for EQ-5D-5L

Baseline scores for the EQ-5D-5L health utility index and visual analogue scale assessment were similar across treatment groups. Results showed that general HRQoL and overall health status were maintained on treatment for patients in each treatment arm without evidence of a difference between them, with the caveat that few patients completed both baseline and follow-up assessments routinely, so the models were underpowered to find any differences. HRQoL and overall health status were generally similar in both treatment arms over the course of the study (Figure 16 and Figure 17).

Figure 16: Mean Change (± SE) From Baseline on the EQ-5D-5L Health Utility Score VAS Score Domain for Patients in Cohort 1 of the THOR Trial (ITT Analysis Set; January 15, 2023 Clinical Cut-Off)

ECOG PS = Eastern Cooperative Oncology Group Performance Status; HRQoL = health-related quality of life; ITT = intention-to-treat; PRO = patient-reported outcome; SE = standard error; VAS = visual analogue scale.

Notes: The primary model includes treatment, time point (categorical), treatment-by-time interaction, baseline PRO score, region, ECOG PS, and disease distribution. The model includes all scheduled visits through cycle 11.

The EQ-5D-5L is a generic health measure that consists of 5 items covering 5 domains, with lower scores indicating better quality of life. A health utility score is also calculated, with values anchored at 0 (a state as bad as being dead) and 1 (full health).

Patients with available PROs data in the ITT population were used.

Source: 2023 Clinical Study Report for the THOR study, cohort 1.⁴⁷

Harms

The safety of erdafitinib and chemotherapy in the THOR trial were assessed in the safety analysis set. Safety was measured with a monitoring medical review of AE reports and the results of vital sign measurements, electrocardiograms, physical examinations, clinical laboratory tests, ECOG PS, ophthalmologic examinations, Amsler grid test, and other safety evaluations at specified time points. The baseline value for safety assessment was defined as the value collected at the time closest to, but before, administration of the first dose of the study drug. In instances where there were multiple records at a given visit date for lab parameters associated with disease assessment, the most recent lab value was selected as the unique lab value for analysis.

AEs are those considered in the trial to be treatment-emergent adverse events (TEAEs), which were coded in accordance with the Medical Dictionary for Regulatory Activities (MedDRA), version 24.1, using the lower-level term as the description most closely related to the investigator’s terminology, a preferred term describing a group of closely related lower-level terms, and the system organ class, which is the broad category that includes related preferred terms. The severity of all AEs was graded in accordance with the National Cancer Institute Common Terminology Criteria for Adverse Events, version 4.03.

Figure 17: Mean Change (± SE) From Baseline on the EQ-5D-5L Health Utility Score Health Utilities Index Domain for Patients in Cohort 1 of the THOR Trial (ITT Analysis Set; January 15, 2023, Clinical Cut-Off)

ECOG PS = Eastern Cooperative Oncology Group Performance Status; HRQoL = health-related quality of life; ITT = intention-to-treat; PRO = patient-reported outcome; SE = standard error.

Notes: The primary model includes treatment, time point (categorical), treatment-by-time interaction, baseline PRO score, region, ECOG PS, and disease distribution. The model includes all scheduled visits through cycle 11.

The EQ-5D-5L is a generic health measure that consists of 5 items covering 5 domains, with lower scores indicating better quality of life. A health utility score is also calculated, with values anchored at 0 (a state as bad as being dead) and 1 (full health).

Patients with available PROs data in the ITT population were used.

Source: 2023 Clinical Study Report for the THOR study, cohort 1.⁴⁷

Adverse Events

At least 1 AE was experienced by 133 of 135 patients (98.5%) in the erdafitinib group. The most frequently reported AEs by preferred term, reported by at least 10% of patients in the erdafitinib group, were hyperphosphatemia (80.0%), diarrhea (62.2%), and stomatitis (48.1%), as depicted in Table 20, Table 21, and Table 22.

In the chemotherapy group, at least 1 AE was experienced by 109 of 112 patients (97.3%). The most frequently reported AEs by preferred term, reported by at least 10% of patients in the chemotherapy group, were anemia (32.1%), constipation (27.7%), and asthenia (25.0%).

Grade 3 to Grade 4 AEs

A total of 85 of 135 patients (63.0%) in the erdafitinib group and 72 of 112 patients (64.3%) in the chemotherapy group experienced at least 1 grade 3 or grade 4 AE. Grade 3 and grade 4 blood system AEs were more common in the chemotherapy group, and no patients in the erdafitinib group experienced neutropenia, febrile neutropenia, or leukopenia.

Table 20: Summary of AEs in Cohort 1 of the THOR Trial (Safety Analysis Set; January 15, 2023, Clinical Cut-Off)

AE = adverse event.

Note: For AEs with an incidence of ≥ 10% in any treatment group, patients are counted only once for any given event, regardless of the number of times they actually experienced the event. AEs are coded using the Medical Dictionary for Regulatory Activities (MedDRA), version 24.1.

Source: 2023 Clinical Study Report for the THOR study, cohort 1.⁴⁷

Table 21: Summary of Deaths in Cohort 1 of the THOR Trial (ITT Analysis Set; January 15, 2023, Clinical Cut-Off)

ITT = intention-to-treat.

^aIncludes deaths that were very likely, probably, or possibly related to the study drug.

^bIncludes doubtful, unknown, or missing relationship to the study drug.

Source: 2023 Clinical Study Report for the THOR study, cohort 1.⁴⁷

Table 22: AEs Leading to Death in Cohort 1 of the THOR Trial (Safety Analysis Set; January 15, 2023, Clinical Cut-Off)

AE = adverse events.

Note: Patients are counted only once for any given event, regardless of the number of times they actually experienced the event. The event experienced by the patient with the worst toxicity is used. If a patient has missing toxicity for a specific AE, the patient is only counted in the total column for that AE. AEs are coded using the Medical Dictionary for Regulatory Activities (MedDRA), version 24.1.

Source: 2023 Clinical Study Report for the THOR study, cohort 1.⁴⁷

Serious Adverse Events

In cohort 1 of the THOR trial, 56 of 135 patients (41.5%) in the erdafitinib treatment group and 47 of 112 patients (42.0%) in the chemotherapy treatment group experienced an SAE, as described in Table 23. The most frequently reported SAE by preferred term reported by at least 2% of patients in the erdafitinib group were urinary tract infection (4.4%) and hematuria (3.7%). The most frequently reported SAE by preferred term reported by at least 2% of patients in the chemotherapy group were febrile neutropenia (6.3%) and febrile bone marrow aplasia (3.6%).

Table 23: Summary of SAEs in Cohort 1 of the THOR Trial (Safety Analysis Set; January 15, 2023, Clinical Cut-Off)

AE = adverse event; SAE = serious adverse event.

Note: Patients are counted only once for any given event, regardless of the number of times they actually experienced the event. AEs are coded using the Medical Dictionary for Regulatory Activities (MedDRA), version 24.1.

Source: 2023 Clinical Study Report for the THOR study, cohort 1.⁴⁷

Withdrawals Due to Adverse Events

In cohort 1 of the THOR trial, in the erdafitinib group, AEs led to the discontinuation of the study drug in 19 of 135 patients (14.1%) and to an interruption of the study drug in 97 of 135 patients (71.9%) (Table 24). In the chemotherapy group, AEs led to the discontinuation of the study drug in 20 of 112 patients (17.9%) and to an interruption of the study drug in 35 of 112 patients (31.3%). A lower percentage of patients in the erdafitinib group (11 patients [8.1%]) experienced AEs that led to treatment discontinuation considered to be related to the study drug than the chemotherapy group (15 patients [13.4%]).

The most frequently reported AEs leading to treatment discontinuation by preferred term in the erdafitinib group were general physical health deterioration (1.5%) and sudden death (1.5%). The most frequently reported AEs leading to discontinuation by preferred term in the chemotherapy group were febrile bone marrow aplasia (2.7%), general physical health deterioration (1.8%), septic shock (1.8%), and diarrhea (1.8%).

Table 24: Summary of AEs Leading to Treatment Discontinuation in Cohort 1 of the THOR Trial (January 15, 2023, Clinical Cut-Off)

AE = adverse event; TEAE = treatment-emergent adverse event.

Note: Patients are counted only once for any given event, regardless of the number of times they actually experienced the event. AEs are coded using the Medical Dictionary for Regulatory Activities (MedDRA), version 24.1.

Source: 2023 Clinical Study Report for the THOR study, cohort 1.⁴⁷

AEs of Special Interest

Based on input from clinical experts and others involved (e.g., presenter members from the pan-Canadian Oncology Drug Review Expert Review Committee), 3 AEs of special interest were deemed important: CSR and eye disorders, hyperphosphatemia, and nail and/or skin disorders.

CSR has been identified as an AE of special interest and is a known class effect of FGFR inhibitors. This “AEs of special interest” grouped term consists of the following preferred terms: retinal detachment, vitreous detachment, retinal edema, retinopathy, chorioretinopathy, detachment of retinal pigment epithelium, detachment of macular retinal pigment epithelium, macular detachment, serous retinal detachment, subretinal fluid, retinal thickening, chorioretinitis, serous retinopathy, maculopathy, and choroidal effusion. In cohort 1 of the THOR trial, events of CSR were reported for 23 of 135 patients (17.0%) in the erdafitinib treatment group. No patients in the chemotherapy treatment group experienced events of CSR. The most frequent AEs of CSR in the erdafitinib group reported by at least 3% of patients were chorioretinopathy (5.9%), detachment of retinal pigment epithelium (5.2%), and subretinal fluid (3.7%) (Table 25). The majority of serous retinopathy events were grade 1 or 2 (20 of 135 patients [14.8%]). Three patients (2.2%) experienced a grade 3 event. No patients experienced a grade 4 serous retinopathy event. Events of CSR were manageable with dose modifications; dose reductions were reported for 15 patients (11.1%) and dose interruptions were reported for 7 patients (5.2%). Events of CSR were resolved in the majority of patients (16 of 23 patients [69.6%]) at the time of clinical cut-off; however, 7 patients had an ongoing event, the majority of which were grade 1.

Eye disorders (other than CSR), such as conjunctivitis and dry eye, occurred in 57 of 135 patients (42.2%) in the erdafitinib group and in 6 of 112 patients (5.4%) in the chemotherapy group.

Hyperphosphatemia was also considered of special interest by CDA-AMC, in agreement with clinical experts. In cohort 1, 108 of 135 patients (80%) in the erdafitinib group and 0 patients in the chemotherapy group experienced an event of hyperphosphatemia. The majority of cases were considered grade 1 to 2 (101 patients). Only 1 patient was assessed with a grade 4 event.

The overall incidence of nail disorders was reported in 90 patients (66.7%) in the erdafitinib group and in 6 patients (5.4%) in the chemotherapy group. Similarly, skin disorders were reported in 74 patients (54.8%) and 14 patients (12.5%), respectively.

Table 25: Summary of AEs of Special Interest in Cohort 1 of the THOR Trial (Safety Analysis Set; January 15, 2023, Clinical Cut-Off)

AE = adverse event.

Notes: Patients are counted only once for any given event, regardless of the number of times they actually experienced the event. AEs are coded using the Medical Dictionary for Regulatory Activities (MedDRA), version 24.1.

AEs of special interest include the following MedDRA preferred terms: retinal detachment, vitreous detachment, retinal edema, retinopathy, chorioretinopathy, detachment of retinal pigment epithelium, detachment of macular retinal pigment epithelium, macular detachment, serous retinal detachment, subretinal fluid, retinal thickening, chorioretinitis, serous retinopathy, maculopathy, choroidal effusion.

Source: 2023 Clinical Study Report for the THOR study, cohort 1.⁴⁷

Critical Appraisal

Internal Validity

Overall, the THOR study is a well-designed RCT comparing erdafitinib to chemotherapy in patients with la/mUC harbouring FGFR alterations who have been previously treated with anti-PD-1 or anti-PD-L1 therapies.

Although randomization and allocation concealment were properly conducted, some baseline imbalances were observed, including the number of patients who declined chemotherapy (1 in the erdafitinib group versus 18 in the chemotherapy group) and the number of previous lines of chemotherapy received. Although these imbalances may introduce bias into the results, sensitivity analyses that excluded patients who did not initiate chemotherapy (1 versus 18) supported the robustness of the main study findings, minimizing concerns about potential bias that originated from the randomization process.

As an open-label trial, the absence of blinding presents potential bias risks, particularly for subjective outcomes like PROs and AEs. Notably, there were significant differences in the use of concomitant therapies, such as antibiotics, steroids, and gastrointestinal drugs, which could have been influenced by the open-label design. These concomitant treatments might have affected both the reporting and actual event rates of AEs, such as a potential reduction in gastrointestinal events in the treatment arm related to increased gastrointestinal drug use. Even for more objective outcomes, such as OS, the lack of blinding and these treatment imbalances might introduce bias, although the full impact is likely small. It is also important to acknowledge that the open-label design was necessary because of the nature of drug administration, making blinding difficult to implement.

The PRO measures (i.e., FACT-Bl, PGI-S, EQ-5D-5L) were considered valid, reliable, and responsive in this specific population; only the measurement for the time to urinary bladder cancer symptom deterioration lacked specific details on validity, reliability, and responsiveness.

The trial also included an interim analysis that led to early stopping because of demonstrated efficacy. Although this decision is ethically sound, early termination can sometimes result in an overestimation of treatment effects due to the shorter follow-up. However, the use of a prespecified interim analysis with defined stopping rules mitigates some of these concerns.

Last, the long-term effects of erdafitinib remain unknown, owing to the reduced sample sizes and censoring of patients after 12 months, which limited the reliability of assessments conducted beyond this period. This is particularly important for estimates of PFS and OS, where the CIs become notably wide beyond 6 months.

External Validity

The study included participants from 23 countries on multiple continents, providing a diverse population. However, there was an underrepresentation of certain demographic groups, particularly Black patients. In Canada, with its multicultural population, the limited representation of certain ethnic groups in the study may affect the generalizability of the findings. The median age of participants was around 66 to 69 years, which is consistent with the typical age range for patients with UC in Canada.

The study focuses on patients with FGFR3 and/or FGFR2 alterations, which is a subset of patients with UC. The generalizability of the results depends on the availability and accessibility of molecular testing for FGFR alterations in Canada.

The study compares erdafitinib to chemotherapy, specifically docetaxel or vinflunine. In Canada, cisplatin-based chemotherapy remains the standard of care for advanced UC; alternatives depend on patient eligibility. The standard-of-care options in the study are consistent with those available in Canada, further supporting the external validity.

Overall, the clinical experts consulted by CDA-AMC considered the results of the THOR study to be applicable to most patients with la/mUC in Canada, within the submission criteria assessed.

GRADE Summary of Findings and Certainty of the Evidence

Methods for Assessing the Certainty of the Evidence

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

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

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

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

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

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

The GRADE assessments are presented in Table 2.

Long-Term Extension Studies

No long-term extension studies were included in the sponsor’s submission to CDA-AMC.

Indirect Evidence

The contents of this section have been informed by materials submitted by the sponsor. The following information has been summarized and validated by the CDA-AMC review team.

Objectives for the Summary of Indirect Evidence

The indirect comparison submitted by the sponsor aimed to compare the relative efficacy and safety of erdafitinib and enfortumab vedotin in patients with la/mUC who progressed after 1 or 2 prior treatments, at least 1 of which included an anti-PD-L1 drug, with any other platinum-based, or appropriate alternative therapies.

Description of Indirect Comparison

To estimate the efficacy of erdafitinib relative to the other treatment options for patients with la/mUC who received at least 1 prior systemic therapy, 1 of which was ICI therapy, a comprehensive systematic literature review (SLR) was performed that followed the search strategy and criteria assessment described in Table 26.

Table 26: Study Selection Criteria and Methods for Indirect Comparisons