An exploratory analysis of the phase II KEYNOTE-057 trial, presented at the 2024 Society of Urologic Oncology (SUO) annual meeting, investigated post-treatment outcomes in patients with Bacillus Calmette-Guérin (BCG)-unresponsive, high-risk (HR) non-muscle invasive bladder cancer (NMIBC) who did not respond to pembrolizumab monotherapy. The study aimed to assess whether the timing of radical cystectomy following pembrolizumab failure impacts patient outcomes.
Radical cystectomy is the standard treatment for BCG-unresponsive HR NMIBC, but many patients are unsuitable for or decline surgery, opting for bladder-sparing treatments. KEYNOTE-057 previously demonstrated pembrolizumab's potential as a bladder-sparing option in these patients. This analysis focused on patients who experienced disease progression (PD) despite pembrolizumab, examining the impact of subsequent treatments, including radical cystectomy and other bladder-sparing approaches.
The study categorized 144 participants into three groups: early radical cystectomy (within 4 months of confirmed treatment failure), delayed radical cystectomy (surgery >4 months after failure or bladder-sparing treatment before surgery), and no radical cystectomy (bladder-sparing treatment alone or no subsequent therapy). Confirmed non-response to pembrolizumab was defined as persistent or recurrent high-grade NMIBC or progression to muscle-invasive bladder cancer (MIBC) or metastatic disease.
Endpoints included time to radical cystectomy or bladder-sparing treatment, progression-free survival (PFS), metastasis-free survival (MFS), and overall survival (OS). PFS, MFS, and OS were estimated using the Kaplan-Meier method, with no formal hypothesis testing performed. Data cutoff dates were May 30, 2023 (cohort A) and October 20, 2022 (cohort B).
Baseline characteristics showed that patients who did not undergo cystectomy were older (75 years vs. 66–69 years) and had a worse ECOG performance status (1-2: 32% vs. 15-21%). Patients in the early cystectomy group were more likely to have refractory disease (39% vs. 18–26%).
The time from confirmed pembrolizumab non-response to subsequent therapy was numerically longer in participants who did not undergo radical cystectomy or received other subsequent therapy, compared with those who underwent early or delayed surgery.
The 36-month PFS rate was 86% in the no radical cystectomy group, compared to 72% and 68% in the early and delayed surgical groups. Metastasis-free survival rates were also superior in the no radical cystectomy group (at 36 months: 93% versus 87% and 79%, respectively). The highest overall survival rates were in the delayed surgical group, which may reflect a survival bias.
PFS rates were similar regardless of whether disease recurrence occurred within 4 months or ≥4 months of pembrolizumab initiation. Additionally, PFS was not markedly different based on the duration of bladder-sparing treatment in participants who underwent delayed versus no radical cystectomy.
Pathological outcomes were also similar between the early and delayed radical cystectomy groups.
According to Dr. Li, clinical outcomes, including PFS, MFS, and OS, were generally similar between participants with HR NMIBC who underwent early, delayed, or no radical cystectomy after confirmed pembrolizumab non-response. The time from pembrolizumab non-response to subsequent therapy was numerically longer in participants who did not undergo a radical cystectomy or received other subsequent therapy, compared with those who underwent early or delayed surgery. PFS was not markedly different based on duration of bladder-sparing treatment in participants who underwent a delayed or no radical cystectomy, or between participants who experienced disease recurrence within 4 months versus ≥4 months of pembrolizumab initiation in all participants. Participants who underwent early versus delayed radical cystectomy had similar pathological outcomes.
Dr. Li cautioned that the results should be interpreted carefully due to the post hoc nature of the analysis, the small sample size of the subgroups, and potential biases related to disease characteristics, prognosis, and physician/patient preferences influencing treatment decisions. However, the data suggest that implementing other second-line bladder-sparing therapies after pembrolizumab non-response may be a feasible option to delay or avoid radical cystectomy.
