Predictors of early termination in phase 2–3 genitourinary oncology trials: A two-decade analysis of clinicaltrials.gov.

Abstract

331 Background: Early termination of clinical trials delays progress and wastes resources. Predictors of such discontinuation remain poorly characterized. We analyzed two decades of genitourinary (GU) oncology trials to identify trial-level factors associated with completion versus early termination. Methods: Interventional Phase 2–3 GU oncology trials (prostate, bladder, kidney) registered on ClinicalTrials.gov between 2005–2025 were included if conducted in U.S. adults and listed as Completed or Terminated. Withdrawn, suspended, and observational studies were excluded. The primary outcome was trial completion (Completed vs. Terminated). Predictors included trial design (phase, randomization, and masking), primary sponsor type (large pharmaceutical company vs. small/non-industry), intervention category, geographic scope (multisite vs. multinational), and presence of a Data Monitoring Committee (DMC). Associations were evaluated using multivariable logistic regression with robust standard errors. Results: Among 1,597 Phase 2–3 GU oncology trials (445 [27.9%] terminated; 1,152 [72.1%] completed), the leading causes of termination were poor accrual (41.8%), unknown/other (21.6%), and sponsor cancellation (8.5%). On multivariable analysis, prostate cancer trials were more likely to complete than bladder studies (OR 1.84, 95% CI 1.29–2.61, p<0.001). Higher completion odds were also observed for multisite (OR 1.58, 95% CI 1.19–2.10, p=0.002), multinational (OR 1.73, 95% CI 1.16–2.56, p=0.006), and large pharma–sponsored trials (OR 1.67, 95% CI 1.01–2.75, p=0.045). Conversely, randomized designs (OR 0.67, 95% CI 0.52–0.88, p=0.002) and trials with a DMC (OR 0.69, 95% CI 0.53–0.90, p=0.008) were linked to early termination. Trial phase, disease stage, masking, intervention type, line of therapy, and number of arms were not significantly correlated with completion. Conclusions: About one-quarter of Phase 2–3 GU oncology trials end prematurely. Prostate, multisite, multinational, and large industry-sponsored studies are more likely to be completed, whereas randomized designs with DMC oversight are at higher risk of early termination. These findings highlight modifiable design and operational factors to improve feasibility and success in future GU oncology trials. Trial characteristics by completion status (N = 1,597). Characteristic, n (%) Total Terminated Completed Disease site  Bladder 274 (17.2) 91 (20.4) 183 (15.9)  Kidney 354 (22.2) 109 (24.5) 245 (21.3)  Prostate 969 (60.7) 245 (55.1) 724 (62.8) Phase 3 (vs 2) 262 (16.4) 58 (13.0) 204 (17.7) Randomized 844 (52.8) 244 (54.8) 600 (52.1) Large Pharma primary sponsor 194 (12.1) 30 (6.7) 164 (14.2) Multisite (>1 site) 960 (60.1) 233 (52.4) 727 (63.1) Multinational 346 (21.7) 68 (15.3) 278 (24.1) DMC present 816 (51.1) 261 (58.7) 555 (48.2) Drug-based intervention 1,417 (88.7) 406 (91.2) 1,011 (87.8)