Visualizing Baseline Eligibility Across Locally Advanced or Metastatic Urothelial Cancer Trials and Real-World Cohorts.

Abstract

Clinical trials in locally advanced or metastatic urothelial carcinoma (mUC) often enroll younger and fitter patients than those treated in routine practice, limiting applicability to real-world populations. We developed a simple descriptive metric, the eligibility gap (EG) score, to place trial and real-world cohorts on a common scale and summarize differences in baseline eligibility profiles. Pivotal clinical trials were identified through a PubMed search, and three real-world datasets from Japan, the United States, and Spain were included. The EG score used three consistently reported domains: age, ECOG performance status, and cisplatin eligibility. Each domain was scaled from 0 to 33.3 points and summed to yield a score from 0 to 99.9, with higher scores indicating younger, fitter populations. Robustness was assessed across alternative model specifications, including prespecified weighting schemes, random-weight simulations, alternative age and ECOG scoring functions, and one-domain-out analyses. EG scores were generally higher in cisplatin-based trials (74.8-90.1) than in immune checkpoint inhibitors (ICI)-based trials (52.3-81.2) and real-world cohorts (35.1-45.5), while the cisplatin-ineligible trial had the lowest score (24.5). Alternative age transformations and a linear ECOG model did not change ranking. In one-domain-out analyses, omission of ECOG preserved ranking, whereas omission of age or cisplatin-fitness produced minor shifts. Across prespecified weighting schemes, rank-order concordance remained high (Spearman's rho = 0.982-1.000), and 10 000 random-weight simulations showed a median rho of 0.982. The EG score may provide a simple descriptive summary of eligibility domains across mUC studies, with stable cohort ordering across alternative model specifications.