Visualizing eligibility gaps between muscle-invasive bladder cancer trials and real-world patients.

OBJECTIVES

Clinical studies in muscle-invasive bladder cancer (MIBC) often enroll younger and fitter patients than those treated in routine practice, limiting generalizability. We developed a simple eligibility gap (EG) score to quantify baseline eligibility differences between clinical studies and real-world cohorts.

METHODS

Eight clinical studies and eight real-world datasets were analyzed. The EG score was based on three commonly reported determinants: age, Eastern Cooperative Oncology Group performance status (ECOG PS), and cisplatin eligibility, each scaled from 0 to 33.3 points (total range: 0-99.9). Higher scores indicate younger, fitter, and more cisplatin-eligible populations. Known-groups validity was assessed by comparing EG scores across prespecified study categories expected to differ in eligibility selectivity. Sensitivity analyses using alternative fixed weighting schemes and random-weight simulations were performed to examine score robustness.

RESULTS

Baseline characteristics differed between clinical studies and real-world cohorts. Trial populations were generally younger, had better ECOG PS, and showed higher cisplatin eligibility than real-world cohorts. EG scores followed the expected ordering across prespecified study categories, with the highest values in cisplatin-fit trials (86), intermediate values in real-world cohorts (66), and the lowest values in cisplatin-unfit/refusal trials (44), consistent with known-groups validity. Domain-specific analyses showed larger differences in age and cisplatin eligibility than in ECOG PS. Across alternative weighting scenarios, rank correlations with the equal-weight model remained high (ρ = 0.976-0.994), and the expected category-level ordering was preserved.

CONCLUSIONS

The EG score describes eligibility differences between clinical studies and real-world MIBC populations and may aid interpretation of trial evidence across heterogeneous populations.