Comparison of Relapse Risk Prediction Models During Active Surveillance in Stage I Seminoma: A Retrospective Multicenter Cohort Study

To improve the prediction of relapse for clinical stage I seminoma, we aimed to validate the Boormans model (incorporating tumor size, rete testis invasion, and lymphovascular invasion) against the conventional classification (tumor size > 4 cm/rete testis invasion). Consecutive cSI seminoma patients managed with active surveillance after orchidectomy across multiple centers were retrospectively analyzed (1994-2024). Part of the data was provided by the prospective Swiss-Austrian-German Testicular Cancer Cohort Study Group. The primary outcome was relapse. Among the 1025 patients, 116 (11.3%) relapsed during a median follow-up of 45 months (95% CI 42-48). In the descriptive multivariable analysis, TS (HR 1.03, 95% CI 1.01-1.04) and RTI (HR 2.73, 95% CI 1.83-4.06) were found to be independent predictors of relapse. The Boormans model classified 645 (62.9%) men as low-, 356 (34.7%) as intermediate-, and 24 (2.3%) men as high-risk candidates. The 5-year relapse risk was 0.07 (95% CI 0.04-0.09), 0.24 (0.19-0.29), and 0.35 (0.10-0.53) for the low-, intermediate-, and high-risk groups, respectively. The Boormans model outperformed the conventional model (C-index 0.66 vs. 0.61; Δ0.06, p < 0.001). Despite improved discrimination, clinical utility was limited across risk-adapted treatment scenarios compared with the conventional model. The Boormans model demonstrated superior discrimination but limited gains in clinical utility. Integrating novel biomarkers may enable more accurate, risk-adapted management.