Evaluating prognostic models for stage I seminoma within the randomised trial of imaging and surveillance in seminoma testis (TRISST).

Abstract

501 Background: Prognosis in stage I seminoma is excellent. Most patients are cured by orchiectomy and, for the 15%-20% that relapse, salvage treatment is generally successful (survival approaching 100%). Relapse risk is reduced with adjuvant carboplatin but may unnecessarily expose young patients to long-term health risks. Effective and reliable risk stratification could guide use of adjuvant treatment and potentially reduce intensity of surveillance for the lowest risk patients. We used data from a UK multicentre, phase III trial to evaluate prognostic factors for relapse in stage I seminoma, including validation of a recently proposed model from the European Association of Urology (EAU)(1). Methods: TRISST used a randomised, factorial design to evaluate imaging schedules (3 vs 7 scans) and modality (MRI vs CT) in surveillance for men who had undergone orchidectomy for stage I seminoma (n=669). In this secondary analysis, model selection was used to fit a multivariable Cox regression to TRISST data to investigate factors predicting time to relapse, including: age, tumour mass size, rete testis invasion (RTI), T stage, lymphovascular invasion (LVI), side of tumour, and tumour markers (LDH, β-HCG). The final model was used to divide the cohort into risk groups. Relapse risk was also estimated according to EAU prognostic groups. Harrell’s C-index was used to measure model fit. Results: Based on TRISST data, tumour size ≥4cm and pT3 disease were associated with the highest risk (multivariable hazard ratio for ≥4cm vs <2cm: HR=4.00, 95% confidence interval 2.00-8.01; pT3 vs pT1: HR=3.89, 1.77-8.57). Patients with one or both of these features (24% of the cohort) had 5-year relapse rate of 22.0%. Medium risk (age<30 years and/or 2-4cm and/or pT2) and low risk (age≥30 years, <2cm and pT1) groups had 5-year relapse rates 12.0% and 2.3% respectively. The EAU model was a good fit to the TRISST data (C-index=0.62); 66%, 33% and 2% fell into the very low, low and high-risk EAU groups respectively with 5-year relapse rates of 10.1%, 16.1% and 45.5%. Conclusions: In stage I seminoma, relapse risk in patients managed with surveillance following orchiectomy is low for the majority of patients. Tumour size is consistently shown to be the best predictor of relapse. The EAU prognostic model, validated here, identifies a small group (>5cm, with RTI and LVI) with particularly high risk who may benefit from adjuvant therapy. Reference: 1. Boormans J, et al. Journal of Clinical Oncology 2023; 41 (6_suppl): 410. Clinical trial information: ISRCTN65987321 .