Evaluating Disease-Free Survival (DFS) As a Surrogate Endpoint (SE) for Overall Survival (OS) in Muscle-Invasive Urothelial Carcinoma (MIUC): An Analysis of Surveillance, Epidemiology, and End Results (SEER)-Medicare Data


Leung L1, Kanters S1, Pourrahmat MM1, Friedman H2, Navaratnam P3, Reardon G4, Patel M5, Teitsson S6, Kurt M7, Lotan Y8
1Evidinno Outcomes Research Inc., Vancouver, BC, Canada, 2DataMed Solutions LLC, New York, NY, USA, 3DataMed Solutions LLC, New York City, NY, USA, 4Keck Graduate Institute, Claremont, CA, USA, 5Bristol Myers Squibb, Summit, NJ, USA, 6Bristol Myers Squibb, Uxbridge, UK, 7Bristol Myers Squibb, Princeton, NJ, USA, 8University of Texas Southwestern Medical Center, Dallas, TX, USA

Presentation Documents

OBJECTIVES: To assess DFS as an SE for OS in adults with MIUC who have undergone radical resection using SEER-Medicare registry data.

METHODS: Patients who had radical resection between 2009-2017 following primary diagnosis of MIUC at high risk of recurrence were analyzed (N=1038; median follow-up=29.5 months). Individual-level association was measured via Spearman’s rank correlation. Patients were categorized in 18 exclusive clusters to define synthetic treatment and control arms based on treatments received and baseline characteristics (e.g., age, sex, resection date, disease staging, and race/ethnicity). Propensity score matching addressed imbalances in baseline characteristics between treatment and control arms. Treatment effects on each endpoint, measured by hazard ratios (HRs), were computed via Cox models and the correlation between them, assessed by coefficient of determination, was estimated via weighted linear regression. Accuracy of the surrogacy equation in predicting OS HR from DFS HR was assessed internally via cross validation and externally on 9 randomized controlled trials (RCTs).

RESULTS: Patients were mostly male (72%), non-Hispanic white (85.5%), older than age 65 (median age=74.1 years) and in stages 2, 3, and 4 (29.3%, 43.6%, and 27.1%). In addition to resection, patients received neoadjuvant (39.1%), adjuvant (13.2%), or perioperative (2.6%) therapy consisting mainly of combination chemotherapy regimens (96.3%, 92.7%, and 74.1% of patients within each therapy setting, respectively). The remaining patients had resection only with no additional systemic therapies (45.1%). Individual- and treatment-effect-level correlations were estimated as 0.85 (95% CI: 0.79-0.90) and 0.83 (95% CI: 0.56-0.94), respectively. The derived surrogacy equation was log(HROS)=-0.04+1.08×log(HRDFS). The 95% prediction intervals on OS HRs contained observed OS HRs in 89% of the clusters and 67% of the RCTs.

CONCLUSIONS: Correlations between the endpoints, and between the treatment effects on them, were both moderate. Estimated surrogacy equation may assist earlier assessments of OS benefit from DFS benefit for treatment of MIUC in the real-world setting.

Conference/Value in Health Info

2023-05, ISPOR 2023, Boston, MA, USA

Value in Health, Volume 26, Issue 6, S2 (June 2023)




Clinical Outcomes

Topic Subcategory

Relating Intermediate to Long-term Outcomes



Your browser is out-of-date

ISPOR recommends that you update your browser for more security, speed and the best experience on ispor.org. Update my browser now