Assessment of Surrogate Endpoints in Real-World Oncology Studies, a Systematic Literature Review

OBJECTIVES: The use of surrogate endpoints in oncology clinical trials is increasingly accepted as a basis for supporting regulatory and clinical practice decision-making purposes (e.g., when using clinical endpoints is impractical/infeasible). With the growing role of real-world data/evidence (RWD/E) in supporting regulatory and payor decisions, it is critical to establish congruent surrogate endpoints in RWD settings. We conducted a systematic review to characterize surrogate endpoints and their performance/validation in oncology RWE studies.

METHODS: We searched the literature from PubMed (till 2023) using keywords related to oncology, RWD/E, surrogate endpoints, and validation. We included eligible studies if a surrogate endpoint was assessed against its corresponding clinical endpoint. Eligibility screening for each paper was conducted independently by at least two reviewers. We extracted the type of cancer, surrogate endpoint assessed, clinical endpoint validated, and validation-related parameters.

RESULTS: After title and abstract screening, 76 papers out of 643 qualified for full-text screening from which 44 papers were included for data abstraction. The leading three cancer types were prostate cancer (n=15), urogenital cancers (n=7), and colorectal cancer (n=5). Surrogate endpoints included biochemical response (e.g., prostate-specific antigen [PSA]) (n=13), time-based endpoints (i.e., progression/disease/metastatic-free survival (PFS/DFS/MFS)) (n=13), clinical complete/partial response (n=9), and others including pathological response (n=9). Overall survival was the most commonly validated clinical endpoint (n=33). The hazard ratio was the most widely used parameter for validation (n=23), followed by C-statistics (n=4), Spearman correlation (n=3), R square (n=3), and Kappa statistics (n=3), and others (n=7).

CONCLUSIONS: A variety of surrogate endpoints are assessed in real-world oncology studies, especially among prostate and urogenital cancers. PSA changes and PFS/DFS/MFS were mostly assessed endpoints against overall survival. Comprehensive efforts are needed to validate surrogate endpoints in RWE settings.