PS05: Advancing Biomarker-Driven Intermediate Endpoints in Oncology: Current Status and Future Directions

The accelerating pace of oncology drug development has created an urgent need for surrogate endpoints that can enable earlier regulatory decisions while preserving reliable inference on long-term clinical benefit. Although overall survival (OS) remains the gold standard, its evaluation requires prolonged follow-up. Historically, overall response rate (ORR) has supported accelerated approval in many oncology settings. However, with modern therapies achieving ORRs exceeding 80% in several malignancies, the discriminatory power of ORR has diminished. This ceiling effect highlights the need for novel biomarker-driven surrogate endpoints that capture deeper and more durable treatment effects. Across solid tumors, pathologic complete response (pCR) in breast cancer and circulating tumor DNA (ctDNA) across multiple tumor types have emerged as promising biomarker-driven endpoints. Recent FDA Oncology Drugs Advisory Committee (ODAC) discussions on ctDNA-guided treatment in metastatic HR-positive/HER2-negative breast cancer further highlighted both the promise of ctDNA and the remaining challenges in establishing its clinical utility and regulatory acceptance. In hematologic malignancies, minimal residual disease (MRD) negativity has emerged as a leading candidate, providing a sensitive and quantifiable measure of treatment depth, particularly in multiple myeloma (MM). In addition, complete remission (CR) in acute myeloid leukemia (AML) and complete or partial remission (CR/PR) in myelodysplastic syndromes (MDS) have long been recognized as clinically meaningful endpoints and are increasingly being evaluated for their potential to predict long-term clinical outcomes. Statistical validation of surrogate endpoints requires rigorous evaluation of both individual-level and trial-level associations. Established methodologies include Prentice-type criteria, weighted least squares regression, copula-based two-stage models, and estimation of the surrogate threshold effect (STE). While randomized controlled trials (RCTs) remain the preferred setting for surrogate endpoint validation, study-level evidence from single-arm trials may also provide meaningful support, particularly in rare diseases or molecularly defined populations where randomized evidence is difficult to obtain. The successful development of MRD as an intermediate endpoint in multiple myeloma provides an important example of how complementary sources of evidence can inform regulatory decision-making. Beyond endpoint validation, biomarkers are increasingly being incorporated into innovative clinical trial designs, including biomarker-guided treatment strategies such as MRD-guided switching therapy. These designs introduce unique statistical challenges related to endpoint selection, treatment-effect estimation, causal interpretation, and operational considerations, underscoring the need for robust statistical methodology throughout both endpoint development and trial design. This session will highlight the growing role of biomarker-driven surrogate and intermediate endpoints in oncology drug development from clinical, statistical, regulatory, and industry perspectives. Speakers from the FDA and industry will discuss current regulatory expectations for earlier clinical benefit endpoints, contemporary statistical methods for surrogate endpoint evaluation, practical approaches to biomarker validation for intermediate endpoints, and statistical considerations for biomarker-guided clinical trial designs. Through examples spanning both solid tumors and hematologic malignancies, the session will examine emerging evidence, methodological challenges, and future directions for integrating biomarkers into oncology drug development and regulatory decision-making.