PS20: Harmonizing Modern Standards with AI and Digital Health Innovation: A New Era for Oncology PRO Evidence Generation in the Digital Era

The role of patient-reported outcomes (PROs) in oncology has become increasingly important for clinical and regulatory review. While overall survival (OS) and progression-free survival (PFS) remain as traditional efficacy endpoints in oncology trials, PRO data and analysis provide essential insight by characterizing the patient's experience with anti-cancer product(s), including physical symptoms (e.g., pain, fatigue) and functional impairments. PRO data are critical for a comprehensive understanding of the benefit-risk profile of a cancer drug and serve as an important component of regulatory submissions and health technology assessments. However, inconsistencies in PRO estimands and methodology in cancer trials have negatively impacted its efficient use to inform the benefit-risk assessment. Current methods of collecting and analyzing PRO data are often limited by unclear specification of research questions, subjective recall, analytical methods, and improper handling of intercurrent events and missing values. The use of digital tools to collect PROs in real time, proper planning of research questions, and the estimand framework combined with AI/ML techniques, open up new opportunities for incorporation and interpretation of PRO data in cancer drug development. The session will begin with a deep dive into the SISAQOL-IMI international recommendations (published 2025), focusing on the interactive webtool for standardizing PRO-based endpoints in oncology trials. In 2021, the Setting International Standards in Analyzing Patient-Reported Outcomes and Quality of Life Endpoints in Cancer Clinical Trials - IMI (SISAQOL-IMI) was formed through a collaboration among 41 international organizations, including regulatory agencies, health technology assessment bodies, industry, academia, and patient organizations to formulate a set of recommendations to improve the analysis and interpretation of PRO data in cancer clinical trials. The speaker will illustrate utilizing the interactive webtool for setting up PRO research questions, choice of appropriate estimand, handling missing values and missing data under SISAQOL-IMI recommendations. The second presentation will provide an overview of how AI/ML methodologies can enhance the quality of life (QOL) data analysis (e.g., predictive modeling, natural language processing, clustering) along with a case study from oncology. The talk will discuss the potential benefits including improved data handling, pattern recognition, personalized insights, and enhanced predictive power. The talk will conclude by highlighting some challenges such as data quality, transparency, interpretability, and ethical considerations. Following two presentations, a senior member from FDA will discuss the current landscape, including how the heath authority views the importance of the SISAQOL-IMI, PROs, and the future of evidence generation. The speaker will also reflect on the opportunities and challenges for regulatory acceptance of innovative PRO analysis methodologies in oncology. This session bridges the gap between established standards and the next frontier of evidence generation. It will provide a roadmap for statisticians to generate robust, interpretable, and high-impact evidence based on PRO data that meet regulatory expectations. With eminent speakers from academia, industry and FDA, the presentations, case study analyses, and open discussions will encourage participants for a lively discussion.