Leveraging AI-Enabled Tumor Assessment Tools on Radiological Images to Evaluate Treatment Effect and Support Clinical Trial Endpoints in Solid Tumors

Summarise with VIRAT
Authors: Hillary Andrews, Fabian Rengier, Andrea Pomrehn Myers, External Many further external authors
Cited by: 0

Summary

No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.

Summarise with VIRAT Free to try with your Science Catalogs account.

Abstract

Accurate and consistent tumor measurement is fundamental to evaluating treatment response in oncology clinical trials. For more than 25 years, RECIST has provided a practical and widely adopted framework for these assessments, however there are limitations. Rapid advances in artificial intelligence (AI) offer an opportunity to modernize tumor response assessment. AI-enabled tumor assessment tools can detect, segment, and quantify tumors across the entire body with greater consistency and improved biological detail. To realize this potential in drug development, the field must converge on clear standards for how AI-enabled tumor assessment tools are used, validated, and interpreted. Key questions include what is being measured, how results should be compared across tools, and how novel endpoints should be defined and qualified for regulatory use, particularly within pathways like FDA’s Accelerated Approval, where reliable early indicators of benefit are essential. To drive alignment, Friends of Cancer Research convened a multi-stakeholder working group to evaluate the current landscape of AI-enabled tumor assessment tools and propose a stepwise roadmap for integration into clinical trials. This white paper outlines: • Limitations of current imaging-based endpoints and opportunities to improve upon RECIST • Emerging AI-driven approaches including enhanced RECIST, radiomics, volumetric analysis, and growth kinetics • Necessary components of a framework to validate a novel AI-imaging based endpoint as an early endpoint, including outstanding questions for defining the endpoint and considerations for a meta-analysis By advancing and aligning on methodological standards, the oncology community can enable AI-based tumor assessments to supplement, and in some cases surpass, traditional measurement approaches in clinical trials. This shift has the potential to improve endpoint precision, reduce trial timelines and costs, and ultimately accelerate patient access to safe and effective therapies.

Fields of Study & Subjects

Radiomics and Machine Learning in Medical Imaging Artificial Intelligence in Healthcare and Education Medical Imaging Techniques and Applications

Publication details

Journal / source
The Novartis Repository (Novartis)
Publication date
1 January 2026
Language
EN