By Peter J. O’Dwyer, MD (left)
and Mitchell D. Schnall, MD, PhD

As we approach the end of a turbulent year for the cancer research community, ECOG-ACRIN is pleased to close out 2025 on a high note. At the San Antonio Breast Cancer Symposium (SABCS) earlier this month, researchers presented the initial findings from a major, multi-year collaboration between our Group and Caris Life Sciences^® (Caris). The endeavor aims to transform recurrence risk assessment in early-stage breast cancer through artificial intelligence (AI). The public-private partnership pairs ECOG-ACRIN’s extensive clinical trial expertise and biorepository resources with Caris’ comprehensive MI Cancer Seek^® whole exome and whole transcriptome profiling, whole slide imaging, and advanced machine learning platforms.

Through this collaboration, the research teams developed multimodal AI models to more precisely stratify recurrence risk in early-stage breast cancer by integrating histopathologic imaging, clinical, and molecular data generated from TAILORx—one of the world’s largest and most rigorously annotated breast cancer research repositories. This level of multimodal integration is unprecedented at this scale in early breast cancer prognostication. Across multiple analytic evaluations, the models demonstrated enhanced prognostic performance compared with existing recurrence-risk assessment methods, highlighting their potential to support more personalized treatment decision-making.

During General Session 1, Dr. Joseph Sparano presented Multimodal Artificial Intelligence (AI) Models Integrating Image, Clinical, and Molecular Data for Predicting Early and Late Breast Cancer Recurrence in TAILORx. In this project, researchers developed and prospectively validated a multimodal model integrating pathomic imaging, clinical, and expanded molecular data from 4,462 TAILORx tumor specimens. The study found that molecular data best predicted early, but not late, recurrence, imaging data best predicted late recurrence, and together they provided the strongest prediction of distant recurrence out to 15 years. These findings support the development of a new diagnostic test to better personalize recurrence risk assessment for women with HR-positive, HER2-negative, node-negative breast cancer.

During Rapid Fire Session 3, Dr. Terry Mamounas presented A Multimodal-Multitask Deep Learning Model Trained in NSABP B-42 and Validated in TAILORx for Late Distant Recurrence Risk in HR+ Early Breast Cancer. This presentation featured a model designed to predict late distant recurrence and help guide decisions about extended endocrine therapy beyond five years. Originally developed and validated in the NSABP B-42 trial, the model was externally validated in 4,300 TAILORx patients, where it demonstrated strong prognostic performance independent of traditional clinical factors. Using routine pathology slides and clinical data, the approach offers a scalable, cost-effective alternative to genomic assays and may help better identify patients who can safely avoid or benefit from extended therapy.

This public-private partnership between ECOG-ACRIN, the National Cancer Institute, and Caris represents a methodological, logistical, and collaborative integration of datasets from the historically impactful TAILORx trial to further extend the benefits for breast cancer patients. The advance in personalized medicine afforded in this work, in turn, helps to advance the potential of AI to refine treatment and improve outcomes.

From all of us at ECOG-ACRIN as the year winds down, we celebrate your dedication and resilience in changing times and wish you peace and relaxation over the holidays. We are confident that 2026 will be as interesting, challenging, and productive for all of us!

Read the December 2025 issue here.