Approximately 13.1% of women in the United States will be diagnosed with breast cancer during their lifetime, making it the most common cancer among American women.1 Although breast cancer mortality has declined substantially over recent decades, diagnostic and treatment challenges persist, in part because most patients present without known risk factors.1, 2 These realities highlight the need for accurate, efficient clinical tools that support timely identification, risk stratification, and management of breast cancer patients.
With increasing clinical complexity and a growing pressure to individualize care, the reliability of Artificial Intelligence (AI) tools to predict patient outcomes has become a critical question among clinicians. At the 2025 San Antonio Breast Cancer Symposium (SABCS), Artera, the developer of multimodal AI (MMAI)-based prognostic and predictive cancer tests, presented three abstracts addressing this very question. Specifically, whether MMAI can successfully integrate clinical and histopathological data to 1) accurately predict the risk of distant metastasis, 2) serve as an alternative to genomic risk assays, and 3) enable personalized care by identifying patients projected to benefit from chemotherapy treatment.
The following summaries provide evidence supporting MMAI as a tool to aid clinicians in making precise, data-driven treatment decisions while advancing the delivery of truly personalized oncology care.
Breast cancer is the most common type of cancer among women, affecting approximately 13.1% (1 in 8) of women in the US.1
The breast cancer death rate has dropped by 44% due to advances in earlier detection and treatment.1
Most women who develope breast cancer have no known risk factors.1, 2
PD11-01: Development of a Multi-Modal Artificial Intelligence (MMAI) Model for Predicting Distant Metastasis in HR+ Early-Stage Invasive Breast Cancer.3 (Dec. 12, 2025)
This study (Abstract #1251), from the NSABP Foundation and the University of Pittsburgh, developed and validated a pathology-based MMAI prognostic biomarker that integrates clinical variables with AI-driven histopathologic features to predict the risk of distant metastasis in HR+ early-stage breast cancer. Across two independent Phase 3 validation cohorts, the locked MMAI model demonstrated superior prognostic performance compared to a clinical comparator model and maintained independent association with distant metastasis (after adjustment for age, tumor size, and nodal status). The model showed consistent performance across nodal and menopausal subgroups and stratified patients into low-, intermediate-, and high-risk categorizations with clearly separated 10-year distant metastasis-free rates. These findings provide evidence in favor of MMAI as a practical, non-tissue-destructive prognostic tool. Additionally, MMAI has the potential to enhance risk stratification and personalized clinical decision-making in HR+ early breast cancer.
(PS3-04-08): Independent Validation of a Pathology-Based Multimodal Artificial Intelligence Biomarker for Predicting Risk of Distant Metastasis in Postmenopausal, Estrogen Receptor-Positive, Early-Stage Breast Cancer Patients: Analysis of the ABCSG Trial 8.4 (Dec. 11, 2025)
This external validation study (Abstract #1410), from the Medical University of Vienna and ABCSG, evaluated a previously developed MMAI prognostic model in a large prospective Phase 3 trial of postmenopausal patients with ER+/HER2-negative early breast cancer who received only endocrine therapy. In this lower-risk cohort, the locked MMAI model independently predicted distant metastasis, stratifying patients into low-, intermediate-, and high-risk groups with distinct 10-year distant metastasis-free rates despite adjustment for standard clinical factors. The MMAI’s prognostic performance was consistent across clinical and pathological subgroups; the image-only component remained independently associated with the risk of distant metastasis. These findings provide evidence in favor of MMAI as a validated, cost-effective, non-genomic alternative for risk stratification and treatment decision-making in ER+/HER2-negative early breast cancer.
(RF3-03): Evaluation of a digital pathology-based multimodal artificial intelligence model for prognosis and prediction of chemotherapy benefit in node-negative, hormone receptor-positive breast cancer patients: analysis of the NSABP B-20 trial.5 (Dec. 10, 2025)
This study (Abstract #3685), from the NSABP Foundation and the UPMC Hillman Cancer Center, evaluated the MMAI prognostic model in the NSABP B-20 Phase 3 trial of node-negative, HR+, early-stage breast cancer to assess risk stratification for distant metastasis and to predict any potential chemotherapy benefit. Using locked scores, MMAI demonstrated strong independent prognostic value in patients treated with tamoxifen alone, significantly stratifying distant metastasis risk across predefined low-, intermediate-, and high-risk groups. While MMAI was not predictive of chemotherapy benefit in the overall cohort, an age-dependent exploratory analysis revealed that patients over the age of 50 (with intermediate/high MMAI risk) displayed a meaningful reduction in distant metastasis with the addition of chemotherapy. No meaningful reduction was observed among low-risk patients. These findings suggest that MMAI may serve as a practical, lower-cost alternative to genomic assays for guiding adjuvant chemotherapy decisions in older patients with HR+, node-negative early breast cancer.
Collectively, these SABCS 2025 presentations suggest that MMAI can deliver robust, independently validated prognostic insight across multiple early breast cancer populations and integrate seamlessly with routine clinical and pathologic data. The consistent performance of MMAI across trials, risk groups, and treatment contexts supports its potential to complement, or act as an alternative to, genomic assay-based risk stratification and treatment selection. As the field of oncology moves toward increasingly individualized care, MMAI represents a clinically actionable tool.
Sources:
- American Cancer Society. Breast Cancer Facts & Figures 2024-2025. Atlanta: American Cancer Society; 2024. Accessed 22 Jan. 2026.
- Shockney, Lillie D. “Risk Factors.” National Breast Cancer Foundation, 2024, www.nationalbreastcancer.org/breast-cancer-risk-factors/. Accessed 22 Jan. 2026.
- “Poster Spotlight 11: Applying AI to Pathology and Risk Stratification – San Antonio Breast Cancer Symposium.” San Antonio Breast Cancer Symposium – the San Antonio Breast Cancer Symposium® Is Designed to Provide State-of-The-Art Information on the Experimental Biology, Etiology, Prevention, Diagnosis, and Therapy of Breast Cancer and Premalignant Breast Disease to an International Audience of Academic and Private Physicians and Researchers., 8 Oct. 2025, sabcs.org/events/2025/poster-spotlight-11-applying-ai-to-pathology-and-risk-stratification/. Accessed 22 Jan. 2026.
- “Poster Session 3 – San Antonio Breast Cancer Symposium.” San Antonio Breast Cancer Symposium – the San Antonio Breast Cancer Symposium® Is Designed to Provide State-of-The-Art Information on the Experimental Biology, Etiology, Prevention, Diagnosis, and Therapy of Breast Cancer and Premalignant Breast Disease to an International Audience of Academic and Private Physicians and Researchers., 8 Oct. 2025, sabcs.org/events/2025/poster-session-3/. Accessed 22 Jan. 2026.
- “Rapid Fire 3 – San Antonio Breast Cancer Symposium.” San Antonio Breast Cancer Symposium – the San Antonio Breast Cancer Symposium® Is Designed to Provide State-of-The-Art Information on the Experimental Biology, Etiology, Prevention, Diagnosis, and Therapy of Breast Cancer and Premalignant Breast Disease to an International Audience of Academic and Private Physicians and Researchers., 8 Oct. 2025, sabcs.org/events/2025/rapid-fire-3/. Accessed 22 Jan. 2026.
