News & Events
News
25 March 2026
Innovative AI Model Predicts Breast Cancer Recurrence from Standard Biopsy Slides
This is the first AI model shown to predict treatment benefit in breast cancer directly from pathology samples.
By decoding hidden patterns in routine biopsies, this AI breakthrough advances personalized precision medicine for breast cancer patients.
The challenge of predicting whether cancer will return after surgery is a critical hurdle in oncology. In a groundbreaking study published in The Lancet Oncology, a multidisciplinary team from the Technion – Israel Institute of Technology has developed an artificial intelligence (AI) model capable of predicting recurrence in hormone receptor-positive, HER2-negative, early-stage breast cancer with high accuracy. The research was led by Dr. Gil Shamai and Prof. Ron Kimmel from the Taub Faculty of Computer Science, in collaboration with Prof. Dvir Aran from the Faculty of Biology, highlighting the power of integrating computational science with biological insights to advance precision medicine.
Beyond the Human Eye: Digital Pathology Meets AI
The research team focused on hormone receptor-positive, HER2-negative, early-stage breast cancer, the most common subtype of the disease. Currently, pathologists examine tissue slides under a microscope to determine the cancer's stage and type, but much of the biological information remains "hidden" in the complex patterns of the tissue. The new AI tool, trained on thousands of digitized slides from hundreds of patients, can identify subtle morphological features that are nearly invisible to the human eye. By analyzing these patterns, the algorithm can categorize patients into high-risk or low-risk groups for recurrence shortly after their initial diagnosis.
Personalizing Treatment Strategies
One of the most significant implications of this study is its potential to guide clinical decisions. Patients identified by the AI as "high-risk" may benefit from more aggressive follow-up or early intervention with systemic therapies, such as chemotherapy or targeted endocrine treatments, even if their cancer was caught at an early stage. Conversely, "low-risk" patients might be spared the toxicity of unnecessary treatments. The model proved its robustness by successfully predicting outcomes across different medical centers and diverse patient populations, demonstrating its readiness for potential clinical integration.
A Cost-Effective and Accessible Solution
Unlike many advanced diagnostic tools that require expensive genetic sequencing or specialized equipment, this AI model utilizes standard Hematoxylin and Eosin (H&E) stained slides—the same slides already produced in every pathology lab worldwide. This makes the technology highly accessible and cost-effective, particularly for healthcare systems in developing regions. By turning a routine diagnostic step into a powerful predictive tool, the researchers have created a scalable method to improve patient survival rates and optimize the allocation of medical resources in the fight against breast cancer.
The Research has been published in the The Lancet Oncology
