Artificial Intelligence Enhances Identification of Candidates for Targeted Breast Cancer Therapies

Advancements in artificial intelligence (AI) are paving the way for more precise detection of breast cancer patients who could benefit from targeted treatment options. A recent study, soon to be presented at the American Society of Clinical Oncology (ASCO) conference, highlights how AI tools can assist pathologists in identifying low and ultralow levels of HER2 protein expression in breast tumors. These levels are critical because they determine eligibility for HER2-targeted therapies, including monoclonal antibodies and antibody-drug conjugates.

Currently, only about 20% of breast cancers show high HER2 expression sufficient for targeted treatment. However, the innovative AI model, known as ComPath, enabled pathologists across ten countries to improve their accuracy in HER2 scoring during multiple sessions. Results demonstrated a nearly 22% increase in accuracy, raising correct assessments from approximately 67% to nearly 89%. Notably, AI assistance reduced misclassification of ultralow HER2 levels by over 25%, significantly increasing the likelihood that patients with low-level HER2 tumors are correctly identified for potentially beneficial therapies.

Dr. Marina De Brot, the lead researcher from the A.C. Comargo Cancer Center in São Paulo, explained that recognizing even low or ultralow HER2 expression can open the door for patients to receive innovative treatments, such as antibody-drug conjugates, which directly target tumor cells. These therapies have traditionally been limited to patients with high HER2 levels.

The study involved 105 pathologists evaluating nearly 2,000 biopsy readings, with AI supporting roughly one-third of these evaluations. Even experienced pathologists benefited from AI, highlighting its potential as a valuable tool rather than a replacement for medical expertise. Dr. Julian Hong from the University of California emphasized that AI can help clinicians work more efficiently and accurately, ultimately improving patient treatment outcomes.

Looking forward, the research team plans to integrate AI more broadly into routine cancer care, aiming to enhance diagnostic precision and expand treatment options for breast cancer patients worldwide. This research underscores the promising future of AI in oncology, facilitating more personalized, effective cancer therapies.