According to results from a multinational study, use of an artificial intelligence (AI) tool during the assessment of HER2 status improved the accuracy of identifying HER2-low or HER2-ultralow breast cancer, reducing the risk of misclassifying these tumors as HER2-null.

These data will be presented at 2025 American Society for Clinical Oncology (ASCO) Annual Meeting in Chicago, Illinois.

According to Marina De Brot, MD, PhD, AC Camargo Cancer Center, Sao Paolo, Brazil, “Roughly 65% of breast tumors once called HER2-negative actually demonstrate some level of HER2 expression and belong to the subgroups now classified as HER2-low or HER2-ultralow breast cancers.” Approximately 1 in 3 HER2-ultralow breast cancers can be mistakenly classified as HER2-null, usually leading oncologists to not recommend HER2-targeted therapy with antibody drug conjugates.

This study included 105 pathologists from 10 different countries in Asia and South America. These pathologists performed a HER2 assessment of 20 digital breast cancer cases, both with and without AI assistance. Over 5 sessions, pathologists completed a total of 1940 readings during 3 separate exams. An AI-supported digital training platform, ComPath Academy, was provided only during the third exam. All readings were compared against ground-truth immunohistochemistry (IHC) scores from a central reference center.

Scoring sensitivity went from 76% to 90% with the use of AI-assistance. Agreement between pathologist and central reference scores improved from 76.3% without AI assistance to 89.6% with AI assistance. Accuracy in correctly identifying cases as HER2-positive, HER2-low, HER2-ultralow, and HER2-null improved from 66.7% without AI assistance to 88.5% with AI assistance. Without the use of AI-assistance, 29.5% of HER2-ultralow cases were misclassified as HER2-null, compared to 4% of cases with the use of AI-assistance.

ASCO Expert in AI, Julian Hong, MD, MS, University of California, San Francisco, commented, “These findings shed light on the promising role for AI in oncology, not as a replacement for the physician, but as a powerful tool to help us work smarter and faster to deliver high-quality, more personalized care.”

Dr De Brot concluded, “Our study provides the first multinational evidence that artificial intelligence can help close a critical diagnostic gap and open the door to new therapies like antibody-drug conjugates for a majority of patients who, until recently, had not been offered these options.”

Source:

ASCO Media Team. Artificial intelligence assistance can help improve accuracy in identifying HER2-low and HER2-ultralow breast cancers, avoid misclassification. Press Release.

Mulder D, Shaaban A, Lacroix-Triki M, et al. Use of artificial intelligence-assistance software for HER2-low and HER2-ultra low IHC interpretation training to improve diagnostic accuracy of pathologists and expand patients’ eligibility for HER2-targeted treatment. Presented at 2025 ASCO Annual Meeting. May 31-June 3, 2025; Chicago, IL. Abstract #1014