In a significant leap forward for personalized breast cancer treatment, a new study highlights the transformative potential of artificial intelligence (AI) in identifying a much larger cohort of women who could benefit from HER2-targeted therapies. According to the American Cancer Society, only about 20% of breast cancers were classified as HER2-positive, meaning they expressed high enough levels of the HER2 protein to qualify for these highly effective treatments. However, groundbreaking research slated for presentation at the American Society of Clinical Oncology (ASCO) annual meeting reveals that AI could extend these life-saving therapies to an additional 65% of women by accurately detecting low or even ultralow levels of HER2 expression.
The HER2 protein plays a crucial role in stimulating the growth of tumor cells. Therapies that specifically target HER2, such as monoclonal antibodies and newer antibody-drug conjugates, work by either blocking the protein’s activity or by using it as a direct pathway to deliver chemotherapy to cancer cells. Dr. Marina De Brot, lead researcher and head pathologist at the A.C. Camargo Cancer Center in São Paolo, Brazil, emphasizes the critical need for accurate HER2 detection: “Some of these tumors could be treated with HER2-targeted drugs, but only if we detect their HER2 expression levels.”
Background notes from the researchers indicate that at least 55% of breast cancers contain low levels of HER2, and another 10% possess ultralow levels. These previously undetectable or misclassified cases represent a vast untapped potential for targeted treatment. The new AI technology, named ComPath, promises to bridge this diagnostic gap. Dr. De Brot states, “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.”
The study involved 105 pathologists from 10 countries across Asia and South America, who utilized ComPath to assist in their HER2 scoring of breast cancer biopsies. Across five sessions, pathologists performed a total of 1,940 readings, with one in three readings conducted with AI assistance. The results were compelling: AI significantly improved pathologists’ accuracy in identifying HER2 status by nearly 22%, raising it from approximately 67% to just under 89%. Crucially, AI assistance reduced the misclassification rate of ultralow HER2 cases as HER2-negative by over 25%. Without AI, nearly 30% of these cases were miscategorized, but with AI, this figure dropped to a mere 4%, preventing patients from missing out on potentially life-extending therapies.
Dr. Julian Hong, medical director of radiation oncology informatics at the University of California-San Francisco, who reviewed the findings, underscored the importance of accurate HER2 scoring for optimal patient care. He noted that the international study demonstrates how an AI-assisted approach improves HER2 scoring, particularly in scenarios that directly impact treatment decisions. Dr. Hong’s perspective highlights the collaborative role of AI in healthcare: “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.” The researchers plan to integrate AI into routine cancer care further to assess its impact on patient treatment and outcomes.
What this means for Black women
The advent of AI-assisted HER2 detection holds profound implications for all women, but particularly for Black women who face unique disparities in breast cancer outcomes. Historically, Black women have a higher incidence of more aggressive breast cancer subtypes, including triple-negative breast cancer (TNBC), which typically does not respond to HER2-targeted therapies. However, recent research suggests that a subset of what was previously considered TNBC may actually express low levels of HER2, making them potential candidates for these targeted treatments.
Systemic inequities in healthcare access, quality of care, and implicit bias can lead to delays in diagnosis, less comprehensive testing, and limited access to specialized treatments for Black American women. The current reliance on manual HER2 scoring, which as the study indicates, can be prone to misclassification, exacerbates these disparities. If ultralow HER2 levels are consistently missed, Black women may disproportionately lose out on effective targeted therapies due to diagnostic oversight.
The AI-driven improvement in HER2 scoring accuracy—reducing misclassifications of ultralow HER2 levels from nearly 30% to 4%—could directly address this disparity. By providing a more objective and consistent diagnostic tool, AI can help ensure that Black American women, regardless of their socioeconomic status or geographic location, receive accurate HER2 status assessments. This improved accuracy means more Black women could be identified as eligible for HER2-targeted therapies, potentially including the newer antibody-drug conjugates that deliver chemotherapy directly to cancer cells. These treatments could offer a lifeline for those with historically difficult-to-treat breast cancers.
Furthermore, the “powerful tool” aspect of AI, as described by Dr. Hong, suggests that it can enhance the efficiency and precision of pathologists, potentially reducing the burden on healthcare systems and allowing for faster, more accurate diagnoses. This could translate to quicker treatment initiation, which is crucial for aggressive cancers often seen in Black women. By mitigating the risk of human error and providing a more standardized approach to HER2 testing, AI has the potential to narrow the gap in breast cancer outcomes for Black Americans, moving towards a future where personalized, effective treatment is accessible to all. The findings from this study, though preliminary until peer-reviewed publication, offer a significant beacon of hope for advancing health equity in oncology.