Researchers have developed an artificial intelligence model that can predict the Oncotype DX 21-gene recurrence score (RS) for breast cancer patients using digitized histopathology slides and basic clinical data. The study, led by Dr. David Shamai and colleagues, was published in The Lancet Oncology.
The AI model was trained and validated on data from over 3,600 patients across multiple international cohorts. It analyzes standard hematoxylin and eosin (H&E)-stained tissue slides, which are routinely available, to estimate the recurrence score, a genomic test that guides chemotherapy decisions for hormone receptor-positive, HER2-negative early-stage breast cancer.
In validation, the model showed strong correlation with the actual genomic test results. This approach could help identify patients who are likely to have a low recurrence score, potentially reducing the need for expensive genomic testing and making personalized treatment guidance more accessible where such testing is unavailable.
The authors emphasize that this tool is intended to assist, not replace, standard diagnostic procedures and clinical judgment. Further prospective studies are needed to confirm its clinical utility before widespread adoption.
