AI Risk Scores Link to Low-Grade Malignancy Before Clinical Detection

Breast cancer remains the most frequently diagnosed malignancy and a leading cause of cancer death among women globally [1]. While mortality rates have declined by 40 percent since 1989 due to improved treatment and early detection, the incidence of hormone receptor positive and local stage disease continues to rise [2]. To address these trends, clinicians are increasingly exploring artificial intelligence as a tool to enhance screening accuracy and predict future malignancy risk [3]. Current evidence suggests that AI models can identify high risk patients with a degree of accuracy that often exceeds traditional clinical risk factors [3]. However, the specific pathological features that these algorithms detect in the years preceding a diagnosis remain poorly understood. A new study now investigates how these automated risk scores correlate with specific tumor characteristics to better understand the clinical relevance of early imaging signals.