MIT’s Regina Barzilay has used AI to improve breast cancer detection and diagnosis. Machine learning tools predict if a high-risk lesion identified on needle biopsy after a mammogram will upgrade to cancer at surgery, potentially eliminating unnecessary procedures.
In current practice, when a mammogram detects a suspicious lesion, a needle biopsy is performed to determine if it is cancer. Approximately 70 percent of the lesions are benign, 20 percent are malignant, and 10 percent are high-risk.
Using a method known as a “random-forest classifier,” the AI model resulted in 30 per cent fewer surgeries, compared to the strategy of always doing surgery, while diagnosing more cancerous lesions (97 per cent vs 79 per cent) than the strategy of only doing surgery on traditional “high-risk lesions.”
Trained on information about 600 high-risk lesions, the technology looks for data patterns that include demographics, family history, past biopsies, and pathology reports.
MGH radiologists will begin incorporating the method into their clinical practice over the next year.
Join ApplySci at Wearable Tech + Digital Health + Neurotech Silicon Valley on February 26-27, 2018 at Stanford University, featuring: Vinod Khosla – Justin Sanchez – Brian Otis – Bryan Johnson – Zhenan Bao – Nathan Intrator – Carla Pugh – Jamshid Ghajar – Mark Kendall – Robert Greenberg – Darin Okuda – Jason Heikenfeld