Machine Learning facilitates assessment of shaved lumpectomy margins

Drs. Timothy D’Alfonso, David Ho, and Lee K.Tan from Memorial Sloan Kettering in NY discuss their recent Modern Pathology study describing a new machine learning algorithm that can help pathologists assess shaved margins from lumpectomy specimens. A uniquely developed Deep Multi-Magnification Network (DMMN) was utilized in Whole Slide Images (WSI) in the hope to triage negative margins and allow pathologists to focus on shaves that are positive for DCIS and/or Invasive carcinoma.