A deep learning system for the diagnosis and classification of brain tumors

Takeaway

  • This deep learning system using patients’ MRI data may assist neuroradiologists in more accurate classification and diagnosis of brain tumors.

Why this matters

  • The detection and classification of tumors mainly rely on MRI and biopsy. Identification of tumor types from MRI data has been challenging because of the similar phenotypes of multiple tumor classes on MRI scans. Given the invasiveness of biopsies, non-invasive diagnostic methods are preferred.

  • An accurate, reliable preoperative determination of tumor types using a deep learning model may facilitate more rapid clinical decision-making and better treatment planning.