Machine learning and magnetic resonance imaging: Differentiating benign from malignant osseous tumors

Abstract

Radiomics and machine learning applied to the isolated cartilaginous bone lesion on magnetic resonance imaging (MRI) is increasingly useful to distinguish malignant versus benign bone lesions, to determine whether repetitive imaging over time, to determine dynamic expansion, or immediate excisional biopsy are indicated.

Keywords: bone; chondrosarcoma; enchondroma; radiomics; tumor.