Deep learning based similarity-consistency abnormality detection (SCAD) model for classification of MRI patterns of multiple myeloma (MM) infiltration

We developed a one-class semi-supervised similarity-consistency abnormality detection model (SCAD) to classify vertebras with different MRI patterns (normal, focal, variegated and diffused pattern). We trained our SCAD model using vertebras with the normal pattern, and deployed the trained SCAD model to vertebras with the three non-normal patterns. The results showed that our SCAD model achieved a test AUC of 0.71, 0.79 and 0.88 in differentiating the focal, variegated and diffuse patterns from the normal pattern, respectively.