SpineNet: automatically pinpointing classification evidence in spinal MRIs

Conference item

We describe a method to automatically predict radiological scores in spinal Magnetic Resonance Images (MRIs). Furthermore, we also identify and localize the pathologies that are the reasons for these scores. We term these pathological regions the ``evidence hotspots'. Our contributions are two fold: (i) a Convolutional Neural Network (CNN) architecture and training scheme to predict multiple radiological scores on multiple slice sagittal MRIs. The scheme uses multi-task CNN training with augmentation, and handles the class imbalance common in medical classification tasks. (ii) the prediction of a heat-map of evidence hotspots for each score. For both of these, all that is required for training is the class label of the disc or vertebrae, no stronger supervision (such as slice labels) is needed. We report state-of-the-art and near-human performances across multiple radiological scorings including: Pfirrmann grading, disc narrowing, endplate defects, and marrow changes.

Authors: Jamaludin, A; Kadir, T; Zisserman, A

• Publication status: Accepted
• Peer review status: Peer reviewed
• Publisher: Springer International Publishing AG
• Host title: MICCAI 2016: Medical Image Computing and Computer-Assisted Intervention – MICCAI 2016
• Journal: 19th International Conference on Medical Image Computing and Computer Assisted Intervention
• Volume: 9901
• Pages: 166-175
• Series: Lecture Notes in Computer Science
• Publication date: 2016-10-02
• Acceptance date: 2016-03-17
• Event location: Athens Greece
• DOI: 10.1007/978-3-319-46723-8_20
• ISBN: 9783319467238
• Keywords: SBTMR