A novel fully automated mri-based deep learning method for classification of IDH mutation status in brain gliomas

Background: Isocitrate dehydrogenase (IDH) mutation status has emerged as an important prognostic marker in gliomas. Currently, reliable IDH mutation determination requires invasive surgical procedures. The purpose of this study was to develop a highly-accurate, MRI-based, voxel-wise deep-learning IDH-classification network using T2-weighted (T2w) MR images and compare its performance to a multi-contrast network. Methods: Multi-parametric brain MRI data and corresponding genomic information were obtained for 214 subjects (94 IDH-mutated, 120 IDH wild-type) from The Cancer Imaging Archive (TCIA) and The Cancer Genome Atlas (TCGA). Two separate networks were developed including a T2w image only network (T2-net) and a multi-contrast (T2w, FLAIR, and T1 post-contrast), network (TS-net) to perform IDH classification and simultaneous single label tumor segmentation. The networks were trained using 3D-Dense-UNets. A three-fold cross-validation was performed to generalize the networks' performance. ROC analysis was also performed. Dice-scores were computed to determine tumor segmentation accuracy. Results: T2-net demonstrated a mean cross-validation accuracy of 97.14% +/- 0.04 in predicting IDH mutation status, with a sensitivity of 0.97 +/- 0.03, specificity of 0.98 +/- 0.01, and an AUC of 0.98 +/- 0.01. TS-net achieved a mean cross-validation accuracy of 97.12% +/- 0.09, with a sensitivity of 0.98 +/- 0.02, specificity of 0.97 +/- 0.001, and an AUC of 0.99 +/- 0.01. The mean whole tumor segmentation Dice-scores were 0.85 +/- 0.009 for T2-net and 0.89 +/- 0.006 for TS-net. Conclusion: We demonstrate high IDH classification accuracy using only T2-weighted MRI. This represents an important milestone towards clinical translation.