Generative Adversarial Networks for Weakly Supervised Generation and Evaluation of Brain Tumor Segmentations on MR Images
This addresses the problem of reducing annotation burden for radiologists in medical imaging, though it is incremental as it builds on existing GAN and weakly supervised techniques.
The paper tackles brain tumor segmentation in MR images by proposing a weakly supervised GAN-based method that uses only binary image-level labels, achieving a test Dice coefficient of 83.91% and enabling pathology classification with a test AUC of 93.32%, comparable to using true segmentations.
Segmentation of regions of interest (ROIs) for identifying abnormalities is a leading problem in medical imaging. Using machine learning for this problem generally requires manually annotated ground-truth segmentations, demanding extensive time and resources from radiologists. This work presents a weakly supervised approach that utilizes binary image-level labels, which are much simpler to acquire, to effectively segment anomalies in 2D magnetic resonance images without ground truth annotations. We train a generative adversarial network (GAN) that converts cancerous images to healthy variants, which are used along with localization seeds as priors to generate improved weakly supervised segmentations. The non-cancerous variants can also be used to evaluate the segmentations in a weakly supervised fashion, which allows for the most effective segmentations to be identified and then applied to downstream clinical classification tasks. On the Multimodal Brain Tumor Segmentation (BraTS) 2020 dataset, our proposed method generates and identifies segmentations that achieve test Dice coefficients of 83.91%. Using these segmentations for pathology classification results with a test AUC of 93.32% which is comparable to the test AUC of 95.80% achieved when using true segmentations.