We present an Expectation-Maximization (EM) Regularized Deep Learning (EMReDL) model for weakly supervised tumor segmentation. The proposed framework is tailored to glioblastoma, a type of malignant tumor characterized by its diffuse infiltration into the surrounding brain tissue, which poses significant challenge to treatment target and tumor burden estimation using conventional structural MRI. Although physiological MRI provides more specific information regarding tumor infiltration, the relatively low resolution hinders a precise full annotation. This has motivated us to develop a weakly supervised deep learning solution that exploits the partial labelled tumor regions. EMReDL contains two components: a physiological prior prediction model and EM-regularized segmentation model. The physiological prior prediction model exploits the physiological MRI by training a classifier to generate a physiological prior map. This map is passed to the segmentation model for regularization using the EM algorithm. We evaluated the model on a glioblastoma dataset with the pre-operative multiparametric and recurrence MRI available. EMReDL showed to effectively segment the infiltrated tumor from the partially labelled region of potential infiltration. The segmented core tumor and infiltrated tumor demonstrated high consistency with the tumor burden labelled by experts. The performance comparisons showed that EMReDL achieved higher accuracy than published state-of-the-art models. On MR spectroscopy, the segmented region displayed more aggressive features than other partial labelled region. The proposed model can be generalized to other segmentation tasks that rely on partial labels, with the CNN architecture flexible in the framework.
updated: Thu Jul 15 2021 16:44:07 GMT+0000 (UTC)
published: Thu Jan 21 2021 18:14:43 GMT+0000 (UTC)