Location-Guided Multi-Modality Collaborative Learning for Prostate Tumor Segmentation

Prostate tumor segmentation from multi-modality magnetic resonance (MR) images is indispensable for diagnosis and treatment of prostate cancer. Existing methods typically overlook the characteristic distribution of prostate tumors in MR images and also the efficacy of extracted features from different MR modalities in the tumor segmentation task. In this paper, we address these limitations with a novel localization-to-segmentation framework. First, the localization stage locates tumor slices precisely via the label order consistency (LOC) strategy, by explicitly utilizing tumor-distribution prior. Then, in the subsequent segmentation stage, we develop an attention-based multi-modality collaborative learning (MCL) module, which extracts high-level modality-specific features while focusing on aggregating complementary features across modalities, to segment tumors from localized tumor slices. Experimental results demonstrate that our method achieves state-of-the-art segmentation performance on an in-house prostate MRI dataset, especially for tumors with low contrast to surrounding tissues.