MS-Net: Multi-Site Network for Improving Prostate Segmentation with Heterogeneous MRI Data

Abstract

Automated prostate segmentation in MRI is highly demanded for computer-assisted diagnosis. Recently, a variety of deep learning methods have achieved remarkable progress in this task, usually relying on large amounts of training data. Due to the nature of scarcity for medical images, it is important to effectively aggregate data from multiple sites for robust model training, to alleviate the insufficiency of single-site samples. However, the prostate MRIs from different sites present heterogeneity due to the differences in scanners and imaging protocols, raising challenges for effective ways of aggregating multi-site data for network training. In this paper, we propose a novel multi-site network (MS-Net) for improving prostate segmentation by learning robust representations, leveraging multiple sources of data. To compensate for the inter-site heterogeneity of different MRI datasets, we develop Domain-Specific Batch Normalization layers in the network backbone, enabling the network to estimate statistics and perform feature normalization for each site separately. Considering the difficulty of capturing the shared knowledge from multiple datasets, a novel learning paradigm, i.e., Multi-site-guided Knowledge Transfer, is proposed to enhance the kernels to extract more generic representations from multi-site data. Extensive experiments on three heterogeneous prostate MRI datasets demonstrate that our MS-Net improves the performance across all datasets consistently, and outperforms state-of-the-art methods for multi-site learning.

IEEE TMI, 2020

References in corpus (4)

• Confounding variables can degrade generalization performance of radiological deep learning models
• Learning multiple visual domains with residual adapters
• Boundary-weighted Domain Adaptive Neural Network for Prostate MR Image Segmentation
• Deep Attentive Features for Prostate Segmentation in 3D Transrectal Ultrasound

Cited by in corpus (20)

• Domain Generalization: A Survey
• MedViT: A Robust Vision Transformer for Generalized Medical Image Classification
• Semi-supervised Medical Image Classification with Relation-driven Self-ensembling Model
• Towards Data-Efficient Learning: A Benchmark for COVID-19 CT Lung and Infection Segmentation
• Contrastive Cross-site Learning with Redesigned Net for COVID-19 CT Classification
• CLIP in Medical Imaging: A Survey
• Cross-Site Severity Assessment of COVID-19 from CT Images via Domain Adaptation
• Federated Domain Generalization: A Survey
• Domain-incremental Cardiac Image Segmentation with Style-oriented Replay and Domain-sensitive Feature Whitening
• Generalizable multi-task, multi-domain deep segmentation of sparse pediatric imaging datasets via multi-scale contrastive regularization and multi-joint anatomical priors
• Conquering Data Variations in Resolution: A Slice-Aware Multi-Branch Decoder Network
• A diffusion MRI tractography atlas for concurrent white matter mapping across Eastern and Western populations
• Fighting the scanner effect in brain MRI segmentation with a progressive level-of-detail network trained on multi-site data
• A-Eval: A Benchmark for Cross-Dataset Evaluation of Abdominal Multi-Organ Segmentation
• The use of deep learning in interventional radiotherapy (brachytherapy): a review with a focus on open source and open data
• A New Perspective to Boost Performance Fairness for Medical Federated Learning
• Dynamic Data Augmentation via MCTS for Prostate MRI Segmentation
• On the Generalization Properties of Deep Learning for Aircraft Fuel Flow Estimation Models
• A Principled Approach to Failure Analysis and Model Repairment: Demonstration in Medical Imaging
• Multi-Task, Multi-Domain Deep Segmentation with Shared Representations and Contrastive Regularization for Sparse Pediatric Datasets