Towards contrast- and pathology-agnostic clinical fetal brain MRI segmentation using SynthSeg

2025-04-14

AI Summary

This study addresses poor generalizability of fetal brain MRI segmentation models caused by domain shifts—including physiological states, acquisition parameters, and pathological deformations. To this end, we propose a robust, contrast-agnostic, and pathology-agnostic segmentation framework built upon SynthSeg. Our method innovatively integrates adaptive training-time sampling, domain randomization for synthetic data generation, and multi-scale data augmentation to explicitly model anatomical variability. Notably, this is the first work to achieve pathology-independent automatic segmentation on fetal brain MRI exhibiting severe anatomical abnormalities. Quantitatively, our approach achieves statistically significant improvement in Dice score on abnormal test cases (p < 1e−4). The proposed framework establishes a new paradigm for cross-center, cross-pathology fetal brain structural analysis and delivers a practical, deployable tool for clinical and research applications.

Technology Category

• Computer Vision: Segmentation
• Machine Learning: Transfer, Domain Adaptation, Multi-Task Learning
• Intelligent Robots: Multimodal Perception & Sensor Fusion

Application Category

• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Search and Retrieval-Augmented AI: Agentic search
• Responsible Web: Machine-in-the-loop, human agency and autonomy

Abstract

Magnetic resonance imaging (MRI) has played a crucial role in fetal neurodevelopmental research. Structural annotations of MR images are an important step for quantitative analysis of the developing human brain, with Deep learning providing an automated alternative for this otherwise tedious manual process. However, segmentation performances of Convolutional Neural Networks often suffer from domain shift, where the network fails when applied to subjects that deviate from the distribution with which it is trained on. In this work, we aim to train networks capable of automatically segmenting fetal brain MRIs with a wide range of domain shifts pertaining to differences in subject physiology and acquisition environments, in particular shape-based differences commonly observed in pathological cases. We introduce a novel data-driven train-time sampling strategy that seeks to fully exploit the diversity of a given training dataset to enhance the domain generalizability of the trained networks. We adapted our sampler, together with other existing data augmentation techniques, to the SynthSeg framework, a generator that utilizes domain randomization to generate diverse training data, and ran thorough experimentations and ablation studies on a wide range of training/testing data to test the validity of the approaches. Our networks achieved notable improvements in the segmentation quality on testing subjects with intense anatomical abnormalities (p<1e-4), though at the cost of a slighter decrease in performance in cases with fewer abnormalities. Our work also lays the foundation for future works on creating and adapting data-driven sampling strategies for other training pipelines.

Problem

• Automated fetal brain MRI segmentation across diverse domain shifts
• Improving segmentation for pathological cases with anatomical abnormalities
• Enhancing domain generalizability using data-driven train-time sampling

Innovation

• Data-driven train-time sampling strategy
• SynthSeg framework adaptation
• Enhanced domain generalizability