Enhancing Corpus Callosum Segmentation in Fetal MRI via Pathology-Informed Domain Randomization
This addresses data scarcity for rare fetal brain pathologies like corpus callosum dysgenesis, enabling more accurate segmentation and biomarker extraction for clinical assessment.
The paper tackles the problem of corpus callosum segmentation in fetal MRI for rare pathology cases with limited annotated data by proposing a pathology-informed domain randomization method that generates synthetic data from healthy scans. The approach reduces corpus callosum length estimation error from 10.9 mm to 0.7 mm in pathology cases while maintaining performance on healthy and other pathology cases.
Accurate fetal brain segmentation is crucial for extracting biomarkers and assessing neurodevelopment, especially in conditions such as corpus callosum dysgenesis (CCD), which can induce drastic anatomical changes. However, the rarity of CCD severely limits annotated data, hindering the generalization of deep learning models. To address this, we propose a pathology-informed domain randomization strategy that embeds prior knowledge of CCD manifestations into a synthetic data generation pipeline. By simulating diverse brain alterations from healthy data alone, our approach enables robust segmentation without requiring pathological annotations. We validate our method on a cohort comprising 248 healthy fetuses, 26 with CCD, and 47 with other brain pathologies, achieving substantial improvements on CCD cases while maintaining performance on both healthy fetuses and those with other pathologies. From the predicted segmentations, we derive clinically relevant biomarkers, such as corpus callosum length (LCC) and volume, and show their utility in distinguishing CCD subtypes. Our pathology-informed augmentation reduces the LCC estimation error from 1.89 mm to 0.80 mm in healthy cases and from 10.9 mm to 0.7 mm in CCD cases. Beyond these quantitative gains, our approach yields segmentations with improved topological consistency relative to available ground truth, enabling more reliable shape-based analyses. Overall, this work demonstrates that incorporating domain-specific anatomical priors into synthetic data pipelines can effectively mitigate data scarcity and enhance analysis of rare but clinically significant malformations.