Robustness of breast lesion segmentation under MRI undersampling improves with k-space-aware deep learning

2026-05-21

AI Summary

This work addresses the degraded performance of MRI-based breast lesion segmentation under k-space undersampling or noise by proposing a 3D U-Net architecture that fuses k-space and image-domain information to learn segmentation directly from raw k-space data in an end-to-end manner. By jointly modeling frequency-domain filtering and image-domain localization mechanisms, the method maintains segmentation accuracy comparable to fully sampled conditions while significantly enhancing robustness to accelerated acquisition and k-space noise. Evaluated across four 3D U-Net variants on both synthetic and real k-space data through cross-validation, the proposed hybrid model consistently achieves higher Dice scores than image-domain-only baselines under moderate to severe undersampling and noise, demonstrating the effectiveness and complementary benefits of explicit k-space awareness in segmentation tasks.

Abstract

Purpose: To assess whether breast lesion segmentation can be learned directly from acquired MRI k-space, and whether doing so improves robustness when data are accelerated or noisy. Materials and Methods: This retrospective study used public breast dynamic contrast-enhanced MRI (DCE-MRI) datasets with acquired and synthetic k-space, together with a within-dataset synthetic control. We compared four 3D U-Net variants: a hybrid k-space-to-image model, a native k-space model, and magnitude and complex image-space baselines. Models were evaluated under increasing undersampling and added complex Gaussian k-space noise. The primary outcome was patient-level Dice similarity coefficient under cross-validation, with the hybrid model prespecified as the main comparison against the magnitude image-space baseline. Results: At full sampling, the hybrid and image-space models performed similarly. As acceleration increased, the hybrid model retained substantially more segmentation accuracy and significantly outperformed the magnitude image-space baseline across moderate to high undersampling levels. The same pattern was observed when noise was added directly to k-space: the hybrid model degraded more slowly, whereas the image-space baseline failed under heavier noise. This advantage was reproduced in the within-dataset synthetic control. Feature analysis suggested that the k-space stage and image-space stage played complementary roles, with frequency-domain filtering concentrated before image-domain lesion localization. Conclusion: K-space-aware deep learning improves the robustness of breast lesion segmentation under MRI undersampling and k-space noise, while matching image-space methods at full sampling.

Problem

• breast lesion segmentation
• MRI undersampling
• k-space noise
• robustness
• deep learning

Innovation

• k-space-aware learning
• MRI undersampling
• breast lesion segmentation
• hybrid deep learning
• robustness