Preserving Cardiac Integrity: A Topology-Infused Approach to Whole Heart Segmentation

TL;DR

A new topology-preserving module that is integrated into deep neural networks that achieves anatomically plausible segmentation by using learned topology-preserving fields, which are based entirely on 3D convolution and are therefore very effective for 3D voxel data.

Abstract

Whole heart segmentation (WHS) supports cardiovascular disease (CVD) diagnosis, disease monitoring, treatment planning, and prognosis. Deep learning has become the most widely used method for WHS applications in recent years. However, segmentation of whole-heart structures faces numerous challenges including heart shape variability during the cardiac cycle, clinical artifacts like motion and poor contrast-to-noise ratio, domain shifts in multi-center data, and the distinct modalities of CT and MRI. To address these limitations and improve segmentation quality, this paper introduces a new topology-preserving module that is integrated into deep neural networks. The implementation achieves anatomically plausible segmentation by using learned topology-preserving fields, which are based entirely on 3D convolution and are therefore very effective for 3D voxel data. We incorporate natural constraints between structures into the end-to-end training and enrich the feature representation of the neural network. The effectiveness of the proposed method is validated on an open-source medical heart dataset, specifically using the WHS++ data. The results demonstrate that the architecture performs exceptionally well, achieving a Dice coefficient of 0.939 during testing. This indicates full topology preservation for individual structures and significantly outperforms other baselines in preserving the overall scene topology.

Figures (5)

• Figure 1: Examples of cardiac images and WHS results: (a) displays the three orthogonal views of a cardiac CT image and its corresponding WHS result, (b) shows example cardiac MRI data and the WHS result. LV: left ventricle; RV: right ventricle; LA: left atrium; RA: right atrium; Myo: myocardium of LV; AO: ascending aorta; PA: pulmonary artery.
• Figure 2: Multi-class topological constraints for WHS
• Figure 3: An overview of the proposed method. TPM encodes priori knowledge between the different classes (e.g., Myo and LA classes in the WHS++ dataset follow the exclusion constraint). Key voxels N are identified and used for the new loss.
• Figure 4: Qualitative results compared with the baselines. The first row is CT data and the second row is MRI data. Colors for the classes correspond to the ones used in Fig. \ref{['fig3']}
• Figure 5: Violin plot of Dice scores of the whole heart segmentation on CT and MRI dataset by the four methods.