Carotid atherosclerotic lesion analysis in 3D based on distance encoding in mesh representation

The purpose of this study is to support stroke risk analysis, evaluation of therapy effectiveness, and lesion progression through a comprehensive assessment of carotid atherosclerotic lesions in 3D based on automatic segmentation and visualization of quantitative parameters.

We propose a novel method for extracting the pathologically thickened regions from 3D vessel wall segmentations using distance encoding on the inner and outer wall mesh to enable atherosclerotic lesion analysis. A case-specific and constant threshold was evaluated and applied to extract lesions from a dataset of 202 T1-weighted black-blood MRI scans of subjects with up to 50% stenosis. Applied to baseline and one-year follow-up data, the method supports detailed morphology analysis over time, including volume quantification, aided by improved visualization via mesh unfolding. The extracted region was also used to analyze the signal intensity distribution within the lesion region.

We successfully extracted lesion regions from 297 carotid arteries, capturing a wide range of shapes with volumes ranging from 3.61 to $996.9{\text{mm}}^3$ . The use of a constant threshold of 1.6 mm showed an intraclass correlation of 0.861 for the lesion volume and a median average surface distance of 0.594 mm in the validation set.

The proposed method enables the extraction of lesion meshes from 3D vessel wall segmentation masks, enabling a correspondence between baseline and one-year follow-up examinations. Unfolding the lesion meshes enhances visualization, while the mesh-based analysis allows quantification of morphologic parameters and an analysis of the signal intensities in the lesion region.