Clinically Aligned Geometry Constraints for Robust IVUS Vessel Boundary Segmentation

2026-06-17

AI Summary

This study addresses the limitations of traditional intravascular ultrasound (IVUS) vessel segmentation methods, which often prioritize pixel-level overlap metrics at the expense of boundary drift and topological errors, thereby compromising the accuracy of clinical geometric measurements. To overcome this, the authors propose GeoCat, a novel network that, for the first time, integrates differentiable clinical geometric constraints—such as diameter, angle, and area—into end-to-end training. GeoCat employs dual Cartesian–polar encoders, a cross-domain attention mechanism, and multi-frame temporal fusion to directly optimize consistency between anatomical structures and clinical parameters. Evaluated on 12,242 IVUS frames, the method achieves a Dice score of 0.93, reduces the 95% Hausdorff distance to 0.14 mm, yields a topological error rate of only 1.0%, and attains diameter errors of 0.13–0.16 mm and angular errors of approximately 8°, substantially enhancing the reliability of plaque burden quantification.

Abstract

Intravascular ultrasound (IVUS) lumen and external elastic membrane (EEM) segmentation is important for quantitative coronary plaque burden assessment. Errors in lumen or EEM delineation directly propagate to plaque area, plaque burden and geometric measurements. However, standard methods prioritising overlap scores often suffer from boundary drift and topology errors, leading to inaccurate clinical measurements. We present GeoCat, a geometry-consistent network that processes 5-frame IVUS clips using dual Cartesian-polar encoders with cross-domain attention and temporal fusion. A differentiable geometry consistency loss directly supervises clinically relevant descriptors including diameters, orientations, and cross-sectional areas. The model is trained on 12,242 annotated frames from 146 patients acquired with two commercial IVUS systems. We evaluate performance using both segmentation accuracy and plaque-relevant clinical metrics, including Dice/IoU, boundary measures(95HD (mm), ASSD), topology violation rate, and clinical geometry errors (dmax/dmin, angles, and areas). On our dataset, GeoCat achieves a Dice of 0.93, reduces 95HD to 0.14 mm, and lowers topology violations to 1.0%. Importantly, it significantly improves geometric fidelity, yielding diameter errors of 0.13-0.16 mm and angular errors of ~8 degrees, supporting reliable plaque burden quantification.

Problem

• IVUS segmentation
• boundary drift
• topology errors
• plaque burden
• clinical geometry

Innovation

• geometry-consistent segmentation
• cross-domain attention
• differentiable geometric loss
• IVUS boundary delineation
• clinical metric supervision