Beyond Volume Overlap: Surface Matching for Topology-Aware Coronary Artery Segmentation

Rafael Velasquez; Esther Puyol-Antón; Pablo Arbeláez

cs.CVSep 22, 2026

Beyond Volume Overlap: Surface Matching for Topology-Aware Coronary Artery Segmentation

Authors: Rafael Velasquez, Esther Puyol-Antón, Pablo Arbeláez

Organizations: Universidad de los Andes, Bogotá, Colombia · School of Biomedical Engineering & Imaging Sciences, King’s College London, London, UK

Abstract

Accurate coronary artery segmentation on coronary computed tomography angiography (CCTA) is essential for diagnosing coronary artery disease. Deep networks are conventionally trained and evaluated with the Dice coefficient, but volume-overlap metrics are poorly suited to thin, tubular anatomy: since most voxels belong to a few thickproximal segments, a missing distal branch barely affects Dice despite severely disrupting the connectivity required for clinical use. We introduce a surface metric that matches predicted and reference surface points via bipartite assignment under a localized, vessel-radius tolerance, reporting precision, recall, and F1 with decoupled false positives (spurious branches) and false negatives (missed branches) a distinction the symmetric Dice cannot make. With it we show that a strong Dice-trained baseline omits far more vessel surface than it hallucinates, an asymmetry its high Dice hides. Building on this, we propose a differentiable surface loss that simultaneously suppresses spurious mass and recovers absent structure, validated by fine-tuning three backbones (nnU-Net, SwinUNETR, NexToU) on two public benchmarks (Image-CAS, ASOCA). Against a matched-epoch control, it significantly improves surface F1 by recovering missed distal vessels at comparable Dice. Our findings argue for measuring and optimizing the vessel surface, not the volume it overlaps. Code

Figures & tables

Figure 1: Our two contributions, evaluated on identical error modes. (a) The surface metric matches predicted (blue, dashed) and reference (gold) surface points under a per-point radius tolerance $r_{i}$ , decoupling missed branches (FN) from spurious branches (FP) while accepted matches count as true positives (green). (b) The surface loss corrects both: a removal term drives the spurious branch down ( $\hat{p}\!\to\!0$ ) and a creation term grows the missed branch ( $\hat{p}\!\to\!1$ ), each weighted by a distance field (Eqs. 2 , 3 ). Both terms act directly on surface points, ensuring a thin distal branch receives equal gradient magnitude as a thick proximal trunk.

Vessel calibre	Missed (FN) rate $\downarrow$	Vol. share
Thin distal ( $r<0.8$ mm)	$39.5\%$	$1.2\%$
Medium ( $0.8\leq r<1.5$ mm)	$12.5\%$	$24.0\%$
Thick proximal ( $r\geq 1.5$ mm)	$4.5\%$	$74.9\%$

Table 1: Surface error by vessel calibre on the Dice+CE baseline (

n=200

). The missed (false-negative) rate rises sharply toward thin distal branches, yet those branches carry a negligible share of the volume that drives Dice. Radius from the reference distance transform; per-voxel volume share assigned by nearest medial-axis radius.

Backbone	Loss	Dice $\uparrow$	clDice $\uparrow$	$\beta_{0}$ -err $\downarrow$	FP $\downarrow$	FN $\downarrow$	$F_{1}\uparrow$
nnU-Net	Dice+CE	80.99	86.23	6.14	$14{,}972$	$22{,}124$	79.50
	clDice [ 6 ]	81.64 ‡	86.78	5.14	$14{,}180$	$21{,}542$	80.20 †
	cbDice [ 8 ]	81.92 ‡	86.99	4.53	$13{,}835$	$21{,}089$	80.50 †
	Ours	82.05 ‡	87.12	4.05	$\mathbf{13{,}215}$	$\mathbf{15{,}798}$	82.60 †
SwinUNETR	Dice+CE	79.23	83.28	20.44	$19{,}717$	$24{,}215$	77.30
	clDice [ 6 ]	80.33 ‡	85.14	14.42	$19{,}354$	$23{,}985$	78.50 †

Table 2: ImageCAS comparison across backbones. Dice, clDice,

F_{1}

higher is better;

\beta_{0}

-err, FP (spurious), FN (missed) lower is better. The decoupled FP/FN show which error each method reduces. Dice/clDice/

F_{1}

on a

0

–

100

scale; bold = best per backbone group.

\ddagger

\dagger

: significant over Dice+CE on Dice/

F_{1}

(Wilcoxon,

p<0.05

, Holm).

Backbone	Method	Dice $\uparrow$	clDice $\uparrow$	$\beta_{0}$ -err $\downarrow$	FP $\downarrow$	FN $\downarrow$	$F_{1}\uparrow$
nnU-Net	Best baseline	88.14	87.81	8.75	$4{,}270$	$5{,}157$	85.40
nnU-Net	Ours	88.54	87.98	7.22	$\mathbf{3{,}985}$	$\mathbf{4{,}520}$	87.10
SwinUNETR	Best baseline	85.11	84.77	15.62	$4{,}663$	$8{,}288$	81.90
SwinUNETR	Ours	85.74	85.12	13.25	$\mathbf{4{,}283}$	$\mathbf{7{,}235}$	82.50
NexToU	Best baseline	86.89	86.27	7.88	$4{,}559$	$5{,}639$	84.30
NexToU	Ours	87.25	86.90	6.54	$\mathbf{4{,}231}$	$\mathbf{4{,}725}$	84.90

Table 3: ASOCA second-dataset replication, strongest baseline vs. our method per backbone. Columns as in Table 2 . With

n=8

test cases we report point estimates only; the consistent direction corroborates the ImageCAS findings.

Figure 2: Qualitative comparison across the three backbones (rows) and the competing objectives (columns: reference, our method, clDice, cbDice, and the Dice+CE baseline). Predictions are coloured against the reference: green = correctly segmented (TP), red = missed reference vessel (FN), blue = spurious predicted vessel (FP). Our loss recovers missed terminal vessel (less red) across all three backbones while keeping over-segmentation (blue) rare.

Explore similar work

ImageCAS-X: a dataset and benchmark for coronary artery segmentation and centerline extraction in coronary CT angiography
Aug 31, 2026Kit M. Bransby, Esther Øksnebjerg, Kristoffer Kjær +7Medical Image Segmentation Coronary Artery Segmentation
DTU Compute, Technical University of Denmark, Kongens Lyngby, Denmark · Cardiovascular Research Unit, Copenhagen University Hospital – Rigshospitalet, Copenhagen, Denmark
CARDIAG: A Dense Segment Classification Benchmark of Deep Learning Architectures for Coronary Angiography
Jul 24, 2026Dominik Bernard Lau, Hubert Malinowski, Jerzy Szyjut +5Coronary Angiography Medical Image Segmentation
Gdańsk University of Technology · NASK National Research Institute · Medical University of Gdańsk
Multi-View Deformable Convolution Meets Visual Mamba for Coronary Artery Segmentation
Mar 23, 2026Xiaochan Yuan, Pai ZengCoronary Artery Segmentation 3D Medical Image Segmentation
Sichuan Agricultural University

Aug 31, 2026J/K move · Enter open · S save

ImageCAS-X: a dataset and benchmark for coronary artery segmentation and centerline extraction in coronary CT angiography

Kit M. Bransby, Esther Øksnebjerg, Kristoffer Kjær, Jacob Kirkeby+6

DTU Compute, Technical University of Denmark, Kongens Lyngby, Denmark · Cardiovascular Research Unit, Copenhagen University Hospital – Rigshospitalet, Copenhagen, Denmark

Medical Image Segmentation Coronary Artery Segmentation Computed Tomography Medical Image Benchmarks

Open paper AskPDF arXiv

Accurate segmentation of the coronary vessel lumen is a prerequisite for quantitative assessment of atherosclerotic plaque and perivascular adipose tissue in coronary computed tomography angiography (CCTA). Cardiologists rely on semi-automated methods for this task because manual vessel tracing and segmentation are labour-intensive. Although many automated methods have been proposed, their validation remains limited by the lack of large, high-quality publicly available datasets. We provide a new dataset of voxel-wise annotations of the vessel lumen and coronary segments, alongside centerlines, and mesh surfaces for 800 scans from the publicly available ImageCAS dataset. Using this dataset, we benchmark established lumen segmentation methods against inter-observer variability, stratifying performance by disease, image quality, coronary dominance, coronary segment, vessel diameter, and lumen attenuation. These labels allow segmentation accuracy to be described in anatomical and clinical context rather than reported as a single aggregate score. The dataset supports the development and validation of methods for lumen segmentation, plaque and perivascular quantification, and haemodynamic modelling.

Abstract

Explore similar work

ImageCAS-X: a dataset and benchmark for coronary artery segmentation and centerline extraction in coronary CT angiography

CARDIAG: A Dense Segment Classification Benchmark of Deep Learning Architectures for Coronary Angiography

Multi-View Deformable Convolution Meets Visual Mamba for Coronary Artery Segmentation