PaSAL: A Deep Learning Pipeline for Pulmonary Artery-Vein Segmentation and Anatomical Labeling in Thoracic CT

Jasper Eppink, Hoel Kervadec, Julian Capelleveen, Joost Verhoeff, Suresh Senan, Omar Bohoudi

Keywords: Pulmonary artery–vein segmentation, Pulmonary anatomical labeling, Thoracic CT, Deep learning, Graph-based learning

TL;DR: We present PaSAL, a pipeline that automatically segments pulmonary arteries and veins and assigns anatomical labels in CT scans using a combined deep-learning and graph-based approach.

Abstract:

We present PaSAL, a deep learning pipeline for pulmonary artery–vein segmentation and anatomical labeling in thoracic CT. PaSAL combines a nnU-Net-based binary vessel segmentation model with a graph-based anatomical labeling framework that assigns 19 clinically defined vascular classes. The pipeline integrates vessel enhancement, skeletonization, and topology-aware label propagation to produce anatomically coherent outputs.

PaSAL is trained on the HiPaS and PTL public datasets and evaluated on an external set of 63 clinical scans from Amsterdam UMC. On HiPaS, PaSAL achieves Dice scores of $89.5\%$ (arteries) and $88.1\%$ (veins). On PTL, voxel-level anatomical labeling accuracy reaches $90.1\%$ for arteries and $82.7\%$ for veins. Expert review confirms high anatomical plausibility and clinical utility, while showing weak correlation between standard quantitative metrics and perceived quality.

To our knowledge, PaSAL is the first method to jointly perform artery–vein segmentation and anatomical labeling in CT. The results demonstrate robust performance across diverse anatomical presentations, including post-treatment scans, and establish PaSAL as a strong baseline for vascular analysis in medical imaging.

Primary Subject Area: Segmentation

Secondary Subject Area: Application: Other

Registration Requirement: Yes

Visa & Travel: No

Read CFP & Author Instructions: Yes

Originality Policy: Yes

Single-blind & Not Under Review Elsewhere: Yes

LLM Policy: Yes

Midl Latex Submission Checklist: Ensure no LaTeX errors during compilation., Replace NNN with your OpenReview submission ID., Includes \documentclass{midl}, \jmlryear{2026}, \jmlrworkshop, \jmlrvolume, \editors, and correct \bibliography command., Did not override options of the hyperref package., Did not use the times package., Use the correct spelling and format, avoid Unicode characters, and use LaTeX equivalents instead., Any math in the title and abstract must be enclosed within $...$., Did not override the bibliography style defined in midl.cls and did not use \begin{thebibliography} directly to insert references., Avoid using \scalebox; use \resizebox when needed., Included all necessary figures and removed *unused* files in the zip archive., Removed special formatting, visual annotations, and highlights used during rebuttal., All special characters in the paper and .bib file use LaTeX commands (e.g., \'e for é)., No separate supplementary PDF uploads., Acknowledgements, references, and appendix must start after the main content.

Submission Number: 333