Abstract
Mucus plugs are clinically significant airway lesions in chronic obstructive pulmonary disease (COPD), but automated segmentation is hindered by two main challenges: data scarcity due to annotation difficulties, and attenuation overlap with adjacent soft tissues, such as vessels and nodules. We address these challenges using two strategies: first, we implement a two-class framework utilising concurrent airway segmentation to anatomically guide plug predictions within the airway lumen; second, we incorporate an augmentation cohort with naturally abundant and larger mucus plugs, providing a significant boost to positive training signal. Evaluated on expert-curated annotations across 85 internal scans, our approach achieves 83.2% sensitivity and 85.7% precision for plugs in airways ≥ 2 mm, with robust generalisation to an independent lung cancer screening cohort (70.3% sensitivity, 67.0% precision). These findings demonstrate that dataset-level augmentation with a high plug burden cohort and anatomical guidance effectively mitigates the key bottlenecks in mucus plug segmentation.
Links to Paper and Supplementary Materials
Main Paper (Open Access Version): https://papers.miccai.org/miccai-2026-sat/paper/TIA_023.pdf
SharedIt Link: Not yet available
SpringerLink (DOI): Not yet available
Supplementary Material: Not Submitted
Link to Open Review
Open Review Page: Not Available
BibTex
@InProceedings{WanSha_An_MICCAISAT2026,
author = { Wang, Shanshan AND McCabe, John AND Chen, Yunjie AND Azimbagirad, Mehran AND Vasudev, Pardeep AND Baxter, Gabrielle AND Lim, Jason AND Zahid, Azib AND Egashira, Ryoko AND Szmul, Adam AND Yang, Tianqi AND Cheng, Daryl AND Kinney, Greg L. AND Callister, Matthew E. J. AND Lynch, David A. AND Janes, Sam M. AND Alexander, Daniel C. AND Jacob, Joseph},
title = { { An automated segmentation model for Mucus Plug quantification in Chest CT } },
booktitle = {Medical Image Computing and Computer Assisted Intervention -- MICCAI 2026 Workshops and Challenges},
year = {2026},
publisher = {Springer Nature Switzerland},
volume = {LNCS 17269},
month = {pending},
page = {pending}
}