Enhancing auto-segmentation accuracy and robustness through multiple deep learning systems
PLOS One · 19.11.2026 · 21 (11) · e0363053
DOI: 10.1371/journal.pone.0363053 ↗Kurz gefasst
Fachartikel in der Zeitschrift PLOS One, online erschienen am 19.11.2026. Verfasst von Ying Zhang und 5 weiteren Autor:innen. Beteiligte Einrichtung: Department of Radiation Oncology& Medical Artificial Intelligence and Automation (MAIA) Lab, University of Texas Southwestern Medical Center, Dallas,…. Hauptthemen laut MEDLINE: Deep Learning, Tomography, X-Ray Computed, Image Processing, Computer-Assisted. Die Arbeit nennt 19 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Foundation for the National Institutes of Health.
Abstract (Originalsprache)
Purpose
Deep learning auto-segmentation (DLAS) can improve the efficiency and consistency of clinical contouring; however, no single DLAS system performs optimally across all cases, organs, and image slices, and each system may produce occasional poor-quality or outlier contours. Because multiple DLAS systems are available in some institutions, this proof-of-concept study evaluated a multi-system fusion framework designed to leverage their complementary strengths and generate more accurate and robust auto-contours, with potential extension to consensus generation from multiple manual contours.
Methods
Using 35 abdominal CT datasets from three institutions, contours of the stomach, duodenum, liver, and right kidney were generated by three commercial DLAS systems and one non-commercial research DLAS model. Two fusion approaches were evaluated: conventional Simultaneous Truth and Performance Level Estimation (STAPLE), which combined all available DLAS contours, and a selecting-and-combining DLAS (scDLAS) approach. For scDLAS, previously developed contour quality classification (CQC) models used supervised machine learning and seven quantitative metrics to classify each contour slice as acceptable, requiring minor edits, or requiring major edits. Only contours in the highest available quality category were retained for STAPLE fusion. scDLAS configurations incorporating different numbers of DLAS systems were evaluated against the reference contours using the Dice similarity coefficient (DSC), 95th-percentile Hausdorff distance (HD95), and changes in the distribution of contour-quality categories.
Results
No individual DLAS system consistently achieved the best performance across all four organs. STAPLE generally matched or outperformed the individual systems while maintaining comparable or lower interpatient variability. For the stomach, duodenum, liver, and right kidney, respectively, STAPLE achieved mean ± SD DSC values of 0.88 ± 0.10, 0.70 ± 0.13, 0.95 ± 0.02, and 0.92 ± 0.04, with corresponding HD95 values of 10.8 ± 12.6, 18.0 ± 12.3, 5.5 ± 2.2, and 6.8 ± 9.6 mm. The best-performing scDLAS configuration increased the mean DSC to 0.90 ± 0.09, 0.74 ± 0.14, 0.96 ± 0.02, and 0.94 ± 0.03, respectively. The highest duodenal DSC was achieved by 3scDLAS, whereas 4scDLAS achieved or tied the highest DSC for the other organs.
Conclusion
The proposed framework demonstrates the feasibility of leveraging complementary outputs from multiple DLAS systems to improve the accuracy and robustness of auto-segmented contours. These findings support further evaluation of multi-system fusion as a strategy for generating more reliable initial contours for subsequent clinical review and editing.
Schlagwörter
MeSH-Schlagwörter (MEDLINE)
Autor:innen und Institutionen
- Department of Radiation Oncology& Medical Artificial Intelligence and Automation (MAIA) Lab, University of Texas Southwestern Medical Center, Dallas, Texas, United States of America.
- Department of Radiation Medicine, MedStar Georgetown University Hospital, Washington, District of Columbia, United States of America.
- Department of Radiation Oncology, Medical College of Wisconsin, Milwaukee, Wisconsin, United States of America.
- Department of Radiation Oncology, Virginia Commonwealth University, Richmond, Virginia, United States of America.
- Department of Radiation Oncology, University of California, Los Angeles, California, United States of America.
ORCID-Profile: Ying Zhang
Förderung
- Foundation for the National Institutes of Health
Zitieren
APA
Zhang, Y., Zarenia, M., Sarosiek, C., Yuan, L., Paulson, E., & Qi, X. S. (2026). Enhancing auto-segmentation accuracy and robustness through multiple deep learning systems. PLOS One, 21(11), e0363053. https://doi.org/10.1371/journal.pone.0363053
Vancouver
Zhang Y, Zarenia M, Sarosiek C, Yuan L, Paulson E, Qi XS. Enhancing auto-segmentation accuracy and robustness through multiple deep learning systems. PLOS One. 2026;21(11):e0363053. doi:10.1371/journal.pone.0363053
BibTeX
@article{Zhang2026,
author = {Zhang, Ying and Zarenia, Mohammad and Sarosiek, Christina and Yuan, Lulin and Paulson, Eric and Qi, X Sharon},
title = {Enhancing auto-segmentation accuracy and robustness through multiple deep learning systems},
journal = {PLOS One},
year = {2026},
volume = {21},
number = {11},
pages = {e0363053},
doi = {10.1371/journal.pone.0363053}
}
EUVIMED