Automatic uterus segmentation in transvaginal ultrasound using U-Net and nnU-Net
PLOS One · 2025
DOI: 10.1371/journal.pone.0336237 ↗Kurz gefasst
Fachartikel in der Zeitschrift PLOS One, 2025. Verfasst von Dilara Tank und 5 weiteren Autor:innen. Beteiligte Einrichtung: Department of Obstetrics and Gynecology, Amsterdam University Medical Center, location Vrije Universiteit Amsterdam, Amsterdam, North-Holland, the Ne…. Hauptthemen laut MEDLINE: Uterus, Image Processing, Computer-Assisted, Adenomyosis. Die Arbeit wurde bisher 3-mal zitiert (Crossref) und nennt 35 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).
Abstract (Originalsprache)
Purpose
Transvaginal ultrasound (TVUS) is pivotal for diagnosing reproductive pathologies in individuals assigned female at birth, often serving as the primary imaging method for gynecologic evaluation. Despite recent advancements in AI-driven segmentation, its application to gynecological ultrasound still needs further attention. Our study aims to bridge this gap by training and evaluating two state-of-the-art deep learning (DL) segmentation models on TVUS data. Materials and methods An experienced gynecological expert manually segmented the uterus in our TVUS dataset of 124 patients with adenomyosis, comprising still images (n = 122), video screenshots (n = 472), and 3D volume screenshots (n = 452). Two popular DL segmentation models, U-Net and nnU-Net, were trained on the entire dataset, and each imaging type was trained separately. Optimization for U-Net included varying batch size, image resolution, pre-processing, and augmentation. Model performance was measured using the Dice score (DSC).
Results
U-Net and nnU-Net had good mean segmentation performances on the TVUS uterus segmentation dataset (0.75 to 0.97 DSC). We observed that training on specific imaging types (still images, video screenshots, 3D volume screenshots) tended to yield better segmentation performance than training on the complete dataset for both models. Furthermore, nnU-Net outperformed the U-Net across all imaging types. Lastly, we report the best results using the U-Net model with limited pre-processing and augmentations.
Conclusions
TVUS datasets are well-suited for DL-based segmentation. nnU-Net training was faster and yielded higher segmentation performance; thus, it is recommended over manual U-Net tuning. We also recommend creating TVUS datasets that include only one imaging type and are as clutter-free as possible. The nnU-Net strongly benefited from being trained on 3D volume screenshots in our dataset, likely due to their lack of clutter. Further validation is needed to confirm the robustness of these models on TVUS datasets. Our code is available on https://github.com/dilaratank/UtiSeg .
Schlagwörter
MeSH-Schlagwörter (MEDLINE)
Autor:innen und Institutionen
- Department of Obstetrics and Gynecology, Amsterdam University Medical Center, location Vrije Universiteit Amsterdam, Amsterdam, North-Holland, the Netherlands.
- Department of Medical Informatics, University of Amsterdam, Amsterdam, North-Holland, the Netherlands.
- Department of Computer Science, University of Cambridge, Cambridge, United Kingdom.
- Amsterdam Reproduction and Development, Amsterdam, the Netherlands.
ORCID-Profile: Dilara Tank
Zitieren
APA
Tank, D., Schor, B. G. S., Trommelen, L. M., Huirne, J. A. F., Calixto, I., & de Leeuw, R. A. (2025). Automatic uterus segmentation in transvaginal ultrasound using U-Net and nnU-Net. PLOS One. https://doi.org/10.1371/journal.pone.0336237
Vancouver
Tank D, Schor BGS, Trommelen LM, Huirne JAF, Calixto I, de Leeuw RA. Automatic uterus segmentation in transvaginal ultrasound using U-Net and nnU-Net. PLOS One. 2025. doi:10.1371/journal.pone.0336237
BibTeX
@article{Tank2025,
author = {Tank, Dilara and Schor, Bianca G S and Trommelen, Lisa M and Huirne, Judith A F and Calixto, Iacer and de Leeuw, Robert A},
title = {Automatic uterus segmentation in transvaginal ultrasound using U-Net and nnU-Net},
journal = {PLOS One},
year = {2025},
doi = {10.1371/journal.pone.0336237}
}
EUVIMED