Automated Segmentation of Liver and Liver Tumors with SwinUNETR and UNET Neural Networks on 18F-FDG PET/CT
Both SwinUNETR and Residual UNET achieved excellent liver segmentation accuracy and clinically acceptable tumor segmentation performance on <sup>18</sup>F-FDG PET/CT, with SwinUNETR showing slightly better performance in liver volumetric measurements. These open-source models could be integrated into clinical workflows to automate segmentation tasks, facilitate treatment planning for liver-directed therapies, and support reproducible quantitative imaging analyses.
Authors
- Merve Ağcıoğlu Atalay (ORCID: https://orcid.org/0000-0001-6172-6866)
- Burak Demir (ORCID: https://orcid.org/0000-0002-0966-9988)
- Fikret Ertek (ORCID: https://orcid.org/0009-0008-8068-310X)
- Hatice Kübra Yurtçu (ORCID: https://orcid.org/0000-0002-2615-1056)
Institutions
- University of Health Science (KH)
- Sağlık Bilimleri Üniversitesi (TR)
- University of Health Sciences Antigua (AG)
Publication Details
- Journal
- Molecular Imaging and Radionuclide Therapy
- Published
- 2025-12-01
- DOI
- https://doi.org/10.4274/mirt.galenos.2025.58751
- Primary Topic
- Medical Imaging Techniques and Applications
- Type
- article
- Field-Weighted Citation Impact
- 0.00