Endo-MedSAM: a promptable vision foundation model adaptation for uterus segmentation on pelvic MRI in endometriosis

Rawan AlSaad, Shima Albasha, Rajat Thomas
In: Frontiers in Reproductive Health · 2026 · vol. 8 · doi:10.3389/frph.2026.1790980 · PMID:42266222 · W7162344626
article OA: gold CC0
🔓 Open OA copy View on OpenAlex View on PubMed View at publisher
⚙ AI-generated summary by claude@2026-06, 2026-06-07 ⓘ

Endo-MedSAM, an adapted MedSAM-2 model, achieves robust uterus segmentation on endometriosis pelvic MRI across institutions, significantly outperforming zero-shot performance with bounding-box and point prompts.

One-sentence paraphrase of the abstract; not a substitute for reading it. No clinical advice. How this works

⚙ AI-generated deep summary by claude@2026-06, 2026-06-07 ⓘ

This paper studied whether a promptable medical vision foundation model can segment the uterus on pelvic T2-weighted MRI in patients with endometriosis, using a slice-wise adaptation of MedSAM-2 called Endo-MedSAM. Using 74 endometriosis-focused subjects (3,449 slices) from two institutions (multicenter D1 with multiple scanners/sites and single-center D2 on one Philips scanner), the model was fine-tuned and evaluated with bounding-box prompts and low-interaction 1- or 2-click point prompts, with performance measured by 3D Dice and HD95 after reconstructing 3D volumes. Endo-MedSAM achieved mean 3D Dice around 0.81–0.88 with bounding-box prompts and 0.68–0.76 with point prompts, improving over zero-shot MedSAM-2 by roughly 0.27–0.34 absolute Dice for bounding-box prompting, with the largest gains for bounding boxes; the main caveat is that segmentation labels were derived from manual rater contouring with variable number of raters and consensus thresholding for D1. This paper is centrally about endometriosis—specifically, adapting and validating Endo-MedSAM for uterus segmentation on pelvic MRI in endometriosis across heterogeneous and single-center imaging datasets.

Read from the paper's body, not the abstract. Not a substitute for reading the paper. No clinical advice. How this works

Abstract

Background Endometriosis is a common gynecologic condition in which pelvic MRI plays an important role in diagnosis and preoperative assessment. AI-enabled automated uterus segmentation on pelvic MRI could support endometriosis care by enabling standardized volumetric measurements and quantitative imaging analyses. However, developing robust AI models for this task is challenging because endometriosis often distorts pelvic anatomy through adhesions, uterine displacement, and coexisting fibroids or adenomyosis. Although promptable medical vision foundation models, such as MedSAM-2, provide a promising framework for interactive segmentation, their zero-shot performance on endometriosis MRI remains limited. Objective To develop Endo-MedSAM, a uterus-focused adaptation of MedSAM-2 for pelvic MRI in endometriosis, and to systematically evaluate its segmentation performance across institutions and prompting strategies. Methods We used a pelvic MRI dataset comprising 74 subjects and 3,449 T2-weighted slices from two institutions (D1, multicenter; D2, single-center). Endo-MedSAM was initialized with MedSAM-2 weights and fine-tuned by training the prompt encoder and mask decoder while unfreezing the final layers of the image encoder. Slice-wise predictions were then reconstructed into 3D volumes for volumetric evaluation. Performance was assessed using slice-level and 3D Dice scores and the 95th percentile Hausdorff distance (HD95). Results Endo-MedSAM was evaluated in three configurations: training on the single-center D2 cohort and testing on the multicenter D1 cohort, training on D1 and testing on D2, and using a mixed D1 + D2 split with 75% of patients for training and 25% for testing. Across these settings, Endo-MedSAM achieved mean 3D Dice scores of 0.81–0.88 with bounding-box prompts and 0.68–0.76 with 1-click and 2-click point prompts. Compared with zero-shot MedSAM-2, this represented absolute 3D Dice improvements of approximately 0.27–0.34 for bounding-box prompting. Endo-MedSAM also showed markedly better performance across all prompting modes, with the largest gains observed for bounding-box prompting. Conclusion Endo-MedSAM achieved robust uterus segmentation on endometriosis pelvic MRI, consistently outperforming zero-shot MedSAM-2 across both multicenter and single-center settings while supporting bounding-box and low-interaction point prompting. Clinically, this can enable faster and more reproducible uterus delineation, reduce manual contouring burden, standardize measurements across scanners and sites, and support downstream quantitative imaging workflows in endometriosis care.

My notes (saved in your browser only)

Condition tags

endometriosisadenomyosis

Citation neighborhood

Papers in the corpus that this work cites (lower rings, blue) and that cite this one (upper rings, green). Dot size scales with the paper's in-corpus citation count — bigger dot = more influential within the endo/adeno field. Click a dot to open that paper. [ expand to 2 hops ] — adds papers reached through this work's immediate citers/citees. Heavier; up to 60 extra dots.

References (22)

Source provenance

europepmc
last seen: 2026-06-11T06:19:48.454388+00:00
openalex
last seen: 2026-06-10T17:14:06.276822+00:00
License: CC0 · commercial use OK