A Staged Deep Learning Framework for Endometriosis-Aware Pelvic MRI Analysis Using UT-EndoMRI

A Staged Deep Learning Framework for Endometriosis-Aware Pelvic MRI Analysis Using UT-EndoMRI DOI: https://doi.org/10.33411/IJIST/2040Keywords: Deep Learning, Endometriosis, Pelvic Mri, Segmentation, Ut-EndomriAbstract Endometriosis remains challenging to assess on pelvic MRI because lesions are often small, heterogeneous, and poorly distinguished from surrounding anatomy. This study developed a reproducible two-stage framework using the public UT-EndoMRI dataset to investigate whether lesion-aware segmentation can provide informative representations for subject-level endometriosis-aware prediction. In Stage 1, SegResNet, SwinUNETR, and SwinUNETR with lesion-aware refinement were compared for multi-class pelvic MRI segmentation. The refined SwinUNETR achieved the best test macro-Dice of 0.0786 and macro IoU of 0.0447, although ovary and lesion segmentation remained weak. The selected model was subsequently frozen and used to generate segmentation-informed features for Stage 2. A cohort of 121 subjects comprising 110 proxy/inferred positive and 11 proxy negative cases was evaluated using repeated five-fold stratified cross-validation over five repeats with L2-regularized logistic regression. The combined uncalibrated feature model achieved a balanced accuracy of 0.8215 ± 0.0798, MCC of 0.3953 ± 0.1172, and AUC of 0.8597 ± 0.0578. Subject-aggregated estimates were balanced accuracy 0.8364 (95% CI: 0.7955–0.8818), MCC 0.3968 (95% CI: 0.3407–0.4765), and AUC 0.8620 (95% CI: 0.7793–0.9306; permutation p = 0.00498). Isotonic calibration improved probability quality, reducing the Brier score to 0.0653 and ECE to 0.0380. These findings indicate that segmentation-informed representations can support internal subject-level prediction beyond coarse MRI summaries; however, the small negative class, weak lesion segmentation, proxy/inferred labels, possible target–predictor dependency, and absence of external validation preclude claims of clinical diagnostic validity. References C. M. Becker et al., “ESHRE guideline: endometriosis,” Hum. Reprod. open, vol. 2022, no. 2, 2022, doi: 10.1093/HROPEN/HOAC009. B. Sivajohan, M. Elgendi, C. Menon, C. Allaire, P. Yong, and M. A. Bedaiwy, “Clinical use of artificial intelligence in endometriosis: a scoping review,” NPJ Digit. Med., vol. 5, no. 1, p. 109, Dec. 2022, doi: 10.1038/S41746-022-00638-1. S. Guerriero et al., “Artificial intelligence (AI) in the detection of rectosigmoid deep endometriosis,” Eur. J. Obstet. Gynecol. Reprod. Biol., vol. 261, pp. 29–33, Jun. 2021, doi: 10.1016/j.ejogrb.2021.04.012. G. Maicas et al., “Deep learning to diagnose pouch of Douglas obliteration with ultrasound sliding sign,” Reprod. Fertil., vol. 2, no. 4, p. 236, Oct. 2021, doi: 10.1530/RAF-21-0031. K. Miao, Q. Lv, L. Zhang, N. Zhao, and X. Dong, “Discriminative diagnosis of ovarian endometriosis cysts and benign mucinous cystadenomas based on the ConvNeXt algorithm,” Eur. J. Obstet. Gynecol. Reprod. Biol., vol. 298, pp. 135–139, Jul. 2024, doi: 10.1016/j.ejogrb.2024.05.010. H. Wang et al., “Human-AI collaborative multi-modal multi-rater learning for endometriosis diagnosis,” Phys. Med. Biol., vol. 70, no. 1, Jan. 2024, doi: 10.1088/1361-6560/AD997E. A. Bondarenko et al., “Object Detection in Laparoscopic Surgery: A Comparative Study of Deep Learning Models on a Custom Endometriosis Dataset,” Diagnostics 2025, Vol. 15, Page 1254, vol. 15, no. 10, p. 1254, May 2025, doi: 10.3390/DIAGNOSTICS15101254. A. Hernández et al., “Deep Learning to Measure the Intensity of Indocyanine Green in Endometriosis Surgeries with Intestinal Resection,” J. Pers. Med., vol. 12, no. 6, Jun. 2022, doi: 10.3390/JPM12060982. D. Raimondo et al., “Application of Deep Learning Model in the Sonographic Diagnosis of Uterine Adenomyosis,” Int. J. Environ. Res. Public Health, vol. 20, no. 3, Feb. 2023, doi: 10.3390/IJERPH20031724. U. Tore et al., “Diagnosis of Endometriosis Based on Comorbidities: A Machine Learning Approach,” Biomedicines, vol. 11, no. 11, Nov. 2023, doi: 10.3390/BIOMEDICINES11113015. S. Bendifallah et al., “Machine learning algorithms as new screening approach for patients with endometriosis,” Sci. Reports 2022 121, vol. 12, no. 1, pp. 639-, Jan. 2022, doi: 10.1038/s41598-021-04637-2. A. Goldstein and S. Cohen, “Self-report symptom-based endometriosis prediction using machine learning,” Sci. Reports 2023 131, vol. 13, no. 1, pp. 5499-, Apr. 2023, doi: 10.1038/s41598-023-32761-8. C. R. Nezhat et al., “Real world perspectives on endometriosis disease phenotyping through surgery, omics, health data, and artificial intelligence,” npj Women’s Heal. 2025 31, vol. 3, no. 1, pp. 8-, Feb. 2025, doi: 10.1038/s44294-024-00052-w. X. Liang et al., “A Multi-Modal Pelvic MRI Dataset for Deep Learning-Based Pelvic Organ Segmentation in Endometriosis,” Sci. Data 2025 121, vol. 12, no. 1, pp. 1292-, Jul. 2025, doi: 10.1038/s41597-025-05623-3. X. Liang et al., “UTHealth - Endometriosis MRI Dataset (UT-EndoMRI)”, doi: 10.5281/ZENODO.13749613. Olaf Ronneberger, Philipp Fischer, Thomas Brox., “U-Net: Convolutional Networks for Biomedical Image Segmentation,” arXiv, vol. 1, 2015, doi: https://doi.org/10.48550/arXiv.1505.04597. A. Hatamizadeh, V. Nath, Y. Tang, D. Yang, H. R. Roth, and D. Xu, “Swin UNETR: Swin Transformers for Semantic Segmentation of Brain Tumors in MRI Images,” Lect. Notes Comput. Sci. (including Subser. Lect. Notes Artif. Intell. Lect. Notes Bioinformatics), vol. 12962 LNCS, pp. 272–284, Jan. 2022, doi: 10.1007/978-3-031-08999-2_22. F. Isensee, P. F. Jaeger, S. A. A. Kohl, J. Petersen, and K. H. Maier-Hein, “nnU-Net: a self-configuring method for deep learning-based biomedical image segmentation,” Nat. Methods 2020 182, vol. 18, no. 2, pp. 203–211, Dec. 2020, doi: 10.1038/s41592-020-01008-z. Z. Liu et al., “Swin Transformer: Hierarchical Vision Transformer using Shifted Windows,” Proc. IEEE Int. Conf. Comput. Vis., pp. 9992–10002, 2021, doi: 10.1109/ICCV48922.2021.00986. M. J. Cardoso et al., “MONAI: An open-source framework for deep learning in healthcare,” Nov. 2022, Accessed: Sep. 06, 2026. [Online]. Available: https://arxiv.org/pdf/2211.02701 S. Mittal, A. Tong, S. Young, and P. Jha, “Artificial intelligence applications in endometriosis imaging,” Abdom. Radiol. 2025 5010, vol. 50, no. 10, pp. 4901–4913, Apr. 2025, doi: 10.1007/S00261-025-04897-W. M. Haber, M. Montebello, G. P. Gauci, F. Zarb, and K. B. Grima, “Bridging the Gap Between Artificial Intelligence and Clinical Readiness in Endometriosis Diagnosis: A Systematic Review,” J. Minim. Invasive Gynecol., vol. 33, no. 9, pp. 1242–1250, Sep. 2026, doi: 10.1016/j.jmig.2026.04.016. A. S. Tejani et al., “Checklist for Artificial Intelligence in Medical Imaging (CLAIM):2024 Update,” https://doi.org/10.1148/ryai.240300, vol. 6, no. 4, May 2024, doi: 10.1148/RYAI.240300. G. S. Collins et al., “TRIPOD+AI statement: Updated guidance for reporting clinical prediction models that use regression or machine learning methods,” BMJ, vol. 385, p. e078378, Apr. 2024, doi: 10.1136/BMJ-2023-078378. B. Vasey et al., “Reporting guideline for the early-stage clinical evaluation of decision support systems driven by artificial intelligence: DECIDE-AI,” Nat. Med. 2022 285, vol. 28, no. 5, pp. 924–933, May 2022, doi: 10.1038/s41591-022-01772-9. L. Maier-Hein et al., “Metrics reloaded: recommendations for image analysis validation,” Nat. Methods 2024 212, vol. 21, no. 2, pp. 195–212, Feb. 2024, doi: 10.1038/s41592-023-02151-z. Olaf Ronneberger, Philipp Fischer, Thomas Brox, “U-Net: Convolutional Networks for Biomedical Image Segmentation,” arXiv:1505.04597, 2015, [Online]. Available: https://arxiv.org/abs/1505.04597 Downloads Published How to Cite Issue Section License Copyright (c) 2026 50sea This work is licensed under a Creative Commons Attribution 4.0 International License.