{"paper_id":"5ae534f2-cc8f-4396-9c2a-2efb0c97e21f","body_text":"Abstract\nBackground and purpose\nEndometriosis affects 5–10% of women of reproductive age. Despite its prevalence, diagnosing endometriosis through imaging remains challenging. Advances in deep learning (DL) are revolutionizing the diagnosis and management of complex medical conditions. This study aims to evaluate DL tools in enhancing the accuracy of multi-sequence MRI-based detection of endometriosis.\nMethod\nWe gathered a patient cohort from our institutional database, composed of patients with pathologically confirmed endometriosis from 2015 to 2024. We created an age-matched control group that underwent a similar MR protocol without an endometriosis diagnosis. We used sagittal fat-saturated T1-weighted (T1W FS) pre- and post-contrast and T2-weighted (T2W) MRIs. Our dataset was split at the patient level, allocating 12.5% for testing and conducting seven-fold cross-validation on the remainder. Seven abdominal radiologists with experience in endometriosis MRI and complex surgical planning and one women’s imaging fellow with specific training in endometriosis MRI reviewed a random selection of images and documented their endometriosis detection.\nResults\n395 and 356 patients were included in the case and control groups respectively. The final 3D-DenseNet-121 classifier model demonstrated robust performance. Our findings indicated the most accurate predictions were obtained using T2W, T1W FS pre-, and post-contrast images. Using an ensemble technique on the test set resulted in an F1 Score of 0.881, AUROCC of 0.911, sensitivity of 0.976, and specificity of 0.720. Radiologists achieved 84.48% and 87.93% sensitivity without and with AI assistance in detecting endometriosis. The agreement among radiologists in predicting labels for endometriosis was measured as a Fleiss’ kappa of 0.5718 without AI assistance and 0.6839 with AI assistance.\nConclusion\nThis study introduced the first DL model to use multi-sequence MRI on a large cohort, showing results equivalent to human detection by trained readers in identifying endometriosis.\nKey findings\nEndometriosis diagnosis using multi-sequence MRI and a 3D-DenseNet-121 deep-learning model showed promising results. The model achieved F1 Score: 0.881, AUROCC: 0.911, sensitivity: 0.976, and specificity: 0.720.\nAI slightly improved radiologists’ sensitivity from 84.48 to 87.93% and increased their agreement (Fleiss’ kappa: 0.5718 to 0.6839), highlighting its potential to enhance clinical detection of endometriosis while reducing reading time.\nHighlights\nDeep-learning models leveraging multi-sequence MRI significantly enhance endometriosis detection accuracy, surpassing traditional methods and improving radiologists’ performance, marking an advance in imaging-based diagnostics.\nAbstractSection Graphical abstractSimilar content being viewed by others\nData availability\nNo datasets were generated or analysed during the current study.\nAbbreviations\n- CNNs:\n-\nConvolutional Neural Networks\n- DL:\n-\nDeep Learning\n- MRI:\n-\nMagnetic Resonance Imaging\n- AI:\n-\nArtificial Intelligence\n- ROS:\n-\nReactive Oxygen Species\n- T2W:\n-\nT2-weighted imaging\n- T1W:\n-\nT1-weighted images\n- CE-T1W:\n-\nContrast-enhanced T1-weighted images\n- AUROCC:\n-\nArea Under the ROC Curve\n- US:\n-\nUltrasonography or Ultrasound\n- FS:\n-\nFat saturated\nReferences\nBennett GL, Slywotzky CM, Cantera M, Hecht EM. Unusual manifestations and complications of endometriosis–spectrum of imaging findings: pictorial review. AJR Am J Roentgenol. 2010;194: WS34–46.\nLu PY, Ory SJ. Endometriosis: current management. Mayo Clin Proc. 1995;70: 453–463.\nHeaps JM, Nieberg RK, Berek JS. Malignant neoplasms arising in endometriosis. Obstet Gynecol. 1990;75: 1023–1028.\nChamié LP, VanBuren WM, Xiao L, Pires Franco IV, Feldman MK, Causa Andrieu P, et al. Postmenopausal endometriosis: Clinical insights and imaging considerations. Radiographics. 2024;44: e240046.\nOzkan S, Murk W, Arici A. Endometriosis and infertility: epidemiology and evidence-based treatments. Ann N Y Acad Sci. 2008;1127: 92–100.\nMeuleman C, Vandenabeele B, Fieuws S, Spiessens C, Timmerman D, D’Hooghe T. High prevalence of endometriosis in infertile women with normal ovulation and normospermic partners. Fertil Steril. 2009;92: 68–74.\nFrankel LR. A 10-Year Journey to Diagnosis With Endometriosis: An Autobiographical Case Report. Cureus. 2022;14: e21329.\nPractice Committee of the American Society for Reproductive Medicine. Endometriosis and infertility: a committee opinion. Fertil Steril. 2012;98: 591–598.\nCoccia ME, Nardone L, Rizzello F. Endometriosis and Infertility: A Long-Life Approach to Preserve Reproductive Integrity. Int J Environ Res Public Health. 2022;19. doi:https://doi.org/10.3390/ijerph19106162\nSimoens S, Dunselman G, Dirksen C, Hummelshoj L, Bokor A, Brandes I, et al. The burden of endometriosis: costs and quality of life of women with endometriosis and treated in referral centres. Hum Reprod. 2012;27: 1292–1299.\nBourgioti C, Preza O, Panourgias E, Chatoupis K, Antoniou A, Nikolaidou ME, et al. MR imaging of endometriosis: Spectrum of disease. Diagn Interv Imaging. 2017;98: 751–767.\nBecker CM, Bokor A, Heikinheimo O, Horne A, Jansen F, Kiesel L, et al. ESHRE guideline: endometriosis. Hum Reprod Open. 2022;2022: hoac009.\nHirsch M, Begum MR, Paniz É, Barker C, Davis CJ, Duffy J. Diagnosis and management of endometriosis: a systematic review of international and national guidelines. BJOG. 2018;125: 556–564.\nAgarwal SK, Chapron C, Giudice LC, Laufer MR, Leyland N, Missmer SA, et al. Clinical diagnosis of endometriosis: a call to action. Am J Obstet Gynecol. 2019;220: 354.e1–354.e12.\nPractice bulletin no. 114: management of endometriosis. Obstet Gynecol. 2010;116: 223–236.\nYoung SW, Jha P, Chamié L, Rodgers S, Kho RM, Horrow MM, et al. Society of Radiologists in Ultrasound Consensus on Routine Pelvic US for Endometriosis. Radiology. 2024;311: e232191.\nTong A, VanBuren WM, Chamié L, Feldman M, Hindman N, Huang C, et al. Recommendations for MRI technique in the evaluation of pelvic endometriosis: consensus statement from the Society of Abdominal Radiology endometriosis disease-focused panel. Abdom Radiol (NY). 2020;45: 1569–1586.\nGuerriero S, Condous G, van den Bosch T, Valentin L, Leone FPG, Van Schoubroeck D, et al. Systematic approach to sonographic evaluation of the pelvis in women with suspected endometriosis, including terms, definitions and measurements: a consensus opinion from the International Deep Endometriosis Analysis (IDEA) group. Ultrasound Obstet Gynecol. 2016;48: 318–332.\nBazot M, Daraï E. Diagnosis of deep endometriosis: clinical examination, ultrasonography, magnetic resonance imaging, and other techniques. Fertil Steril. 2017;108: 886–894.\nKido A, Himoto Y, Moribata Y, Kurata Y, Nakamoto Y. MRI in the diagnosis of endometriosis and related diseases. Korean J Radiol. 2022;23: 426–445.\nTong A, Cope AG, Waters TL, McDonald JS, VanBuren WM. Best practices: Ultrasound versus MRI in the assessment of pelvic endometriosis. AJR Am J Roentgenol. 2024; 1–16.\nVanBuren W, Feldman M, Shenoy-Bhangle AS, Sakala MD, Young S, Chamie LP, et al. Radiology state-of-the-art review: Endometriosis imaging interpretation and reporting. Radiology. 2024;312: e233482.\nMurphy AA, Santanam N, Parthasarathy S. Endometriosis: a disease of oxidative stress? Semin Reprod Endocrinol. 1998;16: 263–273.\nDonnez J, Binda MM, Donnez O, Dolmans M-M. Oxidative stress in the pelvic cavity and its role in the pathogenesis of endometriosis. Fertil Steril. 2016;106: 1011–1017.\nVan Langendonckt A, Casanas-Roux F, Donnez J. Oxidative stress and peritoneal endometriosis. Fertil Steril. 2002;77: 861–870.\nCacciottola L, Donnez J, Dolmans M-M. Can Endometriosis-Related Oxidative Stress Pave the Way for New Treatment Targets? Int J Mol Sci. 2021;22. doi:https://doi.org/10.3390/ijms22137138\nDefrère S, Van Langendonckt A, Vaesen S, Jouret M, González Ramos R, Gonzalez D, et al. Iron overload enhances epithelial cell proliferation in endometriotic lesions induced in a murine model. Hum Reprod. 2006;21: 2810–2816.\nVan Langendonckt A, Casanas-Roux F, Dolmans M-M, Donnez J. Potential involvement of hemoglobin and heme in the pathogenesis of peritoneal endometriosis. Fertil Steril. 2002;77: 561–570.\nVan Langendonckt A, Casanas-Roux F, Eggermont J, Donnez J. Characterization of iron deposition in endometriotic lesions induced in the nude mouse model. Hum Reprod. 2004;19: 1265–1271.\nWyatt J, Fernando SM, Powell SG, Hill CJ, Arshad I, Probert C, et al. The role of iron in the pathogenesis of endometriosis: a systematic review. Hum Reprod Open. 2023;2023: hoad033.\nVan Langendonckt A, Casanas-Roux F, Donnez J. Iron overload in the peritoneal cavity of women with pelvic endometriosis. Fertil Steril. 2002;78: 712–718.\nMuñoz M, Villar I, García-Erce JA. An update on iron physiology. World J Gastroenterol. 2009;15: 4617–4626.\nWood JC. Use of magnetic resonance imaging to monitor iron overload. Hematol Oncol Clin North Am. 2014;28: 747–64, vii.\nAmin K, Mileto A, Kolokythas O. MRI for Liver Iron Quantification: Concepts and Current Methods. Semin Ultrasound CT MR. 2022;43: 364–370.\nGhugre NR, Wood JC. Relaxivity-iron calibration in hepatic iron overload: probing underlying biophysical mechanisms using a Monte Carlo model. Magn Reson Med. 2011;65: 837–847.\nPedregosa F, Varoquaux G, Gramfort A, Michel V, Thirion B, Grisel O, et al. Scikit-learn: Machine Learning in Python. J Mach Learn Res. 2011;12: 2825–2830.\nRouzrokh P, Khosravi B, Faghani S, Moassefi M, Vera Garcia DV, Singh Y, et al. Mitigating Bias in Radiology Machine Learning: 1. Data Handling. Radiol Artif Intell. 2022;4: e210290.\nZhang K, Khosravi B, Vahdati S, Faghani S, Nugen F, Rassoulinejad-Mousavi SM, et al. Mitigating Bias in Radiology Machine Learning: 2. Model Development. Radiol Artif Intell. 2022;4: e220010.\nFaghani S, Moassefi M, Madhavan AA, Mark IT, Verdoorn JT, Erickson BJ, et al. Identifying Patients with CSF-Venous Fistula Using Brain MRI: A Deep Learning Approach. AJNR Am J Neuroradiol. 2024;45: 439–443.\nPolak G, Barczyński B, Wertel I, Kwaśniewski W, Bednarek W, Derewianka-Polak M, et al. Disrupted iron metabolism in peritoneal fluid may induce oxidative stress in the peritoneal cavity of women with endometriosis. Ann Agric Environ Med. 2018;25: 587–592.\nGanz T. Macrophages and iron metabolism. Microbiol Spectr. 2016;4. doi:https://doi.org/10.1128/microbiolspec.MCHD-0037-2016\nAllavena G, Carrarelli P, Del Bello B, Luisi S, Petraglia F, Maellaro E. Autophagy is upregulated in ovarian endometriosis: a possible interplay with p53 and heme oxygenase-1. Fertil Steril. 2015;103: 1244–51.e1.\nBruyere C, Maniou I, Habre C, Kalovidouri A, Pluchino N, Montet X, et al. Pelvic MRI for Endometriosis: A Diagnostic Challenge for the Inexperienced Radiologist. How Much Experience Is Enough? Acad Radiol. 2021;28: 345–353.\nMaicas G, Leonardi M, Avery J, Panuccio C, Carneiro G, Hull ML, et al. Deep learning to diagnose pouch of Douglas obliteration with ultrasound sliding sign. Reprod Fertil Dev. 2021;2: 236–243.\nGuerriero S, Pascual M, Ajossa S, Neri M, Musa E, Graupera B, et al. Artificial intelligence (AI) in the detection of rectosigmoid deep endometriosis. Eur J Obstet Gynecol Reprod Biol. 2021;261: 29–33.\nGanjizadeh A, Zawada S, Langer S, et al. Visualizing Clinical Data Retrieval and Curation in Multimodal Healthcare AI Research: A Technical Note on RIL-workflow. J Digit Imaging Inform Med. 2024;37(6):1239–1247. https://doi.org/10.1007/s10278-024-00977-3\nRouzrokh P, Khosravi B, Faghani S, Moassefi M, Vera Garcia DV, Singh Y, et al. Mitigating Bias in Radiology Machine Learning: 1. Data Handling. Radiol Artif Intell. 2022;4: e210290.\nAcknowledgements\nFunding: None.Statement and Declaration: Competing Interest: Dr. VanBuren has received funding from the following societies to support travel for educational lectures at: the American Association of Gynecological Laparoscopists (AAGL), the American Society of Reproductive Medicine (ASRM), the American Roentgen Ray Society (ARRS), the World Endometriosis Congress (WEC) and from the Society for Women’s Health Research (SWHR) to advocate for endometriosis on Capitol Hill. She is the founder of the International Endometriosis Imaging Congress.\nAuthor information\nAuthors and Affiliations\nContributions\nM.M. and S.F. authored the initial draft of the manuscript and revised it based on feedback provided by B.E. and W.V. The development of the AI model was conducted by M.M. and S.F., under the supervision of B.E. Data collection and curation were contributed by C.C., W.V., T.B., and S.G. All other authors participated in the radiological review of the imaging data and the detection of endometriosis. The final manuscript was reviewed and approved by all authors.\nCorresponding author\nEthics declarations\nCompeting interests\nThe authors declare no competing interests.\nAdditional information\nPublisher’s note\nSpringer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.\nRights and permissions\nSpringer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.\nAbout this article\nCite this article\nMoassefi, M., Faghani, S., Colak, C. et al. Advancing endometriosis detection in daily practice: a deep learning-enhanced multi-sequence MRI analytical model. Abdom Radiol 51, 238–249 (2026). https://doi.org/10.1007/s00261-025-04942-8\nReceived:\nRevised:\nAccepted:\nPublished:\nVersion of record:\nIssue date:\nDOI: https://doi.org/10.1007/s00261-025-04942-8","source_license":"CC0","license_restricted":false}