{"paper_id":"7c889323-1474-4449-9f6c-945ad3920b7f","body_text":"Content uploaded by Keira MacDonald\nAuthor content\nAll content in this area was uploaded by Keira MacDonald on Jan 14, 2026\nContent may be subject to copyright.\nZU064-05-FPR output 14 January 2026 14:2\n1\nUsing Natural Language Processing to Support\nEarly Detection of Endometriosis\nKeira Sian MacDonald1,2\n1University of Western Ontario, London, ON, Canada\n2University of Warwick, CV4 7AL, UK\n(e-mail: Keira.MacDonald@Warwick.ac.uk)\nAbstract\nEndometriosis is a chronic inflammatory condition that affects approximately 190 million women\nwho are assigned female at birth. Endometriosis lesions cause severe, often debilitating pain. De-\nspite its prevalence and impact, endometriosis remains one of the most misdiagnosed conditions\nin women’s health. Endometriosis diagnoses are often delayed several years due to the subjective\nnature of its symptoms. As a result, clinically relevant information exists primarily in patient nar-\nratives. This study explores natural language processing (NLP) techniques to extract meaningful\ndiagnostic signals from patient narratives. The aim of this study is to support earlier identification\nof individuals at risk of endometriosis. Anonymous patient narratives are analysed using NLP to\nparse common symptom descriptions and pain descriptors, which is used to train supervised machine\nlearning models to differentiate narratives associated with diagnosed endometriosis and control cases.\nModel interpretability techniques are applied to identify language patterns associated with risk.\nThe expected outcome is that NLP-based models will identify language patterns characteristic of\nendometriosis, including cyclic pain descriptions and evolving symptom predictions over time. These\nmodels are expected to outperform baseline heuristic approaches based on isolated symptom indi-\ncators. This suggests that patient narratives contain underutilized diagnostic signals relevant to early\nclinical consideration. This study highlights the feasibility of using NLP-driven analysis of patient-\nreported data as a scalable and non-invasive support tool for earlier identification of endometriosis. It\ndemonstrates the potential of computational methods to combat diagnostic delays in conditions that\nmay be widely experienced but underrepresented, providing a possible framework to support other\nunderdiagnosed diseases.\n1 References\n[1] Bontempo AC, Mikesell L. 2020. Patient Perceptions of Misdiagnosis of Endometrio-\nsis: Results from an Online National Survey. Diagnosis (Berl). May 26;7(2):97-106. doi:\n10.1515/dx-2019-0020. PMID: 32007945.\n[2] Horne A. W., Missmer S. A. 2022. Pathophysiology, Diagnosis, and Management of\nEndometriosis. BMJ: British Medical Journal, 379, 1–19. https://www.jstor.org/stable/27424071.\n[3] Ranney, B. 1952. Endometriosis. The American Journal of Nursing, 52(12), 1465–1467.\nhttps://doi.org/10.2307/3459462.\nZU064-05-FPR output 14 January 2026 14:2","source_license":"CC0","license_restricted":false}