{"paper_id":"465e1468-0b41-4920-89ae-ba204ee4127e","body_text":"Content uploaded by Keira MacDonald\nAuthor content\nAll content in this area was uploaded by Keira MacDonald on Apr 01, 2026\nContent may be subject to copyright.\nUniversity of Warwick\nCoventry, West-Midlands, UK\nDepartment of\nComputer Science\nUsing Natural Language Processing to Support Early Detection of Endometriosis\nKeira MacDonald\nUniversity of Warwick, UK, University of Western Ontario, Canada\nIntroduction\nEndometriosis is a chronic inflammatory condition that affects approximately 190 million\nwomen who are assigned female at birth. Endometriosis lesions cause severe, debilitating pain,\nyet remains one of the most misdiagnosed conditions in women’s health. As a result, clinically\nrelevant information exists primarily in patient narratives. By using natural language\nprocessing (NLP) techniques to parse patient narratives, supervised machine learning models\ncan be used to differentiate control cases and cases of diagnosed endometriosis. This\ndemonstrates the potential of using computational methods to combat diagnostic delays in\ncommon, underdiagnosed diseases.\nMethods\nFigure 1. Endometriosis misdiagnosis rate\n(Bontempo and Mikesell 2020).\nDataset & Preprocessing: a publicly available dataset was used (Jackson 2025). It contains\n10,000 synthetic but realistic data with clinical indicators of endometriosis. Data was cleaned by\nhandling missing values and standardizing categorical variables. Patient narratives were\nnormalized through tokenization, lowercasing, and removal of stopwords.\nImpact & Applications\nThis work demonstrates the potential of NLP-driven models to support an early identification of endometriosis using patient-\nreported narratives. An earlier detection could reduce diagnostic delays, improve patient outcomes, and enable more timely\nintervention in a condition that is underdiagnosed. The proposed approach can serve as a non-invasive clinical decision support\ntool, which can assist healthcare providers in identifying high-risk patients based on symptom language patterns. This has value in\ncases where traditional diagnostic pathways rely heavily on subjective symptom reporting.\nBecause this model operates on text data, it has potential to be integrated into digital health platforms, online symptom checkers, or\nAcknowledgements\nBontempo A, Mikesell L. 2020. Patient Perceptions of Misdiagnosis of Endometriosis: Results from an Online National Survey. Diagnosis 7(2):\n97-106. https://doi.org/10.1515/dx-2019-0020.\nGoogle. 2025. Machine Learning: Text Classification. https://developers.google.com/machine-learning/guides/text-classification/step-3\nHorne A. W., Missmer S. A. 2022. Pathophysiology, Diagnosis, and Management ofEndometriosis. BMJ: British Medical Journal, 379, 1–19.\nhttps://www.jstor.org/stable/27424071.\nJackson A. 2025. Endometriosis Dataset. Kaggle. https://www.kaggle.com/datasets/michaelanietie/endometriosis-dataset/data.\nRanney, B. 1952. Endometriosis. The American Journal of Nursing, 52(12), 1465–1467.https://doi.org/10.2307/3459462.\nTeh D. 2022. Not Just Period Pain. HealthMatch. https://healthmatch.io/blog/not-just-period-pain-the-endometriosis-crisis-that-affects-1-in-10-\nwomen.\nFigure 3. Total time taken from symptoms to medical visits (Teh, 2022).\nNatural Language Features & Model Development: features were\nextracted from symptom descriptions using tokenization and\nvectorization, and frequency-based keyword extraction. The\nsupervised machine learning classifier was trained to predict the\npresence of endometriosis. Logistic regression and random forest\nmodels were evaluated. Accuracy, precision, recall, and F1-score\nwas used to measure the effectiveness of NLP-based detection.\nFigure 2. An example of preparing data for machine learning through tokenization and\nvectorization (Google 2025).\nelectronic health records to provide scalable, low-cost preliminary screening\nfor endometriosis risk. This could vastly reduce the time taken between onset\nsymptoms to consultation to diagnosis. This model also provides a possible\nframework to support other underdiagnosed diseases.","source_license":"CC0","license_restricted":false}