Developing symptom-based predictive models of endometriosis as a clinical screening tool: results from a multicenter...

Area of Science:

• Gynecology
• Surgical Oncology
• Medical Diagnostics

Background:

• Endometriosis diagnosis often involves a lengthy process, delaying treatment.
• Symptom-based prediction models could potentially streamline the diagnostic pathway.

Purpose of the Study:

• To develop and validate models predicting endometriosis in women before their first laparoscopy.
• To assess the accuracy of these symptom-based predictive models.

Main Methods:

• A prospective, observational, two-phase study involving 1,396 women undergoing laparoscopy.
• Data collected via a 25-item questionnaire assessing symptoms and patient characteristics.
• Logistic regression and ROC curve analysis used for model development and validation across two datasets.

Main Results:

• Models demonstrated relatively poor prediction for any-stage endometriosis (AUC=68.3).
• Models accurately predicted stage III and IV endometriosis (AUC=84.9, sensitivity 82.3%, specificity 75.8%).
• Menstrual dyschezia and history of ovarian cysts were strong predictors.

Conclusions:

• Symptom-based models offer valuable prediction for advanced endometriosis but are less effective for early stages.
• These predictive tools may aid in prioritizing surgical investigations and reducing diagnostic delays.
• Further validation in diverse populations is encouraged.