A preoperative predictive model for stage IV endometriosis.

Area of Science:

• Gynecology
• Surgical Oncology
• Reproductive Medicine

Background:

• Endometriosis is a chronic condition causing pain and infertility, often requiring surgery.
• Accurate preoperative identification of stage IV endometriosis is crucial for surgical planning and patient safety.
• Existing non-invasive predictive models for advanced endometriosis are limited.

Purpose of the Study:

• To develop and validate a predictive model for stage IV endometriosis.
• To identify non-invasive indicators for predicting advanced endometriosis preoperatively.

Main Methods:

• Retrospective study of 280 patients undergoing endometriosis surgery.
• Logistic regression analysis to identify predictors of stage IV endometriosis.
• Development and validation of a predictive model using key clinical variables.

Main Results:

• The predictive model incorporated visual analogue scale (VAS) pain score ≥4, painful uterosacral ligament nodularity, and bilateral endometriosis.
• The model demonstrated good predictive performance with an Area Under the Curve (AUC) of 0.777.
• Key predictors included VAS pain score (p=0.002), uterosacral ligament nodularity (p=0.015), and bilateral endometriosis (p=0.002).

Conclusions:

• A simple, non-invasive predictive model can effectively identify patients with stage IV endometriosis.
• Preoperative assessment including pain scores, pelvic examination, and imaging aids in predicting advanced disease.
• Improved preoperative diagnosis enhances surgical planning, reduces difficulty, and improves safety for endometriosis treatment.