Exploring cervicovaginal microbiome differences between single and multiple endometrial polyps: implications for...

Area of Science:

• Reproductive Medicine
• Microbiome Research
• Gynecological Oncology

Background:

• Endometrial polyps (EPs) are common gynecological conditions with varying recurrence rates.
• Distinguishing between single and multiple EPs is crucial for treatment decisions, but current methods like laparoscopy are invasive.
• Non-invasive diagnostic methods are needed to accurately categorize EP subtypes preoperatively.

Purpose of the Study:

• To characterize and compare the reproductive tract microbiome in patients with single versus multiple endometrial polyps (EPs).
• To identify potential microbial biomarkers for the non-invasive differentiation of single and multiple EPs.
• To develop a predictive model for EP subtype classification.

Main Methods:

• Metagenomic sequencing was performed on vaginal and cervical samples from 27 patients with single EPs and 22 with multiple EPs.
• Community State Types (CSTs) and specific bacterial species were analyzed and compared between groups.
• A random forest model was utilized to identify key microbiota and clinical parameters for differentiating EP subtypes.

Main Results:

• Single EP cases showed lower percentages of CST I and II compared to controls and multiple EP cases.
• Specific bacteria, including *Sneathia amnii*, *Mesorhizobium* sp., *Acinetobacter baumannii*, and *Pasteurella multocida*, were differentially abundant between single and multiple EPs.
• A random forest model achieved an AUC of 0.861 in differentiating single from multiple EPs, highlighting potential non-invasive biomarkers.

Conclusions:

• The cervicovaginal microbiome composition differs significantly between single and multiple endometrial polyps.
• Identified microbial signatures and pathways offer potential for developing non-invasive diagnostic tools for EP classification.
• These findings pave the way for personalized treatment strategies based on non-invasive EP subtyping.