# Exploring cervicovaginal microbiome differences between single and multiple endometrial polyps: implications for non-invasive classification

## TL;DR

This study explores differences in the cervicovaginal microbiome between single and multiple endometrial polyps, aiming to develop non-invasive classification methods.

## Contribution

The study identifies specific microbial signatures and uses a random forest model to distinguish single from multiple endometrial polyps non-invasively.

## Key findings

- Single endometrial polyps show distinct microbiome profiles with lower CST I and II and higher Sneathia amnii.
- Bacterial species like Mesorhizobium sp., Acinetobacter baumannii, and Pasteurella multocida are enriched in single polyp cases.
- A random forest model achieved 0.861 AUC in differentiating single and multiple endometrial polyps using microbiome and clinical data.

## Abstract

Single and multiple endometrial polyps (EP) are common gynecological conditions with differing recurrence rates, influencing clinical treatment decisions. This study aimed to characterize the reproductive tract microbiome in both subtypes to support the development of methods for the non-invasive categorization of EPs. Using metagenomic sequencing, we analyzed vaginal and cervical samples from 27 reproductive-aged patients with single EP and 22 with multiple EP. Compared with controls and multiple EP cases, single EP vaginal and cervical samples exhibited a lower percentage of community state types (CST) I and II. Sneathia amnii was identified as a characteristic species in both the vagina (P = 0.0051) and cervix (P = 0.0398) of single EP patients compared with controls. Mesorhizobium sp. (vaginal P = 0.0110, cervical P = 0.0210), Acinetobacter baumannii (vaginal P = 8.0 × 10−5, cervical P = 0.0314), and Pasteurella multocida (vaginal P = 0.0173, cervical P = 0.0210) were enriched in single EP compared with multiple EP. Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analysis of vaginal microbiome revealed unique pathways in single EP, including aminoacyl-tRNA biosynthesis, pantothenate and coenzyme A (CoA) synthesis, pyrimidine metabolism, glycolysis/gluconeogenesis, and biosynthesis of phenylalanine, tyrosine, and tryptophan. Using a random forest model, we further selected microbiota and clinical parameters to differentiate single and multiple EPs, thus achieving an area under curve (AUC) of 0.861. Our findings characterized the composition of the cervicovaginal microbiota of single and multiple EPs and proposed biomarkers for their non-invasive classification based on a random forest model.

The prevalence rate of endometrial polyps (EPs), a common gynecological condition, varies between 7.8% and 34.9%. Multiple EPs are associated with higher recurrence rates and chronic endometritis than single EPs and thus require more aggressive clinical interventions. However, only laparoscopic surgery can accurately identify single and multiple polyps. Non-invasive adjunctive diagnostic methods can aid in altering surgical indications preoperatively. Using metagenomic sequencing, we thoroughly analyzed the vaginal and cervical samples of 27 single EP and 22 multiple EP patients of reproductive age. We then identified distinct microbial patterns in the single and multiple samples, which were crucial for understanding EP pathogenesis and its association with gynecological health. Using a random forest model, key bacterial taxa that differentiate single and multiple EPs were identified with high accuracy. These could potentially serve as non-invasive diagnostic biomarkers. This research delineates the cervicovaginal microbiome of the reproductive tract in EP patients, offering a basis for developing non-invasive diagnostic tools and personalized treatment strategies.