# Identification of a weighted urinary microbial signature for bladder cancer discrimination

## TL;DR

This study identifies a urinary microbial signature that can help distinguish bladder cancer patients from healthy individuals, potentially leading to a non-invasive diagnostic tool.

## Contribution

The novel contribution is the development of a weighted composite index (WCI) based on urinary microbiome data for bladder cancer discrimination.

## Key findings

- The WCI showed significant differences between bladder cancer patients and healthy controls at both genus and species levels.
- WCIs outperformed individual taxa in distinguishing between the two groups, indicating their potential as a robust biomarker.
- In-silico validation confirmed the models' high sensitivity and stability, though larger cohorts are needed for confirmation.

## Abstract

Growing evidence from microbiome studies has demonstrated associations between dysbiosis and cancers, including bladder cancer (BCa). Our recent works on urobiome revealed a different microbial composition in BCa patients compared to controls. The aim of this work was to create a Weighted Composite Index (WCI) to distinguish BCa-affected patients (mBCa) from healthy controls (mHC) in a cohort of male aged over 50 years.

Urobiome data from 51 subjects (27 mBCa and 24 mHC) were analyzed. Random Forest (RF) classifier was trained to identify genera and species which significantly contributed to discriminating between mBCa and mHC group. A weighted normalization approach was used to compute separate WCIs at genus and species levels and in-silico validation test were performed to assess the models’ robustness.

the WCI was calculated for each patient at both genera and species levels, showing a significant difference between the two groups (p < 0.0001) in both comparisons. WCIs showed superior discriminative performances compared to any individual taxon used for the model construction. Despite the need for validation in larger independent cohorts, the in-silico validation pipeline showed a stable high sensitivity of the models.

Our findings identified a candidate urinary microbial signature in a biomarker discovery setting associated with bladder cancer. This hypothesis-generating approach may contribute to the identification of a non-invasive biomarker, which requires validation in larger, independent cohorts before clinical application.