Urinary volatile organic compounds and faecal microbiome profiles in colorectal cancer.

ABSTRACT

AIM: Volatile organic compounds (VOCs) are potential biomarkers for diagnosing colorectal cancer (CRC). We characterized urinary VOCs from CRC patients, their spouses/cohabitors (spouses) and first-degree relatives (relatives) to determine any differences. Correlation with stool-derived microbiomes was also undertaken. METHODS: Urine from 56 CRC patients, 45 spouses and 37 relatives was assayed using liquid chromatography, field asymmetric ion mobility spectrometry (FAIMS), mass spectrometer technology. Analysis was performed using five-fold cross-validation and a random forest classifier. Faecal microbiome 16S rRNA was sequenced using Illumina MiSeq protocols and analysed using UPARSE and QIIME pipelines. VOC and microbiome profiles were also compared before and after cancer treatment. RESULTS: Urinary VOC profiles of CRC patients were indistinguishable from either spouses or relatives. When spouses and relatives were grouped together to form a larger non-cancer control group (n = 82), their VOC profiles became distinguishable from those of CRC patients (n = 56) with 69% sensitivity and specificity, area under the curve 0.72 (P < 0.001). Microbiome analysis identified > 1300 operational taxonomic units across all groups. The analysis of similarity R value was 0.067 (P < 0.001), with significantly different bacterial abundances in 82 operational taxonomic units (6.2%) by Kruskal-Wallis testing. CRC patients' VOC or stool microbiome profiles were unchanged after treatment. CONCLUSION: Although CRC patients' urinary VOC profiles cannot be differentiated from those of spouses or relatives they can be differentiated from a larger non-cancer control group. Comparison of the groups' microbiomes confirmed differences in bacterial species abundance. The current FAIMS-based assay can detect a unique, but modest, signal in CRC patients' urinary VOCs, which remains unaltered after treatment.