Investigating the oral microbiome in retrospective and prospective cases of prostate, colon, and breast cancer.

The human microbiome has been proposed as a useful biomarker for several different human diseases including various cancers. To answer this question, we examined salivary samples from two Canadian population cohorts, the Atlantic Partnership for Tomorrows Health project (PATH) and Albertas Tomorrow Project (ATP). Sample selection was then divided into both a retrospective and prospective case control design examining individuals with prostate, breast, or colon cancer. In total 89 retrospective and 260 prospective cancer cases were matched to non-cancer controls and saliva samples were sequenced using 16S rRNA gene sequencing to compare bacterial diversity, and taxonomic composition. We found no significant differences in alpha or beta diversity across any of the three cancer types and two study designs. Although retrospective colon cancer samples did show evidence on visual clustering in weighted beta diversity metrics. Differential abundance analysis of individual taxon showed several taxa that were associated with previous cancer diagnosis in all three groupings within the retrospective study design. However, only one genus (Ruminococcaceae UCG-014) in breast cancer and one ASV (Fusobacterium periodonticum) in colon cancer was identified by more than one differential abundance (DA) tool. In prospective cases of disease three ASVs were associated with colon cancer, one ASV with breast cancer, and one ASV with prostate cancer. None overlapped between the two different study cohorts. Attempting to identify microbial signals using Random Forest classification showed relatively low levels of signal in both prospective and retrospective cases of breast and prostate cancer (AUC range: 0.394-0.665). Contrastingly, colon cancer did show signal in our retrospective analysis (AUC: 0.745) and in one of two prospective cohorts (AUC: 0.717). Overall, our results indicate that it is unlikely that reliable oral microbial biomarkers of disease exist in the context of both breast and prostate cancer. However, they do suggest that further research into the relationship between the oral microbiome and colon cancer could be fruitful. Particularly in the context of early disease progression and risk of cancer development.