Gut microbiome health company Viome has repurposed its mRNA analysis and machine learning technology to develop a saliva test to screen for head and neck cancers. Viome’s diagnostic test, which received Breakthrough Device designation from the US Food and Drug Administration (FDA) in May, is based on results from a study combining microbial RNA (the metatranscriptome) in saliva using next-generation sequencing with machine learning to identify biomarkers linked to oral cancers. The algorithm spotted early signs of oral cancer with a 83% sensitivity and a specificity of nearly 98%. The risk of oral cancer increases in people over 50 or with a history of smoking or excessive alcohol consumption, and multiple studies in the past decade have revealed associations between changes in the oral microbiome and cancer. To develop a diagnostic tool, scientists at the Viome Research Institute, the company’s research arm, used microbiome transcriptome profiling to study the saliva microbiome from 71 patients with diagnosed oral premalignant conditions or cancer and 171 healthy controls. mRNA sequencing provides an indication of the genes that are expressed within a community, focusing on bacterial function and activity (rather than metagenomics as assessed by 16S ribosomal RNA or DNA). One finding is that high amounts of transcripts from periodontal bacteria Fusobacterium, Prevotella and Prophyromonas in saliva samples and another eight microbial species involved in biofilm formation are associated with tumorigenesis. Each year, 350,000–400,000 patients worldwide are newly diagnosed with oral cancers. With no early detection tools, diagnosis is mostly at late stages, and as a result, the five-year survival has remained at 40% for four decades. But before the microbial metatranscriptome can serve as an accurate, non-invasive diagnostic to prevent oral cancers, prospective studies in well-characterized patient groups will be necessary. Viome has yet to receive FDA approval for the oral cancer test.