Prediction of post-treatment toxicities using the pre-treatment oral microbiome in the context of definitive chemoradiation for oropharyngeal cancer.

Abstract

e18099 Background: Definitive chemoradiation for head and neck cancers is a common treatment with a high toxicity burden. There is an urgent need to identify toxicity biomarkers and develop mitigation or amelioration strategies. Recently, the microbiome was observed to associate with various cancer treatment outcomes, including response and toxicity. In this prospective trial, we aimed to determine whether the oral microbiome at baseline could predict treatment-related toxicities after chemoradiation. Methods: We collected pre-treatment oral swabs from 34 patients on the OROMIC trial (OSU-20074). Patients received definitive chemoradiation (CRT) in 33-35 fractions for oropharyngeal cancer. Patient swabs were collected using an OMR-110 kit containing a nucleic acid stabilizing solution. Samples were processed for total RNAseq, human ribodepleted, and sequenced by Illumina Nextseq 2×150bp to a depth greater than 50 million reads/sample. Metatranscriptomes were aligned to a custom database using Kraken2. We compared differences in oral microbiome based on toxicity grade (low-grade [CTCAE 0-1] vs high-grade [CTCAE 2-3]) to evaluate correlations with acute toxicity and microbes in this cohort. Results: The most common treatment-related toxicities were dysphagia, secretions, oral mucositis, and dermatitis and patients who developed high-grade for these adverse effects had higher abundances of Kineobactrum , Musicloa , and Opituts genera pre-treatment (adjusted p values < 0.0001). Conversely, the presence of Stanieria and Paroceanicella genera seemed protective and related to low grade dermatitis, secretions, and dysgeusia (p value < 0.0001). Several microbes demonstrated more variable relationship to acute toxicity. For example, Ureaplasma demonstrated an association with high-grade mucositis, but also correlated with low-grade dysphagia and secretions (p value < 0.0001). Similarly, Trabulseilla appeared to relate to high-grade dysgeusia and low-grade xerostomia (p value < 0.0001). Conclusions: Overall, a widespread diversity was observed between the relationship between the pre-treatment microbial abundance in the oral cavity and acute toxicity severity in patients undergoing definitive chemoradiation for oropharynx cancer. While the presence of some organisms seems consistently detrimental and the presence of others is generally protective, others still demonstrate variability in influencing acute toxicity outcomes. These findings need to be validated and further analysis of the microbial composition and flux in the oral microbiome is required through longitudinal data which we will present as this study continues to mature. These findings will help in identifying potential biomarkers and therapeutic targets within the oral to improve response to acute toxicity and adverse effects from to chemoradiation for oropharynx.