Validation of a clinically accessible prognostic biomarker for oropharynx cancer using molecular and spatial data

PROJECT SUMMARY Oropharynx cancer (OPC) incidence has risen annually by 2.7% in the U.S. and is expected to be among the most common cancers over the next few decades. The vast majority of OPC is now reliably linked to human papillomavirus exposure (HPV+). Because HPV+ OPC typically has favourable outcomes, de-intensification trials have been aggressively advanced to minimize treatment-related toxicity and morbidity. However, while most patients respond well and might benefit from treatment de-escalation, there remain 10-38% of HPV+ OPC patients who experience adverse outcomes including recurrence and death. Unfortunately, current de- escalation trials remain indiscriminate in their ability to stratify these patients as there is no marker that is available to classify HPV+ patients further than the dichotomous measure of HPV positivity. The goal of this proposal is to validate a novel prognostic biomarker that captures tumor biology and heterogeneity in >1,000 diverse HPV+ OPC patients from the United States, Europe and South America to guide clinical decision-making and reduce treatment-related toxicity and morbidity. The specific aims are to 1) analytically validate HPVhet scores to establish reliability across diverse patient populations, establish risk groups to maximise prognostic accuracy, and externally validate informative HPVhet scores on outcomes; and 2) translate the HPVhet score as an imaging biomarker through detection from standard hematoxylin and eosin (H&E) stains through optimization of a validated, advanced machine-learning algorithm, OPSCCNet. The OPSCCNet algorithm will combine the molecular HPVhet score with related spatial features for effective interpretation of HPVhet on a spatial scale. Validation of a clinically useful prognostic biomarker that utilizes molecular and spatial data to capture tumor heterogeneity in >1,000 diverse HPV+ OPC patients could lead to a paradigm shift in how these patients are treated and in their quality of life after treatment. Importantly, the inclusion of a range of clinical settings and patient populations ensures scientific rigor and health equity in these findings.