The incidence of oropharyngeal squamous cell carcinoma (OPSCC) is rising in younger patients because of its association with human papilloma virus (HPV) oncogenesis. Compared with carcinogen-associated OPSCC, cancer related to HPV infection behaves differently on a molecular level and has a better prognosis. Although this distinction is clinically relevant, it cannot be made by standard diagnostic imaging techniques. Instead, HPV status of a tumor is currently established invasively by analyzing tissue specimens for over-expression of p16 or the presence of HPV DNA. In a retrospectively analyzed cohort (target accrual of 30 patients), we use magnetic resonance imaging texture analysis together with machine learning methodologies in order to noninvasively classify whether OPSCC is HPV-related.

Source: ASU-Mayo seed