Evaluation of the eighth TNM classification on p16-positive oropharyngeal squamous cell carcinomas in the Netherlands and the importance of additional HPV DNA testing

Background: Oropharyngeal squamous cell carcinomas (OPSCCs) are traditionally caused by smoking and excessive alcohol consumption. However, in the last decades high-risk human papillomavirus (HPV) infections play an increasingly important role in tumorigenesis. HPV-driven OPSCCs are known to have a more favorable prognosis, which has led to important and marked changes in the recently released TNM-8. In this 8th edition, OPSCCs are divided based on p16 immunostaining, with p16 overexpression as surrogate marker for the presence of HPV. The aims of this study are to evaluate TNM-8 on a Dutch consecutive cohort of patients with p16-positive OPSCC and to determine the relevance of additional HPV DNA testing. Patients and methods: All OPSCC patients without distant metastases at diagnosis and treated with curative intent at VU University Medical Center (2000-2015) and Erasmus Medical Center (2000-2006) were included (N = 1204). HPV status was determined by p16 immunostaining followed by HPV DNA PCR on the p16-immunopositive cases. We compared TNM-7 and TNM-8 using the Harrell's C index. Results: In total, 388 of 1204 (32.2%) patients were p16-immunopositive. In these patients, TNM-8 had a markedly better predictive prognostic power than TNM-7 (Harrell's C index 0.63 versus 0.53). Of the 388 p16-positive OPSCCs, 48 tumors (12.4%) were HPV DNA-negative. This subgroup had distinct demographic, clinical and morphologic characteristics and showed a significantly worse five-year overall survival compared with the HPV DNA-positive tumors (P < 0.001). Conclusions: TNM-8 has a better predictive prognostic power than TNM-7 in patients with p16-positive OPSCC. However, within p16-positive OPSCCs, there is an HPV DNA-negative subgroup with distinct features and a worse overall survival, indicating the importance to perform additional HPV DNA testing when predicting prognosis and particularly for selecting patients for de-intensified treatment regimens.