Patient Selection in Human Papillomavirus Related Oropharyngeal Cancer: The Added Value of Prognostic Models in the New TNM 8th Edition Era

Background: With the growing interest in treatment de-intensification trials for human papillomavirus positive (HPV+) oropharyngeal squamous cell carcinoma (OPC), prognostic models have become essential for patient selection. The aim of this paper is to validate and compare the prognostic ability of the TNM 8th edition and previous published risk group classifications of Ang et al. and Rietbergen et al. and to derive a patient selection classification for de-intensification trials. Materials: Patients with HPV+ OPC treated with curative (chemo)radiotherapy between 2004 and 2017 were classified according to the TNM 8th edition, the model of Ang et al. and of Rietbergen et al. HPV status was determined by p16 immunohistochemistry staining. Overall survival was estimated using the Kaplan-Meier method and groups were compared using the log-rank test. Harrell's C-index was used as measure of discriminative performance. Results: A total of 333 OPC were identified of whom 100 were HPV+. The median follow-up was 63.7 months. The 5-year overall survival (5Y-OS) of stage I, II and III were 91.6, 55.2, and 38.0%. There was a significant difference between stage I vs. II and III. The Harrell's C-index for TNM 8th edition stage was 0.67. Including only HPV+ OPC, the Harrell's C-index for the model of Ang and Rietbergen were both 0.62. We combined the main prognostic factors defining the low risk groups in the three models, stage I, low comorbidity and ≤ 10 pack years, into one new low risk group to identify patients who may benefit from de-intensification trials. Intermediate risk was defined as stage I with high comorbidity or >10 pack years, high risk as stage II-III. The 5Y-OS were 100, 85.7, and 51.3%. The Harrell's C-index for the new classification model was 0.67. Conclusion: Although TNM 8th edition provides better OS stratification than the 7th edition, it is not performant enough for patient selection, neither are the models from Ang et al. and Rietbergen et al. Therefore, we propose a patient selection classification for de-intensification trials based on the new TNM classification 8th edition, comorbidity and smoking pack years. In addition, this study emphasizes the importance of patient selection and personalized treatment for HPV+OPC.