Long-term outcome of re-irradiation for recurrent or second primary head and neck cancer: A multi-institutional study of AIRO-Head and Neck working group

Background To report the long-term outcome of patients undergoing re-irradiation (re-RT) for a recurrent or second primary head and neck cancer (RSPHNCs) in seven Italian tertiary centers, while testing the Multi-Institution Reirradation (MIRI) recursive partitioning analysis (RPA) recently published. Methods We retrospectively analyzed 159 patients. Prognostic factors for overall survival (OS) selected by a random forest model were included in a multivariable Cox analysis. To externally validate MIRI RPA, we estimated the Kaplan-Meier group-stratified OS curves for the whole population. Results Five-year OS was 43.5% (median follow-up: 49.9 months). Nasopharyngeal site, no organ dysfunction, and re-RT volume <36 cm(3) were independent factors for better OS. By applying the MIRI RPA to our cohort, a Harrell C-Index of 0.526 was found indicating poor discriminative ability. Conclusion Our data reinforce the survival benefit of Re-RT for selected patients with RSPHNC. MIRI RPA was not validated in our population.