Abstract
BACKGROUND: Post-radiation nasopharyngeal necrosis (PRNN) is a severe complication following re-irradiation in patients with recurrent nasopharyngeal carcinoma (NPC). This study aimed to explore the association between nasopharyngeal microbiota and PRNN in patients with recurrent NPC undergoing re-irradiation and to evaluate the predictive value of the microbiota for PRNN. METHODS: This retrospective study collected data from 113 patients with recurrent NPC who underwent re-irradiation at the Sun Yat-sen University Cancer Center (SYSUCC) between January 2020 and November 2022. Patients were divided into necrosis and non-necrosis groups based on the development of necrosis after re-irradiation. 5R 16S rRNA sequencing of nasopharyngeal biopsy tissues conducted before re-irradiation was used to assess microbiota composition, diversity, and functional predictions. Clinical features and selected microbial markers were used in a random forest model to predict the occurrence of PRNN. RESULTS: Of the 113 patients with recurrent NPC who underwent re-irradiation, 60 developed PRNN, while 53 did not. Proteobacteria and Firmicutes were the dominant phyla in the nasopharyngeal microbiota of all the patients with recurrent NPC. The necrosis group exhibited significantly higher alpha diversity and distinct beta diversity than the non-necrosis group did. A predictive model that combined clinical features (gross tumor volume [GTV]) with microbiome characteristics achieved an AUC of 87.9% in the training set and 86.9% in the test set, demonstrating robust predictive performance. CONCLUSIONS: Nasopharyngeal microbial diversity prior to re-irradiation was significantly higher in the necrosis group. Our predictive model, integrating clinical and microbial features, demonstrated strong performance in predicting PRNN and offers a promising tool for early intervention and prevention strategies.