Abstract
BACKGROUND: Few population-based studies on nasopharyngeal carcinoma (NPC) prognosis exist. Such real-world data with complete follow-up are needed to enable accurate estimation of survival probabilities and monitor progress in early detection and treatment. METHODS: In a population-based study of 2529 patients with incident NPC in southern China diagnosed in 2010-2013, we developed a passive-active-passive follow-up strategy involving data linkage and direct contact to track patients' vital status, cause of death, and migration through December 31, 2018. We calculated 5-year survival probabilities by the Kaplan-Meier method, and estimated avoidable deaths comparing hypothetical scenarios and real-world observations. RESULTS: Early-stage (I-II) patients accounted for 11·5%, and 21·1% were treated in medical-university-affiliated or province-level hospitals. With a mean follow-up time of 5·5 years after diagnosis (98.3% complete), 5-year overall and NPC-specific survival probabilities were 70·1% and 74·5%. The 5-year overall survival probabilities for stage I, II, III, IVa, IVb, and IVc were 91·3%, 87·8%, 79·8%, 63·9%, 57·7%, and 34·4%, respectively. Survival of non-metastatic (I-IVb) population-based patients diagnosed in 2010-2013 was comparable to that of a hospital-based southern Chinese NPC cohort diagnosed in 1997-2007. Based on observed data, if all patients were diagnosed at stages I-II, then total avoidable deaths within 5 years after diagnosis would be 156 per 1000 patients. CONCLUSIONS: Real-world population-based NPC survival lags behind that of large hospital-based cohorts, but earlier diagnosis has the potential to substantially reduce mortality from NPC. Our results highlight the importance of improving the accessibility of health resources and promoting early detection of NPC.