PURPOSE: This study aimed to investigate the ability of thyroid volumes after intensity-modulated radiotherapy (IMRT) to predict the incidence of radiation-induced hypothyroidism/primary hypothyroidism (HT)/(PHT) in nasopharyngeal carcinoma (NPC) patients. METHODS AND MATERIALS: 404 NPC patients were retrospectively enrolled from January 2015 to January 2019. Thyroid volumes were calculated based on magnetic resonance imaging. Volume of thyroid after IMRT within 3 months was defined as Vt-after. Least absolute shrinkage and selection operator (LASSO) regression, random forest analysis, and multivariate Cox proportional hazards were performed to identify optimal predictors of PHT. Then, these predictors were employed to construct the risk stratification using the recursive partitioning analysis (RPA). RESULTS: A smaller Vt-after was significantly associated with a higher risk of HT/PHT (hazard ratio (HR): 2.046, p <
0.001
HR: 3.214, p <
0.001
respectively). Vt-after and thyroid volumes at 1, 2, and 3 years after IMRT demonstrated superior predictive ability compared to the thyroid volume before IMRT for predicting PHT within the subsequent 2-year period. The combination of Vt-after and thyroid volume spared from 45 Gy (VS45), both of which wereindependent factors for predicting PHT, was incorporated into the RPA risk stratification called PRA-VS45. The PRA-VS45 emerged as a highly discriminative tool, with an area under the curve (AUC) of 0.716 on predicting PHT at 5 years. CONCLUSIONS: This study identified thyroid volumes after IMRT as valuable parameters to predict radiation-related PHT in NPC patients. The PRA-VS45 has the potential to be clinical applications for predicting PHT.