头颈部肿瘤放疗患者急性放射性口腔黏膜炎风险列线图模型的构建OA

Objective To construct a risk prediction model for acute radiation-induced oral mu-cositis(RIOM)in patients undergoing radiotherapy for head and neck cancer.Methods A retro-spective cohort study was conducted,enrolling 332 patients with head and neck cancer who received radiotherapy between March 2022 and December 2025.Patients were randomly divided into training set(n=232)and validation set(n=100)at a 7∶3 ratio.Demographic data,disease-related data,treatment-related data,oral status,and laboratory parameters were collected,with the occurrence of acute RIOM during radiotherapy as the outcome variable.Least absolute shrinkage and selection operator(LASSO)regression was used to screen predictive variables.Multivariable logistic regres-sion was applied to identify independent influencing factors and to construct a nomogram model.The discriminative ability of the model was evaluated using the receiver operating characteristic(ROC)curve;calibration was assessed using the Hosmer-Lemeshow test combined with 1 000 bootstrap re-sampling;clinical net benefit was evaluated using decision curve analysis.Results Among the 332 patients,124 developed acute RIOM,with an incidence rate of 37.35％.Compared with other pri-mary sites,patients with oropharyngeal and hypopharyngeal primary tumors had a higher risk of a-cute RIOM.Diabetes,planned concurrent chemotherapy,planned total prescription dose ≥60 Gy,pre-radiotherapy oral infection,and elevated neutrophil-to-lymphocyte ratio(NLR)were associated with an increased risk of acute RIOM,while an elevated prognostic nutritional index(PNI)was as-sociated with a decreased risk(P＜0.05).In the training set,the area under the curve(AUC)of the baseline clinical model was 0.711(95％CI:0.640-0.782).The combined model improved the AUC to 0.752(95％CI:0.686-0.818,DeLong P=0.032),with a continuous NRI of 0.318(95％CI:0.086-0.550,P=0.008)and an IDI of 0.052(95％CI:0.017-0.087,P=0.004).In the validation set,the AUC of the baseline clinical model was 0.748(95％CI:0.646-0.849),and the combined model increased the AUC to 0.794(95％CI:0.698-0.891,DeLong P=0.041),with a continuous NRI of 0.294(95％CI:0.041-0.547,P=0.023)and an IDI of 0.061(95％CI:0.012-0.110,P=0.015).In the training set,the combined model a-chieved an AUC of 0.752(95％CI:0.686-0.818),with a sensitivity of 74.71％(95％CI:64.25％-83.42％)and a specificity of 68.97％(95％CI:60.76％-76.38％).In the validation set,the AUC was 0.794(95％CI:0.698-0.891),with a sensitivity of 78.38％(95％CI:61.79％-90.17％)and a specificity of 76.19％(95％CI:63.79％-86.02％).The Hosmer-Lemeshow test P-values for the training and validation sets were 0.908 and 0.193,respectively,and the cali-bration curves demonstrated good agreement between predicted and observed probabilities.Decision curve analysis showed that the model provided a clinical net benefit across threshold probability ran-ges of 0.05 to 0.90 in the training set and 0.08 to 0.88 in the validation set.Conclusion The nomogram constructed in this study,based on routine clinical indicators,demonstrates moderate dis-criminative ability,good calibration,and potential clinical utility.It may assist in risk stratification for acute RIOM prior to the first course of radiotherapy;however,multicenter external validation is still warranted.