构建乳腺癌保乳术后放疗致急性放射性皮炎（ARD）发生风险预测模型并验证其效果。

选择宝鸡高新医院乳腺科于2020年1月至2024年10月接受乳腺癌保乳术后放疗患者123例作为研究对象，按照7:3比例分别纳入训练集（n=86）、验证集（n=37）。根据训练集患者是否发生ARD分为ARD组与无ARD组。多因素Logistic回归分析ARD发生危险因素，构建预测模型，采用受试者工作特征曲线（ROC）验证该模型在临床应用中的价值。

接受乳腺癌保乳术后放疗患者86例，发生ARD患者67例，发生率为77.91%。经单因素分析，ARD组和无ARD组体重指数、吸烟史、饮酒史、合并高血压、合并冠心病、病理类型、肿瘤-淋巴结-转移（TNM）分期、三阴性乳腺癌、人表皮生长因子受体-2（HER2）、雌激素受体（ER）和孕激素受体（PR）比较差异无统计学意义（P>0.05）；ARD组年龄显著低于无ARD组[（42.31±7.67）比（56.27±9.98），P<0.001]；ARD组合并糖尿病占比显著高于无ARD组[（34.33%比10.53%），P<0.05]；ARD组营养不足占比显著高于无ARD组[（43.28%比10.53%），P<0.01]；ARD组放疗剂量显著高于无ARD组[（57.82±3.24）比（48.97±4.35），P<0.001]。多因素Logistic回归分析结果显示，年龄、营养状况、合并糖尿病、放疗剂量，为影响放疗致ARD发生的独立危险因素（P<0.05）。构建放疗致ARD预测模型：P=-2.22-2.35×年龄+1.75×营养状况+2.68×合并糖尿病+1.99×放疗剂量。采用ROC曲线验证结果显示，曲线下面积为0.876（P=0.000，95%CI：0.828~0.924），敏感度89.90%，特异度71.00%。

年龄、营养状况、合并糖尿病和放疗剂量是影响乳腺癌保乳术后放疗致ARD发生的重要因素，根据上述因素拟定相应的风险预测模型，且该模型预测效果良好，值得进一步深入分析。

To explore the establishment and validation of a risk prediction model for acute radiation dermatitis (ARD) after breast-conserving surgery for breast cancer.

A total of 123 patients who received radiotherapy after breast-conserving surgery for breast cancer in our hospital from January 2020 to October 24 were selected as the study subjects, and were included in the training set (n=86) and the validation set (n=37) according to the ratio of 7:3. According to whether patients in the training set have developed ARD, they were divided into the ARD group and the non ARD group. The occurrence of ARD in patients was recorded, and multivariate logistic regression analysis was conducted to identify risk factors for ARD, to construct a predictive model, and to validate the value of the model in clinical applications using ROC curves.

Of the 86 patients who received radiotherapy after breast-conserving surgery for breast cancer, 67 patients developed ARD (77.91%). By univariate analysis, there was no significant difference between ARD group and non ARD group in body mass index, smoking history, drinking history, hypertension, coronary heart disease, pathological type, TNM stage, triple negative breast cancer, HER2, ER and PR (P>0.05). The age in ARD group was significantly lower than that in non ARD group (42.31±7.67 vs 56.27±9.98, P<0.001); The ARD group combined with diabetes was significantly higher non ARD (34.33% vs 10.53%, P<0.05); The ARD group proportion of malnutrition was significantly higher non ARD group (43.28% vs 10.53%, P<0.01); The ARD group radiation dose was significantly higher non ARD group (57.82±3.24 vs 48.97±4.35, P<0.001). Multivariate logistic regression analysis showed that age, nutritional status, combined diabetes, radiotherapy dose entered the regression model (P<0.05), which were independent risk factors for the occurrence of radiation-induced ARD. A prediction model of radiation-induced ARD was established: P=-2.22-2.35×age+1.75×nutritional status+2.68×combined diabetes+1.99×radiation dose. The ROC curve was used to validate the predictive value of the above model for ARD occurrence. The results showed that the area under the curve was 0.876 (P=0.000, 95%CI0.828-0.924), with a sensitivity of 89.90% and a specificity of 71.00%.

Age, nutritional status, combined diabetes, radiation dose are important factors that affect the occurrence of ARD caused by radiotherapy after breast-conserving surgery for breast cancer. According to the above factors, the corresponding risk prediction model is developed, and the prediction effect of the model is good, which is worth further analysis.