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.