乳腺癌放疗致慢性放射性皮肤损伤的危险因素分析及列线图模型构建OA
Risk factor analysis and nomogram model construction of chronic radiation injury of skin caused by breast cancer radiotherapy
目的 分析影响乳腺癌放疗致慢性放射性皮肤损伤(CRIS)发生的危险因素并构建列线图预测模型.方法 回顾性选取2017年4月至2022年4月在唐山市人民医院放化科接受放疗的乳腺癌根治术后患者860例,随机分为训练集602例和验证集258例.根据CRIS发生情况将训练集分为CRIS组104例和非CRIS组498例,采用logistic回归分析影响CRIS发生的危险因素并构建列线图预测模型.通过评价校准曲线和临床决策曲线完成模型内部验证.通过评价模型预测训练集和验证集CRIS发生的效能完成外部验证.结果 CRIS组患者年龄、体重指数、有化疗史比例、肿瘤直径均大于非CRIS组(均P＜0.05),血清白蛋白水平低于非CRIS组(P＜0.05).高龄(OR=4.495,95％CI:1.146～6.934)、高体重指数(OR=3.710,95％CI:1.048～5.395)、低血清白蛋白水平(OR=7.471,95％CI:1.935～13.275)、有化疗史(OR=6.025,95％CI:1.845～9.284)、大肿瘤直径(OR=8.306,95％CI:2.004～15.934)均是乳腺癌放疗致CRIS发生的独立危险因素(均P＜0.05).内部验证结果显示列线图预测模型区分能力和一致性均较为理想,可提供高于其他独立预测因子的临床净收益.外部验证结果显示模型预测训练集和验证集CRIS发生的效能基本一致.结论 高龄、高体重指数、低血清白蛋白水平、有化疗史、大肿瘤直径均是乳腺癌放疗致CRIS发生的独立危险因素,基于此构建的CRIS发生风险列线图预测模型对该病的防治有一定的应用价值.
Objective To analyze the risk factors for developing chronic radiation-induced skin injury(CRIS)caused by radiotherapy for breast cancer and to construct nomogram prediction model.Methods A retrospective study was conducted on 860 patients who underwent radical mastectomy and received radiotherapy in the Department of Radiotherapy and Chemotherapy of Tangshan People's Hospital from April 2017 to April 2022.These patients were randomly divided into a training set of 602 cases and a validation set of 258 cases.The training set was further divided into a CRIS group of 104 cases and a non-CRIS group of 498 cases based on the occurrence of CRIS.Logistic regression analysis was used to identify the risk factors for developing CRIS,and a nomogram prediction model was constructed.Internal validation of the model was completed by evaluating the calibration curve and clinical decision curve.External validation was completed by assessing the model's ability to predict the occurrence of CRIS in both the training and validation sets.Results The patients in the CRIS group had older age,higher body mass index and chemotherapy history proportion,and larger tumor diameter than those in the non-CRIS group(all P＜0.05),while serum albumin level was lower(P＜0.05).Advanced age(OR=4.495,95％CI:1.146-6.934),high body mass index(OR=3.710,95％CI:1.048-5.395),low serum albumin level(OR=7.471,95％CI:1.935-13.275),history of chemotherapy(OR=6.025,95％CI:1.845-9.284),and relatively large tumor diameter(OR=8.306,95％CI:2.004-15.934)were all the independent risk factors for the occurrence of CRIS during breast cancer radiotherapy(all P＜0.05).Internal validation results showed that the nomogram prediction model had satisfactory discrimination and consistency,providing higher clinical net benefit than other independent predictors.External validation results indicated that the model's performance in predicting the occurrence of CRIS in the training and validation sets was essentially consistent.Conclusion Advanced age,high body mass index,low serum albumin level,history of chemotherapy,and large tumor diameter are all the independent risk factors for the occurrence of CRIS caused by radiotherapy in breast cancer.The risk nomogram prediction model for CRIS occurrence constructed based on these factors has certain application value for the prevention and treatment of this disease.
崔文韬;黄昱霖;迟子锋;苗明昌;杨海芳;田龙;陈璇
063000 唐山市人民医院放化科063000 唐山市人民医院放化科河北医科大学第四医院放疗科河北医科大学第四医院放疗科063000 唐山市人民医院放化科063000 唐山市人民医院放化科063000 唐山市人民医院放化科
乳腺癌慢性放射性皮肤损伤影响因素列线图模型预测
Breast cancerChronic radiation injury of skinInfluencing factorNomogram modelPrediction
《浙江医学》 2026 (6)
599-603,5
河北省医学科学研究课题计划资助(20261069)
评论