女性乳腺癌患者术后上肢水肿的危险因素分析及其列线图预测模型的构建

Analysis of risk factors for upper limb edema in female breast cancer patients after operation and construction of a nomogram prediction model

目的:分析女性乳腺癌术后上肢水肿(LE)的危险因素,构建女性乳腺癌术后上肢水肿的列线图预测模型。方法:回顾性分析2013年1月—至2020年12月行手术治疗的女性乳腺癌患者的相关临床资料,采用χ^(2)检验或t检验初步分析与LE相关的因素,再进一步采用Logistic回归分析筛选出导致LE的独立危险因素,应用R软件构建列线图预测模型,通过绘制校准曲线评价模型预测结局概率与实际概率的一致程度,以C-index表示。绘制受试者工作特征(ROC)曲线,以曲线下面积(AUC)评估列线图预测模型对LE的预测能力。结果:489例女性乳腺癌患者纳入研究,术后LE发生率为16.2%。单因素分析结果显示,术后LE与体重指数(BMI)(t=-3.716,P<0.001)、糖尿病(χ^(2)=36.784,P<0.001)、高血压病(χ^(2)=4.525,P=0.033),肿瘤TNM分期(χ^(2)=3.955,P=0.047)、术后功能锻炼依从性(χ^(2)=7.011,P=0.030),新辅助化疗(χ^(2)=4.025,P=0.045),腋窝淋巴结清除(χ^(2)=5.197,P=0.023),放疗(χ^(2)=8.068,P=0.005)有关。Logistic多因素分析结果表明,BMI(OR=1.182,95%CI:1.083~1.291,P<0.001),糖尿病(OR=5.126,95%CI:2.924~8.989,P<0.001),术后功能锻炼依从性(OR=0.242,95%CI:0.083~0.701,P=0.009),新辅助化疗(OR=2.181,95%CI:1.101~4.319,P=0.025),行腋窝淋巴结清除(OR=1.996,95%CI:1.156~3.447,P=0.013),放疗(OR=1.986,95%CI:1.109~3.556,P=0.021)是LE发生的独立危险因素,基于此构建出列线图预测模型,模型总分220分,通过校准曲线可见该列线图模型的预测概率与实际概率有较好的一致性,C-index值为0.775(95%CI:0.716~0.833),所绘制的ROC曲线显示其曲线下面积为0.775,提示该列线图预测模型对LE有较强的预测能力。结论:BMI、糖尿病、术后功能锻炼依从性、新辅助化疗、腋窝淋巴结清除、放疗是LE发生的独立危险因素,基于这6项独立危险因素构建的列线图模型,可以直观地预测术后出现LE的风险。

Objective: To analyze the risk factors of upper arm Lymphedema(LE) after breast cancer surgery, and to construct a nomogram prediction model for upper arm Lymphedema after breast cancer surgery. Methods: The clinical data of female breast cancer patients who underwent surgical treatment from January 2013 to December 2020 in the Department of Thyroid and breast surgery of our hospital were retrospectively analyzed. The chi-square test or T test was used to analyze the factors related to LE, and then the independent risk factors leading to LE were screened out by Logistic regression analysis. Then, R software was used to construct the prediction model of the nomogram, and the consistency between the probability of the predicted outcome and the actual observation probability was evaluated by the calibration curve, which was represented by the consistency index(C-index). The area under the Receiver Operating Characteristic(ROC) curve(AUC) was used to evaluate the ability of the lipopograph prediction model to predict LE. Results: 489 female breast cancer patients were included in the study, and the incidence of upper arm Lymphedema was 16.2%.Univariate analysis showed that postoperative upper arm Lymphedema(LE) was related to BMI(t=-3.716, P<0.001), diabetes(χ~2=36.784, P<0.001), hypertension(χ~2=4.525, P=0.033), tumor TNM stage(χ~2=3.955, P=0.047), Postoperative compliance with functional exercise(χ~2=7.011, P=0.030), neoadjuvant chemotherapy(χ~2=4.025, P=4.025), axillary lymph node dissection(χ~2=5.197, P=0.023), radiotherapy(χ~2=8.068, P=0.005)(P<0.05). Logistic multivariate analysis showed that BMI(OR: 1.182, 95%CI: 1.083-1.291, P=0.000), diabetes mellitus(OR=5.126, 95% CI: 2.924-8.989, P<0.001), postoperative compliance to functional exercise(OR=0.242, 95% CI: 0.083-0.701, P=0.009), neoadjuvant chemotherapy(OR=2.181, 95% CI: 1.101-4.319, P=0.025), axillary lymph node dissection(OR=1.996,95% CI: 1.156~3.447, P=0.013), radiotherapy(OR=1.986, 95% CI: 1.109-3.556, P=0.021) were an independent risk factor for postoperative upper arm Lymphedema in female breast cancer patients. The total score of the model was 220 points, and the calibration curve analysis showed that the predicted probability of the model was consistent with the actual probability, with the C-index value was 0.775(95% CI: 0.716-0.833). The ROC curve drawn showed that the area under the curve(AUC) was 0.775,suggesting that the model had a good prediction ability for LE. Conclusions: BMI, diabetes, postoperative functional exercise compliance, neoadjuvant chemotherapy, axillary lymph node dissection,and radiotherapy are independent risk factors for postoperative upper arm Lymphedema in female breast cancer patients. The nomogram prediction model constructed based on these 6 independent risk factors can intuitively predict the risk of postoperative upper arm Lymphedema(LE).