验证Bevilacqua乳腺癌术后淋巴水肿风险预测模型的临床适用性及可行性。

回顾性分析2010年1月至2015年12月203例乳腺癌患者临床资料，临床数据分析使用统计学软件SPSS 24.0。Cox回归模型分析乳腺癌患者术后发生上肢淋巴水肿的危险因素，以P<0.05为有统计学意义；绘制ROC曲线，以曲线下面积检验模型预测效果；应用Hosmere-Lemeshow检验评估预测值与实际值的校准程度，以P>0.05为预测模型校准能力较好，预测与实际没有区别。

所有患者随访共计62～86个月，中位随访时间70个月。术后5年内共发生上肢淋巴水肿患者45例(22.2%)。Cox回归模型分析结果显示，高身体质量指数(BMI)、接受过新辅助化疗、全腋窝淋巴结清扫、接受过放疗是上肢淋巴水肿的独立危险因素。Becilacqua上肢淋巴水肿风险预测模型ROC曲线分析结果显示，模型AUC值为0.711，95%CI(0.651～0.760)，有较好的的预测效果。Hosmer-Lemeshow检验结果显示，风险预测模型预测风险与实际无明显差异(P＝0.262)，校准能力较好，与实际差别不大。

Bevilacqua术后6个月淋巴水肿风险预测模型的准确性及适用性较高，可用于临床对乳腺癌保乳术后淋巴水肿的预测，可为预防淋巴水肿的发生制定干预决策提供参考。

To investigate the clinical applicability and feasibility of the Bevilacqua risk prediction model for postoperative lymphedema in patients after breast conserving surgery for breast cancer.

From January 2010 to December 2015, clinical data of 203 breast cancer patients were analyzed retrospectively. Statistical analysis were performed by using SPSS24.0 software. Cox regression model was used to analyze the risk factors of postoperative lymphedema of patients’ upper extremity after breast conserving surgery. A P value of < 0.05 was considered as statistically significant difference. ROC curve was used to evaluate the prediction effect of the Bevilacqua model, Hosmer lemeshow test was used to verify the calibration ability of the Bevilacqua model. The calibration ability of the predictive model was considered as no difference between forecast and reality with a P value of >0.05.

The follow-up ranged from 62 to 86 months (median=70 months) . 45 patients (22.2%) had lymphedema of upper extremity within 5 years after operation. Cox regression model analysis showed that high BMI, neoadjuvant chemotherapy, total axillary lymph node dissection, and radiotherapy were independent risk factors for lymphedema of upper extremity. The ROC curve analysis of Bevilacqua model showed that the AUC value of the model was 0.711, 95% CI (0.651-0.760), which had a good prediction effect. Hosmer lemeshow test showed that there was no significant difference (P=0.262) between the predicted risk and the actual risk, with a good calibration ability.

The accuracy and applicability of the Bevilacqua model for predicting the risk of lymphedema in 6 months after operation are good, which could be used to predict the occurrence of lymphedema in breast cancer patients and could help to make intervention decisions for the prevention of lymphedema.