Development and Validation of an Intraoperative Nomogram to Predict Breast Cancer-Related Lymphedema Based on the Arm Lymphatics Distribution.

Abstract

Preoperatively determining those patients who are at high risk of encountering breast cancer-related lymphedema (BCRL) is still not well understood. This study aimed to develop a simple intraoperative nomogram for BCRL, incorporating a protective factor. Overall, 320 breast cancer patients at Zhongnan Hospital (training set) and 221 patients at Dongfeng General Hospital (external validation cohort) treated between November 2017 and December 2018 were included. Prior to axillary lymph node dissection (ALND), 1mL (2.5 mg) of indocyanine green was administered to the area of the internal bicipital sulcus of the upper limb. The fluorescence image was obtained and the proportion of arm lymph flow above the level of the axillary vein was calculated. Multivariate logistic regression was performed using this proportion together with clinical data. A nomogram was then constructed and assessed for its discrimination and calibration ability and clinical utility in the training and external validation sets. The cumulative incidence of BCRL was 18.7% (60/320), with a median follow-up of 29months (20-34). In the multivariate logistic regression analysis, body mass index, taxane, radiotherapy, and proportion of arm lymph flow above the level of the axillary vein were identified as independent risk factors. In the training and validation cohorts, the calibration curve performed well (p=0.721 and p=0.315, respectively), and the area under the receiver operating characteristic curve values were 0.829 (95% confidence interval [CI] 0.773-0.885) and 0.804 (95% CI 0.732-0.877), respectively. High-risk patients could be identified intraoperatively with this nomogram, and timely intervention could be performed with preservation of the arm lymphatics.