A New Model for Predicting Nonsentinel Lymph Node Metastasis in Early-Stage Breast Cancer Using MMP15. (18th August 2022)

• Record Type: Journal Article
• Authors: Zeng, Xue; Li, Yubing; Sun, Chaonan; Liu, Zhuang; Zhao, Jiaming; Ma, Xinchi; Zhang, Yanyu; Zhang, Na
• Other Names: Gundamaraju Rohit Academic Editor.
• Is Part Of: Journal of oncology. Volume 2022(2022)
• Publication Date: 2022-08-18
• Subjects: Oncology -- Research -- Periodicals; Tumors -- Periodicals; Neoplasms; Oncology -- Research; Tumors; Periodicals; Periodicals; 616.994
• DOI: 10.1155/2022/8675705
• Languages: English
• ISSNs: 1687-8450

Abstract

Abstract : Background . In early-stage breast cancer (BC) patients, 40–70% of lymph node metastases are limited to the sentinel lymph nodes (SLNs). Patients at low risk for nonsentinel lymph node (NSLN) metastasis should be exempt from axillary lymph node dissection (ALND) or regional lymph node radiotherapy (RNI). Methods . The present study included 237 female early-stage BC patients with positive SLNs who received ALND. Based on the clinicopathological factors of the 158 patients in the training cohort, multivariate analysis was used to determine the independent risk factors for NSLN metastasis, which were used to establish the NSLN metastasis prediction model. The calibration and discrimination of this model were tested with the training and validation cohorts and compared to the Memorial Sloan Kettering Cancer Center (MSKCC) model. Results . Tumor size, neural invasion, lymphovascular invasion, expression of matrix metalloproteinase 15 (MMP15) in the cytoplasm, and the number of positive SLNs were statistically significant by multivariate analysis (P < 0.05 ), which were used to establish the new model. The MSKCC model was verified by the training cohort, and the area under the receiver-operating characteristic (ROC) curve was 0.733 (95% CI: 0.650–0.816), which was less than that of the new model (0.824; 95% CI: 0.760–0.889). The area under the ROC curve in the validation cohort for the new model was 0.773 (95% CI: 0.669–0.877), and the calibration performed well. The false-negative rates were 3.2%, 6.5%, and 14.5% for the predicted probability cut-offs of 50%, 60%, and 70%, respectively. Conclusions . The new model included five variables: tumor size, neural invasion, lymphovascular invasion, cytoplasmic MMP15 expression, and the number of positive SLNs. The model with a cut-off of 60% could accurately identify low-risk patients with NSLN metastasis.