An exploratory study of whether axillary lymph node dissection can be avoided in breast cancer patients with positive lymph nodes
Fei Ren,Chenxuan Yang,Jiaxiang Liu,Kexin Feng,Qingyao Shang,Xiyu Kang,Ruixuan Zhang,Li Li,Shuangtao Zhao,Xin Wang,Xiang Wang
DOI: https://doi.org/10.21037/tcr-23-1639
2024-02-29
Abstract:Background: Breast cancer patients with positive axillary lymph nodes usually require axillary lymph node dissection (ALND), with many postoperative complications, such as lymphedema. For these patients, whether sentinel lymph node biopsy (SLNB) can replace ALND has been a research hotspot in the field of breast cancer. This study developed two risk stratification models for predicting the clinical outcomes of breast cancer patients with positive axillary lymph nodes receiving SLNB alone or ALND to determine which patients could potentially avoid ALND. Methods: A total of 21,942 breast cancer patients, including a training set (n=15,362) and a testing set (n=6,580), were enrolled in this study from Surveillance, Epidemiology, and End Results (SEER) between 2000 and 2017. The risk factors associated with breast cancer-specific survival (BCSS) and overall survival (OS) were evaluated using multivariate Cox regression analysis and then integrated into nomograms and risk stratification models examined by receiver operating characteristic (ROC) curves and calibration curves. The survival discrepancies were compared between the SLNB and ALND subgroups with different risk scores with Kaplan-Meier plots. Results: In multivariate Cox regression analyses, grade, marital status, T stage, radiotherapy and lymph node metastasis (GMTRL) were independent risk factors in breast cancer patients with both OS and BCSS status in the ALND cohort from the training set. Nomograms have been developed based on these factors to predict 3- and 5-year OS and BCSS in patients with ALND. Calibration curves and ROC curves in both the training and testing sets confirmed the excellent overall predictive performance of the nomograms. Furthermore, we developed two risk stratification models based on OS and BCSS status, revealing that patients with low GMTRL scores might avoid ALND in both OS and BCSS status [OS: hazard ratio (HR) =0.929, 95% confidence interval (CI): 0.841-1.027, P=0.150; BCSS: HR =0.953, 95% CI: 0.831-1.094, P=0.495], but patients with moderate (OS: HR =0.756, 95% CI: 0.666-0.859, P<0.001; BCSS: HR =0.643, 95% CI: 0.537-0.768, P<0.001) and high GMTRL scores could not (OS: HR =0.719, 95% CI: 0.549-0.940, P=0.014; BCSS: HR =0.731, 95% CI: 0.549-0.974, P=0.031). Conclusions: Breast cancer patients with positive nodes could be treated with SLNB alone rather than ALND without affecting prognosis based on GMTRL scores. Patients with high or moderate GMTRL scores benefited greatly from ALND, but not for patients with low GMTRL scores. This study may assist clinicians in tailoring treatments.