BACKGROUND: It has been determined that axillary lymph node dissection after the detection of limited axillary lymph node metastasis does not improve the prognosis of patients with breast cancer. Thus, a need exists for less-invasive axillary surgery. However, it remains unclear whether a predictive model based on preoperative data would be sufficient to accurately predict the probability of pN2-N3 (> 3 lymph node metastases). We sought to develop an easy-to-use scoring system to distinguish between pN0-N1 (0-3 lymph node metastases) and pN2-N3 using only preoperative data and validate its predictive performance. PATIENTS AND METHODS: We retrospectively identified 2687 patients diagnosed with cT1-3cN0-N1 who had undergone surgery in our hospital from 2013 to 2019. We evaluated the risk factors associated with pN2-N3 by logistic regression analysis and developed a scoring system. Predictive performance was assessed by calculating the receiver operating characteristic area under the curve (AUC) and was validated using K-fold cross-validation. RESULTS: We identified 1987 patients with stage pN0, 522 with pN1, and 178 with pN2-N3. Multivariate analysis revealed tumor size, number of suspicious lymph nodes on axillary ultrasound examination, histologic type, histologic grade, and receptor status were significant risk factors for pN2-N3. The AUC value was 0.87, and the mean AUC of the 10-fold cross-validation was 0.88. When the cutoff score was set at 6, the negative predictive value for excluding patients with pN2-N3 was 98.4%. CONCLUSION: Our easy-to-use scoring system could be useful to preoperatively identify patients at lower risk of pN2-N3 and avoid unnecessary axillary lymph node dissection.
BACKGROUND: It has been determined that axillary lymph node dissection after the detection of limited axillary lymph node metastasis does not improve the prognosis of patients with breast cancer. Thus, a need exists for less-invasive axillary surgery. However, it remains unclear whether a predictive model based on preoperative data would be sufficient to accurately predict the probability of pN2-N3 (> 3 lymph node metastases). We sought to develop an easy-to-use scoring system to distinguish between pN0-N1 (0-3 lymph node metastases) and pN2-N3 using only preoperative data and validate its predictive performance. PATIENTS AND METHODS: We retrospectively identified 2687 patients diagnosed with cT1-3cN0-N1 who had undergone surgery in our hospital from 2013 to 2019. We evaluated the risk factors associated with pN2-N3 by logistic regression analysis and developed a scoring system. Predictive performance was assessed by calculating the receiver operating characteristic area under the curve (AUC) and was validated using K-fold cross-validation. RESULTS: We identified 1987 patients with stage pN0, 522 with pN1, and 178 with pN2-N3. Multivariate analysis revealed tumor size, number of suspicious lymph nodes on axillary ultrasound examination, histologic type, histologic grade, and receptor status were significant risk factors for pN2-N3. The AUC value was 0.87, and the mean AUC of the 10-fold cross-validation was 0.88. When the cutoff score was set at 6, the negative predictive value for excluding patients with pN2-N3 was 98.4%. CONCLUSION: Our easy-to-use scoring system could be useful to preoperatively identify patients at lower risk of pN2-N3 and avoid unnecessary axillary lymph node dissection.