A model to predict pathologic complete response of axillary lymph nodes to neoadjuvant chemo(immuno)therapy in patients with clinically node-positive breast cancer.

BACKGROUND: Between 20% and 42% of patients with clinically node-positive breast cancer achieve a pathologic complete response (pCR) of axillary lymph nodes after neoadjuvant chemotherapy or immunotherapy, or both, (chemo[immuno]therapy). Hypothetically, axillary lymph node dissection (ALND) may be safely omitted in these patients. This study aimed to develop a model for predicting axillary pCR in these patients.

PATIENTS AND METHODS: We retrospectively identified patients with clinically node-positive breast cancer who were treated with neoadjuvant chemo(immuno)therapy and ALND between 2005 and 2012 in 5 hospitals. Patient and tumor characteristics, neoadjuvant chemo(immuno)therapy regimens, and pathology reports were extracted. Binary logistic regression analysis was used to predict axillary pCR with the following variables: age, tumor stage and type, hormone receptor and human epidermal growth factor receptor 2 (HER2) status, and administration of taxane and trastuzumab. The model was internally validated by bootstrap resampling. The overall performance of the model was assessed by the Brier score and the discriminative performance by receiver operating characteristic (ROC) curve analysis.

RESULTS: A model was developed based on 291 patients and was internally validated with a scaled Brier score of 0.14. The area under the ROC curve of this model was 0.77 (95% confidence interval [CI], 0.71-0.82). At a cutoff value of predicted probability ≥ 0.50, the model demonstrated specificity of 88%, sensitivity of 43%, positive predictive value (PPV) of 65%, and negative predictive value (NPV) of 75%.

CONCLUSION: This prediction model shows reasonable accuracy for predicting axillary pCR. However, omitting axillary treatment based solely on the nomogram score is not justified. Further research is warranted to noninvasively identify patients with axillary pCR.