Factors Associated with Nodal Positivity Following Neoadjuvant Systemic Therapy in Breast Cancer Patients Who are Initially Node-Negative on MRI.

Abstract

The objective of this study is to determine the rate of positive axillary lymph nodes on surgical specimens (ypN+) following neoadjuvant systemic therapy (NST) in patients who were classified as node negative (cN0) on diagnostic dynamic contrast-enhanced MRI and to identify clinicopathological factors that predict ypN+ status. This retrospective study included consecutive cases of newly diagnosed invasive breast cancer patients who had breast MRI for locoregional evaluation prior to NST at our institution between 1 February 2013 and 1 February 2018. Clinically node-positive or recurrent breast cancers were excluded. Clinicopathologic features, molecular subtype data, pre- and post-NST imaging results, and definitive surgery results were recorded. Patient and tumor histopathologic characteristics' association with cN0 status was assessed with chi-squared, Fisher's exact, and Mann-Whitney U tests. We developed a model to predict axillary metastasis using tenfold cross-validation with least absolute shrinkage and selection operation for variable selection, reporting mean area under the curve and 95% confidence intervals (CI). Of 971 patients with breast cancer who underwent magnetic resonance imaging (MRI), 194 (20.5%) met the inclusion criteria. The mean age was 53 years (SD ± 11.5). All patients underwent axillary surgical staging after NST (174 [89.7%] sentinel lymph node biopsy alone, 2 [1.0%] axillary lymph node dissection alone, and 18 [9.3%] both). Residual nodal disease (ypN+) was identified in 27 patients (13.9%: ypN1 in 24 [12.4%], ypN2-3 in 3 [1.6%], median, 1 [1-14]). Patients with triple negative subtype had a lower ypN+ rate (4.8%) compared with the human epidermal growth factor 2 (HER2)-positive (11.8%) and hormone receptor-positive (25.0%) subtypes (p < 0.01). All patients with pathologic complete response in the primary tumor were ypN0. Our model showed modest performance for predicting ypN+ status (AUC: 0.62, 95% CI 0.50-0.75). In our cohort, breast cancer patients who were assessed as cN0 by breast MRI prior to receiving NST had a ypN+ rate of 13.9%, with a median of a single metastatic node.