Abstract P3-13-06: Development and validation of a nomogram predicting pathological axillary status (ypN0 vs. ypN+) in a subgroup of patients converting from cN+ to ycN0 through neoadjuvant therapy (NAT) – A transSENTINA substudy

Abstract

Abstract Background: Particularly among patients converting from cN+ to ycN0 status through neoadjuvant therapy (NAT) the optimal method and extent of axillary staging is unclear. The aim of this analysis was to develop a nomogram predicting the probability of positive axillary status (ypN+) after PST among these patients based on clinical and pathological parameters. Methods:Patients converting from cN+ to ycN0 due to PST included in a prospective study (SENTINA, Arm C) were included. Univariate and multivariate analyses were carried out to evaluate the association between 14 clinical/pathological parameters and pathological axillary status (ypN0 vs ypN+) using logistic regression models. Model accuracy, sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV) were assessed applying leave-one-out cross-validation (LOOCV) and ROC analyses. Different cut-points were evaluated. Calculations were performed using the SAS Software (Version 9.4, SAS Institute Inc., Cary, NC, USA.). Results: Arm C contained 553 patients, 369 patients were evaluable with respect to the above parameters. Univariate analyses revealed a significant association between pathological axillary status and ER status (odds ratio (OR) 4.05, 95% confidence interval (95%CI) 2.81-5.83), PR status (OR 3.07, 95%CI 2.16-4.36), multifocality (OR 2.37, 95%CI 1.57-3.58), lymphovascular invasion (OR 8.61, 95%CI 5.12-14.46), detection of a SLN after NAT (OR .56, 95%CI .36-.87), detection method (IHC vs routine: OR .46, 95%CI .27-.78; IHC vs serial HE: OR .72, 95%CI .49-1.07; serial hematoxylin eosin (HE) vs routine: OR .639, 95%CI .39-1.04), clinical tumor size (OR 1.051, 95%CI 1.03-1.07) and pCR-status in the breast (ypT0 and ypTis vs others, OR .11, 95%CI .08-.17). A multivariate model was fitted including significant clinical parameters. Stepwise backward variable selection was carried out resulting in a model including ER status (OR 3.81, 95%CI 2.25-6.44), multifocality (OR 2.22, 95%CI 1.26-3.92), LVI (OR 9.16, 95%CI 4.68-17.90), detection of a SLN after NAT (OR .50, 95%CI .26-.95) and clinical tumor size (OR 1.03, 95%CI 1.01-1.06). In LOOCV, this model demonstrated an accuracy of 73% (sensitivity 73%, specificity 72%, PPV 75%, NPV 70%) using .5 as cut-off. Based on the performed ROC analysis an area under the curve (AUC) of 0.81 was calculated. Conclusion: A model using ER status, multifocality, LVI, detection of a SLN after NAT and clinical tumor size was built to predict pathological axillary status (ypN+) with a high accuracy. If successfully validated based upon an independent dataset, this nomogram could allow advising patients for / against axillary surgery in case of clinical axillary conversion after NAT. Citation Format: Liedtke C, Kolberg H-C, Kerschke L, Goerlich D, Bauerfeind I, Fehm T, Fleige B, Hauschild M, Helms G, Lebeau A, Schmatloch S, Schrenk P, Schwentner L, Staebler A, von Minckwitz G, Loibl S, Untch M, Kuehn T. Development and validation of a nomogram predicting pathological axillary status (ypN0 vs. ypN+) in a subgroup of patients converting from cN+ to ycN0 through neoadjuvant therapy (NAT) – A transSENTINA substudy [abstract]. In: Proceedings of the 2016 San Antonio Breast Cancer Symposium; 2016 Dec 6-10; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2017;77(4 Suppl):Abstract nr P3-13-06.