Abstract P2-08-12: Integration of tumor size and grade with the breast cancer index (BCI) for prediction of distant recurrence in hormone receptor-positive breast cancer with 1-3 positive lymph nodes

Abstract

Abstract Background: BCI is a genomic signature (Molecular Grade Index (MGI) and HOXB13/IL17BR (H/I)) that significantly predicts risk of distant recurrence (DR) in hormonal receptor-positive, lymph node negative (LN-) breast cancer. As previously shown in the TransATAC and MA.14 studies, BCI was also prognostic for DR in lymph node-positive (LN+) patients. Here, a distinct BCI model that integrates tumor size and grade was evaluated for prediction of DR in women with 1-3 lymph node positive disease. Methods: 219 primary tumor samples from hormonal receptor-positive patients with 1-3 positive lymph nodes treated with 5 years of tamoxifen or anastrozole were examined. Women with four or more positive lymph nodes were excluded. BCI was combined with tumour size and grade into a comprehensive risk score, BCIN+. Kaplan-Meier (KM) estimates of 10 year DR and hazard ratios (HR) and 95% confidence intervals (CI) were estimated. Change in likelihood ratio 2 (LR-2) values were used to measure prognostic information of each variable alone or combined in new score. New cutpoints for low versus high risk groups for the new model were determined to ensure the low risk patients had minimal 10-year residual disease. Results: In 219 LN+ patients, BCI alone provided substantial additional prognostic information to tumor size (LR-2=11.83, P=0.0006) and grade (LR-2=8.33, P=0.004). Both clinical variables provided additional significant prognostic information to BCI alone (BCI alone: LR-2=9.59, P=0.0004; T: LR-2=7.09, P=0.008; G: LR-2=27.59, P<0.0001). Integration of tumor size and grade with BCI (BCIN+) provided additional highly significant prognostic information compared to BCI alone (Interquartile HR=3.15 [95% CI: 1.54-6.04]; LR-2=33.89, P<0.0001). A cut-point for a very low risk group in this LN+ population was determined, and included 55 (25%) women with no DR within 10 years. In contrast, 51 (31.1%) women developed a DR in the high risk group (N=164). 10-year DR risk for those in the high risk group was 35.4% (95% CI 28.0-44.1%). Discussion: Integration of tumor size and grade significantly enhanced the prognostic ability of BCI to predict 10 year DR risk in hormonal receptor-positive patients with 1-3 positive nodes. A significant number of patients have been identified to have a very low 10-year risk for DR, who may choose to safely forego unnecessary adjuvant chemotherapy or extended adjuvant endocrine therapy. Validation of BCIN+ in other datasets is ongoing. Citation Format: Sestak I, Zhang Y, Schroeder B, Dowsett M, Sgroi D, Cuzick J, Schnabel CA. Integration of tumor size and grade with the breast cancer index (BCI) for prediction of distant recurrence in hormone receptor-positive breast cancer with 1-3 positive lymph nodes. [abstract]. In: Proceedings of the Thirty-Eighth Annual CTRC-AACR San Antonio Breast Cancer Symposium: 2015 Dec 8-12; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2016;76(4 Suppl):Abstract nr P2-08-12.