Body mass index (BMI), tumor subtype, and relapse-free survival (RFS) in CALGB 9741 (Alliance).

Abstract

1032 Background: Obesity is a predictor of poor outcomes in women with early-stage breast cancer (BC). Some reports suggest that obesity is associated with aggressive tumor histology. We examined the relationship between BMI at diagnosis and PAM50 subtype, and explored the interaction between BMI and subtype on prognosis in early BC. Methods: CALGB 9741 evaluated dose-density and sequence in node-positive BC. All patients received doxorubicin, cyclophosphamide and paclitaxel dosed by actual body weight without cap or dose adjustment. The primary endpoint was RFS. Height and weight at diagnosis were abstracted from patient records; the PAM50 assay was performed using the Nanostring platform. Association between PAM50 and BMI was assessed by a chi-squared test. The prognostic value of BMI conditional on PAM50 was tested as a continuous variable using Cox models for RFS adjusted for number of involved nodes, tumor size, menopausal status, drug sequence, and dose density. Results: Baseline height and weight were available for 1909 of 2005 enrolled patients; 1272 also had subtype determination by PAM50. Distribution of subtypes differed significantly by BMI (p=0.03), with all weight groups having similar proportions of Basal and HER-2 Enriched subtypes and with obese patients having lower rates of Luminal A (30 vs 35%) and higher rates of Luminal B (39 vs 26%) tumors as compared to normal-weight individuals. In multivariate analyses, BMI and subtype were independent predictors of RFS (p=0.011 and p<0.001, respectively). Exploratory analyses did not show a significant interaction between subtype and the relationship between increased BMI and RFS (p=0.15), although largest differences were seen among Luminal cancers (Table). Conclusions: Biologic subtypes were distributed differently in obese and non-obese individuals. The association between BMI and RFS did not differ significantly by subtype in this study. More work is needed to further explore the interaction of BMI and subtype, especially in a larger cohort of Luminal tumors. [Table: see text]