Molecular subtypes, body mass index (BMI), and their time-varying prognostic impact in node-positive breast cancer (BC): Pooled analysis from the WSG AM-01 and -02 trials.

Abstract

1021 Background: Obesity has been reported to be associated with aggressive tumor characteristics and poor prognosis in early BC in mostly unselected collectives. This study was performed to investigate the prognostic impact of obesity taking into account molecular subtypes in randomized node-positive BC cohort treated by chemotherapy (Cht). Methods: Pooled analysis in 2353 patients was done using data from the WSG-AM-01 (high risk BC with >9 positive lymph nodes (LN), treated by either tandem high dose or dose-dense Cht) and AM-02 trials (intermediate risk BC with 1−3 positive LN treated by CEF/CMF or EC-Doc). Data on hormone receptors (HR) and HER2 were available in 2045 pts, molecular subtypes (luminal A vs. B (by Ki-67 or HER2), basal/non-basal triple negative (TN) BC and HER2) by central pathology review in 865 pts. Time-varying Cox proportional hazards regression models were used to assess the prognostic impact of BMI after adjusting for other factors. Results: Median follow up was 59 months. BMI was available in 2302 pts (median BMI 24.9 kg/m2). 133 pts were found to be severely obese (BMI>35 kg/m2). Higher BMI was moderately associated with increasing age and larger tumor size. Interestingly, there was a significant association between higher BMI and TNBC as well higher Ki-67 levels, yet only in younger (<50 years) pts. No further association between molecular subtypes and BMI in both age subgroups could be observed. Only severe (>35 kg/m2) obesity was predictive for relapse risk by univariate analysis (HR=1.5; p=0.02). However if multivariate analysis was performed (including LN, age, grade, HR, HER2, TNBC, obesity and time-varying interaction between obesity and late relapse (>5 years), severe obesity was highly predictive only for late (HR=3,638, p<0.000) relapse. Conclusions: In this pooled analysis of two randomized chemotherapy trials, higher BMI was associated with TNBC only in younger patients. By time-varying multivariate analysis, only severe obesity was predictive for late relapse and next to counselling on life style changes, this factor may be suitable as a possible marker for patient selection regarding extended adjuvant therapy.