19P Her2 Extracellular Domain Levels in Patients With Her2-Positive Early Breast Cancer Experiencing Disease Progression: Case-Control Comparison of Samples from the Hera Study

Abstract

ABSTRACT Background Levels of serum-soluble human epidermal growth factor receptor 2 (HER2) extracellular domain (ECD) are elevated in some patients with HER2-positive breast cancer. Serum ECD levels have been proposed as a potential biomarker of response to given therapies or an indication of disease progression. Serum ECD elevations >20% from baseline have been associated with disease progression. We performed a nested case-control study to test this hypothesis in patients with HER2-positive early breast cancer (EBC). Methods Serum for the measurement of ECD was collected from 992 patients in the Herceptin Adjuvant (HERA) trial (N Engl J Med 2005; 353: 1659-72). Samples were taken at baseline, weeks 25 and 52, and months 18, 24, 36, 48 and 60, from patients in any of the three HERA trial arms. Serum was stored until assayed using the Siemens Serum HER-2/neu Test. Cases were identified as the last sample before the date of first local, regional or distant recurrence. Two control samples without recurrence prior to that of the case were selected to form case-control triplets, matched on time from randomisation to ECD sample, nodal and neoadjuvant chemotherapy status, histological grade, and ER/PgR status. Statistical analyses included conditional logistic regression (CLR) and the Wilcoxon signed-ranked test. Results A total of 160 case-control triplets were considered. The median ECD level in case samples was 2.4% higher than in control samples. CLR indicated that ECD samples with a higher ECD level had a higher probability of being a case sample than a control sample (p 20% in 44 (27.5%) and ≤20% in 116 (72.5%) case-control triplets. The Wilcoxon test did not indicate that the difference in mean ECD level between cases and controls was >20%. Conclusions Higher ECD levels were indicative of case samples rather than control samples. However, the mean difference between ECD levels in case samples and control samples was not >20%. These data were therefore insufficient to support the use of the specified ECD level as a predictive biomarker in patients with EBC. Disclosure M. Dowsett: I serve on Advisory Boards for Roche and receive grant income form them. M. Procter: Marion Procter's institution received funding from Roche in relation to the HERA trial. E. de Azambuja: Honoraria as Roche speaker; travel grant from Roche (ASCO 2011). M.J. Scullion: Roche employee. All other authors have declared no conflicts of interest.