Lack of efficacy of adjuvant lapatinib in HER2-negative breast cancer (HER2–ve BC): Analysis of patients in the TEACH trial.

Abstract

628 Background: Benefit from trastuzumab (T) in patients (pts) with HER2–ve tumors by central laboratory FISH testing was shown in an exploratory analysis of the NSABP B-31trial and confirmed by levels of expression of HER2 mRNA. DNA sequencing studies demonstrate that undetected HER2 mutations may drive tumor growth. It is hypothesized that pts with BC deemed as HER2–ve based on amplification may still benefit from anti-HER2 therapy. We undertook an exploratory analysis of disease free survival (DFS) in pts in TEACH whose tumors were HER2–ve or borderline by central FISH testing. Methods: TEACH, a randomized, double-blind, placebo (P)-controlled trial in 33 countries, evaluated lapatinib (L) in reducing relapse risk in T-naive pts pretreated with chemotherapy for HER2+ BC. L showed a hazard ratio (HR) for DFS compared with P of 0.83 (95% confidence interval [CI] 0.70-1.00); P=.053 in the ITT population (n=3147) and in pts with HER2+ BC by central FISH [HR: 0.82 (0.67-1.00); P=.04]. We conducted an exploratory analysis of DFS in pts with borderline or HER2–ve tumors by central FISH. Estimated DFS times were calculated using a stratified log-rank test. HR was estimated by a stratified Cox regression model. Results: 657 pts (21%) did not have centrally confirmed HER2+ BC (L: 341 and P: 316) in the ITT population: 425 had negative (HER2:CEP-17 ratio <1.8) or borderline (≥1.8–<2.0) FISH testing (L: 218 and P: 207), 216 were unevaluable due to insufficient tissue or sample failed to hybridize (L: 114 and P: 102) and 16 were not centrally tested. In the 425 pts with centrally confirmed HER2–ve or borderline FISH results with 27 events in L and 34 in P, the HR for DFS in L compared with P was 0.94 (0.56-1.57). The percentage of pts with DFS at 4 yrs was 85.0% (79.5%-90.4%) in L and 81% (75.1%-87.1%) in P. Conclusions: Although L showed efficacy in pts with centrally confirmed HER2+ BC in the TEACH trial, our analysis did not show benefit in pts whose tumors were HER2–ve or borderline by central FISH. The TEACH trial shows once again that quality control of HER2 testing is crucial and central laboratory testing should be considered for non-specialized centers worldwide. Clinical trial information: NCT00374322.