Abstract S5-1: Biomarker analyses in CLEOPATRA: A phase III, placebo-controlled study of pertuzumab in HER2-positive, first-line metastatic breast cancer (MBC)

Abstract

Listen

Translate article

Abstract Background: The anti-HER2 humanized monoclonal antibodies trastuzumab (T) and pertuzumab (P) have complementary mechanisms of action. CLEOPATRA showed that P+T+docetaxel (D) significantly improved progression-free survival (PFS) compared with placebo (Pla)+T+D in the first-line treatment of HER2-positive MBC (HR = 0.62, 95% CI 0.51–0.75; p < 0.001), with a similar safety profile. A panel of biomarkers (BMs) was assessed in tumor tissue and in serum samples to explore their potential predictive and/or prognostic value. Methods: Samples were collected at baseline prior to the first dose of study drug (N = 808; 58–99% of which were assessable for BMs due to tissue availability/technical reasons). HER2, HER3, IGF1R, PTEN, and pAKT were assessed by immunohistochemistry and a modified H-score was derived for each. HER1, HER2, HER3, AREG, and betacellulin tumor mRNA levels were assessed by quantitative reverse transcription polymerase chain reaction (qRT-PCR) and c-myc by fluorescence in situ hybridization. PCR-based methods were used for mutational analyses of 8 hotspots within PIK3CA on tumor DNA, and for assessments of FCγR polymorphisms in DNA extracted from whole blood. Serum HER2 extracellular domain, AREG, EGF, and TGFα were detected by ELISA. Exploratory subgroup analyses of PFS by BM level were performed for each BM. Median values were used as cutoffs for high vs low levels, except for PTEN, PIK3CA mutation, and c-myc amplification, where cutoffs were applied following a biologic rationale. Univariate Cox regression analyses assessed the prognostic value on a background of HER2-targeted treatment (relationship of each BM to clinical outcome, both arms pooled) and predictive value (association of BMs with treatment benefit) of each BM. Analyses were exploratory; therefore no adjustment for multiplicity was made. Results: Significant prognostic variables identified were: HER2 protein (p = 0.05), HER2 mRNA (p = 0.008), HER3 mRNA (p = 0.03), and PI3KCA mutation (p = 0.0001). The HR was <1 for each, indicating that patients with high levels of HER2 protein, HER2 mRNA, HER3 mRNA, or wild type PIK3CA had a better prognosis. Irrespective of prognostic impact, the PFS benefit of P+T+D, as observed in the overall study population, was maintained following individual adjustment for each BM (HR range=0.61–0.66), including PIK3CA (HR = 0.66). Analyses exploring the predictive value showed that all patients derived a benefit from P+T+D treatment compared with Pla+T+D, regardless of the expression of any assessed candidate BMs. HR point estimates were <1.0 for all BM subgroups assessed. Summary: The BM analysis from CLEOPATRA is consistent with results from previous phase II studies: HER2 remains the only marker suitable for use when selecting patients for anti-HER2 therapy. The current analyses did not identify any additional predictive markers that would allow refinement of the HER2-positive target population treated with P+T+D. The lack of a HER2 treatment-naive control group may partly explain the lack of a signal. The PI3K mutational status may identify a poor prognostic group within the HER2-positive population that could be appropriate for clinical trials of HER2-targeted molecules in combination with PI3K pathway-targeted agents. Citation Information: Cancer Res 2012;72(24 Suppl):Abstract nr S5-1.