Clinical characterization, prognostic, and predictive values of HER2-low in patients with early breast cancer in the PALLAS trial (ABCSG-42/AFT-05/BIG-14-13/PrE0109).

ABSTRACT

BACKGROUND: Bidirectional crosstalk between HER2 and estrogen receptor (ER) pathways may influence outcomes and the efficacy of endocrine therapy (ET). Low HER2 expression levels (HER2-low) have emerged as a predictive biomarker in patients with breast cancer (BC). METHODS: PALLAS is an open, international, phase 3 study evaluating the addition of palbociclib for 2 years to adjuvant ET in patients with stage II-III ER-positive/HER2-negative BC. To assess the impact of HER2 expression on patient outcomes in the phase III PALLAS trial, we analyzed (1) the association between rate of HER2-low with demographic and clinicopathological parameters, (2) the prognostic value of HER2-low status on invasive disease-free survival (iDFS), distant relapse-free survival (DRFS), and overall survival (OS) and (3) HER2 expression's value as a predictive biomarker of response to palbociclib. HER2-low was defined as HER2 immunohistochemistry (IHC) 1 + or IHC 2 + with negative in situ hybridization (ISH). All pathologic evaluation was performed locally. Prognostic and predictive power of HER2 were assessed with Cox models. RESULTS: From the original PALLAS intention-to-treat population (N = 5753), 5304 patients (92.2%) were included in this analysis. Among these, 2254 patients (42.5%) were classified as having HER2 IHC 0 (HER2-0), and 3050 (57.5%) as having HER2-low disease (1838 with IHC 1 + and 1212 with IHC 2 +). Median follow-up was 59.8 months. HER2-low prevalence varied significantly across 21 participating countries (range 16.7% to 75.6%; p < 0.001) and was more frequent in patients enrolled in North America (63.1%) than in Europe (53.4%) or other regions (53.4%) (p < 0.001). HER2 status was not significantly associated with iDFS in a multivariable Cox model (hazard ratio 0.93, 95% confidence interval 0.81 - 1.06). No significant interaction was observed between treatment arm and HER2 status for iDFS (p = 0.43). Similar results were obtained for DRFS and OS. CONCLUSIONS: In this large, prospective, global patient cohort, no differences were observed in clinical parameters, prognosis, or differential benefit from palbociclib between HER2-0 and HER2-low tumors. Significant geographic variability was observed in the prevalence of HER2-low status, suggesting a high degree of variation in pathologic assessment of HER2 expression without impact on outcomes.