Abstract S5-3: PHARE Trial results of subset analysis comparing 6 to 12 months of trastuzumab in adjuvant early breast cancer

Abstract

Abstract Background: Since 2005, One year trastuzumab (T) treatment has been providing survival benefit to patients with early breast cancer and HER2 overexpression. However, the optimal duration of T has been debatable due to concerns for cardiac toxicity and results from the FinHer trial which showed that 9 weeks of T provided a similar magnitude of benefit than the 1-year treatment. The French National Cancer Institute (INCa) initiated an academic randomised non-inferiority trial aiming to compare a shorter T exposure of 6 months versus the standard 12 months. This trial, named PHARE for ‘Protocol for Herceptin® as Adjuvant therapy with Reduced Exposure’ [NCT00381901] was approved byan independent ethics committee with regularly planned IDMC meetings. Patients and methods: Patients with HER2 + early breast cancer who received at least 4 cycles of (neo)-adjuvant chemotherapy were eligible. Randomization 1:1 using a minimisation algorithm was stratified on concomitant or sequential T administration with chemotherapy, oestrogen receptors (ER) status and center. The primary objective was to compare disease free survival (DFS). Overall survival (OS) and cardiac toxicity were investigated as secondary aims. An absolute loss of 2% in DFS in the experimental arm was defined as the non-inferiority margin (1.15 in relative terms) and required 3400 patients with alpha=0.05 and 80% power. Results: Between 5/2006 and 7/2010, 3382 patients were randomized to 6 or 12 months of T following the IDMC recommendation for accrual interruption and extended follow-up. Disease and treatment characteristics were well balanced between the 2 arms: median age 55 years (range 21–86), median tumour size 20 mm (0–270), node involvement 45%, SBR grade III 56%, ER positive 58%, radiotherapy 88%; concomitant T administration 58%, anthracyclin and taxane containing chemotherapy 73%. The data-base was locked on July 30th, 2012. The 95% HR confidence interval was ranged between 1.09–1.66 for DFS between the 2 arms and it included the pre-specified non-inferiority margins of 1.15. Thus the results were inconclusive regarding the non-inferiority (p = 0.14). A subset analysis based on stratification factors will be presented at SABCS. Citation Information: Cancer Res 2012;72(24 Suppl):Abstract nr S5-3.