e12107 Background: Curebest™ 95GC Breast (95GC) is one of the multi-gene assays to predict prognosis of node negative and estrogen receptor (ER) - positive breast cancer patients, developed using 95 gene-set without overlap with that used in Oncotype DXⓇ(ref 1). It has been shown to have the capability to classify the “intermediate” patients determined using Recurrence Online (microarray-based simulation model for Oncotype DXⓇ) but was validated only using the data from single institute and public database. Here we report the result of the first multi-center validation study for this multi-gene assay. Methods: ER-positive and T1-2/N0/M0 breast cancer patients who received adjuvant hormonal therapy were enrolled retrospectively. Fresh frozen tissues were applied to the assay, resulting classification into “L” and “H”, which was used for the validation on 5 year recurrence free survival (5Y-RFS) data of each patient. Results: 73 cases out of 150 enrolled cases were eligible and analyzed. 46 patients were classified as “L” whose 5Y-RFS was 96.5% (95%CI:89.5-98.9) while 27 patients were classified as “H” whose 5Y-RFS was 79.0% (95%CI:63.6-88.5). There was a statistically significant difference between RFS of “L” and “H” group by Log-Rank test (p = 0.0016). Significant association with 95GC were seen in histological grade (p = 0.0012), Recurrence Online (p < 0.001) and PAM50 (p < 0.001). The assay could classify the patients of histological grade 2, intermediate group by Recurrence Online (RS > 17, RS < 31) and Luminal B patients into “L” and “H”. Conclusions: Curebest™ 95GC Breast was well validated by this first multi-centered retrospective study on 5Y-RFS of the ER positive, node-negative patients who received only hormonal therapy in adjuvant setting. This result indicates the usefulness of 95GC as a novel multi-gene assay, as it can classify target patients into 2 groups, “H” and “L” according to predicted prognosis of 5Y-RFS. Reference: 1. Naoi et al. Breast Cancer Res Treat (2011) 128:632-641