P1-07-08: PCR Genomic Grade in Breast Cancer: A New Tool for Daily Practice.

Abstract

Background: Genomic grade (GG) is a 97-gene signature which improves the accuracy and prognostic value of histological grade (HG) in invasive breast cancer (IBC) (Sotiriou JNCI 2006). It is particularly useful in stratifying HG2 tumors into GG-1 (low grade) and GG-3 (high grade) tumors. In order to extend the applicability of GG on routinely-used formalin-fixed-paraffin-embedded (FFPE) samples, the original microarray MapQuant Dx® signature was converted into a RQ-PCR assay (PCR-GG). The study aimed at validating the grading classification performance of the newly developed PCR-GG test. Methods: Subsets of genes derived from the original GG were selected by testing their grading classification and prognostic performances using different classifiers on 15 independent public microarray data sets. 25 most-performing genes were quantified by RQ-PCR using a TaqMan assay on a training set consisting of 91 early IBC cases graded acc. to the Elston-Ellis recommendations, and classification performance of various combinations assessed. Analytical performance of the final reduced signature was assessed on 4 samples (5 RNA extractions and 3 RQ-PCR per sample) and used to define a 95% CI around the GG-1/GG-3 cut-off. Cases with a PCR-GG score falling into the 95% CI were considered “Equivocal” (Eq). Concordance with the MapQuant Dx® GG was assessed on 44 paired FFPE/frozen samples of the training set. Grade classification performance of the PCR-GG test was evaluated on an independent validation set of 396 FFPE samples retrieved from the Department of Pathology (2004 to 2010) of Institut Jules Bordet (Brussels, Be). FFPE samples from early invasive, N0, ER+, HER2−, ductal or lobular BC containing ≥30% inv. tumor cells were selected. Classification performance was evaluated on the whole cohort and by histological subtype. Results: The PCR-GG signature is based on the expression of 6 reporter and 3 reference genes. Most of the genes are overexpressed in grade 3 tumors and are associated with proliferation. The 9-gene PCR-GG showed a high concordance with the original GG microarray test (95%). 388 eligible cases were available for validation. 336 samples contained ≥30% inv. tumor cells and 322 were successfully amplified, corresponding to a 96% technical success rate. The HG1, HG2 and HG3 distribution in the validation cohort were 27%, 55% and 18% resp. PCR-GG reclassified 84% of all tumors (Table 1) with an overall concordance of 90% (92% HG1/GG-1 concordance and 88% HG3/GG-3). 82% of HG2 tumors were reclassified into 74% GG-1 and 26% GG-3. Overall concordance was 91% in the Ductal subset (comprising 84% of all cases) and 89% in the Lobular subset. Conclusion: The newly developed PCR-GG test composed of 9 genes can be performed on FFPE samples. It is highly concordant with HG and recapitulates the classification performance of the original 97-gene microarray GG test with 84% reclassified tumors. Further validation is planned on the BIG 1–98 cohort. Applicability of GGI on FFPE samples will open its use in routine clinical practice. Citation Information: Cancer Res 2011;71(24 Suppl):Abstract nr P1-07-08.