Abstract PD02-07: Next-generation sequencing of FFPE breast cancers demonstrates high concordance with FISH in calling HER2 amplifications and commonly detects other clinically relevant genomic alterations

Abstract

Abstract Background: As more therapies targeting genomic alterations become available, next-generation sequencing (NGS) is increasingly performed in tumor types where mutational status may drive treatment choice. In addition to its ability to identify base substitutions, insertions and deletions across entire exons, NGS can detect relevant copy number changes such as amplification of HER2 in breast tumors. However, for NGS to be clinically applicable, it must reliably analyze FFPE tumor samples and show concordance with the best current diagnostic methods. Methods: To confirm a clinical role for NGS in detecting copy number alterations, we identified 35 FFPE invasive breast carcinomas previously tested for HER2 status by FISH, including 15 HER2 positives (≥7 copies) and 20 HER2 negatives (<4 copies) and sequenced 3,230 exons of 182 cancer genes including HER2, in a CLIA certified lab (Foundation Medicine). Average coverage depth of >900X uniquely-mapping reads was obtained. Sequence data were analyzed for HER2 copy number (blinded to FISH results) based on a statistical model using allele frequencies and coverage depth of HER2 exons versus a process-matched normal control, classifying cases as HER2 positive (≥6 average copies), HER2 negative (<4 copies), intermediate (4–5 copies) or unknown (<20% tumor purity). The data were also analyzed for additional clinically relevant genomic alterations. Results: High concordance was noted between HER2 copy number status determined by FISH and NGS: 30 of the 35 samples were classified as positive or negative by NGS, 1 was classified as intermediate and 4 as unknown due to low purity. Using FISH as a gold standard, NGS HER2 calls demonstrated an accuracy of 97% (29/30, 95% CI 83–99%), 93% sensitivity (13/14, 95% CI 69–99%) and 100% specificity (16/16, 95% CI 81–100%). One discordant case was noted (FISH positive, NGS negative). Furthermore, NGS revealed 70 additional alterations (38 base substitutions, 10 insertions/deletions, 22 copy number alterations) in 23 cancer genes (an average of 2.0 alterations per sample). Genomic alterations that predict sensitivity or resistance to approved or experimental targeted therapies and thus plausibly guide treatment decisions were found in 69% of patients. These include PIK3CA (16 cases, PI3 kinase/mTOR inhibitors), PTEN (3 cases, PI3K/AKT/mTOR inhibitors), KRAS (1 case, resistance to cetuximab and panitumumab), and NF1 (1 case, mTOR/MAPK inhibitors) plus amplifications of CCND1 (4 cases, CDK4 inhibitors), FGFR1 (3 cases, FGF inhibitors) and MCL1 (3 cases, BCL-2 inhibitors, resistance to anti-tubulin therapies). Four cases included co-amplification of RARA with HER2. Conclusions: We conclude that HER2 status can be reliably determined by NGS on FFPE breast cancers and that NGS uncovers additional actionable genomic alterations that could impact disease management in a high proportion of patients. Further evaluation of NGS as a guide to therapy in breast cancer is warranted. Citation Information: Cancer Res 2012;72(24 Suppl):Abstract nr PD02-07.