Massively Parallel Sequencing in NSCLC: Comparison to Traditional Hot Spot Analysis for Selection of Approved and Novel Targeted Therapies
Jeffrey Ross, Alex Parker, Mirna Jarosz, Sean Downing, Roman Yelensky, Doron Lipson, Philip Stephens, Gary Palmer, Maureen Cronin, Christine Sheehan. Albany Medical College, Albany, NY; Foundation Medicine Inc., Cambridge, MA
Background: The recent introduction of next generation (NGS) DNA sequencing to clinical samples has enabled the discovery of novel and unanticipated genomic-derived targets of therapy response and resistance for patients with NSCLC.
Design: DNA was extracted from 4 x 10 μm FFPE sections from 45 primary NSCLC (26 female; 19 male; mean age 68 years; 24% Stage I; 13% Stage II; 5% Stage III; 16% Stage IV; 46% Stage unknown). The exons of 145 cancer-related genes were fully sequenced using the Illumina HiSeq 2000 (Illumina, San Diego, CA) to at an average sequencing depth of 253X and evaluated for point mutations, insertions/deletions (indels), specific genomic rearrangements and copy number alterations (CNA). Samples included 5% fluid cellblocks; 5% regional lymph nodes; 3% pericardial biopsy and 87% lung biopsies or resections. There were 42 adenocarcinomas (27acinar, 12 lepidic, 2 mucinous, 1 papillary), 1 large cell carcinoma, and 2 squamous cell carcinomas. In 23 adenocarcinomas, the NGS results were compared with commercial laboratory allele-specific PCR genotyping on the same tissue blocks.
Results: In the comparison study of EGFR status, the NGS result was concordant with commercial laboratory genotyping in 23/23 (100%) cases. In 22 additional NSCLC samples, NGS revealed 53 total genomic alterations, including 14 (64%) base substitutions, 2 (9%) INDELs, 6 (27%) CNA, and 0 (0%) rearrangements. Genomic alterations associated with sensitivity or resistance to targeted therapies for NSCLC were found in 16/22 (73%) of cases including 10 KRAS, 4 STK11, 3 JAK2, 2 PIK3CA, 2 BRAF, 2 EGFR, 1 NF1, 1 TSC1, 1 TSC2, 1 CCNE1, 1 PTCH, 1 CDK4, 1 CCND1, 1 BRCA2, 1 CDKN2A, and 1 ATM mutation. In comparison with the COSMIC database, NGS results were similar for most genes except for a lower rate of EGFR mutations (9% vs. 21%), a higher rate of KRAS mutations (41% vs. 16%) and an unprecedented rate of JAK2 mutations (14% vs. 0%). 7/22 (32%) of the NSCLC had 2 or more potentially actionable alterations after NGS.
Conclusions: Deep sequencing of clinical NSCLC samples is completely concordant with traditional hot-spot genotyping and also uncovers an unexpected number of genomic alterations that could influence therapy selection for this disease. Broad-based, deep sequencing of cancer-related genes results in sensitive detection of all classes of genomic alterations in NSCLC and can reveal actionable genomic abnormalities that inform treatment decisions.
Monday, March 19, 2012 2:45 PM
Platform Session: Section D, Monday Afternoon