Genomic profiling of esophagogastric (EG) tumors in clinical practice.

Abstract

57 Background: Esophagogastric (EG) cancer is a phenotypically heterogeneous disease. The Cancer Genome Atlas (TCGA) identified four distinct subsets: 1) Epstein Barr Virus (EBV) tumors with PIK3CA mutations, PD-L1/2 amplification (amp), 2) Tumors with Microsatellite instability (MSI-H), 3) chromosomally unstable (CIN) tumors with frequent amplifications, 4) chromosomally stable/diffuse type tumors with RHOA mutations. We explore the utility/feasibility of genomic profiling of EG tumors in routine clinical practice. Methods: Patients (pts) with metastatic EG adenocarcinoma undergoing treatment at MSKCC were consented for molecular testing under institutional protocol. Archival formalin fixed paraffin embedded (FFPE) samples were analyzed using an on-site 341 cancer-associated gene bait capture, next generation sequencing (NGS) assay. Results: Of 70 analyzed EG tumors, 66 (94%) harbored at least one genomic alteration, the most frequent being TP53 mutations (77%), HER2 amp (30%) and EGFR amp (10%). Alterations in PI3K/AKT/mTOR (9%) and G1-S1 cell cycle regulators (CDK12 (11%), CDKN2A (10%)) were also observed. A comparison with the TCGA results revealed an over-representation of the CIN subtype (65% vs 50% in TCGA). We found very few EBV or MSI tumors (3% each), with the remaining 29% being chromosomally stable. Data is reported in the clinical medical record and maintained in the MSKCC-internal cBioPortal for Cancer Genomics ( http://cbioportal.org ) (Gao, et al.). The portal is a web resource for exploring, visualizing, and analyzing multidimensional cancer genomics data and when available, to link these data to therapeutic interventions and clinical outcomes. Conclusions: CIN tumors with frequent amplifications represent majority of metastatic EG cancer at our institution. Optimal development of target-specific therapy for EG tumors will require genomic characterization. Comprehensive molecular profiling is feasible in routine clinical practice and has created a roadmap for pt stratification and genome-guided trial development.