Neuroendocrine carcinoma of the esophagus: A genomic landscape study.

642 Background: Esophageal neuroendocrine carcinoma (ENEC) is a rare and aggressive upper GI tract malignancy when compared with the far more frequent esophageal adenocarcinoma (EAC). Methods: 194 ENEC and 9920 EAC underwent hybrid capture based comprehensive genomic profiling (CGP) to identify all classes of genomic alterations (GA). Microsatellite instability status (MSI) and tumor mutation burden (TMB) were determined from the sequencing data. PD-L1 expression was determined by IHC using the Dako TPS score (0% = negative; 1-49% = low positive and >50% = high positive). Results: Although the mean (63.6 vs 64.8 yrs) and median (65 yrs for both) ages of the 2 groups were nearly identical, the ENEC group featured significantly more female patients (20.6% vs 13.2%; p=.002). The number of driver GA per tumor sample was slightly higher in the EAC group (NS). MSI-High status was infrequent in both groups (2.1% vs 2.6%; NS), the relative low median and mean TMBs were also similar with the TMB >10, % frequency slightly lower in the ENEC group (7.2% vs 8.4%; NS) and low PD-L1 expression was higher in the ENEC group but low sample numbers impacted significance (25.0% vs 12.3%; NS). Both ENEC and EAC featured high frequencies of GA in TP53 (84.5% vs 86.9%; NS) with additional GA in RB1 relatively unique for the ENEC (40.2% vs 1.9%; p<.0001). Nearly all ENEC had TP53 GA when RB1 GA were identified. CDKN2A GA were more common in EAC (25.8% vs 33.0%; p=.011), but when MTAP GA were identified in the CDKN2A mutated cases, MTAP deletion were equally identified (9.0% vs 9.5%; NS). ERBB2 GA were mostly amplifications and significantly more frequent in the EAC group (5.7% vs 21.8%; p<.0001). Another GA more frequently associated with EAC included KRAS (4.1% vs 24.3%; p<.0001) and KRAS G12C (0% vs 0.6%; NS). GA more frequently identified in the ENEC included SMARCA4 (20.1% vs 5.1%; p<.0001), APC (13.9% vs 8.4%; p=.006) and CTNNB1 (13.9% vs 3.5%; p<.0001). Additional noteworthy low frequency GA equally identified in both groups included NOTCH1 (6.7% vs 5.5%; NS), CDH1 (3.1% vs 2.7%; NS) and BRAF (3.1% vs 1.6%; p=.013). Conclusions: Although similar in many respects such as in the identification of immunotherapy biomarkers (MSI and TMB status), ENEC patients are more often female and are characterized by a different genomic landscape when compared with EAC. Genomic landscape of ENEC. ENEC (194 cases) EAC (9920 cases) p-value TP53 + RB1 36.6% 1.5% <.0001 CDKN2A 25.8% 33.0% =0.011 SMARCA4 20.1% 5.1% <0.0001 ERBB2 5.7% 21.8% <0.0001 MTAP 9.0% 9.5% NS APC 13.9% 8.4% =0.006 CTNNB1 13.9% 3.5% <0.0001 KRAS 4.1% (0% G12C) 24.3% (2.5% G12C) <0.0001 BRAF 3.1% 1.6% =0.013