Treatment of puNETs represents a clinical challenge. Although there is some evidence concerning genomic profile of puNETs, targetable molecular alterations remain unclear. WES provides insight into the genomic landscape of puNETs offering new opportunities for a precision medicine.
Fresh-frozen tumor-normal pairs from 35 typical carcinoid (TC), 4 atypical carcinoid (AC) and 9 large-cell neuroendocrine carcinoma (LCNEC) patients were consecutively collected from three European centers (GR, CCML, ICH) between Feb 2010 and Nov 2013. All specimens were reviewed by expert pathologists and those with more than 65% tumor cells on H&E were selected. Genomic DNA (gDNA) was extracted from tumor and matched normal tissue. After normalization and quality control, gDNA was captured using in-solution enrichment methodology (Human All Exon V5 + UTR–75 Mb, Agilent Technologies). Exome enriched libraries were sequenced on an Illumina HiSeq 2000 with a paired-end 2 x 100 bp protocol. Variants were called using VarScan2 against the reference genome hg19 (GRCh37). After filtering based on frequency, variants were annotated using SnpEff and SnpSift with dbSNP and dbNSFP.
Fifty-nine % of TC and AC were females, 89% of LCNEC were males. Median age was 57 (18-83) yrs, 26 (54%) stage I, 16 (24%) stage II, 3 (6%) stage III and 3 (6%) stage IV. On average, 11.6 Gb of sequence were produced per sample, aiming a mean coverage of 72 X. A median of 277 (10-8470) somatic variants per sample was observed. Overall, preliminary analysis showed several somatic variants in histone modifiers as MEN1 (n = 9), EZH2 (n = 6) and HDAC5 (n = 5). Variants in genes involving SWI/SNF complex as ARID, BCL-2 and SMARCA have also been found. Regarding the mTOR pathway, several variants were observed in PIK3 family. However, none of the alterations mentioned above were enriched in any subtype. Nevertheless, TP53 (n = 6) and RB1 (n = 3) variants were observed exclusively in LCNEC.
Our preliminary overview of the molecular landscape in puNETS provides the basis for further analysis. Some of these variants might potentially represent an actionable target. An exhaustive bioinformatic analysis will be presented.
Clinical trial identification
Legal entity responsible for the study
Georges Mathé/ESMO fellowship
All authors have declared no conflicts of interest.