694PDiagnostic value of digital multiplexed detection of single nucleotide variants in pancreatic cancer specimens collected by endoscopic ultrasound fine-needle aspiration

• Published in: Annals of oncology Vol. 30; no. Supplement_5
• Main Authors: Cazacu, I M, Semaan, A, Stephens, B M, Swartzlander, D B, Guerrero, P A, Singh, B S, Lungulescu, C V, Raileanu, S, Danciulescu, M M, Roy-Chowdhuri, S, Maitra, A, Saftoiu, A, Bhutani, M S
• Format: Journal Article
• Language: English
• Published: Oxford University Press 01.10.2019
• ISSN: 0923-7534; 1569-8041
• DOI: 10.1093/annonc/mdz247.021

Abstract Background Current diagnostic approaches for pancreatic cancer (PC) mostly rely on cytologic examination of endoscopic ultrasound fine-needle aspiration (EUS-FNA) samples. However, in a subset of cases, PC diagnosis remains inconclusive due to low tumour cellularity. Molecular analysis of EUS-FNAs might be used as an auxiliary tool to strengthen diagnosis in samples with suboptimal cytology. This study aimed to evaluate the diagnostic utility of a single nucleotide variants (SNV) assay using molecular barcode technology performed on EUS-FNAs. Methods 28 EUS-FNA samples of pancreatic masses (19 PC, 9 non-malignant lesions) were analyzed. FNAs were collected in RNAlater and stored at -80 C. Mutational status was evaluated using the Nanostring Vantage 3D™ DNA SNV Solid Tumor Panel, utilizing digital enumeration of unique barcoded probes to detect 104 SNV from 24 genes of clinical significance. 5ng of tumor-derived DNA was subjected to multiplexed preamplification and hybridization of variant-specific probes to unique fluorescent barcodes. A multiplex KRAS assay (G12/13) droplet digital PCR (ddPCR) was used to confirm SNVs. Results The SNV assay detected at least one variant in 18/19 (95%) PC samples. One PC case harbored 3 SNVs. Among the PC samples, KRAS variants (G12D, G12V, G12R, Q61H, Q61L) were detected in 17 (90%) cases, EGFR in 1 (5%), and PIK3CA in 1 (5%). All KRAS mutations were also detected by ddPCR. Diagnostic accuracy of cytology alone for PC was 68% (19/28). 32% of the FNAs were inconclusive; at least one SNV was detected in 5/6 inconclusive FNAs with a final diagnosis of PC. No SNV was identified in the remaining 3 inconclusive cases diagnosed as chronic pancreatitis. Combining cytology and SNV analysis for inconclusive cases increased the diagnostic accuracy to 96% (27/28). Conclusions Nanostring SNV assay combined with cytology can enhance the diagnostic power of EUS-FNA, especially in inconclusive cases, preventing repeat biopsies, unnecessary resections for benign disease or delay in PC patients care. Given the low DNA input, digital data output and rapid turn-around time, this novel technology may be instrumental for the preoperative molecular diagnosis of PC. Legal entity responsible for the study The authors. Funding Has not received any funding. Disclosure All authors have declared no conflicts of interest.