A miRNA assay for the classification of benign and neoplastic lesions in pancreatic fine needle aspirates.

Abstract

10614 Background: Improved diagnosis of pancreatic ductal adenocarcinoma (PDAC) versus benign pancreatic lesions is an urgent clinical need. We previously reported a laboratory-developed test based on the differential expression of 2 miRNAs, miR-196a and miR-217, with a sensitivity and specificity of 95% in formalin-fixed, paraffin-embedded (FFPE) specimens with ≥60% tumor content. Here, we report the development and validation of a miRNA-based research test optimized for low % tumor specimens and un-enriched fine needle aspirates (FNA). Methods: RNA was extracted using methods developed and validated by the Asuragen Services Laboratory from frozen, FFPE or FNA pancreatic specimens. One to 3 endoscopic ultrasonography-guided FNAs were collected, stored, and shipped in RNARetain, an RNA stabilization solution formulated in 1 mL single-use vials. Expression levels of up to 12 miRNAs were interrogated using TaqMan qRT-PCR assays. Results: A set of 211 specimens, including 94 PDAC, 96 chronic pancreatitis, and 21 normal, was used to develop a miRNA classifier to discriminate benign and low tumor content PDAC specimens. Training on 95 FFPE specimens showed that expression analysis of miR-196a, -210, and -375 improved by 2- to 3-fold the detection rate of PDAC in specimens with 5-88% tumor. The classifier performance was further verified on 4 independent sets of specimens (n=116): frozen with high tumor content (80-100%), FFPE with high tumor content (60-95%), FFPE with low tumor content (10-60%), and FNA. The signature alongside the remaining 9 miRNAs, is now being evaluated in a multi-site prospective study using FNAs collected in RNARetain; 133 specimens have been analyzed to date and we will report the relative performance of the top miRNA classifiers. Conclusions: Benign lesions and PDAC can be accurately classified based on the expression level of a few miRNAs in specimens with 5-100% tumor and in FNA. Testing of pre-operative FNAs collected in RNARetain as part of routine clinical evaluation could in the future improve the differential diagnosis of PDAC and potentially reduce the rate of indeterminate or false negative calls.