Utility of MicroRNAs in Differentiating Pancreatic Cancer and Autoimmune Pancreatitis

Abstract

Introduction: Distinguishing pancreatic ductal adenocarcinoma (PDAC) from autoimmune pancreatitis (AIP) remains a challenge. Better diagnostic markers are needed to diagnose AIP. MicroRNAs (miRNA) are important biomarkers in diagnosis of pancreatic cancer. Aim: This pilot study seeks to identify whether a 7-miRNA panel developed for distinguishing PDAC from benign pancreatic lesions in fine needle aspirates (FNA) can also help differentiate PDAC from chronic pancreatitis (CP) and AIP. Methods: 34 formalin fixed, paraffin embedded (FFPE) specimens were selected: 1) 15 PDAC, 2) 13 AIP 3) 6 CP. Total RNA extraction was performed using protocols developed at Asuragen followed by interrogation of expression levels of previously published miRInform Pancreas Test consisting of a 7-miRNA panel (miR-24, -96, -130b, -135b, -148a, -196a and -375) using TaqMan® MicroRNA Assays. The test score was calculated for each sample using proprietary bioinformatics algorithms with 0.5 to 1, inclusive, consistent with PDAC, and the results compared to histological evaluation. Results: Mean age of the 34 patients was 54±17 years with 50% female. Symptoms were present in 78% of AIP (abdominal pain), 80% of CP (abdominal pain, recurrent pancreatitis in 1 patient), and 50% of PDAC patients (painless jaundice and abdominal pain) (p=0.6). Serum CA 19-9 was not elevated in patients with AIP or CP while abnormal in 1 patient with PDAC. Cytology was available in 4 AIP with none diagnostic; 1 was suspicious for adenocarcinoma and 1 consistent with adenocarcinoma. Cytology was performed in 1 CP patient and showed atypical cells; 75% of PDAC patients had diagnostic cytology. miRNA expression profiling was successfully completed on 32 FFPE samples (2 AIP samples did not yield adequate RNA). None of the 11 AIP and 6 CP specimens had a test score indicating PDAC, while 9 of 15 PDAC samples had a score consistent with PDAC. Sensitivity of the 7-miRNA classifier was 60% [95% CI 32-84], specificity 100% [95% CI 80-100], positive predictive value 100% [95% CI 66-100], and negative predictive value 74% [95% CI 52-90]. Conclusion: - The 7-miRNA classifier miRInform Pancreas test can help differentiate AIP and CP from PDAC in FFPE specimens. - Future implication: Endoscopic ultrasound-guided FNA of pancreatic masses combined with the 7-miRNA classifier may improve the ability to differentiate AIP from pancreatic adenocarcinoma.