microRNA Biomarkers as Potential Diagnostic Markers for Cancer - Genomic Biomarkers for Pharmaceutical Development - page 100

important tools for diagnosis and classification of this disease. In fact, a number of early
studies identified distinct miRNA expression profiles of pancreatic cancer and non-malignant
pancreatic diseases using frozen and FFPE tissues, leading to identification of key potential
diagnostic miRNAs.
A small number of candidate miRNAs, including miR-196a, miR-196b, miR-135b, miR-150,
miR-155, miR-21, miR-210, miR-221 and miR-222, have been reported as significantly elevated
in PDAC, with a much lesser contribution from those that are down-regulated (e.g., miR-130b,
miR-148a, miR-148b, miR-216, miR-217, and miR-96) [85-91] . Detection of some miRNAs has
been associated with different stages of pancreatic intraepithelial neoplasias (PanINs), the pre-
cursor lesions involved in the progression to PDAC. For example, expression of miR-196a was
shown to progressively increase with the PanIN grade, with peak production occurring at the
stage of PanIN-2 and -3 [92] . The closely related miR-196b was shown to be specifically over-
expressed in PanIN-3 lesions (carcinoma in situ ) as compared to lower-grade PanINs and nor-
mal pancreas [93] , while miR-148a was reported to be down-regulated earlier in the course of
pancreatic carcinogenesis (PanIN-1b) via hypermethylation of the DNA region encoding miR-
148a, and persists at that level through the lesion's progression into PDAC [94] .
Notably, miR-196a along with miR-217, a marker uniquely associated with normal pancre-
atic cells, became crucial components of the first miRNA-based LDT, the miR Inform Pancreas
test. This test was developed using FFPE specimens to measure the increased proportion of
ductal adenocarcinoma cells (indicated by up-regulation of miR-196a) relative to the decline
in the number of acinar cells observed in PDAC (resulting in reduced expression levels of
miR-217). The score was defined as the difference in expression (ΔCt) between miR-196a and
miR-217. The specimens were classified as either benign or PDAC according to a cutoff point
of 0.5 ΔCt, where ΔCt > 0.5 indicated a diagnostic negative (benign) and ΔCt ≤ 0.5 indicated
a diagnostic positive (PDAC). During the clinical validation that was performed in accord-
ance with CLIA and CAP guidelines, the miR Inform Pancreas test showed a high sensitivity
and specificity of 95% (95% CI: 76-100) and 95% (95% CI: 83-99), respectively, for diagnosis
of PDAC [14] . Although this test has a limited clinical utility because it is based on resected
tissue, it demonstrates the diagnostic capability of miRNAs confirmed by the gold standard:
histology on resected tissue.
5.3.1.2 Improving Diagnosis and Treatment of Cystic Lesions
Pancreatic cystic neoplasms are discovered more frequently because of the increase in the
use of high-resolution radiological imaging for the evaluation of abdominal trauma and other
disorders [95] . The prevalence of unsuspected, asymptomatic pancreatic cysts identified dur-
ing imaging can be as high as 2.6% [96] . Intraductal papillary mucinous neoplasms (IPMN) are
the most common cystic precursor lesions of PDAC. However, current pre-operative modali-
ties including imaging, endoscopy, and biochemical cyst fluid analysis fail to accurately assess
the cysts' grade of dysplasia and presence of invasion without resorting to resection and pre-
cise histopathologic examination. For example, CT and MRI imaging features have only a 50%
diagnostic accuracy. EUS can only differentiate mucinous from non-mucinous lesions in about
50% of the cases [97-100] , which is important because mucin is characteristic of lesions that
have malignant potential, such as IPMN and mucinous cystic neoplasms (MCN). Cytologic
evaluation of aspirated cystic fluid from FNA is often performed during EUS, but it has at best
50% overall accuracy, which decreases dramatically in smaller pancreatic cysts [98,101-103] .