Abstract Pancreatic ductal adenocarcinoma (PDAC) is the most lethal among solid tumors. Only a subset of patients benefit from chemotherapy, and identification of novel prognostic factors is warranted. Recently, miRNAs emerged as a pivotal class of regulators for multiple genes involved in cancer, including PDAC (Giovannetti et al, Cancer Res 2010), and miRNA expression pattern classified tumors better than mRNA data (Lu et al, Nature 2005), suggesting their use for diagnostic purposes. However, high-throughput arrays detecting more than 1000 miRNAs are opening new opportunities to evaluate whether their profiles can also predict clinical outcome. Therefore, the present study evaluated whether comprehensive miRNA expression profiling by microarray technology can predict overall survival (OS) in resected PDAC patients. High-resolution miRNA profiles were obtained with the Toray's 3D-GeneTM miRNA chip, detecting 1200 types of human miRNA. RNA was isolated from formaldehyde fixed paraffin embedded primary tumors of 26 stage-pT3N1Mx homogeneously treated patients (gemcitabine 1000 mg/m2/day on days 1/8/15, every 28 days), selected according to their outcome (OS <12 months vs. OS >30 months). Microarray slides of the 19 samples that passed RNA quality check were scanned using 3D-GeneTM Scanner 3000 with appropriate photomultiplier settings for red channel. Each microarray was scanned three times, and then merged into one dataset analyzed. Highly stringent statistics included t-test, Spearman ranked correlation to generate a distance matrix, and iterative approaches. Unsupervised hierarchical analysis revealed that the PDAC specimens clustered according to their long/short OS classification, while the feature selection algorithm RELIEF identified the top 10 discriminating miRNAs between the two groups. MiR211 emerged as the best discriminating miRNA, with significantly higher expression in long vs. short OS patients. The expression of this miRNA was subsequently assessed by quantitative RT-PCR in an independent cohort of laser microdissected tumors from 60 resected stage-pT3N1Mx PDAC patients treated with the same gemcitabine regimen. Patients with low miR211 expression, according to median value (12.8 a.u. calculated as ΔCt vs. RNU6), had a significantly shorter median OS (14.8, 95%CI=13.1-16.5, vs. 25.7 months, 95%CI=16.2-35.1, log-rank P=0.004). Multivariate analysis demonstrated that low miR211 expression was an independent factor of poor prognosis for OS (hazard ratio 2.3, P=0.03) after adjusting for all the factors influencing clinical outcome. In conclusion, through comprehensive microarray analysis and PCR validation we identified miR211 as prognostic factor in resected PDAC. This miRNA was already associated with melanoma invasiveness, but our results prompt further prospective studies as well as research on miR211 biological role in PDAC. Citation Format: {Authors}. {Abstract title} [abstract]. In: Proceedings of the 103rd Annual Meeting of the American Association for Cancer Research; 2012 Mar 31-Apr 4; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2012;72(8 Suppl):Abstract nr 186. doi:1538-7445.AM2012-186
Read full abstract