Study
PLCO
(Learn more about this study)
Project ID
2017-9002
Initial CDAS Request Approval
Jul 27, 2017
Title
Circulating miRNAs as Biomarkers for Pancreatic Cancer Early Detection and Risk Assessment
Summary
Understanding the etiology and identifying markers for early detection are a top priority for reducing pancreatic cancer incidence and mortality. MicroRNAs (miRNAs), small, noncoding ribonucleic acids that regulate gene expression, plays a critical role in carcinogenesis and cancer progression. In 2014, a JAMA publication reported the discovery of 2 blood miRNA-based diagnostic indices capable of distinguishing pancreatic cancer patients from healthy controls and chronic pancreatitis patients (AUC: 0.83 to 0.75). Although these findings are very promising, the study was limited by using only healthy young controls, measuring miRNAs in whole blood, and including very few early-stage cases. Thus, the results need to be replicated in more rigorously designed studies. Furthermore, the utility of these indices and other circulating miRNAs for early detection and risk assessment of pancreatic cancer across different ethnic populations needs to be evaluated. Recently, we completed a pilot study in which we measured 800 miRNAs in pre-diagnosis plasma samples from 185 pancreatic cancer cases and the same number of individually-matched controls selected from the PLCO participants. All cases were diagnosed within 5 years after blood donation. We found a significant case-control difference for 54 miRNAs (P<0.05 based on paired test). Of these, 16 retained significance after FDR adjustment; three had an FDR-adjusted P-value less than 1x E-5 (one of the miRNAs was included in the JAMA-reported diagnosis indices). All 16 FDR-adjusted significant miRNAs were significantly associated with pancreatic cancer risk among cases diagnosed less than 2 years and 2-4.99 years after blood draw; 9 miRNAs showed a slightly stronger association in the former group. The Ingenuity Pathway Analysis (IPA) pathway analysis revealed a clear pattern of over-representation of cancer-related pathway/genes in the targets of almost all top 16 miRNAs. Of note, several significantly enriched signaling pathways, such as HGF Signaling, ERK5 and Estrogen-mediated S-phase Entry, were observed for multiple top miRNAs, indicating that those miRNAs may contribute to etiology or progression of pancreatic cancer via similar biological pathways. The results of our pilot study not only show the substantial potential for miRNAs to serve as biomarkers for pancreatic cancer early detection and risk assessment, but also for understanding underlying biological mechanisms. To validate these promising findings and to evaluate the time window during which the discriminative/predictive miRNAs are most informative for cancer early detection and/or risk assessment, we propose to conduct a comprehensive follow-up study that will utilize resources from an additional four prospective cohort studies, i.e., pancreatic cancer cases and controls from the Southern Community Cohort Study, the Multi-ethnic Cohort Study, the Shanghai Women’s Health Study, and the Shanghai Men’s Health Study. We will also include in this study any PLCO pancreatic cancer cases that were not included in our pilot study and those who had repeated pre-diagnosis plama samples. Individual matched controls will be selected from the PLCO. Circulating miRNA in pre-diagnosis plasma samples will be measured using the NanoString nCounter Human V2 miRNA Expression Assay. A nested case-control study design will be employed.
Aims
Specific aims for the study are:
1). To validate promising pancreatic cancer-related miRNAs found in the pilot study using additional PLCO resources, i.e., pancreatic cancer cases not included in the pilot study and matched controls, and pancreatic cancer cases and matched controls from the SCCS, MEC, SWHS and SMHS.
Premise and hypotheses to be tested: 1). Pancreatic cancer-related miRNA changes are released into the bloodstream and are detectable years before the manifestation of clinical symptoms. 2). Aberrant expression or/and function of miRNAs have been found in a broad range of human diseases/conditions, including in conditions that are themselves suspected risk factors of pancreatic cancer, such as infectious diseases, chronic inflammatory diseases, diabetes, and metabolic disorders. Circulating levels of miRNAs thus may serve as risk assessment markers for pancreatic cancer. 3). The cancer biology of humans is similar across racial/ethnic groups. miRNAs related to the initiation and progression of pancreatic cancer are robust and can serve as biomarkers for cancer early detection and risk assessment across different ethnic populations.
2). To evaluate the time window during which levels of circulating miRNAs can serve as biomarkers for early diagnosis and/or are most predictive of the risk for pancreatic cancer. Repeated blood samples from the PLCO will be used to define the most informative time window for cancer early detection and risk assessment. These findings will be validated in the SCCS, MEC, SWHS and SMHS by time window analyses.
Premise and hypotheses to be tested: 1). miRNAs secreted by cancer cells increase with progression of the disease and their associations with disease risk are strengthened as measurements are taken closer to the cancer diagnosis; 2). miRNAs related to disease development may exist many years prior to cancer diagnosis and their associations with pancreatic cancer risk do not increase appreciably when measurements are taken closer to the cancer diagnosis.
3). To evaluate whether adding the confirmed miRNAs to the existing pancreatic predictive model improves risk prediction.
Premise and hypothesis to be tested: the addition of disease-predictive biomarkers will increase the sensitivity and specificity of a risk prediction model.
Collaborators
Xiao Ou Shu (Vanderbilt University)
Veronica W Setiawan (University of Southern California)
