Polygenic and multifactorial scores for pancreatic ductal adenocarcinoma risk prediction.
Galeotti AA., Gentiluomo M., Rizzato C., Obazee O., Neoptolemos JP., Pasquali C., Nentwich M., Cavestro GM., Pezzilli R., Greenhalf W., Holleczek B., Schroeder C., Schöttker B., Ivanauskas A., Ginocchi L., Key TJ., Hegyi P., Archibugi L., Darvasi E., Basso D., Sperti C., Bijlsma MF., Palmieri O., Hlavac V., Talar-Wojnarowska R., Mohelnikova-Duchonova B., Hackert T., Vashist Y., Strouhal O., van Laarhoven H., Tavano F., Lovecek M., Dervenis C., Izbéki F., Padoan A., Małecka-Panas E., Maiello E., Vanella G., Capurso G., Izbicki JR., Theodoropoulos GE., Jamroziak K., Katzke V., Kaaks R., Mambrini A., Papanikolaou IS., Szmola R., Szentesi A., Kupcinskas J., Bursi S., Costello E., Boggi U., Milanetto AC., Landi S., Gazouli M., Vodickova L., Soucek P., Gioffreda D., Gemignani F., Brenner H., Strobel O., Büchler M., Vodicka P., Paiella S., Canzian F., Campa D.
BACKGROUND: Most cases of pancreatic ductal adenocarcinoma (PDAC) are asymptomatic in early stages, and the disease is typically diagnosed in advanced phases, resulting in very high mortality. Tools to identify individuals at high risk of developing PDAC would be useful to improve chances of early detection. OBJECTIVE: We generated a polygenic risk score (PRS) for PDAC risk prediction, combining the effect of known risk SNPs, and carried out an exploratory analysis of a multifactorial score. METHODS: We tested the associations of the individual known risk SNPs on up to 2851 PDAC cases and 4810 controls of European origin from the PANcreatic Disease ReseArch (PANDoRA) consortium. Thirty risk SNPs were included in a PRS, which was computed on the subset of subjects that had 100% call rate, consisting of 839 cases and 2040 controls in PANDoRA and 6420 cases and 4889 controls from the previously published Pancreatic Cancer Cohort Consortium I-III and Pancreatic Cancer Case-Control Consortium genome-wide association studies. Additional exploratory multifactorial scores were constructed by complementing the genetic score with smoking and diabetes. RESULTS: The scores were associated with increased PDAC risk and reached high statistical significance (OR=2.70, 95% CI 1.99 to 3.68, p=2.54×10-10 highest vs lowest quintile of the weighted PRS, and OR=14.37, 95% CI 5.57 to 37.09, p=3.64×10-8, highest vs lowest quintile of the weighted multifactorial score). CONCLUSION: We found a highly significant association between a PRS and PDAC risk, which explains more than individual SNPs and is a step forward in the direction of the construction of a tool for risk stratification in the population.