Do Polygenic Risk Scores Add to Clinical Data in Predicting Pancreatic Cancer? A Scoping Review

Abstract

Abstract Background: Polygenic risk scores (PRS) summarize an individual's germline genetic risk, but it is unclear whether PRS offer independent information for pancreatic cancer risk prediction beyond routine clinical data. Methods: We searched 8 databases from database inception to March 10, 2023 to identify studies evaluating the independent performance of pancreatic cancer–specific PRS for pancreatic cancer beyond clinical risk factors. Results: Twenty-one studies examined associations between a pancreatic cancer–specific PRS and pancreatic cancer. Seven studies evaluated risk factors beyond age and sex. Three studies evaluated the change in discrimination associated with the addition of PRS to routine risk factors and reported improvements (AUCs: 0.715 to 0.745; AUC 0.791 to 0.830; AUC from 0.694 to 0.711). Limitations to clinical applicability included using source populations younger/healthier than those at risk for pancreatic cancer (n = 10), exclusively of European ancestry (n = 13), or controls without relevant exposures (n = 1). Conclusions: While most studies of pancreatic cancer–specific PRS did not evaluate the independent discrimination of PRS for pancreatic cancer beyond routine risk factors, three that did showed improvements in discrimination. Impact: For pancreatic cancer PRS to be clinically useful, they must demonstrate substantial improvements in discrimination beyond established risk factors, apply to diverse ancestral populations representative of those at risk for pancreatic cancer, and use appropriate controls.