Hsieh MH, et al. - Given that a higher risk of pancreatic cancer has been suggested for patients with type 2 diabetes mellitus (T2DM), researchers used two models to predict pancreatic cancer risk in this study population. They used data obtained from the National Health Insurance Research Database of Taiwan. The prediction models, including the available possible risk factors for pancreatic cancer, were used. As training set and as the test set, 97.5% and 2.5% of the data were used, respectively. Using Python (Version 3.7.0), logistic regression (LR), and artificial neural network (ANN), models were implemented. The LR and the ANN models were compared with respect to the F1, precision, and recall. Findings revealed that a more accurate prediction of pancreatic cancer was enabled by LR model vs the ANN model. LR model appeared to be a good fit with 0.727 area under the receiver operating characteristic (ROC) curve in the prediction of pancreatic cancer.