Table 1.

Using multivariable methods, logistic regression, two proof-of-concept predictive models were developed and validated using leave-one-out cross validation (LOOCV), one for CAC progression and one for systolic blood pressure. Each model was trained and tested 46 times, resulting in all patients belonging to the training sample 45 times and the test sample once. An example is seen in the Table

	Patient 1	Patient 2	Patient 3	Patient 45	Patient 46
Model 1	Train	Train	Train	Train	Test
Model 2	Train	Train	Train	Test	Train
Model 3	Train	Train	Train	Train	Train
Model 45	Train	Test	Train	Train	Train
Model 46	Test	Train	Train	Train	Train