Table 1.

Components of biomarker-based model validation

•Ascertainment of true outcome
•Proper model building, including internal validation
•Avoidance of overfitting in the model building
•Consideration of biomarker assay reproducibility
•External validation is crucial
•External validation necessary for determining generalizability
•Choosing appropriate performance measures dependent on outcome and prediction type
•Avoiding overuse of cut-points in risk prediction