Loading
Analytical validation shows a test measures what it claims, reliably; clinical validation shows the measurement actually predicts the patient outcome it is meant to.
A biomarker in medicine is a measurable indicator of the presence or severity of a disease state (Wikipedia). Before it is used, a test must show analytical validity (accuracy, precision, reproducibility of the measurement itself) and clinical validity (that the result relates to the clinical condition or outcome in the intended population), and ideally clinical utility, that acting on it improves outcomes. For an AI model the same ladder applies: reproducible outputs on fixed inputs, then discrimination and calibration on external cohorts, then a trial of using it.
Shares Research use only (RUO), Software as a medical device (SaMD), Cancer AI vocabulary (CanSim terms map) and the tag cansim-terms.
Shares Calibration: reliability diagrams and the Brier score, Biomarker, Cancer AI vocabulary (CanSim terms map) and the tag cansim-terms.
Shares Calibration: reliability diagrams and the Brier score, Cancer AI vocabulary (CanSim terms map) and the tag cansim-terms.
Shares Calibration: reliability diagrams and the Brier score, Cancer AI vocabulary (CanSim terms map) and the tag cansim-terms.
Shares External validation, Cancer AI vocabulary (CanSim terms map) and the tag cansim-terms.
Shares External validation, Cancer AI vocabulary (CanSim terms map) and the tag cansim-terms.
Shares External validation, Cancer AI vocabulary (CanSim terms map) and the tag cansim-terms.
Shares External validation, Cancer AI vocabulary (CanSim terms map) and the tag cansim-terms.