Biomarker Evaluation and Clinical Development

  • Melissa Assel Memorial Sloan Kettering Cancer Center
  • Andrew Vickers Memorial Sloan Kettering Cancer Center
Keywords: Biomarkers, Prediction Modeling, Prostate cancer, clinical utility, Decision Analysis, Discriminiation, Calibration, Net Benefit


Most candidate biomarkers are never adopted into clinical practice. The likelihood that a biomarker with good predictive properties will be incorporated into urologic decision-making and will improve patient care can be enhanced by following established principles of biomarker development. Studies should follow the REMARK guidelines, should have clinically relevant outcomes, and should evaluate the biomarker on the same patients to whom the biomarker would be applied in practice. It is also important to recognize that biomarker research is comparative: the question is not whether a biomarker provides information, but whether it provides better information than is already available. Continuous biomarkers should not be categorized above or below a fixed cutpoint: risk prediction allows for individualization of care. The risk predictions must be calibrated, that is, close to a patient’s true risk, and decision analysis is required to determine whether using the biomarker in clinical practice would change decisions and improve outcomes. Finally, impact studies are needed to evaluate how use of the biomarker in the real world affects outcomes.

How to Cite
Assel, M., & Vickers, A. (2020). Biomarker Evaluation and Clinical Development. Société Internationale d’Urologie Journal, 1(1), 16-22. Retrieved from
Molecular Biomarkers in Urologic Oncology