A prognostic biomarker is one that indicates an increased (or decreased) likelihood of a future clinical event, disease recurrence or progression in an identified population. Prognostic biomarkers are measured at a defined baseline, which may include a background treatment. Many familiar examples of prognostic biomarkers occur in clinical contexts where an individual is diagnosed with a disease or condition and there is interest in assessing the likelihood of a future clinical event. Examples of future events include death, disease progression, disease recurrence, or development of a new medical condition. In oncology, biomarkers such as tumor size, number of lymph nodes positive for tumor cells, and presence of metastasis have traditionally been used to indicate prognosis. Increasingly, molecular indicators or signatures measured on tumors are being used in lieu of, or in addition to, these clinicopathologic characteristics. For patients who have previously suffered a heart attack, elevated blood pressure, evidence of diabetes, elevated LDL cholesterol, and low HDL cholesterol are examples of biomarkers that indicate an increased risk for another heart attack. For individuals with hypertension, concomitant evidence of diabetes is associated with an increased likelihood of cardiovascular events. The prognostic biomarker’s association with outcome is present without reference to different interventions (i.e., predicts increased likelihood of an event without an intervention). However, the presence or strength of a prognostic association may vary depending on the specific clinical setting (e.g., background therapy, stage of disease) and particular endpoint of interest, so it is important that prognostic biomarkers be described in the proper context.

Prognostic biomarkers are often used as eligibility criteria in clinical trials to identify patients who are more likely to have clinical events or disease progression. Thus, they are widely used as enrichment factors in drug development (U.S. Food and Drug Administration 2012). Many clinical trials of medical interventions have as their endpoint either an event rate or time-to-event. The statistical power for a time-to-event endpoint to assess treatment effect in a controlled clinical trial is driven by the planned effect size (i.e., hazard ratio for a time to event endpoint) and the planned number of events. Enrichment with patients who have a higher likelihood of experiencing an event will therefore increase statistical power. Analogous to this situation is the use of susceptibility/risk biomarkers for enrichment of prevention trial populations. In a treatment setting, prognostic biomarkers can contribute to decisions about whether or how aggressively to intervene with the treatment.

The term prognostic has not been used consistently in the biomedical community. Some have applied the term only in the clinical context of individuals who have already been diagnosed with a disease or other medical condition. Others would include among prognostic biomarkers those that indicate, for apparently healthy individuals, the likelihood of a future diagnosis or disease. This document makes a distinction between prognostic biomarker and susceptibility/risk biomarker, the latter defined here as applying to individuals without clinically apparent disease (or medical condition).