Testing the test
When a trial is enriched by a response predictor such as a pharmacogenomic (PG) marker, FDA may require the sponsor to demonstrate the clinical utility of the marker. Such a prospective efficacy enrichment design would need to give an independent test of the specificity and sensitivity of the PG test. By comparing the "PG tested" and "Not PG tested" arms, the clinical value of the marker is evaluated. By including participants who do not receive the PG test, the design addresses the possibility of false negatives and/or cases in which the toxicity must be evaluated in nonresponders because pretreatment selection is impossible. Source: adapted from a slide presented by Robert Temple, director of FDA's Office of Medical Policy