Impact of misclassification in genomic enrichment designs.

Stefano Vezzoli,
Biostatistician CROS NT

Diagnostic assays needed to assess genomic characteristics are prone to misclassification error. We will evaluate the impact of the misclassification in the simplified setting of a non-adaptive genomic enrichment design in terms of Type I error rate and power. We will show that misclassification leads to a smaller statistical power for showing superiority and to an increase of Type 1 error probability in non-inferiority trials. In an adaptive setting, the misclassification does not affect the inference on the unselected population; however, the results obtained highlight that assay inaccuracy may invalidate the inference about the marker. 10.1002/sim.5541).