Abstract
We consider evaluation and comparison of the diagnostic accuracy of biomarkers with continuous test outcomes, possibly correlated due to repeated measurements. We develop nonparametric group sequential testing procedures to evaluate and compare the area of biomarkers under their receiver operating characteristic curves, with either independent or paired test outcomes. These procedures rely on the construction of a two-dimensional statistic of Whitehead (Statist. Med. 1999; 18:2271-2286) so that design methods based on Brownian motion can be applied.