SAMPLE-SIZE DETERMINATION FOR DIAGNOSTIC-ACCURACY STUDIES INVOLVING BINORMAL ROC CURVE INDEXES

Receiver operating characteristic (ROC) curves and their associated in dices are valuable tools for the assessment of the accuracy of diagnos tic tests. The area under the ROC curve is a popular summary measure o f the accuracy of a test. The full area under the ROC curve, however, has been criticized because it gives equal weight to all false positiv e error rates. Alternative indices include the area under the ROC curv e in a particular range of false positive rates ('partial' area) and t he sensitivity of the test for a single fixed false positive rate (FPR ). We present a unified approach for computing sample size for binorma l ROC curves and their indices. Our method uses Taylor series expansio ns to derive approximate large-sample estimates of the variance and co variance of binormal ROC curve parameters. Several examples from diagn ostic radiology illustrate the proposed method. (C) 1997 by John Wiley & Sons, Ltd.