Sample size for Comparison of two ROC curves

Determine the required sample size for a test comparing the areas under two ROC curves (AUC). The test assesses whether the diagnostic accuracy of two tests differs significantly when both are applied to the same subjects. When the two tests are applied to independent groups rather than paired subjects, enter 0 for both correlation coefficients.

Input
Type I error - Alpha:
Type II error - Beta:

Area under ROC curve 1:
Area under ROC curve 2:
Correlation in positive group:
Correlation in negative group:
Ratio of sample sizes in negative/positive groups:
Result
Number of cases required in positive group:
Number of cases required in negative group:
Total sample size (2 groups together):

For details, see the MedCalc manual: Sample size: Comparison of two ROC curves.

How to cite this page

MedCalc Software Ltd. Sample size for Comparison of two ROC curves. https://www.medcalc.org/en/calc/sample-size-comparison-of-2-roc-curves.php (Version 24.0; accessed October 12, 2026)

Trusted by researchers worldwide — download the full MedCalc statistical software and perform advanced statistical analysis in seconds. No registration required.