A Dunnett-Type Test and Its Sample Size Calculation for Comparing K ROC Curves with a Control

Area of Science:

• Biostatistics
• Medical Diagnostics
• Biomarker Discovery

Background:

• Diagnostic biomarkers are crucial for distinguishing between malignant and benign conditions.
• Evaluating continuous biomarkers involves assessing their performance using the Area Under the Curve (AUC) of the receiver operating characteristic curve.
• Comparing multiple experimental biomarkers against a control introduces multiplicity issues.

Purpose of the Study:

• To propose a non-parametric statistical testing procedure for comparing K experimental biomarkers against a single control.
• To develop a sample size calculation method for this comparison, accounting for multiplicity.
• To evaluate the performance of the proposed method through simulations.

Main Methods:

• A novel non-parametric statistical test is developed to compare K experimental biomarkers with a control.
• A sample size formula is derived, incorporating AUC values, correlation coefficients, disease prevalence, type I error rate, and statistical power.
• Simulations are used to assess the accuracy of the type I error rate control and the proposed sample size calculation.

Main Results:

• The proposed statistical test accurately controls the overall type I error rate.
• The sample size calculation method effectively maintains the specified statistical power.
• The method provides a robust approach for biomarker comparison in diagnostic studies.

Conclusions:

• The developed statistical procedure offers a reliable method for comparing multiple diagnostic biomarkers against a control.
• The sample size calculation ensures adequate power for detecting significant differences.
• This work contributes to the rigorous evaluation of diagnostic biomarkers.