Three-Way Mixed Effect ANOVA to Estimate MRMC Limits of Agreement

When evaluating the clinical performance of a medical imaging device, multi-reader multi-case (MRMC) analysis is usually applied to account for both case and reader variability. For task that equates quantitative measurement, an agreement such as limits (LOA) method can be used compare different measurement methods. In this work, we introduce four types comparisons; these differ depending on whether measurements are within or between readers modalities. A three-way mixed effect ANOVA model estimate variances individual differences, which essential step estimating LOA. To verify estimates LOA, propose hierarchical simulate MRMC data. Two simulation studies were conducted validate LOA variance estimates. From results, conclude our unbiased, uncertainty estimation drops number cases increases rises value true increases.