Abstract:

In diagnostic radiology, the multireader multicase (MRMC) design and the free-response receiver operating characteristics (FROC) method are often used in combination. The cross-correlated data generated by the MRMC-FROC study leads to difficulties in the corresponding analysis, and the need to include covariates in the model further complicates the subsequent analysis. In this paper, we propose a regression approach based on three new measures with good interpretability. The correlation structure of the original test results is taken directly into account in the estimation procedure. The proposed method also allows the inclusion of continuous or discrete covariates. Consistent and asymptotically normal estimators are derived for the new measures. Simulation studies are conducted to evaluate the performance of the proposed approach. The simulation results show that the regression approach performs well under a wide range of scenarios. We also apply the proposed regression approach to a diagnostic study of computer-aided diagnosis in lung cancer.

(Link: https://onlinelibrary.wiley.com/doi/10.1002/sim.9400)

Papers Published in March and April 2022：

[1] Equivalence tests for ratio of means in bioequivalence studies under crossover design. He Y, Deng Y, You C, Zhou X-H. Statistical Methods in Medical Research. April 2022. doi:10.1177/09622802221093721.

[2] Liu, X, Sun, J, Zhou, X-H. A novel regression method for the analysis of multireader multicase-free-response receiver operating characteristics studies. Statistics in Medicine. 2022; 1- 17. doi:10.1002/sim.9400

[3] Dong R, Hu T, Zhang Y, Li Y, Zhou XH. Assessing the Transmissibility of the New SARS-CoV-2 Variants: From Delta to Omicron. Vaccines (Basel). 2022 Mar 24;10(4):496. doi: 10.3390/vaccines10040496. PMID: 35455246; PMCID: PMC9026126.

[4] Yang, Bing et al. “An Asia-specific variant of human IgG1 represses colorectal tumorigenesis by shaping the tumor microenvironment.” The Journal of clinical investigation vol. 132,6 (2022): e153454. doi:10.1172/JCI153454