Sample size calculation is an important factor to evaluate the reliability of the diagnostic test. In this paper, a case study of the clinical diagnostic test of artificial intelligence for identification of liver contrast-enhanced ultrasound was performed to conduct two-category and multi-categories studies. Based on sensitivity and specificity, the sample size was then estimated in combination with the statistical characteristics of disease incidence, test level and one/two-sided test. Eventually, the sample size was corrected by integrating the factors of the proportion of training/test dataset and the dropout rate of cases in the medical image recognition system. Moreover, the application of Sample Size Calculator, MedCalc, PASS, and other software can accelerate sample size calculation and reduce the amount of labor.
Stepped wedge cluster randomized trials (SW-CRT) is a kind of cluster randomized controlled trial mainly applied in the field of public health policy that has emerged in recent years, which has gradually attracted the attention of workers in the field of health and wellness. At present, this trial method is not widely used at home and abroad, and there are various ways of sample size calculation and statistical analysis. This paper describes the principles, categories, and differences between SW-CRT and traditional randomized controlled trials, and outlines sample size calculation and statistical analysis methods. In general, SW-CRT is characterized by clustering, cross-design, and measurement of results at multiple time points. In terms of sample size calculation, it is necessary to distinguish between clusters with the same and different sizes, and commonly used sample size calculation procedures can be implemented in Stata, R, and SAS software, as well as in fixed online websites, including the "Steppedwedge" program, the "swCRTdesign" program, the "Swdpwr" program, the "CRTpowerdist" program, and the "Shiny CRT Calculator" tool and so on. Based on the design characteristics of SW-CRT, the researcher should also consider the confounding factors of time effects and repeated measurements of result. Therefore, the statistical analysis methods are often based on generalized linear mixed model (GLMM) and generalized estimating equations (GEE). However, most of the above models have been proposed based on cross-sectional studies, there is a lack of statistical methods for queue design and SW-CRT with transitional period now, and more comprehensive methodological exploration is still needed in the future.