5. Statistical analysis

Statistical analysis was performed using MedCal ver. 18.2.1 (Med-Calc, Mariakerke, Belgium) and IBM SPSS ver. 22.0 (IBM Corp., Armonk, NY, USA). Intraclass correlation coefficients (ICCs) were used to assess the reliability index in intrarater and interrater reliability analyses. Two-way random effects, consistency and absolute agreement, and a multiple-raters model were used in this study based on a mean-rating (k=5). Koo and Li [11] suggested that ICC values less than 0.5 indicate poor reliability, values between 0.5 and 0.75 indicate moderate reliability, values between 0.75 and 0.9 indicate good reliability, and values greater than 0.90 indicate excellent reliability. The four methods were compared using the Tukey test. Bland and Altman analysis was used to determine the systematic biases when the differences between two methods were plotted against the average of two methods. Passing-Bablok regression was performed to describe a linear regression with no special assumptions referring to the distribution of the samples and the measurement error. The slope was calculated with a 95% confidence interval (CI). The Spearman rank correlation coefficient was used to measure the strength of the monotonic relationship between paired data. The interpretation of correlation coefficients was assessed using five levels: very weak (0.00<r≤0.19), weak (0.20<r≤0.39), moderate (0.40<r≤0.59), strong (0.60<r≤0.79), and very strong (0.80<r≤1.00). Mountain plots were generated by calculating the percentile of each ranked difference variable for the four staining methods.