*4.8. Statistical Analysis*

Data normality was assessed using the Shapiro–Wilk test [138]. Since our experimental data did not follow a normal distribution, microRNA levels were compared between groups using the Kruskal–Wallis one-way analysis of variance with post-hoc test for the comparison among multiple groups. The significance level was established at a *p*-value of *p* < 0.05.

Receivers operating characteristic (ROC) curves were constructed to calculate the area under the curve (AUC) and the best cut-off point for particular microRNA was used in order to calculate the respective sensitivity at 90.0% specificity (MedCalc Software bvba, Ostend, Belgium). For every possible threshold or cut-off value, the MedCalc ® v16.8.4 program reports the sensitivity, specificity, likelihood ratio positive (LR+), likelihood ratio negative (LR −).

To select the optimal combinations of microRNA biomarkers logistic regression was used (MedCalc ® v16.8.4 program, MedCalc Software bvba, Ostend, Belgium). The logistic regression procedure allows to analyze the relationship between one dichotomous dependent variable and one or more independent variables. Another method to evaluate the logistic regression model makes use of ROC curve analysis. In this analysis, the power of the model's predicted values to discriminate between positive and negative cases is quantified by the area under the ROC curve (AUC). To perform a full ROC curve analysis the predicted probabilities are first saved and next used as a new variable in ROC curve analysis. The dependent variable used in logistic regression then acts as the classification variable in the ROC curve analysis dialog box.

Correlation between variables was calculated using the Spearman's rank correlation coefficient (ρ). Spearman's rank correlation coefficient, a nonparametric measure of rank correlation, assesses how well the relationship between two variables can be described using a monotonic function.

If the correlation coefficient value ranges within <0.5; 1.0>, there is a strong positive correlation. The significance level was established at a *p*-value of *p* < 0.05.

Box plots encompassing the median (dark horizontal line) of log-normalized gene expression values for particular microRNAs were generated using Statistica software (version 9.0; StatSoft, Inc., Tulsa, OK, USA). The upper and lower limits of the boxes represent the 75th and 25th percentiles, respectively. The upper and lower whiskers indicate the maximum and minimum values that are no

more than 1.5 times the span of the interquartile range (range of the values between the 25th and the 75th percentiles). Outliers are marked by circles and extremes by asterisks.
