*2.6. Statistical Analyses*

SPSS 26.0 (IBM, Armonk, New York, NY, USA) and GraphPad Prism 9.0 (GraphPad Software Inc., San Diego, CA, USA) were used for statistical analysis and figure production, respectively. For all variables, the Kolmogorov–Smirnov normal distribution test was performed first. Continuous variables with a normal distribution were expressed as mean ± SD. Continuous variables that were not normally distributed were expressed as median and interquartile intervals. Categorical variables were expressed as numbers and percentages. The Kappa consistency analysis was used to assess the two physicians' subjective score, and the Wilcoxon signed-ranks test was used to evaluate the image conspicuity difference. The sensitivities, specificities, positive predictive values (PPVs), negative predictive values (NPVs), and accuracies of PET/MR, MR, and PET/CT for diagnosis were determined in accordance with the gold standard. According to the R × C chi-square test and the fisher exact probability method, the difference in diagnostic performance of PET/MR, MR, PET/CT was tested. The Wilcoxon signed-rank test was used to test the difference of PET/MR and PET/CT parameters (SUVmax, SUVmean, and lymph node diameter) for the same lymph nodes subgroup. The independent-samples Mann–Whitney U test was used to analyze the differences of these parameters between malignant and benign lymph nodes. Spearman's correlation method was used to analyze the correlation between PET/MR and PET/CT parameters and calculate the coefficient of determination (r2). Bland–Altman analysis was used to compare the two techniques.
