2.6. Sample size calculation and statistical analysis

We used an accurate binomial test to calculate the required sample size. According to the previous randomized control studies and meta-analyses, the threshold for the histological accuracy rate with a standard FNA or FNB needle with a reverse bevel was 77%. We assumed the diagnostic accuracy of histology with this needle to be 90%. To achieve a statistical power of 0.80 for exact binomial tests with a 2-sided type-I error of 0.05, our minimum sample size was 60 patients. Since we expected some patients to be nonevaluable, we considered enrolling 70 patients.

We expressed out results as numbers and percentages for categorical variables or as medians and ranges for continuous variables. For the analysis of overall diagnostic yield, Clopper & Peason exact confidence interval was used for binary outcome. We considered a P < .05 as statistically significant. We used the SAS software version 9.4 (SAS Institute, Cary, NC) to perform all the statistical analyses. The final clinical data obtained with follow-ups until March 2019.