Controlled Trial Comparing the Performance of 22 Gauge Versus 25 Gauge EUS-FNA Needles (FNA-22G-25G)

Overview

The aim of the study is to compare the performance characteristics of EUS-FNA 22 Gauge needle and EUS FNA 25 Gauge needle in terms of cellularity and diagnostic yield for diagnosis of various pathologies, including lymph nodes, pancreatic, luminal and other lesions outlined by EUS.

Tissue acquisition by Endoscopic ultrasound plays a central role in diagnosis of several pathologies; this is achieved by obtaining cytological material with fine needle aspiration (FNA) needles that are able to provide smears of malignant cells. Several needles are at present available including 19 Gauge, 22 Gauge and 25 Gauge needles. A 22 Gauge needle is at present the standard needle and most of the EUS-FNA results are based on FNA with this needle. Twenty five Gauge needles presently available on the market are distinctly small, highly flexible and may be easier to penetrate hard pancreatic tumors. There even is anecdotal messages that this size of needle may be able to traverse relatively large vascular structures without major risk. However, there is very scant data on whether a 25 G needle is comparable to a standard 22 G needle in terms of specimen cellularity and quality. There is only one small randomized trial published as an abstract by Lee et al reporting no statistical significant difference in cellularity between the two groups of needle size as judged by 2 different cytologists. However, the conclusion is dubious because considerable bias may be present since both needles were used in the same lesion. There is therefore a need for a properly designed randomized study comparing different needle sizes. This study is a prospective randomized multicenter study. Patients referred to one of the participating departments for an EUS examination will be included prospectively. Linear EUS examination will be done with a linear echoendoscope for lymph nodes, mediastinal, pancreatic and other intra-abdominal masses. Tumor and lymph node characteristics will be carefully described: size (long axis), echogeneity (hypoechoic, hyperechoic, mixed), and echostructure (homogenous, in-homogenous, calcifications), shape (round, oval, triangular), border (regular, irregular). EUS-guided cell sampling will be performed with a 22 G (Medi-Globe, Sonotip II) and 25 G (Medi-Globe, Sonotip II) needles under EUS and Colour-Doppler guidance. Needle order will be selected randomly, one single needle pr. Lesion. Three passes will be performed with either of the needles after randomization, using 6 uniform to and fro movements on every pass with continuous 10cc suction for the reason of standardization. The material will be expelled on separate numbered glass slides. The cytopathologist will review the material after staining blinded to the needle size and will be using standardized criteria for evaluation of cellularity. Doubtless, this multicenter randomized controlled trial comparison, as any other experimental method, needs with necessity a serious statistical evaluation, bringing the benefits of improving the reliability and credibility of the findings thus obtained. Accordingly, both exploratory data analysis and statistical inference will be performed. Technically, the expected number of lesions to be included in the study will be 220, 110 in each group. Block randomization (block size = 4) will be used to ensure exactly equal treatment numbers at certain equally spaced points in the sequence of patients assignments (Knuth shuffle algorithm for random permutations). For diagnostic tests the sensitivity, specificity positive and negative predictive value and diagnostic accuracy will be calculated and calculated separately for EUS-FNA during 1st pass, 2nd pass and 3rd pass by comparing the results with the final diagnosis. All results will be expressed as mean, standard deviation (SD) and confidence interval. Testing hypotheses will be used afterwards. An a priori power analysis will be performed to determine the appropriate sample size in order to achieve adequate power for the statistical tests subsequently used. The data from both groups will be compared by standard comparison tests chi-square test or Fishers exact test where appropriate. The level of statistical significance is stated as 0.05.