Prospective comparative study of the EUS guided 25‐gauge FNA needle with the 19‐gauge Trucut needle and 22‐gauge FNA needle in patients with solid pancreatic masses

Abstract
Background and Study Aims: The aim of this prospective study was to compare fine‐needle aspiration guided by endoscopic ultrasonography (EUS‐FNA) using 25‐gauge and 22‐gauge needles with the EUS‐guided 19‐gauge Trucut needle biopsy (EUS‐TNB) in patients with solid pancreatic mass. Patients and Methods: Twenty‐four consecutive patients with pancreatic mass underwent biopsies by both EUS‐FNA and EUS‐TNB. Three needles were compared with respect to technical success rate, tissue size obtained, overall diagnostic accuracy and accuracy for histological and cytological diagnosis. Results: The 25‐gauge EUS‐FNA was technically easier and obtained superior overall diagnostic accuracy than the 22‐gauge and Trucut needles, especially in lesions of the pancreas head and uncinate process. Overall accuracy for the 25‐gauge, 22‐gauge and Trucut needle was 91.7%, 79.7% and 54.1%, respectively. Accuracy for cytological diagnosis irrespective the site of lesions with 25‐gauge, 22‐gauge and Trucut needles was 91.7%, 75.0%, and 45.8%, respectively. For uncinate masses, it was 100%, 33.3%, and 0.0%, respectively. These differences were significant. Among technically successful patients, the accuracy for histological diagnosis using the 25‐gauge was significantly inferior (P < 0.05) to 22‐gauge and Trucut needles and the rates were 45.8%, 78.9% and 83.3%. Conclusions: The 25‐gauge FNA needle was significantly superior in terms of technical success rate and overall diagnostic accuracy, especially for the head and uncinate lesions, compared to the 22‐gauge and Trucut needles and could be considered ‘the best choice needle for cytological diagnosis’ of solid pancreatic lesions. If histological diagnosis is required, the 22‐gauge FNA needle and Trucut needle may be advantageous for use in head/uncinate and body/tail lesions, respectively.

This publication has 23 references indexed in Scilit: