Endoscopy 2019; 51(09): 811-812
DOI: 10.1055/a-0967-1829
© Georg Thieme Verlag KG Stuttgart · New York

Which needle and technique should we use for endoscopic ultrasound-guided liver biopsy? A work in progress

Referring to Hasan MK et al. p. 818–824
Alberto Larghi
Unita’ di Endoscopia Digestiva Chirurgica, Fondazione Policlinico Universitario A. Gemelli, IRCCS, Rome, Italy
› Institutsangaben
Weitere Informationen


28. August 2019 (online)

Endoscopic ultrasound-guided liver biopsy (EUS-LB) is gaining increasing popularity for evaluating patients with abnormal cholestatic liver parameters who have been referred to EUS to rule out extrahepatic biliary obstruction [1]. Liver parenchyma sampling has always been performed with the intent of obtaining histological samples, which need to be at least 20 mm long and contain at least 11 complete portal tracts (CPTs) for an accurate evaluation [1]. Initial experience with the Tru-Cut biopsy needle was unsatisfactory, while subsequent utilization of standard 19-G fine needle aspiration (FNA) and specifically designed 19-G fine needle biopsy (FNB) needles demonstrated a diagnostic accuracy ranging from 85 % to 100 % [1]. The only potential limitation with large diameter needles could be that community-based endosonographers, who are unfamiliar with their use, would be reluctant to utilize them to transduodenally sample the right liver lobe. Thus, a needle that satisfied both the experts and the individual endosonographers working outside centers of expertise would be highly valuable to facilitate the widespread use of EUS-LB.

“[...] more studies with improved designs are needed to establish the most efficient needle(s) and sampling technique for performing EUS-guided liver biopsy.”

In this issue of Endoscopy, Hasan et al. [2] report the results of a prospective evaluation of the performance of a 22-G FNB needle (Acquire; Boston Scientific) in 40 patients with abnormal liver chemistries. Bilobar EUS-LB, utilizing the fanning technique and with limited elevator use to prevent tissue fragmentation, was done after bile duct obstruction had been excluded. The stylet was slowly re-inserted while the assembly was kept straight to avoid further fragmentation. The endosonographers performed two passes in the left lobe and one in the right, as well as, a visual core tissue assessment. No suction was applied in the first two passes, but the protocol allowed suction during the third pass in cases where no visible cores had been retrieved. Visual core adequacy was obtained in 119/120 passes (99.2 %) and no patient required suction during the third pass. The retrieved samples from all patients had sufficient length and contained adequate numbers of CPTs for accurate histopathological evaluation. Nonalcoholic fatty liver disease with fibrosis was the most frequent diagnosis. Moderate post-processing tissue distortion, with crush and compression of CPT structures because of the small needle size, was observed. Six patients (15 %) reported mild self-limiting post-procedural pain; one patient required hospitalization to rule out myocardial infarction; another patient developed self-limiting fever. Finally, an additional patient died within 24 hours; this was most probably unrelated to EUS-LB [2]. The authors concluded that the use of EUS-LB with a 22-G FNB needle was safe and highly accurate in patients with abnormal liver chemistry and in whom biliary obstruction had been ruled out. Therefore, they proposed the use of the 22-G FNB needle, together with the technique that avoided sample fragmentation, as an alternative to the 19-G needle, but emphasized the need for a direct comparison between the devices in order to draw definitive conclusions [2].

In a recent randomized crossover study by Mok et al., bilobar EUS-FNB with a fanning technique was performed to compare a standard 19-G FNA needle (Expect Flexible 19G; Boston Scientific) versus a 22-G FNB needle (SharkCore; Medtronic) in the same patient [3]. Endosonographers primed the needles with heparin and applied 20-mL suction. The samples acquired with the 19-G needle were significantly more adequate (88 % vs. 68 %, P = 0.03), longer (9.1 mm vs. 6.6 mm, P = 0.02), and less fragmented (20.7 % vs. 29.9 %, P = 0.01) than those obtained with the 22-G FNB needle.

How can we explain the differences in performance of the 22-G FNB needle in the two abovementioned studies [2] [3]? Is the 19-G needle really superior to the 22-G FNB in this clinical setting? Which technique should be preferred for EUS-LB?

Variations in the design of needle tips does not seem responsible for the performance differences between the two 22-G FNB needles, considering previous experience in pancreatic solid lesion sampling [4].

A study on EUS-LB in human cadaveric liver tissue revealed better performance of the 19-G FNB SharkCore needle in terms of CPT numbers, but not in terms of overall adequacy, as compared to the 22-G SharkCore [5]. Moreover, the 22-G FNB SharkCore showed better efficiency than the standard 19 G FNA needle [5]. From a technical point of view, three passes with fanning were the only independent predictor of the number of CPTs [5].

In the human cadaveric study [5], suction did not affect sample adequacy. Hasan et al. did not apply any suction but developed a technique to limit tissue fragmentation. This technique restricted elevator utilization, and the stylet was slowly re-inserted while keeping the assembly straight to retrieve samples [2]. However, a moderate degree of post-processing tissue distortion was observed, which might impair histopathological evaluation of acquired samples, especially by community pathologists with limited expertise.

In my personal opinion, the technique for performing EUS-FNB should be different from that for EUS-FNA; it should be more similar to a “Menghini” type biopsy, with limited actuation passes (usually one) at a maximum of three different lesion sites [6]. The wet suction technique showed high effectiveness for EUS-LB, using the 19-G Sharkcore or a standard 19-G FNA needle even with a single pass and one actuation, as reported in a retrospective study on 165 patients [7] and in a prospective study on 40 patients [8]. In this latter study, the wet technique with a heparin-primed needle improved tissue adequacy compared with dry suction techniques [8].

In conclusion, more studies with improved designs are needed to establish the most efficient needle(s) and sampling technique for performing EUS-LB. These studies need to be randomized, multicentric to assess reproducibility, and should involve endosonographers not only from tertiary care centers but also from the community, after they have had adequate training to familiarize them with transduodenal sampling using large caliber needles. Such information is vital for the expansion of EUS-LB to other clinical indications, such as patients with a suspicion of advanced fibrosis of unknown cause with elevated values on transient elastography (> 13.6 kPa) and low platelets count (< 150 000) who undergo upper endoscopy to rule out esophagogastric varices, and, in the near future, to patients who undergo EUS-guided portal pressure gradient measurement without an histological diagnosis of their underlying liver disease.