Surgical Endoscopy and Other Interventional Techniques

https://doi.org/10.1007/s00464-020-08053-x

Endoscopic ultrasound‑guided parenchymal liver biopsy: a systematic

review and meta‑analysis
Bulent Baran1 · Santosh Kale2 · Prithvi Patil3,5 · Bijun Kannadath4,5 · Srinivas Ramireddy3,5 · Ricardo Badillo3,5 ·
Roy Tomas DaVee3,5 · Nirav Thosani3,5 

Received: 11 April 2020 / Accepted: 29 September 2020

© Springer Science+Business Media, LLC, part of Springer Nature 2020

Abstract
Background  Endoscopic ultrasonography (EUS)-guided liver biopsy is a novel technique to obtain adequate liver samples
for diagnosis of liver parenchymal diseases. There are studies that have evaluated the feasibility and safety of EUS-guided
parenchymal liver biopsy (EUS-LB), however, factors that can influence specimen quality are yet to be determined. Our aim
was to determine the diagnostic accuracy of EUS-LB and evaluate factors associated with specimen quality.
Methods  We performed a detailed search of PubMed/MEDLINE and Web of Science™ databases to identify studies in
which results of EUS-guided liver parenchymal biopsies were reported published up to July 2020. A random effects model
was used to estimate pooled values (mean ± SE) for total specimen length (TSL) and complete portal tracts (CPT). Subgroup
analyses were applied to find out the procedural factors associated with better specimen quality using Cochran’s Q test. A
total of 10 meta-analyses were done focusing on international studies. Total of 1326 patients who underwent EUS-LB. EUS-
LBs performed for suspicion of parenchymal liver disease. Pooled mean values for TSL and CPT with subgroup analyses.
Results  Twenty-three studies with a total of 1326 patients were included in our meta-analysis. Overall pooled mean TSL and
CPT were 45.3 ± 4.6 mm and 15.8 ± 1.5, respectively. In subgroup analysis, core biopsy needles proved to better in terms
of CPT than fine-needle aspiration needles (18.4 vs 10.99, p = 0.003). FNB with slow-pull or suction technique provided a
similar TSL (44.3 vs 53.9 mm, p = 0.40), however, slow-pull technique was better in terms of CPT (30 vs 14.6, p < 0.001).
Heterogeneity was present among the studies. Another limitation is the low number randomized control trials.
Conclusion  EUS-guided parenchymal liver biopsy is a good alternative to other methods of liver sampling. Using FNB
needles with a slow-pull technique can provide better results.

Keywords  Endoscopic ultrasonography · Fine-needle biopsy · Liver biopsy · Liver parenchymal disease · Liver histology

Although the use of liver biopsy has decreased with the

IRB Approval: Systematic Reviews and Meta-Analysis are emerging non-invasive markers and techniques to evaluate
exempted from IRB approval. liver fibrosis, histopathologic examination of liver tissue is
necessary to confirm the type of liver injury [1]. Percutane-
Electronic supplementary material  The online version of this
article (https​://doi.org/10.1007/s0046​4-020-08053​-x) contains ous liver biopsy (PLB) has been the main technique for liver
supplementary material, which is available to authorized users. sample acquisition in diagnosis of parenchymal liver dis-
eases and mass lesions. PLB involves transection of the liver
* Nirav Thosani capsule and there are certain clinical factors that increase the
Nirav.thosani@uth.tmc.edu
risk of subcapsular and/or intraperitoneal bleeding which
1
Koc University School of Medicine, Istanbul, Turkey includes patients taking antiplatelets or anticoagulants,
2
Nassau University Medical Center, East Meadow, NY, USA patients with coagulopathy, thrombocytopenia, morbid obe-
3
sity and high-volume ascites [2, 3]. Transjugular liver biopsy
UTHealth McGovern Medical School, Houston, TX, USA
(TLB) was developed to facilitate obtaining liver tissue in
4
University of Arizona, Tucson, AZ, USA these patients with increased risk of bleeding, but even this
5
Center for Interventional Gastroenterology at the University technique is not free of adverse events [4].
of Texas (iGUT), McGovern Medical School at UTHealth,
6400 Fannin, Suite 1400, Houston, TX 77030, USA

13
Vol.:(0123456789)
 Surgical Endoscopy

Endoscopic ultrasonography (EUS) has become an patients included, (2) the outcome measure for liver sample
important tool for gastroenterologists both for diagnostic quality was reported as total specimen length (TSL) in mm
and therapeutic purposes. It allows for the identification and/or number of complete portal tracts (CPT) (3) human
of liver lesions as well as making it possible to perform studies involving adult patients who has not been included
fine-needle aspiration (FNA) or biopsy (FNB) [5, 6]. It also in another publication. The studies that do not report either
appears to be superior in the detection of very small hepatic of those outcome measures were excluded from the study.
lesions (< 1 cm) that are not identified by cross-sectional The studies that included patients or cohorts overlapping
imaging. This has important implications for staging of can- with other published study cohorts were determined using
cer that can affect the treatment strategy. Although the use the name of participating study center(s), investigator(s) and
of EUS for liver sampling is as safe and effective as PLB patient enrollment time frames.
[7], it cannot visualize the entire organ and is limited to
the most proximal regions of both liver lobes. Recently, a Data extraction
new role for EUS-guided parenchymal liver biopsy (EUS-
LB) has emerged and feasibility of this technique have been We extracted the data onto a preset data extraction form
evaluated in multiple studies that showed diagnostic yield of (Microsoft Excel, Microsoft Corporation, Redmond,
90–100% and low rate of adverse events [8]. In those stud- Wash). This form was pilot tested using randomly selected
ies, EUS-LB was performed using different needles (tru-cut, articles to check if all variables of interests were success-
FNA or FNB) and tissue acquisition methods (slow-pull or fully extracted. From each article following information
suction) that makes describing an optimal technique more were recorded: first author’s last name, journal and year
complicated. In this study, we aimed to determine the overall of publication, country of origin, study setting and design,
diagnostic accuracy and safety of EUS-LB, and to evaluate sample size, types, sizes and brands of EUS needles used,
technical factors associated with the specimen quality. tissue acquisition method, number of passes/actuations, and
median and/or mean ± standard deviation (SD) values for
outcome measures. In studies that only reported median
Methods outcome measures, data for range and/or interquartile range
were also recorded. After completing the data extraction, we
Data sources and search strategy reviewed the data extraction form for accuracy.

This systemic review was performed using the current guide- Quality criteria
lines for conducting a systematic review [9]. We performed
a literature search of PubMed/MEDLINE and Web of Sci- The latest PRISMA guideline for systemic review and meta-
ence™ databases using terms as follows: [(endoscopic ultra- analysis of diagnostic test accuracy studies states that those
sound) OR (EUS)] AND (guided) AND (liver biopsy). The studies may be at risk of bias and there also can be concerns
papers and abstracts published until July 2020 in English regarding applicability [10]. Therefore, it recommends indi-
language were included in the study. There was no restric- cating the methods of assessing risk of bias and applicability.
tion of study design but studies with <  6 patients were Current quality assessment methods focus on randomization,
excluded from the analysis. We checked the reference list selection bias of the arms in the study, concealment of allo-
of published clinical reviews and original articles to find cation, and blinding of outcome to evaluate the quality of the
relevant published articles. Two independent reviewers (B.B. clinical trials with the control arm [11]. There is no consen-
and S.K.) screened the title and abstract of retrieved articles. sus or criteria to evaluate the quality of the studies without a
Studies that meet the inclusion criteria were selected for a control arm. Most of the studies focusing on EUS-LB were
full-text review. Disagreements between the two reviewers either retrospective or non-randomized prospective studies
were resolved by a discussion in consecutive meetings with that reported outcome measures in terms of specimen qual-
the participation of investigators. ity. Therefore, for this systematic review and meta-analysis,
we selected studies based on our predefined inclusion and
Inclusion and exclusion criteria exclusion criteria and completeness of data reporting in the
studies.
The study population consisted of patients with clinical evi-
dence or suspicion of parenchymal liver disease who under- Statistical analysis
went EUS-LB procedure. Studies that satisfies the following
criteria were included in the meta-analysis: (1) randomized The Comprehensive Meta-Analysis (CMA) software ver-
controlled trials, prospective or retrospective cohort stud- sion V2 (Biostat Inc, Englewood, NJ) was used in this
ies, case–control studies and case series with more than 5 meta-analysis. We calculated the pooled mean ± SE values

13
Surgical Endoscopy

for predefined outcome measures of TSL (mm) and CPT (n) review articles, 1 study included pediatric patients [20], 9
using the information reported in the studies. The studies were non-human, cadaveric or ex vivo studies, and 14 pub-
differed in methods of reporting their outcome measures. lications were case reports, correspondences or editorials.
In studies where mean ± SD of outcome measures were Fifty-one articles were eligible for further full-text screen-
reported or can be calculated from individual case values; ing. Eleven publications were excluded for being conference
those data were inputted on the datasheet. In studies that abstract of a full-text paper, and 10 studies were excluded
only report median and range and/or interquartile range for being previous publication including the same patient
(IQR), mean and SD needed to be estimated by methods cohort either partially or completely. Five studies did not
previously described by Luo et al. [12] for “mean” and Wan provide sufficient information to calculate or estimate mean
et al. [13] or Shi et al. [14] for SD when appropriate. In these and/or SD of outcome measures [21–25], and 2 studies were
studies that is from the same group, the authors suggested excluded due to the low number of patients included [26,
three specific methods for estimation of mean and SD to 27]. In total, 23 articles were selected for meta-analysis
perform meta-analysis from clinical studies using sample [28–43] of which 7 of them was presented in an abstract
size, median, range and/or IQR. In the present study, we form [44–50].
used each one of those three methods for individual stud- Two publications by Sey et al. [40] and Patel et al. [39]
ies according to the availability of median, range and/or were conducted in the same institution including the same
IQR as follows: method 1 for median and range, method cohort from 2007 to 2014 and 2007–2017, respectively.
2 for median and IQR, and method 3 for median, range The former study only reported TSL, and Patel et al. had
and IQR. Heterogeneity was assessed using I2 measure of results for CPT, therefore both studies were included in the
inconsistency and Cochran’s Q test [15–17]. Heterogeneity respective analyses. In the prospective uncontrolled study
was determined to be present when the value I2 was > 50% by Diehl et al. [31], 110 patients with elevated transami-
or if the Q-statistic was significant at a p < 0.05. Because nases or parenchymal liver diseases underwent EUS-LB by
significant heterogeneity was detected among the studies a a 19-gauge FNA needle. In the study by the same group
random effects model was used to estimate pooled values (Pineda et al.) [51], same patient cohort of 110 were used
(mean ± SE) for TSL and CPT. Subgroup analyses were in a retrospective setting in comparison to a prospectively
applied to find out the procedural factors associated with recruited group of patients that underwent PLB and TLB by
better specimen quality using Cochran’s Q test with a fully interventional radiology in the same institution [51]. Those
random effects analysis. A random effects model is used two studies were considered as a single study and informa-
to combine studies within each subgroup and to combine tion in both studies were extracted to use for analysis. A
subgroups for yielding the overall effect. The study-to-study flow-diagram summarizing the study selection process is
variance ­(tau2) is assumed to be the same for all subgroups shown in Fig. 1.
which is computed within subgroups and then pooled across
subgroups. Comparisons for diagnostic yield is performed
using Pearson Chi-square test. A two-tailed p value of < 0.05 EUS technique
was considered significant.
Since this is a Meta-Analysis/Systematic Review, no IRB Diagnostic EUS was performed in most of the studies
approval is needed from the institution beforehand hereby. with curvilinear echoendoscopes including Olympus GF-
UC140P, GF-UC160-AL5, GF-UE160-AL5, GF-UC160P-
AT8, GF-UC140-AL5, GF-UCT140-AL5, GF-UCT180, GF-
Results UC30P, GFUC140P or GF-UCT140P (Olympus America,
Center Valley, PA) and Pentax EG-3870UTK [29] (Pentax
Literature search Medical Co, Montvale, NJ, USA).
In 12 studies EUS-LB was solely performed to the left
A total of 978 titles and/or abstracts were initially retrieved liver lobe, and both left and right lobes were biopsied in 6
using the search strategy. Two more articles were retrieved studies. In the study by Diehl et al. [31], there were 3 groups
using manual search and review of reference lists [18, 19]. in terms of biopsied liver lobe: both lobes (68 patients), left
Of them, 258 articles were found to be duplicates either lobe (34 patients) and right lobe (8 patients). In the study
published as an abstract or were present in both databases. by Hasan et al. 40 patients included were intended to be
Of remaining 722 articles, 622 excluded by 2 independent biopsied from both left and right lobes, but 2 patients were
reviewers (B.B. and S.K.) after preliminary review of titles biopsied from the left lobe only because of surgically altered
and abstracts, which left 100 articles for detailed evaluation anatomy [34]. Right, left or both lobes were biopsied vari-
of full-texts. Among those papers, 49 articles failed to meet ably in the remaining 5 studies.
the predefined selection criteria: 25 were meta-analysis or

13
 Surgical Endoscopy

Fig. 1  Flow-diagram of the study selection process

Variable types of 19G EUS needles were used with Meta‑analysis

different tissue acquisition techniques. Foor-Pessin et al.
used both standard and core needles but did not men- Twenty-three studies with a total of 1326 patients and 1488
tioned the number of patients per needle-type. Median liver biopsies were included in the meta-analysis. There
number of passes was variable among the studies and were 14 case series and retrospective studies, 2 randomized
ranged from 1 to 4. Details of EUS techniques used in controlled [30, 45] and 2 randomized crossover studies [35,
individual studies are summarized in Table 1. 37], 5 prospective uncontrolled studies [29, 31, 34, 36, 44].

13
Surgical Endoscopy

Table 1  Summary of characteristics, techniques, equipment and diagnostic yield of studies evaluating feasibility of EUS-LB
Index Author/year Study type No of Approach Type of Needle brand Technique No of Diagnostic
patients needle passes yield (%)
(median)

1 Gleeson/2008 Retrospective 9 Left lobe EUS-TB Q-C Tru-cut 2 100

2 Stavropoulos/2012 Case series 22 Left lobe EUS-FNA EchoTip 19G Suction 2 91
3 Gor/2014 Case series 10 Left lobe EUS-FNA Expect 19G Not 3 100
mentioned
4 Diehl/2015a Prospective 110 Variablec EUS-FNA Expect 19G Suction 2 98
cohort or Expect (20 ml)
flex 19G
5 Sey/2016b Retrospective 75 Left lobe EUS-TB vs Q-C v Tru-cut vs 3 vs 2 83
FNB ProCore Suction
(10 ml)
6 Shah/2017 Retrospective 24 Left lobe EUS-FNB SharkCore Slow-pull 2 96
19G
7 Foor-Pessin/2017 Retrospective 15 Left lobe EUS-FNA Not Not 2 100
or FNB mentioned mentioned
8 Rombaoa/2018 Retrospective 10 Left lobe EUS-FNB Acquire 19G Not 2 100
mentioned
9 Ching- RCT​ 40 Both lobes EUS-FNB 19G FNB Suction 1 100
Companioni/2018 needle (20 ml)
10 El Chafic/2018 Retrospective 52 Left lobe EUS-FNB 19G FNB Wet suction 2 Not provided
needle
11 Shimizu/2018 Retrospective 44 Left lobe EUS-FNA EchoTip 19G Variable 1–2 100
12 Mok/2018 Prospective 40 Left lobe EUS-FNA Expect flex Suction 3 98
crossover 19G (variable)
study
13 Mok/2019 Randomized 20 Both lobes EUS-FNA Expect flex Wet suction 4 88 vs 68
crossover or FNB 19G v with heparin
study Sharkcore
22G
14 Hasan/2019 Prospective 40 Both lobes EUS-FNB Acquire 22G No suction 3 100
cohort
15 Shuja/2019 Retrospective 69 Both lobes EUS-FNA Expect flex Wet suction 3 100
19G
16 Ching- RCT​ 40 Both lobes EUS-FNA Expect flex Wet 2 100
Companioni/2019 vs FNB 19G v suction
Acquire with
19G heparin
17 Bazerbachi/2019 Prospective 21 Left lobe EUS-FNB SharkCore Suction 2 100
cohort 22G (10 ml)
18 Bhat/2019 Prospective 39 Left lobe EUS-FNB SharkCore Slow-pull 1–2 95
cohort 19G vs
22G
19 Ellis/2019 Retrospective 39 Variablec EUS-FNB SharkCore Wet suction 1 95
or Acquire
19G
20 Ali/2020 Retrospective 30 Left or EUS-FNB SharkCore Wet suc- 2 100
right lobe 19G or tion with
22G heparin
21 Patel/2020b Retrospective 135 Left or both EUS-TB vs Q-C v Suction 2 65
lobes FNB (22 Acquire v
vs 19G) ProCore
v Expect
flex

13
 Surgical Endoscopy

Table 1  (continued)
Index Author/year Study type No of Approach Type of Needle brand Technique No of Diagnostic
patients needle passes yield (%)
(median)
22 Hashimoto/2020 Randomized 22 Left or EUS-FNB Acquire 19G Wet suction 2 100
crossover right lobe vs Shark-
study Core 19G
23 Nieto/2020 Retrospective 420 Both lobes EUS-FNB Acquire 19G Wet suction 2 100
vs Shark-
Core 19G

No number, EUS-TB endoscopic ultrasonography-guided tru-cut biopsy, FNA fine-needle aspiration, FNB fine-needle biopsy, RCT​ randomized
controlled trial, Q-C Quick-Core, 19G FNB needle (SharkCore or Acquire 19G)
a
 Studies by Diehl and Pineda et al. reported same cohort of 110 patients
b
 Studies by Sey et al. and Patel et al. are from the same institution and have cohort overlap. Data from Sey et al. was used in analyses for TSL
only
c
 Right, left or both liver lobes were biopsied

The study characteristics of individual studies are shown in p = 0.21), yet Fork and Franseen-tip needles performed
Table 1. superior than reverse-bevel core needle (ProCore; Cook
Two recent studies by Mok et al. [36, 37] used a crossover Medical, Bloomington, IN, US) (96%, 768/797 vs 89.6%,
design in their respective cohorts. In the first study, inves- 52/58, respectively, p < 0.013). Foor-Pessin et al. used
tigators compared wet-heparin, dry-heparin and dry-needle either 19G FNA or FNB needles in their cohort with-
EUS-LB techniques in prospectively included 40 patients. out providing details, therefore the study was excluded
Each patient underwent 3 passes with each technique in a from the analyses regarding needle-type. After exclud-
crossover design (120 biopsies in total), and best specimen ing patients who underwent left lobe EUS-TB because of
quality was obtained by wet-heparin technique versus dry significantly lower performance compared with standard
techniques. The latter study compared nitinol 19G FNA nee- and core-type needles, pooled diagnostic yield for bilobar
dle (Expect Flexible 19G, Boston Scientific, Marlborough, approach resulted better than single lobar EUS-LB (99.7%
MA, US) with 22G fork-tip core needle (SharkCore, Covi- 704/706 vs 91.4%, 544/595, respectively, p < 0.001). The
dien-Medtronic Inc, Minneapolis, MN, US) in 20 patients studies by Patel et al. and El Chafic et al. were excluded
using a randomized crossover design. Left and right lobes of from this analysis regarding lobar approach due to missing
the liver were biopsied using both needle types for all study information.
patients, resulting in a total of 80 biopsies. The results were Outcome measures of TSL were available in 20 studies
reported per needle basis. Another randomized crossover and CPT were available in 22 studies. Overall pooled mean
study by Hashimoto et al., included 22 patients who were TSL and CPT were 45.3 ± 4.6 mm and 15.8 ± 1.5, respec-
randomized to either left or right liver lobe for biopsy, as tively (Fig. 2). In subgroup analysis, pooled values of TSL
well as the sequence of needle used (19G franseen-tip— and CPT for FNA and core needles were as follows: TSL;
Acquire; Boston Scientific, Marlborough, MA, US—vs 19G 48.9 ± 7.3 vs 51.9 ± 5.2 mm, p = 0.74; CPT; 10.99 ± 1.97 vs
fork-tip). We included all techniques in the meta-analysis to 18.4 ± 1.6, respectively, p = 0.003 (Fig. 3).
ensure a per biopsy/needle analysis with a pooled compari- FNB with or without suction (no stylet or slow-pull tech-
son of different techniques. nique) provided following pooled TSL and CPT values,
The diagnostic yield of EUS-LB was reported in all respectively: TSL; 53.9 ± 4.7 vs 44.3 ± 10.6 mm, p = 0.40
studies except the study by El Chafic et  al. which was and CPT; 14.6 ± 1.7 vs 30 ± 4.1, p < 0.001 (Supplementary
published as a conference abstract [46]. The number of Fig. 1).
patients with inadequate/non-diagnostic liver biopsy The comparison of 19G vs 22G EUS needles in terms
were 100. Overall diagnostic yield of EUS-LB was 93% of pooled TSL and CPT values were as follows, respec-
(1336/1436). Diagnostic yield achieved by EUS-FNB nee- tively: TSL; 52.6 ± 4.5 vs 41.6 ± 9.7 mm, p = 0.30 and CPT;
dles (standard or core needles) was found to be better than 14.7 ± 1.44 vs 17.3 ± 3.18, p = 0.47 (Supplementary Fig. 2).
EUS-TB (95%, 1271/1332 vs 63%, 65/104, respectively, The analysis of studies using a Fork-tip or Franseen-tip
p < 0.001). EUS-FNB technique carried out using a stand- FNB needle for EUS-LB, were as follows, respectively: TSL;
ard or core-type needle resulted with comparable diag- 39.3 ± 6.8 vs 75.9 ± 9.9 mm, p = 0.002 and CPT; 17.7 ± 3.47
nostic yield (94%, 436/462 vs 96%, 820/855, respectively, vs 22.3 ± 3.97, p = 0.38 (Supplementary Fig. 3).

13
Surgical Endoscopy

Fig. 2  Forrest plots of studies for A total specimen length (TSL) and B complete portal tracts (CPT)

13
 Surgical Endoscopy

Fig. 3  Forrest plots for A total specimen length (TSL) and B complete portal tracts (CPT) for comparison of EUS-guided fine-needle aspiration
(FNA) needle versus core needle biopsy

13
Surgical Endoscopy

Adverse events introduction of novel EUS-FNB needles into the market,

fork and franseen-tip needles have become the needle of
In all but one study by Patel et al. [39] post-procedural choice in most centers when core tissue is needed for histo-
adverse events or complications were mentioned. Seven pathological examination. In our meta-analysis, franseen-tip
studies including 179 patients reported that there were no needle provided a longer TSL compared with fork-tip needle,
complications in any patient after the procedure. In the which is most likely associated with an outlier result from 1
remaining 16 studies that included 1147 patients, 111 (9.7%, study [30] in which both liver lobes were biopsied. Also, the
111/1147) experienced post-procedural complications. Most comparable number of CPTs obtained by both core needles
common post-procedural complication was abdominal pain suggests similar diagnostic specimen quality.
that was reported in 102 patients. Significant complica- A liver biopsy specimen measuring ≥ 15 mm and contain-
tions that required emergency visit, inpatient observation or ing ≥ 6 CPTs is generally considered adequate for the histo-
hospitalization developed in 15/1326 (1.1%) patients. Six logic diagnosis of parenchymal liver disease [53, 54]. How-
patients developed bleeding and/or subcapsular hematoma ever, stricter requirements of specimen length of ≥ 20 mm
that required emergency admission or inpatient observation with ≥ 11 CPTs for reliable evaluation of liver disease have
with administration of intravenous analgesia. One patient been recommended by several authorities [1, 55]. This
developed bile-leak [38] and 1 death was reported in the threshold is difficult to achieve in daily practice, even with a
first 24 h after the procedure in whom the cause of death large-bore PLB needle and often requires a minimum of two
was unclear [34]. passes which may increase the risk of complications with the
percutaneous route [56]. In the recent years, multiple studies
clearly demonstrated that EUS-LB has become an important
Discussion alternative to PLB to acquire liver sample in patients with
suspected parenchymal liver disease. Our meta-analysis con-
In this study, we reviewed the published literature on the fea- firms that EUS-LB is a feasible technique to obtain adequate
sibility, safety and utility of EUS-LB in patients undergoing liver samples for diagnosis of parenchymal liver diseases
parenchymal liver biopsy. Pooled analysis of studies showed (Table 2). However, it has been difficult to define the optimal
a high diagnostic yield of more than 90% which is compa- method of EUS-LB regarding the brand or type of needle,
rable with the yield of PLB [52]. Moreover, after excluding tissue acquisition technique such as suction or stylet slow-
studies using EUS-TB technique with QuickCore™ needle pull and the number of passes. The level of heterogeneity
which is not available in the market anymore due to the of studies included, which is clearly a limitation for this
technical difficulties associated with higher failure rates, meta-analysis, was related with the intention of investigators
diagnostic yield increased to ≥ 95% using EUS-FNB tech- having the primary motivation of proving the feasibility and
nique with either standard or core-type needles. The higher safety of the concept instead of further developing it.
diagnostic yield can be attributed to a better TSL and CPT In a recent systematic review and meta-analysis, Mohan
performed by the EUS-FNB compared with EUS-TB tech- et al. [57] also evaluated the role of EUS-guided liver biopsy,
nique. Also, novel fork or franseen-tip needles performed and they showed that EUS-LB is both safe and effective
better than reverse-bevel (ProCore™) needles, which indi- which is confirmed by our results. However, our meta-anal-
cates the reverse-beveled structure of Procore™ needles can ysis is the most recent one with more studies and patients
become a disadvantage in EUS-LB. The comparison of FNB
with or without suction showed comparable pooled TSL, yet
Table 2  Comparison of liver sample adequacy requirements for opti-
higher number of CPTs in patients who underwent EUS- mal diagnosis and staging of liver diseases
FNB without suction (slow-pull or no stylet). This finding
EASL AASLD EUS-guided
can be explained by the lower negative pressure leading to a fine-needle
less fragmented sample or most likely the heterogeneity of ­biopsya
studies included. Nevertheless, significantly higher number
of CPTs may favor “no suction”, which needs evaluation in Liver specimen ≥ 15 mm ≥ 20 mm 51 mm
length
a better designed RCT. Standard 19G FNA needles did as
Number of ≥ 6 ≥ 11 15
good as core-type needles in terms of TSL but core needles complete portal
provided a higher number of CPTs, that is most likely associ- tracts
ated with the less fragmented liver tissue obtained. Although
AASLD American Association for the Study of Liver Diseases, EASL
it can be assumed that a larger caliber biopsy needle would European Association for the Study of the Liver, EUS endoscopic
provide less fragmentation and a higher specimen quality, ultrasonography
our meta-analysis showed no significant difference for 19G a
 According to the results of the current meta-analysis excluding tru-
or 22G FNB needles in terms of TSL and CPTs. With the cut needle

13
 Surgical Endoscopy

included, and previous meta-analyses did not consider Compliance with ethical standards 
cohort overlaps while selecting studies to include. Moreo-
ver, Mohan et al. did not consider TSL and CPT which are Disclosures  Dr. Nirav Thosani reports other from Boston Scientific,
the most important quality indicators for liver biopsy. Lastly, other from Medtronic, primarily as a consultant outside the submitted
work. Drs. Bulent Baran, Santosh Kale, Prithvi Patil, Bijun Kannadath,
they did not include the studies by Mok et al. published in Srinivas Ramireddy, Ricardo Badillo and Tomas Davee have no con-
2018 and 2019. flicts of interest or financial ties to disclose.
EUS-LB may not be a first-line approach given its cost
and requisite for sedation. However, its evident safety, high
specimen quality and diagnostic yield in patients already
undergoing endoscopic procedures may expand the role of
References
EUS in the diagnosis of liver parenchymal disease. Fur-
thermore, the ability to perform a combined endoscopy and 1. Rockey DC, Caldwell SH, Goodman ZD, Nelson RC, Smith AD
EUS-LB has the potential to save health care costs. However, (2009) American Association for the Study of Liver D. Liver
a cost-effectiveness analysis is required to further elucidate biopsy. Hepatology 49:1017–1044
this issue. 2. Boyum JH, Atwell TD, Schmit GD, Poterucha JJ, Schleck CD,
Harmsen WS et al (2016) Incidence and risk factors for adverse
EUS-LB may also be useful in patients with contrain- events related to image-guided liver biopsy. Mayo Clin Proc
dications to the use of conventional techniques including 91:329–335
morbid obesity, ascites, lack of cooperation, and hemangio- 3. Seeff LB, Everson GT, Morgan TR, Curto TM, Lee WM, Ghany
mas or cysts. EUS provides a high-resolution image that MG et al (2010) Complication rate of percutaneous liver biopsies
among persons with advanced chronic liver disease in the HALT-
allows visualization while allowing real-time monitoring and C trial. Clin Gastroenterol Hepatol 8:877–883
passage of the biopsy needle into the liver. Doppler capa- 4. Kalambokis G, Manousou P, Vibhakorn S, Marelli L, Cholongitas
bility of the linear echoendoscope further allows intrahe- E, Senzolo M et al (2007) Transjugular liver biopsy—indications,
adequacy, quality of specimens, and complications—a systematic
patic vessels to be avoided during biopsy, that significantly
review. J Hepatol 47:284–294
contributes to safety of the procedure. Pooled analysis of 5. Nguyen P, Feng JC, Chang KJ (1999) Endoscopic ultrasound
the studies showed that adverse event rate is < 10%, and (EUS) and EUS-guided fine-needle aspiration (FNA) of liver
clinically significant complication rate was around 1%, both lesions. Gastrointest Endosc 50:357–361
6. Jhala NC, Jhala D, Eltoum I, Vickers SM, Wilcox CM, Chhieng
are acceptable to consider this invasive procedure as safe as
DC et al (2004) Endoscopic ultrasound-guided fine-needle aspira-
conventional percutaneous route. Nevertheless, an important tion biopsy: a powerful tool to obtain samples from small lesions.
limitation to our meta-analysis is that most of the studies Cancer 102:239–246
were retrospective and not designed to capture all adverse 7. tenBerge J, Hoffman BJ, Hawes RH, Van Enckevort C, Giovannini
M, Erickson RA et al (2002) EUS-guided fine needle aspiration of
events and potential complications.
the liver: indications, yield, and safety based on an international
One additional concern when liver biopsy specimens are survey of 167 cases. Gastrointest Endosc 55:859–862
obtained is that some disease processes may not affect the 8. Shah ND, Baron TH (2018) Endoscopic ultrasound and the liver:
liver uniformly. Sampling error which is the most impor- current applications and beyond. J Hepatobiliary Pancreat Sci
25:171–180
tant limitation of liver biopsy can be minimized by obtain-
9. Moher D, Liberati A, Tetzlaff J, Altman DG, Group P (2009) Pre-
ing separate biopsy specimens from the right and left lobes ferred reporting items for systematic reviews and meta-analyses:
of the liver. The published literature confirms that bilobar the PRISMA statement. J Clin Epidemiol 62:1006–12
approach under EUS guidance is safe and this meta-analysis 10. McInnes MDF, Moher D, Thombs BD, McGrath TA, Bossuyt PM,
the P-DTAG et al (2018) Preferred reporting ıtems for a system-
shows that specimen quality tends to be better than unilobar
atic review and meta-analysis of diagnostic test accuracy studies:
approach. the PRISMA-DTA statement. JAMA 319:388–396
In conclusion, EUS-LB is safe and effective in patients 11. Whiting PF, Rutjes AW, Westwood ME, Mallett S, Deeks JJ,
undergoing endoscopic procedures for any indication. This Reitsma JB et al (2011) QUADAS-2: a revised tool for the qual-
ity assessment of diagnostic accuracy studies. Ann Intern Med
technique provides more than adequate samples with poten-
155:529–536
tially improved patient satisfaction. Using an FNB needle 12. Luo D, Wan X, Liu J, Tong T (2018) Optimally estimating the
with slow-pull technique may provide better specimen sample mean from the sample size, median, mid-range, and/or
quality and higher diagnostic yield. This technique will not mid-quartile range. Stat Methods Med Res 27:1785–1805
13. Wan X, Wang W, Liu J, Tong T (2014) Estimating the sample
replace PLB but may emerge as a desirable alternative with
mean and standard deviation from the sample size, median, range
expanded use. and/or interquartile range. BMC Med Res Methodol 14:135
14. Shi J, Luo D, Weng H, Zeng XT, Lin L, Chu H et al (2020) Opti-
mally estimating the sample standard deviation from the five-
number summary. Res Synth Methods 11:641–654
15. Higgins JP, Thompson SG, Deeks JJ, Altman DG (2003) Meas-
uring inconsistency in meta-analyses. BMJ 327:557–560

13
Surgical Endoscopy

16. Deeks JJ (2001) Systematic reviews in health care: systematic a 19-gauge non-Tru-Cut FNA needle. Gastrointest Endosc
reviews of evaluations of diagnostic and screening tests. BMJ 79:170–172
323:157–162 34. Hasan MK, Kadkhodayan K, Idrisov E, Ali S, Rafiq E, Ben-Ami
17. Higgins JP, Thompson SG (2002) Quantifying heterogeneity in Shor D et al (2019) Endoscopic ultrasound-guided liver biopsy
a meta-analysis. Stat Med 21:1539–1558 using a 22-G fine needle biopsy needle: a prospective study.
18. Nakanishi Y, Mneimneh WS, Sey M, Al-Haddad M, DeWitt JM, Endoscopy 51:818–824
Saxena R (2015) One hundred thirteen consecutive transgastric 35. Hashimoto R, Lee DP, Samarasena JB, Chandan VS, Guo W, Lee
liver biopsies for hepatic parenchymal diseases: a single-insti- JG et al (2020) Comparison of two specialized histology nee-
tution study. Am J Surg Pathol 39:968–976 dles for endoscopic ultrasound (EUS)-guided liver biopsy: a pilot
19. Saab S, Phan J, Jimenez MA, Grotts JF, Walters L, Hathaway study. Dig Dis Sci. https​://doi.org/10.1007/s1062​0-020-06391​-3
KA et al (2017) Endoscopic ultrasound liver biopsies accurately 36. Mok SRS, Diehl DL, Johal AS, Khara HS, Confer BD, Mudireddy
predict the presence of fibrosis in patients with fatty liver. Clin PR et al (2018) A prospective pilot comparison of wet and dry
Gastroenterol Hepatol Off Clin Pract J Am Gastroenterol Assoc heparinized suction for EUS-guided liver biopsy (with videos).
15:1477–1478 Gastrointest Endosc 88:919–925
20. Johal AS, Khara HS, Maksimak MG, Diehl DL (2014) Endo- 37. Mok SRS, Diehl DL, Johal AS, Khara HS, Confer BD, Mudireddy
scopic ultrasound-guided liver biopsy in pediatric patients. PR et al (2019) Endoscopic ultrasound-guided biopsy in chronic
Endosc Ultrasound 3:191–194 liver disease: a randomized comparison of 19-G FNA and 22-G
21. Aggarwal SN, Magdaleno T, Shah H, Shah S, Goonewardene S FNB needles. Endosc Int Open 7:E62–E71
(2019) Endoscopıc ultrasound-guıded fıne needle lıver bıopsy: 38. Nieto J, Dawod E, Deshmukh A, Penn E, Adler D, Saab S (2020)
a sıngle-center experıence at a large, quartenary care center. EUS-guided fine-needle core liver biopsy with a modified one-
Gastrointest Endosc 89:AB305 pass, one-actuation wet suction technique comparing two types of
22. Earasi AG, Lynch S, Jahann DA, Sauer BG, Wang AY, Uppal EUS core needles. Endosc Int Open 8:E938–E943
D et al (2019) Transjugular vs. EUS-guıded lıver bıopsy: whıch 39. Patel HK, Saxena R, Rush N, Patel SK, Dasari CS, Mneimneh W
modalıty ıs better? Gastrointest Endosc 89:AB326 et al (2020) A comparative study of 22G versus 19G needles for
23. Sampath K, Gordon SR, Toor A, Dinani A, Dickson RC, Suri- EUS-guided biopsies for parenchymal liver disease: are thinner
awinata A et  al (2017) Endoscopic ultrasound-guided liver needles better? Dig Dis Sci. https​://doi.org/10.1007/s1062​0-020-
biopsy: comparing the diagnostic yield and safety between the 06165​-x
19-gauge and 22-gauge needle. Gastrointest Endosc 85:AB479 40. Sey MS, Al-Haddad M, Imperiale TF, McGreevy K, Lin J, DeWitt
24. Sarkar A, Shahid H, Tyberg A, Andalib I, Kahaleh M, JM (2016) EUS-guided liver biopsy for parenchymal disease: a
Gaidhane M et  al (1090A) Endoscopic ultrasound (EUS) comparison of diagnostic yield between two core biopsy needles.
guided liver biopsy: a single center experience. Hepatology Gastrointest Endosc 83:347–352
70:1089A–1090A 41. Shah ND, Sasatomi E, Baron TH (2017) Endoscopic ultrasound-
25. Shahshahan M, Gertz H, Fakhreddine AY, Datta A, Eysselein VE, guided parenchymal liver biopsy: single center experience of a
Reicher S et al (2017) Endoscopic ultrasound-guided liver biopsy new dedicated core needle. Clin Gastroenterol Hepatol Off Clin
versus percutaneous and trans-jugular liver biopsy for evaluation Pract J Am Gastroenterol Assoc 15:784–786
of liver parenchyma. Gastrointest Endosc 85:AB490 42. Shuja A, Alkhasawneh A, Fialho A, Fialho A, Shukri A, Harris
26. Mathew A (2007) EUS-guided routine liver biopsy in selected C et al (2019) Comparison of EUS-guided versus percutaneous
patients. Am J Gastroenterol 102:2354–2355 and transjugular approaches for the performance of liver biopsies.
27. Phillips B, Lo DY, Forsthoefel K, Speck O, Turner B (2018) Ade- Dig Liver Dis Off J Ital Soc Gastroenterol Ital Assoc Study Liver
quacy of endoscopic ultrasound-guided liver fine needle biopsies 51:826–830
using a 20-gauge needle. Am J Gastroenterol 113:S1230–S1231 43. Stavropoulos SN, Im GY, Jlayer Z, Harris MD, Pitea TC, Turi GK
28. Ali AH, Panchal S, Rao DS, Gan Y, Al-Juboori A, Samiullah S et al (2012) High yield of same-session EUS-guided liver biopsy
et al (2020) The efficacy and safety of endoscopic ultrasound- by 19-gauge FNA needle in patients undergoing EUS to exclude
guided liver biopsy versus percutaneous liver biopsy in patients biliary obstruction. Gastrointest Endosc 75:310–318
with chronic liver disease: a retrospective single-center study. J 44. Bhat I, Malvar G, Calmet F, Bonder A, Nasser IA, Sawhney M
Ultrasound 23:157–167 et al (2019) EUS-guıded lıver bıopsy usıng 22G FNB vs. 19G
29. Bazerbachi F, Vargas EJ, Matar R, Storm AC, Mounajjed TM, FNB needles: a prospectıve comparıson of core specımen sıze
Topazian MD et al (2019) EUS-guided core liver biopsy sam- and dıagnostıc adequacy. Gastrointest Endosc 89:AB293
pling using a 22-gauge fork-tip needle: a prospective blinded trial 45. Ching-Companioni R, Johal AS, Confer B, Khara HS, Forster
for histologic and lipidomic evaluation in nonalcoholic fatty liver E, Diehl DL (2018) One versus three needle actuations for EUS-
disease. Gastrointest Endosc 90:926–932 guided liver biopsy: a prospective randomized trial. Am J Gastro-
30. Ching-Companioni RA, Diehl DL, Johal AS, Confer BD, Khara enterol 113:S420
HS (2019) 19 G aspiration needle versus 19 G core biopsy needle 46. El Chafic AH, Mubarak MF, Shah J, Therapondos G, Romero R,
for endoscopic ultrasound-guided liver biopsy: a prospective ran- Evans J et al (2018) Endoscopic ultrasound-guided liver biopsy
domized trial. Endoscopy 51:1059–1065 compared to percutaneous and transjugular liver biopsy: a tertiary
31. Diehl DL, Johal AS, Khara HS, Stavropoulos SN, Al-Haddad M, center experience. Am J Gastroenterol 113:S421
Ramesh J et al (2015) Endoscopic ultrasound-guided liver biopsy: 47. Ellis DJ, Therapondos G, Shah JN, Cohen A, Galliano GE,
a multicenter experience. Endosc Int Open 3:E210–E215 Romero RV et al (2019) Simultaneous biliary and liver paren-
32. Gleeson FC, Clayton AC, Zhang L, Clain JE, Gores GJ, Rajan chymal evaluation in post liver-transplant population using endo-
E et al (2008) Adequacy of endoscopic ultrasound core needle scopic ultrasound-guided liver biopsy and endoscopic retrograde
biopsy specimen of nonmalignant hepatic parenchymal disease. cholangiopancreatography in a single session approach. Am J
Clin Gastroenterol Hepatol Off Clin Pract J Am Gastroenterol Gastroenterol 114:S501–S502
Assoc 6:1437–1440 48. Foor-Pessin C, Bittner K, Kothari S, Sharma M, Zhou ZR,
33. Gor N, Salem SB, Jakate S, Patel R, Shah N, Patil A (2014) Sharma A et al (2017) Histologic yield of endoscopic ultrasound
Histological adequacy of EUS-guided liver biopsy when using guided liver biopsy compared to percutaneous and transjugular

13
 Surgical Endoscopy

approaches: a single-center retrospective review. Gastrointest needle liver biopsy evaluation of patients with chronic hepatitis
Endosc 85:AB485–AB486 C. Clin Gastroenterol Hepatol 3:930–935
49. Rombaoa C, Chen AM (2018) The safety and feasıbılıty of 54. European Association for Study of L, Asociacion Latinoameri-
endoscopıc ultrasound-guıded parenchymal lıver bıopsy at a large cana para el Estudio del H (2015) EASL-ALEH Clinical Prac-
communıty hospıtal. Gastrointest Endosc 87:AB458 tice Guidelines: non-invasive tests for evaluation of liver disease
50. Shimizu T, Fortinsky KJ, Tsujino T, Lu YX, Chin MA, Samar- severity and prognosis. J Hepatol 63:237–264
asena JB et al (2018) EUS guıded lıver bıopsy ın benıgn lıver 55. Fryer E, Wang LM, Verrill C, Fleming K (2013) How often do our
dısease usıng heparanızed needle technıque (HNT). Gastrointest liver core biopsies reach current definitions of adequacy? J Clin
Endosc 87:AB346 Pathol 66:1087–1089
51. Pineda JJ, Diehl DL, Miao CL, Johal AS, Khara HS, Bhanushali A 56. Al Knawy B, Shiffman M (2007) Percutaneous liver biopsy in
et al (2016) EUS-guided liver biopsy provides diagnostic samples clinical practice. Liver Int 27:1166–1173
comparable with those via the percutaneous or transjugular route. 57. Mohan BP, Shakhatreh M, Garg R, Ponnada S, Adler DG (2019)
Gastrointest Endosc 83:360–365 Efficacy and safety of EUS-guided liver biopsy: a systematic
52. Sporea I, Sirli R, Popescu A, Cornianu M, Manciu C, Focsa M review and meta-analysis. Gastrointest Endosc 89:238–246.e3
(2007) The quality of the fragment obtained by liver biopsy for
staging chronic hepatitis. J Gastrointest Liver Dis 16:263–266 Publisher’s Note Springer Nature remains neutral with regard to
53. Schiano TD, Azeem S, Bodian CA, Bodenheimer HC Jr, Merati jurisdictional claims in published maps and institutional affiliations.
S, Thung SN et al (2005) Importance of specimen size in accurate

13