Endoscopy 2014; 46(12): 1063-1070
DOI: 10.1055/s-0034-1377559
Original article
© Georg Thieme Verlag KG Stuttgart · New York

Core needle versus standard needle for endoscopic ultrasound-guided biopsy of solid pancreatic masses: a randomized crossover study

Geoffroy Vanbiervliet
1   Université de Nice Sophia Antipolis, Faculté de Médecine and Centre Hospitalier Universitaire de l’Archet 2, Pôle digestif, Nice, France
,
Bertrand Napoléon
2   Hôpital Privé Mermoz, Gastroentérologie, Lyon, France
,
Marie Christine Saint Paul
3   Centre Hospitalier Universitaire Pasteur, Anatomopathologie, Nice, France
,
Charlotte Sakarovitch
4   Centre Hospitalier Universitaire Cimiez, Département de la recherche clinique, Nice, France
,
Marc Wangermez
5   Centre Hospitalier Universitaire de Poitiers, Gastroentérologie, Poitiers, France
,
Philippe Bichard
6   Hôpitaux Universitaires de Genève, Gastroentérologie, Genève, Switzerland
,
Clément Subtil
7   Centre hospitalier Universitaire du Haut-Lévêque, Gastroentérologie, Pessac, France
,
Stéphane Koch
8   Centre hospitalier Universitaire de Besançon, Gastroentérologie, Besançon, France
,
Philippe Grandval
9   Assistance publique des hôpitaux de Marseille, Hôpital de la Timone, Gastroentérologie, Marseille, France
,
Rodica Gincul
10   Hospices civils de Lyon, Gastroentérologie, Lyon, France
,
David Karsenti
11   Clinique de Bercy, Pôle digestif, Charenton le Pont, France
,
Laurent Heyries
12   Assistance publique des hôpitaux de Marseille, Hôpital de la Conception, Gastroentérologie, Marseille, France
,
Jean-Christophe Duchmann
13   Centre Hospitalier Général de Compiègne, Gastroentérologie, Compiègne, France
,
Jean François Bourgaux
14   Centre Hospitalier Universitaire Carémeau, Gastroentérologie, Nîmes, France
,
Michaël Levy
15   Assistance Publique des hôpitaux de Paris, Hôpital Mondor, Gastroentérologie, Créteil, France
,
Gilles Calament
16   Centre Hospitalier Universitaire de la Cavale Blanche, Gastroentérologie, Brest, France
,
Fabien Fumex
2   Hôpital Privé Mermoz, Gastroentérologie, Lyon, France
,
Bertrand Pujol
2   Hôpital Privé Mermoz, Gastroentérologie, Lyon, France
,
Christine Lefort
2   Hôpital Privé Mermoz, Gastroentérologie, Lyon, France
,
Laurent Poincloux
17   Centre Hospitalier Universitaire Estaing, Gastroentérologie, Clermont Ferrand, France
,
Maël Pagenault
18   Centre Hospitalier Universitaire de Pontchaillou, Gastroentérologie, Rennes, France
,
Eduardo Aimé Bonin
19   Assistance publique des hôpitaux de Marseille, Hôpital Nord, Gastroentérologie, Marseille, France
,
Monique Fabre
20   Gustave Roussy, Pathologie Morphologique, Villejuif, France
,
Marc Barthet
19   Assistance publique des hôpitaux de Marseille, Hôpital Nord, Gastroentérologie, Marseille, France
› Author Affiliations
Further Information

Corresponding author

Geoffroy Vanbiervliet, MD MSc
Hôpital de L'Archet 2 – Pôle digestif
151 Route de Saint Antoine de Ginestière
CS 23079
06202 Nice
France   
Fax: +33-492-035928   

Publication History

submitted: 15 September 2013

accepted after revision: 12 June 2014

Publication Date:
06 August 2014 (online)

 

Background and study aims: A new core biopsy needle for endoscopic ultrasound (EUS)-guided sampling has recently been developed. The aim of this prospective multicenter study was to compare this needle with a standard needle in patients with solid pancreatic masses.

Patients and methods: Consecutive patients with solid pancreatic masses referred to 17 centers for EUS-guided sampling were included. Each patient had two passes with a standard 22G needle and a single pass with a 22G core needle performed in a randomized order. Samples from both needles were separately processed for liquid-based cytology and cell-block preparation and were assessed independently by two blinded expert pathologists. The primary endpoint was the accuracy of the detection of malignancy. The reference standard was based on further cytohistological analysis obtained under ultrasound or computed tomography scanning, endoscopic or surgical guidance, and/or by clinical follow-up with repeated imaging examinations for at least 12 months. The secondary endpoints were the rate of technical failure and the quality of the cytohistological samples obtained.

Results: Of the 80 patients included (49 men; mean age 67.1 ± 11.1), 87.5 % had final malignant diagnoses (adenocarcinoma n = 62, 77.5 %). There was no difference between the needles in diagnostic accuracy (standard needle 92.5 % vs. core needle 90 %; P = 0.68) or technical failure. Both pathologists found the overall sample quality significantly better for the standard needle (expert 1, P = 0.009; expert 2, P = 0.002).

Conclusions: The diagnostic accuracy of EUS sampling for solid pancreatic masses using standard and core needles seems comparable but with a better overall histological sample quality for the former.

ClinicalTrial.gov identifier: NCT01479803.


#

Introduction

Endoscopic ultrasound (EUS)-guided fine needle aspiration (EUS-FNA) is a validated and recommended technique for tissue diagnosis of pancreatic masses [1]. However, the diagnostic performance of EUS-FNA for pancreatic lesions is lower than that for other abdominal or mediastinal lesions, with a sensitivity of 85 % – 90 % and an accuracy of 78 % – 94 % [2] [3] [4] [5] [6]. This can be explained by several factors, including the technical difficulty of the procedure, the experience of the endoscopists and cytopathologists, and the tumor size and consistency [7] [8]. The high rate of false-negative samples leads to repeated procedures and delayed care. Several measures have been proposed to improve sampling and diagnostic accuracy [9] [10]. A real-time sample adequacy evaluation from an onsite cytopathologist has been reported to increase the yield of samples by 10 % – 15 % [11]. Nevertheless, because of increased expenses and a longer procedure time, an onsite cytopathologist may not be available at all centers [12].

EUS-FNA for pancreatic masses is usually performed with a 22G aspiration needle, which allows a very limited sample of tissue to be obtained, thereby potentially reducing diagnostic performance [13]. In addition, European Society of Gastrointestinal Endoscopy (ESGE) guidelines do not recommend the use of 19G needles for transduodenal biopsies [10].

Recently, a new core biopsy needle has been developed in order to improve sample quality (ProCore; Wilson-Cook Medical, Winston-Salem, North Carolina, USA). Sampling with this new needle is referred to in the literature as EUS-guided fine needle biopsy (EUS-FNB), as opposed to sampling with standard needles referred to as EUS-FNA. Prospective reports regarding the use of 19G EUS-FNB in a non-comparative study yielded a final accuracy of 89.4 % for solid pancreatic lesions [14]. Other potential advantages are a lower number of needle passes required to establish the diagnosis and a reduction in procedure time.

To deal with the issue of the 19G technical limit for transduodenal procedures, the FNB device has been produced as a 22G device with a special method of use proposed by the designer.

The primary endpoint of this study was to compare the diagnostic accuracy of EUS-guided sampling using the 22G standard needle with that using the 22G core needle in patients with solid pancreatic masses. The secondary endpoints were the rate of technical failure and the cytohistological quality of the samples obtained using each needle type.


#

Materials and methods

Study design and population

This multicenter prospective study evaluated consecutive patients with solid pancreatic masses referred for EUS-guided sampling to 17 expert endoscopy centers from January to October 2012. Written informed consent for the EUS procedure was obtained from all patients. The study was carried out in accordance with the Helsinki Declaration and was approved by the ethical committee Sud Mediterranée V of the University Hospital of Nice, France (number 11.052) and the French National Security Drug Agency (ANSM 2011-A00578 – 33). It is registered in the ClinicalTrial.gov database (NCT01479803) and followed the STARD and CONSORT statements.

Adult patients with a suspicion of a pancreatic mass based on clinical symptoms (pain, jaundice, weight loss) and/or radiological findings (a pancreatic mass and/or dilated common bile duct and/or a dilated pancreatic duct) were eligible. Patients were excluded from the study if they had predominantly cystic pancreatic lesions (cystic component of more than 50 % of the mass on imaging), according to the ESGE guidelines [1], coagulation disorders (prothrombin time [Quick value] < 60 %, partial thromboplastin time > 42 seconds, and platelets < 60 000/mm3), or were pregnant.


#

EUS-guided biopsy

EUS-guided biopsies were performed on patients who were under general propofol-induced anesthesia and positioned in the left lateral decubitus position using a convex-array echoendoscope (Pentax EG-3870UTK, Pentax Europe GmbH, Hamburg, Germany; Olympus UCT-140, Olympus Europe Inc., Hamburg, Germany; or Fujinon EG-530UT, Fujinon Europe GmbH, Willich, Germany). All biopsies were carried out by senior endoscopists, each with experience of more than 250 procedures.

Computer-generated randomization assignments that established the order of needle use were placed in sealed envelopes and opened locally during the procedure when the patient matched the inclusion criteria. Both needles were used in each patient. The first needle used was the 22G standard FNA needle (EchoTip; Cook Medical Inc., Limerick, Ireland) in group 1, with the 22G FNB needle used first in group 2.

The mass was initially defined endosonographically and the area was scanned using color and pulsed Doppler to detect interposed vessels and/or vascularization in the lesion. The needle was then advanced into the target lesion under ultrasound guidance.

According to the manufacturer’s instructions, for EUS-FNB procedures, once the lesion had been penetrated by the FNB needle, the stylet was removed and suction was applied for 20 seconds using a 10-mL syringe. The needle was then moved back and forth through the lesion 10 times using the fanning method, while maintaining suction throughout the pass [15]. Suction was stopped before the needle was removed. Only one needle pass was performed.

For the EUS-FNA procedure, after the stylet had been removed, suction was applied and the needle was moved back and forth through the lesion 10 times using the fanning method [15]. Suction was stopped before the needle was removed. Two needle passes were performed.

The tissue samples obtained with each needle were placed in a preweighed bottle with a liquid-based preparation (CytoLyt ThinPrep; Hologic Corp, Bedford, Massachusetts, USA) by flushing the needle with 5 mL of CytoLyt solution (one bottle per needle; two bottles per patient). So as not to distort the comparative measurements of the final weight of the specimens, no direct air-dried smears were made. The vials containing the tissue samples were immediately sent for prospective cytohistological analysis, by cytotechnologists trained in liquid-based cytology, to the University Hospital of Nice, France.


#

Cytohistological preparations

Each bottle of CytoLyt was weighed before and after filling with tissue. Each bottle (containing the sample from one needle) was prepared for cytological and a histological analysis as follows:

For liquid-based cytology, the specimen collected in CytoLyt solution (a transport medium) was centrifuged at 1200g for 5 minutes. The supernatant was decanted and the cell button was then resuspended into a fixative solution (PreservCyt solution; Cytyc Corp., Marlborough, Massachusetts, USA) and fixed for at least 30 minutes. A portion of the resuspended PreservCyt cell-fluid sample (20 mL) was transferred to a ThinPrep2000 automated slide processor (Hologic, Marlborough, Massachusetts, USA) according to the manufacturer’s recommendations. The slides were immediately fixed in 95 % ethanol and were stained by the Papanicolaou method.

For histology sample preparation, the entire residual PreservCyt fluid sample was used for paraffin cell-block preparation and for immunohistochemical testing. Cell blocks were prepared with the Shandon Cytoblock kit (Shandon, Pittsburgh, Pennsylvania, USA) according to the modified technology developed at the Gustave Roussy cytology unit, Villejuif, France (M.F., unpublished method). The PreservCyt fluid sample was centrifuged at 1500 rpm for 10 minutes. Two to four drops of reagent 2 of the Cytoblock kit were added to the pellet, which was resuspended and incubated for 1 minute, and then two to four drops of reagent 1 of the Cytoblock kit were added, causing a polymerization of frost. After 1 minute, the button of cells formed was put in a MESH cassette (Sakura, Alphen aan den Rijn, The Netherlands), then fixed in formalin and embedded in paraffin. The sections were cut at 4 μm and were stained with hematoxylin and eosin (H&E). Antibodies for pan-cytokeratin, synaptophysin, chromogranin A, P53, and Ki67 were applied when necessary.


#

Cytohistological analysis

The biopsy specimens were independently reviewed by two experienced cytohistopathologists (M.C.S.P. and M.F. who report more than 300 fine needles aspirations per year), who were informed of the location of the fine needle aspiration. They were blinded to the type of needle and its order of use; however, they evaluated the specimens from both needles with the knowledge that they were from the same patient. An additional consensus reading was performed in discordant cases.

A visual analog scale (VAS) and a sample scoring system (Mair score) were used to assess the cytological/histological quality [16]. The two sampling techniques were compared using five objective parameters contained in the Mair score: (i) the amount of diagnostic cellular material present; (ii) the retention of an appropriate architecture and a cellular arrangement; (iii) the degree of cellular degeneration; (iv) the degree of cellular trauma; and (v) the volume of obscuring background blood and clots (contamination).

The VAS of any unsatisfactory sample was taken into account in the global quality analysis (an intention-to-diagnose analysis). The Mair scores were calculated only for cases where adequate material for the cytological and histological preparations had been obtained from each needle (per protocol analysis). Finally, each cytopathologist evaluated their global satisfaction and preference between the samples collected from each needle (sample A > sample B, sample A = sample B, sample A < sample B).


#

Definitions and endpoints

The primary endpoint was to compare the overall accuracy for the detection of malignancy using 22G EUS-FNA and the 22G EUS-FNB needles. The secondary endpoints were the rate of technical failure and the cytological/histological quality of the obtained samples. The cytological and histological findings were classified as follows: (i) samples that were negative, indeterminate, and suspicious for malignancy were considered negative; (ii) only samples that were unequivocally positive for malignancy were considered positive. (iii) neuroendocrine neoplasms were considered malignant [17]. Any specimens that contained inadequate material (unsatisfactory) were not excluded from the analysis and were considered negative for malignancy (an intention-to-diagnose analysis). A corrected analysis including only those cases with adequate material was also carried out.

The gold-standard diagnosis was composite and was based on at least one of the following reference methods: (i) the unequivocal cytological and/or histological positivity of EUS-FNA and/or EUS-FNB; (ii) a final definite benign or malignant cytological and/or histological diagnosis based on percutaneous (ultrasonography-guided or computed tomography [CT]-guided specimens), endoscopic (endoscopic retrograde cholangiopancreatography [ERCP] biliary cytology), or surgical resection specimens from operated patients; (iii) cytological or histological findings without any proof of malignancy and clinical follow-up with repeated imaging for at least 12 months. The combined safety of use (both needles being used on the same mass) was evaluated during follow-up. Technical failure was defined as the inability to perform the procedure, including the need to change the needle device (type or size).


#

Follow-up

Patients were monitored postoperatively for at least 6 hours before discharge. Any symptoms reported by the patient during the recovery and follow-up time were carefully assessed and documented by the endoscopist. Serious adverse events were defined as: complications resulting in a physician or emergency department visit; an episode of hospitalization or extension of an existing hospital stay; significant disability or death [18]. Serum amylase and lipase were measured for patients with abdominal pain, and an abdominal CT scan was indicated if symptoms persisted.

All patients were followed-up by phone calls or physical examination at 1 month and 6 months. Repeat and/or alternative procedures were required before the first follow-up visit (at 1 month) in patients where false-negative diagnoses were suspected. An additional phone call was made 1 year after the procedure to patients who were negative for malignancy after 6 months.


#

Statistical analysis

Given the literature, we hypothesized a 70 % overall accuracy with the EUS-FNA needle and an 88.5 % accuracy with the EUS-FNB needle [19]. The sample size was calculated in order to detect this 18.5 % difference in diagnostic accuracy between the needles with a McNemar’s test for equality of paired proportions with a 0.05 two-sided significance level and an 80 % power. In the case of paired studies with a binary endpoint, in addition to the usual hypothesis, the proportion of discordant pairs expected is needed to calculate the sample size. As recommended by Machin [20], we assumed that the response with the EUS-FNA needle was independent of the response with the EUS-FNB needle in each subject (conservative hypothesis regarding the sample size needed). We obtained a proportion of 34 % of discordant pairs expected, resulting in a sample size of 72 patients using a sample-size calculator (Nquery Advisor 7.0 software) for a paired study with binary endpoint. Anticipating some possible loss to follow-up, we added an additional 10 % to the sample size, leading to a requirement for 80 patients.

No sequence effect due to the crossover design was expected in this trial. Nevertheless, an order effect due to the use of the needles was assessed before each analysis. In case of a suspected order effect, Prescott’s test was used to compare paired proportion; otherwise, McNemar’s test was used. Continuous variables were compared using a paired t test (no order effect was found). All the tests were two-sided and the significance level was set to 5 %. Statistical analyses were performed using SAS Enterprise Guide 4.1 (Copyright 1999 – 2006 by SAS Institute Inc., Cary, North Carolina, USA).


#
#

Results

A total of 80 patients (49 men, 31 women; mean age 67.1 ± 11.1, range 41 – 86) were enrolled in the study and randomized ([Fig. 1]). Eight patients (10 %) had a family history of pancreatic cancer. Chronic aspirin intake was recorded in eight patients (10 %) at the time of the procedure.

Zoom Image
Fig. 1 Flow diagram showing the randomization and outcomes of the 80 study participants with solid pancreatic masses who were investigated by endoscopic ultrasound (EUS)-guided fine needle aspiration (EUS-FNA) and EUS-guided fine needle biopsy (EUS-FNB). SD, standard deviation; CT, computed tomography.

The EUS features of the tumors are presented in [ Table 1 ]. The tumors were most frequently located in the pancreatic head and the uncinate process (n = 50; 62.5 %). The mean tumor size was 33.9 mm (range 11 – 60 mm) with a large majority of the lesions being 2 cm or larger (n = 73; 91 %). Signs of chronic pancreatitis upstream to the mass were observed according to the Rosemont classification in 24 patients (30 %). Color and pulsed Doppler signals were detected in 15 lesions (19 %) in favor of vascularized tumors.

Table 1

Endoscopic ultrasound (EUS) features of the lesions and final diagnosis in 80 patients with solid pancreatic masses.

Characteristic

Result

Number of lesions per patient, n (%)

 Single

 Multiple

 

76 (95)

 4 (5)

Size of mass on EUS, mean ± SD (range), mm

33.9 ± 10.8 (11 – 60)

Location, n (%)

 Head/uncinate

 Body/tail

 

50 (62.5)

30 (37.5)

Definitive histology, n (%)

 Adenocarcinoma

 Neuroendocrine neoplasm

 Metastasis (small cell lung carcinoma)

 Chronic pancreatitis

 Autoimmune pancreatitis

 Serous cystadenoma

 

62 (77.5)

 6 (7.5)

 2 (2.5)

 6 (7.5)

 3 (3.75)

 1 (1.25)

SD, standard deviation.

The final diagnoses of all the pancreatic masses are listed in [Table 1], with 70 lesions (87.5 %) being considered definitively malignant and 10 lesions (12.5 %) benign. A combined cytohistological analysis for both needles allowed the final diagnosis (malignant or benign) in 70 patients.

Among the 10 patients who were negative for a malignant diagnosis using EUS-guided sampling, four were confirmed as adenocarcinoma and six as benign lesions (autoimmune pancreatitis [n = 3], chronic pancreatitis [n = 2], and serous cystadenoma [n = 1]). The adenocarcinomas were confirmed by CT-guided biopsy (n = 1), ultrasound-guided biopsy (n = 1), pathology of a surgical pancreatic resection (n = 1), and lastly cytology of biliary brushings from the common bile duct taken during ERCP. The autoimmune pancreatitis was confirmed in two patients by therapeutic response to corticosteroid and by pathological specimen evaluation after a surgical pancreatic resection in the third. Repeated clinical and imaging follow-up led to the diagnosis of focal chronic pancreatitis in two patients and an atypical serous cystadenoma in one patient.

Primary endpoint

The accuracy for combined analysis of cytology and histology is presented in [Table 2]. No significant difference in accuracy was observed between sampling with the standard needle and with the core needle (92.5 % vs. 90 %; P = 0.68). Overall (intention-to-diagnose, including all cases) and corrected (cases with adequate material only) sensitivities, specificities, and predictive values did not reveal significant differences between the two sampling methods. Discordance with the gold-standard malignant diagnosis was reported in nine patients, all of whom had adenocarcinoma: EUS-FNA positive and EUS-FNB negative (n = 4); EUS-FNA negative and EUS-FNB positive (n = 1); both sampling methods negative for a malignant diagnosis (n = 4). The only false-positive result for malignancy was observed with EUS-FNA (specificity 90 %) and involved a patient with autoimmune pancreatitis. In this patient, the EUS-FNB was not conclusive (indeterminate).

Table 2

Diagnostic performance of 22G endoscopic ultrasound (EUS)-guided fine needle aspiration (EUS-FNA) and 22G EUS-guided fine needle biopsy (EUS-FNB) based on combined cytological and histological analysis in 80 patients with solid pancreatic masses.

Performance indicator,

Intention-to-diagnose (all patients; n = 80)

Samples with adequate material (n = 70)

EUS-FNA

EUS-FNB

EUS-FNA

EUS-FNB

Overall accuracy[1], % (cases/number for analysis)

[95 % CI]

92.5 % (74/80)

[85 % – 96 %]

90 % (72/80)

[81 % – 95 %]

98.6 % (69/70)

[92 – 99.7 %]

94.3 % (66/70)

[86 % – 98 %]

Sensitivity, % (cases/number for analysis)

[95 % CI]

92.9 % (65/70)

[84 % – 97 %]

88.6 % (62/70)

[79 % – 94 %]

98.5 % (64/65)

[92 % – 99.7 %]

93.8 % (61/65)

[85 % – 98 %]

Specificity, % (cases/number for analysis)

[95 % CI]

90 % (9/10)

[60 % – 98 %]

100 % (10/10)

[72 % – 100 %]

100 % (5/5)

[56.5 % – 100 %]

100 % (5/5)

[56.5 % – 100 %]

Positive predictive value, % (cases/number for analysis)

[95 % CI]

98.5 % (65/66)

[92 % – 100 %]

100 % (62/62)

[94 % – 100 %]

100 % (64/64)

[94 % – 100 %]

100 % (61/61)

[94 % – 100 %]

Negative predictive value, % (cases/number for analysis)

[95 % CI]

64.3 % (9/14)

[39 % – 84 %]

55.5 % (10/18)

[34 % – 75 %]

83.3 % (5/6)

[44 % – 97 %]

55.6 % (5/9)

[27 % – 81 %]

CI, confidence interval.

1 P value using McNemar’s test: intention-to-diagnose P = 0.68; cases with adequate material P = 0.25


Agreement of the diagnoses between both needles was obtained in a total of 71 patients, resulting in a high level of concordance (kappa coefficient 0.68).


#

Secondary endpoints

[Table 3] and [Table 4] summarize the results relevant to the secondary endpoints. The overall quality of the cytohistological analysis (VAS) is shown with the average assessments of the two expert pathologists ([Table 3]). No technical failures were observed with either needle and technical difficulty rates were identical.

Table 3

Technical assessment of the samples obtained by endoscopic ultrasound (EUS)-guided fine needle aspiration (EUS-FNA) and EUS-guided fine needle biopsy (EUS-FNB) in patients with solid pancreatic masses.

Variable

EUS-FNA

(n = 80)

EUS-FNB

(n = 80)

P value

Sample weight, mean ± SD, g

0.374 ± 0.245

 0.272 ± 0.174

0.01

Unsatisfactory sample, n (%)

 Cytology

 Histology

 Both

 

17 (21.3)

10 (12.5)

 5 (6.3)

 

14 (17.5)

24 (30.0)

 9 (11.3)

 

0.44

0.0028

0.01

Overall quality

VAS, mean ± SD

 Cytology

 Histology

 

4.61 ± 3.33

4.90 ± 3.14

 

 4.27 ± 3.05

 3.92 ± 3.39

 

0.42

0.0033

Mair score, mean ± SD

 Expert 1

  Cytology

  Histology

 Expert 2

  Cytology

  Histology

 

 

6.06 ± 1.51

6.41 ± 2.03

 

6.33 ± 1.82

6.04 ± 2.01

 

 

 5.87 ± 1.74

 6.21 ± 2.23

 

 6.16 ± 1.84

 6.03 ± 2.48

 

 

0.45

0.046

 

0.47

0.45

VAS, visual analogical scale; SD, standard deviation.

The weight of the material obtained in the EUS-FNA group was significantly higher ([Table 3]). Fewer overall unsatisfactory samples were reported with EUS-FNA (5 vs. 9; P = 0.01), especially with regard to histology (10 vs. 24; P = 0.0028). The analysis of sample quality showed significant superiority of EUS-FNA with two passes with regard to the histological analysis ([Table 3]). In contrast, no difference was observed between EUS-FNA and EUS-FNB in terms of cytological analysis.

The comparison of the distribution between the different methods (cytology, histology, both or none of the two) used to establish the final diagnosis in each cohort showed no significant difference (χ² test, P = 0.1033) ([Table 4]). Histological analysis (alone or in combination with cytology) was possible significantly more often with EUS-FNA: histology-based diagnosis was possible in only 53 /80 patients (66 %) with EUS-FNB and in 66 /80 patients (82.5 %) with EUS-FNA (χ² test, P = 0.0185).

Table 4

Comparison of the methods used to establish the final diagnosis in samples taken by endoscopic ultrasound (EUS)-guided fine needle aspiration (EUS-FNA) and EUS-guided fine needle biopsy (EUS-FNB).

Diagnosis provided by:

EUS-FNA

(n = 80)

EUS-FNB

(n = 80)

P value

Neither, n (%)

 8 (10.0)

13 (16.3)

0.1033[1]

Cytology alone, n (%)

 6 (7.5)

14 (17.5)

Histology alone, n (%)

10 (12.5)

 6 (7.5)

Both, n (%)

56 (70.0)

47 (58.8)

1 χ² test.


The overall evaluation of the satisfaction and the preference between the samples according the cytopathologists showed a significant difference in favor of EUS-FNA (expert 1, P = 0.009; expert 2, P = 0.002). In 43 and 47 cases for expert 1 and expert 2, respectively, the evaluation was better on the sample collected with EUS-FNA; in 15 and 12 cases, respectively, the evaluation was comparable; and in 22 and 21 cases, respectively, the evaluation was better on the sample obtained with EUS-FNB.

In terms of sample quality for histological analysis, using the Mair score, the EUS-FNA group showed a significantly higher yield of diagnostic cell clusters (P = 0.024), appropriate architecture (P = 0.05), and total score (P = 0.046) for at least one expert (expert 1). In contrast, no difference was observed concerning the histological total score for the expert 2 (P = 0.4568) ([ Table 3 ]). These results were not influenced by aspirin intake (in particular the measure of contamination and the volume of obscuring background blood and clots in the Mair score) or by the presence of significant intratumoral necrosis (noted in 14 patients).


#

Complications

There was one episode of mild bleeding after EUS-FNA (1.25 %) and none after EUS-FNB. In total six patients (7.5 %) experienced early complications: abdominal pain requiring analgesics following the EUS procedure without elevation of the serum lipase level (n = 4; 5 %); mild transient bleeding at the puncture site after EUS-FNA sampling (n = 1; 1.25 % of patients; 0.4 % of needle passes); an aspiration event that developed into aspiration pneumonia (n = 1). Serious adverse events occurred in six patients (7.5 %) during the month following the procedure, including cholangitis due to biliary obstruction (n = 5) and aspiration pneumonia (n = 1). No deaths or late complications relating to the usage of the device and/or to the EUS-guided sampling procedure had been recorded at the end of the study.


#

Follow-up

Eighteen patients (22.5 %) died before the end of the study as a consequence of their tumors. The mean duration of the follow-up was 193.2 ± 82.9 days (range 16 – 367 days) for the overall population. The mean duration of the follow-up was 102.6 ± 57.9 days (16 – 204 days) for the deceased population (n = 18) and 219 ± 67.3 days (124 – 367 days) for the patients who were alive at the end of the study (n = 62). Data on chemotherapy, radiation use, and hospice care were not reported and were unavailable in this cohort of patients. No patients were lost to follow-up during the study and only the deceased patients did not complete the entire follow-up.


#
#

Discussion

This multicenter prospective randomized study demonstrated that the diagnostic accuracy when making one pass with a 22G FNB device seems comparable to two passes with a 22G FNA needle for the diagnosis of solid pancreatic masses, with a diagnostic accuracy at least of 90 % for each device. The technical performance was also comparable; however, the quality and quantity of the tissue specimen obtained with one pass of the 22G FNB device was significantly inferior to that obtained with two passes of the 22G FNA needle.

The 22G FNB needle was recently developed to maximize tissue sampling while minimizing the risk of technical failure. Previous evaluations of this device were, however, scarce and discordant [14] [21]. Moreover, in one series, the design of the trial in one single center was not optimal, because it compared two different patients and two kind of lesions, which was likely to result in a significant bias [21].

Our multicenter study failed to demonstrate any superiority in diagnostic yield using the FNB device. In order to reduce patient population bias, we performed the EUS-guided sampling procedure with both needles in the same pancreatic mass, in a randomized crossover fashion. The absence of statistically significant difference between the devices is not due to a lower technical feasibility when using the EUS-FNB as we observed no technical difficulties when performing transgastric or transduodenal passes with the 22G EUS-FNB needle.

Recently the 19G EUS-FNB needle has demonstrated greater feasibility with only two failures among 35 patients with lesions approached via the transduodenal route [14]. However, this study involved only experts in EUS and, in many of the patients examined, they had to push the EUS-FNB needle out of the endoscope when they were in the stomach before advancing it towards the duodenum.

Despite the absence of a statistical difference when compared with EUS-FNA in the current study, one pass only for EUS-FNB showed a high rate of diagnostic accuracy (90 %). This could be explained by the use of the fanning technique, which has recently shown a significant superiority in a randomized trial over the standard approach for EUS-FNA in solid pancreatic masses [15]. Another explanation could be the combined cytology and histology approach recommended by expert cytopathologists, which are complementary, and the implementation of methods to improve the diagnostic value of the EUS-guided samples [22]. Liquid-based cytology was also selected because it was the only technique that allowed comparative analysis of the weight of the samples. Moreover, the ThinPrep method for cytology has a superior sensitivity and accuracy in pancreatic cytology compared with direct smears, as demonstrated in a large series [23]. Finally this result confirms the 88.5 % diagnostic accuracy achieved with the same device in a prospective multicenter study, in which a single needle pass was performed [24].

A strength of this study was its accurate evaluation of the comparative cytohistological quality of the samples obtained, with the reporting experts based in one single institution. By using such an approach, we tried to minimize interobserver bias.

Surprisingly, in these conditions, the tissue samples obtained with EUS-FNB were lighter than those obtained with EUS-FNA, and histology-based diagnosis was possible in only 53 /80 patients (66 %). This rate was lower than the 83 % and 88.5 % reported previously with the same 22G EUS-FNB needle and lower than the 92.9 % noted with the 19G EUS-FNB device [14] [24].

In addition, the quality and the quantity of the specimens collected using 22G EUS-FNB appeared to be significantly inferior for histological analysis. Such inferiority in the quality of the sample was mainly due to the number and the low quality of diagnostic cell clusters as highlighted by the Mair score. Another explanation for these findings is that the beveled needle-tip and the suction technique used with this new needle device could lead to abundant digestive tract contamination. As we did not differentiate the two passes with the FNA needle, the comparison was only possible between one pass of FNB and two passes of FNA.

With regard to the mean weight, we would assume that the amount of material would have been greater with two passes of the FNB needle than with two passes of the FNA needle. Further studies need to be carried out to evaluate the contribution of two passes in increasing the diagnostic accuracy of EUS-FNB.

There were only minor complications (no pancreatitis) reported in our study, confirming the results of earlier studies evaluating 22G and 19G EUS-FNB. Our data were concordant with the well-established safety of the EUS-FNA technique, which has complication rates ranging between 1 % and 2.5 % [18]. Nevertheless, we excluded from this study the puncture of pancreatic cysts, which is well known to present a higher risk of bleeding and pancreatitis. Moreover, despite its cutting method, the use of EUS-FNB does not increase the risk of bleeding, even in patients using aspirin, as reported by eight of our patients (10 %).

This study has several limitations. First, our standardized protocol for EUS-FNB allowed only a single needle pass, as per the manufacturer’s instructions. This limitation was also imposed by the ethics committee.

A major advantage of the beveled needle would be the higher cost-effectiveness of performing a single pass, while maintaining high diagnostic accuracy, in a shorter procedural time. This was recently suggested by Hucl et al. [25] who showed that 22G EUS-FNB required fewer passes to provide an adequate sample and offered a potentially shorter procedure time. This contrasts with other findings, in which more needle passes (five to seven) increased the diagnostic accuracy by increasing tissue acquisition, and gave better sensitivity and specificity [26]. Nevertheless, Bang et al. [21] failed to find any difference between EUS-FNA and EUS-FNB, even using a multiple-passes protocol.

We cannot conclude from the present study that using 22G FNB decreases the number of passes; the trial was designed to show a statistical superiority and was not a non-inferiority study. We cannot conclude that the diagnostic accuracy of EUS-FNB is comparable to the EUS-FNA procedure using a 22G needle because of this statistical consideration. Nevertheless, it is suggested by our results.

Moreover, we do not know what the accuracy of a single FNA pass was and it cannot be excluded that it was as good as the accuracy of a single EUS-FNB pass. In any case, we may conclude that EUS-FNB cannot be more efficient than FNA performed with the best procedure. Therefore the choice of procedure should be based on other factors, such as cost and feasibility in difficult anatomical condition. We can draw no conclusion about the time potentially saved as we did not record the respective times for the EUS-FNB and EUS-FNA procedures.

Second, the number of passes using EUS-FNA was low: limited to two per patient. The limitations with regard to the number of needle passes seemed to be overcome by using a combined cytological and histological analysis, which yielded a high diagnostic accuracy rate in the underlying study. This was also demonstrated by Möller et al. [27] in a large series of solid pancreatic masses, where adequate specimens were obtained in 98.9 % of patients using only one or two passes.

Another limitation is the potentially reduced statistical power in this study caused by the low number of discordant cases. Agreement on the performance of both needles shows a high level of concordance.

A bias of histopathological interpretation cannot be excluded: specimens from both EUS-FNA and EUS-FNB were evaluated together and the pathologists knew that they were from the same patient/tumor. However, because CytoLyt is not a preservative and technical preparation in two different centers could have created bias, we maintained the blinded origin only of the specimen and not of patient. Furthermore obtaining a fast diagnosis was crucial to allow treatment to be started in patients with lesions highly suspicious of pancreatic cancer.

Finally the last issue was to rule out malignancy in patients with a negative biopsy. Although the gold-standard reference method for the diagnosis of a pancreatic mass remains surgery, this approach can be questioned particularly for benign lesions and/or multiple negative biopsies. In our study, the collection of surgical samples from patients with multiple negative pancreatic biopsies was not possible for ethical reasons. Therefore, such patients were clinically followed-up for at least 12 months with repeated imaging performed. Furthermore the reference test was composite and joined by the unequivocally positive EUS-guided samples. Although EUS-FNA is now recommended as the first-line diagnostic procedure in suspected pancreatic cancer (solid pancreatic mass) in the ESGE guidelines [1], we cannot exclude a potential incorporation bias.

In conclusion, these results suggest that the diagnostic accuracy and technical feasibility of one pass with a 22G EUS-FNB needle are comparable with two passes with a 22G EUS-FNA needle. Despite being easier to use and having an overall diagnostic accuracy approaching 90 %, the EUS-FNB device provided a lower overall quality of histological sample. Further trials should be performed to evaluate whether multiple passes with a 22G EUS-FNB device can significantly increase the yield compared with conventional use of the 22G FNA needle.


#
#

Competing interests: None.

  • References

  • 1 Dumonceau JM, Polkowski M, Larghi A et al. Indications, results, and clinical impact of endoscopic ultrasound (EUS)-guided sampling in gastroenterology: European Society of Gastrointestinal Endoscopy (ESGE) Clinical Guideline. Endoscopy 2011; 43: 897-912
  • 2 Puli SR, Bechtold ML, Buxbaum JL et al. How good is endoscopic ultrasound-guided fine-needle aspiration in diagnosing the correct etiology for a solid pancreatic mass? A meta-analysis and systematic review. . Pancreas 2013; 42: 20-26
  • 3 Chen G, Liu S, Zhao Y et al. Diagnostic accuracy of endoscopic ultrasound-guided fine-needle aspiration for pancreatic cancer: a meta-analysis. Pancreatology 2013; 13: 298-304
  • 4 Hewitt MJ, McPhail MJ, Possamai L et al. EUS-guided FNA for diagnosis of solid pancreatic neoplasms: a meta-analysis. Gastrointest Endosc 2012; 75: 319-331
  • 5 Affolter KE, Schmidt RL, Matynia AP et al. Needle size has only a limited effect on outcomes in EUS-guided fine needle aspiration: a systematic review and meta-analysis. Dig Dis Sci 2013; 58: 1026-1034
  • 6 Hartwig W, Schneider L, Diener MK et al. Preoperative tissue diagnosis for tumours of the pancreas. Br J Surg 2009; 96: 5-20
  • 7 Siddiqui AA, Brown LJ, Hong SK et al. Relationship of pancreatic mass size and diagnostic yield of endoscopic ultrasound-guided fine needle aspiration. Dig Dis Sci 2011; 56: 3370-3375
  • 8 Binmoeller KF, Rathod VD. Difficult pancreatic mass FNA: tips for success. Gastrointest Endosc 2002; 56: S86-S91
  • 9 Gimeno-Garcia AZ, Elwassief A. How to improve the success of endoscopic ultrasound guided fine needle aspiration cytology in the diagnosis of pancreatic lesions. J Interv Gastroenterol 2012; 2: 31-36
  • 10 Polkowski M, Larghi A, Weynand B et al. Learning, techniques, and complications of endoscopic ultrasound (EUS)-guided sampling in gastroenterology: European Society of Gastrointestinal Endoscopy (ESGE) Technical Guideline. Endoscopy 2012; 44: 190-206
  • 11 Iglesias-Garcia J, Dominguez-Munoz JE, Abdulkader I et al. Influence of on-site cytopathology evaluation on the diagnostic accuracy of endoscopic ultrasound-guided fine needle aspiration (EUS-FNA) of solid pancreatic masses. Am J Gastroenterol 2011; 106: 1705-1710
  • 12 Pellise UrquizaM, Fernandez-Esparrach G, Sole M et al. Endoscopic ultrasound-guided fine needle aspiration: predictive factors of accurate diagnosis and cost-minimization analysis of on-site pathologist. Gastroenterol Hepatol 2007; 30: 319-324
  • 13 Wittmann J, Kocjan G, Sgouros SN et al. Endoscopic ultrasound-guided tissue sampling by combined fine needle aspiration and trucut needle biopsy: a prospective study. Cytopathology 2006; 17: 27-33
  • 14 Iglesias-Garcia J, Poley JW, Larghi A et al. Feasibility and yield of a new EUS histology needle: results from a multicenter, pooled, cohort study. Gastrointest Endosc 2011; 73: 1189-1196
  • 15 Bang JY, Magee SH, Ramesh J et al. Randomized trial comparing fanning with standard technique for endoscopic ultrasound-guided fine-needle aspiration of solid pancreatic mass lesions. Endoscopy 2013; 45: 445-450
  • 16 Mair S, Dunbar F, Becker PJ et al. Fine needle cytology--is aspiration suction necessary? A study of 100 masses in various sites. . Acta Cytol 1989; 33: 809-813
  • 17 Eltoum IA, Chen VK, Chhieng DC et al. Probabilistic reporting of EUS-FNA cytology: Toward improved communication and better clinical decisions. Cancer 2006; 108: 93-101
  • 18 Eloubeidi MA, Tamhane A, Varadarajulu S et al. Frequency of major complications after EUS-guided FNA of solid pancreatic masses: a prospective evaluation. Gastrointest Endosc 2006; 63: 622-629
  • 19 Napoleon B, Alvarez-Sanchez MV, Gincoul R et al. Contrast-enhanced harmonic endoscopic ultrasound in solid lesions of the pancreas: results of a pilot study. Endoscopy 2010; 42: 564-570
  • 20 Machin D. Sample size tables for clinical studies. 3rd edn. Chichester, UK. Hoboken, NJ: Wiley-Blackwell; 2008
  • 21 Bang JY, Hebert-Magee S, Trevino J et al. Randomized trial comparing the 22-gauge aspiration and 22-gauge biopsy needles for EUS-guided sampling of solid pancreatic mass lesions. Gastrointest Endosc 2012; 76: 321-327
  • 22 Kopelman Y, Marmor S, Ashkenazi I et al. Value of EUS-FNA cytological preparations compared with cell block sections in the diagnosis of pancreatic solid tumours. Cytopathology 2011; 22: 174-178
  • 23 Volmar KE, Vollmer RT, Routbort MJ et al. Pancreatic and bile duct brushing cytology in 1000 cases: review of findings and comparison of preparation methods. Cancer 2006; 108: 231-238
  • 24 Larghi A, Iglesias-Garcia J, Poley JW et al. Feasibility and yield of a novel 22-gauge histology EUS needle in patients with pancreatic masses: a multicenter prospective cohort study. Surg Endosc 2013; 27: 3733-3738
  • 25 Hucl T, Wee E, Anuradha S et al. Feasibility and efficiency of a new 22G core needle: a prospective comparison study. Endoscopy 2013; 45: 792-798
  • 26 LeBlanc JK, Ciaccia D, Al-Assi MT et al. Optimal number of EUS-guided fine needle passes needed to obtain a correct diagnosis. Gastrointest Endosc 2004; 59: 475-481
  • 27 Moller K, Papanikolaou IS, Toermer T et al. EUS-guided FNA of solid pancreatic masses: high yield of 2 passes with combined histologic-cytologic analysis. Gastrointest Endosc 2009; 70: 60-69

Corresponding author

Geoffroy Vanbiervliet, MD MSc
Hôpital de L'Archet 2 – Pôle digestif
151 Route de Saint Antoine de Ginestière
CS 23079
06202 Nice
France   
Fax: +33-492-035928   

  • References

  • 1 Dumonceau JM, Polkowski M, Larghi A et al. Indications, results, and clinical impact of endoscopic ultrasound (EUS)-guided sampling in gastroenterology: European Society of Gastrointestinal Endoscopy (ESGE) Clinical Guideline. Endoscopy 2011; 43: 897-912
  • 2 Puli SR, Bechtold ML, Buxbaum JL et al. How good is endoscopic ultrasound-guided fine-needle aspiration in diagnosing the correct etiology for a solid pancreatic mass? A meta-analysis and systematic review. . Pancreas 2013; 42: 20-26
  • 3 Chen G, Liu S, Zhao Y et al. Diagnostic accuracy of endoscopic ultrasound-guided fine-needle aspiration for pancreatic cancer: a meta-analysis. Pancreatology 2013; 13: 298-304
  • 4 Hewitt MJ, McPhail MJ, Possamai L et al. EUS-guided FNA for diagnosis of solid pancreatic neoplasms: a meta-analysis. Gastrointest Endosc 2012; 75: 319-331
  • 5 Affolter KE, Schmidt RL, Matynia AP et al. Needle size has only a limited effect on outcomes in EUS-guided fine needle aspiration: a systematic review and meta-analysis. Dig Dis Sci 2013; 58: 1026-1034
  • 6 Hartwig W, Schneider L, Diener MK et al. Preoperative tissue diagnosis for tumours of the pancreas. Br J Surg 2009; 96: 5-20
  • 7 Siddiqui AA, Brown LJ, Hong SK et al. Relationship of pancreatic mass size and diagnostic yield of endoscopic ultrasound-guided fine needle aspiration. Dig Dis Sci 2011; 56: 3370-3375
  • 8 Binmoeller KF, Rathod VD. Difficult pancreatic mass FNA: tips for success. Gastrointest Endosc 2002; 56: S86-S91
  • 9 Gimeno-Garcia AZ, Elwassief A. How to improve the success of endoscopic ultrasound guided fine needle aspiration cytology in the diagnosis of pancreatic lesions. J Interv Gastroenterol 2012; 2: 31-36
  • 10 Polkowski M, Larghi A, Weynand B et al. Learning, techniques, and complications of endoscopic ultrasound (EUS)-guided sampling in gastroenterology: European Society of Gastrointestinal Endoscopy (ESGE) Technical Guideline. Endoscopy 2012; 44: 190-206
  • 11 Iglesias-Garcia J, Dominguez-Munoz JE, Abdulkader I et al. Influence of on-site cytopathology evaluation on the diagnostic accuracy of endoscopic ultrasound-guided fine needle aspiration (EUS-FNA) of solid pancreatic masses. Am J Gastroenterol 2011; 106: 1705-1710
  • 12 Pellise UrquizaM, Fernandez-Esparrach G, Sole M et al. Endoscopic ultrasound-guided fine needle aspiration: predictive factors of accurate diagnosis and cost-minimization analysis of on-site pathologist. Gastroenterol Hepatol 2007; 30: 319-324
  • 13 Wittmann J, Kocjan G, Sgouros SN et al. Endoscopic ultrasound-guided tissue sampling by combined fine needle aspiration and trucut needle biopsy: a prospective study. Cytopathology 2006; 17: 27-33
  • 14 Iglesias-Garcia J, Poley JW, Larghi A et al. Feasibility and yield of a new EUS histology needle: results from a multicenter, pooled, cohort study. Gastrointest Endosc 2011; 73: 1189-1196
  • 15 Bang JY, Magee SH, Ramesh J et al. Randomized trial comparing fanning with standard technique for endoscopic ultrasound-guided fine-needle aspiration of solid pancreatic mass lesions. Endoscopy 2013; 45: 445-450
  • 16 Mair S, Dunbar F, Becker PJ et al. Fine needle cytology--is aspiration suction necessary? A study of 100 masses in various sites. . Acta Cytol 1989; 33: 809-813
  • 17 Eltoum IA, Chen VK, Chhieng DC et al. Probabilistic reporting of EUS-FNA cytology: Toward improved communication and better clinical decisions. Cancer 2006; 108: 93-101
  • 18 Eloubeidi MA, Tamhane A, Varadarajulu S et al. Frequency of major complications after EUS-guided FNA of solid pancreatic masses: a prospective evaluation. Gastrointest Endosc 2006; 63: 622-629
  • 19 Napoleon B, Alvarez-Sanchez MV, Gincoul R et al. Contrast-enhanced harmonic endoscopic ultrasound in solid lesions of the pancreas: results of a pilot study. Endoscopy 2010; 42: 564-570
  • 20 Machin D. Sample size tables for clinical studies. 3rd edn. Chichester, UK. Hoboken, NJ: Wiley-Blackwell; 2008
  • 21 Bang JY, Hebert-Magee S, Trevino J et al. Randomized trial comparing the 22-gauge aspiration and 22-gauge biopsy needles for EUS-guided sampling of solid pancreatic mass lesions. Gastrointest Endosc 2012; 76: 321-327
  • 22 Kopelman Y, Marmor S, Ashkenazi I et al. Value of EUS-FNA cytological preparations compared with cell block sections in the diagnosis of pancreatic solid tumours. Cytopathology 2011; 22: 174-178
  • 23 Volmar KE, Vollmer RT, Routbort MJ et al. Pancreatic and bile duct brushing cytology in 1000 cases: review of findings and comparison of preparation methods. Cancer 2006; 108: 231-238
  • 24 Larghi A, Iglesias-Garcia J, Poley JW et al. Feasibility and yield of a novel 22-gauge histology EUS needle in patients with pancreatic masses: a multicenter prospective cohort study. Surg Endosc 2013; 27: 3733-3738
  • 25 Hucl T, Wee E, Anuradha S et al. Feasibility and efficiency of a new 22G core needle: a prospective comparison study. Endoscopy 2013; 45: 792-798
  • 26 LeBlanc JK, Ciaccia D, Al-Assi MT et al. Optimal number of EUS-guided fine needle passes needed to obtain a correct diagnosis. Gastrointest Endosc 2004; 59: 475-481
  • 27 Moller K, Papanikolaou IS, Toermer T et al. EUS-guided FNA of solid pancreatic masses: high yield of 2 passes with combined histologic-cytologic analysis. Gastrointest Endosc 2009; 70: 60-69

Zoom Image
Fig. 1 Flow diagram showing the randomization and outcomes of the 80 study participants with solid pancreatic masses who were investigated by endoscopic ultrasound (EUS)-guided fine needle aspiration (EUS-FNA) and EUS-guided fine needle biopsy (EUS-FNB). SD, standard deviation; CT, computed tomography.