A prospective, comparative trial to optimize sampling techniques in EUS-guided FNA of solid pancreatic masses

f

ORIGINAL ARTICLE: Clinical Endoscopy

A prospective, comparative trial to optimize sampling techniques inEUS-guided FNA of solid pancreatic masses

Jun Kyu Lee, MD, PhD,�,1 Jong Hak Choi, MD,�,2 Kwang Hyuck Lee, MD,2 Kwang Min Kim, MD,2

Jae Uk Shin, MD,2 Jong Kyun Lee, MD, PhD,2 Kyu Taek Lee, MD, PhD,2 Kee-Taek Jang, MD, PhD3

Seoul, Republic of Korea

Background: There is no standardization of the use of suction during puncturing of a target in pancreaticEUS-guided FNA (EUS-FNA). It is also debatable whether expressing aspirates from the needle by the traditionalmethod of reinserting the stylet is more effective than by air flushing, which is easier and safer.

Objective: To optimize sampling techniques in pancreatic EUS-FNA.

Design: Prospective, comparative trial.

Setting: Tertiary-care referral center.

Patients: Eighty-one consecutive patients with solid pancreatic masses.

Intervention: Four punctures were performed for each mass in random order by a 2 � 2 factorial design.Sample quality and diagnostic yield were compared between samples with suction (S�) versus no suction (S-)and expressed by reinserting the stylet (RS) versus air flushing (AF).

Main Outcome Measurements: Sample quality by the number of diagnostic samples, cellularity, bloodiness,and air-drying artifact; diagnostic yield by accuracy, sensitivity, and specificity.

Results: The number of diagnostic samples (72.8% vs 58.6%; P � .001), cellularity (odds ratio [OR] 2.12; 95%confidence interval [CI], 1.37-3.30; P � .001), bloodiness (OR 1.46; CI, 1.28-1.68; P � .001), accuracy (85.2% vs75.9%; P � .004), and sensitivity (82.4% vs 72.1%; P � .005) were higher in S� than in S-. Bloodiness was lowerin AF than in RS (OR 1.16; CI, 1.03-1.30; P � .017).

Limitations: Single-center trial, 2 kinds of needle gauges, and no immediate cytopathology evaluation.

Conclusion: Puncturing with suction and expressing by air flushing may be used preferentially in pancreaticEUS-FNA because they were more effective and convenient techniques. (Clinical trial registration number:NCT01354795.) (Gastrointest Endosc 2013;77:745-51.)

ctgy

r

h

R

CItSS

Rp

EUS-guided FNA (EUS-FNA) is proved to be safe anduseful for tissue sampling of solid pancreatic masses.1,2

The diagnostic yield of pancreatic EUS-FNA is, however,lower than that of other organs or tissues, with accuracy of78% to 94% and sensitivity of 64% to 95%.3-7 It is importantor patients with pancreatic cancer to receive pathology

Abbreviations: AF, samples expressed by air flushing; EUS-FNA, EUS-guided FNA; RS, samples expressed by reinserting the stylet; S�, sampleswith suction during puncturing; S-, samples with no suction duringpuncturing.

DISCLOSURE: This study was supported by the Samsung Medical CenterClinical Research Development Program grant CRS-110-21-1. No otherfinancial relationships relevant to this publication were disclosed.

*Drs Lee and Choi contributed equally to this article.

Copyright © 2013 by the American Society for Gastrointestinal Endoscopy
0016-5107/$36.00 g
www.giejournal.org V

onfirmation because most of them will undergo chemo-herapy and/or radiation therapy instead of curative sur-ery.8 Therefore, further improvement of the diagnosticield is necessary.

The diagnostic yield of EUS-FNA depends on the expe-iences of an endosonographer, the sizes and types of

ttp://dx.doi.org/10.1016/j.gie.2012.12.009

eceived July 9, 2012. Accepted December 13, 2012.

urrent affiliations: Department of Internal Medicine, Dongguk Universitylsan Hospital, College of Medicine, Dongguk University (1), Division of Gas-roenterology, Department of Medicine (2), Department of Pathology (3),amsung Medical Center, Sungkyunkwan University School of Medicine,eoul, Republic of Korea.

eprint requests: Kwang Hyuck Lee, MD, Division of Gastroenterology, De-artment of Medicine, Samsung Medical Center, 50 Irwon-dong, Gangnam-
u, Seoul 135- 710, Republic of Korea.
olume 77, No. 5 : 2013 GASTROINTESTINAL ENDOSCOPY 745

http://dx.doi.org/10.1016/j.gie.2012.12.009

foodbetsdsr

ifwttrnasnse

O

daroscceonassi

adto1

lasbe

S

Optimal sampling techniques in pancreatic EUS-FNA Lee et al

needles, the methods of cytopathology preparations, theavailability of immediate cytopathology evaluation, andthe expertise of a cytopathologist.9 In addition to theseactors, sampling techniques have a pivotal role.10 Becausef lack of evidence, however, there is no standardizationf the use of suction during puncturing of a target. It is alsoebatable whether expressing aspirates from the needley the traditional method of reinserting the stylet is moreffective than by air flushing, which is easier and safer.11 Inhis prospective, comparative trial, we aimed to optimizeampling techniques by comparing sample quality andiagnostic yield between samples with suction (S�) ver-us no suction (S-) during puncturing and expressed byeinserting the stylet (RS) versus by air flushing (AF).

PATIENTS AND METHODS

PatientsPatients with solid pancreatic masses, which were di-

agnosed with CT or magnetic resonance imaging, wereprospectively enrolled at Samsung Medical Center (Seoul,Korea) from September 2010 to March 2011. Patients withthe following conditions were excluded: synchronous le-sions to be aspirated; coagulation disorder (prothrombintime-international normalized ratio �1.5, activated partialthromboplastin time �50 seconds, platelet count �50,000/mm3); history of acute pancreatitis in the preceding 4 weeks;pregnancy; and refusal or inability to provide informedconsent. Patients were monitored closely for possiblecomplications after the procedure. The Institutional Re-view Board approved this trial, and written informed con-sents for voluntary participation were obtained from allpatients before they entered the study. This trial was reg-istered at ClinicalTrials.gov (NCT01354795).

EUS-FNA procedureEUS-FNA was performed with two kinds of needle

gauges (Endocoil with 22-gauge and Echotip with 25-gauge; Cook endoscopy, Winston-Salem, NC). The choiceof a needle was made of an operator’s own will to achievethe safest and most successful puncturing. A mass waspunctured 4 times with the same needle. The needledevice was passed through the biopsy channel of theechoendoscope and advanced into a target lesion underUS guidance. After the stylet was removed, a 10-mL sy-ringe was attached to the hub of the needle for puncturingwith suction, and no syringe was used in cases of punc-turing with no suction. Moving the needle back and forthwithin the lesion was repeated approximately 10 times foreach pass. Suction was applied during the movements andreleased before removal of the needle to avoid contami-nation of GI mucosa and contents for a puncture withsuction. After retracting the needle into the catheter, weexpressed aspirated material in the needle onto glassslides by reinserting the stylet into the needle slowly or by
applying air pressure by using a 10-mL syringe. Air flush- t
746 GASTROINTESTINAL ENDOSCOPY Volume 77, No. 5 : 2013

ng was done without delay and in a slow, controlledashion to prevent drying and splattering. Four puncturesere performed for each mass in random order according

o computer-generated random orders with the followingechniques: puncturing with suction and expressing byeinserting the stylet; with suction and by air flushing; witho suction and by reinserting the stylet; with no suctionnd by air flushing. Smeared slides were fixed in an ab-olute alcohol solution. Smears for cytopathology exami-ations were done by endosonographers trained in thelide preparation techniques. Immediate cytopathologyvaluation was not available.

utcomes of interest and definitionSample quality was assessed by means of the number of

iagnostic samples, cellularity, bloodiness, and air-dryingrtifact. A diagnostic sample was defined as a set of aspi-ates containing adequate cellular material for cytopathol-gy analysis of a mass. Cellularity was graded into 4 levels:canty, no cellular smear; low, �2 clusters of malignantells with a minimum of 10 cells; moderate, about 2 to 4lusters of malignant cells with a minimum of 10 cellsach; high, �4 clusters of malignant cells with a minimumf 10 cells each. Bloodiness also was graded into 4 levels:one, absent blood cells; low, a few blood cells withoutffecting cytopathology diagnosis; moderate, partially ob-cured by blood cells but possible cytopathology diagno-is; high, obscured by blood cells leading to inadequatenterpretation.

Diagnostic yield was evaluated by accuracy, sensitivity,nd specificity, which were calculated by using a finaliagnosis that was defined as a diagnosis obtained fromhe integration of all the results of cytopathology, biopsy,r surgical pathology, or clinical observation after at least2 months with necessary follow-up studies.

The degree of cytopathologic atypia was graded into 5evels: definite for malignancy, suspicious for malignancy,typical, negative for malignancy, inadequate for diagno-is. An experienced cytopathologist (K.-T. J.), who waslinded to the use of suction during puncturing and thexpression techniques, interpreted all of the slides.

tatistical analysisA 2 � 2 factorial design with a 2-way analysis was used

Take-home Message

● In pancreatic EUS-guided FNA, puncturing a target withsuction and expressing aspirates from the needle by airflushing may be used preferentially because they weremore effective and convenient techniques according tothis prospective trial with 81 patients having solidpancreatic masses.

o evaluate effectiveness of the two separate techniques

www.giejournal.org

http://www.clinicaltrials.gov

otpfpaEppecdTf

Lee et al Optimal sampling techniques in pancreatic EUS-FNA

simultaneously; that is, the samples for which suction wasapplied during puncturing (expressed either by reinsertingthe stylet or air flushing) were compared with the samplesfor which suction was not applied, and the samples ex-pressed by reinserting the stylet (with or without suction)were compared with the samples expressed by air flush-ing. The 2-way analysis was chosen because the 2 tech-niques should be performed separately in chronologicsequence and would not interact with each other, al-though all of these 4 methods were to be performed inevery patient. Also, it should be mentioned that there wereno interactions involving other technical factors such asthe order of sampling and the needle size, excludingpatient factors that were uncontrollable in nature. Wecalculated descriptive statistics for sensitivity, specificity,accuracy, and the distribution of cellularity and bloodi-ness. The mean values and standard deviation (SD) ofcontinuous variables were presented as mean (SD) andcategorical variables as frequencies. To calculate P valuesfor 2-way comparisons between each method, we usedgeneralized estimating equations with logit-link and un-structured correlation matrix, taking into account the cor-related nature of the 4 samples obtained from the samepatient. P values were further adjusted for multiple com-parisons, and odds ratios (OR) with confidence intervals(CI) were calculated where relevant. These generalizedestimating equations with cumulative logit adjusted forpotential confounders also were used to verify our as-sumption that the use of suction and the expression tech-nique would not interact with each other, nor were thereany interactions involving order of sampling or needlesize.

For the sample size calculations, we expected that thediagnostic performances of the different methods weresimilar. As a consequence, we designed our study as anequivalence study of alternative methods. Also, becausethe objective of each method was to identify tumor cells insamples obtained from the same patient, we tried to esti-mate differences in sensitivity and specificity betweenmethods by comparisons within each patient. We assumedthat when a method had a sensitivity of 80% and a spec-ificity of 80% to identify tumor cells, the 2 methods wouldbe considered equivalent if they could be performedwithin 20% of one another (range of equivalence of 0.80).Also, because about 75% of patients were expected tohave a final diagnosis of malignancy, the calculated sam-ple size was 77, with a power of 90% and a 2-sidedsignificance level of 5%. Data were analyzed by using SPSS18.0 for Windows (SPSS Inc, Chicago, Ill).

RESULTS

Patient characteristicsA total of 85 patients were eligible during the study

period. Two patients were excluded due to refusal. An- i


ther 2 were omitted from the analysis because the in-ended procedures could not be completed because ofoor cooperation. Therefore, the final analyses were per-ormed for a total of 324 punctures from 81 consecutiveatients. Baseline characteristics and the final diagnosisre summarized in Table 1. One patient whose result ofUS-FNA was atypical cells was found to have chronicancreatitis after surgery. Of 4 cases with negative cyto-athology results, 1 patient was diagnosed with pancreaticndocrine tumor and another with metastatic renal cellarcinoma after surgery. The other 2 patients were finallyiagnosed as having pancreatic cancer during follow-up.here were no procedure-related adverse events exceptor 2 patients who developed mild acute pancreatitis and

TABLE 1. Baseline characteristics and the finaldiagnoses of the 81 patients who underwent EUS-guided FNA of solid pancreatic masses

Age, mean (SD), y 59.11 (1.7)

Sex (male/female) 51/30

Tumor location, no (%)

Uncinate process 6 (7)

Head 21 (26)

Neck 16 (20)

Body 28 (35)

Tail 10 (12

Tumor size, mean (SD), cm 3.1 (1.2)

Final diagnosis

Malignant, no. (%) 65 (80)

Pancreatic cancer 53

Metastasis 5

Neuroendocrine tumor, malignant 4

Solid-pseudopapillary neoplasm,malignant

2

Miscellaneous 1

Benign, no. (%) 16 (20)

Chronic pancreatitis 3

Tuberculosis 3

Benign lymphoid tissue 2

Neuroendocrine tumor, benign 2

Solid-pseudopapillary neoplasm,malignant

2

Miscellaneous 4

SD, Standard deviation.

mproved with conservative treatment.


NS7..

1s

d(P

D

qinigscf

iFpsdctwai

comp


OutcomesThe number of diagnostic samples (118 [72.8%] of 162

vs 95 (58.6%) of 162; P � .001), cellularity (OR 2.12; 95%CI, 1.37-3.30; P � .001), and bloodiness (OR 1.46; CI,1.28-1.68; P � .001) were higher in S� than in S- (Table 2).

o air-drying artifact was observed in either group. Also,� was superior to S- in terms of accuracy (85.2% vs5.9%; P � .004) and sensitivity (82.4% vs 72.1%; P �005), although specificity was similar (95.8% vs 100%; P �999).

Bloodiness was greater in RS than in AF (OR 1.16; CI,.03-1.30; P � .017), although the number of diagnosticamples (108 [66.7%] of 162 vs 105 [64.8%] of 162; P �

.608), cellularity (OR 0.99; CI, 0.86-1.14; P � .870), andair-drying artifact (none for both; P � .999) were notifferent (Table 3). There were no differences in accuracy79.6% vs 81.5%; P � .582), sensitivity (75.7% vs 78.8%;

� .455), and specificity (100% vs 95.8%; P � .999)between RS and AF. Also, it was ascertained that the use ofsuction and the expression technique would not interactwith each other, nor were there any interactions involvingorder of sampling or needle size (Supplementary Tables

TABLE 2. Sample quality and diagnostic yield according to thesolid pancreatic masses

With suc(n � 16

Sample quality

Diagnostic samples, no (%) 118 (72

Cellularity, OR (95% CI) 2.12 (1.37-

High, no (%) 29 (17.

Moderate, no (%) 58 (35.

Low, no (%) 42 (25.

Scanty, no (%) 33 (20.

Blood contamination, OR (95% CI) 1.46 (1.28-

Scanty, no (%) 20 (12.

Low, no (%) 88 (54.

Moderate, no (%) 44 (27.

High, no (%) 10 (6.2

Air-drying artifact, no (%) 0 (0)

Diagnostic yield (%)

Accuracy 85.2

Sensitivity 82.4

Specificity 95.8

OR, Odds ratio; CI, confidence interval.*P values were calculated from generalized estimation equations with logit-linof the 4 samples obtained from each patient and further adjusted for multiple

S1-S3, available online at www.giejournal.org). d


ISCUSSION

This prospective, comparative trial showed that sampleuality was better when suction was used during punctur-ng of a target than when no suction was used because theumber of diagnostic samples and cellularity were highern S� than in S-. The diagnostic yield turned out to bereater when suction was used because the accuracy andensitivity of S� were higher than those of S-. For theomparisons of expression techniques, there were no dif-erences except for lower bloodiness in AF than in RS.

It is controversial whether the use of suction wouldmprove sample quality and/or diagnostic yield in EUS-NA. Currently, it is usual practice to use suction duringuncturing of a target. Thomson12 supports the use ofuction by suggesting that the purpose of suction is not toraw cells into the needle but to hold the tissue against theutting edge of the needle as it is moved through theissue. On the other hand, it is possible that suction wouldorsen sample quality by bringing in more blood as wells more cells. As yet, the evidence for clarifying this issues limited. Bhutani et al13 published the first article that

f suction during puncturing of a target in EUS-guided FNA of

Without suction(n � 162) P value*

95 (58.6) .001

1 �.001

7 (4.3)

51 (31.5)

60 (37.0)

44 (27.2)

1 �.001

48 (29.6)

85 (52.5)

28 (17.3)

1 (0.6)

0 (0) .999

75.9 .004

72.1 .005

100 .999

unstructured correlation matrix, taking into account the correlated naturearisons.

use o

tion2)

.8)

3.30)

9)

8)

9)

4)

1.68)

3)

3)

2)

)

k and

iscussed the use of suction and reported that continuous

www.giejournal.org

http://www.giejournal.org

tpielrb

iptacttw

cbHtlIie

tigitigcficrcr

comp


rather than intermittent suction provided optimal cellular-ity in EUS-FNA of mediastinal lymph nodes. Puri et al14

performed a controlled trial in which 52 masses wererandomized to with or without suction and showed thatsensitivity and negative predictive value were higher whensuction was used. Wallace et al,15 however, concluded thathe traditional technique of applying suction did not im-rove diagnostic accuracy and worsened specimen blood-

ness in a study with 46 masses. Most of the patientsnrolled in the studies by Puri et al and Wallace et al hadymph nodes, and the data about pancreatic cancer—elatively lower cellularity from dense infiltration of fi-rotic tissue makes the histologic diagnosis difficult16—are

much more limited. In a single-arm observational study byLarghi et al17 with 27 masses, 17 of which were pancreatic,t was found that tissue acquisition by use of high negativeressure suction had a high yield for the retrieval of coreissue samples. Storch et al18 conducted the only compar-tive study, with 53 solid masses, 23 of which were pan-reatic. Four passes were performed for each mass, andhe first 2 passes were done with suction and the addi-ional 2 passes with no suction. They concluded that there

TABLE 3. Sample quality and diagnostic yield according to the

Reinsertin(n �

Sample quality

Diagnostic samples, no (%) 108

Cellularity, OR (95% CI) 0.99 (0.

High, no (%) 18 (

Moderate, no (%) 55 (

Low, no (%) 48 (

Scanty, no (%) 41 (

Blood contamination, OR (95% CI) 1.16 (1.

Scanty, no (%) 28 (

Low, no (%) 86 (

Moderate, no (%) 42 (

High, no (%) 6 (

Air-drying artifact, no (%) 0

Diagnostic yield (%)

Accuracy 79

Sensitivity 75

Specificity 1

OR, Odds ratio; CI, confidence interval.*P values were calculated from generalized estimation equations with logit-linof the 4 samples obtained from each patient and further adjusted for multiple

ere no differences in sample quality and diagnostic ac- i


uracy and that the decision to use suction or not shoulde left to the discretion of an individual endosonographer.owever, the sample sizes of these studies were too small

o draw firm conclusions. Our trial enrolled a sufficientlyarge number of patients to provide 90% statistical power.n addition, the use of suction or no suction was random-zed for each pass in our trial, unlike in the study by Storcht al.

Furthermore, there is a more severe lack of results ofhe optimal expression technique despite their possiblenfluences, as suggested by Savides.9 Wani et al19 sug-ested that the use of a stylet may be useful for expressionn a retrospective study that compared specimens ob-ained with and without a stylet. However, their mainnterest was the effects of a stylet as a needle traversed theut wall, and the influences on expression were not dis-ussed in detail. To the best of our knowledge, ours is therst randomized trial that prospectively compared the out-omes of 2 expression techniques prospectively. Cur-ently, reinserting the stylet is a technique of the mostommon use, but it is labor intensive and may increase theisk of accidental needle stick injury.9-11 Instead, air flush-

ssion technique in EUS-guided FNA of solid pancreatic masses

stylet Air flushing(n � 162) P value*

105 (64.8) .608

4) 1 .870

18 (11.1)

54 (33.3)

54 (33.3)

36 (22.2)

0) 1 .017

40 (24.7)

87 (53.7)

30 (18.5)

5 (3.1)

0 (0) .999

81.5 .582

78.8 .455

95.8 .999

unstructured correlation matrix, taking into account the correlated naturearisons.

expre

g the162)

(66.7)

86-1.1

11.1)

34.0)

29.6)

25.3)

03-1.3

17.3)

53.1)

25.9)

3.7)

(0)

.6

.7

00

k and

ng is an easier and safer method to express aspirated


wcit

sav

A

mBt

11

11

1

1

1

1

1

1

2

2

2

2

2


material. However, the material may spread out uncon-trollably or clot, in addition to the risk of air-drying artifact.According to our results, air flushing in a slow, controlledfashion was superior to reinserting the stylet becausebloodiness was lower in AF than in RS, although no air-drying artifact was noticed in either group. It is thoughtthat the traditional technique of reinserting the stylet is stillvaluable because it can be reserved for cases in whichaspirates cannot be expelled because of drying or clotting.This also is supported by the work of Sahai et al,20 inwhich the results of EUS-FNA with and without the styletwere prospectively compared in 135 solid masses. Theyexpressed all of the samples by air flushing and discussedthat clotting was rare if aspirated material was expelledwithout inordinate delay, which was in agreement withour observation.

Our trial has a few limitations. At first, we used 2 kindsof needle gauges, which might have confounded results.However, it is unlikely that the influences were significantbecause a mass was punctured 4 times with the sameneedle, and each puncture was analyzed with generalizedestimating equations considering matching, which wassupported by the supplementary data. Also, a large num-ber of studies concluded that there were no differences inthe diagnostic yield between 25-gauge and 22-gaugeneedles.21-26 With respect to another aspect of the designof the study that could prompt concern, there were noindications of any interactions involving order of samplingin our own separate examination of outcomes by order ofadministration. Second, immediate cytopathology evalua-tion, which is one of the important factors determiningdiagnostic yield, was not used in our trial.27 There are stillmany centers, like ours, that do not use immediate cyto-pathology evaluation because of increased procedure timeand cost.28 We believe our findings also would be note-

orthy to an endosonographer working with an on-siteytopathologist by suggesting the techniques for decreas-ng the number of needle passes, which is directly relatedo additional use of resources.

In conclusion, puncturing during suction and expres-ion by air flushing may be used preferentially in pancre-tic EUS-FNA because they were more effective and con-enient techniques.

CKNOWLEDGMENT

The authors wish to thank Eliseo Guallar, MD (Depart-ent of Epidemiology and Medicine, Johns Hopkinsloomberg School of Public Health) for his contribution in

he statistical analysis of the data.

REFERENCES

1. Gan SI, Rajan E, Adler DG, et al. Role of EUS. Gastrointest Endosc 2007;66:425-34.

2. Dumonceau JM, Polkowski M, Larghi A, et al. Indications, results, and
clinical impact of endoscopic ultrasound (EUS)-guided sampling in gas-

troenterology: European Society of Gastrointestinal Endoscopy (ESGE)Clinical Guideline. Endoscopy 2011;43:897-912.

3. Gress F, Gottlieb K, Sherman S, et al. Endoscopic ultrasonography-guided fine-needle aspiration biopsy of suspected pancreatic cancer.Ann Intern Med 2001;134:459-64.

4. Fritscher-Ravens A, Topalidis T, Bobrowski C, et al. Endoscopicultrasound-guided fine-needle aspiration in focal pancreatic lesions: aprospective intraindividual comparison of two needle assemblies. En-doscopy 2001;33:484-90.

5. Harewood GC, Wiersema MJ. Endosonography-guided fine needle as-piration biopsy in the evaluation of pancreatic masses. Am J Gastroen-terol 2002;97:1386-91.

6. Eloubeidi MA, Chen VK, Eltoum IA, et al. Endoscopic ultrasound-guidedfine needle aspiration biopsy of patients with suspected pancreatic can-cer: diagnostic accuracy and acute and 30-day complications. Am J Gas-troenterol 2003;98:2663-8.

7. Savides TJ, Donohue M, Hunt G, et al. EUS-guided FNA diagnostic yieldof malignancy in solid pancreatic masses: a benchmark for quality per-formance measurement. Gastrointest Endosc 2007;66:277-82.

8. Hidalgo M. Pancreatic cancer. N Engl J Med 2010;362:1605-17.9. Savides TJ. Tricks for improving EUS-FNA accuracy and maximizing cel-

lular yield. Gastrointest Endosc 2009;69:S130-3.0. Erickson RA. EUS-guided FNA. Gastrointest Endosc 2004;60:267-79.1. Rastogi A, Wani S, Gupta N, et al. A prospective, single-blind, random-

ized, controlled trial of EUS-guided FNA with and without a stylet. Gas-trointest Endosc 2011;74:58-64.

2. Thomson HD. Thin needle aspiration biopsy. Acta Cytol 1982;26:262-3.3. Bhutani MS, Suryaprasad S, Moezzi J, et al. Improved technique for per-

forming endoscopic ultrasound guided fine needle aspiration of lymphnodes. Endoscopy 1999;31:550-3.

4. Puri R, Vilmann P, Saftoiu A, et al. Randomized controlled trial ofendoscopic ultrasound-guided fine-needle sampling with or with-out suction for better cytological diagnosis. Scand J Gastroenterol2009;44:499-504.

5. Wallace MB, Kennedy T, Durkalski V, et al. Randomized controlled trial ofEUS-guided fine needle aspiration techniques for the detection of ma-lignant lymphadenopathy. Gastrointest Endosc 2001;54:441-7.

6. Bachem MG, Schunemann M, Ramadani M, et al. Pancreatic carcinomacells induce fibrosis by stimulating proliferation and matrix synthesis ofstellate cells. Gastroenterology 2005;128:907-21.

7. Larghi A, Noffsinger A, Dye CE, et al. EUS-guided fine needle tissue ac-quisition by using high negative pressure suction for the evaluation ofsolid masses: a pilot study. Gastrointest Endosc 2005;62:768-74.

8. Storch IM, Sussman DA, Jorda M, et al. Evaluation of fine needle aspira-tion vs. fine needle capillary sampling on specimen quality and diagnos-tic accuracy in endoscopic ultrasound-guided biopsy. Acta Cytol 2007;51:837-42.

9. Wani S, Gupta N, Gaddam S, et al. A comparative study of endoscopicultrasound guided fine needle aspiration with and without a stylet. DigDis Sci 2011;56:2409-14.

0. Sahai AV, Paquin SC, Gariepy G. A prospective comparison of endo-scopic ultrasound-guided fine needle aspiration results obtained in thesame lesion, with and without the needle stylet. Endoscopy 2010;42:900-3.

1. Siddiqui UD, Rossi F, Rosenthal LS, et al. EUS-guided FNA of solid pan-creatic masses: a prospective, randomized trial comparing 22-gaugeand 25-gauge needles. Gastrointest Endosc 2009;70:1093-7.

2. Camellini L, Carlinfante G, Azzolini F, et al. A randomized clinical trialcomparing 22G and 25G needles in endoscopic ultrasound-guided fine-needle aspiration of solid lesions. Endoscopy 2011;43:709-15.

3. Fabbri C, Polifemo AM, Luigiano C, et al. Endoscopic ultrasound-guidedfine needle aspiration with 22- and 25-gauge needles in solid pancreaticmasses: a prospective comparative study with randomisation of needlesequence. Dig Liver Dis 2011;43:647-52.

4. Yusuf TE, Ho S, Pavey DA, et al. Retrospective analysis of the utility of
endoscopic ultrasound-guided fine-needle aspiration (EUS-FNA) in
www.giejournal.org

2

2

2

2


pancreatic masses, using a 22-gauge or 25-gauge needle system: amulticenter experience. Endoscopy 2009;41:445-8.

5. Sakamoto H, Kitano M, Komaki T, et al. Prospective comparative study ofthe EUS guided 25-gauge FNA needle with the 19-gauge Trucut needleand 22-gauge FNA needle in patients with solid pancreatic masses. JGastroenterol Hepatol 2009;24:384-90.

6. Lee JH, Stewart J, Ross WA, et al. Blinded prospective comparison of the
performance of 22-gauge and 25-gauge needles in endoscopic

ultrasound-guided fine needle aspiration of the pancreas andperi-pancreatic lesions. Dig Dis Sci 2009;54:2274-81.

7. Klapman JB, Logrono R, Dye CE, et al. Clinical impact of on-site cytopa-thology interpretation on endoscopic ultrasound-guided fine needleaspiration. Am J Gastroenterol 2003;98:1289-94.

8. Aithal GP, Anagnostopoulos GK, Tam W, et al. EUS-guided tissue sampling:comparison of “dual sampling” (Trucut biopsy plus FNA) with “sequential sam-
pling” (Trucut biopsy and then FNA as required). Endoscopy 2007;39:725-30.
GIE on Twitter

GIE now has a Twitter aacount. Followers will learn when the new issues areposted and will receive up-to-the-minute news as well as links to author inter-views, podcasts, and articles. Search on Twitter for @GIE_Journal and follow allof GIE’s tweets.


Documents

A prospective, comparative trial to optimize sampling techniques in EUS-guided FNA of solid pancreatic masses