- Email: [email protected]
Role of endoscopic ultrasound-guided fine-needle aspiration in the diagnosis of solid pancreatic and peripancreatic lesions: is onsite cytopathology necessary?
HPB
DOI:10.1111/j.1477-2574.2010.00180.x
ORIGINAL ARTICLE
Role of endoscopic ultrasound-guided fine-needle aspiration in the diagnosis of solid pancreatic and peripancreatic lesions: is onsite cytopathology necessary? P. Thomas Cherian, Prasoon Mohan, Abdel Douiri, Philippe Taniere, Rahul K. Hejmadi & Brinder S. Mahon Department of Endoscopy and Imaging, Queen Elizabeth, University Hospital Birmingham NHS Trust, Edgbaston, Birmingham, UK
Abstract
hpb_180
389..395
Objectives: The reported median diagnostic yield from endoscopic ultrasound (EUS) fine-needle aspiration (FNA) cytology is 78% (range 39–93%). The aim of this study is to describe a single-centre experience in the diagnostic work-up of solid pancreatic and peripancreatic masses without the benefit of an onsite cytopathologist. Methods: In a consecutive series of 429 EUS examinations performed over a 12-month period by a single operator, 108 were on non-cystic pancreatic or biliary lesions. Data were collected prospectively and the accuracy of FNA was assessed retrospectively using either surgery or repeat imaging as the benchmark in the presence or absence of malignancy. Results: Of the 108 FNAs, 102 (94%) were diagnostic, four were falsely negative (FN) and two were atypical and considered equivocal. There were 78 pancreatic lesions, of which 65 were true positives (TP), 11 true negatives (TN) and two FN, giving an overall accuracy of 97% (76/78). Of nine periampullary lesions, two were TP, six were TN and one was FN, giving an overall accuracy of 89% (8/9). The sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV) and accuracy of EUS-FNA for pancreatic and periampullary lesions combined were 96%, 100%, 100% [95% confidence interval (CI) 95–100%], 85% (95% CI 62–97%) and 97%, respectively. There were 21 bile duct lesions, of which 10 were TP, eight TN, two atypical and one FN, giving an overall accuracy of 86% (18/21). The sensitivity, specificity, PPV, NPV and accuracy of EUS-FNA for biliary lesions were 91%, 100%, 100% (95% CI 69–100%), 91% (95% CI 59–100%) and 95%, respectively. Conclusions: The diagnostic accuracy of EUS-FNA for pancreatic lesions in our series was 97% and the PPV for the three subgroups of lesion type was 100%; these figures are comparable with the best rates reported in the literature, despite the absence of onsite cytopathology. These rates are potentially a direct result of high-volume practice, dedicated endosonography and cytopathology. These results show that it is possible to achieve high rates of accuracy in places where logistical issues make it impossible to maintain a cytopathologist in the endoscopy suite. In addition, our results contribute to the limited, collective global experience on the effectiveness of EUS-FNA in periampullary and biliary lesions.
Keywords pancreatic mass, cytology, biliary tract, investigation, endoscopic ultrasound Received 17 January 2010; accepted 31 March 2010
Correspondence Brinder S. Mahon FRCR, Consultant Clincal Lead for Endoscopic Ultrasound Service, University Hospital Birmingham NHS Trust, Edgbaston, Birmingham B15 2TH, UK. Tel: + 44 121 472 1311. Fax: + 44 121 627 5886. E-mail: [email protected]
This paper is based on an oral presentation given at the International Hepato-Pancreato-Biliary Association Meeting, 27 February to 2 March 2008, Mumbai.
HPB 2010, 12, 389–395
Introduction In recent years, endoscopic ultrasound (EUS) and EUS-guided fine-needle aspiration (EUS-FNA) have gained acceptance in the
© 2010 International Hepato-Pancreato-Biliary Association
HPB
390
diagnosis of pancreatic and biliary pathology. Endoscopic ultrasound allows for the visualization of the pancreatic head, neck and uncinate process from the duodenum, and the neck, body and tail from the stomach. High-frequency ultrasound and the close proximity of the transducer produce high-resolution images of the pancreas and peripancreatic structures and can detect small lesions that are discriminated with difficulty by computed tomography (CT). Although several series have quoted high rates of accuracy for EUS-FNA in the diagnosis of pancreatic lesions,1–10 the literature on EUS-FNA in biliary and ampullary lesions is limited.11–17 Our study aims to assess the diagnostic accuracy of EUS-FNA in pancreatic, periampullary and biliary lesions in a single, high-volume centre in the absence of an onsite cytopathologist. The paper also analyses the false negative cases in detail to uncover learning points. We hope these results will contribute to the limited, collective global experience on the effectiveness of EUS-FNA in periampullary and biliary lesions, as data on
large numbers of these relatively rare lesions are difficult to amass in individual centres performing EUS-FNA.
Materials and methods All patients who underwent EUS-FNA performed for pancreatic, biliary and periampullary lesions in a single hepato-biliary referral centre, over a 1-year period between March 2006 and March 2007, were identified and reviewed (Figs 1 and 2). Our patient dataset represented a preselected cohort of patients who presented with atypical histories or symptoms and equivocal CT findings (e.g. no double duct sign or mass lesion) or who required a definitive diagnosis although deemed unresectable or unfit for surgery. Those patients with typical symptoms (e.g. painless obstructive jaundice) and a resectable pancreatic mass on CT underwent surgery without EUS. Prospectively collected data from the EUS database and hospital records were used for analysis. All cystic
Total annual UGI EUS-FNA, n = 429
Pancreatic and peripancreatic lesions, n = 196
Cystic pancreatic lesions, n = 88
Non-pancreatic lesions, n = 233
Excluded
Study cohort, n = 108
Solid pancreatic lesions, n = 78; biliary tract lesions, n = 21; periampullary lesions, n = 9
Excluded
Figure 1 Workload in our department over the study year and the selection of the present study population. UGI, upper gastrointestinal; EUS-FNA, endoscopic ultrasound-guided fine-needle aspiration
Solid lesions, n = 108
Biliary lesions, n = 21
Head of pancreas, n = 50 (64%)
Pancreatic lesions, n = 78
Neck of pancreas, n = 13 (17%)
Ampullary lesions, n = 9
Body of pancreas, n = 10 (13%)
Tail of pancreas, n = 5 (6%)
Figure 2 Distribution of solid pancreatic lesions
HPB 2010, 12, 389–395
© 2010 International Hepato-Pancreato-Biliary Association
HPB
391
(A)
(B)
Figure 3 Cytology showing adenocarcinoma cells: (A) pap stain; (B) clot preparation
pancreatic lesions (cystic tumours and pseudocysts) were excluded as the diagnostic work-up for these lesions differed from that for solid tumours. Endoscopic ultrasound was performed under conscious sedation (using midazolam and fentanyl) by a single experienced operator using a curved linear array echo endoscope (Olympus GF-UCT240-AL5; Olympus Corp., Tokyo, Japan), which allows the visualization of the needle tract in real time. The instrument has colour Doppler imaging built in, which helps with the identification and avoidance of vascular structures and, in some situations, aids diagnosis by quantifying the vascularity of the lesion concerned. When the transducer had been positioned, the needle catheter system (EchoTip, 22-gauge needle; Wilson-Cook Medical, Inc., Winston-Salem, NC, USA) was inserted through the working channel of the echo endoscope and the needle advanced into the target lesion under real-time ultrasound guidance. A negative suction of 0–40 ml was applied to the FNA needle using a syringe and the needle was then carefully moved back and forth within the lesion. The aspirate was then expressed onto glass slides and both air-dried and alcohol-fixed smears were prepared by the endoscopist. A further aspirate was usually expressed into 15 ml of CytoRich Red fixative. No cytopathologist was available in the endoscopy suite to examine the adequacy of the sample. One to four needle passes were performed as necessary to achieve an adequate sample as determined by the echo endoscopist. All smears were subsequently examined by two experienced, dedicated cytopathologists with a special interest in pancreatic FNAs. Figures 3 and 4 show example cytology. The results of the FNAs were compared with the final diagnoses as determined by specimen histology in patients undergoing resection, and by serial imaging and clinical follow-up in nonsurgical patients. Lesions were considered to be benign if there was no evidence of progression or resolution on serial imaging at 12 months. Any lesion which progressed on serial imaging or which resulted in clinical deterioration or death was considered malignant. In the current study, results were deemed to represent: a true positive (TP) if both EUS-FNA and histology, serial imaging or follow-up suggested malignancy; a true negative (TN) if both EUS-FNA and histology or serial imaging suggested absence of
HPB 2010, 12, 389–395
malignancy; a false positive (FP) if EUS-FNA showed tumour and the resection specimen found no malignancy or no increase in mass on imaging at 12 months, and a false negative (FN) if EUSFNA found no malignancy, but specimen histology, serial imaging or follow-up confirmed the presence of malignancy. It should be noted that the lesions were categorized into the three groups defined above (pancreatic, biliary and periampullary lesions) according to EUS findings rather than cross-sectional imaging. Although lesions within the bile duct or ampulla were relatively easy to categorize, intrapancreatic lesions were more difficult. Here, the criterion of ‘mass centring on the bile duct or ampulla’ (i.e. the site of uniform circumferential growth) was used to assess the likely origin of the tumour. Therefore, it is possible that some crossover may have occurred because some bile duct cancers can be eccentrically centred around the bile duct.
Results During the study period, 429 EUS-FNAs were performed for various upper gastrointestinal lesions, of which 196 were for pancreatic and peripancreatic pathologies. Eighty-eight of these were cystic lesions and were therefore excluded from the study, leaving 108 EUS-FNAs [78 (72%) were solid pancreatic, 21 (19%) were bile duct and nine (8%) were periampullary lesions] in the final study cohort (Figs 1 and 2). All but two of the 108 EUS were completed as day patient procedures. Overall, 31 patients (29%) had negative cytology (including ‘atypical’ results) on FNA; nine (29%) of these patients underwent surgery and their diagnoses were confirmed by histology, whereas the remaining 22 (71%) underwent serial imaging rather than surgery. Solid pancreatic lesions In the 78 patients with solid pancreatic lesions, 50 (64%) of the lesions were in the head, 13 (17%) in the body, 10 (13%) in the neck and five (6%) in the tail of the pancreas (Fig. 2). The median size of the lesions was 30 mm (range 20–80 mm). The median number of passes during biopsy was two (range 1–4). Among the pancreatic lesions, 65 (83%) EUS-FNAs yielded TP results after comparison with the final diagnosis. The other 13 pancreatic
© 2010 International Hepato-Pancreato-Biliary Association
HPB
392
(A)
(B)
(D)
(C)
Figure 4 Clusters of cells with round nuclei and moderate cytoplasm confirming neuroendocrine neoplasm: (A) pap stain; (B) clot preparation; (C) positive synaptophysin; (D) positive chromogranin
Table 1 Pre- and post-test probabilities of endoscopic ultrasound-guided fine-needle aspiration in the three subgroups. Pre-test probabili-
ties are based on those reported by Heiken21 Pre-test probability
EUS-FNA performance
Disease prevalence in the sample (95% CI)
Sensitivity (95% CI)
Specificity (95% CI)
PPV (95% CI)
Post-test probabilities NPV (95% CI)
Overall performance Accuracy (95% CI)
Pancreatic lesions
86% (76–93%)
97% (90–100%)
100% (72–100%)
100% (95–100%)
85% (55–98%)
97% (91–99%)
Periampullary lesions
33% (8–70%)
67% (9–99%)
100% (54–100%)
100% (16–100%)
86% (42–100%)
89% (67–100%)
Bile duct lesions
52% (30–74%)
91% (59–100%)
100% (69–100%)
100% (69–100%)
91% (59–100%)
95% (81–100%)
Pancreatic and periampullary lesions
80% (71–88%)
96% (88–99%)
100% (81–100%)
100% (95–100%)
85% (62–97%)
97% (92–99%)
EUS-FNA, endoscopic ultrasound-guided fine-needle aspiration; PPV, positive predictive value; NPV, negative predictive value; 95% CI, 95% confidence interval
lesions included 11 TN, two FN and no FP results. The other patients were followed up by imaging for ⱖ12 months. One of the two FN patients underwent a Whipple’s pancreaticoduodenectomy and the other a jejunal resection. Their resection specimen histology confirmed both the lesions to be gastrointestinal stromal tumours (GISTs). Two FNA samples showed atypical cells on cytology, which were later confirmed to represent chronic pancreatitis and pancreatic abscesses. The sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV) and accuracy of EUS-FNA for pancreatic lesions are shown in Table 1 and Fig. 5.
HPB 2010, 12, 389–395
Periampullary lesions Nine EUS-FNAs were performed for periampullary lesions. These yielded two TP, six TN, no FP and one FN result for a lesion which was later proven to be an adenocarcinoma following a Whipple’s procedure. The sensitivity, specificity, PPV, NPV and accuracy of EUS-FNA for periampullary lesions alone and combined with pancreatic lesions are shown in Table 1. Common bile duct lesions A total of 21 patients underwent EUS-FNA for bile duct lesions. Of these, 17 patients had undergone an endoscopic retrograde cho-
© 2010 International Hepato-Pancreato-Biliary Association
HPB
393
100% 80% FalseP
60%
FalseN TrueN
40%
TrueP
20%
pu lla ry Pe ria m
tra ct lia ry Bi
Pa nc re at ic
0%
Figure 5 Results for each of the three study cohorts according to lesion location. Results for two patients in the bile duct cohort were indeterminate and are not included. FalseP, false positive; FalseN, false negative; TrueN, true negative; TrueP, true positive
langiopancreatography (ERCP) prior to EUS. The EUS-FNA results were TP in 10 patients, TN in eight and FN in one. Despite being categorized as FN, a repeat EUS-FNA confirmed the presence of a cholangiocarcinoma in this particular patient. In two patients, the EUS-FNA samples were indeterminate. Both of these were later confirmed to be benign on repeat EUS-FNA performed at 6 and 12 weeks followed by serial imaging. Of the eight patients with TN results, three underwent surgery and diagnosis was confirmed by histology. Diagnoses in the remaining patients were again confirmed by repeat EUS-FNA at 6 and 12 weeks, and imaging follow-up for 12 months. The sensitivity, specificity, PPV, NPV and accuracy of EUS-FNA for biliary lesions are shown in Table 1. In the current series one patient suffered post-procedural haemorrhage. This patient developed a peri-portal haematoma and required surgical exploration. Another patient needed an overnight stay in hospital for pain. No patients re-presented after discharge from the day-case unit with pain or post-procedural pancreatitis and neither was there any procedure-related mortality. The final histology included, in addition to the adenocarcinomas, GISTs (n = 3), adenosquamous tumour (n = 1), neuroendocrine tumours (n = 10) and tuberculosis (n = 1).
Discussion Endoscopic ultrasound-guided FNA has emerged as a major diagnostic adjuvant in the management of pancreatic and biliary lesions. Because of their anatomical location, relatively small pancreatic and bile duct lesions are often not amenable to percutaneous guided biopsy, although this limitation depends on the experience of the individual operator and centre. Endoscopic ultrasound-guided FNA is able to obtain tissue diagnosis in such lesions in a relatively simple and safe manner, almost obviating the
HPB 2010, 12, 389–395
need for surgery for tissue diagnosis. In addition, EUS-FNA significantly decreases the likelihood of tumour seeding via the needle track as the tissue planes traversed are usually part of the potential surgical field. Endoscopic ultrasound-guided FNA of pancreatic lesions was first reported by Vilmann et al. in 1992,18 since when several large series have reported sensitivities of EUS-FNA ranging from 74% to 96% and specificity of close to 100%. Rocca et al. reported a series of 246 patients in which the sensitivity and specificity of EUS-FNA were 79% and 60%, respectively.1 In a series of 300 EUS-FNAs, Varadarajulu et al. noted that sensitivity of EUS-FNA was significantly lower (74% vs. 91%) in patients with chronic pancreatitis.2 The diagnostic accuracy and sensitivity of EUS-FNA for solid pancreatic lesions in the current series are comparable with the best rates reported in the literature. In both of the two patients in whom FN results were obtained, the endoscopist noted concern about either the sample or the lesion at the time of the procedure. In the first patient, EUS showed a highly vascular mass in the head of pancreas and FNA obtained a bloody aspirate which did not show any malignant cells. In view of the significant clinical suspicion, this patient went on to have a pancreaticoduodenectomy and the final histology confirmed a GIST. Endoscopic ultrasound in the second patient showed a dilated pancreatic duct up to the level of the pancreatic head (bile duct thickening was noted despite the absence of a stricture) and an ill-defined hypoechoic lesion in the head of the pancreas, which was aspirated. Imaging follow-up with serial CT demonstrated a pancreatic head mass of increasing size and eventual liver metastases consistent with malignancy. Co-morbidity had prevented initial surgery. Despite the lack of malignant cells on cytology, endoscopic and clinical features prompted further investigation. Although in theory it is possible that an onsite cytopathologist might have enforced further attempts to obtain representative cells, in practice the lack of such a cytopathologist did not change our course of management as no better aspirates were achieved in the first patient despite multiple attempts and it was felt unnecessary to proceed with further attempts in the second because the patient was unfit for intervention. Endoscopic ultrasound-guided FNA of biliary tract lesions has been reported by many series to be useful, especially in determining the presence and extent of the often small lesions and the involvement of regional lymph nodes.11–17 However, it is generally acknowledged that EUS of the proximal bile duct is technically more demanding as a result of the spiral relationship of the duct to the duodenum as it travels past the duodenum. DeWitt et al. studied 24 patients with a proximal biliary stricture and negative ERCP brush biopsy.11 Endoscopic ultrasound enabled the visualization of a mass in 23 (96%) patients and EUS-FNA had a sensitivity of 77%.11 A review by Meara et al. of 46 patients with suspected biliary tree malignancy found EUS-FNA to have sensitivity of 87% and specificity of 100%.12 Sensitivity reported in other studies varies between 43% and 89% and most of these studies were conducted in circumstances that included an onsite
© 2010 International Hepato-Pancreato-Biliary Association
HPB
394
cytopathologist. In the current series, we found one FN result in a patient in whom the US appearance was indeed suggestive of a cholangiocarcinoma, but whose aspirate did not show malignant cells, even when the procedure was repeated a month later. The final diagnosis (cholangiocarcinoma) was confirmed by histology following surgery. The two cases in which cytology was indeterminate were later confirmed to be benign with repeat EUS-FNA at 6 and 12 weeks and serial imaging. Despite the results detailed above, it is important to highlight the nine patients in the present study who underwent surgery despite negative cytology and in whom diagnoses were confirmed by histology at surgery. The procedures carried out in these patients included gastric bypass (n = 2) and Roux-en-Y biliary bypass (n = 7). All of these patients had chronic pancreatitis or acute recurrent pancreatitis. In our experience, although chronic pancreatitis is not diagnosed on cytopathological grounds, cytopathology still has a very high PPV in detecting malignancy. One of the difficulties for the EUS operator in this subgroup of patients concerns knowing where to target sampling in the context of an abnormal gland. The current study did not, however, specifically look at this subset of pancreatic pathology on this occasion. The literature on the use of EUS-FNA for ampullary tumours is limited. In a series of 35 patients, Defrain et al. reported the sensitivity and specificity of EUS-FNA for ampullary tumours to be 82% and 100%, respectively.19 In the current series, the one patient in the periampullary lesion subgroup in whom an FN result was obtained went on to have a resection and was found to have a small focus of adenocarcinoma within the lesion, which would have been unlikely to be sampled even with an onsite cytopathologist. By comparison, the diagnostic accuracy and sensitivity of EUS-FNA for pancreatic lesions were greater than for biliary lesions, which is in line with the published literature. Endoscopic ultrasound-guided FNA is highly operatordependent.20 In the current series, a single endoscopist performed all the procedures, which may have resulted in greater consistency and contributed to the better diagnostic accuracy and sensitivity achieved. No cytopathologist was present during the procedure to evaluate the adequacy of the samples. Despite this, diagnostic rates in the current series are comparable with the best rates reported in the literature and most reported studies utilized an onsite cytopathologist. The PPV for all three subgroups was 100%, which, because it is greater than the actual prevalence in each subgroup, suggests that EUS-FNA increases confidence in the confirmation of the diagnosis. By comparison, CT has some limitations in detecting these lesions, with published PPVs of 76–90%.21 In two instances, bloody or inadequate samples were obtained, both of which were recognized by the endoscopist at the time of biopsy. In all instances of false interpretations, the endoscopist expressed concerns about the sample at the time of the procedure. It is to be noted that some of the credit for the results of the present study reflect the experience of dedicated hepato-biliary pathologists, who were able to make definite diagnoses on the basis of limited samples.
HPB 2010, 12, 389–395
Conclusions The diagnostic accuracy of EUS-FNA for pancreatic lesions in the current series was 97% and the PPVs for the three subgroups were 100%. These rates are comparable with the best rates reported in the literature, despite the absence of onsite cytopathology. We believe that these results are a direct result of high-volume practice and the substantial experience of dedicated endosonographers and cytopathologists. These results indicate that it is possible to achieve high rates of accuracy, even when logistical issues make it impossible to maintain a cytopathologist in the endoscopy suite. Conflicts of interest None declared. References 1. Rocca R, De Angelis C, Daperno M, Carucci P, Ravarino N, Bruno M et al. (2007) Endoscopic ultrasound-fine needle aspiration (EUS-FNA) for pancreatic lesions: effectiveness in clinical practice. Dig Liver Dis 39:768– 774. 2. Varadarajulu S, Tamhane A, Eloubeidi MA. (2005) Yield of EUS-guided FNA of pancreatic masses in the presence or the absence of chronic pancreatitis. Gastrointest Endosc 62:728–736; quiz 751, 753. 3. Iglesias-Garcia J, Dominguez-Munoz E, Lozano-Leon A, Abdulkader I, Larino-Noia J, Antunez J et al. (2007) Impact of endoscopic ultrasoundguided fine needle biopsy for diagnosis of pancreatic masses. World J Gastroenterol 13:289–293. 4. Raut CP, Grau AM, Staerkel GA, Kaw M, Tamm EP, Wolff RA et al. (2003) Diagnostic accuracy of endoscopic ultrasound-guided fine-needle aspiration in patients with presumed pancreatic cancer. J Gastrointest Surg 7:118–126; discussion 127–128. 5. Ylagan LR, Edmundowicz S, Kasal K, Walsh D, Lu DW. (2002) Endoscopic ultrasound guided fine-needle aspiration cytology of pancreatic carcinoma: a 3-year experience and review of the literature. Cancer 96:362–369. 6. Hernandez LV, Mishra G, Forsmark C, Draganov PV, Petersen JM, Hochwald SN et al. (2002) Role of endoscopic ultrasound (EUS) and EUSguided fine needle aspiration in the diagnosis and treatment of cystic lesions of the pancreas. Pancreas 25:222–228. 7. Chhieng DC, Jhala D, Jhala N, Eltoum I, Chen VK, Vickers S et al. (2002) Endoscopic ultrasound-guided fine-needle aspiration biopsy: a study of 103 cases. Cancer 96:232–239. 8. O'Toole D, Palazzo L, Arotçarena R, Dancour A, Aubert A, Hammel P et al. (2001) Assessment of complications of EUS-guided fine-needle aspiration. Gastrointest Endosc 53:470–474. 9. Chang KJ, Nguyen P, Erickson RA, Durbin TE, Katz KD. (1997) The clinical utility of endoscopic ultrasound-guided fine-needle aspiration in the diagnosis and staging of pancreatic carcinoma. Gastrointest Endosc 45:387–393. 10. Wiersema MJ, Vilmann P, Giovannini M, Chang KJ, Wiersema LM. (1997) Endosonography-guided fine-needle aspiration biopsy: diagnostic accuracy and complication assessment. Gastroenterology 112:1087–1095. 11. DeWitt J, Misra VL, Leblanc JK, McHenry L, Sherman S. (2006) EUSguided FNA of proximal biliary strictures after negative ERCP brush cytology results. Gastrointest Endosc 64:325–333. 12. Meara RS, Jhala D, Eloubeidi MA, Eltoum I, Chhieng DC, Crowe DR et al.
© 2010 International Hepato-Pancreato-Biliary Association
HPB
395
(2006) Endoscopic ultrasound-guided FNA biopsy of bile duct and gall-
Thonke F et al. (2004) EUS-guided fine-needle aspiration of suspected
bladder: analysis of 53 cases. Cytopathology 17:42–49.
hilar cholangiocarcinoma in potentially operable patients with negative
13. Rösch T, Hofrichter K, Frimberger E, Meining A, Born P, Weigert N et al. (2004) ERCP or EUS for tissue diagnosis of biliary strictures? A prospective comparative study. Gastrointest Endosc 60:390–396. 14. Byrne MF, Gerke H, Mitchell RM, Stiffler HL, McGrath K, Branch MS et al. (2004) Yield of endoscopic ultrasound-guided fine-needle aspiration of bile duct lesions. Endoscopy 36:715–719. 15. Lee JH, Salem R, Aslanian H, Chacho M, Topazian M. (2004) Endoscopic ultrasound and fine-needle aspiration of unexplained bile duct strictures. Am J Gastroenterol 99:1069–1073.
brush cytology. Am J Gastroenterol 99:45–51. 18. Vilmann P, Jacobsen GK, Henriksen FW, Hancke S. (1992) Endoscopic ultrasonography with guided fine needle aspiration biopsy in pancreatic disease. Gastrointest Endosc 38:172–173. 19. Defrain C, Chang CY, Srikureja W, Nguyen PT, Gu M. (2005) Cytologic features and diagnostic pitfalls of primary ampullary tumours by endoscopic
ultrasound-guided
fine-needle
aspiration
biopsy.
Cancer
105:289–297. 20. Savides TJ, Donohue M, Hunt G, Al-Haddad M, Aslanian H, Ben-
16. Eloubeidi MA, Chen VK, Jhala NC, Eltoum IE, Jhala D, Chhieng DC et al.
Menachem T et al. (2007) EUS-guided FNA diagnostic yield of malig-
(2004) Endoscopic ultrasound-guided fine needle aspiration biopsy of
nancy in solid pancreatic masses: a benchmark for quality performance
suspected cholangiocarcinoma. Clin Gastroenterol Hepatol 2:209– 213. 17. Fritscher-Ravens A, Broering DC, Knoefel WT, Rogiers X, Swain P,
HPB 2010, 12, 389–395
measurement. Gastrointest Endosc 66:277–282. 21. Heiken JP. (2001) Carcinoma of the pancreas: detection and staging using CT and MRI. Cancer Imaging 2:19–22.
© 2010 International Hepato-Pancreato-Biliary Association
DOI:10.1111/j.1477-2574.2010.00180.x
ORIGINAL ARTICLE
Role of endoscopic ultrasound-guided fine-needle aspiration in the diagnosis of solid pancreatic and peripancreatic lesions: is onsite cytopathology necessary? P. Thomas Cherian, Prasoon Mohan, Abdel Douiri, Philippe Taniere, Rahul K. Hejmadi & Brinder S. Mahon Department of Endoscopy and Imaging, Queen Elizabeth, University Hospital Birmingham NHS Trust, Edgbaston, Birmingham, UK
Abstract
hpb_180
389..395
Objectives: The reported median diagnostic yield from endoscopic ultrasound (EUS) fine-needle aspiration (FNA) cytology is 78% (range 39–93%). The aim of this study is to describe a single-centre experience in the diagnostic work-up of solid pancreatic and peripancreatic masses without the benefit of an onsite cytopathologist. Methods: In a consecutive series of 429 EUS examinations performed over a 12-month period by a single operator, 108 were on non-cystic pancreatic or biliary lesions. Data were collected prospectively and the accuracy of FNA was assessed retrospectively using either surgery or repeat imaging as the benchmark in the presence or absence of malignancy. Results: Of the 108 FNAs, 102 (94%) were diagnostic, four were falsely negative (FN) and two were atypical and considered equivocal. There were 78 pancreatic lesions, of which 65 were true positives (TP), 11 true negatives (TN) and two FN, giving an overall accuracy of 97% (76/78). Of nine periampullary lesions, two were TP, six were TN and one was FN, giving an overall accuracy of 89% (8/9). The sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV) and accuracy of EUS-FNA for pancreatic and periampullary lesions combined were 96%, 100%, 100% [95% confidence interval (CI) 95–100%], 85% (95% CI 62–97%) and 97%, respectively. There were 21 bile duct lesions, of which 10 were TP, eight TN, two atypical and one FN, giving an overall accuracy of 86% (18/21). The sensitivity, specificity, PPV, NPV and accuracy of EUS-FNA for biliary lesions were 91%, 100%, 100% (95% CI 69–100%), 91% (95% CI 59–100%) and 95%, respectively. Conclusions: The diagnostic accuracy of EUS-FNA for pancreatic lesions in our series was 97% and the PPV for the three subgroups of lesion type was 100%; these figures are comparable with the best rates reported in the literature, despite the absence of onsite cytopathology. These rates are potentially a direct result of high-volume practice, dedicated endosonography and cytopathology. These results show that it is possible to achieve high rates of accuracy in places where logistical issues make it impossible to maintain a cytopathologist in the endoscopy suite. In addition, our results contribute to the limited, collective global experience on the effectiveness of EUS-FNA in periampullary and biliary lesions.
Keywords pancreatic mass, cytology, biliary tract, investigation, endoscopic ultrasound Received 17 January 2010; accepted 31 March 2010
Correspondence Brinder S. Mahon FRCR, Consultant Clincal Lead for Endoscopic Ultrasound Service, University Hospital Birmingham NHS Trust, Edgbaston, Birmingham B15 2TH, UK. Tel: + 44 121 472 1311. Fax: + 44 121 627 5886. E-mail: [email protected]
This paper is based on an oral presentation given at the International Hepato-Pancreato-Biliary Association Meeting, 27 February to 2 March 2008, Mumbai.
HPB 2010, 12, 389–395
Introduction In recent years, endoscopic ultrasound (EUS) and EUS-guided fine-needle aspiration (EUS-FNA) have gained acceptance in the
© 2010 International Hepato-Pancreato-Biliary Association
HPB
390
diagnosis of pancreatic and biliary pathology. Endoscopic ultrasound allows for the visualization of the pancreatic head, neck and uncinate process from the duodenum, and the neck, body and tail from the stomach. High-frequency ultrasound and the close proximity of the transducer produce high-resolution images of the pancreas and peripancreatic structures and can detect small lesions that are discriminated with difficulty by computed tomography (CT). Although several series have quoted high rates of accuracy for EUS-FNA in the diagnosis of pancreatic lesions,1–10 the literature on EUS-FNA in biliary and ampullary lesions is limited.11–17 Our study aims to assess the diagnostic accuracy of EUS-FNA in pancreatic, periampullary and biliary lesions in a single, high-volume centre in the absence of an onsite cytopathologist. The paper also analyses the false negative cases in detail to uncover learning points. We hope these results will contribute to the limited, collective global experience on the effectiveness of EUS-FNA in periampullary and biliary lesions, as data on
large numbers of these relatively rare lesions are difficult to amass in individual centres performing EUS-FNA.
Materials and methods All patients who underwent EUS-FNA performed for pancreatic, biliary and periampullary lesions in a single hepato-biliary referral centre, over a 1-year period between March 2006 and March 2007, were identified and reviewed (Figs 1 and 2). Our patient dataset represented a preselected cohort of patients who presented with atypical histories or symptoms and equivocal CT findings (e.g. no double duct sign or mass lesion) or who required a definitive diagnosis although deemed unresectable or unfit for surgery. Those patients with typical symptoms (e.g. painless obstructive jaundice) and a resectable pancreatic mass on CT underwent surgery without EUS. Prospectively collected data from the EUS database and hospital records were used for analysis. All cystic
Total annual UGI EUS-FNA, n = 429
Pancreatic and peripancreatic lesions, n = 196
Cystic pancreatic lesions, n = 88
Non-pancreatic lesions, n = 233
Excluded
Study cohort, n = 108
Solid pancreatic lesions, n = 78; biliary tract lesions, n = 21; periampullary lesions, n = 9
Excluded
Figure 1 Workload in our department over the study year and the selection of the present study population. UGI, upper gastrointestinal; EUS-FNA, endoscopic ultrasound-guided fine-needle aspiration
Solid lesions, n = 108
Biliary lesions, n = 21
Head of pancreas, n = 50 (64%)
Pancreatic lesions, n = 78
Neck of pancreas, n = 13 (17%)
Ampullary lesions, n = 9
Body of pancreas, n = 10 (13%)
Tail of pancreas, n = 5 (6%)
Figure 2 Distribution of solid pancreatic lesions
HPB 2010, 12, 389–395
© 2010 International Hepato-Pancreato-Biliary Association
HPB
391
(A)
(B)
Figure 3 Cytology showing adenocarcinoma cells: (A) pap stain; (B) clot preparation
pancreatic lesions (cystic tumours and pseudocysts) were excluded as the diagnostic work-up for these lesions differed from that for solid tumours. Endoscopic ultrasound was performed under conscious sedation (using midazolam and fentanyl) by a single experienced operator using a curved linear array echo endoscope (Olympus GF-UCT240-AL5; Olympus Corp., Tokyo, Japan), which allows the visualization of the needle tract in real time. The instrument has colour Doppler imaging built in, which helps with the identification and avoidance of vascular structures and, in some situations, aids diagnosis by quantifying the vascularity of the lesion concerned. When the transducer had been positioned, the needle catheter system (EchoTip, 22-gauge needle; Wilson-Cook Medical, Inc., Winston-Salem, NC, USA) was inserted through the working channel of the echo endoscope and the needle advanced into the target lesion under real-time ultrasound guidance. A negative suction of 0–40 ml was applied to the FNA needle using a syringe and the needle was then carefully moved back and forth within the lesion. The aspirate was then expressed onto glass slides and both air-dried and alcohol-fixed smears were prepared by the endoscopist. A further aspirate was usually expressed into 15 ml of CytoRich Red fixative. No cytopathologist was available in the endoscopy suite to examine the adequacy of the sample. One to four needle passes were performed as necessary to achieve an adequate sample as determined by the echo endoscopist. All smears were subsequently examined by two experienced, dedicated cytopathologists with a special interest in pancreatic FNAs. Figures 3 and 4 show example cytology. The results of the FNAs were compared with the final diagnoses as determined by specimen histology in patients undergoing resection, and by serial imaging and clinical follow-up in nonsurgical patients. Lesions were considered to be benign if there was no evidence of progression or resolution on serial imaging at 12 months. Any lesion which progressed on serial imaging or which resulted in clinical deterioration or death was considered malignant. In the current study, results were deemed to represent: a true positive (TP) if both EUS-FNA and histology, serial imaging or follow-up suggested malignancy; a true negative (TN) if both EUS-FNA and histology or serial imaging suggested absence of
HPB 2010, 12, 389–395
malignancy; a false positive (FP) if EUS-FNA showed tumour and the resection specimen found no malignancy or no increase in mass on imaging at 12 months, and a false negative (FN) if EUSFNA found no malignancy, but specimen histology, serial imaging or follow-up confirmed the presence of malignancy. It should be noted that the lesions were categorized into the three groups defined above (pancreatic, biliary and periampullary lesions) according to EUS findings rather than cross-sectional imaging. Although lesions within the bile duct or ampulla were relatively easy to categorize, intrapancreatic lesions were more difficult. Here, the criterion of ‘mass centring on the bile duct or ampulla’ (i.e. the site of uniform circumferential growth) was used to assess the likely origin of the tumour. Therefore, it is possible that some crossover may have occurred because some bile duct cancers can be eccentrically centred around the bile duct.
Results During the study period, 429 EUS-FNAs were performed for various upper gastrointestinal lesions, of which 196 were for pancreatic and peripancreatic pathologies. Eighty-eight of these were cystic lesions and were therefore excluded from the study, leaving 108 EUS-FNAs [78 (72%) were solid pancreatic, 21 (19%) were bile duct and nine (8%) were periampullary lesions] in the final study cohort (Figs 1 and 2). All but two of the 108 EUS were completed as day patient procedures. Overall, 31 patients (29%) had negative cytology (including ‘atypical’ results) on FNA; nine (29%) of these patients underwent surgery and their diagnoses were confirmed by histology, whereas the remaining 22 (71%) underwent serial imaging rather than surgery. Solid pancreatic lesions In the 78 patients with solid pancreatic lesions, 50 (64%) of the lesions were in the head, 13 (17%) in the body, 10 (13%) in the neck and five (6%) in the tail of the pancreas (Fig. 2). The median size of the lesions was 30 mm (range 20–80 mm). The median number of passes during biopsy was two (range 1–4). Among the pancreatic lesions, 65 (83%) EUS-FNAs yielded TP results after comparison with the final diagnosis. The other 13 pancreatic
© 2010 International Hepato-Pancreato-Biliary Association
HPB
392
(A)
(B)
(D)
(C)
Figure 4 Clusters of cells with round nuclei and moderate cytoplasm confirming neuroendocrine neoplasm: (A) pap stain; (B) clot preparation; (C) positive synaptophysin; (D) positive chromogranin
Table 1 Pre- and post-test probabilities of endoscopic ultrasound-guided fine-needle aspiration in the three subgroups. Pre-test probabili-
ties are based on those reported by Heiken21 Pre-test probability
EUS-FNA performance
Disease prevalence in the sample (95% CI)
Sensitivity (95% CI)
Specificity (95% CI)
PPV (95% CI)
Post-test probabilities NPV (95% CI)
Overall performance Accuracy (95% CI)
Pancreatic lesions
86% (76–93%)
97% (90–100%)
100% (72–100%)
100% (95–100%)
85% (55–98%)
97% (91–99%)
Periampullary lesions
33% (8–70%)
67% (9–99%)
100% (54–100%)
100% (16–100%)
86% (42–100%)
89% (67–100%)
Bile duct lesions
52% (30–74%)
91% (59–100%)
100% (69–100%)
100% (69–100%)
91% (59–100%)
95% (81–100%)
Pancreatic and periampullary lesions
80% (71–88%)
96% (88–99%)
100% (81–100%)
100% (95–100%)
85% (62–97%)
97% (92–99%)
EUS-FNA, endoscopic ultrasound-guided fine-needle aspiration; PPV, positive predictive value; NPV, negative predictive value; 95% CI, 95% confidence interval
lesions included 11 TN, two FN and no FP results. The other patients were followed up by imaging for ⱖ12 months. One of the two FN patients underwent a Whipple’s pancreaticoduodenectomy and the other a jejunal resection. Their resection specimen histology confirmed both the lesions to be gastrointestinal stromal tumours (GISTs). Two FNA samples showed atypical cells on cytology, which were later confirmed to represent chronic pancreatitis and pancreatic abscesses. The sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV) and accuracy of EUS-FNA for pancreatic lesions are shown in Table 1 and Fig. 5.
HPB 2010, 12, 389–395
Periampullary lesions Nine EUS-FNAs were performed for periampullary lesions. These yielded two TP, six TN, no FP and one FN result for a lesion which was later proven to be an adenocarcinoma following a Whipple’s procedure. The sensitivity, specificity, PPV, NPV and accuracy of EUS-FNA for periampullary lesions alone and combined with pancreatic lesions are shown in Table 1. Common bile duct lesions A total of 21 patients underwent EUS-FNA for bile duct lesions. Of these, 17 patients had undergone an endoscopic retrograde cho-
© 2010 International Hepato-Pancreato-Biliary Association
HPB
393
100% 80% FalseP
60%
FalseN TrueN
40%
TrueP
20%
pu lla ry Pe ria m
tra ct lia ry Bi
Pa nc re at ic
0%
Figure 5 Results for each of the three study cohorts according to lesion location. Results for two patients in the bile duct cohort were indeterminate and are not included. FalseP, false positive; FalseN, false negative; TrueN, true negative; TrueP, true positive
langiopancreatography (ERCP) prior to EUS. The EUS-FNA results were TP in 10 patients, TN in eight and FN in one. Despite being categorized as FN, a repeat EUS-FNA confirmed the presence of a cholangiocarcinoma in this particular patient. In two patients, the EUS-FNA samples were indeterminate. Both of these were later confirmed to be benign on repeat EUS-FNA performed at 6 and 12 weeks followed by serial imaging. Of the eight patients with TN results, three underwent surgery and diagnosis was confirmed by histology. Diagnoses in the remaining patients were again confirmed by repeat EUS-FNA at 6 and 12 weeks, and imaging follow-up for 12 months. The sensitivity, specificity, PPV, NPV and accuracy of EUS-FNA for biliary lesions are shown in Table 1. In the current series one patient suffered post-procedural haemorrhage. This patient developed a peri-portal haematoma and required surgical exploration. Another patient needed an overnight stay in hospital for pain. No patients re-presented after discharge from the day-case unit with pain or post-procedural pancreatitis and neither was there any procedure-related mortality. The final histology included, in addition to the adenocarcinomas, GISTs (n = 3), adenosquamous tumour (n = 1), neuroendocrine tumours (n = 10) and tuberculosis (n = 1).
Discussion Endoscopic ultrasound-guided FNA has emerged as a major diagnostic adjuvant in the management of pancreatic and biliary lesions. Because of their anatomical location, relatively small pancreatic and bile duct lesions are often not amenable to percutaneous guided biopsy, although this limitation depends on the experience of the individual operator and centre. Endoscopic ultrasound-guided FNA is able to obtain tissue diagnosis in such lesions in a relatively simple and safe manner, almost obviating the
HPB 2010, 12, 389–395
need for surgery for tissue diagnosis. In addition, EUS-FNA significantly decreases the likelihood of tumour seeding via the needle track as the tissue planes traversed are usually part of the potential surgical field. Endoscopic ultrasound-guided FNA of pancreatic lesions was first reported by Vilmann et al. in 1992,18 since when several large series have reported sensitivities of EUS-FNA ranging from 74% to 96% and specificity of close to 100%. Rocca et al. reported a series of 246 patients in which the sensitivity and specificity of EUS-FNA were 79% and 60%, respectively.1 In a series of 300 EUS-FNAs, Varadarajulu et al. noted that sensitivity of EUS-FNA was significantly lower (74% vs. 91%) in patients with chronic pancreatitis.2 The diagnostic accuracy and sensitivity of EUS-FNA for solid pancreatic lesions in the current series are comparable with the best rates reported in the literature. In both of the two patients in whom FN results were obtained, the endoscopist noted concern about either the sample or the lesion at the time of the procedure. In the first patient, EUS showed a highly vascular mass in the head of pancreas and FNA obtained a bloody aspirate which did not show any malignant cells. In view of the significant clinical suspicion, this patient went on to have a pancreaticoduodenectomy and the final histology confirmed a GIST. Endoscopic ultrasound in the second patient showed a dilated pancreatic duct up to the level of the pancreatic head (bile duct thickening was noted despite the absence of a stricture) and an ill-defined hypoechoic lesion in the head of the pancreas, which was aspirated. Imaging follow-up with serial CT demonstrated a pancreatic head mass of increasing size and eventual liver metastases consistent with malignancy. Co-morbidity had prevented initial surgery. Despite the lack of malignant cells on cytology, endoscopic and clinical features prompted further investigation. Although in theory it is possible that an onsite cytopathologist might have enforced further attempts to obtain representative cells, in practice the lack of such a cytopathologist did not change our course of management as no better aspirates were achieved in the first patient despite multiple attempts and it was felt unnecessary to proceed with further attempts in the second because the patient was unfit for intervention. Endoscopic ultrasound-guided FNA of biliary tract lesions has been reported by many series to be useful, especially in determining the presence and extent of the often small lesions and the involvement of regional lymph nodes.11–17 However, it is generally acknowledged that EUS of the proximal bile duct is technically more demanding as a result of the spiral relationship of the duct to the duodenum as it travels past the duodenum. DeWitt et al. studied 24 patients with a proximal biliary stricture and negative ERCP brush biopsy.11 Endoscopic ultrasound enabled the visualization of a mass in 23 (96%) patients and EUS-FNA had a sensitivity of 77%.11 A review by Meara et al. of 46 patients with suspected biliary tree malignancy found EUS-FNA to have sensitivity of 87% and specificity of 100%.12 Sensitivity reported in other studies varies between 43% and 89% and most of these studies were conducted in circumstances that included an onsite
© 2010 International Hepato-Pancreato-Biliary Association
HPB
394
cytopathologist. In the current series, we found one FN result in a patient in whom the US appearance was indeed suggestive of a cholangiocarcinoma, but whose aspirate did not show malignant cells, even when the procedure was repeated a month later. The final diagnosis (cholangiocarcinoma) was confirmed by histology following surgery. The two cases in which cytology was indeterminate were later confirmed to be benign with repeat EUS-FNA at 6 and 12 weeks and serial imaging. Despite the results detailed above, it is important to highlight the nine patients in the present study who underwent surgery despite negative cytology and in whom diagnoses were confirmed by histology at surgery. The procedures carried out in these patients included gastric bypass (n = 2) and Roux-en-Y biliary bypass (n = 7). All of these patients had chronic pancreatitis or acute recurrent pancreatitis. In our experience, although chronic pancreatitis is not diagnosed on cytopathological grounds, cytopathology still has a very high PPV in detecting malignancy. One of the difficulties for the EUS operator in this subgroup of patients concerns knowing where to target sampling in the context of an abnormal gland. The current study did not, however, specifically look at this subset of pancreatic pathology on this occasion. The literature on the use of EUS-FNA for ampullary tumours is limited. In a series of 35 patients, Defrain et al. reported the sensitivity and specificity of EUS-FNA for ampullary tumours to be 82% and 100%, respectively.19 In the current series, the one patient in the periampullary lesion subgroup in whom an FN result was obtained went on to have a resection and was found to have a small focus of adenocarcinoma within the lesion, which would have been unlikely to be sampled even with an onsite cytopathologist. By comparison, the diagnostic accuracy and sensitivity of EUS-FNA for pancreatic lesions were greater than for biliary lesions, which is in line with the published literature. Endoscopic ultrasound-guided FNA is highly operatordependent.20 In the current series, a single endoscopist performed all the procedures, which may have resulted in greater consistency and contributed to the better diagnostic accuracy and sensitivity achieved. No cytopathologist was present during the procedure to evaluate the adequacy of the samples. Despite this, diagnostic rates in the current series are comparable with the best rates reported in the literature and most reported studies utilized an onsite cytopathologist. The PPV for all three subgroups was 100%, which, because it is greater than the actual prevalence in each subgroup, suggests that EUS-FNA increases confidence in the confirmation of the diagnosis. By comparison, CT has some limitations in detecting these lesions, with published PPVs of 76–90%.21 In two instances, bloody or inadequate samples were obtained, both of which were recognized by the endoscopist at the time of biopsy. In all instances of false interpretations, the endoscopist expressed concerns about the sample at the time of the procedure. It is to be noted that some of the credit for the results of the present study reflect the experience of dedicated hepato-biliary pathologists, who were able to make definite diagnoses on the basis of limited samples.
HPB 2010, 12, 389–395
Conclusions The diagnostic accuracy of EUS-FNA for pancreatic lesions in the current series was 97% and the PPVs for the three subgroups were 100%. These rates are comparable with the best rates reported in the literature, despite the absence of onsite cytopathology. We believe that these results are a direct result of high-volume practice and the substantial experience of dedicated endosonographers and cytopathologists. These results indicate that it is possible to achieve high rates of accuracy, even when logistical issues make it impossible to maintain a cytopathologist in the endoscopy suite. Conflicts of interest None declared. References 1. Rocca R, De Angelis C, Daperno M, Carucci P, Ravarino N, Bruno M et al. (2007) Endoscopic ultrasound-fine needle aspiration (EUS-FNA) for pancreatic lesions: effectiveness in clinical practice. Dig Liver Dis 39:768– 774. 2. Varadarajulu S, Tamhane A, Eloubeidi MA. (2005) Yield of EUS-guided FNA of pancreatic masses in the presence or the absence of chronic pancreatitis. Gastrointest Endosc 62:728–736; quiz 751, 753. 3. Iglesias-Garcia J, Dominguez-Munoz E, Lozano-Leon A, Abdulkader I, Larino-Noia J, Antunez J et al. (2007) Impact of endoscopic ultrasoundguided fine needle biopsy for diagnosis of pancreatic masses. World J Gastroenterol 13:289–293. 4. Raut CP, Grau AM, Staerkel GA, Kaw M, Tamm EP, Wolff RA et al. (2003) Diagnostic accuracy of endoscopic ultrasound-guided fine-needle aspiration in patients with presumed pancreatic cancer. J Gastrointest Surg 7:118–126; discussion 127–128. 5. Ylagan LR, Edmundowicz S, Kasal K, Walsh D, Lu DW. (2002) Endoscopic ultrasound guided fine-needle aspiration cytology of pancreatic carcinoma: a 3-year experience and review of the literature. Cancer 96:362–369. 6. Hernandez LV, Mishra G, Forsmark C, Draganov PV, Petersen JM, Hochwald SN et al. (2002) Role of endoscopic ultrasound (EUS) and EUSguided fine needle aspiration in the diagnosis and treatment of cystic lesions of the pancreas. Pancreas 25:222–228. 7. Chhieng DC, Jhala D, Jhala N, Eltoum I, Chen VK, Vickers S et al. (2002) Endoscopic ultrasound-guided fine-needle aspiration biopsy: a study of 103 cases. Cancer 96:232–239. 8. O'Toole D, Palazzo L, Arotçarena R, Dancour A, Aubert A, Hammel P et al. (2001) Assessment of complications of EUS-guided fine-needle aspiration. Gastrointest Endosc 53:470–474. 9. Chang KJ, Nguyen P, Erickson RA, Durbin TE, Katz KD. (1997) The clinical utility of endoscopic ultrasound-guided fine-needle aspiration in the diagnosis and staging of pancreatic carcinoma. Gastrointest Endosc 45:387–393. 10. Wiersema MJ, Vilmann P, Giovannini M, Chang KJ, Wiersema LM. (1997) Endosonography-guided fine-needle aspiration biopsy: diagnostic accuracy and complication assessment. Gastroenterology 112:1087–1095. 11. DeWitt J, Misra VL, Leblanc JK, McHenry L, Sherman S. (2006) EUSguided FNA of proximal biliary strictures after negative ERCP brush cytology results. Gastrointest Endosc 64:325–333. 12. Meara RS, Jhala D, Eloubeidi MA, Eltoum I, Chhieng DC, Crowe DR et al.
© 2010 International Hepato-Pancreato-Biliary Association
HPB
395
(2006) Endoscopic ultrasound-guided FNA biopsy of bile duct and gall-
Thonke F et al. (2004) EUS-guided fine-needle aspiration of suspected
bladder: analysis of 53 cases. Cytopathology 17:42–49.
hilar cholangiocarcinoma in potentially operable patients with negative
13. Rösch T, Hofrichter K, Frimberger E, Meining A, Born P, Weigert N et al. (2004) ERCP or EUS for tissue diagnosis of biliary strictures? A prospective comparative study. Gastrointest Endosc 60:390–396. 14. Byrne MF, Gerke H, Mitchell RM, Stiffler HL, McGrath K, Branch MS et al. (2004) Yield of endoscopic ultrasound-guided fine-needle aspiration of bile duct lesions. Endoscopy 36:715–719. 15. Lee JH, Salem R, Aslanian H, Chacho M, Topazian M. (2004) Endoscopic ultrasound and fine-needle aspiration of unexplained bile duct strictures. Am J Gastroenterol 99:1069–1073.
brush cytology. Am J Gastroenterol 99:45–51. 18. Vilmann P, Jacobsen GK, Henriksen FW, Hancke S. (1992) Endoscopic ultrasonography with guided fine needle aspiration biopsy in pancreatic disease. Gastrointest Endosc 38:172–173. 19. Defrain C, Chang CY, Srikureja W, Nguyen PT, Gu M. (2005) Cytologic features and diagnostic pitfalls of primary ampullary tumours by endoscopic
ultrasound-guided
fine-needle
aspiration
biopsy.
Cancer
105:289–297. 20. Savides TJ, Donohue M, Hunt G, Al-Haddad M, Aslanian H, Ben-
16. Eloubeidi MA, Chen VK, Jhala NC, Eltoum IE, Jhala D, Chhieng DC et al.
Menachem T et al. (2007) EUS-guided FNA diagnostic yield of malig-
(2004) Endoscopic ultrasound-guided fine needle aspiration biopsy of
nancy in solid pancreatic masses: a benchmark for quality performance
suspected cholangiocarcinoma. Clin Gastroenterol Hepatol 2:209– 213. 17. Fritscher-Ravens A, Broering DC, Knoefel WT, Rogiers X, Swain P,
HPB 2010, 12, 389–395
measurement. Gastrointest Endosc 66:277–282. 21. Heiken JP. (2001) Carcinoma of the pancreas: detection and staging using CT and MRI. Cancer Imaging 2:19–22.
© 2010 International Hepato-Pancreato-Biliary Association