0016-5107/92/3805-0531$03.00 GASTROINTESTINAL ENDOSCOPY Copyright © 1992 by the American Society for Gastrointestinal Endoscopy
Endoscopic needle aspiration biopsy at ERCP in the diagnosis of biliary strictures D. A. Howell, MD, R. P. Beveridge, MD J. Bosco, MD, M. Jones, MD Portland, Maine
We have developed a ball-tipped catheter with a retractable 22-gauge, 7-mm long needle to perform endoscopic needle aspiration (ENA) for cytology and compared this technique to brush cytology of malignant-appearing biliary strictures during ERCP. Of 31 patients, 26 had proven malignant strictures involving the common bile duct and 5 had benign lesions. All 31 patients had ENA and 29 were brushed. Positive ENAs were obtained in 16 of 26 patients (61.5%) and positive brushings in 2 of 24 (8.3%). With the addition of two suspicious ENAs for pancreatic adenocarcinoma, 73% of patients had positive or suspicious cytology for malignancy by combined ENA and brush with a specificity of 100%. Although ENA appeared to be more sensitive in diagnosing cholangiocarcinoma, it proved to be most effective in the diagnosis of pancreatic adenocarcinoma when compared with brush cytology. One patient with cholangiocarcinoma in our series was diagnosed by brush cytology only, with a negative ENA, supporting our recommendation of using both endoscopic brushings and ENA for cytology when evaluating biliary strictures. (Gastrointest Endosc 1992;38:531-535)
Differentiating malignant biliary strictures from benign ductal lesions at ERCP can be a diagnostic challenge. Brush and forceps biopsies at endoscopy and percutaneously have shown some recent success, but have lower sensitivities with sclerotic and, particularly, extrinsic tumors such as pancreatic adenocarcinoma. 1- 7 Modern imaging techniques can be suggestive of malignancy, but in many cases these lesions are small or unsuccessfully visualized, making percutaneous aspiration difficult. Since a precise cytologic diagnosis has direct bearing on subsequent treatment options, a more sensitive modality is needed to obtain tissue from this anatomically difficult area. We have developed a new metal ball-tipped catheter containing a fully retractable aspiration needle which can be accurately positioned below suspicious appearing biliary lesions during ERCP. Endoscopic needle aspiration (ENA) for cytology can then be easily and safely Received November 13, 1991. For revision March 23, 1992. Accepted June 8, 1992. From the Department of Medicine, Division of Gastroenterology and Department of Pathology, Maine Medical Center, Portland, Maine. Reprint requests: Douglas A. Howell, MD, 131 Chadwick Street, Portland, Maine 04102. Presented in part at the annual meeting of the American Society for Gastrointestinal Endoscopy, New Orleans, Louisiana, May 21, 1991. VOLUME 38, NO.5, 1992
obtained by advancing the needle deeply into the lesion and applying suction. In this study we compare this new technique with that of standard brush cytology in the evaluation of malignant-appearing biliary strictures at ERCP. PATIENTS AND METHODS
Between October 1988 and April 1991, 31 consecutive patients presented with signs and symptoms of biliary obstruction who did not have, or had unsuccessful, CT-guided fine needle aspiration cytology for suspected malignancy. This group consisted of 17 men and 14 women ranging in age from 52 to 88 years with a mean age of 68 years. Twentyseven patients initially presented with obstructive jaundice while four patients had right upper quadrant abdominal pain. No mass seen on CT was greater than 5 em, and 12 of 26 (42.6%) had tumors equal to or smaller than 3 em including 6 patients with subsequently biopsy-proven cancers and no visible mass at CT. After overnight fast and appropriate premedication, each patient underwent ERCP with a side-viewing duodenoscope (TJF-lO; Olympus Corporation of America, Lake Success, N. Y.). Informed consent was obtained prior to each evaluation. Diagnostic studies were done in the usual manner and malignant-appearing strictures were present in all patients. After sphincterotomy for eventual large bore endoprosthesis, 31 patients (100%) underwent endoscopic needle aspiration 531
(ENA) and 29 (93.5%) were brushed with a Geenen guidewire tip cytology brush (Wilson-Cook Medical, Inc., Winston-Salem, N. C.) during a single endoscopic procedure. All procedures were performed by one author (D. A. H.). The ENA technique employs a 7 F pre-bent ball-tipped catheter with a retractable 22-gauge, 7-mm long needle (HBAN 22; Wilson-Cook Medical, Inc.) (Fig. 1) which is inserted through the biopsy port of the endoscope. Care is utilized in order to minimize straightening of the catheter and a large biopsy channel endoscope is preferred. Catheter placement is facilitated by prior sphincterotomy done for large bore endoprothesis placement. Once the common bile duct is cannulated, the tip is advanced into appropriate position directly below the stricture under fluoroscopic guidance. The needle is then extended into the tumor and aspirations are obtained using a 20-ml dry syringe. At times, hard tumors can be better penetrated by stiffening the catheter with the insertion of a stainless steel stilette wire. Positioning of the needle for biopsy in a patient with pancreatic carcinoma is shown in Figure 2. In this study endoscopic needle aspirations were obtained by gently moving the needle back and forth within the tumor and by aspirating in several orientations relative to the stricture without withdrawing the needle from the bile duct. We generally chose three sites for biopsy, releasing the
Figure 2. Malignant biliary stricture due to pancreatic adenocarcinoma with the needle positioned (left) and extended (right).
suction between each placement of the needle. After all aspirations were performed, the needle was retracted into the ball tip and the catheter was withdrawn from the endoscope. The contents of the needle were then expressed, using an air-filled syringe, onto slides which were immediately smeared and fixed to prevent air drying artifact. The standard Papanicolaou staining technique was then employed in all cases (Fig. 3). Clearance of the cytology material from the needle was facilitated by running the stainless steel stilette back down the catheter if necessary. Flushing the needle with saline will prevent cell adhesion to the slide and was not utilized. In cases where an excessively bloody or clotted specimen was expressed, cell block technique was occasionally used. The Geenen guidewire tip cytology brush (Wilson-Cook Medical, Inc.) was then placed through the stricture and several brushings were obtained in the usual fashion.!' 4, 7 In two cases, the stricture was not brushed due to difficult guide wire placement. Finally, an endoprosthesis for the relief of obstructive symptoms was placed directly over the guide catheter containing the Geenen brush without removing it. All smears were classified as positive, suspicious, or negative by a single cytologist (M. J.) using standard cytology criteria for pancreatico-biliary lesions. Of note, two patients with initially negative ENA and brushings underwent a second ENA and brush, and one patient underwent a third attempt before a definitive diagnosis was made.
RESULTS
Figure 1. HBAN-22
extended position. 532
biliary aspiration needle in withdrawn and
Of the 31 consecutive patients with malignant-appearing biliary strictures at ERCP, carcinoma was confirmed in 26 and benign disease was found in 5. As outlined in Table 1, there were 19 pancreatic adenocarcinomas, 5 cholangiocarcinomas, 1 gallbladder carcinoma, and 1 metastatic colon carcinoma. The diagnoses were verified at surgery, CT-guided biopsy, auGASTROINTESTINAL ENDOSCOPY
topsy, or by subsequent hospital course. The five benign lesions in our study were ultimately proven at surgery to be four cases of pancreatitis and one Mirizzi syndrome. In the malignant group, 2 of 24 patients successfully brushed had a positive brush cytology (8.3%). Both had cholangiocarcinoma. Sixteen of 26 patients (61.5%) had positive ENA (10 pancreatic adenocarcinomas, 4 cholangiocarcinomas, 1 gallbladder carci-
noma, and 1 metastatic carcinoma). Highly suspicious ENAs were present in 2 of 26 patients (7.7%), both with pancreatic adenocarcinomas. Overall, 18 of 26 patients (69.2%) had positive or suspicious ENA for cytology. Further breakdown of our results reveal that three of five cholangiocarcinomas (60%) were diagnosed with the first endoscopic needle aspiration and 80% were found after two attempts. Brushings for cholangiocarcinoma were positive at initial ERCP in two of four patients brushed, one of whom had had a negative ENA. By combining ENA and brush cytology, 100% of the cholangiocarcinomas were diagnosed in our study population. Both the gallbladder carcinoma and the metastatic colon carcinoma were diagnosed on initial ENA. Brushings of these lesions were negative. In patients with pancreatic adenocarcinoma, 8 of 19 (47%) had positive ENAs on the first attempt. With the addition of two suspicious ENAs the yield for pancreatic adenocarcinoma was 52.6% at initial ERCP. One patient underwent a second attempt while another had two additional attempts before positive ENAs were obtained. Overall, by combining positive and suspicious cytologies, 12 of 19 patients (63.1 %) with pancreatic adenocarcinomas were diagnosed by endoscopic needle aspiration. Conversely, endoscopic brushings for cytology failed to diagnose a single patient with pancreatic adenocarcinoma in our series. Of the remaining seven patients with pancreatic adenocarcinomas who had false negative ENAs, six did not undergo a repeat ERCP for a second attempt at needle aspiration since the diagnosis was made at surgery, at CT biopsy, or by other means. For all types of malignancies in our study group, 19 of 26 patients (73%) had positive or suspicious cytology by combined ENA and brush. There were no false positives or complications attributed to sphincterotomy or biopsy in the study group.
•
DISCUSSION
Malignant-appearing bile duct strictures are commonly encountered at ERCP in patients presenting
Figure 3. Malignant cells aspirated from an adenocarcinoma of the pancreas.
Table 1. Results in patients with proven malignancy Tumor type
No. of patients
Pancreatic adenocarcinoma Cholangiocarcinoma Gallbladder carcinoma Metastatic colon cancer
19 5 1 1
loa 4 1 1
26
16 (61.5)
Total
+ ENA (%) (52.6) (80) (100) (100)
+ Brushings (%)
+ Combined (%)
Ob (0) 2' (50) o (0) o (0)
10 (52.6) 5 (100) 1 (100) 1 (100)
2 (8.3)
17 (65.4)
Does not include two patients with highly suspicious ENAs for pancreatic adenocarcinoma. One pancreatic adenocarcinoma did not undergo brush cytology. , One cholangiocarcinoma did not undergo brush cytology. a
b
VOLUME 38, NO.5, 1992
533
with signs and symptoms of biliary obstruction or pancreatitis. Benign etiologies may be indistinguishable radiologically from malignancy, emphasizing the importance of histologic diagnosis. CT or ultrasound-guided aspiration biopsy using 22gauge needles are commonly employed but require a second or repeat procedure if not done prior to endoscopy, need a sizable target, and, although safe, are frequently uncomfortable. Furthermore, their true yield in unselected cases may be lower than the frequently quoted 80 to 90%.7-10 A recent study reported only 57% positive results in pancreatic adenocarcinomas. 9 Open surgical needle biopsy in this series was a surprisingly low 77%, emphasizing the difficulty in producing adequate sampling in this tumor type. The development of an endoscopic biopsy technique, which could be immediately applied during the initial endoscopic procedure, and which would be technically simple, safe, inexpensive, and frequently successful has been a goal of many investigators. Early reports included pancreatic juice collection after secretin injection 11, 12 and endoscopic bile duct catheter aspiration for cytology, 13 but these techniques suffered from low yields and prolonged procedure times. Forceps biopsy and brush cytology at ERCP have been more recently reported techniques with specificity approaching nearly 100% but variable sensitivity ranging from 8 to 100% generally in fairly small series. 1- s In an effort to increase yield, Ryan 1 reported brushing from the pancreatic duct but sensitivity was a disappointing 30%, below that of common duct brushings, and two patients developed pancreatitis possibly related to the procedure. Cytologic preparations from the proximal flap of removed endoprostheses was recently reported to yield 78.6% positive results but only two pancreatic adenocarcinomas were in the study group.6 This technique may be helpful in patients failing initial tissue sampling, but a delay in cytologic diagnosis of 51.5 days
in the study limits its usefulness in treatment planning. Percutaneous catheters have allowed repeated brushing of cholangiocarcinoma with an initial yield of 40%, rising to 62% overall;7 however, few patients today are managed in this fashion. A summary of the reported results of endoscopic tissue sampling techniques is presented in Table 2. Obviously, sensitivity has been particularly low in pancreatic carcinoma likely due to its extrinsic nature which compresses the duct epithelium but does not destroy it, explaining the strictured but patent ductal lumen. Our complete absence of positive brushings from these patients is well below other reports but may be in part explained by pre-selection which favored smaller earlier tumors not seen or successfully biopsied at CT scanning. Sampling deep to the epithelium by needle aspiration might logically be expected to increase the yield of positive cytologies in this setting. Since the aspirations are done from the lower aspect of the tumor and not from within the stricture itself, we do not feel the technique adversely affects the yield of subsequent brush cytology by producing blood contamination or artifact. Our low brush yield suggests further studies of the brush sensitivity in early stage carcinomas may be needed. Endoscopic fine needle aspiration transmurally has been successful in the diagnosis of esophagogastric 14 and pancreatic malignancies. IS To be successful, pancreatico-biliary tumors had to involve the duodenum or retrogastric region and be identified endoscopically. Placement of a retractable needle for aspiration biopsy into the common bile duct has not been feasible until the unique physical aspects of retrograde cannulation were incorporated into the design. The specifically engineered pre-bend with a metal ball tip enabled successful cannulation in all study patients immediately following small sphincterotomy for large bore endoprosthesis placement. No complications of sphincterotomy were noted. The ball tip permitted
Table 2. Sensitivity of endoscopic techniques in tissue sampling Sensitivity for (%) Series" Ryan! Scudera et al! Foutch et aU Rustgi et al. 4 Aabakken et al. 5 Leung et al. B
Technique utilized
Pancreatic adenocarcinoma
Cholangiocarcinoma
Malignant ductal strictures
Brush Brush Brush Brush Forceps Forceps Scraping biopsy
6/20 (30) 5/10 (50) 0/1 (0) NAb NA 1/7 (14.3) Not specified
4/9 (44.4) 2/2 (100) 4/4 (100) 2/4 (50) 4/4 (100) Not specified Not specified
20/45 7/14 4/7 2/5 4/5 3/10 11/14
Total
12/38 (31.6)
16/23 (69.6)
51/100 (51)
(44.4) (50) (57) (40) (80) (30) (78.6)
" All studies report total specificity of 100%. b NA, not applicable. 534
GASTROINTESTINAL ENDOSCOPY
sliding of the tip over the tissues into position as the endoscope was advanced in more difficult cases. ENA after needle positioning is not technically difficult, adds minimal time to the ERCP procedure, and has caused no morbidity in our hands. A drawback is the requirement of the sphincterotomy, since cannulation through an intact sphincter would be most difficult, although we have had success in a few situations where sphincterotomy was not employed. The yield of ENA would be expected to rise with repeated attempts. In this study, we did not request a cytologist to be present at the procedure, but this is a consideration, since immediate cytologic analysis could guide the need for repeat aspirations if initial samples are inconclusive. ENA is a promising new technique for collection of cytologic material at therapeutic ERCP. Combining it with brush cytology yielded 19 of 26 (73%) positive or suspicious cytology overall in the study. More striking, however, was the contribution of ENA (69.2%) as compared with brush (8.3%) in diagnosing malignant biliary strictures. ACKNOWLEDGMENTS
The authors are grateful for the assistance of Pat Studley, Rochelle Petersen, and Maureen Puckett.
REFERENCES 1. Ryan M. Cytologic brushings of ductal lesions during ERCP. Gastrointest Endosc 1991;37:139-42.
VOLUME 38, NO.5, 1992
2. Scudera PL, Koizumi J, Jacobson 1M. Brush cytology evaluation of lesions encountered during ERCP. Gastrointest Endosc 1990;36:281-4. 3. Foutch PG, Harlan JR, Kerr D, Sanowski RA. Wire-guided brush cytology: a new endoscopic method for diagnosis of bile duct cancer. Gastrointest Endosc 1989;35:243-7. 4. Rustgi AK, Kelsey PB, Guelrud M, Saini S, Schapiro RH. Malignant tumors of the bile ducts: diagnosis by biopsy during endoscopic cannulation. Gastrointest Endosc 1989;35:248-51. 5. Aabakken L, Karesen R, Serck-Hanssen A, Osnes M. Transpapillary biopsies and brush cytology from the common bile duct. Endoscopy 1986;18:49-51. 6. Leung JWC, Sung JY, Chung SCS, Chan KM. Endoscopic scraping biopsy of malignant biliary strictures. Gastrointest Endosc 1989;35:65-6. 7. Rabinovitz M, Zajko AB, Hassanein T, et al. Diagnostic value of brush cytology in the diagnosis of bile duct carcinoma: a study in 65 patients with bile duct strictures. Hepatology 1990;12:747-52. 8. Athlin L, Blind BJ, Angstrom T. Fine needle aspiration biopsy of pancreatic masses. Acta Chir Scand 1990;156:91-4. 9. Parsons L, Palmer C. How accurate is fine-needle biopsy in malignant neoplasm ofthe pancreas? Arch Surg 1989;124:6813.
lD. Ekberg 0, Bergenfeldt M, Aspelin P, et al. Reliability of ultrasound guided fine needle biopsy of pancreatic masses. Acta Radiol 1988;29:535-9. 11. Endo Y, Moric T, Tanura H, Okuda S. Cytodiagnosis of pancreatic malignant tumors by aspiration under direct vision, using a duodenal fiberscope. Gastroenterology 1974;67:944-51. 12. Hatfield ARW, Whittaker R, Gibbs DD. The collection of pancreatic fluid for cytodiagnosis using a duodenoscope. Gut 1974;15:305-7. 13. Roberts-Thomson IC, Hobbs JB. Cytodiagnosis of pancreatic and biliary cancer by endoscopic duct aspiration. Med J Aust 1979;1:370-2. 14. Kochhar R, Rajwanshi A, Malik AK, Gupta SK, Mehta SK. Endoscopic fine needle aspiration biopsy of gastroesophageal malignancies. Gastrointest Endosc 1988;34:321-3. 15. Tsuch R, Henmi T, Kondo N, Akashi M, Harada N. Endoscopic aspiration biopsy of the pancreas. Gastroenterology 1977;73:1050-2.
535
Endoscopic needle aspiration biopsy at ERCP in the diagnosis of biliary strictures D. A. Howell, MD, R. P. Beveridge, MD J. Bosco, MD, M. Jones, MD Portland, Maine
We have developed a ball-tipped catheter with a retractable 22-gauge, 7-mm long needle to perform endoscopic needle aspiration (ENA) for cytology and compared this technique to brush cytology of malignant-appearing biliary strictures during ERCP. Of 31 patients, 26 had proven malignant strictures involving the common bile duct and 5 had benign lesions. All 31 patients had ENA and 29 were brushed. Positive ENAs were obtained in 16 of 26 patients (61.5%) and positive brushings in 2 of 24 (8.3%). With the addition of two suspicious ENAs for pancreatic adenocarcinoma, 73% of patients had positive or suspicious cytology for malignancy by combined ENA and brush with a specificity of 100%. Although ENA appeared to be more sensitive in diagnosing cholangiocarcinoma, it proved to be most effective in the diagnosis of pancreatic adenocarcinoma when compared with brush cytology. One patient with cholangiocarcinoma in our series was diagnosed by brush cytology only, with a negative ENA, supporting our recommendation of using both endoscopic brushings and ENA for cytology when evaluating biliary strictures. (Gastrointest Endosc 1992;38:531-535)
Differentiating malignant biliary strictures from benign ductal lesions at ERCP can be a diagnostic challenge. Brush and forceps biopsies at endoscopy and percutaneously have shown some recent success, but have lower sensitivities with sclerotic and, particularly, extrinsic tumors such as pancreatic adenocarcinoma. 1- 7 Modern imaging techniques can be suggestive of malignancy, but in many cases these lesions are small or unsuccessfully visualized, making percutaneous aspiration difficult. Since a precise cytologic diagnosis has direct bearing on subsequent treatment options, a more sensitive modality is needed to obtain tissue from this anatomically difficult area. We have developed a new metal ball-tipped catheter containing a fully retractable aspiration needle which can be accurately positioned below suspicious appearing biliary lesions during ERCP. Endoscopic needle aspiration (ENA) for cytology can then be easily and safely Received November 13, 1991. For revision March 23, 1992. Accepted June 8, 1992. From the Department of Medicine, Division of Gastroenterology and Department of Pathology, Maine Medical Center, Portland, Maine. Reprint requests: Douglas A. Howell, MD, 131 Chadwick Street, Portland, Maine 04102. Presented in part at the annual meeting of the American Society for Gastrointestinal Endoscopy, New Orleans, Louisiana, May 21, 1991. VOLUME 38, NO.5, 1992
obtained by advancing the needle deeply into the lesion and applying suction. In this study we compare this new technique with that of standard brush cytology in the evaluation of malignant-appearing biliary strictures at ERCP. PATIENTS AND METHODS
Between October 1988 and April 1991, 31 consecutive patients presented with signs and symptoms of biliary obstruction who did not have, or had unsuccessful, CT-guided fine needle aspiration cytology for suspected malignancy. This group consisted of 17 men and 14 women ranging in age from 52 to 88 years with a mean age of 68 years. Twentyseven patients initially presented with obstructive jaundice while four patients had right upper quadrant abdominal pain. No mass seen on CT was greater than 5 em, and 12 of 26 (42.6%) had tumors equal to or smaller than 3 em including 6 patients with subsequently biopsy-proven cancers and no visible mass at CT. After overnight fast and appropriate premedication, each patient underwent ERCP with a side-viewing duodenoscope (TJF-lO; Olympus Corporation of America, Lake Success, N. Y.). Informed consent was obtained prior to each evaluation. Diagnostic studies were done in the usual manner and malignant-appearing strictures were present in all patients. After sphincterotomy for eventual large bore endoprosthesis, 31 patients (100%) underwent endoscopic needle aspiration 531
(ENA) and 29 (93.5%) were brushed with a Geenen guidewire tip cytology brush (Wilson-Cook Medical, Inc., Winston-Salem, N. C.) during a single endoscopic procedure. All procedures were performed by one author (D. A. H.). The ENA technique employs a 7 F pre-bent ball-tipped catheter with a retractable 22-gauge, 7-mm long needle (HBAN 22; Wilson-Cook Medical, Inc.) (Fig. 1) which is inserted through the biopsy port of the endoscope. Care is utilized in order to minimize straightening of the catheter and a large biopsy channel endoscope is preferred. Catheter placement is facilitated by prior sphincterotomy done for large bore endoprothesis placement. Once the common bile duct is cannulated, the tip is advanced into appropriate position directly below the stricture under fluoroscopic guidance. The needle is then extended into the tumor and aspirations are obtained using a 20-ml dry syringe. At times, hard tumors can be better penetrated by stiffening the catheter with the insertion of a stainless steel stilette wire. Positioning of the needle for biopsy in a patient with pancreatic carcinoma is shown in Figure 2. In this study endoscopic needle aspirations were obtained by gently moving the needle back and forth within the tumor and by aspirating in several orientations relative to the stricture without withdrawing the needle from the bile duct. We generally chose three sites for biopsy, releasing the
Figure 2. Malignant biliary stricture due to pancreatic adenocarcinoma with the needle positioned (left) and extended (right).
suction between each placement of the needle. After all aspirations were performed, the needle was retracted into the ball tip and the catheter was withdrawn from the endoscope. The contents of the needle were then expressed, using an air-filled syringe, onto slides which were immediately smeared and fixed to prevent air drying artifact. The standard Papanicolaou staining technique was then employed in all cases (Fig. 3). Clearance of the cytology material from the needle was facilitated by running the stainless steel stilette back down the catheter if necessary. Flushing the needle with saline will prevent cell adhesion to the slide and was not utilized. In cases where an excessively bloody or clotted specimen was expressed, cell block technique was occasionally used. The Geenen guidewire tip cytology brush (Wilson-Cook Medical, Inc.) was then placed through the stricture and several brushings were obtained in the usual fashion.!' 4, 7 In two cases, the stricture was not brushed due to difficult guide wire placement. Finally, an endoprosthesis for the relief of obstructive symptoms was placed directly over the guide catheter containing the Geenen brush without removing it. All smears were classified as positive, suspicious, or negative by a single cytologist (M. J.) using standard cytology criteria for pancreatico-biliary lesions. Of note, two patients with initially negative ENA and brushings underwent a second ENA and brush, and one patient underwent a third attempt before a definitive diagnosis was made.
RESULTS
Figure 1. HBAN-22
extended position. 532
biliary aspiration needle in withdrawn and
Of the 31 consecutive patients with malignant-appearing biliary strictures at ERCP, carcinoma was confirmed in 26 and benign disease was found in 5. As outlined in Table 1, there were 19 pancreatic adenocarcinomas, 5 cholangiocarcinomas, 1 gallbladder carcinoma, and 1 metastatic colon carcinoma. The diagnoses were verified at surgery, CT-guided biopsy, auGASTROINTESTINAL ENDOSCOPY
topsy, or by subsequent hospital course. The five benign lesions in our study were ultimately proven at surgery to be four cases of pancreatitis and one Mirizzi syndrome. In the malignant group, 2 of 24 patients successfully brushed had a positive brush cytology (8.3%). Both had cholangiocarcinoma. Sixteen of 26 patients (61.5%) had positive ENA (10 pancreatic adenocarcinomas, 4 cholangiocarcinomas, 1 gallbladder carci-
noma, and 1 metastatic carcinoma). Highly suspicious ENAs were present in 2 of 26 patients (7.7%), both with pancreatic adenocarcinomas. Overall, 18 of 26 patients (69.2%) had positive or suspicious ENA for cytology. Further breakdown of our results reveal that three of five cholangiocarcinomas (60%) were diagnosed with the first endoscopic needle aspiration and 80% were found after two attempts. Brushings for cholangiocarcinoma were positive at initial ERCP in two of four patients brushed, one of whom had had a negative ENA. By combining ENA and brush cytology, 100% of the cholangiocarcinomas were diagnosed in our study population. Both the gallbladder carcinoma and the metastatic colon carcinoma were diagnosed on initial ENA. Brushings of these lesions were negative. In patients with pancreatic adenocarcinoma, 8 of 19 (47%) had positive ENAs on the first attempt. With the addition of two suspicious ENAs the yield for pancreatic adenocarcinoma was 52.6% at initial ERCP. One patient underwent a second attempt while another had two additional attempts before positive ENAs were obtained. Overall, by combining positive and suspicious cytologies, 12 of 19 patients (63.1 %) with pancreatic adenocarcinomas were diagnosed by endoscopic needle aspiration. Conversely, endoscopic brushings for cytology failed to diagnose a single patient with pancreatic adenocarcinoma in our series. Of the remaining seven patients with pancreatic adenocarcinomas who had false negative ENAs, six did not undergo a repeat ERCP for a second attempt at needle aspiration since the diagnosis was made at surgery, at CT biopsy, or by other means. For all types of malignancies in our study group, 19 of 26 patients (73%) had positive or suspicious cytology by combined ENA and brush. There were no false positives or complications attributed to sphincterotomy or biopsy in the study group.
•
DISCUSSION
Malignant-appearing bile duct strictures are commonly encountered at ERCP in patients presenting
Figure 3. Malignant cells aspirated from an adenocarcinoma of the pancreas.
Table 1. Results in patients with proven malignancy Tumor type
No. of patients
Pancreatic adenocarcinoma Cholangiocarcinoma Gallbladder carcinoma Metastatic colon cancer
19 5 1 1
loa 4 1 1
26
16 (61.5)
Total
+ ENA (%) (52.6) (80) (100) (100)
+ Brushings (%)
+ Combined (%)
Ob (0) 2' (50) o (0) o (0)
10 (52.6) 5 (100) 1 (100) 1 (100)
2 (8.3)
17 (65.4)
Does not include two patients with highly suspicious ENAs for pancreatic adenocarcinoma. One pancreatic adenocarcinoma did not undergo brush cytology. , One cholangiocarcinoma did not undergo brush cytology. a
b
VOLUME 38, NO.5, 1992
533
with signs and symptoms of biliary obstruction or pancreatitis. Benign etiologies may be indistinguishable radiologically from malignancy, emphasizing the importance of histologic diagnosis. CT or ultrasound-guided aspiration biopsy using 22gauge needles are commonly employed but require a second or repeat procedure if not done prior to endoscopy, need a sizable target, and, although safe, are frequently uncomfortable. Furthermore, their true yield in unselected cases may be lower than the frequently quoted 80 to 90%.7-10 A recent study reported only 57% positive results in pancreatic adenocarcinomas. 9 Open surgical needle biopsy in this series was a surprisingly low 77%, emphasizing the difficulty in producing adequate sampling in this tumor type. The development of an endoscopic biopsy technique, which could be immediately applied during the initial endoscopic procedure, and which would be technically simple, safe, inexpensive, and frequently successful has been a goal of many investigators. Early reports included pancreatic juice collection after secretin injection 11, 12 and endoscopic bile duct catheter aspiration for cytology, 13 but these techniques suffered from low yields and prolonged procedure times. Forceps biopsy and brush cytology at ERCP have been more recently reported techniques with specificity approaching nearly 100% but variable sensitivity ranging from 8 to 100% generally in fairly small series. 1- s In an effort to increase yield, Ryan 1 reported brushing from the pancreatic duct but sensitivity was a disappointing 30%, below that of common duct brushings, and two patients developed pancreatitis possibly related to the procedure. Cytologic preparations from the proximal flap of removed endoprostheses was recently reported to yield 78.6% positive results but only two pancreatic adenocarcinomas were in the study group.6 This technique may be helpful in patients failing initial tissue sampling, but a delay in cytologic diagnosis of 51.5 days
in the study limits its usefulness in treatment planning. Percutaneous catheters have allowed repeated brushing of cholangiocarcinoma with an initial yield of 40%, rising to 62% overall;7 however, few patients today are managed in this fashion. A summary of the reported results of endoscopic tissue sampling techniques is presented in Table 2. Obviously, sensitivity has been particularly low in pancreatic carcinoma likely due to its extrinsic nature which compresses the duct epithelium but does not destroy it, explaining the strictured but patent ductal lumen. Our complete absence of positive brushings from these patients is well below other reports but may be in part explained by pre-selection which favored smaller earlier tumors not seen or successfully biopsied at CT scanning. Sampling deep to the epithelium by needle aspiration might logically be expected to increase the yield of positive cytologies in this setting. Since the aspirations are done from the lower aspect of the tumor and not from within the stricture itself, we do not feel the technique adversely affects the yield of subsequent brush cytology by producing blood contamination or artifact. Our low brush yield suggests further studies of the brush sensitivity in early stage carcinomas may be needed. Endoscopic fine needle aspiration transmurally has been successful in the diagnosis of esophagogastric 14 and pancreatic malignancies. IS To be successful, pancreatico-biliary tumors had to involve the duodenum or retrogastric region and be identified endoscopically. Placement of a retractable needle for aspiration biopsy into the common bile duct has not been feasible until the unique physical aspects of retrograde cannulation were incorporated into the design. The specifically engineered pre-bend with a metal ball tip enabled successful cannulation in all study patients immediately following small sphincterotomy for large bore endoprosthesis placement. No complications of sphincterotomy were noted. The ball tip permitted
Table 2. Sensitivity of endoscopic techniques in tissue sampling Sensitivity for (%) Series" Ryan! Scudera et al! Foutch et aU Rustgi et al. 4 Aabakken et al. 5 Leung et al. B
Technique utilized
Pancreatic adenocarcinoma
Cholangiocarcinoma
Malignant ductal strictures
Brush Brush Brush Brush Forceps Forceps Scraping biopsy
6/20 (30) 5/10 (50) 0/1 (0) NAb NA 1/7 (14.3) Not specified
4/9 (44.4) 2/2 (100) 4/4 (100) 2/4 (50) 4/4 (100) Not specified Not specified
20/45 7/14 4/7 2/5 4/5 3/10 11/14
Total
12/38 (31.6)
16/23 (69.6)
51/100 (51)
(44.4) (50) (57) (40) (80) (30) (78.6)
" All studies report total specificity of 100%. b NA, not applicable. 534
GASTROINTESTINAL ENDOSCOPY
sliding of the tip over the tissues into position as the endoscope was advanced in more difficult cases. ENA after needle positioning is not technically difficult, adds minimal time to the ERCP procedure, and has caused no morbidity in our hands. A drawback is the requirement of the sphincterotomy, since cannulation through an intact sphincter would be most difficult, although we have had success in a few situations where sphincterotomy was not employed. The yield of ENA would be expected to rise with repeated attempts. In this study, we did not request a cytologist to be present at the procedure, but this is a consideration, since immediate cytologic analysis could guide the need for repeat aspirations if initial samples are inconclusive. ENA is a promising new technique for collection of cytologic material at therapeutic ERCP. Combining it with brush cytology yielded 19 of 26 (73%) positive or suspicious cytology overall in the study. More striking, however, was the contribution of ENA (69.2%) as compared with brush (8.3%) in diagnosing malignant biliary strictures. ACKNOWLEDGMENTS
The authors are grateful for the assistance of Pat Studley, Rochelle Petersen, and Maureen Puckett.
REFERENCES 1. Ryan M. Cytologic brushings of ductal lesions during ERCP. Gastrointest Endosc 1991;37:139-42.
VOLUME 38, NO.5, 1992
2. Scudera PL, Koizumi J, Jacobson 1M. Brush cytology evaluation of lesions encountered during ERCP. Gastrointest Endosc 1990;36:281-4. 3. Foutch PG, Harlan JR, Kerr D, Sanowski RA. Wire-guided brush cytology: a new endoscopic method for diagnosis of bile duct cancer. Gastrointest Endosc 1989;35:243-7. 4. Rustgi AK, Kelsey PB, Guelrud M, Saini S, Schapiro RH. Malignant tumors of the bile ducts: diagnosis by biopsy during endoscopic cannulation. Gastrointest Endosc 1989;35:248-51. 5. Aabakken L, Karesen R, Serck-Hanssen A, Osnes M. Transpapillary biopsies and brush cytology from the common bile duct. Endoscopy 1986;18:49-51. 6. Leung JWC, Sung JY, Chung SCS, Chan KM. Endoscopic scraping biopsy of malignant biliary strictures. Gastrointest Endosc 1989;35:65-6. 7. Rabinovitz M, Zajko AB, Hassanein T, et al. Diagnostic value of brush cytology in the diagnosis of bile duct carcinoma: a study in 65 patients with bile duct strictures. Hepatology 1990;12:747-52. 8. Athlin L, Blind BJ, Angstrom T. Fine needle aspiration biopsy of pancreatic masses. Acta Chir Scand 1990;156:91-4. 9. Parsons L, Palmer C. How accurate is fine-needle biopsy in malignant neoplasm ofthe pancreas? Arch Surg 1989;124:6813.
lD. Ekberg 0, Bergenfeldt M, Aspelin P, et al. Reliability of ultrasound guided fine needle biopsy of pancreatic masses. Acta Radiol 1988;29:535-9. 11. Endo Y, Moric T, Tanura H, Okuda S. Cytodiagnosis of pancreatic malignant tumors by aspiration under direct vision, using a duodenal fiberscope. Gastroenterology 1974;67:944-51. 12. Hatfield ARW, Whittaker R, Gibbs DD. The collection of pancreatic fluid for cytodiagnosis using a duodenoscope. Gut 1974;15:305-7. 13. Roberts-Thomson IC, Hobbs JB. Cytodiagnosis of pancreatic and biliary cancer by endoscopic duct aspiration. Med J Aust 1979;1:370-2. 14. Kochhar R, Rajwanshi A, Malik AK, Gupta SK, Mehta SK. Endoscopic fine needle aspiration biopsy of gastroesophageal malignancies. Gastrointest Endosc 1988;34:321-3. 15. Tsuch R, Henmi T, Kondo N, Akashi M, Harada N. Endoscopic aspiration biopsy of the pancreas. Gastroenterology 1977;73:1050-2.
535
