Fine-Needle Aspiration Biopsy of Solid Pancreatic Masses: Comparison of CT and Endoscopic Sonography Guidance

Comparison of Guidance Techniques for Pancreatic Mass Biopsy Abdominal Imaging Original Research A C D E M N E U T R Y L I A M C A I G O F I N G Sukru Mehmet Erturk 1 Koenraad J. Mortelé 1 Kemal Tuncali 1 John R. Saltzman 2 Roger Lao 1 Stuart G. Silverman 1 Erturk SM, Mortelé KJ, Tuncali K, Saltzman JR, Lao R, Silverman SG Keywords: biopsy, CT, endoscopic sonography, fine-needle aspiration biopsy, imaging-guided biopsy, oncologic imaging, pancreas, pancreatic cancer DOI:10.2214/AJR.05.1657 Received September 16, 2005; accepted after revision October 25, 2005. 1 Division of Abdominal Imaging and Intervention, Department of Radiology, Brigham and Women s Hospital, Harvard Medical School, 75 Francis St., Boston, MA 02115. Address correspondence to S. M. Erturk. 2 Division of Gastroenterology, Department of Medicine, Brigham and Women s Hospital, Harvard Medical School, Boston, MA. AJR 2006; 187:1531 1535 0361 803X/06/1876 1531 American Roentgen Ray Society Fine-Needle Aspiration Biopsy of Solid Pancreatic Masses: Comparison of CT and Endoscopic Sonography Guidance OBJECTIVE. Imaging-guided biopsies of solid pancreatic masses are performed with either CT or endoscopic sonography at our institution. We compared test characteristics of fineneedle aspiration biopsies guided using CT with those guided using endoscopic sonography and secondarily evaluated for an effect of mass size. MATERIALS AND METHODS. Of 70 solid pancreatic masses, 43 (mean size, 4.4 cm; range, 1.5 10.3 cm) underwent fine-needle (20- to 22-gauge) aspiration biopsy with CT guidance and 27 (mean size, 2.3 cm; range, 1.0 5.0 cm) underwent fine-needle (22-gauge) aspiration biopsy with endoscopic sonography guidance. The diagnostic rate, sensitivity, and negative predictive value (NPV) for each technique were compared using Fisher s exact test before and after stratifying masses by size as small ( 3 cm) or large (> 3 cm). RESULTS. The overall diagnostic rate, sensitivity, and NPV of fine-needle aspiration biopsies guided using CT (97.7%, 94.9%, and 60%, respectively) were not significantly different from those guided using endoscopic sonography (88.9%, 85%, and 57.1%, respectively). Among small masses, the diagnostic rate and sensitivity for biopsies guided using CT (100% and 100%, respectively) were not significantly different from those for biopsies guided using endoscopic sonography (90.9% and 93.8%, respectively). Among large masses, the diagnostic rate and sensitivity (96.6% and 92.3%, respectively) for biopsies guided using CT were not significantly different from those for biopsies guided using endoscopic sonography (83.3% and 50%, respectively). CONCLUSION. When biopsying solid pancreatic masses with fine needles, procedures guided with CT and those guided with endoscopic sonography have similar test characteristics regardless of mass size. ancreatic cancer is the fourth leading cause of cancer-related deaths P in the United States [1, 2]. Approximately 28,000 new cases of pancreatic cancer are diagnosed each year [1, 2]. Fine-needle aspiration biopsy is performed typically to diagnose the cause of a pancreatic mass identified on CT, MRI, or sonography [3]. Biopsy is useful both to characterize the mass as benign or malignant and, if malignant, to diagnose the type of malignancy [4]. For example, among malignant pancreatic tumors, including adenocarcinomas, islet cell carcinomas, and lymphomas, each carries a different prognosis and is treated differently. CT-guided fine-needle aspiration biopsy is an established technique for the diagnosis of pancreatic masses [5 7]. Endoscopic sonography also has been used to guide pancreatic mass biopsy [1, 2, 8 10]. The accuracy rates for fine-needle aspiration biopsy with both guidance techniques are similar, with sensitivities up to 100% for CT guidance and 96% for endoscopic sonography guidance [1, 2, 11 13]. However, endoscopic sonography guidance has been postulated to be more effective than CT guidance, especially for the diagnosis of small pancreatic masses [2, 8, 11]. Nevertheless, to our knowledge, this claim has been based on anecdotal descriptions and has not been evaluated [3, 13]. The purpose of our study was to compare test characteristics of pancreatic fine-needle aspiration biopsies guided using CT with those guided using endoscopic sonography and secondarily to evaluate for an effect of mass size. Materials and Methods Subjects Institutional review board approval was obtained to review the medical records and images of all patients who underwent pancreatic mass aspira- AJR:187, December 2006 1531

tion biopsy during a 4-year period. A total of 57 masses in 57 patients were biopsied under CT guidance. Of these, 49 masses were solid. All were biopsied using fine needles (20- to 22-gauge). Four masses were also biopsied with 18-gauge needles; these were excluded because we consider fine needles to include needles that are 19-gauge or thinner. Two patients were excluded because of inadequate follow-up. The remaining 43 masses in 43 patients were included in the study. The patients ranged in age from 19 to 89 years (mean age, 65.1 years). Twenty-four (55.8%) of them were women, and 19 (44.2%) were men. Among 50 masses in 50 patients that underwent endoscopic biopsy during the same 4-year period, 30 were solid. All were biopsied using 22-gauge needles. Three masses were excluded because of inadequate follow-up. The remaining 27 masses were included in the study. The patients ranged in age from 34 to 88 years (mean age, 61.2 years). Sixteen (59.3%) of them were women, and 11 (40.7%) were men. Pancreatic Masses The maximum in-plane diameters of the pancreatic masses biopsied using CT guidance were measured on contrast-enhanced CT images obtained before or after the biopsy procedure (mean time difference between CT and biopsy, 8.6 days). The mean diameter of the pancreatic masses biopsied using CT guidance was 4.4 cm (range, 1.5 10.3 cm): 14 masses (32.6%) were equal to or smaller than 3 cm in maximum dimension (mean, 2.4 cm; range, 1.5 3 cm), and 29 masses (67.4%) were larger than 3 cm (mean, 5.4 cm; range, 3.1 10.3 cm). Twenty-four masses (55.8%) were located in the head of the pancreas, 11 (25.6%) in the body, and eight (18.6%) in the tail. The maximum diameters of pancreatic masses biopsied under endoscopic sonographic guidance were measured on static sonographic images obtained during the biopsy procedure. The mean diameter of pancreatic masses that were biopsied using endoscopic sonography was 2.3 cm (range, 1.0 5.0 cm). Twenty-two (81.5%) of the masses were equal to or smaller than 3 cm in maximum dimension (mean, 2.0 cm; range, 1.00 3.00 cm). Five (18.5%) of the masses were larger than 3 cm (mean, 3.9 cm; range, 3.3 5.0 cm). Seventeen masses (63.0%) were located in the head of the pancreas, 10 (37.0%) in the body, and none in the tail. Biopsy Procedures CT-guided fine-needle aspiration biopsy procedures were performed using a CT scanner capable of CT fluoroscopy (Somatom Plus 4/Carevision, Siemens Medical Solutions). CT fluoroscopic guidance was used for biopsy of 40 masses (93%), and CT guidance without CT fluoroscopy was used for biopsy of the remaining three masses (7%). In 31 biopsies (72.1%), only 20-gauge needles were used, and in 12 biopsies (27.9%) only 22-gauge needles were used. Biopsy needles were placed using the tandem technique: An initial needle was placed and was then followed by the placement of additional needles in tandem alongside the initial needle. A cytotechnologist was present during all the procedures to evaluate the initial specimens for adequacy. The endoscopic sonographically guided fineneedle aspiration biopsies were performed by gastroenterologists using a sonography scanner (ProSound SSD-5000 HD, Aloka) with curvilinear array echoendoscopes (GF-UC140P-AL5 and GF- UCT140-AL5, Olympus) integrated into the unit. After endosonographic localization, the mass was punctured with 22-gauge needles using either a transgastric approach if the mass was located in the body or tail of the pancreas or a transduodenal approach if it was located in the head of the pancreas. A cytotechnologist was not routinely present during these procedures. Cytopathologic Examination and Final Diagnosis The cytology reports were reviewed, and the fine-needle aspiration biopsy results were categorized as positive or negative, both of which were considered diagnostic, or as nondiagnostic. Results were categorized as positive if the cytopathology report described the specimen as containing malignant cells or cells suspicious for malignancy. All the biopsy results that were derived from specimens containing malignant cells were considered truepositive. Positive results that were derived from specimens containing suspicious cells were considered true-positive if a subsequent surgical biopsy confirmed that diagnosis. Cases for which the cytopathology report described specimens as containing no malignant cells or cells consistent with a benign process were categorized as negative. Patients with biopsy results that were categorized as negative were followed. Negative results were considered true-negative if the results were confirmed on a subsequent surgical biopsy or if the mass showed regression or lack of growth at imaging follow-up (minimum of 12 months). Nondiagnostic results included biopsy reports that described the specimens as inadequate. Statistical Analysis Diagnostic rate (proportion of procedures that resulted in biopsy reports that were categorized as positive or negative) was calculated with 95% CIs. Sensitivity and NPV, along with 95% CIs, were calculated for only the biopsies with diagnostic specimens. The proportions of small and large masses in the CT guidance and endoscopic sonography guidance groups were calculated. The diagnostic rate, sensitivity, and NPV with 95% CIs were calculated for small and large masses in the two study groups. All intergroup differences were investigated using Fisher s exact test. The differences were considered significant if the p value was less than 0.05. Results All but one (97.7%) of 43 biopsies in the CT-guided fine-needle aspiration biopsy group were diagnostic (Table 1). The biopsy of a 4.5- cm infiltrative pancreatic mass was nondiagnostic. Subsequent fine-needle aspiration biopsy of an enlarged peripancreatic lymph node revealed lymphoma, and a presumptive diagnosis of lymphomatous involvement of the pancreas was made after the mass regressed after the patient underwent treatment with chemotherapy. Of the 42 masses with diagnostic results, 37 (88.1%) were malignant and five (11.9%) were benign. Among the malignant cases, 33 (89.2%) were diagnosed as adenocarcinoma, one as anaplastic giant cell tumor, and one as melanoma metastasis. In the remaining two masses, the diagnosis was malignant islet cell tumor of the pancreas. In three of the cases of adenocarcinoma, cells suspicious for adenocarcinoma were reported, and the diagnosis was confirmed with surgical biopsy. Of the five masses with benign results, one was diagnosed as a well-differentiated islet cell tumor; this mass was classified as benign, was left untreated, and remained stable on CT for 22 months after the biopsy (Table 2). Another mass was diagnosed as extramedullary hematopoiesis of the pancreas. This diagnosis was corroborated on a subsequent bone marrow biopsy that revealed a chronic myeloproliferative disorder. The mass had not changed 12 months after the initial biopsy. In the remaining three cases that had benign results, benign-appearing pancreatic ductal and acinar cells were reported. One proved to be chronic pancreatitis with CT follow-up; this result was classified as true-negative. The remaining two proved to be pancreatic ductal adenocarcinoma at surgical resection and were classified as false-negative. There were a total of 37 true-positive, three true-negative, and two false-negative results. The sensitivity and NPV of CT-guided fineneedle aspiration biopsy were 94.9% and 60.0%, respectively. All but three of the 27 biopsies performed under endoscopic sonographic guidance were diagnostic (88.9%) (Table 1). One of the three masses that yielded nondiagnostic 1532 AJR:187, December 2006

Comparison of Guidance Techniques for Pancreatic Mass Biopsy TABLE 1: Summary of Results of Fine-Needle Aspiration Biopsy of Pancreatic Masses Under CT Guidance Versus Endoscopic Sonography Guidance Cytopathology Final Diagnosis (No. of Masses) Test Characteristic (%) Guidance Technique Result Malignant Benign Diagnostic Rate Sensitivity Negative Predictive Value CT 97.7 (43/44; 87.9 99.6) 94.9 (37/39; 83.1 98.6) 60 (3/5; 23.1 88.2) Malignant 37 0 Benign 2 3 Endoscopic sonography 88.9 (24/27; 71.9 96.2) 85 (17/20; 63.9 94.8) 57.1 (4/7; 25 84.1) Malignant 17 0 Benign 3 4 Nondiagnostic 2 1 Note Numbers in the parentheses are proportions and 95% CIs. TABLE 2: Pancreatic Masses with Negative Results at Fine-Needle Aspiration Biopsy Guidance Technique Cytopathology Result Final Diagnosis Method Used to Establish Final Diagnosis CT Benign acinar cells Adenocarcinoma Surgical biopsy Benign reactive ductal cells Adenocarcinoma Surgical biopsy Well-differentiated islet cell tumor Islet cell tumor No growth at 12-month CT follow-up Extramedullary hematopoiesis Extramedullary hematopoiesis No growth at 12-month CT follow-up Benign acinar cells Chronic pancreatitis No growth at 12-month CT follow-up Endoscopic sonography Benign reactive glandular cells Adenocarcinoma CT-guided fine-needle aspiration biopsy Benign reactive glandular cells Adenocarcinoma Cytologic analysis of pleural fluid Benign epithelial cells Adenocarcinoma CT-guided fine-needle aspiration biopsy of metastatic liver lesion Benign epithelial cells Peripancreatic fat necrosis Exploratory laparotomy Benign-appearing ductal cells Chronic pancreatitis Surgical biopsy Benign glandular cells Chronic pancreatitis Surgical biopsy Benign ductal cells Pancreatitis No growth at 12-month CT follow-up biopsy results was diagnosed as benign manifestation of chronic pancreatitis because the mass was unchanged on CT follow-up 30 months after the initial biopsy. The other two masses were diagnosed as pancreatic adenocarcinoma at surgical resection in one and at CT-guided fine-needle aspiration biopsy in the other. Of the 24 masses with diagnostic results, 17 (70.8%) were pancreatic adenocarcinoma and seven (29.2%) were benign. In five of the cases of adenocarcinoma, cells suspicious for adenocarcinoma were reported, and the diagnosis was confirmed with surgical biopsy. Of the seven masses that yielded benign biopsy results, three were proven with surgical biopsy and one with CT follow-up (Table 2). Two of the remaining three masses proved to be pancreatic ductal adenocarcinoma: One was proven with a CT-guided biopsy of the same pancreatic mass and the other with a CT-guided biopsy of a metastatic liver lesion. In the third patient, analysis of pleural fluid showed malignant cells consistent with adenocarcinoma. These three cases were classified as false-negative. There were a total of 17 true-positive, four true-negative, and three false-negative results. The sensitivity and NPV of endoscopic biopsy were 85% (17/20) and 57.1% (4/7), respectively. There were no complications in either group. There were no significant differences in diagnostic rate, sensitivity, and NPV between the guidance techniques. Mass Size Fourteen (32.6%) of the 43 masses biopsied under CT guidance were small and 29 (67.4%) were large. Twenty-two (81.5%) of the 27 masses biopsied under endoscopic sonography guidance were small and five (18.5%) were large (p < 0.01). Biopsy procedures guided using CT had higher diagnostic rates and sensitivities among both small and large masses than those guided using endoscopic sonography (Table 3); however, the differences did not reach statistical significance. For small masses biopsied under CT guidance, the NPV was not calculated because there were no false-negative results. The NPV was not calculated for large masses biopsied under endoscopic sonography guidance because there were no true-negative results. AJR:187, December 2006 1533

TABLE 3: Results of Pancreatic Mass Biopsies Stratified by Mass Size and Guidance Technique Discussion Percutaneous CT-guided fine-needle aspiration biopsy is an established means of diagnosing the cause of pancreatic masses [2, 9, 12]. The procedure may be performed on an outpatient basis, and complication rates are low, ranging from 3% to 6.7% [5, 14]. Recently, an accuracy of 81% was reported for CT-guided fine-needle aspiration biopsy of pancreatic masses [3]. Endoscopic sonographically guided fine-needle aspiration biopsy is an alternative technique for biopsying pancreatic masses [3]. Endoscopic sonography was developed in the 1990s to overcome limitations of transabdominal sonography of the pancreas caused by intervening gas and fat [1]. The positioning of a highfrequency ultrasound transducer in direct proximity to the pancreas by way of the stomach or duodenum produces high-resolution images of the pancreas [2, 9]. Endoscopic sonographically guided fine-needle aspiration biopsy of the pancreas has been performed over the past 10 years [15], and recently an accuracy rate of 76% was reported [3]. When an appropriate indication for pancreatic mass biopsy is identified, the question often arises as to which guidance method, CT or endoscopic sonography, is appropriate to use to biopsy the mass. We sought to examine test characteristics using each guidance technique Final Diagnosis (No. of Masses) Test Characteristic (%) Mass Size (Diameter) Guidance Technique Cytopathology Malignant Benign Diagnostic Rate Sensitivity Negative Predictive Value Small ( 3 cm) CT 100 (14/14; 78.5 100) 100 (13/13; 77.2 100) NC Malignant 13 0 Benign 0 1 Nondiagnostic 0 0 Endoscopic sonography 90.9 (20/22; 72.2 97.4) 93.8 (15/16; 71.7 98.9) 80 (4/5; 37.6 96.4) Malignant 15 0 Benign 1 4 Nondiagnostic 1 1 Large (> 3 cm) CT 96.6 (28/29; 82.8 99.4) 92.3 (24/26; 75.9 97.9) 50 (2/4; 15 85) Malignant 24 0 Benign 2 2 Endoscopic sonography 83.3 (4/5; 37.6 96.4) 50 (2/4; 15 85) NC Malignant 2 0 Benign 2 0 Note Numbers in parentheses are proportions and 95% CIs, respectively. NC = not calculated. and the effect of mass size to help determine which technique to use in clinical practice. We found that CT-guided and endoscopic biopsy had diagnostic rates of 97.7% and 88.9%, respectively. Our results were comparable with those of previous studies. Tillou et al. [5] reported a diagnostic rate of 96.5% for diagnosing pancreatic masses via CT and transabdominal sonographically guided fineneedle aspiration biopsy. Recently, Shin et al. [8] reported a diagnostic rate of 86.8% for endoscopic sonographically guided fine-needle aspiration biopsy of pancreatic masses. To the best of our knowledge, there are only two published studies that directly compare CT-guided and endoscopic sonographically guided fine-needle aspiration biopsies of pancreatic masses [3, 16]. In a retrospective study, Qian and Hecht [13] suggested that CT-guided biopsies may be more sensitive for diagnosing malignancy than endoscopic sonographically guided biopsies. In their study, CT-guided biopsies and endoscopic sonographically guided biopsies had sensitivities of 71% and 42%, respectively; the authors did not report whether the difference was statistically significant [13]. Likewise, in the study of Mallery et al. [3], CT- and transabdominal sonographically guided pancreatic biopsies (80%) had a higher sensitivity than endoscopic sonographically guided biopsies (74%); however, the difference was not statistically significant. We also found higher sensitivity for CT guidance (94.9%) compared with endoscopic sonographic guidance (85%). However, the difference did not reach statistical significance in our study either. In the study of Qian and Hecht [13], the NPVs of CT- and endoscopic sonographically guided fine-needle aspiration biopsies were similar: 41% and 45%, respectively. Mallery et al. [3] reported NPVs of 23% and 27% for fine-needle aspiration biopsies performed under CT and endoscopic sonographic guidance, respectively. We found NPVs of 60% and 57.1% for CT-guided and endoscopic biopsy, respectively. The sensitivity and NPV of both techniques in our study were higher than those in the studies of Qian and Hecht [13] and Mallery et al. [3]. The reason for these discrepancies in results might be that cystic pancreatic masses were included in those studies, whereas we focused on solid pancreatic masses and excluded cystic ones from our analysis. The diagnostic workup of most cystic pancreatic masses involves analysis of cystic fluid for biochemical and tumor markers rather than cytology [17 19]; thus, the 1534 AJR:187, December 2006

Comparison of Guidance Techniques for Pancreatic Mass Biopsy accuracy of fine-needle aspiration biopsy is related to both the ability to position a needle in a mass and the accuracy of the biochemical analysis of the cystic fluid [20]. The criteria for establishing a malignant diagnosis based on cystic fluid analysis and its precise role have not yet been completely established [21]. Both Qian and Hecht [13] and Mallery et al. [3] reported that endoscopic sonography was used more frequently than CT to biopsy small pancreatic masses. Also in our study, the frequency of small masses biopsied under endoscopic sonographic guidance (81.5%) was significantly higher than the frequency of those biopsied under CT guidance (32.6%). In fact, since the first report of the use of endoscopic biopsy for the diagnosis of pancreatic cancer in 1994 [15], several groups of researchers have suggested that endoscopic sonographically guided fine-needle aspiration biopsy should be more accurate than CT-guided fine-needle aspiration biopsy, especially for the diagnosis of small pancreatic masses [3, 13, 22]. To evaluate the effect of mass size on biopsy performance, we stratified our study results by mass size and found that there were no significant differences in test characteristics for the guidance techniques after the data were stratified by mass size; small masses were not biopsied more effectively under endoscopic sonography guidance than under CT guidance. Complication rates are low for both CTguided and endoscopic sonographically guided fine-needle aspiration biopsy of the pancreas. Recently, in a meta-analysis, Chen et al. [7] reported complication rates of 4% for CT-guided procedures and 2% for endoscopic sonographically guided procedures [7]. We encountered no complications. The most important limitation of our study is that we used a nonrandomized retrospective design. As a result, there may have been a referral bias that led to a significantly higher frequency of small masses biopsied under endoscopic sonographic guidance. Some referring physicians might have preferred that small pancreatic masses be biopsied under endoscopic sonographic guidance rather than CT guidance on the basis of other investigators opinions in prior reports [3, 13, 22] that endoscopic sonographically guided biopsy is more accurate than CT-guided biopsy for the diagnosis of small masses. However, to our knowledge, no published study has found a statistically significant difference in test characteristics among small masses. Indeed, the test characteristics for small masses in our study were similar for both guidance techniques. A second limitation is that a cytotechnologist was not routinely present during the endoscopic sonographically guided procedures to examine the adequacy of the sample. The diagnostic rate achieved with CT guidance was almost 100%. Although the diagnostic rate of biopsies performed under endoscopic sonographic guidance was slightly lower, there was no significant difference between the guidance techniques. Nevertheless, the diagnostic rate of endoscopic sonographically guided procedures might have been improved if a cytotechnologist had examined the initial specimens for adequacy. In summary, the diagnostic rate and sensitivity of CT-guided fine-needle aspiration biopsy of solid pancreatic masses were slightly higher than those of endoscopic biopsy; however, there were no significant differences between the techniques, and the NPVs of both techniques were almost equal. Hence, we conclude that when evaluating a solid pancreatic mass that needs a nonsurgical biopsy, either CT or endoscopic sonography may be chosen to guide the biopsy: CT has the advantage of not requiring oral intubation with an endoscope; endoscopic sonography the advantage of not requiring percutaneous puncture. References 1. Wallace MB, Woodward T, Raimondo M. EUS and fine-needle aspiration for pancreatic cancer. Dig Endosc 2004; 16:193 196 2. Harewood GC, Wiersema MJ. Endosonographyguided fine needle aspiration biopsy in the evaluation of pancreatic masses. Am J Gastroenterol 2002; 97:1386 1391 3. Mallery JS, Centeno BA, Hahn PF, Chang Y, Warshaw AL, Brugge WR. Pancreatic tissue sampling guided by EUS, CT/US, and surgery: a comparison of sensitivity and specificity. Gastrointest Endosc 2002; 56:218 224 4. Voss M, Hammel P, Molas G, et al. Value of endoscopic ultrasound guided fine needle aspiration biopsy in the diagnosis of solid pancreatic masses. Gut 2000; 46:244 249 5. Tillou A, Schwartz MR, Jordan PH Jr. Percutaneous needle biopsy of the pancreas: when should it be performed? World J Surg 1996; 20:283 286 6. Linder S, Blasjo M, Sundelin P, von Rosen A. Aspects of percutaneous fine-needle aspiration biopsy in the diagnosis of pancreatic carcinoma. J Surg 1997; 174:303 306 7. Chen VK, Arguedas MR, Kilgore ML, Eloubeidi MA. A cost-minimization analysis of alternative strategies in diagnosing pancreatic cancer. Am J Gastroenterol 2004; 99:2223 2234 8. Shin HJ, Lahoti S, Sneige N. Endoscopic ultrasound-guided fine-needle aspiration in 179 cases: the M. D. Anderson Cancer Center experience. Cancer 2002; 96:174 180 9. Agarwal B, Abu-Hamda E, Molke KL, Correa AM, Ho L. Endoscopic ultrasound-guided fine needle aspiration and multidetector spiral CT in the diagnosis of pancreatic cancer. Am J Gastroenterol 2004; 99:844 850 10. Jhala NC, Jhala D, Eltoum I, et al. EUS-guided fine-needle aspiration biopsy: a powerful tool to obtain samples from small lesions. Cancer 2004; 25:239 246 11. Rodriguez J, Kasberg C, Nipper M, Schoolar J, Riggs MW, Dyck WP. CT-guided needle biopsy of the pancreas: a retrospective analysis of diagnostic accuracy. Am J Gastroenterol 1992; 87:1610 1613 12. Edoute Y, Lemberg S, Malberger E. Preoperative and intraoperative fine needle aspiration cytology of pancreatic lesions. Am J Gastroenterol 1991; 86:1015 1019 13. Qian X, Hecht JL. Pancreatic fine needle aspiration: a comparison of computed tomographic and endoscopic ultrasonographic guidance. Acta Cytol 2003; 47:723 726 14. Brandt KR, Charboneau JW, Stephens DH, Welch TJ, Goellner JR. CT- and US-guided biopsy of the pancreas. Radiology 1993; 187:99 104 15. Chang KJ, Albers CG, Erickson RA, Butler JA, Wuerker RB, Lin F. EUS-guided fine needle aspiration of pancreatic carcinoma. Am J Gastroenterol 1994; 89:263 266 16. DeMay RM. Pancreas. In: De May RM, ed. The art and science of cytopathology. Chicago, IL: ASCP Press, 1996:1053 1082 17. MacCarty RL. Cyst fluid analysis and imaging of pancreatic cystic lesions. AJR 1995; 164:820 821 18. Lewandrowski KB, Southern JF, Pins MR, Compton CC, Warshaw AL. Cyst fluid analysis in the differential diagnosis of pancreatic cysts: a comparison of pseudocysts, serous cystadenomas, mucinous cystic neoplasms, and mucinous cystadenocarcinoma. Ann Surg 1993; 217:41 47 19. Hammel P, Levy P, Voitot H, et al. Preoperative cyst fluid analysis is useful for the differential diagnosis of cystic lesions of the pancreas. Gastroenterology 1995; 108:1230 1235 20. Eloubeidi MA, Chen VK, Eltoum IA, et al. EUSguided fine needle aspiration biopsy of patients with suspected pancreatic cancer: diagnostic accuracy and acute and 30-day complications. Am J Gastroenterol 2003; 98:2663 2668 21. Khalid A, Finkelstein S, McGrath K. Molecular diagnosis of solid and cystic lesions of the pancreas. Gastroenterol Clin North Am 2004; 33:891 906 22. Brugge WR. Pancreatic fine needle aspiration: to do or not to do? JOP 2004; 5:282 288 AJR:187, December 2006 1535