Power Doppler Endoscopic Ultrasonography for the Differential Diagnosis Between Pancreatic Cancer and Pseudotumoral Chronic Pancreatitis

Transcription

1 Article Power Doppler Endoscopic Ultrasonography for the Differential Diagnosis Between Pancreatic Cancer and Pseudotumoral Chronic Pancreatitis Adrian Săftoiu, MD, PhD, Carmen Popescu, MD, Sergiu Cazacu, MD, Daniela Dumitrescu, MD, Claudia Valentina Georgescu, MD, PhD, Mihai Popescu, MD, Tudorel Ciurea, MD, PhD, Florin Gorunescu, MD, PhD Objective. The accuracy of endoscopic ultrasonography (EUS) and EUS-guided fine-needle aspiration for the differential diagnosis of pancreatic masses is variable in the literature, being as low as 75% in some studies. The aim of the study was to assess the accuracy of power Doppler EUS for the differential diagnosis between pancreatic cancer and pseudotumoral chronic pancreatitis. Methods. We included 42 consecutive patients with pancreatic tumor masses (27 men and 15 women) examined by EUS between January 2002 and August Endoscopic ultrasonographic procedures included power Doppler EUS as well as EUS-guided fine-needle aspiration in all patients. Final diagnosis of pancreatic cancer was confirmed in 29 patients on the basis of a combination of information provided by imaging tests, follow-up of at least 6 months, and laparotomy in 18 patients for diagnostic or palliative reasons. Results. Sensitivity and specificity of the absence of power Doppler signals inside the suggestive pancreatic mass were 93% and 77%, respectively, with accuracy of 88%. Moreover, the addition of the information provided by the presence of peripancreatic collaterals improved the sensitivity and specificity to 97% and 92%, with accuracy of 95%. Conclusions. Power Doppler EUS provides useful information for the differential diagnosis of pancreatic masses. The results were in concordance with previous studies that showed a hypovascular pattern of pancreatic carcinoma, as well as the formation of collaterals in advanced cases due to the invasion of the splenic or portal veins. Further studies of dynamic EUS with contrast agents are necessary to better characterize pancreatic masses. Key words: chronic pancreatitis; endoscopic ultrasonography; pancreatic cancer; power Doppler ultrasonography. Abbreviations CT, computed tomography; EUS, endoscopic ultrasonography; FNA, fine-needle aspiration; ROC, receiver operating characteristic; TUS, transabdominal ultrasonography Received August 16, 2005, from the Departments of Gastroenterology (A.S., S.C., T.C.), Radiology and Imaging (D.D., M.P.), Pathology (C.V.G.), and Biostatistics (F.G.) and Cytology Laboratory (C.P.), University of Medicine and Pharmacy Craiova, Craiova, Romania. Revision requested September 6, Revised manuscript accepted for publication November 3, Dr Săftoiu thanks Prof Dr Peter Vilmann for critical review of the manuscript as well as for help and advice in mastering the endoscopic ultrasonographically guided fine-needle aspiration technique. Address correspondence to Adrian Săftoiu, MD, PhD, Department of Gastroenterology, University of Medicine and Pharmacy Craiova, strada Horia 11, Craiova, Dolj , Romania. adry@umfcv.ro Pancreatic cancer is the 10th most common malignancy, being the fourth leading cause of cancer death. Most of the patients are diagnosed by imaging studies that visualize pancreatic masses. 1 Because of the variable sensitivity and specificity, different methods are currently used: transabdominal ultrasonography (TUS), contrast-enhanced computed tomography (CT), magnetic resonance cholangiopancreatography, endoscopic ultrasonography (EUS), and endoscopic retrograde cholangiopancreatography. 2 However, the differential diagnosis of pancreatic cancer and pseudotumoral chronic pancreatitis is particularly difficult despite the progress of imaging techniques. Tissue diagnosis (cytologic or histologic examination) is essential in these cases, with EUS-guided fine-needle 2006 by the American Institute of Ultrasound in Medicine J Ultrasound Med 2006; 25: /06/$3.50

2 Power Doppler Endoscopic Ultrasonography of Pancreatic Masses aspiration (EUS-FNA) usually performed. Endoscopic ultrasonographically guided fine-needle aspiration is becoming the standard for obtaining cytologic diagnosis, although the sensitivity of EUS-FNA for the differential diagnosis of pancreatic masses is variable in the literature, being as low as 75% in some studies. 3 Moreover, the sensitivity was reported to be unacceptably lower (54%) in the context of chronic pancreatitis, whereas surgical resection was still necessary to confirm the diagnosis. 4 Consequently, EUS-FNA in patients with pancreatic masses has a low negative predictive value, and its ability to differentiate between pancreatic cancer and pseudotumoral chronic pancreatitis is limited. Pancreatic ductal adenocarcinoma has a poor prognosis, with a 5-year survival rate for patients undergoing curative resection of less than 10% to 20%. 5 Hence, the early diagnosis of pancreatic carcinoma in patients with chronic pancreatitis becomes crucial to establish the management and to improve prognosis. 6,7 The aim of our study was to prospectively assess the accuracy of linear EUS with power Doppler imaging to differentiate between pancreatic cancer and chronic pancreatitis in patients with pancreatic masses visualized by other imaging techniques. Materials and Methods Study Design and Patients The study prospectively included all the patients with a previous suspicion of pancreatic masses by ultrasonography or CT examined in the Department of Gastroenterology of the University of Medicine and Pharmacy Craiova over a 30- month period from January 2002 to August Thus, 42 consecutive patients with a suspicion of pancreatic tumor masses underwent power Doppler EUS, followed by EUS-FNA in all cases. The following information was gathered for each patient: age, sex, clinical results, other imaging procedures, EUS findings and diagnosis, TNM staging, type of therapy (chemotherapy or curative or palliative surgery), complications, followup (type and duration in months), and final pathologic diagnosis, where available. All patients provided written informed consent to undergo the EUS examination and were followed for a minimum of 6 months or until death occurred. The study did not use any procedures beyond our hospital protocol because all imaging tests (including EUS-FNA) are current procedures used in the daily clinical treatment of patients. All EUS examinations with EUS-FNA were performed by 2 gastroenterologists (A.S. and S.C.) according to a common protocol. Patients were placed in the left lateral decubitus position, and sedation with midazolam or propofol was applied. A final diagnosis of pancreatic cancer was confirmed in 29 patients on the basis of a combination of information provided by imaging tests (including EUS-FNA and cytologic confirmation of malignancy), follow-up of at least 6 months by repeated clinical, ultrasonographic, or CT examinations, as well as laparotomy with pathologic results in 18 patients, for diagnostic or palliative reasons. The other 13 patients had a diagnosis of chronic pseudotumoral pancreatitis and were followed for at least 6 months, by ultrasonography, CT, or repeated EUS-FNA. Power Doppler EUS Power Doppler EUS procedures were performed with a linear echoendoscope (GF-UCT 140 AL5; Olympus Optical Co, Ltd, Tokyo, Japan) coupled with the corresponding ultrasonography system (ProSound 5000; Aloka Co, Ltd, Tokyo, Japan). An ideal scanning plane of the mass was chosen, and efforts were made to interrogate the whole area and to assess the presence or absence of vascular signals. A careful optimization of the power Doppler technique was performed, and the gain was set to avoid the appearance of noise. The gain was also set as a function of the flow in the portal vein, which was assumed to be laminar and fluent. The procedures were recorded on a DVD recorder for latter review. The presence or absence of power Doppler signals inside the pancreatic mass was reviewed by both endoscopists (A.S. and S.C.), and discordant cases were solved by consensus. Furthermore, the examinations were further reviewed by 2 readers (D.D. and T.C.) who were absent during the EUS examinations and were blinded to the results of other imaging procedures (TUS, CT, and cytologic examination of EUS-FNA). The presence or absence of collateral circulation was also reviewed in all cases. To establish whether there were differences between the 2 groups of reviewers, a receiver operating characteristic (ROC) analysis was done to establish the accuracy of power Doppler EUS for the differential diagnosis (see below). 364 J Ultrasound Med 2006; 25:

3 Săftoiu et al EUS-FNA and Cytologic Analysis The EUS-FNA procedures were performed according to a common protocol, which included a minimum of 3 passes with a minimum of 10 to-and-fro movements in each pass that were performed under continuous aspiration, using a technique previously published in detail. 8 The needle (Olympus NA-10J-1) was directly visualized under real-time conditions during the procedure, thus ensuring the correct placement inside the suggestive pancreatic mass. Endoscopic ultrasonographically guided fine-needle aspiration was performed in various locations within the lesion, both in the center and periphery, by subtle adjustments of the needle and scope positions (using both the up-down wheel and the elevator) to maximize the likelihood of obtaining diagnostic material. If multiple lesions were sampled in the same patient, EUS-FNA was performed in a successive order (distant metastases, ascites, local lymph nodes, and, finally, the primary tumor) to prevent a potential false-positive upstaging of the disease. 9 Between 3 and 6 passes were necessary in each patient to obtain sufficient material, that is, not only blood or scant cellularity samples. A cytopathologist (C.P.) was in attendance during all the procedures, prepared the smears, verified the adequacy of specimens, and advised as to the need of additional passes. The aspirated material obtained by EUS-FNA was sprayed onto glass slides, and the resulting monolayer smears were colored by both Giemsa and Papanicolaou stains in all cases. Thus, the Papanicolaou stain was done after wet fixation in ethanol for at least 5 minutes, whereas the Giemsa stain was done after dry fixation and postfixation with methanol. The smears were individually characterized concerning the cells, nuclei, nucleoli, and chromatin and were further interpreted as benign, atypical probably benign, atypical probably malignant, suggestive of malignancy, and malignant. A conclusion of negative or positive concerning the diagnosis of malignancy was recorded, with all suggestive cases regarded as malignant. Immunocytochemistry was used in selected cases for the confirmation of the pancreatic origin (acinar or ductal), exclusion of metastases (eg, gastric carcinomas, breast carcinomas, pancreatic lymphoma, and pulmonary carcinomas) and the diagnosis of neuroendocrine differentiations or neuroendocrine tumors. Statistical Analysis All results are expressed as mean ± SD. The sensitivity, specificity, positive predictive value, negative predictive value, and accuracy for the differential diagnosis between pancreatic cancer and chronic pseudotumoral pancreatitis were calculated for power Doppler EUS, as well as for EUS-FNA, by a comparison of the results of these techniques with the final diagnosis. Receiver operating characteristic curve analysis was done separately for power Doppler EUS results of the 2 groups of readers (blinded and unblinded) and also for the results of EUS-FNA. In ROC analysis, the differential diagnosis implies a classification of the results as either positive (pancreatic cancer) or negative (chronic pancreatitis) based on a level of confidence of each case on an ordinal scale (1 5). 10 Consequently, the results of power Doppler EUS and EUS-FNA were classified by the examiners as definitely negative (1), probably negative (2), possibly positive (3), probably positive (4), or definitely positive (5). Receiver operating characteristic analysis thus displays the range of tradeoffs between true-positive and false-positive rates possible with the test. This method fits well to our task because image assessments are made subjectively, and implicit criteria for assessment may vary among examiners. Results The study included 42 consecutive patients with a clinical suspicion of pancreatic tumor masses (27 men and 15 women) examined by EUS, between January 2002 and August The patients were included prospectively based on a suspicion of a pancreatic mass by ultrasonography or CT. Endoscopic ultrasonographic procedures included color and power Doppler ultrasonography as well as EUS-FNA in all patients. The characteristics of the patient subgroups and EUS examinations are presented in Tables 1 3. The diagnosis of pancreatic cancer was confirmed in 29 patients on the basis of a combination of information provided by imaging tests (including EUS with FNA biopsy and cytologic confirmation of malignancy), clinical follow-up of 6 months, and laparotomy in 18 patients for diagnostic or palliative reasons (Tables 1 and 2). Thus, the subgroup of patients with pancreatic cancer included 21 men and 8 women with a J Ultrasound Med 2006; 25:

4 Power Doppler Endoscopic Ultrasonography of Pancreatic Masses Table 1. Pancreatic Cancer Without Chronic Pancreatitis (n = 25) Size, Mass Power Collaterals TNM Stage EUS-FNA Final Diagnosis Case Age, y Sex mm (Location) Doppler (Location) (CT + EUS) Result (Follow-up) Follow-up, mo 1 56 F 23 Head Absent Absent T1 N0 M0 Positive Surgery M 35 Head Absent Head T4 N1 M0 Positive Surgery M 45 Head Absent Head, body T4 N1 M1 Positive CT, EUS-FNA F 17 Head Absent Absent T1 N0 M0 Negative Surgery M 30 Head Absent Absent T2 N0 M0 Positive Surgery M 40 Head Absent Absent T4 N1 M1 Positive CT, EUS-FNA M 35 Head Absent Head T4 N1 M1 Positive Surgery M 45 Head Absent Absent T4 N1 M1 Positive CT, EUS-FNA M 40 Head Absent Head T4 N1 M0 Positive CT, EUS-FNA F 40 Body Absent Body T4 N1 M0 Positive Surgery M 55 Body Absent Absent T4 N1 M0 Positive Surgery M 20 Uncinate Absent Absent T2 N0 M0 Negative Surgery M 25 Head Absent Absent T4 N0 M0 Positive CT, EUS-FNA F 30 Head Absent Absent T4 N0 M0 Positive Surgery M 50 Head Absent Head T4 N1 M0 Positive CT, EUS-FNA F 18 Head Absent Absent T2 N1 M0 Positive Surgery M 70 Body Absent Absent T4 N1 M0 Positive Surgery F 20 Head Absent Absent T2 N0 M0 Positive Surgery F 40 Head Absent Absent T4 N0 M0 Positive Surgery M 40 Body Absent Absent T3 N1 M0 Positive Surgery M 35 Body Absent Body, tail T4 N1 M1 Positive CT, EUS-FNA F 32 Head Present Absent T4 N1 M0 Positive Surgery M 50 Head Present Head T4 N1 M0 Positive Surgery F 50 Uncinate Absent Absent T4 N1 M1 Positive CT, EUS-FNA M 40 Head Absent Absent T4 N1 M0 Positive Surgery 6 F indicates female; and M, male. mean age ± SD of 62.2 ± 12.2 years. The pancreatic tumors were located in the head (20 patients [69%]), uncinate process (2 patients [6.9%]), body (5 patients [17.2%]), or tail (2 patients [6.9%]) of the pancreas. Four patients had clinical and EUS criteria of chronic pancreatitis (Table 2). The diagnosis of pancreatic cancer in patients with chronic pancreatitis was initially suspected on imaging (presence of a mass by ultrasonography, CT, and EUS) and confirmed by EUS-FNA in 3 patients and surgery with pathologic results in another patient. The imaging findings in this subgroup of patients did not differ from those in the patients with pseudotumoral chronic pancreatitis, with the exception of power Doppler EUS. The tumors appeared by EUS as hypoechoic masses, inhomogeneous, of variable dimensions between 17 and 50 mm in diameter. Results of endoscopic ultrasonographically guided fineneedle aspiration (Figure 1A) with cytologic (Figure 1B) and immune cytochemical analysis were positive in 26 cases of pancreatic cancer and false-negative in 3 patients. There were no false-positive results of EUS-FNA in the patients with chronic pancreatitis. Consequently, the sensitivity, specificity, and accuracy of EUS-FNA were 90%, 100%, and 93%, respectively. Although the positive predictive value was 100%, the negative predictive value was only 81%. There were no complications recorded after EUS-FNA. Table 2. Pancreatic Cancer With Chronic Pancreatitis (n = 4) Size, Mass Power Collaterals TNM Stage EUS-FNA Final Diagnosis Case Age, y Sex mm (Location) Doppler (Location) (CT + EUS) Result (Follow-up) Follow-up, mo 1 67 M 35 Tail Absent Body, tail T4 N0 M0 Negative Surgery M 50 Head Absent Body T4 N1 M1 Positive CT, EUS-FNA F 20 Head Absent Absent T2 N0 M0 Positive CT, EUS-FNA M 40 Tail Absent Absent T4 N1 M0 Positive CT, EUS-FNA 9 F indicates female; and M, male. 366 J Ultrasound Med 2006; 25:

5 Săftoiu et al Table 3. Chronic Pancreatitis (n = 13) Size, Mass Power Collaterals EUS EUS-FNA Final Diagnosis Case Age, y Sex mm (Location) Doppler (Location) Criteria* Result (Follow-up) Follow-up, mo 1 37 F 40 Tail Present Absent 5 Negative Surgery M 30 Head Present Absent 4 Negative Ultrasonography, CT F 25 Head Absent Body 6 Negative Ultrasonography, CT F 40 Head Present Absent 6 Negative EUS-FNA M 45 Head Present Absent 4 Negative Ultrasonography, CT M 15 Head Absent Absent 5 Negative Surgery M 30 Head Absent Absent 6 Negative Ultrasonography, CT F 28 Head Present Absent 4 Negative EUS-FNA M 33 Head Present Absent 5 Negative EUS-FNA M 50 Head Present Absent 7 Negative Surgery M 30 Head Present Absent 4 Negative EUS-FNA M 25 Head Present Absent 4 Negative Ultrasonography, CT F 22 Body Present Head 6 Negative EUS-FNA 28 F indicates female; and M, male. *EUS criteria of chronic pancreatitis included inhomogeneity of parenchyma, hyperechoic foci or bands, honeycomb appearance of the parenchyma, irregular pancreatic duct, pancreatic duct dilatation (>3, 2, and 1 mm at the level of head, body, and tail, respectively), hyperechoic margins of the pancreatic duct, calcifications, and pseudocysts. The patients with pancreatic tumors had absent power Doppler signals inside the tumor mass in 27 cases (Figure 2), whereas 2 patients had hypervascular tumors (Figure 3). Sensitivity and specificity of the absence of power Doppler signals inside the suggestive pancreatic mass were 93% and 77%, with accuracy of 88%. The positive predictive value and negative predictive value were 90% and 83%, respectively. Moreover, the addition of the information provided by the presence of peripancreatic or periduodenal collaterals (Figure 4) improved the sensitivity and specificity to 97% and 92%, with accuracy of 95%. The positive predictive value and negative predictive value were 97% and 92%, respectively. The mean follow-up of the patients with pancreatic cancer was 8.9 ± 6.7 months and consisted of repeated clinical, ultrasonographic, or CT examinations. Moreover, 13 patients were dead or had a diagnosis of liver metastases during the follow-up. Chronic pseudotumoral pancreatitis was diagnosed in 13 patients, who were subsequently followed for a minimum of 6 months (Table 3). Although the patients had suggestive masses on Figure 1. A, Pancreatic head adenocarcinoma of approximately mm in diameter (arrows), visualized as a hypoechoic, inhomogeneous mass, with indefinite margins. Portal vein invasion was evident posterior to the pancreatic head (arrowheads). Endoscopic ultrasonographically guided FNA with real-time visualization of the needle inside the tumor was subsequently performed. B, The cytologic smears showed clumps of atypical cells (arrows) with a small island of normal acinar cells (arrowheads) in a background of erythrocytes (Papanicolaou stain). A B J Ultrasound Med 2006; 25:

6 Power Doppler Endoscopic Ultrasonography of Pancreatic Masses Figure 2. Hypoechoic tumor mass due to pancreatic adenocarcinoma of approximately mm, visualized at the level of the pancreatic head (arrows), with absent power Doppler signals and retrograde dilatation of the main pancreatic duct (arrowheads). Figure 3. Hypoechoic, inhomogeneous mass due to pancreatic adenocarcinoma, of approximately mm, visualized at the level of the pancreatic head (arrows), with indefinite margins and power Doppler signals inside. ultrasonography or CT and even inhomogeneous, hypoechoic masses on the EUS examination, they also had increased vascularization depicted by power Doppler imaging inside the tumor mass in 10 of 13 patients (Figure 5). The patients with concomitant chronic pancreatitis and pancreatic cancer were the most difficult to examine because of the inhomogeneous pancreatic parenchyma, with hyperechoic foci and bands. In these cases, however, the absence of power Doppler signals inside the suggestive pancreatic mass and the presence of scattered signals inside the pancreatic parenchyma (Figure 6A) were helpful for the correct diagnosis, confirmed by EUS-FNA (Figure 6B). The presence of pancreatic cancer was excluded by a combination of imaging tests (ultrasonography, CT, and EUS-FNA) repeated during the mean follow-up of 19.1 ± 7.5 months. Because of the suspicion of pancreatic cancer, EUS-FNA was repeated in 5 of 13 patients, again with negative results. Separate ROC analyses were conducted for the results of power Doppler EUS, with the same results in both blinded and unblinded readers, as well as for the interpretation of EUS-FNA (Figure 7). The differential diagnosis consisted of a classification of the results as either positive (pancreatic cancer) or negative (chronic pancreatitis) based on an ordinal scale of 1 to 5 in each case. The areas under the ROC curve were (95% confidence interval, ) for EUS-FNA and (95% confidence interval, ) for power Doppler EUS. The difference was not statistically significant (P =.1342). Discussion Endoscopic ultrasonography coupled with EUS- FNA is very useful for the detection of pancreatic tumors, differential diagnosis between benign and malignant tumors, and staging and resectability. 1 Endoscopic ultrasonography was previously reported to be superior to TUS and conventional CT, especially for the detection of small pancreatic tumors of less than 2 to 3 cm. 2,11 Endoscopic ultrasonography was further compared with helical CT, being either equivalent or superior for the diagnosis of pancreatic cancer. 12,13 A recent prospective study assessed multidetector CT and EUS for the detection of Figure 4. Hypoechoic tumor due to pancreatic adenocarcinoma, of mm, visualized at the level of the pancreatic body (arrows), with absent power Doppler signal inside. Collateral circulation was visualized in power Doppler mode in front of the pancreatic body. 368 J Ultrasound Med 2006; 25:

7 Săftoiu et al Figure 5. Chronic pancreatitis with a hypoechoic, inhomogeneous mass at the level of the pancreatic head (arrows), of approximately mm. Power Doppler EUS showed intense vascularization of the inflammatory mass (arrows), whereas EUS-FNA results were negative for atypical cells. pancreatic cancer, concluding that EUS is superior because it identifies tumors that are undetected by CT. 14 However, even EUS can miss diffusely infiltrative pancreatic tumors or neoplasms in the setting of chronic pancreatitis. Hence, in case of a high suspicion of pancreatic neoplasm, follow-up EUS after 2 to 3 months might be necessary. 15 Although EUS has high sensitivity for the detection of pancreatic masses, it has a limited ability to differentiate between inflammatory masses and cancer. 16 The next step was to use EUS-FNA for a correct discrimination between pancreatic tumor masses. Accuracy of EUS with EUS-FNA for the differential diagnosis of pancreatic masses is variable, usually 90% to 95% but as low as 75% in some studies. 3,17 21 Most studies report high specificity (close to 100%) and positive predictive value. However, the negative predictive value was reported to be low because of falsenegative results, which may influence treatment of the patients, with disastrous results. 3 The same results were obtained in our study, with excellent accuracy of 93% but a lower negative predictive value of only 81%. Three patients had negative EUS-FNA findings despite multiple passes, probably because of sampling errors caused by the scant cellularity of specimens due to the small size of the lesions (2 patients) or the presence of chronic pancreatitis in another patient. Other tumor-related factors may also decrease the cellularity of samples, including extensive fibrosis (desmoplastic reaction) and necrosis, as well as the degree of differentiation (welldifferentiated tumors require a larger number of needle passes than moderately or poorly differentiated tumors). 8,21,22 The diagnostic yield of EUS-FNA in pancreatic cancer is also dependent on the technique and the experience of the team (endoscopist and cytopathologist). In our study, the ability of imaging methods (EUS with power Doppler imaging) was similar to that of EUS-FNA for the differential diagnosis of pancreatic cancer and pseudotumoral inflammatory masses. Thus, the accuracy of the absence of power Doppler signals inside the mass in diagnosing pancreatic cancer was 88%, with a negative predictive value of 83%, comparable with the accuracy and negative predictive value of EUS-FNA of 93% and 81%, respectively. Figure 6. A, Chronic pancreatitis with an inhomogeneous pancreatic head containing calcifications, with a suggestion of a pancreatic head tumor mass (arrows). The mass had no power Doppler signals inside, whereas the diagnosis of adenocarcinoma was confirmed by EUS-guided FNA. B, The cytologic smears showed tridimensional clumps of atypical cells, indicating the presence of adenocarcinoma (arrows), in a background of erythrocytes (Giemsa stain). A B J Ultrasound Med 2006; 25:

8 Power Doppler Endoscopic Ultrasonography of Pancreatic Masses Figure 7. Comparative ROC analysis of power Doppler EUS and EUS-FNA based on a classification of the results as either positive (pancreatic cancer) or negative (chronic pancreatitis), determined according to an ordinal scale of 1 to 5 in each case. Receiver operating characteristic analyses also showed excellent accuracy for power Doppler EUS and EUS-FNA, with a difference between areas under the ROC curves that was not statistically significant (0.975 and 0.857; P =.1342). Moreover, the addition of the information provided by the appearance of collaterals enhanced the diagnostic value of power Doppler ultrasonography, with better accuracy of 95% and a higher negative predictive value of 92%. Consequently, EUS with power Doppler imaging provides information about the etiology of the tumor mass, even in the absence of tissue confirmation that may occur in patients with chronic pancreatitis and pseudotumoral inflammatory masses. The high negative predictive value might also allow a decrease of the number of unnecessary surgical interventions, thus reducing the associated morbidity and mortality. The ability of power Doppler EUS to differentiate pancreatic masses has to be viewed as an adjunct to the other imaging techniques rather than a replacement of tissue confirmation. Currently, there is no imaging method that can reliably provide this capability in patients with pancreatic masses, especially in the setting of chronic pancreatitis. Nevertheless, categorizing the risk of malignancy is very important for the clinical decision-making process and subsequent treatment (follow-up CT and EUS, repeated EUS-FNA, or surgery), especially in the cases with negative EUS-FNA findings, in which a diagnosis of pancreatic cancer cannot be excluded. Dynamic imaging is increasingly used for the differential diagnosis of pancreatic masses. 2 Thus, contrast-enhanced power Doppler ultrasonography was used for the transabdominal characterization of pancreatic masses Pancreatic carcinoma was described as usually hypovascular compared with the rest of the parenchyma, whereas inflammatory masses are isovascular or hypervascular. New progress in ultrasonography systems and transducer technology has facilitated the application of contrast-enhanced coded phase (pulse) inversion harmonic or wideband ultrasonographic techniques for a better characterization of pancreatic masses These techniques showed the presence of small tumor vessels in approximately 65% of pancreatic ductal carcinomas. The pancreatic carcinoma masses were relatively hypovascular compared with surrounding parenchyma. However, poor visualization of the tumor mass and operator dependence limit the transabdominal visualization of the pancreas, disturbing the examinations in up to one third of patients because of large body habitus, overlying bowel gas, or presence of ascites. 2 The introduction of new techniques, such as phase inversion tissue harmonic imaging, did not significantly increase the sensitivity for the detection of pancreatic lesions compared with conventional B-mode ultrasonography (70% versus 60%; P =.46). 29 In fact, all these methods based on TUS are limited only for patients in whom the pancreatic tumor and parenchyma can be clearly observed in a single view. 26 Contrast-enhanced EUS was also used for the differential diagnosis of pancreatic tumor masses for a better assessment of perfusion in the pancreatic tissue and inside the mass. 30,31 Pancreatic carcinoma was again shown to be relatively hypovascular compared with surrounding pancreatic tissue, whereas markedly hypervascular lesions were inflammatory masses. Moreover, the results were comparable with cytopathologic results (percutaneously or EUS guided) with high sensitivity (94%) and specificity (100%). 31 The results of our study showed that power Doppler EUS was an accurate discriminative test, even without the enhancement produced by contrast agents, especially when the information provided by the presence of collaterals was added. Although collaterals might also appear in chronic pancreatitis because of segmentary portal hypertension, the relative frequency is inferior to the frequency of collateral appearance in patients with pancreatic cancer J Ultrasound Med 2006; 25:

9 Săftoiu et al Consequently, combining the information provided by the absence of power Doppler signals and the presence of pancreatic collaterals yielded the best accuracy for the differential diagnosis between pseudotumoral chronic pancreatitis and pancreatic cancer. One limitation of this study was the absence of histopathologic confirmation in 11 patients with pancreatic cancer who did not undergo surgery. However, the diagnosis of pancreatic cancer was confirmed in 29 patients on the basis of a combination of EUS-FNA (positive cytologic results in 26 patients), laparotomy with biopsies (positive pathologic results in 16 patients), and mean follow-up of 8.9 months (including 13 patients who were dead or had a diagnosis of liver metastases). The other 13 patients with a diagnosis of chronic pseudotumoral pancreatitis based on imaging studies and negative EUS-FNA findings were followed for a mean of 19.1 months by ultrasonography, CT, or repeated EUS with FNA. Another limitation was that the endoscopists were not blinded to the results of ultrasonographic or CT scanning obtained before the EUS-FNA examination combined with power Doppler imaging. Consequently, a considerable selection bias might have influenced the results. However, we included consecutive patients who had a suggestion of a pancreatic mass, whereas the results of EUS with power Doppler imaging were reviewed by other authors blinded by the results of previous imaging studies. The pathologists were also blinded to the results of imaging tests. The third limitation was the small number of patients with chronic pseudotumoral pancreatitis, which might have influenced the statistical significance of the positive and negative predictive values, as well as the overall accuracy of the imaging methods. The presence of power Doppler signals inside the inflammatory masses is variable as a function of inflammation and necrosis, with a considerable bias that might have been determined by the small number of patients included. In conclusion, power Doppler EUS provides useful information for the differential diagnosis of pancreatic masses. The results of our study are in concordance with those of previous studies that showed a hypovascular pattern of pancreatic carcinoma, as well as the formation of collaterals in advanced cases due to the invasion of the splenic or portal veins. Further prospective studies of dynamic EUS with contrast agents are necessary to better characterize pancreatic masses. References 1. Horwhat JD, Gress FG. Defining the diagnostic algorithm in pancreatic cancer. JOP J Pancreas [serial online] 2004; 5: Takhar AS, Palaniappan P, Dhinga R, Lobo DN. Recent developments in diagnosis of pancreatic cancer. BMJ 2004; 329: Eloubeidi MA, Chen VK, Eltoum IA, et al. Endoscopic ultrasound-guided fine-needle aspiration biopsy of patients with suspected pancreatic cancer: diagnostic accuracy and acute and 30-day complications. Am J Gastroenterol 2003; 98: Fritscher-Ravens A, Brand L, Knofel WT, et al. Comparison of endoscopic ultrasound-guided fine needle aspiration for focal pancreatic lesions in patients with normal parenchyma and chronic pancreatitis. Am J Gastroenterol 2002; 97: Carpelan-Holmström M, Nordling S, Pukkala E, et al. Does anyone survive pancreatic ductal adenocarcinoma? A nationwide study re-evaluating the data of the Finnish Cancer Registry. Gut 2005; 54: Talamini G, Bassi C, Falconi M, et al. Early detection of pancreatic cancer following the diagnosis of chronic pancreatitis. Digestion 1999; 60: Canto MI, Goggins M, Yeo CJ, et al. Screening for pancreatic neoplasia in high-risk individuals: an EUS-based approach. Clin Gastroenterol Hepatol 2004; 2: Vilmann P, Hancke S. A new biopsy handle instrument for endoscopic ultrasound guided biopsy. Gastrointest Endosc 1996; 43: Chang KJ. Maximizing the yield of EUS-guided fine-needle aspiration. Gastrointest Endosc 2002; 56:S28 S Eng J. ROC Analysis: Web-Based Calculator for ROC Curves. Baltimore, MD: Johns Hopkins University. Available at: Accessed September 15, Gress PG, Hawes RH, Savides TJ, et al. Role of EUS in the preoperative staging of pancreatic cancer: a large singlecenter experience. Gastrointest Endosc 1999; 50: Legmann P, Vignaux O, Dousset B, et al. Pancreatic tumors: comparison of dual-phase helical CT and endoscopic sonography. AJR Am J Roentgenol 1998; 170: Mertz HR, Sechopoulos P, Delbeke D, et al. EUS, PET, and CT scanning for evaluation of pancreatic adenocarcinoma. Gastrointest Endosc 2000; 52: DeWitt J, Devereaux B, Chriswell M, et al. Comparison of endoscopic ultrasonography and multidetector computed tomography for detecting and staging pancreatic cancer. Ann Intern Med 2004; 141: Bhutani MS, Gress FG, Giovannini M, et al. The NO Endosonographic Detection of Tumor (NEST) study: a case series of pancreatic cancers missed on endoscopic ultrasonography. Endoscopy 2004; 36: J Ultrasound Med 2006; 25:

10 Power Doppler Endoscopic Ultrasonography of Pancreatic Masses 16. Brand B, Pfaff T, Binmoeller KF, et al. Endoscopic ultrasound for differential diagnosis of focal pancreatic lesions, confirmed by surgery. Scand J Gastroenterol 2000; 35: Agarwal B, Abu-Hamda E, Molke KL, et al. Endoscopic ultrasound-guided fine needle aspiration and multidetector spiral CT in the diagnosis of pancreatic cancer. Am J Gastroenterol 2004; 100: Chang KJ, Nguyen P, Erickson RA, et al. The clinical utility of endoscopic ultrasound-guided real-time fine-needle aspiration in the diagnosis and staging of pancreatic carcinoma. Gastrointest Endosc 1997; 45: Bhutani MS, Hoffman BJ, van Velse A, Hawes RH. Contrastenhanced endoscopic ultrasonography with galactose microparticles: SHU508A (Levovist). Endoscopy 1997; 29: Becker D, Strobel D, Bernatik T, Hahn EG. Echo-enhanced and power Doppler EUS for the discrimination between focal pancreatitis and pancreatic carcinoma. Gastrointest Endosc 2001; 53: Eloubeidi MA, Iseman DT, Chen VK, et al. Prevalence and significance of periduodenal venous collaterals in patients evaluated for pancreaticobiliary disorders by endosonography. Endoscopy 2003; 35: Faigel DO, Ginsberg GG, Bentz JS, et al. Endoscopic ultrasound-guided real-time fine-needle aspiration biopsy of the pancreas in cancer patients with pancreatic lesions. J Clin Oncol 1997; 15: Wiersema MJ, Vilmann P, Giovannini M, et al. Endosonography-guided fine-needle aspiration biopsy: diagnostic accuracy and complication assessment. Gastroenterology 1997; 112: Eloubeidi MA, Jhala D, Chhieng DC, et al. Yield of endoscopic ultrasound-guided fine-needle aspiration biopsy in patients with suspected pancreatic carcinoma. Cancer 2003; 99: Erickson RA. EUS-guided FNA. Gastrointest Endosc 2004; 60: Ueno N, Tomiyama T, Tano S, et al. Contrast-enhanced color Doppler ultrasonography in diagnosis of pancreatic tumor: two case reports. J Ultrasound Med 1996; 15: Koito K, Namieno T, Nagakawa T, Morita K. Inflammatory pancreatic masses: differential from ductal carcinomas with contrast-enhanced sonography using carbon dioxide microbubbles. AJR Am J Roentgenol 1997; 169: Rickes S, Unkrodt K, Neye H, Ocran KW, Wermke W. Differentiation of pancreatic tumors by conventional ultrasound, unenhanced and echo-enhanced power Doppler sonography. Scand J Gastroenterol 2002; 37: Oshikawa O, Tanaka S, Ioka T, et al. Dynamic sonography of pancreatic tumors: comparison with dynamic CT. AJR Am J Roentgenol 2002; 178: Kitano M, Kudo M, Maekawa K, et al. Dynamic imaging of pancreatic diseases by contrast enhanced coded phase inversion harmonic ultrasonography. Gut 2004; 53: Ohshima T, Yamaguchi T, Takeshi I, et al. Evaluation of bloody flow in pancreatic ducal carcinoma using contrastenhanced, wide-band Doppler ultrasonography: correlation with tumor characteristics and vascular endothelial growth factor. Pancreas 2004; 28: Hohl C, Schmidt T, Haage P, et al. Phase-inversion tissue harmonic imaging compared with conventional B-mode ultrasound in the evaluation of pancreatic lesions. Eur Radiol 2004; 14: J Ultrasound Med 2006; 25: