Endoscopic ultrasonography in the diagnosis and staging of pancreatic neoplasms (original) (raw)

Review Article

Endoscopic Ultrasonography in the Diagnosis and Staging of Pancreatic Neoplasms

Petr Protiva, Anand V. Sahai, and Banke Agarwal*
St. Louis University School of Medicine, 3635 Vista Avenue, St. Louis, MO 63110

Summary

Abstract

Background. Pancreatic cancer is the fourth leading cause of death among Americans. Twenty-eight thousand cases of pancreatic cancer are diagnosed annually and about the same number of patients die of pancreatic cancer every year. Most patients with pancreatic cancer are diagnosed when the tumor is 3 cm or more in diameter. Most pancreatic cancers are metastatic at the time of diagnosis and the median survival is only 18−2018-20 mo. Overall, actual 5-yr survival is about 10%10 \% and has not changed much over several decades. Curative surgical resection is currently believed to offer the only chance of long-term survival in these patients. Difficulty in the early diagnosis of pancreatic cancer remains a major obstacle in improving outcomes in these patients. Screening for pancreatic cancer is currently not recommended. However, recent developments in endoscopic ultrasound (EUS) and cytological EUS-guided fine-needle aspiration (FNA) allow early diagnosis of pancreatic cancer with histological confirmation. EUS-FNA in conjunction with helical CT also provides reliable preoperative staging of pancreatic tumors. EUS with FNA therefore appears to be a promising tool in the fight against pancreatic cancer.

Burden of Suffering

The incidence of pancreatic cancer in the United States is 28,000 cases per year with 8.5 deaths/ 100,000. The majority of patients are between 65−7965-79 years of age. Ninety percent of cancers of the pancreas are adenocarcinomas. The majority of pancreatic cancers are approx 3 cm in diameter at initial diagnosis and are diagnosed when they become symptomatic or are detected incidentally in a small number of patients (1). Eighty-ninety percent of pancreatic cancers have regional or distant involvement at the time of initial diagnosis (2). Biologically, pancreatic adenocarcinoma behaves aggressively and only 5−10%5-10 \% of the patients are

[1]alive at 5 yr from diagnosis. Curative surgical resection offers the only chance of cure in pancreatic cancer.

Computed tomography (CT) scan, magnetic resonance imaging (MRI), and abdominal ultrasound (US), which were until recently the only imaging tests for the detection and staging of pancreatic cancers are not sufficiently accurate. Consequently, only 5−25%5-25 \% of patients taken for surgery actually undergo curative resection (1,3)(1,3). Even in cases of surgical resection with clear margins, the 5-yr survival is about 25%(1,3,4)25 \%(1,3,4). Whipple’s resection, currently the standard of care for resection of pancreatic cancer, is associated with considerable morbidity and mortality. Therefore, there is a need for improvements in strategies for diagnosis and staging of pancreatic cancer to: 1) identify early pancreatic cancers for curative surgical resection, and 2) reliably identify patients with surgically unresectable disease, to obviate

1. *Author to whom all correspondence and reprint requests should be addressed: Banke Agarwal, Director, Therapeutic Endoscopy, St. Louis University School of Medicine, 3635 Vista Avenue, St. Louis, MO 63110. Email: Bagarwal@pol.net ↩︎

the need for a potentially debilitating surgical procedure.

Meaningful improvement of the early diagnosis of pancreatic cancer implies that pancreatic tumors be detected when they are still confined to the pancreas. Unfortunately, most of the pancreatic cancer, less than 2 cm in size are asymptomatic and a small number of these patients present for medical attention with nonspecific symptoms such as dyspepsia, back pain, abdominal pain, and depression. Screening of asymptomatic individuals for pancreatic cancer is controversial and currently not recommended. A study of a mass screening of more than 10,000 asymptomatic persons for pancreatic cancer in Japan, using either ultrasonography alone or CA 19-9 plus elastase-1, found the likelihood of pancreatic cancer given a positive screening test to be 0.5%0.5 \%. Only one of the four cancers discovered could be curably resected (5).

Diagnosis of Pancreatic Cancer

Endoscopic Retrograde Cholangiopancreatography (ERCP)

Most patients with pancreatic cancer who have obstructive jaundice as the presenting symptoms undergo an ERCP for biliary drainage and to rule out choledocholithiasis. The majority ( 85−90%85-90 \% ) of biliary strictures in patients with obstructive jaundice without evidence of choledocholithiasis are malignant, usually because of pancreatic cancer, cholangiocarcinoma, or metastases. The sensitivity of ERCP for diagnosing pancreatic cancer varies from 85 to 95%95 \%, respectively. Adamek et al. reported a sensitivity of 81%81 \% and specificity of 88%88 \% with positive predictive value of 89%89 \% for malignant lesions in the pancreatic head (6). Another prospective controlled trial reported a sensitivity and specificity for diagnosing pancreatic cancer of 70%70 \% and 94%94 \%, respectively (7). The “double duct sign,” which was regarded as pathognomic of periampullary cancers seems to be only 85%85 \% specific (8).

ERCP (using stricture morphology and brush cytology) is not accurate in reliably determining the etiology of biliary strictures. Identification of patients with benign biliary strictures remains a major challenge. In view of 85−90%85-90 \% likelihood of cancer in patients with obstructive jaundice and common bile duct strictures, these patients are often taken to the
operating room to attempt resection of tumor. Reliable preoperative diagnosis and determination of resectability is highly desirable in these patients as 1) 10−15%10-15 \% of these patients do not have cancer, 2) 50−80%50-80 \% of patients with pancreatic cancer have unresectable disease ( 9−129-12 ), 3) patients with incomplete resection have no survival benefit (1,2), and 4) considerable morbidity ( 40%40 \% or more) and mortality (1−10%)(1-10 \%) are associated with Whipple resection (2,13,14)(2,13,14).

Helical CT

Helical CT allows thin section scanning of the pancreas, allowing a much higher resolution than can be achieved with earlier generation of CT scanners. Several studies have documented that helical CT with the proper contrast-enhancement technique is indeed an improvement over conventional imaging techniques in the diagnosis of pancreatic tumors (15-18). With dramatic increases in computing power of CT scanners, it is now possible to have a three-dimensional (3-D) reconstruction of the vascular anatomy. This allows radiologists to determine whether there is infiltration of major peri-pancreatic vessels by pancreatic tumors (19). Helical CT scanners are still not widely available. Nonetheless, Helical CT with 3-D reconstruction represents a major advance in the diagnosis and treatment of pancreatic cancers. The technique is discussed elsewhere in this issue.

Endoscopic Ultrasound

Endoscopic ultrasound (EUS) has now graduated from being an experimental technique to becoming invaluable in diagnosis and management of pancreatic cancers. EUS is performed using specialized endoscopes (echo-endoscopes) that have ultrasound transducers mounted on their tips. As a result, the ultrasound transducers can be positioned in the stomach or the duodenum in close proximity to pancreas. Because of smaller distances between the ultrasound transducer and pancreas, much higher ultrasound frequencies ( 7.5 and 12 MHz vs 5 MHz or lower used with abdominal ultrasound) can be used to image the pancreas. In addition, some echoendoscopes have Doppler capabilities, so that the vascular structures can be reliably identified. An important advance in last few years has been the ability to perform EUS-guided fine-needle aspiration (FNA) of

Fig. 1. (A) Radial echoendoscope. (B) EUS image of pancreatic tumor using radial echoendoscope.
suspicious lesions with needles that are passed through the biopsy channel of the echoendoscopes. In adequately trained hands, EUS-guided FNA is safe and provides adequate specimens for reliable cytologic diagnosis.

Two types of endoscopes are currently availableradial and linear array. The radial echoendoscope uses a mechanically rotating single-scanning transducer that allows a 360∘360^{\circ} scanning perpendicular to the long axis of the endoscope shaft. It provides easily interpretable high-resolution cross-sectional images (Fig. 1). The linear-array echoendoscope produces a scanning plane in the long axis of the endoscope shaft. This allows real-time visualization of the FNA needle as is introduced into target lesions (Fig. 2).

EUS in Detection and Diagnosis of Pancreatic Malignancy

Several studies have reported that the sensitivity of EUS in the detection of pancreatic tumors is well above 90%90 \% and is clearly superior to abdominal ultrasound, CT scan, and MRI (Table 1). Because early detection is the key to better outcomes in pancreatic cancer patients, the real diagnostic challenge is to detect small tumors. Table 2 summarizes the sensitivities of various imaging techniques in the detection of pancreatic tumors according to tumor size. EUS has consistently been found to have 90%90 \% or higher sensitivity for detecting tumors less than 2 cm in size. The sensitivities of CT, US abdomen, or MRI for finding these lesions
are less than 70%70 \% and less than 50%50 \% in most studies. Even helical CT with its higher resolution is less sensitive than EUS. Thus, EUS is currently the most reliable test for detection of small pancreatic tumors.

As with other imaging modalities, presence of inflammation in pancreas or its surrounding tissue complicates the interpretation of EUS (20,24)(20,24). Significant pancreatic and peripancreatic inflammation can mask the EUS features of pancreatic tumors and make them impossible to find. Focal chronic pancreatitis can also mimic pancreatic cancer. Several EUS criteria have been suggested to differentiate focal chronic pancreatitis from pancreatic cancer, but reliable diagnosis of focal chronic pancreatitis remains challenging. Becker et al. have recently suggested that reliable discrimination between focal pancreatitis and pancreatic cancer may be possible using a newer type of EUS (25). Though these results are encouraging, more studies are needed to assess the its clinical utility and reliability.

EUS is safe and the complication rates for EUS without FNA is the same as for diagnostic endoscopy, approx 0.6%0.6 \% (26). Complete EUS examination of the pancreas can be successfully performed by adequately trained operators in more than 95%95 \% of patients. Inadequate examination may result from 1) inability to pass the echoendoscope, 2) postsurgical anatomy, particularly in patients with history of Billroth II anastamosis, and 3) gastric outlet obstruction as a result of a tumor.

Fig. 2. (A) Linear echoendoscope. (B) FNA needle protruding from linear echoendoscope. © EUS image of pancreatic tumor using linear echoendoscope. (D) FNA of pancreatic mass.

Table 1
Diagnosis of Pancreatic Cancer-Comparison of EUS (Without FNA) With Abdominal US, CT, and MR

EUS-FNA in the Diagnosis of Pancreatic Cancer

Even though EUS provides higher resolution imaging of the pancreas and allows detection of lesions not seen by other imaging techniques, EUS imaging cannot reliably differentiate malignant lesions from inflammatory mass. The specificity of EUS for diagnosis of pancreatic cancer has varied from 73%73 \% to 100%100 \% in different studies (Table 1). Therefore, surgeons may be reluctant to operate when a lesion is seen only on EUS and the patient has less than classical presentation for
pancreatic cancer. In such cases, EUS may actually complicate clinical decision-making. This limitation has prevented full realization of the benefits of early detection of pancreatic cancer by EUS.

Development of EUS-FNA and its refinement over the past few years has vastly improved the utility of EUS in the management of pancreatic cancer (27). EUS-FNA is performed under direct and continuous ultrasound guidance with a 22 -gage needle, which can be directed into pancreatic mass or a lymph node. The needle can extend as far as 50 mm . Doppler may be used during biopsy to identify vascular structures and avoid puncturing them (28). The first

Table 2
Comparison of Sensitivities of EUS with CT, Helical CT, US, and MR in Detection of Small Pancreatic Tumors

H-CT—Helical CT
size-tumor size in mm .

Table 3
EUS/FNA in Diagnosis of Pancreatic Cancer

encouraging paper on this diagnostic modality reported a successful diagnosis of pancreatic cancer with EUS-guided FNA of a 1.6−cm1.6-\mathrm{cm} small lesion in the head of the pancreas (29). A recent study with 322 patients confirmed EUS-FNA to be safe with complication rate of 1.6%1.6 \% with no severe or fatal incidents (26).

The sensitivity of EUS-FNA in diagnosis of pancreatic cancer has varied from 59.5 to 90%90 \%, though a majority of studies have reported sensitivity upwards of 80%80 \% (Table 3). The specificity now approaches 100%100 \% and false-positive cases are extremely rare. Therefore, FNA should be always
be performed when a small potentially resectable pancreatic cancers is identified by EUS. Tumor size is an important prognostic factor in patients with pancreatic cancer (39) and only patients with tumors less than 3 cm in size have a realistic chance of 5 -yr survival after curative surgical resection (1,3,4,39)(1,3,4,39). By reliably identifying small cancers, EUS-FNA offers hope for improving outcomes in patients with pancreatic cancer. EUS-FNA of the pancreatic lesions may also be preferable to CT-guided biopsy (regardless of the size of lesion) because it is less painful and theoretically there is no risk of tumor seeding of the abdominal wall.

Table 4
TNM Staging of Pancreatic Tumors

The sensitivity of the EUS-FNA is bound to increase in the future because of the following factors: 1) improved tissue sampling with refinements in EUS-FNA techniques and with increasing experience of the endosonographers, and 2) improved interpretation of FNA samples by cytologists because of their accumulated experience and their subspecialization into GI cytopathologists, particularly in high-volume centers. The sensitivity of EUS-FNA in diagnosis of pancreatic cancers can be further increased by having a cytopathologist present either in the same room or in close vicinity to determine the adequacy of specimen. Erikson et al. have reported that the presence of cytologist can increase the sensitivity of EUS-FNA by approx 10−15%10-15 \% and decrease the number of needle passes required (40). At MD Anderson Cancer, we have gone a step further and the author (BA) waits for the cytopathologist to examine the specimen for adequacy and make a preliminary diagnosis before terminating the procedure. This increases the duration of the procedure slightly but has a number of advantages. It obviates the need for repeat procedure if the specimen is deemed inadequate by cytopathologist. Additional needle passes can be made to provide additional tissue if the cytologist thinks that the sample is suspicious but not diagnostic for malignancy. Finally, if the specimen is adequate but negative for malignancy, additional needle passes can be made to minimize sampling errors and make a confident determination of benign etiology of the pancreatic lesion.

EUS in Staging of Pancreatic Cancer

TNM classification is the most widely used staging system for pancreatic cancer. It is summarized in Table 4. The most important goal of preoperative staging of pancreatic cancer is to accurately assess the feasibility of performing curative surgical resection. Patients with surgical resection margins positive for malignancy have a very poor prognosis and have no survival benefit from surgery (1,3,4)(1,3,4). These patients can be effectively palliated by nonsurgical means. It is imperative that pre-operative staging be accurate so as to avoid unnecessary surgeries in patients with unresectable tumors and conversely not to deny surgery to patients with potentially resectable tumors.

The role of EUS in the staging of pancreatic and periampullary cancers has been extensively studied and several studies have suggested that EUS is an effective tool for staging of these cancers. The accuracy of EUS in predicting resectability of the pancreatic cancer is summarized in Table 5. Early studies had suggested that the accuracy of EUS in predicting resectability of pancreatic cancer was 90%90 \% or higher. However, Ahmad et al. in a more recent study found that the accuracy of EUS for predicting resectability was a modest 46%46 \% (43). Buscail et al. similarly observed an accuracy of 65%65 \% in their patient group (41). More studies are needed to critically address the wide discordance between published studies about the accuracy of EUS in determining resectability of pancreatic cancer.

Table 5
Accuracy of EUS in Predicting Resectability of Pancreatic Cancer

EUS for Tumor (T) Staging

Assessment of Tumor Size

The prognosis of pancreatic cancer patients is related to the size of tumor and it is therefore important to correctly assess tumor size. Akahoshi et al. evaluated the accuracy of EUS in determination of size of pancreatic tumors (44). They found that EUS correctly measured the size of 90%90 \% of tumors less than 3 cm but only 30%30 \% of the tumors more than 3 cm in size. Overassessment of size by more than 1 cm occurred in only 10%10 \% of tumor less than 3 cm . On the other hand, underassessment of tumor size occurred in 61%61 \% of tumor more than 3 cm . Therefore, size determination of the tumors should be based on both EUS and CT scan findings, and in case of tumors more than 3 cm , clinicians should rely on size measurement from helical CT scan. However, when tumors are small, particularly when they are not visualized on CT scan, EUS provides reliable assessment of tumor size.

Vascular Infiltration

The involvement of portal vein, portal confluence, superior mesenteric vessels, and celiac axis have a major impact on surgical resectability of pancreatic cancers. Until recently, angiography was the standard of care for evaluating vascular involvement and even though its sensitivity was low, it was highly specific and therefore was invaluable in pre-op staging of patients for Whipple resection. EUS has been shown to be more sensitive than angiography in assessing vascular involvement (45,46)(45,46) and because EUS is less invasive, it offers a better alternative for preoperative vascular staging. However, with the advent of helical CT scan
with reconstructive software to generate CT angiograms, the role of EUS in staging in pancreatic tumor is being challenged. CT angiography is less invasive than EUS and allows impressive 3-D reconstruction of the vascular anatomy around the tumor. One of the major practical advantages of CT angiography is that CT angiograms are visually more appealing than EUS pictures and the surgeons can easily comprehend them and appreciate relationship of the tumor to surrounding vascular structures.

The sensitivity and specificity of commonly used EUS criteria for determination of vascular involvement by pancreatic tumors are summarized in Table 6. In the earlier studies, Yasuda et al. (47) and Rosch et al. (45) found a high sensitivity and specificity of EUS in determining the vascular infiltration. Subsequently Snady et al. (48) and Brugge et al. (49) reported that the specificity of EUS was close to 100%100 \% but the sensitivity was only 19−50%19-50 \%. In a more recent study involving reevaluation of videotapes of EUS examinations performed over several years, Rosch et al. (50) found that when the EUS criteria were more rigorously applied, sensitivity of EUS criteria was close to what had been reported by Snady and Brugge.

How does tumor staging by EUS staging compare with helical CT? and Can EUS findings complement CT assessment of vascular infiltration? The endosonographers directly assess tumor extension into the vessel wall by searching for loss of hyperechoic interface between tumor (Fig. 3A) and vessel or presence of irregularity of vessel wall at the point of contact with tumor (Fig. 3B). These criteria are highly specific but insufficiently sensitive for detecting vascular infiltration (Table 5). On the other hand, the radiologists indirectly “predict” infiltration of

Fig. 3. EUS features of vascular infiltration by pancreatic tumor. (A) Preservation of hyperechoic interface between the tumor and the portal vein (marked by arrows). There is also an enlarged lymph node next to the tumor. (B) Loss of interface between tumor and portal vein and tumor extending in lumen of portal vein.

Table 6
Reliability of EUS Criteria in Detecting Vascular Infiltration

the vessels based on the percentage of circumference of vessel wall in contact the tumor (graded 0−40-4 ). It seems reasonable that in patients with borderline CT staging (CT grade 2-3), presence of EUS evidence of vascular invasion would be very useful information. Midwinter et al. reported that EUS is comparable to helical CT in determi-
nation of venous involvement (portal vein and superior mesenteric vein) but is significantly inferior in determining the involvement of superior mesenteric artery (51). In institutions where surgeons consider arterial encasement but not venous involvement as a contra-indication to curative resections, EUS offers no advantage over helical

Table 7
Accuracy of EUS in Tumor Staging According to T Stage

Fig. 4. Paraduodenal collaterals in patient with 3 cm tumor in the head of pancreas. Paragastric and paraduodenal collaterals are highly specific for portal vein or splenic vein involvement in patients with pancreatic cancer.

CT alone in determining resectability of tumors. However, when portal venous infiltration is used as a contraindication for curative surgical resection, EUS staging can complement staging with helical CT.

So, when should EUS be used to stage pancreatic cancer? In the case of patients with small pancreatic tumor visualized only by EUS, the choice is obvious. The accuracy of EUS is between 90-100% in case of TI tumors but is lower and variable for T2
and T3 tumors (Table 7). EUS staging is reliable for small tumors that are staged T1 by EUS and have clear evidence of no vascular involvement. EUS is also reliable in patients with larger tumors with definite evidence of portal vein involvement based on presence of vascular collaterals (Fig. 4), occlusion of vascular lumen with tumor and loss of interface between the tumor and the vessel.

Nodal Staging

The presence of lymph node metastases has significant inverse correlation with long-survival in pancreatic cancer patients (1,3,4,52,53)(1,3,4,52,53). After curative surgical resection, patients with negative lymph nodes had an estimated 5-yr survival of 35%35 \% and an actual survival of 20%20 \%, whereas patients with lymph nodes metastases had an estimated 5-yr survival of 6%6 \% and actual 5-yr survival of 2%2 \% (52). In the same study, the median survival was 11.5 mo in patients with one or more lymph nodes positive for malignancy and 24 mo in patients with no nodal metastases. Sperti et al. (4) reported 1,3 , and 5 survival of 68%,5%68 \%, 5 \%, and 0%0 \%, respectively, in patients with positive lymph nodes and 90%,35%90 \%, 35 \%, and 19%19 \%, respectively, with uninvolved lymph nodes. Yeo et al. (1) and Nitecki et al. (3) observed similar trends in survival curves in lymph node positive vs lymph node negative group of patients with pancreatic cancer. Though, currently there is no consensus on this issue, there is a case to be made for not performing a Whipple procedure in patients with lymph node metastasis,

Table 8
Accuracy of Nodal Staging of Pancreatic Cancer

Table 9
EUS-FNA Operating Characteristics for Evaluation of Lymph Nodes (35)

particularly to celiac and mediastinal lymph nodes. These patients may be considered for systemic therapy regimens such as neo-adjuvant chemo-radiation in combination with surgery or be candidates for nonsurgical palliative treatment. Accurate and reliable nodal staging of pancreatic adenocarcinomas is therefore crucial.

By EUS imaging, malignant involvement of lymph nodes is suspected if they are hypoechoic, round in shape (vs ovoid or oblong) with sharp margins and are more than 10 mm in size. However, size and morphologic criteria for determining the malignant infiltration of lymph nodes are not accurate. The accuracy of nodal staging by EUS (without FNA), CT scans, or helical CT in published studies is summarized in Table 8. Due to lack of specificity of size or morphologic criteria for enlarged celiac and mediastinal lymph nodes, patients with pancreatic cancer with enlarged lymph nodes visualized by various imaging techniques, often undergo laparoscopy or laparotomy for sampling of the enlarged lymph nodes prior to resection of primary tumor. EUS-FNA can reliably diagnose malignant infiltration of lymph nodes and has the potential to markedly improve the accuracy
of lymph node staging of pancreatic cancer (Table 9 and Fig. 5).

Assessment of Metastasis

The field of examination with EUS is limited and therefore EUS is not reliable in the assessment of metastatic spread of tumor. However, a limited examination of the liver can still be performed with EUS and sometimes liver metastases missed by abdominal US or CT scan are detected by EUS and can be confirmed histologically by EUS-FNA. EUS supplements the M staging performed by CT scan or MRI but is inadequate and inappropriate for determining the metastatic spread of pancreatic tumors.

Conclusion

The overall survival of patients with pancreatic cancer has not changed much in past several decades. This is largely because of difficulty in reliably detecting small pancreatic tumors as small pancreatic tumors are asymptomatic and frequently are missed by conventional imaging techniques including CT scan, abdominal US, and MRI. EUS can detect small

Fig. 5. Enlarged celiac lymph node in a patient with pancreatic cancer. (A) Lymph node visualized on radial scanning. (B) and © FNA of celiac lymph node using linear echoendoscope.
pancreatic tumors missed by other modalities and EUS-FNA permits histological confirmation of malignancy in these lesions. EUS and EUS-FNA
may also improve tumor staging when combined appropriately with helical CT.

References

1 Yeo CJ, Cameron JL, Lillemoe KD, et al. Pancreaticoduodenectomy for cancer of the head of the pancreas. Ann Surg 1996;221:721-733.
2 Warshaw AL, Fernandez-Del Castillo C. Pancreatic carcinoma. N Engl J Med 1992;326:455-465.
3 Nitecki SS, Sarr MG, Colby TV, Heerden JAV. Long term survival after resection for ductal adenocarcinoma of the pancreas. Ann Surg 1995;221:59-66.
4 Sperti C, Pasquali C, Piccoli A, Pedrazzoli S. Survival after resection for ductal adenocarcinoma of pancreas. Br J Surg 1996;83:625-631.
5 Homma T, Tsuchiya R. The study of the mass screening of persons without symptoms and of the screening of outpatients with gastrointestinal complaints or icterus for pancreatic cancer in Japan, using CAI 9-9 and elastase-1 or ultrasonography. Int J Pancreatol 1991;9:119-124.
6 Adamek HE, Albert J, Breer H, et al. Pancreatic cancer detection with magnetic resonance cholangiopancreatography and endoscopic retrograde cholangiopancreatography: a prospective controlled study. Lancet 2000;356: 190-193.
7 Glasbrenner B, Schwarz M, Pauls S, et al. Prospective comparison of endoscopic ultrasound and endoscopic retrograde cholangiopancreatography in the preoperative assessment of masses in the pancreatic head. Dig Surg 2000;17:468-474.
8 Menges M, Lerch MM, Zeitz M. The double duct sign in patients with malignant and benign pancreatic lesions. Gastrointest Endosc 2000;52:74-77.
9 Sener SF, Fremgen A, Menck HR, Winchester DP. Pancreatic cancer. A report of treatment and survival trends for 100313 patients diagnosed from 1985 to 1995, using the national cancer database. J Am Coll Surg 1999;189: 1−71-7.
10 Baumel H, Huguier M, Manderscheid JC, et al. Results of resection for cancer of the exocrine pancreas: a study from the French Association of Surgery. Br J Surg 1994;81: 102-107.
11 Wade TP, Choplin MA, Virgo KS, Johnson FE. Periampullary cancer treatment in US department of Veteran Affairs Hospitals: 1987-1991. Surgery 1994;116:819-826.
12 Takahashi S, Ogata Y, Tsuzuki T. Combined resection of the pancreas and portal vein for pancreatic cancer. BrJ\mathrm{Br} J Surg 1994;81:1190-1193.
13 Sewnath ME, Birjmohun RS, Rauws EA, et al. The effect of pre-operative biliary drainage on post-operative complication after pancreaticoduodenectomy. J Am Coll Surgeons 2001;192:726-734.
14 Povoski SP, Karpeh MD, Jr, Conlon KC, et al. Preoperative biliary drainage: impact on intra-operative bile cultures and infectious morbidity and mortality after pancreaticoduodenectomy. J Gastrointest Surg 1999;3: 496-505.

15 Nishiharu T, Yamashita Y, Abe Y, et al. Local extension of pancreatic carcinoma: assessment with thin section helical CT versus with breath hold fast MR-imaging-ROC analysis. Radiology 1999;212:445-452.
16 Zeman RK, Cooper C, Zeiberg AS, et al. TNM staging of pancreatic carcinoma using helical CT. Am J Roentgenol 1997;169:459-464.
17 Megibow AJ, Zhou XH, Rotterdam H, et al. Pancreatic adenocarcinoma: CT versus MR imaging in the evaluation of resectability—report of diagnostic oncology group. Radiology 1995;195:327-332.
18 Diehl SJ, Lehmann KJ, Sadick M, et al. Pancreatic cancer: value of dual-phase helical CT in assessing resectability. Radiology 1998;206:373-378.
19 Raptopoulos V, Steer ML, Sheiman RG, et al. The use of helical CT and CT angiography to predict vascular involvement from pancreatic cancer: correlation with findings at surgery. Am J Roentgenol 1997;168:971-977.
20 Palazzo L, Roseau G, Gayet B, et al. Endoscopic ultrasonography in the diagnosis and staging of pancreatic adenocarcinoma. Results of a prospective study with comparison to ultrasonography and CT scan. Endoscopy 1993;25:143-150.
21 Muller MF, Meyenberger C, Bertschinger P, et al. Pancreatic tumors: evaluation with endoscopic US, CT and MR imaging. Radiology 1994;190:745-751.
22 Nakaizumi A, Uehara H, Iishi H, et al. Endoscopic ultrasonography in diagnosis and staging of pancreatic cancer. Dig Dis Sci 1995;40:696-700.
23 Legmann P, Vignaux O, Dousset B, et al. Pancreatic tumors: comparison of dual phase helical CT and endoscopic sonography. Am J Roentgenol 1998;170:1315-1322.
24 Rosch T, Lorenz R, Braig C, et al. Endoscopic ultrasound in pancreatic tumor diagnosis. Gastrointest Endosc 1991; 37:347-352.
25 Becker D, Strobel D, Bernatik T, Hahn EG. Echo-enhanced color-and power-Doppler EUS for discrimination between focal pancreatitis and pancreatic carcinoma. Gastrointest Endosc 2001;53:811-813.
26 O’Toole D, Palazzo L, Arotcarena R, et al. Assessment of complications of EUS-guided fine-needle aspiration. Gastrointest Endosc 2001;53:470-474.
27 Chang KJ, Nguyen P, Erikson RA, et al. The clinical utility of endoscopic ultrasound-guided FNA in the diagnosis and staging of pancreatic carcinoma. Gastrointest Endosc 1997;45:387-393.
28 Wiersema MJ, Vilmann P, Giovannini M, et al. Endosonog-raphy-guided fine-needle aspiration biopsy: diagnostic accuracy and complication assessment. Gastroenterology 1997;112:1087-1095.
29 Chang KJ, Albers CG, Erickson RA, et al. Endoscopic ultrasound-guided fine needle aspiration of pancreatic carcinoma. Am J Gastroenterol 1994;89:263-266.
30 Giovannini M, Seitz JF, Monges G, et al. Fine needle aspiration cytology guided by endoscopic ultrasonography: results in 141 patients. Endoscopy 1995;27:171-177.
31 Gress F, Hawes R, Ikenberry S, et al. A prospective evaluation of EUS-guided fine needle aspiration biopsy for
diagnosing pancreatic masses. Gastrointest Endosc 1995: AB666, 1995.
32 Carroll N, Quirk D, Centeno B, et al. Accuracy of EUSguided pancreatic biopsies. Gastrointest Endosc 1997;45: AB580.
33 Wiersema MJ, Vilmann P, Giovannini M, et al. Endosono-raphy-guided fine-needle aspiration biopsy: diagnostic accuracy and complication assessment. Gastroenterology 1997;112:1087-1095.
34 Faigel DO, Ginsberg GG, Bentz JS, et al. Endoscopic ultra-sound-guided real-time fine-needle aspiration biopsy of the pancreas in cancer patients with pancreatic lesions. J Clin Oncol 1997;15:1439-1443.
35 Williams DB, Sahai, AV, Aabakken L, et al. Endoscopic ultrasound guided fine needle aspiration biopsy: a large single center experience. Gut 1999;44:720-726.
36 Fritscher-Ravens A, Izbicki JR, Sriram PVJ, et al. Endosongraphy-guided fine needle aspiration cytology extending the indication of organ preserving pancreatic surgery. Am J Gastroenterol 2000;95:2255-2260.
37 Fritscher-Ravens A, Sriram PVJ, Krause C, et al. Detection of pancreatic metastases by EUS-guided fine needle aspiration. Gastrointest Endosc 2001;53:65-70.
38 Brandwein SL, Farrell JJ, Centeno BA, Brugge WR. Detection and tumor staging of malignancy in cystic, intraductal, and solid tumors of the pancreas by EUS. Gastrointest Endosc 2001;53:722-727.
39 Fortner JG, Klimstra DS, Senie RT, Maclean BJ. Tumor size is the primary prognosticator for pancreatic cancer after regional pancreatectomy. Ann Surg 1996;223: 147−153147-153.
40 Erikson RA, Sayage-Rabie L, Beissner RS. Factors predicting the number of EUS guided fine needle passes for the diagnosis of pancreatic malignancies. Gastrointest Endosc 2000;51:184-190.
41 Buscail L, Pages P, Berthelemy P, et al. Role of EUS in the management of pancreatic and ampullary carcinoma: a prospective study assessing resectability and prognosis. Gastrointest Endosc 1999; 50:34-40.
42 Gress FG, Hawes RH, Savides TJ, et al. Role of EUS in the preoperative staging of pancreatic cancer: a large single-center experience. Gastrointest Endosc 1999; 50:786-791.
43 Ahmad NA, Lewis JD, Ginsberg GG, et al. EUS in preoperative staging of pancreatic cancer. Gastrointest Endosc 2000;52:463-468.
44 Akahoshi K, Chijiiwa Y, Nakano I, et al. Diagnosis and staging of pancreatic cancer by endoscopic ultrasound. Br J Radiol 1998;71:492-496.
45 Rosch T, Braig C, Gain T, et al. Staging of pancreatic and ampullary carcinoma by endoscopic ultrasonography. Comparison with conventional sonography, computed tomography and angiography. Gastroenterology 1992;102: 188-199.
46 Ahmad NA, Kochman ML, Lewis JD, et al. Endosonography is superior to angiography in the pre-operative assessment of vascular involvement among patients with pancreatic carcinoma. J Clin Gastroenterol 2001;32: 54-58.

47 Yasuda K, Mukai H, Nakajima M. Staging of pancreatic carcinoma by endoscopic ultrasonography. Endoscopy 1993;25:151-155.
48 Snady H, Bruckner H, Seigel J, et al. Endoscopic ultrasonographic criteria of vascular invasion by potentially resectable pancreatic tumors. Gastrointest Endosc 1994; 40:326-333.
49 Brugge WR, Lee MJ, Kelsey PB, et al. The use of EUS to diagnose malignant portal venous invasion by pancreatic cancer. Gastrointest Endosc 1996;43:561-567.
50 Rosch T, Dittler HJ, Strobel K, et al. Endoscopic ultrasound criteria for vascular invasion in the staging of cancer in the head of the pancreas; a blind re-evaluation of videotapes. Gastrointest Endosc 2000;52:469-477.

51 Midwinter MJ, Beveridge CJ, Wilsdon JB, et al. Correlation between spiral computed tomography, endoscopic ultrasonography and findings at operation in pancreatic and ampullary tumors. Br J Surg 1999,86: 189-193.
52 Kawarada Y, Das BC, Naganuma T, Isaji S. Surgical treatment of pancreatic cancer. Does extended lymphadenectomy provide a better outcome? J Hepatobiliary Pancreat Surg 2001;8:224-229.
53 Delcore R, Rodriguez FJ, Forster J, et al. Significance of lymph node metastases in patients with pancreatic cancer undergoing curative surgical resection. Am J Surg 1996; 172:463-469.