Dual-Phase Helical CT of Pancreatic Adenocarcinoma: Assessment of Resectability Before Surgery

Transcription

1 Carlos Valls 1 Eduard Andía 1 Anna Sanchez 1 Juan Fabregat 2 Oscar Pozuelo 1 Juan Carlos Quintero 1 Teresa Serrano 3 Francisco Garcia-Borobia 2 Rosa Jorba 2 Received July 10, 2001; accepted after revision October 18, Institut de Diagnòstic per la Imatge, Hospital Duran i Reynals, Ciutat Sanitària i Universitària de Bellvitge, Autovia de Castelldefels km 2, 7, L'Hospitalet de Llobregat, Barcelona, Spain. Address correspondence to C. Valls. 2 Department of Surgery, Hospital Princeps d'espanya Ciutat Sanitaria i Universitaria de Bellvitge, C/Feixa Llarga s/n, L Hospitalet de Llobregat, Barcelona, Spain. 3 Department of Pathology, Hospital Princeps d'espanya, Barcelona, Spain. AJR 2002;178: X/02/ American Roentgen Ray Society Dual-Phase Helical CT of Pancreatic Adenocarcinoma: Assessment of Resectability Before Surgery OBJECTIVE. The aim of our study was to prospectively evaluate the accuracy of dualphase helical CT in the preoperative assessment of resectabiliy in patients with suspected pancreatic cancer using surgical and histopathologic correlation. SUBJECTS AND METHODS. Between January 1999 and December 2000, 76 patients with suspected pancreatic cancer underwent preoperative evaluation and staging with dualphase helical CT (3-mm collimation for pancreatic phase, 5-mm collimation for portal phase). Iodinated contrast material was injected IV (170 ml at a rate of 4 ml/sec); acquisition began at 40 sec during the pancreatic phase and at 70 sec during the portal phase. Three radiologists prospectively evaluated the imaging findings to determine the presence of pancreatic tumor and signs of unresectability (liver metastasis, vascular encasement, or regional lymph nodes metastasis). The degree of tumor vessel contiguity was recorded for each patient (no contiguity with tumor, contiguity of <50%, or contiguity of 50%). RESULTS. Thirty-nine patients with pancreatic adenocarcinoma were surgically explored. Curative resections were attempted in 34 patients and were successful in 25. The positive predictive value for resectability was 73.5%. Nine patients considered resectable on the basis of CT findings were found to be unresectable at surgery because of liver metastasis (n = 5), vascular encasement (n = 2), or lymph node metastasis (n = 2). We found that the overall accuracy of helical CT as a tool for determining whether a pancreatic adenocarcinoma was resectable was 77% (30/39 patients). CONCLUSION. Dual-phase helical CT is a useful technique for preoperative staging of pancreatic cancer. The main limitation of CT is that it may not reveal small hepatic metastases. P ancreatic adenocarcinoma is one of the leading causes of cancer death in the Western world, accounting for 22% of the deaths due to malignant gastrointestinal neoplasms [1]. Despite recent advances in diagnosis and treatment, the prognosis for patients with pancreatic cancer remains extremely poor, with an overall 5-year survival rate of only 3% [2]. The poor outcome of patients with pancreatic cancer has been related to the lack of clinically relevant signs before the disease reaches an advanced stage and the relatively high rates of morbidity and mortality associated with pancreatic resection. Still, surgery is the only curative option in patients with pancreatic adenocarcinoma. Recently, increased surgical experience with this type of operation has led to a dramatic decline in mortality rates for patients with pancreatic cancer. For patients in the hands of experienced surgeons, the mortality rate for pancreaticoduodenectomy is less than 5% [1 3], although postoperative morbidity approaches 50% [3]. In addition, the advent of new venous-resection techniques has expanded the indications for potentially radical surgery in pancreatic cancer. However, in clinical practice, only 5 22% of patients are amenable to surgical resection because of extensive disease at initial evaluation or at surgical exploration [3]. The main role of preoperative staging procedures is to distinguish between potentially resectable and nonresectable patients so that unnecessary surgical procedures may be avoided; survival benefits can be achieved only in patients in whom the tumor can be completely resected. Because of recent improvements in radiologic techniques, a wide range of imaging tools is now available, such as helical CT, MR imaging, endoscopic sonography, endoscopic retrograde cholangiopancreatography, and angiography. However, at this point, no AJR:178, April

2 Valls et al. consensus exists as to the best staging algorithm. The aim of our study was to assess the results of helical CT used as the only procedure in the diagnosis and staging of pancreatic cancer in a large series of patients from a single institution. We sought to avoid invasive procedures such as endoscopic retrograde cholangiopancreatography. Our objective was to evaluate the results of helical CT in preoperative staging, focusing on either arterial or venous vascular invasion. We specifically assessed the relevance of small reticular opacities arising from the tumor and abutting an arterial vessel that we defined as periarterial streaks. Subjects and Methods Patients Between January 1999 and December 2000, preoperative evaluation and staging with dual-phase helical CT were performed prospectively in 76 consecutive patients with clinical, biologic, and sonographic findings that raised suspicions of pancreatic cancer. Our institutional review board approved our study, and informed consent was obtained from all patients. To be included in the study, patients had to have painless obstructive jaundice with sonographic evidence of intra- and extrahepatic bile duct dilatation or sonographically based suspicion of pancreatic mass. In our institution, the imaging algorithm of painless obstructive jaundice begins with sonography. If the sonographic examination discloses intra- and extrahepatic bile duct dilatation, we proceed directly to dual-phase pancreatic helical CT. If the dilatation involves only the intrahepatic biliary tree (with a normal extrahepatic bile duct), we perform dualphase helical CT of the liver and MR cholangiopancreatography. For our study, we specifically refrained from using endoscopic retrograde cholangiopancreatography or biliary drainage before CT to avoid artifacts in the diagnostic study and complications related to infections. Twenty-one patients were not considered candidates for surgical exploration because of poor clinical condition (n = 2) or advanced metastatic disease at CT (n = 19). Fine-needle aspiration biopsy was performed in all patients who did not undergo surgical exploration. Biopsy was not performed in candidates for surgical exploration. Eventually, 55 patients underwent surgical exploration within 1 week of their CT examination. For analysis of preoperative staging, only patients with proven pancreatic adenocarcinoma who underwent surgical exploration were considered. Patients with periampular adenocarcinoma, islet cell cancer, distal bile duct cholangiocarcinoma, or other pancreatic conditions were excluded. Therefore, our final study group consisted of 39 patients with pancreatic adenocarcinoma who underwent surgical exploration. There were 24 men and 15 women with a mean age of 61.1 years (range, years). Helical CT Technique Helical CT was performed with a ProSpeed Plus system (General Electric Medical Systems, Yokogawa, Japan). Unenhanced images were obtained first to locate the upper and lower limits of the pancreas so that the entire gland would be included in the contrastenhanced study. Either ionic (370 mg I/mL) or nonionic (320 mg I/mL) iodinated contrast material (170 ml) was injected at a rate of 4 ml/sec using a power injector (MCT; Medrad, Pittsburgh, PA). Acquisition began at 40 sec for the pancreatic phase and 70 sec for the portal phase. In the first helical acquisition, images of the pancreas were obtained with a 3-mm collimation and a 5-mm interval from the lower to the upper levels of the pancreas. Images were reconstructed at 3-mm intervals. In the second helical acquisition, images of the liver and pancreas were acquired from the dome of the liver to the lower level of the pancreatic gland with a 5-mm collimation and a 7.5-mm interval. Images were reconstructed at a 5-mm interval. Because of tube-heating limitations, we used mas and 120 kv for thin-section pancreatic images and 250 mas for the rest of the study. In patients with suspected hemangiomas or liver cysts observed during the portal phase imaging, delayed scans were also obtained. Before the patients underwent surgical exploration, the CT images were prospectively assessed by at least two of three experienced abdominal radiologists in consensus. The pancreatic and portal phase images were interpreted at the same time. The imaging findings and preoperative CT diagnosis were recorded in an electronic database. Preoperative Staging For the staging procedure, we specifically evaluated signs suggesting or indicating unresectable disease, such as vascular invasion and distant metastasis. Vascular invasion. The degree of tumor-tovessel circumferential contiguity was recorded for Fig. 1. Arterial invasion in 65-year-old woman with obstructive jaundice. Helical CT scan shows large hypodense mass (arrow) in pancreatic head, abutting and encasing superior mesenteric artery. This sign is clear indication of unresectability, and patient was excluded from surgery. each patient (no contiguity with tumor, contiguity of <50%, or contiguity of 50%) for the following vessels: celiac trunk, hepatic artery, superior mesenteric artery, superior mesenteric vein, and portal vein. Tumor-to-vessel contiguity was defined as direct tumor-to-vessel contact with complete obliteration of the fat plane. Arterial invasion was defined as any direct tumor-to-vessel contiguity even if the contiguity was less than 50% (Fig. 1). Small reticular opacities radiating from the tumor and abutting an arterial vessel (so-called periarterial streaks) were not considered arterial invasion. Venous invasion was defined as tumor-to-vessel circumferential contiguity of 50% or more. Tumor-to-vein circumferential contiguity of less than 50% was not considered venous invasion (Fig. 2). Distant metastasis. Liver metastases were defined as nodular low-attenuation lesions with diameters of more than 1 cm without characteristic findings of benign lesions, such as cysts or hemangiomas. Subcentimetric lesions were considered indeterminate and were not treated as metastases. Enlarged lymph nodes (>1.5 cm) outside peripancreatic draining chains of the pancreas were considered metastatic. Substantial ascites and a large number of peritoneal nodules were considered evidence of peritoneal dissemination. Criteria of unresectability. A tumor was considered unresectable if one of the following features was found: distant metastasis either to the liver, regional lymph nodes, omentum, or adjacent organs except the duodenum; arterial invasion; any tumor-to-artery contiguity; or massive venous invasion, that is, tumor infiltration with thrombosis and obliteration of the vessel lumen. Venous invasion without thrombosis or obliteration of the lumen was not considered indicative of unresectable disease, and the surgeon was prepared for venous resection. All patients with questionable findings, such as periarterial streaks, were surgically explored. Fig. 2. Helical CT scan obtained during pancreatic phase in 52-year-old man reveals small hypoattenuating mass (arrow) in pancreatic head. Tumor abuts superior mesenteric vein (arrowhead), but tumor-to-vessel circumferential contiguity is less than 50%, indicating absence of venous invasion. Surgical exploration confirmed absence of venous invasion, and tumor was resected without venous resection. 822 AJR:178, April 2002

3 Helical CT of Pancreatic Adenocarcinoma Results Histologic studies were performed for all 76 patients either at surgery (n = 55) or by percutaneous fine-needle aspiration biopsy (n = 21). Final histologic diagnoses were pancreatic adenocarcinoma (n = 60), carcinoma of the pancreatic ampula (n = 8), distal bile duct cholangiocarcinoma (n = 3), islet cell tumor (n = 3), and pancreatic metastasis (n = 2). Surgical procedures were performed in 39 patients. In 34 patients, radical surgery was planned on the basis of preoperative imaging, whereas gastroenteroanastomosis and palliative surgery were planned in five patients with unresectable disease and duodenal invasion. Curative resections were successful in 25 of 34 instances, a resectability rate of 73.5%. The positive predictive value of helical CT in regard to resectability was also 73.5% (25/34). Nine patients thought to have resectable disease on the basis of CT findings were found to have unresectable disease at surgery. In that subgroup of patients, the reasons for unresectability were liver metastases (n = 5), vascular encasement (n = 2), and lymph nodes metastases (n = 2). The overall accuracy of helical CT in revealing resectability or unresectability of pancreatic adenocarcinoma was 76.9% (30/39). Liver Metastases Liver metastases were found at surgical exploration in eight (20.5%) of 39 patients with pancreatic adenocarcinoma in whom surgery was performed. The reason for unresectability in 55.5% (5/9) of patients believed to be resectable on the basis of CT findings but found to be unresectable at surgery was unsuspected liver metastasis. The average size of the lesions was 8 mm. In two of these patients, retrospective evaluation of helical CT images disclosed two tiny liver lesions. It is interesting that one Fig year-old woman with unresectable pancreatic cancer understaged using helical CT. Note small hepatic cyst (arrowhead) in both pancreatic and portal phases. A, Pancreatic-phase CT scan shows small hypervascular hepatic lesion (arrow) in segment III that was missed at prospective evaluation. B, During portal phase, CT scan obtained at same level as A shows lesion is isoattenuating and no longer visible. of these lesions was hypervascular in the pancreatic phase (Fig. 3). Lymph Node Metastases Eighteen (52.9%) of 34 surgically explored patients had tumoral lymph node involvement confirmed at surgery or pathologic examination. Helical CT accurately identified metastatic lymph node involvement in 16.7% (3/18) of the patients. These patients had lymph nodes larger than 1.5 cm and were resectable because the metastatic lymph nodes affected peripancreatic nodes and were resected en bloc with the primary tumor. In 39.8% (7/18) of the patients with positive lymph nodes at histopathologic examination, helical CT disclosed nonspecific lymph nodes with diameters of less than 1.5 cm. In the remaining 44.4% (8/18) of these patients, helical CT did not reveal any lymph node involvement. Only 5.9% (2/34) of potentially resectable patients were found to be unresectable because of undetected lymph node metastasis. One of these patients had a left paraaortic lymph node of 1.7 cm that was missed prospectively. In the other case, lesions were not detectable in a retrospective evaluation. Vascular Invasion Superior mesenteric vein. Superior mesenteric vein invasion (tumor-to-vessel contiguity, 50%) was observed in six patients. Palliative surgery was performed in three patients who had duodenal obstruction and other CT findings indicating unresectability (superior mesenteric vein invasion or liver metastasis), and therefore, thorough inspection of the superior mesenteric vein was not possible. In another patient, radical surgery was planned, but surgical inspection disclosed one solitary liver metastasis. Therefore, palliative surgery was performed, and the superior mesenteric A vein was not explored. Two patients with more than 50% tumoral vascular contact were successfully resected but required venous resection in one case and venous resection with graft placement in the other case (Fig. 4). Thirty-three patients had no sign of superior mesenteric vein invasion (tumor-to-vessel contiguity, <50%). Eight patients had tumor-tovein contiguity of less than 50%, suggesting no such invasion. Six of these patients were resected but did not undergo venous resection, and two patients were not evaluated because of palliative surgery. Twenty-five patients showed no tumor-to-vessel contact. In 23 of the patients, tumors could be resected. In two patients, the superior mesenteric vein was not evaluated because of palliative surgery Portal vein. Portal vein invasion (tumorto-vessel contact, 50%) was detected in three patients. Two patients underwent palliative surgery, and venous invasion was not evaluated. In the other patient, radical surgery was planned, but surgical inspection disclosed regional lymph node metastasis. Therefore, palliative surgery was performed, and the portal vein was not explored. In 32 patients, no tumor-to-vessel contact was observed. Of these, 25 patients underwent resection. In seven patients, palliative surgery was performed, and the portal vein was not evaluated. In three patients, helical CT did not reveal vascular invasion (tumor-to-to-vessel contact, <50%), but vascular invasion was found at surgery. Patients were successfully resected with venous resection (Fig. 5). Venous invasion as a whole (superior mesenteric vein and portal vein) was identified at surgery in 12 (30.8%) of the 39 patients. Helical CT had a sensitivity of 75% (9/12) in revealing venous invasion. Three patients (25%) with false-negative findings of less than 50% tumor-to-vein contiguity were found to have B AJR:178, April

4 Valls et al. Fig year-old man who had pancreatic adenocarcinoma with venous invasion. Axial helical CT image obtained at level of uncinate process shows large hypodense mass (arrow) with circumferential involvement of superior mesenteric vein. Note that tumor-to-vein contiguity (arrowheads) is greater than 50%, indicating venous invasion. Tumor was successfully resected after en bloc venous resection and terminoterminal anastomosis. Fig. 5. Axial CT scan in 69-year-old woman shows low-density tumor (arrow) in pancreatic head. Tumor-to-vein contiguity is less than 50%, suggesting absence of venous invasion. However at surgical exploration, venous invasion was found, and tumor was resected with partial vein resection. venous invasion at surgical exploration. The negative predictive value of helical CT to exclude venous invasion was 90%. One additional patient had venous resection because of surgically based suspicion of venous invasion, but microscopic examination showed no tumor invasion. Periarterial streaks. In eight patients, we detected reticular opacities or streaks radiating from the tumor and abutting the superior mesenteric artery without definite tumor-to-vessel contact. All these patients were surgically explored. Six of the patients were successfully resected (Fig. 6), and two patients were unre- Fig. 6. Periarterial streaks in resectable pancreatic adenocarcinoma in 61-year-old woman with obstructive jaundice. Transverse contrast-enhanced CT scan reveals low-density lesion (arrow) in pancreatic head. Note periarterial streaks (arrowhead) arising from tumor and in contact with lateral aspect of superior mesenteric artery. 824 sectable (Fig. 7). The positive predictive value of periarterial streaks for unresectability was 25% (2/8). Discussion The indications for radical pancreaticoduodenectomy for pancreatic cancer have expanded over recent years as a result of a dramatic decline in surgical mortality rates [3]. However, according to the findings of a recent surgical series, less than 20% of pancreatic cancers are amenable to surgical resection [4]. The reasons for unresectability include unsuspected Fig. 7. Periarterial streaks in unresectable pancreatic adenocarcinoma in 66-year-old man with obstructive jaundice. Helical CT scan shows hypodense lesion (arrow) in uncinate process and linear reticular opacities arising from tumor and abutting posterior aspect of superior mesenteric artery (arrowhead). At surgical exploration, invasion of mesenteric artery was found. Resection was not possible, leaving grossly positive margin. liver metastasis, vascular invasion, lymph node metastasis, and peritoneal carcinomatosis. Therefore, the main role of preoperative imaging is to distinguish between potentially resectable and clearly unresectable patients so that the optimal treatment for each patient can be ascertained. Various imaging procedures have been proposed for use in preoperative staging of pancreatic cancer, although no consensus has been reached concerning the best imaging algorithm. Contrast-enhanced helical CT has been extensively used as a tool in both diagnosing and staging pancreatic adenocarcinoma [5 8], mainly focusing on metastatic spread to the liver and local invasion of vascular structures. The reported sensitivity of helical CT in revealing pancreatic carcinoma is high, ranging between 89% and 97% [4, 6, 8 10]. Metastatic liver disease is a well-accepted criterion of unresectability for pancreatic adenocarcinoma [11]. Detection of tiny metastatic liver lesions remains the most difficult part of preoperative staging of pancreatic cancer. In our series, 55% (5/9) of the false-negative findings for resectability were due to undetected liver metastasis. In addition, all patients with solid hepatic lesions larger than 1 cm were confirmed as having metastases. However, all patients excluded from surgery because of liver metastasis had several hepatic lesions, some of them larger than 2 cm. In one (20%) of the five patients with liver metastasis, a hypervascular lesion in the pancreatic phase was seen retrospectively to correspond to a hypervascular metastasis (Fig. 3). However, this lesion would be difficult to differentiate from a small hemangioma. In addition, we missed two other tiny metastatic lesions found in this patient. In another patient, a tiny 1-cm hypodense lesion was detected in retrospect. In the other three patients, additional lesions were not found after careful retrospective evaluations. In each of these three patients, only one tiny metastatic lesion was found at surgical exploration. Therefore, metastatic pancreatic cancer might present as an isolated small solid lesion. In that clinical setting, if no other ancillary findings of unresectability are found, it would be necessary to perform diagnostic laparoscopy before excluding the patient from surgery. In our series, preoperative nodal staging was clearly inaccurate. Only 16.6% of the patients with lymph node metastasis had the metastasis detected prospectively. This low detection rate is similar to those reported in previous studies [6, 9] and reflects the fact that AJR:178, April 2002

5 Helical CT of Pancreatic Adenocarcinoma size is not a discriminant feature between metastatic and nonmetastatic lymph nodes because a lymph node of normal size might be infiltrated by tumor. However, in clinical practice, the low detection rate of metastatic lymph nodes has a limited importance because most metastatic lymph nodes are resected at surgery. Arterial invasion is another well-accepted criterion of unresectability [11]. However, peripancreatic venous invasion is a controversial contraindication for pancreaticoduodenectomy. Some surgeons believe that surgery is precluded in cases of tumor invasion of the portal vein or superior mesenteric vein, whereas other surgeons will perform aggressive surgery with mesenteric vein resection [4]. Recent surgical experience in resecting pancreatic adenocarcinoma with venous invasion has not shown histologic features suggesting a worse prognosis [12]. Therefore, assessment of major venous structures is a key factor in preoperative staging of pancreatic cancer and is critical for deciding surgical procedures. In fact, thorough assessment of mesenteric vessels during surgical exploration is extremely difficult because direct inspection of the lateral and posterior wall of mesenteric vessels can only be performed after gastric and pancreatic transections have been performed. At that point, resection is inevitable either with negative margins or by cutting through the tumor [13]. Initial reports [14] with conventional CT suggested that any degree of tumorto-vessel contiguity indicated unresectability of the tumor. However, our subject group mainly consisted of patients with advanced pancreatic cancer, and only seven patients had surgical correlations. Therefore, those vascular invasion criteria may not be applicable to the more limited forms of pancreatic cancer. Recent reports have attempted to grade the degree of vascular invasion. Loyer et al. [7] proposed a classification of vascular involvement in pancreatic ductal adenocarcinoma that divided tumor-to-vessel contiguity into six types, ranging from type A, in which a fat plane is present between the tumor and the vessel, to type F, in which the tumor occludes the vessel. These authors found that 95% of the patients with types A and B were resectable without venous resection, whereas 47% of the patients with type D needed venous resection. Types E and F were unresectable in all cases, and type C was nonspecific. This classification has the merit of being the first attempt to stratify patients with vascular invasion to distinguish the clearly unresectable patients from the potentially resectable patients. However, this classification is complex and relatively subjective and does not provide a clear cut-offs for resectable and unresectable patients. In addition, the authors do not differentiate between venous and arterial invasion. In another series, Lu et al. [15] studied 35 patients with pancreatic adenocarcinoma who underwent surgery. These researchers focused on local staging of pancreatic adenocarcinoma and developed a grading system for vascular tumor invasion on the basis of circumferential contiguity of the tumor to the vessel from no contiguity to more than 75% circumferential involvement. These authors reported a sensitivity of 84% and a positive predictive value and negative predictive value for unresectability of 95% and 93%, respectively, when a threshold of 50% circumferential contiguity of tumor is used as a criterion of unresectability. The main limitation of this study is that only 11 patients were eventually resectable, and most of the surgical correlation was based on venous vessels. This particular limitation is of utmost importance because although venous invasion is a controversial contraindication for radical resection, arterial invasion is a well-accepted unresectability criterion. In our series, arterial invasion was considered a formal contraindication for pancreaticoduodenectomy, and no patients with clear tumor-to-artery contiguity underwent surgery. Conversely, small reticular opacities arising from the tumor and abutting an arterial vessel are occasionally found in preoperative evaluation of pancreatic adenocarcinoma and may raise the question whether arterial invasion is present. We have seen these periarterial streaks in eight patients, all of whom were surgically explored; six patients could be resected, and two could not. Therefore, the positive predictive value of periarterial streaks for unresectability was only 25%. Thus, one cannot rely solely on this finding to exclude patients from surgery, although the surgeon must be made aware that in 25% of the cases, the patient will not be resectable. The sensitivity of helical CT in showing portal or superior mesenteric vein invasion was 75% (9/12) in our series. Five of these patients were successfully resected after venous resection and end-to-end anastomosis in four cases and venous resection with venous graft in the remaining case. In addition, one patient had venous resection for suspected vascular invasion discovered at surgery that was found to be only fibrosis at histopathologic study. In our series, five (83%) of six venous resections had evidence of tumor invasion at microscopic examination. In contradistinction, in the series of Launois et al. [4], only 21% of venous resections showed definite tumor invasion of the venous wall. Because Launois et al. studied 445 patients over a 20-year period, the discrepancy between their rate of tumor venous invasion and ours might be related to our use of better preoperative staging techniques. In conclusion, we find that helical CT used as the only preoperative imaging technique in pancreatic adenocarcinoma provides accurate preoperative staging (positive predictive value for resectability of 77%). Local extension of pancreatic cancer and invasion of adjacent vascular structures are well depicted with dedicated thinsection helical CT. The main limitations of helical CT in preoperative staging are a difficulty in revealing unsuspected tiny liver metastases (55% of the false-negative findings for resectability) and a low rate of revealing lymph node metastasis (16.7%). Possibly, use of multidetector helical CT may improve our results in preoperative staging. Specifically, we believe that obtaining an arterial phase image of the liver before the pancreatic phase may improve the detection of tiny liver metastases. Our experience is that diagnostic endoscopic retrograde cholangiopancreatography is unnecessary in the diagnostic and preoperative workup of patients with suspected pancreatic cancer and should be reserved exclusively for palliative therapy in unresectable patients. In our institution, because of the noninvasiveness of the examination and its wide availability, helical CT has become the routine preoperative imaging technique in patients with suspected pancreatic adenocarcinoma. References 1. Warshaw AL, Fernandez del Castillo C. Pancreatic carcinoma. N Engl J Med 1992;326: Lillemoe KD. Current management of pancreatic carcinoma. Ann Surg 1995;221: Yeo CJ, Cameron JL, Sohn TA, et al. Six hundred fifty consecutive pancreaticoduodenectomies in the 1990s: pathology, complications, and outcomes. Ann Surg 1997;226: Launois B, Stasik C, Bardaxoglou E, et al. Who benefits from portal vein resection during pancreaticoduodenectomy for pancreatic cancer? World J Surg 1999;23: Hommeyer SC, Freeny PC, Crago LG. Carcinoma of the head of the pancreas: evaluation of the pancreaticoduodenal veins with dynamic CT potential for improved accuracy in staging. Radiology 1995;196: Bluemke DA, Cameron JL, Hruban RA, et al. Potentially resectable pancreatic adenocarcinoma: spiral CT. Radiology 1995;197: Loyer EM, David CL, Dubrow RA, Evans DB, Charnsangavej C. Vascular involvement in pan- AJR:178, April

6 Valls et al. creatic adenocarcinoma: reassessment by thinsection CT. Abdom Imaging 1996;21: Freeny PC, Traverso LW, Ryan JA. Diagnosis and staging of pancreatic adenocarcinoma with dynamic computed tomography. Am J Surg 1993; 165: Diehl SJ, Lehman KJ, Sadick M, Lachman R, Georgi M. Pancreatic cancer: value of dual-phase helical CT in assessing resectability. Radiology 1998;206: Nishibaru T, Yamashita Y, Abe Y, et al. Local extension of pancreatic carcinoma: assessment with thin-section helical CT. Radiology 1999;212: Warshaw AL, Gu ZY, Wittenberg J, Walman AC. Preoperative staging and assessment of resectability of pancreatic cancer. Arch Surg 1990;125: Furhman GM, Leach SD, Staley CA, et al. Rationale for en bloc vein resection in the treatment of pancreatic adenocarcinoma adherent to the superior mesenteric portal vein confluence. Ann Surg 1996;223: Cusack JC, Fuhrman GM, Lee JE, Evans DB. Notice to Authors Managing unsuspected tumor invasion of the superior mesenteric portal venous confluence during pancreaticoduodenectomy. Am J Surg 1994; 168: Freeny PC, Marks WM, Ryan JA, et al. Pancreatic ductal adenocarcinoma: diagnosis and staging with dynamic CT. Radiology 1988;166: Lu DS, Reber HA, Krasny RM, Kadell BM, Sayre J. Local staging of pancreatic cancer: criteria for unresectability of major vessels as revealed by pancreatic-phase thin-section helical CT. AJR 1997;168: Authors of accepted manuscripts can now receive galley proofs via . The author s computer system must have version 3.0 (or later) of Adobe Acrobat Reader. The ARRS Web site ( has a link to the Adobe Web site, where authors can download free copies of the most recent version of the reader. 826 AJR:178, April 2002