1 From the Department of Diagnostic Radiology (L.L., Y.A., D.G., P.P.) and Hepatobiliary Surgery Unit (M.D., R. Lapointe, R. Létourneau, A.R.), Centre Hospitalier de l’Université de Montréal, Hôpital Saint-Luc, 1058 Saint-Denis St, Montreal, Quebec, Canada H2X 3J4. Received December 29, 2000; revision requested February 8, 2001; revision received May 15; accepted July 5.
PURPOSE: To determine the accuracy of helical computed tomography
(CT) with CT angiography in identifying vascular invasion by
periampullary neoplasms and to assess the added value of CT
angiography.
MATERIALS AND METHODS: Sixty-nine patients suspected of having periampullary neoplasms were examined. Images from dual phase helical CT with CT angiography were compared with surgical findings in 36 patients. Arterial and venous invasion were assessed separately. Accuracy, positive predictive value (PPV), and negative predictive value (NPV) were determined for CT alone and for CT supplemented with CT angiography.
RESULTS: The accuracy, PPV, and NPV of helical CT with CT angiography in identifying venous invasion was 92% (33 of 36 patients), 86% (12 of 14 patients), and 95% (21 of 22 patients), respectively. When transverse CT images alone were analyzed, accuracy decreased to 69% (25 of 36 patients) (P = .005); PPV and NPV were 63% (five of eight patients) and 71% (20 of 28 patients), respectively. When identifying arterial invasion, the accuracy of CT with CT angiography and of CT alone was 86% (31 of 36 patients). PPV and NPV also were identical at 71% (five of seven patients) and 90% (26 of 29 patients), respectively.
CONCLUSION: CT angiography significantly increases the ability to identify venous invasion when compared with CT alone but does not improve detection of arterial invasion.
Index terms: Pancreas, abnormalities, 770.321 • Pancreas, CT, 770.12112, 770.12115, 770.12116 • Pancreas, neoplasms, 770.321
Despite advances in diagnosis and treatment, periampullary neoplasms
remain a major health problem associated with high mortality.
Carcinoma of the head of the pancreas accounts for a majority
of periampullary neoplasms, with an annual incidence rate of
nine in 100,000 and a case fatality rate of close to 1.0 (
1).
The prognosis is grim, even in patients who undergo tumor resection,
with a median survival of 15.5 months and a 5-year survival
rate of 21% (
2). Other less commonly encountered periampullary
lesions include ampulloma, distal cholangiocarcinoma, and duodenal
neoplasm. Since patients may have inoperable disease at presentation
and the time of diagnosis, staging based on imaging findings
is important to identify a subgroup of patients who might benefit
from surgery and to avoid unnecessary intervention in others.
The advent of helical computed tomography (CT) has improved
our ability to diagnose and evaluate pancreatic and peripancreatic
disease (
3–
7). In the case of pancreatic adenocarcinoma,
evaluation of tumor resectability has been enhanced. Authors
of several studies (
8–
10) have shown that CT angiography
improves the use of helical CT in accurate staging of pancreatic
cancer. Other groups (
5,
11), however, remain skeptical of the
added value of CT angiography. It is not clear whether analysis
of transverse images alone is incomplete and if further postprocessing
of imaging data is essential to access all available information.
The purpose of our study was to determine the accuracy of helical
CT with CT angiography in identifying vascular invasion by periampullary
neoplasm and to assess the added value of CT angiography.
Between October 1996 and September 2000, 69 patients referred
to our center for evaluation of suspected periampullary neoplasm
underwent helical CT. All helical CT was performed with a HiSpeed
Advantage scanner (GE Medical Systems, Milwaukee, Wis). Patients
received 150 mL of 300 mg/dL iodinated nonionic contrast material
(iopromide, Ultravist 300; Berlex Canada, Montreal, Quebec,
Canada) intravenously at a rate of 4 mL/sec. For patients whose
body weight exceeded 80 kg (
n = 6), 180 mL was infused at 5
mL/sec. Acquisition of the arterial phase began 20 seconds after
the beginning of intravenous infusion, and portal phase acquisition
began 60 seconds after the beginning of infusion. Section collimation
was set at 3 mm, and pitch was selected to allow scanning of
the liver and pancreas in one breath hold for each phase (25–30
seconds). Median pitch was 2.0 (range, 1.6–3.0), and the
same pitch was used during arterial and portal phases for each
patient. Data acquired were then reconstructed at 1.5-mm intervals.
In each case, CT angiographic images of the arterial and portal
phases were obtained.
Comparisons with surgical findings were available in 36 patients, who constituted the basis for the study. There were 19 men and 17 women, with average ages of 61.7 years (range, 43.0–74.0 years) and 58.6 years (range, 37.0–72.0 years), respectively. Eighteen patients underwent Whipple pancreatoduodenectomy, whereas 18 patients underwent palliative procedures. Histologic diagnoses in the 36 patients who underwent surgery included pancreatic adenocarcinoma (n = 25), ampulloma (n = 6), cholangiocarcinoma (n = 3), duodenal lymphoma (n = 1), and neuroendocrine tumor (n = 1). Of the remaining 33 patients, one underwent surgery with open biopsy but without resection or adequate evaluation of the vessels. The remaining 32 patients were deemed unable to undergo surgery after CT (ie, had obvious vascular or adjacent organ invasion, adenopathy, or metastatic disease, or were poor surgical risks) and were excluded from the study.
Helical CT scans were analyzed prior to surgery, without knowledge of other imaging findings (ie, angiographic or ultrasonographic images). All images were analyzed by one radiologist (L.L.). Because our imaging department still uses film, transverse images were analyzed both on film and at a workstation. CT angiographic images were obtained and analyzed with Advantage Windows version 2.0 and 3.1 software at Advantage Windows workstations (GE Medical Systems). CT angiograms were generated from transverse data at the time of interpretation and were analyzed concurrently with the transverse images. Since extensive local disease, significant adenopathy, and metastatic disease were not present in the 36 patients in the study, data analysis was limited to vascular involvement. For each case, tumor encasement of the celiac trunk, hepatic artery, and superior mesenteric artery, as well as invasion of the portal vein, splenoportal confluence, and superior mesenteric vein, were assessed. Diagnostic criteria for vascular involvement on the transverse images included vessel occlusion, stenosis caused by tumor contact, or greater than 50% perimeter contact with tumor (12).
A standardized protocol was used to produce CT angiographic images. Arterial invasion was evaluated on maximum intensity projection (MIP) images obtained from the complete data set. CT angiographic images of the venous structures were produced from MIP image renderings obtained from multiplanar volume reconstructions and shaded surface display renderings obtained from the complete data set. The thickness and obliquity of the volume slabs for multiplanar volume reconstructions were selected to enhance visualization of vessel contour at the point of suspected tumor contact. Oblique coronal and sagittal planes were used to visualize the superior mesenteric vein and splenoportal confluence. The workstation we used allowed simultaneous visualization of the transverse, sagittal, coronal, and reconstructed CT angiographic images. It was possible, by moving a cursor along a given vessel on a CT angiographic image, to see the surrounding soft-tissue structures on the corresponding transverse image. Vessel wall irregularities identified on CT angiographic images at points of suspected tumor contact were interpreted as a sign of focal tumor invasion (Fig 1).
| fig.ommitted |
Figure 1a. Adenocarcinoma of the pancreatic head in a 71-year-old woman who did not undergo surgery. (a) Transverse CT image shows tumor (straight arrows) abutting the portal vein (curved arrow) near confluence. (b) Thin-slab MIP image shows contour irregularity (arrow) of the inferior surface of the portal confluence.
| |
| fig.ommitted |
Figure 1b. Adenocarcinoma of the pancreatic head in a 71-year-old woman who did not undergo surgery. (a) Transverse CT image shows tumor (straight arrows) abutting the portal vein (curved arrow) near confluence. (b) Thin-slab MIP image shows contour irregularity (arrow) of the inferior surface of the portal confluence.
| |
Surgery was performed by a team of experienced hepatobiliary
and pancreatic surgeons (M.D., R. Lapointe, R. Létourneau,
A.R.). All relevant transverse and CT angiographic images were
reviewed with the surgeons prior to surgery and were available
to the surgeons in the operating room at the time of surgery.
The surgical findings constituted the standard against which
helical CT images were compared. Institutional review board
approval was obtained for this study; patient informed consent
was not required. Patient charts were consulted, and the surgical
protocol was reviewed to determine if arterial or venous vascular
invasion was present. The surgeons were aware of the study and
evaluated the presence of vascular invasion. In the case of
partial venous resection for suspected limited neoplastic invasion
(one patient), histopathologic confirmation of focal invasion
was obtained.
The accuracies, sensitivities, specificities, positive predictive values (PPVs), and negative predictive values (NPVs) of the transverse CT images alone and of the transverse images supplemented with CT angiographic images for identifying vascular invasion were calculated by using the surgical findings as a reference standard. Arterial and venous invasion were assessed separately. Accuracies were compared by using the McNemar test. Accuracy was defined as the number of true-positive and true-negative cases divided by the total number of patients who underwent surgery (n = 36).
The accuracies, sensitivities, specificities, PPVs, and NPVs
of helical CT with CT angiography and of helical CT alone in
identifying arterial and venous invasion are summarized in the
Table. For evaluating venous invasion, the results of CT alone
and of CT with CT angiography were concordant in 28 of the 36
patients. Both correctly depicted the presence of venous invasion
in 25 patients and resulted in disagreement with surgical findings
in three patients. In the remaining eight patients, CT with
CT angiography depicted venous invasion that was not suspected
on the basis of the transverse CT images alone. With regard
to venous invasion, the accuracy of CT with CT angiography in
identifying tumor ingrowth was 92%, whereas the accuracy of
CT alone was 69%; this difference is statistically significant
(
P = .005). For evaluating arterial invasion, CT images alone
and supplemented with CT angiographic images were concordant
in all cases. There was no significant difference in determining
arterial involvement, whether three-dimensional vascular reconstructions
were used or not. The accuracies of CT with CT angiography and
of CT alone were identical at 86%.
| fig.ommitted |
CT with CT Angiography and CT Alone: Detection of Vascular Invasion Compared with Surgical Findings
| |
CT angiography improves the detection of vessel contour abnormalities
that are less evident on serial images, allowing visualization
of mass effect not seen in the transverse plane (
Figs 2,
3).
Mass effect may not always be due to neoplastic invasion. Inflammatory
changes accompanying neoplastic disease can accentuate mass
effect, leading to displacement and compression of vessels (
Fig 4).
With regard to arterial invasion, arteries may be totally
encased, with no apparent change in vessel caliber, and this
can be demonstrated only on transverse images (
Fig 5). Also,
severely encased vessels can be so stenotic that they fill poorly
and are difficult to identify on CT angiographic images (
Fig 6).
| fig.ommitted |
Figure 2a. Focal invasion of the superior mesenteric vein by pancreatic adenocarcinoma in a 38-year-old woman. (a) Transverse CT image obtained through the level of the pancreatic head shows that tumor (curved arrows) abuts the superior mesenteric vein (straight arrow), without obvious deformation of vessel contour. (b) Thin-slab MIP image of portal phase shows focal indentation (arrow) of the superior mesenteric vein. Tumor invasion was confirmed at surgery at which partial resection was performed, followed by venoplasty.
| |
| fig.ommitted |
Figure 2b. Focal invasion of the superior mesenteric vein by pancreatic adenocarcinoma in a 38-year-old woman. (a) Transverse CT image obtained through the level of the pancreatic head shows that tumor (curved arrows) abuts the superior mesenteric vein (straight arrow), without obvious deformation of vessel contour. (b) Thin-slab MIP image of portal phase shows focal indentation (arrow) of the superior mesenteric vein. Tumor invasion was confirmed at surgery at which partial resection was performed, followed by venoplasty.
| |
fig.ommitted
|
Figure 3a. Adenocarcinoma of the pancreatic head invading the portal confluence in a 73-year-old man. (a) Transverse CT image obtained at the level of the superior mesenteric vein just caudal to the confluence, in which vascular invasion by tumor (arrow) is difficult to appreciate. (b) Shaded surface display of the portal phase shows indentation (arrows) at the confluence of the superior mesenteric and portal veins. Surgical finding confirmed tumor invasion.
| |
fig.ommitted
|
Figure 3b. Adenocarcinoma of the pancreatic head invading the portal confluence in a 73-year-old man. (a) Transverse CT image obtained at the level of the superior mesenteric vein just caudal to the confluence, in which vascular invasion by tumor (arrow) is difficult to appreciate. (b) Shaded surface display of the portal phase shows indentation (arrows) at the confluence of the superior mesenteric and portal veins. Surgical finding confirmed tumor invasion.
| |
| fig.ommitted |
Figure 4a. Adenocarcinoma of the pancreatic head with concomitant inflammatory changes in a 60-year-old man. (a) Transverse CT image shows enlarged pancreatic head (straight arrows), with compression of the superior mesenteric vein (curved arrow). Although no focal lesion is seen, biopsy confirmed the presence of cancer. (b) Thin-slab MIP image shows indentation (arrow) of the superior mesenteric vein. (c) Transverse CT image obtained 8 weeks after b shows size decrease in the pancreatic head (straight arrows) and in mass effect on the superior mesenteric vein (curved arrow).
| |
| fig.ommitted |
Figure 4b. Adenocarcinoma of the pancreatic head with concomitant inflammatory changes in a 60-year-old man. (a) Transverse CT image shows enlarged pancreatic head (straight arrows), with compression of the superior mesenteric vein (curved arrow). Although no focal lesion is seen, biopsy confirmed the presence of cancer. (b) Thin-slab MIP image shows indentation (arrow) of the superior mesenteric vein. (c) Transverse CT image obtained 8 weeks after b shows size decrease in the pancreatic head (straight arrows) and in mass effect on the superior mesenteric vein (curved arrow).
| |
| fig.ommitted |
Figure 4c. Adenocarcinoma of the pancreatic head with concomitant inflammatory changes in a 60-year-old man. (a) Transverse CT image shows enlarged pancreatic head (straight arrows), with compression of the superior mesenteric vein (curved arrow). Although no focal lesion is seen, biopsy confirmed the presence of cancer. (b) Thin-slab MIP image shows indentation (arrow) of the superior mesenteric vein. (c) Transverse CT image obtained 8 weeks after b shows size decrease in the pancreatic head (straight arrows) and in mass effect on the superior mesenteric vein (curved arrow).
| |
| fig.ommitted |
Figure 5a. Adenocarcinoma of the pancreatic head with arterial and venous encasement in a 69-year-old man. (a) Thick-slab MIP image of portal phase shows stenosis (arrow) of the proximal portal vein. (b) Transverse CT image obtained during arterial phase at the level of the origin of the superior mesenteric artery (long straight arrow) shows neoplastic tissue (curved arrow) contacting the superior mesenteric artery and encasing a small vessel (small arrow), which is a replaced right hepatic artery. (c) Thick-slab MIP image from the arterial phase, in which the celiac trunk (large straight arrow) and superior mesenteric artery (curved arrow) are isolated. The proximal portion of a replaced right hepatic artery (small arrows) is not well demonstrated.
| |
| fig.ommitted |
Figure 5b. Adenocarcinoma of the pancreatic head with arterial and venous encasement in a 69-year-old man. (a) Thick-slab MIP image of portal phase shows stenosis (arrow) of the proximal portal vein. (b) Transverse CT image obtained during arterial phase at the level of the origin of the superior mesenteric artery (long straight arrow) shows neoplastic tissue (curved arrow) contacting the superior mesenteric artery and encasing a small vessel (small arrow), which is a replaced right hepatic artery. (c) Thick-slab MIP image from the arterial phase, in which the celiac trunk (large straight arrow) and superior mesenteric artery (curved arrow) are isolated. The proximal portion of a replaced right hepatic artery (small arrows) is not well demonstrated.
| |
| fig.ommitted |
Figure 5c. Adenocarcinoma of the pancreatic head with arterial and venous encasement in a 69-year-old man. (a) Thick-slab MIP image of portal phase shows stenosis (arrow) of the proximal portal vein. (b) Transverse CT image obtained during arterial phase at the level of the origin of the superior mesenteric artery (long straight arrow) shows neoplastic tissue (curved arrow) contacting the superior mesenteric artery and encasing a small vessel (small arrow), which is a replaced right hepatic artery. (c) Thick-slab MIP image from the arterial phase, in which the celiac trunk (large straight arrow) and superior mesenteric artery (curved arrow) are isolated. The proximal portion of a replaced right hepatic artery (small arrows) is not well demonstrated.
| |
| fig.ommitted |
Figure 6a. Adenocarcinoma of pancreas in a 58-year-old woman. (a) Transverse CT image obtained during arterial phase shows neoplastic tissue (small arrows) surrounding the hepatic artery (large straight arrow), which originates from the superior mesenteric artery (curved arrow). (b) Thin-slab MIP image obtained during the arterial phase shows a hepatic artery (straight arrow) of normal caliber; its origin from the superior mesenteric artery (curved arrow) is well demonstrated. Surgery confirmed vascular encasement.
| |
| fig.ommitted |
Figure 6b. Adenocarcinoma of pancreas in a 58-year-old woman. (a) Transverse CT image obtained during arterial phase shows neoplastic tissue (small arrows) surrounding the hepatic artery (large straight arrow), which originates from the superior mesenteric artery (curved arrow). (b) Thin-slab MIP image obtained during the arterial phase shows a hepatic artery (straight arrow) of normal caliber; its origin from the superior mesenteric artery (curved arrow) is well demonstrated. Surgery confirmed vascular encasement.
| |
Despite advances in imaging of pancreatic and periampullary
neoplasms, identification of resectable tumors still constitutes
a significant challenge. In a multiinstitutional study (
13)
completed before the introduction of helical CT, the predictive
value of CT for identifying resectable disease was found to
be as low as 28%. The causes of unresectability include metastases
(ie, hepatic), adenopathy, and vascular invasion. In studies
published before the use of helical CT (
14,
15), the sensitivity
for identifying vascular invasion was 36%–67%. Helical
CT has improved our ability to detect vascular invasion. Diehl
et al (
5) identified vascular invasion with an accuracy of 92%
and an NPV of 88%. To our knowledge, authors of other studies
have not reproduced these results. Bluemke et al (
4) reported
an NPV of only 56%, whereas Raptopoulos et al (
8) found an NPV
of 70% when only transverse images were studied. Investigators
in these two studies specifically assessed the added value of
CT angiography but obtained discordant results. Raptopoulos
et al (
8) reported an increase in NPV from 70% to 96% with CT
angiography, whereas Diehl et al (
5) found that CT angiography
revealed 81% of the vascular invasion detected on transverse
images.
In the absence of vessel occlusion or obvious stenosis, the diagnosis of vascular invasion has been based mostly on the extent of perimeter contact between the vessel and tumor (12,16). Adherence to these criteria still leads to labeling tumors as resectable when, in fact, they are not. Attempts have been made to include vessel morphology in the diagnostic criteria of vascular involvement. Loyer et al (17) have classified tumor contact as forming either a convexity or a concavity against the vessel wall. Hough et al (18) have described a tethered teardrop-shaped superior mesenteric vein as a reliable indicator of unresectability.
Our results show that CT angiography improves detection of vessel contour abnormalities that are less evident on serial transverse images. CT angiography significantly improves the ability to identify venous invasion. CT angiography better demonstrates subtle contour abnormalities of the superior mesenteric vein and the splenoportal confluence, especially in the craniocaudal axis. However, advances in CT technology have improved our ability to visualize tumor and peripancreatic tissue to a degree that further postprocessing is often unnecessary. CT angiography will improve accuracy only by demonstrating subtle abnormalities not evident on transverse images.
Certain considerations limit the accuracy of CT angiography and can lead to false attribution of subtle abnormalities to tumor invasion when there is none. Lower spatial resolution in the z axis can give rise to indistinct vessel margins not due to tumor ingrowth. Furthermore, contrast differences between vessels and adjacent tissue vary from patient to patient, and it is not always possible to clearly distinguish tumor margins from the vessel wall.
The presence of soft tissue in contact with or surrounding vessels may represent true tumor ingrowth but may also be due to a desmoplastic inflammatory reaction (19). It may not be possible to distinguish between these two processes on the basis of imaging data, thus leading to false-positive findings. Because of this issue, detection of subtle vessel abnormalities, especially with respect to venous invasion, should not preclude surgery. In these cases, only laparotomy with surgical exploration will permit definitive staging. Nevertheless, the value of such information preoperatively must not be underestimated.
In the case of venous invasion, some surgeons (20) advocate vascular resection with reanastomosis or grafting, thus prolonging the length and increasing the difficulty of surgery. Some authors (19) claim that the duration of the surgical procedure can be decreased and performance enhanced if the surgeon is aware of possible vascular invasion. To our knowledge, however, there are no published studies linking improved preoperative staging with shortened surgical procedures. Furthermore, it is not clear whether increased accuracy in staging periampullary neoplasms affects the surgeon’s decision to perform surgery, since the demonstration of subtle vascular invasion by tumor does not automatically preclude surgical intervention. Further studies are needed to evaluate these issues.
With regard to arterial invasion, there was no significant difference in accuracies when comparing the ability of CT with CT angiography with that of CT alone to depict unresectable disease. Unlike the venous structures, the celiac trunk, hepatic artery, and superior mesenteric artery are surrounded by fat. The presence of soft tissue in contact with or surrounding these vessels is highly suspicious for tumor encasement, with or without change in vessel caliber. These findings are usually obvious on transverse thin-section images.
In our study, the CT protocol was optimized for vascular imaging and not for imaging of the pancreas and tumor. With single-section helical CT it is not possible to image the arterial, pancreatic, and portal phases independently in a single examination. Since the purpose of the study was to evaluate the use of CT angiography to predict vascular invasion, scanning was limited to the arterial and portal phases. Also, review of patient charts for documenting vascular invasion has limitations; it depends on the completeness of the surgeon’s surgical staging. In the current series of patients, however, the surgeons were aware of the study. Another possible limitation was the use of 50% or more perimeter contact of tumor and vessel as the diagnostic criterion for diagnosing vascular invasion on transverse images. This is based on studies (12,16) showing that the best combination of sensitivity and specificity is obtained with this criterion. It is by no means an absolute value but represents a tradeoff between improved sensitivity and acceptable specificity. Since we had no control over the surgeon’s decision to not perform surgery in 33 of the 69 patients, it must be recognized that a possible selection bias existed.
In conclusion, CT angiography significantly increases the ability to identify venous vascular invasion and augments information drawn from analysis of sequential transverse images, which is still the preferred method of image interpretation. However, limited z-axis resolution, suboptimal contrast, and associated inflammatory changes can limit its usefulness. To further improve the staging of periampullary cancer, a better understanding of enhancement patterns after intravenous infusion is necessary to maximize contrast between vessels and an adjacent tumor. The arrival of multisection CT scanners will undoubtedly modify the acquisition protocols used to study pancreatic and periampullary neoplasms and will improve spatial resolution. Angiographic and three-dimensional reconstructions will become even more useful in interpreting imaging data as resolution in the z plane improves. As soft-copy image reading replaces hard-copy view boxes, rapid three-dimensional modeling of focal areas of interest may soon become part of routine CT scan interpretation.
We thank Marie-Claude Guertin, PhD, for her help with statistical
analysis.
- Gold EB, Goldin SB. Epidemiology of and risk factors for pancreatic cancer. Surg Oncol Clin N Am 1998; 7:67-91.
- Yeo CJ, Cameron Jl, Lillemoe KD, et al. Pancreaticoduodenectomy for cancer of the head of the pancreas: 201 patients. Ann Surg 1995; 221:721-733.
- Bluemke DA, Fishman EK. CT and MR evaluation of pancreatic cancer. Surg Oncol Clin N Am 1998; 7:103-124.
- Bluemke DA, Cameron JL, Hruban RH, et al. Potentially resectable pancreatic adenocarcinoma: spiral CT assessment with surgical and pathological correlation. Radiology 1995; 197:381-385.
- Diehl SJ, Lehmann KJ, Sadick M, et al. Pancreatic cancer: value of dual phase helical CT in assessing resectability. Radiology 1998; 206:373-378.
- Hommeyer SC, Freeny PC, Crabo LG. Carcinoma of the head of the pancreas: evaluation of the pancreaticoduodenal veins with dynamic CT—potential for improved accuracy in staging. Radiology 1995; 196:233-238.
- Ibukuro K, Toshitaka T, Mori K, et al. Peripancreatic veins on thin section (3-mm) helical CT. AJR Am J Roentgenol 1996; 167:1003-1008.
- Raptopolous 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. AJR Am J Roentgenol 1997; 168:971-977.
- Yoshimi F, Hasegawa H, Koizumi S, et al. Application of three-dimensional spiral computed tomographic angiography to pancreatoduodenectomy for cancer. Br J Surg 1995; 82:116-117.
- Zeman RK, Davros WJ, Berman P, et al. Three-dimensional models of the abdominal vasculature based on helical CT: usefulness in patients with pancreatic neoplasms. AJR Am J Roentgenol 1994; 162:1425-1429.
- Baron PL, Kay C, Hoffman B. Pancreatic imaging. Surg Oncol Clin N Am 1999; 8:35-58.
- Lu DSK, 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 Am J Roentgenol 1997; 168:1439-1443.
- Megibow AJ, Zhou XH, Rotterdam H, et al. Pancreatic adenocarcinoma: CT versus MR imaging in the evaluation of resectability—report of the Radiology Diagnostic Oncology Group. Radiology 1995; 195:327-332.
- Vellet AD, Romano W, Bach DB, Passi RB, Taves DH, Munk PL. Adenocarcinoma of the pancreatic ducts: comparative evaluation with CT and MR imaging at 1.5 T. Radiology 1992; 183:87-95.
- Rosch T, Braig C, Gain T, et al. Staging of pancreatic and ampullary carcinoma by endoscopic ultrasonography, computed tomography, and angiography. Gastroenterology 1992; 102:188-199.
- O’ Malley ME, Boland GWL, Wood BJ, Fernandez-del-Castillo C, Warshaw AL, Mueller PR. Adenocarcinoma of the head of the pancreas: determination of surgical unresectability with thin-section pancreatic-phase helical CT. AJR Am J Roentgenol 1999; 173:1513-1518.
- Loyer EM, David CL, Dubrow RA, Evans DB, Charnsangavej C. Vascular involvement in pancreatic adenocarcinoma: reassessment by thin-section CT. Abdom Imaging 1996; 21:202-206.
- Hough TJ, Raptopoulos V, Siewert B, Matthews JB. Teardrop superior mesenteric vein: CT sign for unresectable carcinoma of the pancreas. AJR Am J Roentgenol 1999; 173:1509-1512.
- Yeo CS, Cameron JL. Pancreatic cancer. Curr Probl Surg 1999; 36:57-152.
- Harrison LE, Brennan MF. Portal vein resection for pancreatic adenocarcinoma. Surg Oncol Clin N Am 1998; 7:165-182.
作者：
Luigi Lepanto MD Yervant Arzoumanian MD David 2007-5-12