| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Gastrointestinal Imaging |
1 From the Department of Radiology (M.K., H.H., H.K., M.M.) and the First Department of Internal Medicine (Y.S., H.M.), Gifu University School of Medicine, 40 Tsukasamachi, Gifu 500-8705, Japan. Received April 20, 1999; revision requested June 29; revision received August 5; accepted August 13. Address reprint requests to M.K. (e-mail: masa-gif@umin.ac.jp).
 |
Abstract |
|---|
 TOP Abstract IntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
MATERIALS AND METHODS: Dynamic MR images were obtained after intravenous bolus injection of gadopentetate dimeglumine in 75 patients with no pancreatic malignancies. Test-bolus imaging was performed to determine aortic transit time in individual patients. Patients were randomly assigned so that the middle of k space was acquired at 5, 15, 25, 35, or 45 seconds after arrival of contrast material in the abdominal aorta. Mean signal intensities of pancreas, liver, and peripancreatic vessels were measured, and images were qualitatively assessed by two radiologists.
RESULTS: The best pancreatic enhancement occurred at a delay of 15 seconds (P < .05). The best enhancement of the liver and peripancreatic vessels occurred at 25 seconds or later (P < .05). Qualitatively, the best images of the pancreas were obtained at 15 seconds, and the best images of the liver and peripancreatic vessels were obtained at 25 seconds or later.
CONCLUSION: Biphasic imaging at 15 and 45 seconds or later after arrival of contrast material in the abdominal aorta is a practical method for acquisition of high-quality dynamic gradient-recalled-echo MR images of the pancreas.
Index terms: Liver, MR, 761.121412, 761.121414, 761.121415, 761.12143 • Pancreas, blood supply, 770.92, 95.92 • Pancreas, cysts, 770.312 • Pancreas, fat, 770.299 • Pancreas, MR, 770.121412, 770.121414, 770.121415, 770.12143 • Pancreas, neoplasms, 770.312, 770.32 • Pancreas, surgery, 770.45 • Pancreatitis, 770.291, 770.458
|
Introduction |
|---|
|
TOPAbstractIntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
Previous researchers (15–17) who described dynamic MR imaging of the pancreas used manual injection of contrast material and different amounts of saline solution flush; imaging timing after the contrast material injection also varied among the studies. Recent reports on gadolinium-enhanced MR angiography (20) or dynamic MR imaging of the liver (21) have shown the clinical utility of image-triggering software or test-bolus imaging to help optimize imaging timing to produce images with contrast as intense as possible. We believe these techniques can be applied to dynamic MR imaging of the pancreas, although we are aware of no previous study that has been conducted for the purpose of exploring the optimal imaging delay to obtain dynamic MR images with adequate contrast enhancement of the pancreas and peripancreatic vessels. Therefore, we used quantitative and qualitative assessments in the attempt to determine the delay after arrival of contrast material in the abdominal aorta that would result in optimal pancreatic and peripancreatic vascular enhancement.
|
MATERIALS AND METHODS |
|---|
|
TOPAbstractIntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
The eventual diagnoses in the 85 patients were pancreatic carcinoma in six patients, chronic pancreatitis in 17, pancreatic cyst in three, severe pancreatic lipomatosis in three, status after pancreatic resection because of chronic pancreatitis in one, and healthy pancreas in 55. These diagnoses were determined in conference by three radiologists (M.K, H.K., H.H.) and two gastroenterologists (Y.S., H.M.) on the basis of findings from abdominal US; endoscopic US; CT; T1- and T2-weighted MR imaging; MR cholangiopancreatography; dynamic MR imaging; endoscopic retrograde cholangiopancreatography; 3–6-month follow-up US, CT or MR imaging; serologic tests; or definitive surgery.
Ten patients were excluded from the study during the accumulation of study population: the six with pancreatic carcinoma, the three with severe pancreatic lipomatosis, and one who had undergone pancreatic resection. These patients were excluded because accurate measurements of signal intensity were difficult in pancreatic parenchyma with tumor, atrophy, or defect and because quantitative measurements in these cases might introduce statistically significant bias due to the small number of patients. The remaining 75 patients who formed the study population included 38 men and 37 women aged 20–85 years (mean age, 61 years).
MR imaging was performed with a 1.5-T superconducting magnet (Signa Horizon; GE Medical Systems, Milwaukee, Wis). The system provides a maximum gradient strength of 23 mT · m-1 with a peak slew rate of 120 mT · m-1 · msec-1. All MR images were obtained with a phased-array body multicoil.
The MR imaging protocol consisted of chemical shift–selective fat-suppressed T1-weighted spin-echo imaging (450/8 [repetition time msec/echo time msec], 256 x 192 matrix, two signals acquired, 6.5-minute acquisition time) and non–fat-suppressed respiratory-triggered fast spin-echo T2-weighted imaging (3,750–7,500/88 [effective repetition time msec/effective echo time msec], echo train length of eight to 16, 512 x 256 matrix, three signals acquired, 3.2–4.6-minute acquisition time), breath-hold thick-section (40–70-mm) half-Fourier single-shot fast spin-echo coronal MR cholangiopancreatography (1,596–2,064 [effective]/818–1,034 [effective], 256 x 256 matrix, echo train length of 136, 1-second acquisition time per image), and breath-hold high-spatial-resolution dynamic gadolinium-enhanced fast multiplanar spoiled gradient-recalled-echo (GRE) imaging with steady-state free precession (150/1.6, 110° flip angle, 512 x 224 matrix, receive bandwidth of ±62.5 kHz, one signal acquired, 18 locations per 26 seconds, no chemical shift–selective fat-suppression technique). The k-space lines for the fast multiplanar spoiled GRE sequence were filled with echo data from top to bottom in the phase-encoding direction (sequential view ordering).
For all transverse MR imaging, the section thickness was 5 mm with a 1.5-mm intersection gap, and spatial presaturation pulses were applied superiorly and inferiorly to the imaging volume. Automated shimming was always used before data acquisition. The imaging range was determined so that the imaging volume provided sufficient coverage from the celiac trunk and pancreatic tail to the transverse portion of the duodenum.
Test-Bolus Imaging
To measure aortic transit time, defined as the time from initiation of intravenous injection of contrast material to peak enhancement of the abdominal aorta at the level of the first lumbar vertebral body, single-section fast multiplanar spoiled GRE images (19/1.3, 110° flip angle, 256 x 128 matrix, field-of-view of 32 x 24 cm, receive bandwidth of ±62.5 kHz, one signal acquired, 2-second acquisition time) were obtained every 2 seconds after initiation of an intravenous bolus injection of 1 mL of gadopentetate dimeglumine (Magnevist; Schering, Berlin, Germany) followed by a flush with sterile saline solution. The volume of the flush was calculated according to the patient's body weight by using the following equation: V = 15 + 0.2 x Wbody - 1, where V is the volume of the saline solution in milliliters and Wbody is the patient's body weight in kilograms). Signal intensities of the abdominal aorta on test-bolus GRE images were measured by using operator-defined region-of-interest measurements of mean signal intensity, and the aortic transit time was determined as an odd number (Fig 1).
|
The imaging delay D (in seconds) for first-phase dynamic GRE imaging of the pancreas was determined in individual patients by using the following equation: D = TV-A + t - 13, where TV-A is the aortic transit time and t is the imaging delay (in seconds) between arrival of contrast material in the abdominal aorta and the middle of the dynamic GRE acquisition. Thirteen seconds (one-half the acquisition time of the high-spatial-resolution fast multiplanar spoiled GRE sequence) was subtracted because the central k-space lines were filled with echo data at the middle of the acquisition, and echoes sampled at the middle of the acquisition markedly affected the contrast of the entire image (Fig 1). The delays for second- and third-phase dynamic GRE imaging were fixed at 90 seconds and 3 minutes, respectively, in all patients. All patients were randomly assigned to one of five groups (15 patients per group) with an imaging delay (t) of 5, 15, 25, 35, or 45 seconds for first-phase dynamic GRE MR imaging.
Quantitative Image Analysis
The quantitative measurement was conducted by either of two radiologists (M.K. or H.K.). Operator-defined region-of-interest measurements were performed on nonenhanced and first-phase dynamic gadolinium-enhanced GRE MR images, with mean signal intensity measured in the uncinate process, head, body, and tail of the pancreas; the right and left lobes of the liver; the abdominal aorta; the superior mesenteric artery and vein; and background noise. Signal intensities in the pancreas and liver were measured in areas devoid of focal changes in signal intensity, large vessels, the main pancreatic duct, and prominent artifacts. For measurement of a vessel, a circular region of interest was drawn to encompass as much of the vessel as possible. When the vessel was too small to place a region of interest, the image was magnified up to four times.
In each patient, the four pancreatic measurements and the two liver measurements were averaged to obtain the mean signal intensities of the pancreas and liver, respectively. The SD of background noise was measured in the phase-encoding direction outside the anterior abdominal wall to calculate signal-to-noise ratios for the pancreas, liver, abdominal aorta, and superior mesenteric artery and vein. These signal-to-noise ratios were calculated by dividing the mean signal intensity in the respective organ or vessel by the SD of background noise. Degrees of contrast enhancement in the pancreas, liver, abdominal aorta, and superior mesenteric artery and vein were expressed as contrast-enhancement indexes, which were calculated by subtracting the signal-to-noise ratios on nonenhanced GRE images from those on the first-phase dynamic GRE images.
Qualitative Image Analysis
Two readers (M.K., H.K.), who have served mainly as gastrointestinal radiologists and have interpreted MR images of the abdomen as part of their daily clinical and research practice, independently reviewed the first-phase dynamic GRE images. The images were presented in random order. Each reader evaluated the degrees of contrast enhancement for the pancreas, liver, and peripancreatic vessels by using a five-point scale, with a score of 1 assigned for no enhancement; a score of 2, for poor enhancement; a score of 3, for fair enhancement; a score of 4, for good enhancement; and a score of 5, for excellent enhancement. A score of 5 was assigned when contrast enhancement was sufficient to recognize the pancreatic and liver parenchymas, anatomic boundaries, or peripancreatic vessels. A score of 3 was assigned when contrast enhancement was moderate, but recognition of the pancreatic and liver parenchymas, anatomic boundaries, or of peripancreatic vessels was not precluded. A score of 1 was assigned when contrast enhancement was almost absent and interpretation of contrast-enhanced images was markedly precluded. Scores of 2 or 4 were assigned according to the reader's subjective judgment.
Statistical Analysis
The analysis of variance and multiple comparisons were performed by using the Scheffé criterion (23) as applied in the five groups for patient age and body weight; aortic transit time; contrast-enhancement indexes for the pancreas, liver, abdominal aorta, and superior mesenteric artery and vein; and scores for degree of contrast enhancement of the pancreas, liver, and peripancreatic vessels as subjectively judged by two readers.
The 
statistic was used to assess interobserver variability in the evaluation of the degree of contrast enhancement. We used the nonweighted statistic, with binary data defined in terms of the less-than-50% cutoff level. The degree of disagreement was not factored into the calculation. A value of up to 0.40 indicated positive but poor agreement; a value of 0.41–0.75, good agreement; and a value of greater than 0.75, excellent agreement.
|
RESULTS |
|---|
|
TOPAbstractIntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
|
|
|
|
|
|
|
|
|
|
values for the two independent readers were 0.54, 0.91, and 0.92 for degree of enhancement of the pancreas, liver, and peripancreatic vessels, respectively. Good agreement was obtained with regard to the degree of pancreatic enhancement, and excellent agreement was obtained with regard to the degree of hepatic and peripancreatic vascular enhancement. |
DISCUSSION |
|---|
|
TOPAbstractIntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
In none of these previous studies with CT, however, was the difference in aortic transit time in individual patients taken into consideration. Our results indicated that the aortic transit time ranged widely from 9 to 29 seconds, which suggested that the imaging delay should be optimized according to the time of arrival of contrast material in the abdominal aorta rather than from the initiation of the injection of contrast material. This is especially true for dynamic MR imaging, because the amounts of contrast material and saline solution flush are much less for dynamic MR imaging than for dynamic CT, and the aortic transit time may be more dependent on the individual patient's circulation time with dynamic MR imaging than with dynamic CT.
Although the amounts and injection rates of contrast medium differed between the previous CT studies and the present MR imaging study, the observation that greater pancreatic contrast enhancement was achieved at earlier-phase imaging was similar. Our results showed that the greatest enhancement of the pancreatic parenchyma occurred at 15 seconds after arrival of the contrast material in the abdominal aorta. Imaging at this "pancreatogram" phase may be useful for help in the detection of small hypovascular pancreatic carcinoma because the tumor-to-pancreas contrast may be maximal at this phase. When a small pancreatic tumor that does not deform the pancreatic contour is suspected, "pancreatogram phase" dynamic MR imaging may be of assistance for further diagnosis. It is not clear, however, whether the images obtained at the phase when the pancreas is most strongly enhanced really help improve detection of small pancreatic neoplasms. Keogan et al (27) reported that the acquisition of arterial phase CT images (20–40 seconds after injection) in addition to venous phase images (70–100 seconds after injection) did not result in improved detection of pancreatic malignancy.
The pancreas is a noncapsulated organ that is harbored in the retroperitoneal space, and a contour-deforming pancreatic carcinoma readily invades the retroperitoneal space or peripancreatic vessels. Because preoperative diagnosis of such extrapancreatic involvement affects the determination of resectability and the prognosis, dynamic MR imaging in patients suspected of having pancreatic carcinoma should include imaging at the phase when the peripancreatic arteries and veins are well enhanced (2). To obtain images with sufficient enhancement of the pancreas and peripancreatic vessels at a single session of dynamic MR imaging, a biphasic imaging protocol may be mandatory. If, however, on the basis of our results, biphasic dynamic MR images are to be obtained at the pancreatic- and peripancreatic vascular–enhancement phases, then biphasic imaging should be performed so that the middle of each acquisition is 15 and 45 seconds after arrival of contrast material in the abdominal aorta for adequate enhancement of the pancreas and peripancreatic vessels, respectively (Figs 3, 5).
For high-spatial-resolution fast multiplanar spoiled GRE MR imaging, however, there are only 4 seconds for breathing between the two breath-hold acquisitions at 15 and 45 seconds. We were obliged to compromise with regard to the lack of a sufficient breathing interval by using a smaller image matrix (eg, 256 frequency-encoding and 160 phase-encoding steps), which shortened the acquisition time to 18–20 seconds while maintaining the small section thickness and number of obtainable sections for enough coverage of the pancreas. Another compromise would involve use of a larger matrix and delay of second-phase imaging to later than 45 seconds, because preferable contrast enhancement of the liver and peripancreatic vessels would be expected even on images obtained later than 45 seconds after arrival of contrast material in the aorta. Such problems will be solved, however, if future technical developments allow shorter acquisition times for high-spatial-resolution GRE imaging.
Another way to compromise with high-spatial-resolution dynamic GRE imaging may be to perform uniphasic imaging of the pancreas at 25–35 seconds to obtain images with both pancreatic and peripancreatic vascular enhancement (Fig 4), because the mean contrast-enhancement index of the pancreas reached a plateau at 25–35 seconds, which was probably due to recirculation of the contrast material bolus. This was deduced on the basis of our result that there was a second peak of contrast enhancement of the abdominal aorta at 35 seconds (Fig 2b).
When determining surgical indications for patients with pancreatic carcinoma, detection of hepatic metastases is one of the most important factors. Intense hepatic enhancement was seen at 25, 35, and 45 seconds when a contrast material influx via the portal venous system occurred, which represented the portal venous phase of hepatic enhancement. In the present study, we used a 5-mm section thickness and a 1.5-mm intersection gap to maximize spatial resolution along the z axis for the pancreas and peripancreatic vessels; thus, the imaging volume did not cover the cranial one-third of the liver. However, because most hepatic metastases from pancreatic ductal adenocarcinoma are well depicted on portal venous phase MR images, imaging of the whole liver at this phase should ideally be included in the dynamic MR imaging protocol for the pancreas. When imaging of the liver is mandatory for evaluation of possible liver metastases, we must compromise by decreasing the size of the image matrix or increasing the section thickness and intersection gap to sufficiently cover the whole liver and pancreas during a single breath-hold at GRE MR imaging.
There are some limitations to our study. Triggering software that was compatible with the fast multiplanar spoiled GRE sequence was not available at the time of our study; therefore, we performed test-bolus imaging to measure aortic transit time in individual patients. If triggering software were available, however, dynamic GRE MR imaging of the pancreas would be more feasible in clinical practice.
Various degrees of pancreatic lipomatosis are occasionally seen, especially in elderly or obese patients (28), and a decrease in signal intensity may occur at the pancreatic-enhancement phase due to paradoxical suppression with opposed-phase GRE MR imaging (29). Further, chemical shift–cancellation marginal artifacts on opposed-phase GRE images may degrade image quality or mask extrapancreatic tumor involvements. To remedy this artifact, in-phase GRE MR imaging (echo time, 4.2 msec) or MR imaging with chemical shift–selective fat suppression may be helpful. Use of these techniques, however, results in a trade-off in terms of the number of obtainable sections or the acquisition time.
Our results in patients without tumors may not always be applied in patients with pancreatic carcinoma because underlying pancreatitis or distal obstructive pancreatopathy can alter the perfusion status in the pancreas. Although we did not separately perform qualitative assessment of peripancreatic arterial enhancement owing to limitations in the depiction of small peripancreatic arteries, we consider that arterial phase images may be helpful in evaluation of potential encasement or thrombosis of the celiac or superior mesenteric arteries, as was suggested by our quantitative results.
In conclusion, the best enhancement of the pancreas is achieved with dynamic GRE MR imaging when gadopentetate dimeglumine (0.1 mmol/kg) and the sterile saline solution flush (15 mL) are intravenously injected at a rate of 3 mL/sec, and imaging is initiated so that the middle of the acquisition occurs 15 seconds after arrival of contrast material in the abdominal aorta. The liver and peripancreatic vessels were well enhanced at 25 seconds or later after arrival of contrast material in the abdominal aorta. Biphasic dynamic MR imaging at 15 and 45 seconds (or later) may be a practical way to obtain pancreatic- and peripancreatic vascular–enhancement images, respectively, during a single session of high-spatial-resolution dynamic GRE MR imaging. Another method may be uniphasic high-spatial-resolution dynamic GRE imaging at 25–35 seconds after arrival of contrast material in the aorta. Radiologists should understand how delaying dynamic gadolinium-enhanced MR imaging of the pancreas until after arrival of contrast material in the abdominal aorta affects pancreatic and peripancreatic vascular enhancement.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
Author contributions: Guarantors of integrity of entire study, H.H., H.M.; study concepts, M.K., Y.S.; study design, M.K.; definition of intellectual content, M.K., H.H.; literature research, M.K.; clinical studies, M.K., Y.S., H.K., M.M.; data acquisition, M.K., H.K., M.M.; data analysis, M.K.; statistical analysis, M.K.; manuscript preparation, M.K.; manuscript editing, M.K., Y.S.; manuscript review, H.H., H.M.
|
References |
|---|
|
TOPAbstractIntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  L. L. Chu, B. N. Joe, A. C. A. Westphalen, E. M. Webb, F. V. Coakley, and B. M. Yeh Patient-specific Time to Peak Abdominal Organ Enhancement Varies with Time to Peak Aortic Enhancement at MR Imaging Radiology, December 1, 2007; 245(3): 779 - 787. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Kanematsu, R. C. Semelka, M. Matsuo, H. Kondo, M. Enya, S. Goshima, N. Moriyama, and H. Hoshi Gadolinium-enhanced MR Imaging of the Liver: Optimizing Imaging Delay for Hepatic Arterial and Portal Venous Phases—A Prospective Randomized Study in Patients with Chronic Liver Damage Radiology, November 1, 2002; 225(2): 407 - 415. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| RADIOLOGY | RADIOGRAPHICS | RSNA JOURNALS ONLINE |
