| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Technical Innovation |
1 All authors: Department of Radiology, University Hospital of Geneva, Rue Micheli-du-Crest 24, Geneva, Switzerland 1205.
Received September 6, 2004;
accepted after revision March 3, 2005.
Address correspondence to P.-A. Poletti
(pierre-alexandre.poletti{at}hcuge.ch).
Abstract
 Top Abstract IntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
CONCLUSION. Patients were separated into two groups: P1 (severe) and P2 (mild) acute pancreatitis. Mean perfusion value was 24.8 mL/100 mL/min in the P1 group and 50.5 mL/100 mL/min in the P2 group (p = 0.0016, significant). Our preliminary data suggest that pancreatic perfusion measurement using MDCT with perfusion imaging could help in assessing the severity of acute pancreatitis.
Keywords: CT technique • pancreas • pancreatitis • perfusion CT
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
The purpose of this study was to develop an MDCT protocol with perfusion imaging to assess pancreatic perfusion and to determine if we could evidence a correlation between pancreatic perfusion measured by this technique and the outcome of patients with acute pancreatitis.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
For this workup, 16-MDCT (Mx 8000, Philips Medical Systems) was used with the following protocol: A first acquisition (120 kV, 200 mAs) was obtained without IV injection of contrast media. This series helped in locating the pancreas and setting the four dynamic acquisition slices at the correct level. Dynamic images (90 kV, 100 mAs) were then obtained instead of the usual contrast-enhanced arterial phase. Four slices with a beam collimation of 6 mm were set to cover as much of the pancreas as was possible. Acquisition started simultaneously with IV injection of a 40-mL bolus of iodinated contrast (Ultravist 300 [iopromide], Schering). This sequence was performed during a single breath-hold, as were all sequences in this protocol. Acquisition started simultaneously with bolus injection (injection rate was 5 mL/sec) and lasted for 40 sec at the rate of one image per second. This sequence provided a total of 160 images (40 images for each slice level).
Finally, a last acquisition (venous phase) was performed as in our normal workup routine, 60 sec after IV injection of a complemental dose of 100 mL of contrast media at the rate of 3.5 mL/sec. Dynamic images were then processed using the MxView Independent Multi-Modality Diagnostic Workstation with dedicated MxView v5.0.1 software (Philips).
Quality of the acquisition was estimated from the attenuation curves over time obtained in the different regions of interest. Quality was rated good if reliable measurements could be performed in all three pancreatic anatomic subdivisions (i.e., head, body, and tail) and poor if no reliable measurements could be obtained.
A visual estimation of the severity of the disease was made using Balthazar's criteria [6]. The following parameters were then calculated: perfusion (measured in mL/100 mL/min), peak enhancement intensity (PEI) (measured in Hounsfield units), time to peak (TTP) (measured in sec), and blood volume (BV) (measured in mL/100 mL) (Figs. 1A, 1B, 1C, 1D, 2A, 2B, 2C, and 2D). The CT perfusion software calculates the perfusion based on the maximum slope model. The perfusion is calculated as the average slope of the tissue enhancement divided by the peak enhancement in the aorta, as described initially by Miles et al. [7] and Miles [8].
|
|
|
|
|
|
|
|
An input region of interest (ROI) was set over the abdominal aorta; then tissue attenuation over time was measured in various ROIs set over the pancreas (usually the head, body, and tail). Special care was taken not to include visible vessels and to avoid obviously pathologic zones such as necrosis or cysts. The size of the ROI should be at least 50 pixels to minimize the effect of photon noise and with a sufficient margin to avoid partial volume effects. Thresholds were set in the software's preferences at values such that pixels with an initial attenuation value of less than –50 H or more than 100 H would not be reported to help exclude air and bone tissue.
To minimize risk of operator-dependent bias, all perfusion parameters were calculated twice for every patient with the operators blinded to the final diagnosis. Results of the two sets of measurements were compared, and when there was a significant difference between the two sets of measurements, the source of error was sought.
Measurements of perfusion, PEI, TTP, and BV were done in the three anatomic pancreatic subdivisions (head, body, and tail) whenever possible. In some cases, reliable measurements could be obtained in only one or two of these pancreatic subdivisions. Clinical data such as patient history, initial serum amylase and lipase, subsequent treatment, and final diagnosis at discharge were obtained retrospectively from a charts review and collected in a FileMaker Pro 6 database (FileMaker, Inc.).
Patients with confirmed pancreatitis (P+) were separated in two groups: P1 (severe acute pancreatitis) when subsequent recovery in the ICU, surgical débridement, or drainage was needed and P2 (mild acute pancreatitis) when conservative treatment was sufficient. Patients who needed a cholecystectomy after uncomplicated biliary pancreatitis were also considered as having mild acute pancreatitis because the surgical procedure was not due to the pancreatitis itself.
Statistical analysis was performed using the Wilcoxon rank sum test. For the global level of 5%, we had to make a Bonferroni correction because the error rate would be larger than 5% with multiple testing. So for the single tests we used a level of 0.05 / 8 = 0.00625. We considered that if the p value for a parameter was smaller than 0.00625, there was a significant difference between the patients of the two samples.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Eighty-two patients (50 men and 32 women) were retained for this study. The severity of the suspected acute pancreatitis, using Balthazar's criteria, was rated A in 22, B in 16, C in 21, D in 12, and E in 11 patients. The diagnosis of pancreatitis was retrospectively confirmed in 61 patients (P+) and excluded in 21 (P–). In the P– group, five had a neoplastic disease, three had an uncomplicated cholecystitis, five had an infectious disease involving the abdominal cavity or gastrointestinal tract, one had a mesenteric infarct, one had a small-bowel occlusion, one had gastritis related to portal hypertension, two suffered a polytrauma with suspicion of a pancreatic lesion, and three presented with abdominal pain of unknown origin.
Mean pancreatic perfusion was 61.2 mL/100 mL/min (range, 9.4–228.1 mL) in P– and 47.2 mL/100 mL/min (range, 7.3–157.1 mL) in P+ (p = 0.2362, not significant). Mean PEI was 43.2 H (range, 8.4–79 H) in P– and 9.4 H (range, 7.8–85.1 H) in P+ (p = 0.3103, not significant). TTP was 19 sec (range, 9.6–29.3 sec) in P– and 20.4 sec (range, 10.5–35.2 sec) in P+ (p = 0.3836, not significant). BV was 21.1 mL/100 mL (range, 2.6–48.3 mL) in P– and 22 mL/100 mL (range, 3.1–59.1 mL) in P+ (p = 0.6864, not significant).
In the P+ group, eight patients were considered as having severe acute pancreatitis (P1) and 53 as having mild pancreatitis (P2). Mean perfusion value was 24.8 mL/100 mL/min (range, 15.5–37.8 mL) in P1 and 50.5 mL/100 mL/min (range, 7.3–17.1 mL) in the P2 subgroup (p = 0.0016, significant). Mean PEI was 24.9 H (range, 7.8–31.1 H) in P1 and 41 H (range, 8.6–85.1 H) in P2 (p = 0.0048, significant). TTP was 22.7 sec (range, 16–29.8 sec) in P1 and 20.0 sec (range, 10.5–35.2 sec) in P2 (p = 0.1714, not significant). BV was 10.4 mL/100 mL (range, 5.4–22 mL) in P1 and 23.7 mL/100 mL (range, 3.1–59.5 mL) in P2 (p = 0.0010, significant).
The average effective dose of the MDCT on a 2-cm area was calculated to be 0.0295 x 0.015 x 2 = 0.000885 mSv/mAs for one acquisition. The whole effective dose for 40 acquisitions at 100 mAs was 3.54 mSv.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Recent angiographic studies have suggested that locoregional perfusion alterations play an important role in the pathophysiology of the pancreas, especially in acute pancreatitis [4]. However, subtle changes in pancreatic perfusion can be overlooked by this conventional imaging technique, and to date, no quantitative evaluation of regional pancreatic blood flow can be easily obtained in the clinical setting.
Dynamic CT with perfusion imaging has previously enabled measurement of tissue perfusion including the brain, kidney, heart, liver, and spleen [7, 8]. One study by Miles et al. [9] in 1995 described the first application of this technique in eight normal pancreases. In that study, perfusion values ranged from 125 to 166 mL/100 mL/min. In another study [10], pancreatic perfusion was measured using dynamic CT in 23 patients without pancreatic disease. In that study, pancreatic perfusion ranged from 55.4 to 196.8 mL/100 mL/min. Perfusion data were then correlated with patient demographic data and evidence was found that pancreatic perfusion tends to decline with age.
Although we are aware that regional perfusion can be heterogeneous in cases of focal pancreatitis, we have chosen to calculate the average perfusion in the whole gland to simplify the procedure, and we are aware that doing so could be a limitation of this study. However, angiographic studies of patients with severe acute pancreatitis [5] have shown that there are ischemic changes involving not only the intra- and peripancreatic arteries but also the extrapancreatic splanchnic arteries. These diffuse splanchnic perfusion alterations were present even before necrosis developed and were well correlated with the severity of the disease.
In the P– group, we found a mean perfusion value of 61.2 mL/100 mL/min, with extremes ranging from 9.4 to 228 mL/100 mL/min. However, this group consisted of a heterogeneous population of acutely ill patients ranging in age from 32 to 92 years. Five patients had an abdominal sepsis, one had a mesenteric infarct, one had a small-bowel occlusion, one suffered from portal hypertension, and two suffered abdominal trauma with suspicion of a pancreatic lesion.
Perfusion values ranged from 9.4 mL/100 mL/min (in a 76-year-old man with an atrophic pancreas from chronic pancreatitis who was in a state of septic shock) to 228.1 mL/100 mL/min (in a 32-year-old woman with abdominal abscess from a fistulating desmoid tumor). We think these factors explain the great variability of the measured perfusion parameter in this group. This group cannot be considered a control group, and the values of the perfusion parameters measured in this group are published here for illustration purposes only.
In the P+ group, we found a mean perfusion value of 47.2 mL/100 mL/min, with extremes ranging from 7.3 to 157.1 mL/100 mL/min. There was no statistically significant difference between the P+ group and the P– group, probably because of the heterogeneity of the P– group and great range of parameter values. The purpose of the study was not to find evidence of pancreatitis among control patients but to assess if pancreatic perfusion could be measured using MDCT with dynamic perfusion imaging and to correlate these results with patient outcome.
Mean PEI values were also statistically different between the P1 and P2 groups. This parameter might also be used to differentiate severe and mild acute pancreatitis. The lack of pancreatic parenchymal enhancement, often related to the presence of necrotic tissue, was always considered a sign of severity. Nevertheless, perfusion CT allows for dynamic evaluation of the enhancement pattern over time and the precise determination of the peak value of pancreatic parenchymal enhancement, which cannot be achieved visually on sequential (arterial and venous phase) CT.
Twenty-three percent of the procedures performed in this study yielded unreliable data that could not be used for perfusion measurement. This rather elevated failure rate can be explained partly because the ideal timing of the dynamic image acquisition after contrast media IV injection was found by trial and error. This caused at least 12 procedures to come out with incomplete attenuation curves in the beginning of our study.
Another problem with this technique is that patients must be able to hold their breath for 40 sec, which might be difficult for acutely ill patients. Respiratory motion during dynamic imaging can cause the pancreas to slip in and out of the acquisition slices and the attenuation curve becomes irregular and dented. This might not be a problem in some cases because the perfusion is extrapolated from the slope of the ascending portion of this curve, but it can cause results to be unreliable when motion artifacts affect this portion of the curve. This problem caused 16 procedures to be rejected from our study.
The major cause of mismatch between the two sets of measurements was inappropriate ROI positioning, with either an ROI set over a part of the pancreas that moved during acquisition or an ROI that included a vessel. Care should also be taken to set the dynamic acquisition slices over the pancreas in the same breath-holding conditions as in the acquisition itself.
The statistically significant difference between the P1 group and the P2 group confirms the observation of Inoue et al. [5], who reported that the degree of ischemic change on angiography was associated with the severity of the acute pancreatitis.
This study shows that pancreatic perfusion measurement in patients with acute pancreatitis can be performed in the clinical setting using dynamic MDCT. To date we do not know if these perfusion parameters can have any impact on patient management, and we are currently running a prospective study to validate these preliminary results and find out how to take advantage of these new data. We hope that this new study will help us define a cutoff perfusion value beneath which a patient would be at risk of developing an adverse outcome.
If these preliminary results are confirmed by our upcoming study, this relatively simple imaging technique could be integrated into the usual CT workup of patients with acute pancreatitis. Our hope is that MDCT with perfusion imaging will help in identifying patients who are at risk of developing severe complications before the occurrence of necrosis or other evident morphologic alterations. These patients could then benefit from a more aggressive medical treatment from the beginning.
Acknowledgments
We thank Max Wintermark for his valuable advice on MDCT perfusion
measurement methodology and Anja Federer and Michael Vock, University of Bern,
for the statistical work.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  M. A. Bali, T. Metens, V. Denolin, V. De Maertelaer, J. Deviere, and C. Matos Pancreatic Perfusion: Noninvasive Quantitative Assessment with Dynamic Contrast-enhanced MR Imaging without and with Secretin Stimulation in Healthy Volunteers--Initial Results Radiology, April 1, 2008; 247(1): 115 - 121. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
