Portal de Periódicos da CAPES

Perfusion Computerized Tomography Can Predict Pancreatic Necrosis in Early Stages of Severe Acute Pancreatitis

2007; Elsevier BV; Volume: 5; Issue: 12 Linguagem: Inglês

10.1016/j.cgh.2007.07.014

1542-7714

Yoshihisa Tsuji, Hiroshi Yamamoto, Shujiro Yazumi, Yuji Watanabe, Kazuhiro Matsueda, Hiroyuki Yamamoto, Tsutomu Chiba,

Gallbladder and Bile Duct Disorders

Background & Aims: The mortality rate associated with severe acute pancreatitis (SAP) with necrosis remains high because early prediction of pancreatic necrosis is difficult. We evaluated whether perfusion computerized tomography (CT), which is used to identify early stage ischemia in the brain, could detect ischemic changes in the pancreas in the early stages of SAP and predict development of necrosis. Patients and Methods: Thirty consecutive patients with a diagnosis of SAP according to the Atlanta criteria and whose score was greater than 6 were enrolled in this study. All patients were hospitalized within 3 days after onset of symptoms indicative of acute pancreatitis and underwent perfusion CT. Three weeks later, all patients underwent conventional contrast-enhanced CT to detect progression of their disease. Results: Perfusion CT showed that 10 of the 30 patients had pancreatic ischemia at the time of diagnosis. Contrast-enhanced CT disclosed that pancreatic necrosis developed in 9 of these 10 patients, but not in the 20 patients who did not have pancreatic ischemia. The sensitivity and specificity of perfusion CT for predicting pancreatic necrosis was calculated to be 100% and 95.3%, respectively. Conclusions: Perfusion CT is a useful tool for early detection of ischemic changes in the pancreas that lead to pancreatic necrosis. Background & Aims: The mortality rate associated with severe acute pancreatitis (SAP) with necrosis remains high because early prediction of pancreatic necrosis is difficult. We evaluated whether perfusion computerized tomography (CT), which is used to identify early stage ischemia in the brain, could detect ischemic changes in the pancreas in the early stages of SAP and predict development of necrosis. Patients and Methods: Thirty consecutive patients with a diagnosis of SAP according to the Atlanta criteria and whose score was greater than 6 were enrolled in this study. All patients were hospitalized within 3 days after onset of symptoms indicative of acute pancreatitis and underwent perfusion CT. Three weeks later, all patients underwent conventional contrast-enhanced CT to detect progression of their disease. Results: Perfusion CT showed that 10 of the 30 patients had pancreatic ischemia at the time of diagnosis. Contrast-enhanced CT disclosed that pancreatic necrosis developed in 9 of these 10 patients, but not in the 20 patients who did not have pancreatic ischemia. The sensitivity and specificity of perfusion CT for predicting pancreatic necrosis was calculated to be 100% and 95.3%, respectively. Conclusions: Perfusion CT is a useful tool for early detection of ischemic changes in the pancreas that lead to pancreatic necrosis. Despite recent improvements in the treatment of severe acute pancreatitis (SAP), the mortality rate associated with this condition remains high (eg, as great as 20% in Japan).1Sekimoto M. Takada T. Kawarada Y. et al.JPN guidelines for the management of acute pancreatitis: epidemiology, etiology, natural history, and outcome predictors in acute pancreatitis.J Hepatobiliary Pancreat Surg. 2006; 13: 10-24Crossref PubMed Scopus (177) Google Scholar The initiation of appropriate intensive treatment for SAP, however, often is delayed because the scoring systems used to assess the severity of acute pancreatitis and to predict its prognosis (ie, within the first 24 hours of onset) still are inadequate.2Neoptolemos J.P. Kemppainen E.A. Mayer J.M. et al.Early prediction of severity in acute pancreatitis by urinary trypsinogen activation peptide: a multicentre study.Lancet. 2000; 355: 1955-1960Abstract Full Text Full Text PDF PubMed Scopus (324) Google Scholar SAP often leads to necrotizing pancreatitis—the severest form of pancreatitis—which results from the pancreatic ischemia caused by vasospasm.3Lankisch P.G. Warnecke B. Bruns D. et al.The APACHE II score is unreliable to diagnose necrotizing pancreatitis on admission to hospital.Pancreas. 2002; 24: 217-222Crossref PubMed Scopus (66) Google Scholar, 4Takeda K. Mikami Y. Fukuyama S. et al.Pancreatic ischemia associated with vasospasm in the early phase of human acute necrotizing pancreatitis.Pancreas. 2005; 30: 40-49Crossref PubMed Scopus (12) Google Scholar Moreover, pancreatic necrosis often results in the development of pseudocyst, and the patients with infected pseudocyst show a severe prognosis. Therefore, to improve the mortality rate associated with SAP, early and accurate detection of pancreatic ischemia and prediction of pancreatic necrosis are essential. Perfusion computerized tomography (CT) can differentiate the reversible from irreversible ischemic tissue in brain, and thus has been used widely for diagnosing acute brain stroke.5Latchaw R.E. Yonas H. Hunter G.J. et al.Guidelines and recommendations for perfusion imaging in cerebral ischemia: a scientific statement for healthcare professionals by the writing group on perfusion imaging, from the Council on Cardiovascular Radiology of the American Heart Association.Stroke. 2003; 34: 1084-1104Crossref PubMed Scopus (263) Google Scholar Perfusion CT is visualized by using the hemodynamic data, which is obtained from analysis of blood flow after a bolus injection of less contrast materials than those used for conventional contrast-enhanced CT.6Wintermark M. Maeder P. Verdun F.R. et al.Using 80 kVp versus 120 kVp in perfusion CT measurement of regional cerebral blood flow.AJNR Am J Neuroradiol. 2000; 21: 1881-1884PubMed Google Scholar, 7Axel L. Cerebral blood flow determination by rapid-sequence computed tomography: theoretical analysis.Radiology. 1980; 137: 679-686PubMed Google Scholar, 8Axel L. Tissue mean transit time from dynamic computed tomography by a simple deconvolution technique.Invest Radiol. 1983; 18: 94-99Crossref PubMed Scopus (218) Google Scholar Accordingly, it is tempting to expect that perfusion CT can be applied for detecting pancreatic ischemia and predicting development of pancreatic necrosis in patients with SAP. In the present study, therefore, to evaluate whether perfusion CT accurately can detect the ischemic change of SAP and predict the development of pancreatic necrosis, we prospectively analyzed the outcomes of the 30 consecutive patients with SAP in whom perfusion CT was performed, and evaluated the usefulness of perfusion CT. Patients with a clinical diagnosis of acute pancreatitis who were admitted to the emergency room at Kurashiki Central Hospital within 3 days of onset of symptoms were eligible for the study. The diagnosis of acute pancreatitis was based on physical findings of acute abdominal pain, laboratory findings of abnormal increases in serum amylase and serum lipase levels (more than 3 times higher than normal), and radiographic (ie, unenhanced CT) findings of pancreatitis.9Dervenis C. Johnson C.D. Bassi C. et al.Diagnosis, objective assessment of severity, and management of acute pancreatitis: Santorini Consensus Conference.Int J Pancreatol. 1999; 25: 195-210PubMed Google Scholar Patients whose acute pancreatitis was related to cancer or was a side effect of endoscopic retrograde cholangiopancreatography were excluded from the study. The enrollment period was 1 year, from October 2004 to November 2005. After the diagnosis of acute pancreatitis was confirmed, the severity of his or her condition was classified according to the criteria of the Acute Physiology and Chronic Health Evaluation (APACHE II) scoring system.9Dervenis C. Johnson C.D. Bassi C. et al.Diagnosis, objective assessment of severity, and management of acute pancreatitis: Santorini Consensus Conference.Int J Pancreatol. 1999; 25: 195-210PubMed Google Scholar SAP was diagnosed according to the Atlanta criterion10Bradley 3rd, E.L. A clinically based classification system for acute pancreatitis Summary of the International Symposium on Acute Pancreatitis, Atlanta, GA, September 11–13, 1992.Arch Surg. 1993; 128: 586-590Crossref PubMed Scopus (2487) Google Scholar and an APACHE II score greater than 6.9Dervenis C. Johnson C.D. Bassi C. et al.Diagnosis, objective assessment of severity, and management of acute pancreatitis: Santorini Consensus Conference.Int J Pancreatol. 1999; 25: 195-210PubMed Google Scholar All SAP patients were treated with fluid resuscitation to maintain the patients' central venous pressure greater than 8 mm Hg. In addition, they were treated with protease inhibitor and prophylactic antibiotics (imipenem). Patients who did not have a severe ileus were given enteral feedings through a nasojejunal tube. This program was approved by the Ethics Committee of Kurashiki Central Hospital. After the patients were diagnosed with SAP in the emergency room and had given written informed consent, they were referred for perfusion CT of the pancreas with a 16-slice helical CT scanner (Aquillion 16; Toshiba Medical Systems Co., Tokyo, Japan). After obtaining transaxial images of the upper abdomen, we determined 4 consecutive slices of 8-mm–thick sections that would cover as much of the pancreas as possible. Perfusion CT was performed after a bolus injection of 40 mL iodixanol at a rate of 4 mL/s via cubital vein and the 4 consecutive images were acquired every second for 48 seconds (Figure 1).11Hoeffner E.G. Case I. Jain R. et al.Cerebral perfusion CT: technique and clinical applications.Radiology. 2004; 231: 632-644Crossref PubMed Scopus (310) Google Scholar To visualize pancreatic blood flow (mL/100 g/min) and blood volume (mL/100 g), we used the deconvolution method as described previously.12Wintermark M. Fischbein N.J. Smith W.S. et al.Accuracy of dynamic perfusion CT with deconvolution in detecting acute hemispheric stroke.AJNR Am J Neuroradiol. 2005; 26: 104-112PubMed Google Scholar, 13Bize P.E. Platon A. Becker C.D. et al.Perfusion measurement in acute pancreatitis using dynamic perfusion MDCT.AJR Am J Roentgenol. 2006; 186: 114-118Crossref PubMed Scopus (41) Google Scholar, 14Abe H. Murakami T. Kubota M. et al.Quantitative tissue blood flow evaluation of pancreatic tumor: comparison between xenon CT technique and perfusion CT technique based on deconvolution analysis.Radiat Med. 2005; 23: 364-370PubMed Google Scholar, 15Wintermark M. Thiran J.P. Maeder P. et al.Simultaneous measurement of regional cerebral blood flow by perfusion CT and stable xenon CT: a validation study.AJNR Am J Neuroradiol. 2001; 22: 905-914PubMed Google Scholar This method is used to calculate blood flow and volume in a target organ by injecting contrast material (iodixanol). It now is used widely to assess brain ischemia during acute stroke. All perfusion CT scans were analyzed by using an imaging workstation with commercial perfusion CT analysis software (Toshiba box-MTF) to create color maps of pancreatic blood flow and pancreatic blood volume. The software required placement of small regions of interest (ROIs) on one artery and one vein to generate arterial input functions and venous outflow functions, respectively, for the deconvolution method.8Axel L. Tissue mean transit time from dynamic computed tomography by a simple deconvolution technique.Invest Radiol. 1983; 18: 94-99Crossref PubMed Scopus (218) Google Scholar, 14Abe H. Murakami T. Kubota M. et al.Quantitative tissue blood flow evaluation of pancreatic tumor: comparison between xenon CT technique and perfusion CT technique based on deconvolution analysis.Radiat Med. 2005; 23: 364-370PubMed Google Scholar This process is required because the computer algorithm for deconvolution method is used to compare the shape and height of the time-density curve of each pixel of the CT time series with shape and height of the arterial and venous time-density curves to determine blood flow and volume. We used the abdominal aorta as the input marker and the portal vein as the outflow marker (Figure 2).16Tsuji Y. Watanabe Y. Matsueda K. et al.Usefulness of perfusion computed tomography for early detection of pancreatic ischemia in severe acute pancreatitis.J Gastroenterol Hepatol. 2006; 21: 1506-1508Crossref PubMed Scopus (5) Google Scholar To display the blood flow and blood volume as color maps, the data from the deconvolution calculations were obtained from these processes and converted into color maps. After the data were converted into color maps, the ROIs (circle or oval shape ∼ 20–30 mm in diameter) were established in the most darkly colored area of the pancreas and in the most homogeneously colored area of the liver. Arteries and veins were excluded from the ROIs as much as possible. When the color map of the liver was not homogeneous, data were obtained from 3 ROIs in the most darkly colored area and averaged. Quantitative measurements of relative pancreatic blood flow (PBF), relative pancreatic blood volume (PBV), relative hepatic blood flow (HBF), and relative hepatic blood volume (HBV) then were obtained (Figure 3). To effectively diagnose pancreatic ischemia with perfusion CT, the average PBF (38.4 ± 12.0 mg/100 g/min), PBV (9.94 ± 2.48 mg/100 g), HBF (22.6 ± 3.34 mg/100 g/min), and HBV (6.46 ± 0.45 mg/100 g) were calculated from 5 healthy control subjects (mean age, 28.2 y; 4 males, 1 female). Perfusion CT showed that, in normal subjects, the PBF is always faster than the HBF (Figure 4A) and the PBV is always greater than the HBV (Figure 4D). Accordingly, we considered pancreatic ischemia to be present when the PBF (Figure 5B,arrowheads) was lower than the HBF (Figure 5B, arrow) or when the PBV (Figure 5C, arrowheads) was less than the HBV (Figure 5C, arrow). Conversely, we considered pancreatic ischemia to be absent when the hemodynamic color maps showed that the pancreas was colored more brightly than that of the liver (Figure 5F and G).Figure 5Perfusion CT of representative patients. (A) Dynamic CT showed grade C pancreatitis in patient 7 (Table 1). The density of the pancreas was increased to 47 Hounsfield units and ANP was ruled out. (B and C) Perfusion CT showed that the (B, arrowheads) PBF was slower than the (B, arrows) HBF and that the (C, arrowheads) PBV was less than the (C, arrows) HBV, indicating pancreatic ischemia. (D) Pancreatic necrosis detected by contrast-enhanced CT 3 weeks later. (E) Dynamic CT also showed grade C pancreatitis in patient 28 (Table 1). The density of the pancreas was high, and AEP was diagnosed. (F and G) Similar to patient 7, perfusion CT showed that PBF and PBV were faster and more than HBF and HBF, respectively. (H) However, no evidence of pancreatic necrosis was disclosed by contrast-enhanced CT 3 weeks later in this patient.View Large Image Figure ViewerDownload Hi-res image Download (PPT) The final diagnosis of acute edematous pancreatitis (AEP) or acute necrotizing pancreatitis (ANP) was made by performing conventional contrast-enhanced CT 3 weeks after the onset of symptoms. ANP was diagnosed specifically when pancreatic necrosis was confirmed in the same region in which pancreatic ischemia had been disclosed by the hemodynamic color maps (Figure 5). All statistical analyses were performed by using SPSS 11 (SPSS Inc., Chicago, IL). Differences between groups were analyzed by using the Mann–Whitney U test. P values less than .05 were considered to be statistically significant. A total of 30 patients, 11 females and 19 males ranging in age from 15 to 90 years (mean, 58 y), were diagnosed as having SAP and enrolled in the study. Table 1 summarizes the characteristics of the study group, including an average of maximum C-reactive protein level during the clinical course (average of maximum ± SD = 17.4 ± 8.9 mg/dL), APACHE II scores17Larvin M. McMahon M.J. APACHE-II score for assessment and monitoring of acute pancreatitis.Lancet. 1989; 2: 201-205Abstract PubMed Scopus (550) Google Scholar (average ± SD = 11.5 ± 5.3), and CT grades.18Balthazar E.J. Acute pancreatitis: assessment of severity with clinical and CT evaluation.Radiology. 2002; 223: 603-613Crossref PubMed Scopus (558) Google ScholarTable 1Clinical Characteristics of Patients With SAPCaseSexAge, yEtiology of pancreatitisPancreatic ischemiaC-reactive protein,aThe maximum value of measurements taken for each patient throughout the clinical course. mg/dLAPACHE II scorebThe scoring and grading of the pancreatitis were determined at the time of admission to the emergency room.CT gradebThe scoring and grading of the pancreatitis were determined at the time of admission to the emergency room.1M34Alcoholism+8.56C2M71Alcoholism+20.07C3M74Alcoholism+28.014C4M81OthercAcute episodes of pancreatitis in patients with chronic pancreatitis or hyperlipidemia.+14.017D5M44Alcoholism+17.06D6F74Gallstone+21.011E7M54Alcoholism+26.010C8F76OthercAcute episodes of pancreatitis in patients with chronic pancreatitis or hyperlipidemia.+25.014D9dThis patient died of complications of multiple organ failure with systemic infection after the study.F87Gallstone+29.022E10M53Alcoholism+16.07B11M58OthercAcute episodes of pancreatitis in patients with chronic pancreatitis or hyperlipidemia.–20.09B12M73Gallstone–24.019C13M23Idiopathic–4.47D14F53Alcoholism–24.030D15M45Alcoholism–19.09C16M43Alcoholism–37.07C17F77Idiopathic–12.013B18F26Alcoholism–0.912C19M54Idiopathic–7.311D20F15OthercAcute episodes of pancreatitis in patients with chronic pancreatitis or hyperlipidemia.–8.47B21M50Idiopathic–7.710C22M57Alcoholism–19.011B23M61Idiopathic–19.08E24F54OthercAcute episodes of pancreatitis in patients with chronic pancreatitis or hyperlipidemia.–13.07D25M50Gallstone–25.011E26F69Gallstone–0.610C27F90Gallstone–16.016C28M42Idiopathic–26.08C29M70Gallstone–9.615C30F82Gallstone–25.012DM, male; +, positive for pancreatic ischemia by perfusion CT; F, female; –, negative for pancreatic ischemia by perfusion CT.a The maximum value of measurements taken for each patient throughout the clinical course.b The scoring and grading of the pancreatitis were determined at the time of admission to the emergency room.c Acute episodes of pancreatitis in patients with chronic pancreatitis or hyperlipidemia.d This patient died of complications of multiple organ failure with systemic infection after the study. Open table in a new tab M, male; +, positive for pancreatic ischemia by perfusion CT; F, female; –, negative for pancreatic ischemia by perfusion CT. Only 1 of the patients (Table 1, patient 9) died from systemic infection. Perfusion CT12Wintermark M. Fischbein N.J. Smith W.S. et al.Accuracy of dynamic perfusion CT with deconvolution in detecting acute hemispheric stroke.AJNR Am J Neuroradiol. 2005; 26: 104-112PubMed Google Scholar, 13Bize P.E. Platon A. Becker C.D. et al.Perfusion measurement in acute pancreatitis using dynamic perfusion MDCT.AJR Am J Roentgenol. 2006; 186: 114-118Crossref PubMed Scopus (41) Google Scholar, 14Abe H. Murakami T. Kubota M. et al.Quantitative tissue blood flow evaluation of pancreatic tumor: comparison between xenon CT technique and perfusion CT technique based on deconvolution analysis.Radiat Med. 2005; 23: 364-370PubMed Google Scholar, 15Wintermark M. Thiran J.P. Maeder P. et al.Simultaneous measurement of regional cerebral blood flow by perfusion CT and stable xenon CT: a validation study.AJNR Am J Neuroradiol. 2001; 22: 905-914PubMed Google Scholar was performed an average of 1.7 ± 0.2 (mean ± SD) days after the onset of acute pancreatitis and disclosed pancreatic ischemia in 10 of the 30 patients. On the other hand, contrast-enhanced CT taken 3 weeks after the onset of SAP showed that AEP and ANP finally were diagnosed in 21 and 9 patients, respectively (Table 2). Consequently, of 10 patients with pancreatic ischemia detected by perfusion CT, 9 patients developed ANP, and none of 20 patients without pancreatic ischemia developed ANP (Table 2). In patient 9, pathologic findings confirmed the pancreatic necrosis, the position of which corresponded to that of pancreatic ischemia and that of pancreatic necrosis detected by perfusion CT and contrast-enhanced CT, respectively (Figure 6). Of 8 patients with ANP who survived, all 8 patients developed pseudocyst later, and thus they had significantly longer hospital stays (Table 3).Table 2Sensitivity and Specificity of Perfusion CT for Prediction of Development of ANPTotalAEPANPSensitivity, %Specificity, %Pancreatic ischemia detected by perfusion CT(+)1019(−)20200Total3021910095.3 Open table in a new tab Table 3Comparison of Outcomes of SAP Patients With Pancreatic IschemiaPancreatic ischemia detected by perfusion CTP valueNegative (n = 20)Positive (n = 10)Pancreatic necrosis09.01Pseudocyst08.01Infected pseudocyst05.03Hospital days23.6 ± 10.651.7 ± 56.7.02Death01.66NOTE. Infected pseudocyst was diagnosed by culture of juice obtained from pseudocysts. P values less than .05 are considered to be statistically significant. Open table in a new tab NOTE. Infected pseudocyst was diagnosed by culture of juice obtained from pseudocysts. P values less than .05 are considered to be statistically significant. Based on these findings, the sensitivity and specificity of perfusion CT for predicting ANP was calculated to be 100% and 95.3%, respectively (Table 2). No significant differences were found between the patients with AEP and those with ANP with regard to sex, age, etiology of pancreatitis, severity of disease, days from onset to hospital admission, and treatment type (Table 4). Moreover, hematocrit, blood urea nitrogen, blood sugar, serum C-reactive protein, elastase-1, trypsinogen, interleukin-6 levels, and systematic inflammatory response syndrome score on admission day were not significantly different between patients with AEP and ANP patients.Table 4Comparison of Selected Characteristics of Patients With AEP or ANPAEP (n = 21)ANP (n = 9)P valueSex.84 Male136 Female83AgeaResults are expressed as means ± SD.55 ± 1966 ± 18.13Etiology.27 Alcoholism65 Gallstone62 Idiopathic60 Other32C-reactive protein level, mg/dLaResults are expressed as means ± SD., bThe maximum value of measurements taken for patients throughout the clinical course.15.8 ± 9.421.0 ± 6.9.09APACHE II score on admissionaResults are expressed as means ± SD.11.4 ± 5.411.9 ± 5.4.87CT gradeaResults are expressed as means ± SD.3.2 ± 0.93.8 ± 0.8.17Days from onset of symptoms to admissionaResults are expressed as means ± SD.1.6 ± 0.91.8 ± 1.1.83Treatment.49 Ventilation23 Apheresis21NOTE. P values less than .05 are considered to be statistically significant.a Results are expressed as means ± SD.b The maximum value of measurements taken for patients throughout the clinical course. Open table in a new tab NOTE. P values less than .05 are considered to be statistically significant. In all 5 healthy control subjects, the PBF was faster than the HBF, and the PBV was greater than the HBV (Figure 4A and D). The average PBF was faster than the HBF, the average PBV was greater than the HBV, and the differences between the PBF and HBF values and the PBV and HBV values were statistically significant (P < .05; Table 5). In the 21 patients who finally were diagnosed with AEP, the average PBF also was faster than the HBF and the average PBV was greater than HBV; the differences between these values also were statistically significant (P < .01; Figure 4B and E; Table 5). There were no statistically significant differences of the PBF as well as PBV values between the 21 patients with AEP and the 5 healthy control subjects (Figure 4A, B, D, and E; Table 5). As discussed earlier, only 1 of the 21 patients with AEP (Table 1, patient 10) had pancreatic ischemia detected by perfusion CT (Figure 4B and E, open circle; Table 2), but his PBF and PBV values were similar to his HBF and HBV values, respectively. In contrast, in all 9 subjects with ANP, the PBF and PBV values for each subject were lower than their respective HBF and HBV values as shown by perfusion CT (Figure 4C and F), and the averaged PBF and PBV values were significantly lower than the respective averaged HBF and HBV values for this subgroup (P < .01; Table 5). Statistical significance also was found between the average PBF and PBV values for the patients with ANP and the PBF and PBV values of both the control subjects and the patients with AEP, but not for the average HBF and HBV values (P < .01, Table 5).Table 5Comparison of Blood Flow and Volume Among Control Subjects and Patients With AEP or ANPHealthy control subjects (n = 5)Patients with AEP (n = 21)Patients with ANP (n = 9)PBF (mg/100 g/min)38.4 ± 12.0aP < .05 vs respective mean HBF value.57.3 ± 55.9bP < .01 vs respective mean HBF value.10.9 ± 2.0bP < .01 vs respective mean HBF value., cP < .01 vs the mean PBFs of both the control subjects and the patients with AEP.HBF (mg/100 g/min)22.6 ± 3.3425.2 ± 14.323.5 ± 7.8PBV (mg/100 g)9.94 ± 2.48dP < .05 vs respective mean HBV value.14.5 ± 11.1eP < .01 vs respective mean HBV value.3.5 ± 0.9eP < .01 vs respective mean HBV value., fP < .01 vs the mean PBVs of both the control subjects and the patients with AEP.HBV (mg/100 g)6.46 ± 0.456.3 ± 3.56.7 ± 2.7NOTE. All results are expressed as the mean ± SD.a P < .05 vs respective mean HBF value.b P < .01 vs respective mean HBF value.c P < .01 vs the mean PBFs of both the control subjects and the patients with AEP.d P < .05 vs respective mean HBV value.e P < .01 vs respective mean HBV value.f P < .01 vs the mean PBVs of both the control subjects and the patients with AEP. Open table in a new tab NOTE. All results are expressed as the mean ± SD. Development of pancreatic necrosis is the critical event of acute pancreatitis that determines the prognosis of the patients because it often is accompanied by infection and multiple organ dysfunction syndromes.19Isenmann R. Rau B. Zoellner U. et al.Management of patients with extended pancreatic necrosis.Pancreatology. 2001; 1: 63-68Abstract Full Text PDF PubMed Scopus (28) Google Scholar The overall mortality rate of acute pancreatitis is reported to be between 2.1% and 9.2% worldwide. In Japan, pancreatic necrosis occurs in 10%–15% of patients with SAP, with a mortality rate of 23%.1Sekimoto M. Takada T. Kawarada Y. et al.JPN guidelines for the management of acute pancreatitis: epidemiology, etiology, natural history, and outcome predictors in acute pancreatitis.J Hepatobiliary Pancreat Surg. 2006; 13: 10-24Crossref PubMed Scopus (177) Google Scholar This rate is nearly twice that for patients with SAP who do not develop pancreatic necrosis (ie, 11%).1Sekimoto M. Takada T. Kawarada Y. et al.JPN guidelines for the management of acute pancreatitis: epidemiology, etiology, natural history, and outcome predictors in acute pancreatitis.J Hepatobiliary Pancreat Surg. 2006; 13: 10-24Crossref PubMed Scopus (177) Google Scholar According to a Japanese study,1Sekimoto M. Takada T. Kawarada Y. et al.JPN guidelines for the management of acute pancreatitis: epidemiology, etiology, natural history, and outcome predictors in acute pancreatitis.J Hepatobiliary Pancreat Surg. 2006; 13: 10-24Crossref PubMed Scopus (177) Google Scholar infection develops in 30%–40% of patients with ANP and substantially increases the mortality rate (34% for infected patients vs 7% for uninfected patients). Organ failure increases the mortality rate among patients with ANP even higher, reaching 47%.1Sekimoto M. Takada T. Kawarada Y. et al.JPN guidelines for the management of acute pancreatitis: epidemiology, etiology, natural history, and outcome predictors in acute pancreatitis.J Hepatobiliary Pancreat Surg. 2006; 13: 10-24Crossref PubMed Scopus (177) Google Scholar To improve the overall prognosis of SAP, therefore, it is important to predict the development of necrosis early and to treat patients intensively as soon as possible. By performing perfusion CT of the pancreas within 3 days after onset of symptoms, we were able to detect pancreatic ischemia at a very early stage in 10 of 30 patients with SAP. Importantly, both the sensitivity and specificity of perfusion CT for predicting the development of necrotizing pancreatitis were excellent, reaching 100% and 95.3%, respectively. Indeed, 9 of 10 patients who were diagnosed as having pancreatic ischemia by perfusion CT within 3 days after the onset of the disease finally developed necrotizing pancreatitis as diagnosed by contrast-enhanced CT, and none of 20 patients who were revealed not to have pancreatic ischemia developed necrotizing pancreatitis. In the present study, pancreatic necrosis was confirmed histologically only in 1 patient, but in other patients it was diagnosed by contrast-enhanced CT alone. However, CT for detecting pancreatic necrosis is well established.20British Society of GastroenterologyUnited Kingdom guidelines for the management of acute pancreatitis.Gut. 1998; 42: S1-S13PubMed Google Scholar In this study, except for 1 patient who died as described earlier, all the remaining 8 patients with pancreatic necrosis diagnosed by contrast-enhanced CT developed pseudocyst. Lehman21Lehman G.A. Pseudocysts.Gastrointest Endosc. 1999; 49: S81-S84Abstract Full Text Full Text PDF PubMed Scopus (53) Google Scholar reported that there is a significant overlap between pancreatic necrosis and pseudocyst formation. Other investigators have reported that dynamic contrast-enhanced CT22Kivisaari L. Somer K. Standertskjold-Nordenstam C.G. et al.Early detection of acute fulminant pancreatitis by contrast-enhanced computed tomography.Scand J Gastroenterol. 1983; 18: 39-41Crossref PubMed Scopus (84) Google Scholar, 23Block S. Maier W. Bittner R. et al.Identification of pancreas necrosis in severe acute pancreatitis: imaging procedures versus clinical staging.Gut. 1986; 27: 1035-1042Crossref PubMed Scopus (160) Google Scholar is more accurate than either the Ranson criteria for pancreatitis mortality24Leung T.K. Lee C.M. Lin S.Y. et al.Balthazar computed tomography severity index is superior to Ranson criteria and APACHE II scoring system in predicting acute pancreatitis outcome.World J Gastroenterol. 2005; 11: 6049-6052PubMed Google Scholar or the APACHE II scoring system in predicting the development of pancreatic necrosis, but the accuracy of this imaging technique in predicting necrosis at an early stage of severe acute pancreatitis is not satisfactory.2Neoptolemos J.P. Kemppainen E.A. Mayer J.M. et al.Early prediction of severity in acute pancreatitis by urinary trypsinogen activation peptide: a multicentre study.Lancet. 2000; 355: 1955-1960Abstract Full Text Full Text PDF PubMed Scopus (324) Google Scholar, 25Johnson C.D. Abu-Hilal M. Persistent organ failure during the first week as a marker of fatal outcome in acute pancreatitis.Gut. 2004; 53: 1340-1344Crossref PubMed Scopus (436) Google Scholar The United Kingdom guidelines for the management of acute pancreatitis,20British Society of GastroenterologyUnited Kingdom guidelines for the management of acute pancreatitis.Gut. 1998; 42: S1-S13PubMed Google Scholar the most popular clinical guideline of acute pancreatitis, pointed low sensitivity of contrast-enhanced CT to diagnosis pancreatic necrosis in early stage of SAP. Takeda et al4Takeda K. Mikami Y. Fukuyama S. et al.Pancreatic ischemia associated with vasospasm in the early phase of human acute necrotizing pancreatitis.Pancreas. 2005; 30: 40-49Crossref PubMed Scopus (12) Google Scholar also reported that dynamic contrast-enhanced CT diagnosed pancreatic necrosis only in 38 (37.2%) of 102 patients in the early stages of ANP. Within 3 days after onset of SAP, pancreatic necrosis may be incomplete, and contrast-enhanced CT could diagnose ANP accurately as time goes by.26Larvin M. Chalmers A.G. McMahon M.J. Dynamic contrast enhanced computed tomography: a precise technique for identifying and localising pancreatic necrosis.BMJ. 1990; 300: 1425-1428Crossref PubMed Scopus (85) Google Scholar, 27London N.J. Leese T. Lavelle J.M. et al.Rapid-bolus contrast-enhanced dynamic computed tomography in acute pancreatitis: a prospective study.Br J Surg. 1991; 78: 1452-1456Crossref PubMed Scopus (94) Google Scholar For this reason, the UK guideline recommended that contrast-enhanced CT should be performed at day 3 or later after onset of SAP. However, this waiting time may delay induction of intensive care for SAP, and may influence its prognosis.28Traverso L.W. Kozarek R.A. Pancreatic necrosectomy: definitions and technique.J Gastrointest Surg. 2005; 9: 436-439Crossref PubMed Scopus (75) Google Scholar Prophylactic antibacterial agents have been used to prevent the complication of infectious pancreatic necrosis.29Isenmann R. Runzi M. Kron M. et al.Prophylactic antibiotic treatment in patients with predicted severe acute pancreatitis: a placebo-controlled, double-blind trial.Gastroenterology. 2004; 126: 997-1004Abstract Full Text Full Text PDF PubMed Scopus (414) Google Scholar However, in the necrotic tissue, antibacterial agents could not keep enough density30Buchler M. Malfertheiner P. Friess H. et al.Human pancreatic tissue concentration of bactericidal antibiotics.Gastroenterology. 1992; 103: 1902-1908PubMed Google Scholar to achieve an effect, thereby the effect was limited. If antibacterial agents would have been given to the ischemic tissue before developing necrosis, the antibacterial agents could have kept adequate density to prevent the infection and would have changed the prognosis of SAP. Moreover, early administration of anticoagulation therapy may prevent the development of pancreatic necrosis.31Hackert T. Werner J. Uhl W. et al.Reduction of ischemia/reperfusion injury by antithrombin III after experimental pancreas transplantation.Am J Surg. 2005; 189: 92-97Abstract Full Text Full Text PDF PubMed Scopus (29) Google Scholar Graft pancreatitis occurring after pancreatic transplantation often is caused by pancreatic ischemia, and graft pancreatitis is one of the most severe complications of pancreatic transplantation. Hackert et al31Hackert T. Werner J. Uhl W. et al.Reduction of ischemia/reperfusion injury by antithrombin III after experimental pancreas transplantation.Am J Surg. 2005; 189: 92-97Abstract Full Text Full Text PDF PubMed Scopus (29) Google Scholar reported that anti–thrombin III prevents graft pancreatitis from pancreatic ischemia. Several researchers also reported that continuous regional arterial infusion of nafamostat mesylate, an antiprothrombin agent, prevents developing pancreatic necrosis in early stage pancreatitis.32Takeda K. Matsuno S. Sunamura M. et al.Continuous regional arterial infusion of protease inhibitor and antibiotics in acute necrotizing pancreatitis.Am J Surg. 1996; 171: 394-398Abstract Full Text PDF PubMed Scopus (152) Google Scholar, 33Nakase H. Itani T. Mimura J. et al.Successful treatment of severe acute pancreatitis by the combination therapy of continuous arterial infusion of a protease inhibitor and continuous hemofiltration.J Gastroenterol Hepatol. 2001; 16: 944-945PubMed Google Scholar Taken together, it is possible to improve the prognosis of SAP by diagnosing pancreatic ischemia at an early stage. In addition, we were unable to show in our study that both biochemical markers tested and APACHE II scores could predict necrotizing pancreatitis accurately. Thus, at present, perfusion CT appears to be superior to other clinical markers, including dynamic contrast-enhanced CT results, for predicting necrotizing pancreatitis. Perfusion CT also is technically superior to dynamic CT as a predictive tool. Specifically, perfusion CT evaluates changes in blood flow continuously on the selected 4 serial longitudinal slices, which covered the pancreatic tissue as much as possible. Although Isenmann et al34Isenmann R. Buchler M. Uhl W. et al.Pancreatic necrosis: an early finding in severe acute pancreatitis.Pancreas. 1993; 8: 358-361Crossref PubMed Scopus (108) Google Scholar reported the usefulness of dynamic CT for early detection of pancreatic necrosis, dynamic CT evaluates the blood flow at only 1 or 2 time points in a single CT image (ie, slice). Because the flow of contrast materials into the pancreatic circulation changes serially and dynamically, it is almost impossible to evaluate pancreatic blood flow from dynamic CT images that represent only 1 or 2 time points. Moreover, ROIs were established in the imaged pancreatic tissue that showed the slowest blood flow and the least blood volume on the perfusion CT color map. These technical aspects of perfusion CT enhance its ability to detect pancreatic ischemia. It is noteworthy that there were a considerable overlap of PBF values and PBV values between the patients with AEP and those with ANP in our study (Figure 4). Thus, by evaluating only PBF or PBV, pancreatic ischemia may not be evaluated accurately. To overcome such a drawback, we used measurement of blood flow of not only the pancreas but also the liver parenchyma as a control organ, and compared the blood flow of the 2 organs. By doing so, we achieved greater specificity and sensitivity with perfusion CT for detecting pancreatic ischemia. We acknowledge, however, that we achieved this greater sensitivity and specificity because the HBF and HBV values were unaffected by the underlying pancreatitis. It has been reported that contrast media may deteriorate pancreatitis by inducing microcirculation damage. In this respect, we also want to point out that perfusion CT required significantly less contrast media (40 mL) than is used for conventional dynamic CT (80–100 mL), thereby reducing the potential of side effects35Werner J. Schmidt J. Warshaw A.L. et al.The relative safety of MRI contrast agent in acute necrotizing pancreatitis.Ann Surg. 1998; 227: 105-111Crossref PubMed Scopus (28) Google Scholar (eg, nephrotoxicity) from the contrast media. Perfusion magnetic resonance imaging (MRI) may be an alternative to perfusion CT for detecting pancreatic necrosis. Indeed, MRI has more advantages when compared with conventional CT, including greater sensitivity in detecting ischemia. Moreover, when magnetic resonance angiography is performed in conjunction with MRI, assessment of vessel patency also is possible.36Yamada I. Aung W. Himeno Y. et al.Diffusion coefficients in abdominal organs and hepatic lesions: evaluation with intravoxel incoherent motion echo-planar MR imaging.Radiology. 1999; 210: 617-623Crossref PubMed Scopus (475) Google Scholar Despite these advantages, MRI is underused in the evaluation of patients with SAP. This underutilization can be attributed to practical considerations. Patients with SAP have difficulty withstanding the lengthy MRI procedure. Moreover, all magnetic instruments (eg, sphygmomanometer, saturation monitor, multiple infusion pumps, and so forth), which cannot be taken away from patients with SAP even for a short period, must be removed before the procedure is performed. In emergency situations, perfusion CT is, therefore, a more useful and desirable technique. In summary, we have shown clearly that perfusion CT is very useful in detecting ischemia in the early stages of SAP. We believe that the routine use of perfusion CT will permit more accurate prediction of ANP and may help improve the prognosis of patients with SAP. However, because a considerable number of SAP patients without pancreatic necrosis also die at an early stage as a result of multiorgan failure, further study is required to find better prognostic markers.