Assessment of selective arterial calcium stimulation and hepatic venous sampling to localize insulin-secreting tumours

OBJECTIVE: Non-invasive localization modalities such as ultrasound, computed tomography (CT) or magnetic resonance imaging (MRI) often fail to localize insulinomas smaller than 2 cm in diameter. Recent studies have shown that the selective arterial stimulation and hepatic venous sampling (ASVS) technique using intra-arterial calcium as the insulin secretagogue facilitates the regionalization of such occult insulinomas. This study assesses the sensitivity of ASVS in localizing insulin-secreting tumours. SUBJECTS AND METHODS: Eleven consecutive patients (8 women), aged 29-82 years, were studied over the past 4 years at our hospital. Hyperinsulinaemic hypoglycaemia due to an insulin-secreting tumour was proven in all patients. Calcium gluconate (0.025 mEq/kg body weight) was injected directly into the arteries supplying the pancreas and the liver. Insulin levels were measured in samples taken from the right hepatic vein before and 30, 60 and 120 s after each injection. The ASVS technique was performed in all 11 patients; the results were compared with the surgical findings in 10 patients and the autopsy findings in 1 case. The ASVS results were also compared with the findings of other, previously performed imaging modalities. RESULTS: ASVS correctly localized 4 insulin-secreting tumours to the head, 3 to the body, 1 to the tail, 2 to the tail or body of the pancreas and 1 to the liver. Thus, the sensitivity was 100% (11/11) whereas other localization techniques were less sensitive: 7/11 tumours were detected by angiography, 4/8 by endosonography, 3/8 by CT and 1/6 by MRI. Insulinomas (confirmed by histological examination), sized 4-25 mm, were found in 10 patients. All were cured by selective surgery and remained free of hypoglycaemia over the next 1-4 years of follow-up. An insulin-secreting neuroendocrine tumour in the liver was documented in 1 case at autopsy. CONCLUSIONS: Arterial stimulation and hepatic venous sampling is a very sensitive technique for preoperative localization of insulin-producing tumours. It can help to plan minimally invasive surgery and to select an appropriate strategy for patients suffering from malignant tumours in others.     

Insulinoma: Selective arterial calcium injection performed to confirm localization and complete resection

A case of insulinoma is reported in a patient in whom selective arterial calcium injection (SACI) tests were performed both to confirm tumor localization before surgery and to confirm complete tumor removal during surgery. An 18-year-old woman with hypoglycemic episodes was diagnosed with an insulinoma in the pancreatic body demonstrated by celiac arteriography. In a preoperative SACI test, calcium was injected into the splenic artery (SpA), gastroduodenal artery (GDA), and superior mesenteric artery (SMA). Serum immunoreactive insulin (IRI) and proinsulin levels were measured in hepatic venous samples. IRI was markedly increased after the injection of calcium into the GDA and SMA, while there was no response in IRI levels when calcium was injected into the SpA. Therefore, no occult insulinoma was revealed in the distal area fed by the SpA, although the presence of insulinoma was uncertain in the proximal pancreas. In the intraoperative SACI test, calcium was injected into the celiac artery. Insulin (determined by enzyme immunoassay) and proinsulin levels were measured in portal venous samples before and after resection of the tumor. After resection, these levels decreased in response to the calcium stimuli, confirming complete removal of the insulinoma. The SACI test was helpful to localize the insulinoma and was useful to confirm the complete removal of the tumor.     

Intra-arterial calcium stimulation test for detection of insulinomas: detection rate, responses of pancreatic peptides, and its relationship to differentiation of tumor cells

The selective intra-arterial calcium stimulation test has greatly facilitated the precise regionalization of insulinomas smaller than 2 cm, which noninvasive techniques (ultrasound [US], computed tomography [CT], magnetic resonance imaging [MRI]) often fail to localize. This study examined not only the role of the test in the localization of insulinomas, but also the responsiveness of 3 beta-cell peptides (insulin, C peptide, and proinsulin) and their relationship to the degree of differentiation of the tumor cells, using percentage decrease of both proinsulin/insulin (P/I) and proinsulin/C peptide (P/C) ratios after stimulation as indices. Ten consecutive surgically proven insulinoma patients each received an injection of calcium into the arteries supplying the pancreas after standard selective angiography and beta-cell peptide levels were measured in samples taken from the right hepatic vein before and 30, 60, 90, 120, and 180 seconds after each injection prior to operation. After surgery, the expressions of the calcium sensing receptor (CaSR) on the resected tumors were assessed by immunohistochemistry. Intra-arterial calcium stimulation with sampling either for insulin or for C peptide correctly predicted the site of insulinoma in 8 of 9 patients or in 7 of 8 patients if the 2 big malignant insulinomas were excluded; thus, the detection rate of this test was 89% and 88%, respectively. Calcium administration stimulated a marked and prompt release of insulin and C peptide simultaneously. Both peaked within 30 to 60 seconds, then declined gradually thereafter, remaining above the baseline at 180 seconds. The magnitude of increase correlated well with the corresponding percentage decrease of P/I and P/C ratios. The response of proinsulin was much less. Immunohistochemistry demonstrated variable membraneous staining for CaSR in normal pancreatic islets and in about 9% of the total normal beta cells, whereas staining in tumor cells was only minimally detectable. We conclude that selective intra-arterial calcium stimulation with hepatic venous sampling either for insulin or for C peptide is a highly sensitive method for the preoperative localization of small insulinomas. Calcium injection stimulates a brisk response of insulin, C peptide, and proinsulin simultaneously and the magnitude of increase of both insulin and C peptide appears to be correlated well with the degree of differentiation of the tumor cells. The exact mechanism by which calcium provokes the release of beta-cell peptides is less clear and whether the CaSR is involved in the mechanism of its action requires further study.     

Localization of gastroenteropancreatic tumours by angiography

Neuroendocrine tumours are seen on arteriography as diffusely enhancing masses without tumour vessels and without arteriovenous shunting. In 70 patients with surgically proven tumours, the sensitivity of angiography was 68% for extrapancreatic and 86% for hepatic lesions. Hepatic metastases have always been easier to demonstrate arteriographically than the primary tumour because of the absence of overlying bowel. Portal venous sampling is a sensitive technique for detecting functioning gastroenteropancreatic tumours. Sampling the small veins about the pancreatic head yielded a sensitivity of 62% but this is an invasive procedure in which considerable experience is required. Intra-arterial secretagogue, secretin for gastrinomas and calcium for insulinomas, selectively injected into the pancreatic and the hepatic arteries produce a diagnostic gastrin or insulin gradient respectively. The localization sensitivity of arterial stimulation with venous sampling is 77-89% for gastrinoma and 92% for pancreatic insulinoma. Recently, spiral CT in conjunction with selective intra-arterial rather than intravenous injection of contrast may increase the detection sensitivity of duodenal and pancreatic gastrinomas.     

Abnormalities of proinsulin processing in functioning insulinomas: clinical implications

OBJECTIVE: Abnormal proinsulin processing in insulinomas may result in secretory granules containing both insulin and proinsulin, a finding not encountered in healthy beta-cells. The aim of this study was to test whether such abnormalities in the proinsulin to insulin conversion have clinical implications in patients with hypoglycaemic disorders. DESIGN: Case-series. PATIENTS AND METHODS: Fifteen patients with histologically confirmed insulinoma and two patients with islet cell hyperplasia were included. The immunohistochemical distribution pattern of proinsulin within the tumour cells was classified as Golgi pattern (predominantly perinuclear immunolabelling) or diffuse pattern (immunolabelling in the periphery of the cells, indicating the presence of proinsulin in secretory granules). Data obtained from the 72-h fast and arterial calcium stimulation and hepatic venous sampling (ASVS) test were related to the morphological classification. RESULTS: Six insulinomas exhibited a diffuse proinsulin distribution pattern, while nine insulinomas and two islet cell hyperplasias disclosed a Golgi pattern. Median proinsulin concentrations at the termination of the fast tended to be higher in patients with the diffuse proinsulin distribution pattern than in patients with the Golgi pattern (86.9 vs. 18.8 pmol/l, P = 0.07). Higher insulin (P < 0.005) and proinsulin (P < 0.05) concentrations were significantly correlated with earlier occurrence of hypoglycaemia during the prolonged fast. During the ASVS test, tumours with the diffuse proinsulin distribution pattern exhibited a higher increase in both insulin (median, 37.3- vs. 10.5-fold, P < 0.05) and proinsulin (6.3- vs. 1.6 fold, P < 0.005) concentrations following calcium stimulation than the tumours with the Golgi pattern. CONCLUSIONS: Abnormalities in the proinsulin to insulin conversion in patients with insulinomas and islet cell hyperplasia correlate with impaired regulation of both insulin and proinsulin secretion during the prolonged fast as well as the ASVS test.     

ACCURATE LOCALIZATION OF INSULINOMA USING PERCUTANEOUS TRANSHEPATIC PORTAL VENOUS SAMPLING—USEFULNESS OF SIMULTANEOUS MEASUREMENT OF PLASMA INSULIN AND GLUCAGON LEVELS

By means of percutaneous transhepatic portal venous sampling (PTVS), plasma insulin and glucagon levels were determined simultaneously at various sites of the hepatic portal venous system in two patients with insulinoma and four patients without. In the two cases of insulinoma, an obvious rise of insulin concentration was observed at the vicinity of the tumour, while glucagon levels were not elevated in the blood samples showing an insulin peak. In the remaining four cases without evidence of insulinoma, significant step-ups of plasma insulin occurred in the splenic vein as well, but were accompanied with concomitant elevation of plasma glucagon levels in the same blood samples. Thus, by measuring insulin alone, false positive data may frequently be obtained when PTVS is performed on patients with suspected insulinomas. The simultaneous measurement of insulin and glucagon might be helpful in avoiding such errors.     

Endogenous hyperinsulinemic hypoglycemia: diagnostic strategies, predictive features of malignancy and long-term survival

Diagnostic strategies, malignancy predictors and long-term survival were retrospectively evaluated in patients with hyperinsulinemic hypoglycemia (64 insulinomas). Lower median glycemia was 30 (range 20-53) mg/dl [1.6 (1.1-2.9) mmol/l] with concurrent insulin of 48 (13.2-217) microU/ml and 15 (2-46) microU/ml measured by radioimmunoassay (RIA) and immunofluorimetric assay (IFMA), respectively. All patients with insulinomas had a positive prolonged fast within 48 h. Sensitivity of localization methods was: ultrasonography (US) 23%, computed tomography (CT) 28%, magnetic resonance imaging (MRI) 65%, endoscopic US 75%, arteriography 38%, portal venous sampling 67%, selective arterial calcium stimulation 67%, intraoperative US 94% and palpation 92%. Nine patients (14%) had malignant insulinomas. Age at diagnosis (mean+/-SD, 53.8+/-19 vs 39.4+/-16.3 yr; p=0.03), insulin (1372+/-730 vs 785+/-659% (percentage of the method's diagnostic cut-off; 6 and 3 microU/ml for RIA and IFMA, respectively; p=0.007) and C-peptide levels (9.8+/-2.9 vs 3.9+/-2.8 ng/ml (3.2+/-0.9 vs 1.3+/-0.9 nmol/l; p=0.006), and tumor size (6.2+/-4.1 vs 1.5+/-0.6 cm; p=0.0002) were increased in malignant insulinomas. C-peptide level above 6.1 ng/ml (2.0 nmol/l) had a 100% sensitivity and 96% specificity, and tumor size above 2.6 cm yielded a sensitivity of 88% and specificity of 100% in predicting malignancy. Survival of patients with malignant insulinomas was significantly impaired (16 vs 100% at 5 yr; p=0.0000001). The diagnosis of insulinoma can be made within 48 h of fasting. The association between intraoperative US and palpation evidenced the tumor in 95% of the patients. C-peptide and tumor size were reliable malignancy predictors.     

Insulin responses to selective arterial calcium infusion under hyperinsulinemic euglycemic glucose clamps: case studies in adult nesidioblastosis and childhood insulinoma

Selective arterial calcium stimulation and hepatic venous sampling (ASVS) for insulin secretion is used as a diagnostic procedure in patients with insulinomas or adult nesidioblastosis. In some of those patients, severe hypoglycemia requiring urgent glucose administration occurs during the procedure. Such glucose administration, however, may affect the results and damage the validity of the test. We report two cases of hyperinsulinemic hypoglycemia, in which ASVS tests were successfully performed under hyperinsulinemic euglycemic glucose clamps. A 40-year-old male with nesidioblastosis developed continual severe hypoglycemia several years after a Billroth II-Braun gastrectomy, and continuous glucose infusion could not be stopped even during ASVS tests. A 9-year-old girl with an insulinoma that showed atypical hypovascularity on imaging examinations had ASVS tests under a glucose clamp for safety. Hyperinsulinemic (approximately 100 microU/ml) euglycemic (approximately 90 mg/dl) clamps were achieved by an artificial endocrine pancreas. The insulin analogue lispro was utilized for clamps and endogenous insulin was measured with an assay that does not cross-react with the analogue. Diagnostically significant responses (more than twofold) of insulin secretion were observed under hyperinsulinemic clamps in both cases. The use of the hyperinsulinemic glucose clamp technique during the ASVS test should be considered for maintaining the safety of some hypoglycemic patients.     

Insulinoma accompanied by diabetes mellitus

A 53-year-old type 2 diabetic man was admitted due to spontaneous relatively hyperinsulinemic hypoglycemia. Oral glucose ingestion and arginine tolerance test showed hyperinsulinemic response. Arterial stimulation and venous sampling (ASVS) showed hyperinsulinemic response measured from the splenic artery after calcium gluconate stimulation. Diagnosis was insulinoma in the pancreas feeding from the artery. He has not suffered from spontaneous hypoglycemia since removal of the pancreatic body, tail and spleen. The specimen showed a solitary islet cell tumor. The high homeostasis model assessment of insulin resistance (HOMA-R) levels reflecting insulin resistance and hyperinsulinemic response after operation remained almost unchanged, indicating high insulin resistance and an insulin hypersecreting diabetic patient.     

Insulin determination by specific and unspecific immunoassays in patients with insulinoma evaluated by the arterial stimulation and venous sampling test

OBJECTIVE: The aim of this study was to compare insulin concentrations measured by a traditional radioimmunoassay (RIA) and a more specific enzyme-linked immunosorbent assay (ELISA) in blood samples obtained during the arterial stimulation and venous sampling (ASVS) test in patients with insulinoma. DESIGN: Prospective case series. METHODS: In 14 patients with an insulinoma undergoing an ASVS test, blood samples were obtained from the hepatic vein at baseline and 60 s after calcium injection into an artery supplying the tumour and a control artery (supplying pancreatic tissue without tumour). A selective arterial calcium stimulation was performed in five additional patients without evidence for an insulinoma. We measured insulin by a traditional RIA and a specific immunoassay. RESULTS: In patients with insulinoma, insulin concentrations increased between 2.3- and 24.2-fold (median 8.2-fold) when measured by RIA and between 7.3- and 59.4-fold (median 16) when measured by ELISA following calcium injection into the artery supplying the tumour. Following calcium injection into the control artery, insulin concentrations were 0.6 to 1.3 times (median 1.0) the baseline values by RIA and 0.5 to 2.5 times (median 1.1) the baseline values by ELISA. In patients without insulinoma, insulin concentrations increased following calcium stimulation between 0.7- and 2.1-fold (median 1.3-fold) when measured by RIA and between 0.6- and 4.7-fold (median 1.3-fold) when measured by ELISA. CONCLUSIONS: When insulin is measured by specific immunoassays, a higher cut-off (i.e. five- to sixfold increase) rather than the traditional criterion of a twofold increase should be used to localise an insulinoma during the ASVS test. The increase in insulin concentrations following calcium stimulation is significantly higher when insulin is measured by a specific assay compared with results obtained with traditional RIAs.     

Localizaiton of insulinomas by radioimmunoassay of blood obtained by the transport route.

Transhepatic catheterization of the portal system was performed to obtain blood for RIA of insulin. This technique localized insulinomas in two patients after negative celiac, superior mesenteric, and subselective gastroduodenal arteriograms. The tumors were found in the predicted locations at surgery.     

A case with insulinoma diagnosed and localized preoperatively using contrast-enhanced ultrasonography (CEUS) and arterial stimulation and venous sampling (ASVS)

We report a case with insulinoma diagnosed and localized preoperatively using a combination of contrast-enhanced ultrasonography (CEUS) and arterial stimulation and venous sampling (ASVS). A 76-year-old woman was admitted to our hospital because of hypoglycemic attacks, delirium, and dementia. Fajans' ratio, Grunt's ratio, and Turner's ratio, which are reported to be indexes for endogenous hyperinsulinemia in insulinoma, were all negative. Imaging tests, including computed tomography, magnetic resonance imaging and angiography, failed to detect any abnormalities. CEUS showed a small low echoic lesion in the pancreatic body with blood flow and ASVS showed that the insulin levels in the hepatic vein were extremely increased by calcium injection to the splenic artery, indicating an insulinoma in the pancreatic body preoperatively. An open intra-abdominal operation was performed and an insulinoma was confirmed in the pancreatic body. Enucleation of tumor was undertaken and symptomatic hypoglycemia improved.     

Neuroendocrine tumors of the pancreas.

Neuroendocrine tumors of the pancreas (NETP) are rare. We report our surgical experience of 11 patients with NETP. These included 5 patients with benign insulinomas. Raised serum insulin and C-peptide levels with hypoglycemia were always diagnostic. Ultrasonography, CT, visceral angiography, arterial stimulation and venous sampling, and intraoperative ultrasound localized the tumor in 0/5, 1/5, 3/4, 2/2, 3/3 cases, respectively. The 6 other malignant NETP (one gastrinoma, 2 carcinoids, 3 non-functioning) were managed by pancreatic resection (Whipple's operation = 3, distal pancreatectomy with total gastrectomy = 1, total pancreatectomy = 1, distal pancreatectomy with left nephrectomy and proximal gastrectomy = 1). Two patients died postoperatively. We had 5 major and 2 minor postoperative complications, with 2 deaths. During follow up from 1 to 7 years, one patient with malignant carcinoid tumor died at two and half years, of local recurrence. The other 8 patients are disease-free with good quality of life.     

Incidence and treatment of hepatic artery complications after orthotopic liver transplantation

AIM: To investigate the incidence and treatment of hepatic artery complications after orthotopic liver transplantation. METHODS: From February 1999 to May 2002, orthotopic liver transplantations (OLT) were performed in 72 patients with end-stage liver diseases with an average age of 40.2&#177;13.6 years (ranged from 11 to 68 years), 56 were males and 16 females. The preoperative evaluation for the 72 patients was performed using duplexsonography, abdominal CT scan, and angiography of the hepatic artery. All donor grafts were perfused and preserved in University of Wisconsin solution at 4℃. OLT was performed with standard techniques with or without a veno-venous bypass. Reconstructions of hepatic artery were performed between the branch patches of gastroduodenal/hepatic or splenic/common hepatic artery confluence of the donors and recipients, and an end-to-end anastomosis between other arterial vessels of the donors and recipients was done. Arterial anastomosis was performed with interrupted 7-0/8-0 monofilament polypropylene suture under 3.5 x Ioupe magnification. Diagnosis of the complications of hepatic artery after OLT was based on the clinical presentations, ultrasound findings and arterial angiography. All patients were followed up regularly for duplex ultrasound scan after discharge. RESULTS: The overall incidence of arterial complications in 72 patients after OLTs was 1.4% (1/72). One 3cm pseudoaneurysm at the side of anastomotic site of hepatic artery was found by urgent arteriogram due to hemoperitoneum secondary to bile leakage after OLT. Subsequently the pseudoaneurysm was successfully embolized and the blood flow toward the donor liver in hepatic artery remained. The overall postoperative 30day mortality rate was 8.33%. The one-year survival rate was 83.72% in 50 patients with benign diseases and was 71.43% in 22 patients with malignant diseases following OLT. No death associated with complications of hepatic artery occurred. CONCLUSION: Careful preoperative evaluations and intraoperative microsurgical technique for hepatic artery reconstructions are the keys in prevention of hepatic artery complications after OLT.     

Insulinoma: critical study of methods of localization and strategy

In patients with proven hyperinsulinism, localization of the underlying insulinoma may be difficult. The localization diagnosis may be performed preoperatively using different procedures, such as ultrasonography, computed tomography, selective arteriography of the pancreatic vessels and percutaneous transhepatic blood sampling in the portal venous system. At operation, insulinomas may be detected by inspection and bidigital palpation, pancreatico-sonography and rapid determination of insulin concentration after sampling of blood in pancreatic veins. By discussing the advantages and disadvantages of each localization procedure, the authors propose a strategy fort the detection of pancreatic insulinomas.     

A case of secretin-responsive insulinoma with low serum C-peptide levels

Insulinoma is the most common cause of fasting hypoglycemia resulting from autonomous insulin hypersecretion. A 59-year-old woman who had previously had an insulinoma and had undergone a partial pancreatectomy was admitted to our hospital because of recurrence of hypoglycemia after 27 years. She had two unusual endocrinological features: 1) the serum insulin response to intravenous secretin injection was not impaired, and 2) the serum C-peptide levels and ratios of serum C-peptide to insulin were relatively low. Two pancreatic tumors were readily detectable by computed tomography (CT) and magnetic resonance imaging (MRI). The selective arterial calcium injection (SACI) test showed a hyperinsulinemic response by calcium administration to the gastroduodenal artery. A partial pancreatectomy was done and her hypoglycemia disappeared. Histology revealed that the tumors were composed of monotonous, small round cells that were positive for both insulin and cathepsin B. As previous in vitro studies have shown that C-peptide can be metabolized within human insulinoma cells by proteolytic cleavage by cathepsin B, our patient's low serum C-peptide levels might have been caused by degradation of C-peptide by cathepsin B. According to the data from the literature, the molar ratio of serum C-peptide to insulin is generally decreased in patients with insulinoma than normal subjects. This case highlights the need for careful interpretation of C-peptide levels and the intravenous secretin injection test in the diagnosis of insulinoma.     

胰岛素瘤的术前及术中定位诊断方法的探讨

目的 探讨胰岛素瘤术前、术中定位方法的诊断价值,以提高胰岛素瘤定位的准确性.方法 对75例经手术病理证实的胰岛素瘤患者术前B超、CT、MRI的定位资料进行分析,统计术中触诊、术中超声(IOUS)等方法对胰岛素瘤的定位诊断准确率,并进行总结分析.结果 术前B超检查75例,23例发现肿瘤,诊断正确率30.7%;CT检查60例,24例发现肿瘤,诊断正确率40.0%;行MRI 48例,22例发现肿瘤,诊断正确率为45.8%.而术中扪诊、IOUS定位准确率分别为80.4%和96.4%术中定位率比术前明显提高.结论 胰岛索瘤的术前定位诊断比较困难,术中探查和术中B超检查的准确性很高,可不必强调术前准确定位和多种影像学重复检查.     

Localization of islet-cell hyperplasia: value of pre- and intraoperative arterial stimulation and venous sampling

Arterial stimulation and venous sampling was effective in the localization of β‐cell hyperplasia of the pancreas in the islets of Langerhans in an 84‐year‐old woman. The patient presented with repeated episodes of unconsciousness and hypoglycemia. She was first suspected of having insulinoma, but diagnostic imaging failed to reveal any tumors. Arterial stimulation and venous sampling (ASVS) and percutaneous transhepatic portal venous sampling (PTPS) were performed to localize the tumor. By ASVS, increases in immuno reactive insulin (IRI) were noted in renal vein blood samples (because a splenorenal shunt was present) after splenic arterial stimulation and venous sampling, and PTPS revealed a stepup in IRI from splenic venous blood samples. Preoperative diagnosis suggested β‐cell hyperplasia in the pancreas tail. Intraoperative ultrasound failed to find a tumor. Intraoperative ASVS showed the site of increase IRI as the pancreas tail, so distal pancreatectomy and splenectomy were performed. However, hypoglycemia was observed constantly after this operation. Relaparotomy, causing additional resection, was conducted to confirm the precise location and to remove the residual β‐cell hyperplasia of the pancreas. At the second resection, the existing part of β‐cell hyperplasia was confirmed through intraoperative ASVS, and additional resection of the pancreas body and neck was performed. At this time, complete removal of the residual β‐cell hyperplasia was confirmed through ASVS. The hypoglycemia and impaired consciousness disappeared after the operation, and the patient's blood sugar level was maintained at a normal level. Pathological findings revealed islets of Langerhans hyperplasia extending to 1 cm in the pancreas tail region. We conclude that pre‐ and intraoperative ASVS is a useful test for β‐cell hyperplasia, which is difficult to diagnose through ordinary imaging techniques.     

Diagnostic value of contrast enhanced ultrasound for splenic artery complications following acute pancreatitis

AIM:To assess the value of contrast-enhanced ultrasound(CEUS)in diagnosing splenic artery complications(SACs)after acute pancreatitis(AP).METHODS:One hundred and eighteen patients with AP were enrolled in the study.All patients were examined by CEUS and contrast-enhanced computed tomography(CECT).CECT was accepted as a gold standard for the diagnosis of SACs in AP.The diagnostic accuracy of splenic CEUS and pancreatic CEUS was compared with that of CECT.Splenic infarction was the diagnostic criterion for splenic artery embolism and local dysperfusion of the splenic parenchyma was the diagnostic criterion for splenic arterial stenosis.The incidence of splenic sub-capsular hemorrhage,splenic artery aneurysms,and splenic rupture was all lower than that of SACs.RESULTS:Nine patients were diagnosed as having SACs after AP by CECT among the 118 patients.The patients with SACs were diagnosed with severe acute pancreatitis(SAP).Among them,6 lesions were diagnosed as splenic artery embolism,5 as splenic artery aneurysms,and 1 as splenic arterial stenosis.No lesion was diagnosed by pancreatic CEUS and 5 lesions were diagnosed by splenic CEUS.By splenic CEUS,4 cases were diagnosed as splenic artery embolism and 1 as splenic arterial stenosis.The accuracy of splenic CEUS in diagnosis of SACs in SAP was 41.7%(5/12),which was higher than that of pancreatic CEUS(0%).CONCLUSION:Splenic CEUS is a supplementary method for pancreatic CEUS in AP patients,which can decrease missed diagnosis of SACs.     

Selective intra-arterial calcium injection in the investigation of adult nesidioblastosis: a case report

A 69-year-old man with recurrent hypoglycaemia had inappropriately elevated plasma insulin level during a symptomatic hypoglycaemia, but had a negative prolonged fast. Computerized tomography (CT) of the abdomen revealed a nodular lesion over the body of pancreas, whereas pancreatic arteriography failed to show tumour blush. Hence, arterial stimulation (with calcium) and venous sampling (ASVS) was performed and a brisk response of plasma insulin level was found when calcium was injected both into the splenic and the superior mesenteric arteries. Since no tumour was found during the operation, the patient received subtotal distal pancreatectomy. Pathological examination of the resected tissue disclosed a typical finding of nesidioblastosis. We suggest that selective intra-arterial calcium injection with hepatic venous sampling for insulin gradients is useful for the diagnosis of adult nesidioblastosis. © 1997 John Wiley & Sons, Ltd.