Pancreatic islet cells are part of the diffuse neuroendocrine system of the gut and pancreatic endocrine system. Islet cells commonly are referred to as APUD cells, a name derived from their high amine content and capacity for amine precursor uptake with decarboxylation. 
Tumors of pancreatic islet cells are uncommon. They may manifest as sporadic tumors or as part of certain syndromes, including multiple endocrine neoplasia type 1 (MEN 1) and von Hippel–Lindau (VHL) disease. Islet cell tumors may be functional or nonfunctional; although nonfunctional islet cell tumors are not uncommon at autopsy, most islet cell tumors with clinical manifestations are functional.
Functioning tumors produce a clinical syndrome as a result of excessive hormone production. Clinical features of the syndrome depend on tumor cell type. Pancreatic islet cell tumors may secrete 2 or more polypeptide hormones. Functioning tumors usually are small at presentation, and localizing these tumors can be challenging to the radiologist. Hormonal and biochemical parameters are invaluable for skillful interpretation of the imaging and clinical features and to arrive at a specific diagnosis. Nonfunctioning tumors usually are larger and present because of their size or metastatic spread.
Prognosis for patients with MEN 1 is usually poor. Islet cell tumors in this group are often multiple and malignant. In patients with MEN 1 and Zollinger-Ellison (ZE) syndrome,  surgical cure is usually not possible. However, surgical cure can be achieved in patients with MEN type-I and insulinomas, although recurrences are frequent.
Imaging is used to localize primary and metastatic lesions and to determine resectability or alternative palliative and curative treatment options. This article reviews the role of imaging in the management of islet cell tumors. [3, 4, 5, 6, 7]
Types of islet cell tumors
These are as follows:
Nonfunctioning islet cell tumor 
MEN 1 
No universally agreed-upon algorithm exists in the radiologic investigation of islet cell tumors of the pancreas. The diagnosis usually is made by the clinician, and the role of the radiologist is to localize, demonstrate the extent, and stage the lesion, thereby allowing correct management. The best modality for diagnosis is debatable and depends on the expertise of the radiologist at each institution. Larger lesions are easily identifiable by modern cross-sectional imaging techniques. Smaller lesions are difficult to demonstrate, requiring more sophisticated imaging and meticulous technique.
Endoscopic ultrasonography (EUS) is performed by using a 10-MHz transducer incorporated into an endoscope. Lesions are seen as well-rounded, oval echogenic, or hypoechoic areas in the gland parenchyma. [8, 14]
A retrospective study by Boukhman et al evaluated the sensitivities of different techniques to localize insulinomas as follows  :
Arteriography - 47%
Computed tomography (CT) scanning - 24%
Preoperative ultrasound (US) - 50%
Magnetic resonance imaging (MRI) - 30%
Gadolinium-enhanced MRI - 40%
Transhepatic venous sampling - 55%
Intraoperative palpation - 76%
Intraoperative ultrasonography - 91%
EUS is superior to single-slice spiral CT and should replace the latter for preoperative detection of pancreatic insulinomas. However, because EUS is an invasive study, it should be reserved for patients in whom noninvasive modalities fail to localize the expected tumor. However, multislice CT with 1-mm collimation may prove to be an effective imaging modality. A prospective trial comparing EUS with multislice CT is needed
Plain radiographs have no role in the diagnosis or localization of islet cell tumors, although some tumors may demonstrate calcification in a peripheral curvilinear pattern or a central dystrophic pattern in the region of the pancreas. [19, 20, 21] Vipomas and glucagonomas may reveal some fluid-filled loops and hypotonia. Barium studies are likely to show ulcers in the duodenum in ZE syndrome and evidence of gastritis. Radiographs of a patient with glucagonoma reveal thickened mucosal folds in the small bowel from villous hypertrophy. Vipomas and glucagonomas may show delayed transit on barium studies. A patient with vipoma may have "pancreatic cholera" with an extremely wet small bowel.
Degree of confidence
Plain radiography has no role in the specific diagnosis of pancreatic islet cell tumor. It has a complementary role, as it suggests the diagnosis on the basis of bony changes, renal calculi, or pituitary changes in MEN.
CT is an excellent modality to detect larger tumors. Precontrast CT images demonstrate insulinomas as a hyperdense or isodense mass, as compared with the normal pancreas. IV administration of iodinated contrast demonstrates a variable enhancement pattern for insulinomas in the arterial and portal venous phase. Typically, but not always, intense arterial phase enhancement is observed. [22, 15]
Biphasic CT, including the hepatic arterial dominant and portal venous phase, and MR imaging are similarly effective in the detection of islet cell tumors. Glucagonomas are easily demonstrable as a soft-tissue mass. After IV contrast administration, the tumor reveals intense enhancement with heterogeneity when necrotic areas are present within the mass. Tumors are hypervascular in 90% of patients.
Spiral CT performs poorly in comparison to EUS. A recent study showed an overall sensitivity of 16.7% for spiral CT in the localization of insulinomas. CT can depict metastases as small as 5 mm in diameter and demonstrate vascular invasion. However, multislice helical CT and thin-section 3-dimensional gadolinium-enhanced MRI will likely improve diagnostic yield for both the primary pancreatic lesion and liver metastasis. 
Degree of confidence
Degree of confidence in using CT to detect islet cell tumors is good when the tumors are larger (at least 1.5-2 cm). Sensitivity and specificity are poorer for smaller lesions.
The incidence of false-negative results are probably higher than that of false-positive results, especially for small solitary lesions.
Magnetic Resonance Imaging
MRI has demonstrated a higher sensitivity than CT, offering additional information that may help characterize islet cell tumors. Tumors are hypointense on T1-weighted images and hyperintense on T2-weighted images. Intense enhancement is seen on administration of IV gadolinium. [3, 4, 11, 22, 23, 24, 25, 17]
In the selection of MR imaging sequences for optimal sensitivity, a study has shown that T2-weighted fast spin-echo and spoiled gradient-echo sequences usually suffice. Gadolinium-enhanced sequences are needed only if MRI results are equivocal or negative. MRI performed with gadolinium may be more sensitive to tumor vascularity than CT performed with IV contrast material, thus permitting the detection and characterization of lesions better than CT can.
Gadolinium-based contrast agents (gadopentetate dimeglumine [Magnevist], gadobenate dimeglumine [MultiHance], gadodiamide [Omniscan], gadoversetamide [OptiMARK], gadoteridol [ProHance]) have been linked to the development of nephrogenic systemic fibrosis (NSF), or nephrogenic fibrosing dermopathy (NFD). The disease has occurred in patients with moderate to end-stage renal disease after being given a gadolinium-based contrast agent to enhance MRI or MRA scans.
NSF/NFD is a debilitating and sometimes fatal disease. Characteristics of the disease include red or dark patches on the skin; burning, itching, swelling, hardening, and tightening of the skin; yellow spots on the whites of the eyes; joint stiffness with trouble moving or straightening the arms, hands, legs, or feet; pain deep in the hip bones or ribs; and muscle weakness.
Solid and papillary neoplasms are rare, low-grade malignant tumors more commonly seen in women. The mass is most frequently located in the tail of the pancreas with solid and cystic components. They are hypovascular, show minimal enhancement, or may be completely cystic with a mean size of 10 cm. These tumors may contain blood from hemorrhagic necrosis, giving a hyperintense signal on T1-weighted MR images.
MRI is helpful in characterizing islet cell tumors, which may be markedly hyperintense on T2-weighted and short tau inversion-recovery (STIR) images. They usually demonstrate marked enhancement in the hepatic arterial dominant phase (HAP) after gadolinium administration. Conversely, adenocarcinoma is usually less hyperintense on T2-weighted images and relatively hypovascular as compared with a normal pancreas in the HAP.
In the Mayo clinic series, the most frequent metastasis to the pancreas was from renal cell carcinoma. This easily can be diagnosed by demonstrating the primary renal tumor but may be more problematic if the tumor was resected and was low grade.
Degree of confidence
Degree of confidence is high using MRI in the detection of lesions, especially for insulinomas, even on low-field strength magnets. However, specificity is poor. MRI is less sensitive than EUS in detecting lesions smaller than 1 cm.
Tumors less than 2 cm in diameter are not easily identified by transabdominal ultrasonography (TUS). TUS detection sensitivity is 20-75% for insulinomas and 20-30% for gastrinomas. However, the use of harmonic imaging and US contrast agents may improve sensitivity. Appearances vary depending on the nature of the lesion. Lesions may show cystic change, focal calcifications, or hemorrhage within. TUS is helpful in demonstrating liver metastasis. Hyperechoic metastasis with a pancreatic mass generally suggests an islet cell tumor rather than an adenocarcinoma of the pancreas.
Evidence has shown that EUS can be used to accurately diagnose and localize primary endocrine tumors of the pancreas, especially insulinomas and gastrinomas. A study concluded that EUS is the most accurate preoperative method of localizing and detecting insulinomas.
Degree of confidence
Using TUS, the degree of confidence to rule out an islet cell tumor of the pancreas is poor. Lesions less than 2 cm are easily missed. EUS currently may be the best method to detect these tumors. The degree of confidence is high using EUS to make a diagnosis of islet cell tumor in the presence of appropriate clinical and biochemical evidence.
Lesions that contain somatostatin receptors may be detected by somatostatin receptor imaging (SRI) using octreotide, a somatostatin analogue. The detectability of the tumor directly depends on the number of receptors the tumor contains rather than the tumor size. Since only approximately 60-70% of insulinomas are positive for somatostatin receptors, SRI is not a good technique for detecting insulinomas. 
Single photon emission CT (SPECT) scanning using Indium-111 (111 In)–pentetreotide scintigraphy at 4 hours has shown higher sensitivity than planar somatostatin receptor scan SRI and MRI in the detection of insulinomas. Gastrinomas also reveal a high sensitivity with111 In-pentetreotide. 
Interest has been shown in using positron emission tomography (PET) scanning to detect islet cell tumors. A study revealed a sensitivity rate of 53% in the detection of islet cell tumors. Data suggest that fluorodeoxyglucose (FDG)-PET is limited in that it does not detect some small islet cell tumors but has potential use as a complementary modality for other imaging techniques. [26, 7]
Degree of confidence
The degree of confidence to rule out pancreatic islet cell tumor using somatostatin receptor imaging or FDG-PET is poor at present.
Pancreatic arteriography was regarded as the optimal method for preoperatively localizing insulinomas, although evidence has since indicated that EUS is the most sensitive investigation in localizing insulinomas.  A well-performed selective arteriography has approximately 80% sensitivity in detecting the tumors. Insulinomas as small as 5 mm may be demonstrated on angiography.
Arteriographic findings depend on primary tumor size. If the tumor is smaller than 3 cm, arteriography fails to satisfactorily demonstrate venous shunting. With gastrinomas larger than 3 cm in diameter, arteriography may reveal tumor vessels, early venous filling, and venous occlusion. Liver metastasis is generally hypervascular. Gastric arteries may be enlarged in some patients with early filling in splenic veins.
Arteriography generally has been found to be unsatisfactory for the demonstration of gastrinomas, although gastrinomas are larger than insulinomas. Approximately 15% of gastrinomas are peripancreatic; thus, they are easy to miss at arteriography. Vipomas are characteristically hypervascular, with a reticular neovascularity and microaneurysms. Dilated veins in the peripancreatic region may be seen.
Portal venous sampling is a technique used as an adjunct to arteriography to obtain samples of portal venous blood and measure insulin concentration in patients with suspected insulinoma. Depending on the vein from which the highest concentration of insulin is obtained, the tumor is then localized to the area drained by the vein. This time-consuming technique requires more technical skill than simple angiography.
Selective arterial stimulation testing (SAST) has largely replaced portal venous sampling. First, a catheter is placed into a hepatic vein, and the basal value of the hormone is recorded from a sample. Then, a second catheter is selectively positioned into arteries supplying the pancreas and injected with secretagogues. At 1-2 minutes after injection of the secretagogues, hepatic vein samples are taken and hormone levels estimated. A 2-fold increase in hepatic venous concentration identifies and localizes the tumor to the area supplied by the artery. Calcium is used as a secretagogue for insulinomas and secretin for identifying gastrinomas.
Degree of confidence
Angiography may be useful in planning therapeutic considerations. However, the modality does not provide a satisfactory degree of confidence in demonstrating all islet cell tumors and is not recommended in the primary investigation for detection.
Almost 20% of islet cell tumors may not be demonstrated on angiography because of a combination of factors such as size, vascularity, location, and technique. Therefore, it may be unsuitable for lesion detection.