PDF
ePub
Citation
Print
An islet cell tumor (ICT) of the pancreas is an endocrine tumor that is clinically classified as functioning or nonfunctioning based on its ability to produce hormones. From a pathologist's viewpoint, however, ICTs are indistinguishable from each other (12). The prompt identification and early treatment of ICTs is of importance because it has been shown to improve the prognosis and life expectancy more than other pancreatic malignancies (3).
The most common and classic findings of functioning ICTs on dynamic computed tomography (CT) are hyperattenuating small lesions (50% of them are less than 1.3 cm) in the arterial phase, which later become inconspicuous in the venous phase (3). Up to 70% of nonfunctioning ICTs also have similar characteristics consistent with arterial enhancements (23). Yet, they are usually larger and more necrotic than functioning ICTs or other pancreatic tumors with or without cystic or calcified components (13). So, nonfunctioning ICTs are not easily distinguishable from other solid or cystic pancreatic tumors, which have different prognoses and need different therapeutic approaches (23). In this article, we presented six pancreatic ICTs (two insuliomas and four nonfunctioning tumors), and illustrated a spectrum of enhancement patterns of the lesions in relation with their histopathologic characteristics.
Imaging Protocol
All multiphase dynamic CT examinations were performed with a conventional helical CT scanner (HiSpeed Advantage; General Electric Medical Systems, Milwaukee, WI, USA) or a multichannel unit (Somatom Sensation 16; Siemens, Erlangen, Germany). After the intravenous administration of 120-150 mL of nonionic contrast agent, iopamidol (Pamiray, Dongkook Pharmaceutical, Seoul, Korea) with an automatic injector at a rate of 3 mL/s, all patients underwent dynamic CT after pre-contrast imaging. In the multichannel unit, aortic enhancement was performed at the start of arterial phase imaging for 15 seconds at 100 Hounsfield units using the SmartPrep technique. Next, portal phase imaging was conducted 30 seconds after the start of arterial phase imaging. The delayed or equilibrium phase was conducted three minutes after the infusion of contrast media. All scans were acquired in the cephalocaudal direction with 3-5 mm thick image reconstruction.
MRI was performed in a 1.5 T system (Magnetom Vision; Siemens, Erlangen, Germany) using a protocol that includes fast spin-echo T2-weighted imaging (TR/TE = 4,060/138 ms, 8 mm section thickness, 1.4 intersection gap, 29 echo train length) followed by preand post-contrast dynamic (arterial, portal and 5-minute delayed phase) imaging using the 2D T1-weighted spoiled gradient echo sequences (TR/TE = 140/2.7 ms, flip angle 90°, 8 mm section thickness). The delay time for arterial and portal imaging was determined by a test bolus examination. Next, gadopentatate dimeglumine (0.1 mmol/kg) was administered with an automatic injector at a rate of 2 mL/s for dynamic imaging. Each imaging procedure was performed during a breath-holding period of 19 to 21seconds.
Correlation of Histopathology with the Imaging Findings
Early-Enhancing Type
Pancreatic ICTs are usually demonstrated as small early enhancing lesions, so they are often difficult to distinguish from normal pancreatic parenchyma. With the advent of the recent multi-slice and multiphase imaging, there has been drastic improvement in the diagnostic accuracy of small ICTs. It was reported in one study that 71 to 82% of insulinomas could be visualized on dual phase imaging by thin slice multidetector CT (3). The main diagnostic clue comes from the hypervascular nature of ICTs. At least 83% of ICTs appear more hyperdense or partially hyperdense than the adjacent pancreatic parenchyma, whether they are functional or not (3). The exact timing of peak enhancement, however, sometimes appeared to lag, and the degree of contrast enhancement on delayed phases tended to vary.
Early Arterial Phase Enhancement
Regardless of tumor size, most pancreatic ICTs show strong early enhancement at the same time as abdominal aortic enhancement. This takes place before peak enhancement of the background pancreatic parenchyma related to the higher vascular densities in the lesion (Figs. 1, 2) (4). It is asserted that the optimal phase of detection for small hypervascular enhancing ICTs is the early arterial phase (5). A recent CT perfusion study has suggested that such a higher blood flow correlated with a high intratumoral microvascular density would be a histoprognostic factor for lower malignant potential compared with other pancreatic ICTs with smaller blood flow (6).
Lagged Peak Enhancement
In some cases, tumoral peak enhancement did not occur along with aortic enhancement, and peak enhancement was shown on the parenchymal or portal phase images (Fig. 3) (7). The cause of such lagged peak enhancement has not been well described. The recent perfusion CT study has illustrated a case of a pancreatic endocrine tumor with scanty vasculature and low intratumoral blood flow despite distinguishable contrast enhancement of the lesion (6). Therefore, a lag in the peak contrast enhancement of the tumor after peak arterial enhancement would not be related with high vascular density in pancreatic ICTs, and would ultimately reflect enhancement of the extravascular stroma of the tumor (Fig. 3).
As the inherently shorter time window for arterial enhancement resulted from a small amount of contrast material during the dynamic MRI, we should note that there is the possibility of improper image acquisition causing the missing peak tumoral enhancement.
Sustained Enhancement on Delayed Phase Images
Despite the many early-enhancing ICTs that became isodense or isosignal intensity during the delayed phase of dynamic CT or MRI, they cannot be distinguished from the remaining pancreatic parenchyma; this is especially true for small lesions (3). Regardless of the vascular densities of the tumor, pancreatic ICTs would have variable degrees of interstitial space composed of hyaline or amyloid materials intervening in the tumor cells (8).
In the liver, prolonged enhancement occurs in a mass with a large extent of interstitial space (9). Compared with the normal washout of contrast materials from the tumors without connective tissue stroma (Fig. 1), the presence of connective tissue stroma could explain prolonged contrast enhancement in pancreatic ICTs (Figs. 2, 3).
Hypovascular or Late-Enhancing Type
In a minority of pancreatic ICTs, contrast enhancement is not definitely demonstrated in the arterial or portal phase images of dynamic imaging. Enhancement of such tumors is gradual or considerably delayed. In previous reports, abundant amyloid deposits or prominent desmoplastic reactions with relatively smaller vascular volume, were observed in early hypovascular and rather delayed enhancing insulinomas (8). When a large proportion of the tumoral component is replaced by hypovascular connective tissue stroma, there would be minimal enhancement on the early phases of dynamic imaging. However, gradual retention of contrast material in the extracellular interstitial tissue can explain delayed enhancement distinguished from background pancreatic parenchyma with a deficient interstitial volume (Fig. 4). In conjunction with rare invasive features of infiltrative ICTs which spread to adjacent organs and/or distant metastases (1), it is not always easy to differentiate such tumors from common pancreatic cancers of ductal cell origin (Fig. 5).
Cystic or Hemorrhagic Type
Nonfunctioning ICTs are typically heterogeneous tumors that are larger than functioning ICTs (3). Regardless of functioning status, larger tumors are more likely to have cystic degeneration and central areas of necrosis. Cystic change in pancreatic ICTs has been found occasionally with only a 2-3% incidence among pancreatic ICTs (10). Pathologically, this is explained by the presence of cystic cavities filled with clear fluid or hemorrhage. Moreover, the wall of the cavity is lined with islet cells without sound evidence of necrosis or infarction in most cases (Figs. 3, 6). Despite the secondary features, contrast enhancement of the solid component depends on the tumoral vasculature and fibrotic stroma in the remaining parenchyma (Fig. 3). In the cases of young women, differentiating a cystic or hemorrhagic ICT from a solid pseudopapillary tumor (SPT) is not always easy. However, the more heterogeneous and less hypervascular nature of SPT could be helpful in differentiating them from ICTs on dynamic imaging studies.
Conclusion
In addition to aiding the detection of small lesions, multiphase dynamic CT or MRI can characterize pancreatic ICTs. Pancreatic ICTs have been known as representatives of hypervascular tumors. However, regardless of their size or functional behavior, imaging findings on pancreatic ICTs vary depending on the internal composition. While early hypervascular ICTs show abundant intratumoral vasculature, sustained or delayed enhancement is influenced by the amount of interstitial connective tissue stroma that consists of variable degrees of maturity or degenerative changes. Cystic change takes place during the process of intratumoral hemorrhagic congestion and/or cystic degeneration, which is also one of the representative features of pancreatic ICTs. The aforementioned pathological knowledge on the various enhancement patterns of pancreatic ICTs could help radiologists better understand these tumors and perform a differential diagnosis.
XML Download