| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1 Department of Radiology, Université René Descartes,
Hôpital Cochin, 27 rue du Fg Saint Jacques, 75679 Paris Cedex 14,
France.
2 Department of Surgery, Université René Descartes, Hôpital
Cochin, 75679 Paris Cedex 14, France.
3 Department of Gastroenterology, Université René Descartes,
Hôpital Cochin, 75679 Paris Cedex 14, France.
4 Department of Pathology, Université René Descartes,
Hôpital Cochin, 75679 Paris Cedex 14, France.
Received November 6, 2002;
accepted after revision April 18, 2003.
Address correspondence to H. Gouya.
Abstract
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
MATERIALS AND METHODS. All patients treated in our institution for surgically proven insulinoma between 1987 and 2000 were retrospectively reviewed. Thirty patients with 32 pancreatic insulinomas underwent preoperative CT and endoscopic sonography and were included in the study. These 30 patients also underwent dual-phase thin-section multidetector CT (group 1: n = 15), dual-phase multidetector CT without thin sections (group 2: n = 8), or sequential CT (group 3: n = 7). CT scans were interpreted separately and retrospectively by three radiologists in consensus. Sensitivity values for CT, endoscopic sonography, and a combined protocol were determined.
RESULTS. The overall diagnostic sensitivity for dual-phase helical CT was 94.4% for group 1, 57.1% for group 2, and 28.6% for group 3. Endoscopic sonography showed proven insulinomas in 30 of 32 cases (sensitivity, 93.8%). Differences between dual-phase thin-section CT and endoscopic sonography were not statistically significant. The overall diagnostic sensitivity for combined biphasic thin-section helical CT and endoscopic sonography was 100%.
CONCLUSION. The most effective method for revealing insulinomas is a combined imaging protocol that consists of both dual-phase thin-section multidetector CT and endoscopic sonography.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
More recent tools for detecting endocrine tumors may be of great value. A recent study showed excellent results for MRI, with 85% of lesions detected [3]. The sensitivity in this study may be high because of improved sequence design and better gradients and coils that enable faster imaging sequences. However, to our knowledge, no data from a large group of patients are available in the literature to prove this point. Several studies have shown that endoscopic sonography is the most accurate method available for the diagnosis of islet cell tumors [4–7]. With the advent of multidetector CT (MDCT), the acquisition of two sets of images after the injection of contrast material is now possible. This capability provides an opportunity to scan the pancreas during the arterial phase, which is when enhancement of the pancreatic parenchyma peaks. Theoretically, imaging during the arterial phase could improve the conspicuity of the lesion and, therefore, may increase the sensitivity of CT for the detection of small hypervascular tumors such as insulinomas. Although the results are still sporadic and need to be confirmed in a larger series of cases, multiphase helical CT may be of great value in depicting insulinomas [8–10].
The purposes of this retrospective study were to compare CT with endoscopic sonography for the detection and localization of islet cell tumors of the pancreas and to correlate the imaging results with surgical and histologic findings.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
An insulinoma was suspected on the basis of clinical symptoms of hypoglycemia (neuroglycopenic or adrenergic reactions) and abnormal laboratory test results including the Whipple's triad (low blood levels of glucose, symptoms of hypoglycemia, and immediate resolution after administration of IV glucose). All patients in our study group presented with neuroglycopenic symptoms including confusion, personality changes, diplopia, blurring of vision, or loss of consciousness. The most common adrenergic symptoms were weakness, sweating, and tachycardia (n = 21). High levels of C peptide and serum insulin indicated endogenous hyperinsulinism. The interval between the onset of symptoms and diagnosis ranged from 1 month to 5 years, with a median of 18 months.
CT Acquisition
CT examinations were performed using three techniques. Each patient
underwent either a dualphase helical acquisition (groups 1 and 2: n =
23) or a dynamic sequential acquisition (group 3: n = 7) with a
biphasic injection of contrast material. Each patient received a bolus of 120
mL of low-osmolar iodinated contrast medium (Omnipaque [iohexol], Nycomed,
Oslo, Norway) at a rate varying between 3 and 4 mL/sec.
For 15 patients (from 1997 to 2000: group 1), the CT examination consisted of unenhanced scanning followed by two selective sequential breath-hold acquisitions obtained with a helical CT scanner (CT Twin, Elscint, Haifa, Israel). The first and second helical acquisitions were performed during the arterial dominant phase and the portal venous phase, respectively (20 and 60 sec after the initiation of IV administration of contrast material, respectively). The unenhanced CT acquisition was used to localize the pancreas. The first helical acquisition consisted of thin-section images of the pancreas collimated to 3.2 mm (250 mAs, 120 kV) at a pitch of 1:1. The second helical acquisition included the entire liver and pancreas. A collimation of 6.5 mm (200 mAs, 120 kV) and a pitch of 1:1 were used. Reconstruction increments were 1.6 mm for the arterial phase and 5-mm for the portal venous phase.
Eight patients (from 1991 to 1997: group 2) underwent dual-phase helical CT (HiSpeed Advantage scanner, General Electric Medical Systems, Milwaukee, WI). Helical CT scans were obtained 20 sec (arterial phase) and 70–75 sec (portal venous phase) after the start of the injection of contrast material. Sections were collimated to 6.5 mm (200 mAs, 120 kV) at a pitch of 1:1 for the two phases. The images were obtained from the right side of the diaphragmatic dome to the inferior margin of the pancreas. A preliminary unenhanced acquisition was used to determine the extent of the pancreas.
The remaining seven patients (from 1987 to 1991: group 3) underwent nonhelical CT (9800 scanner, General Electric Medical Systems) because a helical CT scanner was not available. CT scans were obtained during breath-holding with a 5-mm collimation at 5-mm intervals through the liver and the pancreas before and after the injection of contrast material.
Technique for Endoscopic Sonography
From 1992 to 2000, 38 patients, in whom the possibility of an insulinoma
was suspected, underwent endoscopic sonography. Endoscopic sonography was
performed with a side-viewing endoscope (GF-UM 20, Olympus, Tokyo, Japan).
Patients were sedated with meperidine hydrochloride and midazolam. The field
of view was 360° orthogonal to the long axis of the scope. The sonographic
frequency ranged from 7.5 to 12 MHz. Lesions close to the probe (25 mm) were
imaged at 12 MHz, and lesions farther from the probe were imaged at 7.5 MHz.
The instrument was advanced as far as possible into the duodenum, and imaging
was initiated. The different parts of the pancreas were carefully scanned by
slowly withdrawing the instrument. The head of the pancreas, the portal vein,
and the confluence were imaged from the duodenum; and the body and tail of the
pancreas, the splenic vein, and celiac trunk were scanned from the gastric
body and fundus. Transduodenal and transgastric approaches were mandatory to
accurately assess the entire pancreas and the surrounding vessels. The
procedure lasted between 10 and 45 min. No complications occurred during the
examinations.
Technique for Intraoperative Sonography
Intraoperative sonography was performed by a radiologist with a 7-MHz
transducer and was used for only 21 patients because, at the beginning of this
study, intraoperative sonography was not available in our institution. Data
recorded included type of echogenicity (hyperechoic, hypoechoic, isoechoic, or
anechoic), whether the tumor was homogeneous or heterogeneous, and whether
clear margins were present.
Surgical and Histopathologic Analysis
The pancreas was widely exposed by the surgeon to localize the tumor. The
objectives at surgery were first to identify the tumor by careful exploration
of the pancreas both manually (inspection and palpation between finger and
thumb) and by intraoperative sonography and second to remove all tumor tissue
by enucleation or resection. The choice of the surgical procedure depended on
the tumor site in the pancreas and its relationship to the pancreatic vessels
and ducts. Enucleation was the preferred technique for tumors located in the
head of the pancreas.
When lesions were treated with enucleation, the surgeon recorded the location, size, and number.
An experienced endocrine tumors pathologist reviewed all pathology specimens. Samples of tumor were obtained by resection. After resection, tissues were fixed in Bouin's fluid for immunohistochemistry or were immediately frozen in liquid nitrogen for Western blot analyses. The size of each tumor and the number of tumors were recorded. The tumor was analyzed to detect whether malignant cells were present.
Image Analysis
All CT scans were interpreted independently by three radiologists with
experience in this imaging technique in a retrospective fashion by consensus
interpretation. The radiologists were unaware of the findings at endoscopic
sonography and at surgery. Nevertheless, they knew that all the patients were
examined to detect an insulinoma. Specific features analyzed were tumor size,
number of tumors, and location of the lesions. The degree of contrast
enhancement of the tumors and normal pancreas was quantitatively evaluated for
all patients of group 1. Three or more regions of interest that excluded
lesions, edges of the organ, and vascular or ductal structures were obtained
from homogeneous areas of the head, body, and tail of the pancreas on
corresponding unenhanced arterial and portal venous phases for a same section.
Then the attenuation values were averaged, and pancreatic enhancement was
obtained by subtracting the average attenuation value on unenhanced images
from the average attenuation value on contrast-enhanced images. Quantitative
lesions enhancement was obtained in an identical fashion on the corresponding
sections.
One experienced endoscopist and sonographer performed endoscopic sonography. At the time of examination, the endosonographic investigator was not aware of the results of CT. Tumor location, size, and number were recorded in a retrospective fashion on the basis of the reports.
The results of CT and endoscopic sonography were compared with surgical and histopathologic results.
Statistical Analysis
Fisher's exact or chi-square test was used to compare sensitivities. The
nonparametric Mann-Whitney test was used for qualitative analysis, and
p values of 0.05 or less were considered statistically
significant.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Lesions to the left of the midline (n = 15 tumors) were treated with distal pancreatectomy (n = 8) or enucleation (n = 7). Lesions to the right of the midline (n = 17 tumors) were treated with duodenopancreatectomy (n = 7) or enucleation (n = 10).
Histopathologic Findings
Thirty-two insulinomas were found and removed from all 30 patients at the
time of surgery. All tumors were located in the pancreas. Two patients had two
insulinomas (Fig. 1). We found
no statistically significant differences in terms of the location of the
lesions in the pancreas: 11 were located in the head of the pancreas, 10 in
the body, and 11 in the tail. The mean diameter of the lesions was 19.6 mm.
The smallest tumor measured 9 mm, and the largest was 90 mm. The mean diameter
of the benign lesions was 16.5 mm (range, 9–30 mm). Three of the
insulinomas were malignant. The malignant insulinomas ranged from 30 to 90 mm
(mean, 50 mm). The difference in the diameter between benign and malignant
tumors was statistically significant (p = 0.007).
|
Imaging Studies
The sensitivities of endoscopic sonography and CT for depicting insulinomas
are reported in Table 1.
|
Dual-phase helical CT with a thin primary collimation during the arterial phase (group 1) was significantly superior to dual-phase helical CT with a nonthin primary collimation (group 2) (p = 0.05) and dynamic sequential CT (group 3) (p = 0.002) with regard to sensitivity for depicting pancreatic tumors. Helical CT in group 1 failed to depict a 10-mm tumor located in the body of the pancreas (Fig. 2).
|
Endoscopic sonography detected 30 (93.8%) of the 32 lesions. Of the two insulinomas not shown on endoscopic sonography, one was a 12-mm insulinoma in the head of the pancreas, and the other one was a 9-mm insulinoma located in the tail (Fig. 1).
Differences between dual-phase thin-section helical CT and endoscopic sonography were not significant for tumor detection. For these two techniques, the sensitivities were independent of the tumor location and size.
Table 2 lists quantitative enhancement measurements of normal pancreatic parenchyma and pancreatic tumors on thin-section helical CT obtained during the arterial and portal venous phases. Enhancement of the normal pancreas was significantly higher in the arterial phase than in the portal venous phase. Although enhancement of insulinomas was also significantly higher in the arterial phase, the contrast between the normal pancreas and the lesions was significantly higher during the arterial phase, resulting in greater tumor conspicuity and significantly better sensitivity for the detection of insulinomas (11.8% for the portal venous phase and 88.2% for the arterial phase).
|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The surgical strategy depends on the exact location of the tumor, and preoperative detection of functioning neuroendocrine tumors remains a difficult challenge. The most effective technique to detect insulinomas is still under discussion. Different techniques for localization of the tumor have been studied and show a great disparity in success rates. Nevertheless, it appears that technical advances, such as endoscopic sonography and helical CT [3–10], can improve the level of sensitivity of imaging modalities. Some authors have shown that endoscopic sonography is a powerful imaging technique for revealing and staging of tumors in the gastrointestinal tract and the pancreas [15–17]. Other studies have established the accuracy for localization of small pancreatic tumors including neuroendocrine tumors, especially insulinomas, is higher for endoscopic sonography than other imaging modalities [4–7]. Although findings in recent studies have revealed that biphasic helical CT is a sensitive technique to identify insulinomas, the suggestion is based on the analysis of only small series of tumors.
In our patient cohort, the size and location of insulinomas agreed with prior reports [3, 18]. All insulinomas were located in the pancreas. Approximately 90% of the insulinomas were benign and less than 2 cm in diameter. All insulinomas that were malignant had a diameter of 30 mm or greater. None had metastasized at the time of the initial diagnosis. Multiple insulinomas were found in two patients.
In this series, intraoperative palpation enabled tumor localization in 27 of 28 cases. This finding is in agreement with prior reports [19, 20]. Intraoperative sonography revealed 20 of 21 insulinomas. The sensitivity of combined intraoperative sonography and palpation was 100%, as has been reported by another group of researchers [21]. Besides enabling tumor localization, intraoperative sonography may also serve as a guide to the decision between enucleation and resection because it delineates the relationship between the tumor and the main pancreatic duct. However, this technique prolongs surgery and has an associated risk of rupture of the splenic vessels due to mobilization of the pancreas [22]. Preoperative identification of the tumor would avoid these risks and also lead to selection of a better surgical strategy [23].
Our study findings reveal that biphasic helical CT with high-resolution thin sections during the arterial phase provided an accurate and sensitive method for the depiction of insulinomas. These results were significantly better than those for dynamic incremental CT. Helical CT offers numerous advantages compared with sequential CT in the evaluation of exocrine and endocrine pancreatic neoplasm as shown in earlier articles, but to our knowledge this was never shown in a comparative study. This technique has the potential to provide high-spatial-resolution images. Acquisition of high-resolution thin sections is facilitated by the rapid scanning speed of our scanner with its dual-detector rings. Several authors have reported increased detection of pancreatic ductal adenocarcinomas with helical CT [24–27].
Because insulinomas are usually hypervascular and the pancreas is exclusively perfused by the arterial system, scanning during the arterial phase has been suggested to increase the contrast between the tumor and normal pancreas. Therefore, portal venous phase imaging is not the most sensitive method for detecting pancreatic tumors. Recently, biphasic helical CT was presented as a solution for achieving increased sensitivity of CT. Images are obtained during three phases relative to the injection of the contrast agent: unenhanced, arterial, and portal venous phases. We found attenuation differences between tumors and normal pancreas during the arterial and portal venous phases. Our findings revealed a significant increase in the contrast between insulinomas and normal pancreas for the thin-section images obtained in the arterial phase compared with those obtained in the venous phase; the increased contrast between tumor and normal pancreas resulted in a higher sensitivity for the detection of small hypervascular insulinomas. These results are in agreement with those from other studies [8–10]. Several previous studies reported that the detection of hypovascular and hypervascular pancreatic tumors improved when scans from both the arterial and portal venous phases were used [25, 26]. Thin-section high-resolution images are obtained during the arterial phase; these images improved the accuracy of helical CT for the detection of small hypervascular lesions.
No statistical difference was observed in the sensitivity results for
endoscopic sonography, regardless of the diameter and location of the tumors.
Endoscopic sonography has been shown to be highly accurate for the
localization of pancreatic endocrine tumors
[4,
18]. However, the results of
some studies suggest that endoscopic sonography is operator-dependent, a
factor that can adversely affect the accuracy of detection
[5,
7]. Furthermore, the
sensitivity of endoscopic sonography is higher for detecting tumors in the
head or the body compared with the tail of the pancreas. This difference in
sensitivity is most likely because of differences in the ability to visualize
these parts of the pancreas on endoscopic sonography
[28]. Our study findings did
not corroborate these results. The accuracy rates for the detection of
insulinomas in the head, the body and the tail were similar. These results may
reflect the experience of the operators. Indeed, the learning curve with
endoscopic sonography is steep, leading to substantially better results after
performing many examinations. Another explanation is that no small lesions
(
1 cm) were located in the body or the tail of the pancreas. In addition,
these values should be viewed with caution because the number of tumors in our
series was relatively small.
Biphasic helical CT has some potential advantages compared with endoscopic sonography for the localization of small neuroendocrine tumors. The advantages of helical CT include its noninvasiveness and greater overall topography display. Endoscopic sonography is not considered accurate in the detection of liver metastases because of the limited penetration depth of sonography. Nevertheless, one potential criticism of our pancreatic CT protocol is that hypervascular liver metastases may be missed because hepatic scanning occurs during the portal venous phase.
A recent study showed that MDCT of the pancreas is useful for tumor diagnosis and staging [29]. MDCT enables coverage of large volumes in short times using a thin-collimation and high-resolution protocol. The main limitations of single-detector helical CT—specifically, the compromise between volume coverage and slice thickness—are eliminated with MDCT. The use of MDCT will probably improve the detection of endocrine tumors, especially insulinomas, with three-dimensional reformatting in any plane, which will most likely increase the conspicuity of small pancreatic and hepatic lesions.
Previously published studies have reported MRI to be better than CT in revealing small insulinomas [30, 31]. However, these results should be regarded with caution because the number of patients in each of these studies was relatively small. More recently, a prospective MRI investigation, taking advantage of improved sequence design and better gradients and coils that enable faster imaging sequences and higher signal-to-noise ratios, reported a high sensitivity for depicting functional islet cell tumors that are 2 cm or less in diameter [3]; in fact, the sensitivity reported in that study was superior to most MRI results reported in the literature [3].
Our results suggest that the use of these two techniques improves significantly the accuracy of both helical CT and endoscopic sonography alone for the preoperative detection of insulinomas. In our study, we report insulinomas that were located preoperatively on helical CT in patients whom endoscopic sonography had failed to detect any pancreatic tumor. Conversely, one tumor was located with endoscopic sonography and not with helical CT. Despite these excellent results, pancreatic tumors remained undetected in one patient. When these diagnostic procedures were combined and were in agreement, the overall sensitivity was 100%. In these cases, tumor locations on both dual-phase helical CT and endoscopic sonography were confused with surgical examination in all patients (Fig. 3A, 3B). To our knowledge, no data have been previously reported about the complementarity of dual-phase helical CT with thin sections and endoscopic sonography.
|
|
In conclusion, endoscopic sonography is considered the gold standard in preoperative diagnosis of small pancreatic insulinomas. CT detection of small pancreatic insulinomas has improved since the introduction of the helical technique. Consistent with other reports, our study found preoperative detection of pancreatic insulinomas was markedly higher with the use of the optimal biphasic thin-section helical technique. Our results also show that even small tumors can be recognized reliably on biphasic thin-section helical CT and that differences between the two techniques in overall sensitivity were not statistically significant. However, the most effective detection method for small insulinomas is a combined imaging protocol consisting of both triphasic contrast-enhanced helical scanning and endoscopic sonography. Additional studies are required to determine the additive value of MDCT and MRI for preoperative evaluation of pancreatic insulinomas.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  D. Wild, M. Behe, A. Wicki, D. Storch, B. Waser, M. Gotthardt, B. Keil, G. Christofori, J. C. Reubi, and H. R. Macke [Lys40(Ahx-DTPA-111In)NH2]Exendin-4, a Very Promising Ligand for Glucagon-like Peptide-1 (GLP-1) Receptor Targeting J. Nucl. Med., December 1, 2006; 47(12): 2025 - 2033. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 K. K. Kazanjian, H. A. Reber, and O. J. Hines Resection of Pancreatic Neuroendocrine Tumors: Results of 70 Cases Arch Surg, August 1, 2006; 141(8): 765 - 770. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 E. M. Chung, M. D. Travis, and R. M. Conran From the Archives of the AFIP: Pancreatic Tumors in Children: Radiologic-Pathologic Correlation. RadioGraphics, July 1, 2006; 26(4): 1211 - 1238. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K. M. Horton, R. H. Hruban, C. Yeo, and E. K. Fishman Multi-Detector Row CT of Pancreatic Islet Cell Tumors. RadioGraphics, March 1, 2006; 26(2): 453 - 464. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
