| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Clinics for Pediatric Surgery (W.B., H.M.) and Pediatrics (O.B.), Institute for Pathology (M.K.), Charité University Medicine Berlin, D-13353 Berlin, Germany; Department of Anesthesiology and Intensive Care Medicine (C.H.), University Leipzig, D-04103 Leipzig, Germany; Diagnostic and Therapeutic Center Frankfurter Tor (W.M., T.E.), D-10243 Berlin, Germany; Institute of Radiopharmacy (F.F.), Forschungszentrum Dresden Rossendorf e.V., D-01314 Dresden, Germany; Institute for Human Genetics (B.L.-D.), HelmholtzZentrum muenchen, D-85764 Muenchen-Neuherberg, Germany; and Department of Pediatrics and Neonatology (K.M.), Otto von Guericke University Magdeburg, 39120 Magdeburg, Germany
Address all correspondence and requests for reprints to: Dr. Winfried Barthlen, Pediatric Surgery Charite University Medicine Berlin, Campus Virchow Klinikum, Augustenburger Platz 1, Mittelallee 8, D-13353 Berlin, Germany. E-mail: winfried.barthlen{at}charite.de.
| Abstract |
|---|
|
|
|---|
Objective: Our objective was to evaluate the predictive value and accuracy of integrated [18F]fluoro-L-DOPA ([18F]FDOPA) positron emission tomography (PET)-computed tomography (CT) for the surgical therapy of CHI.
Design: This was an observational study.
Setting: The study was performed in the Department of Pediatric Surgery at a university hospital.
Patients: From February 2005 to September 2007, 10 children with the clinical signs of CHI and an increased radiotracer uptake in a circumscribed area of the pancreas in the [18F]FDOPA PET-CT were evaluated.
Interventions: Guided by the [18F]FDOPA PET-CT report, all children underwent partial pancreatic resection, in two cases twice.
Main Outcome Measures: Correlation of the anatomical findings at surgery with the report of the [18F]FDOPA PET-CT, and the results of surgery and clinical outcome were determined.
Results: In nine children the intraoperative situation corresponded exactly to the description of the [18F]FDOPA PET-CT. A limited resection of the pancreas was curative in eight cases at the first surgery, in one case at the second intervention. We observed no diabetes mellitus or exocrine insufficiency in the follow up so far. In one child, hypoglycemia persisted even after two partial resections of the pancreatic head. Histological analysis finally revealed an atypical intermediate form of CHI.
Conclusions: The integrated [18F]FDOPA PET-CT is accurate to localize the lesion in focal CHI and is a valuable tool to guide the surgeon in limited pancreatic resection.
| Introduction |
|---|
|
|
|---|
Two forms are clinically indistinguishable: the focal form, which is found in approximately 30–50% of cases, is characterized by a localized adenomatosis of islet cells within an otherwise regular pancreas; whereas the diffuse form exhibits an islet cell hyperplasia throughout the entire gland (3).
The diffuse form is frequently inherited in an autosomal recessive mode. The focal form of CHI, on the contrary, is associated with a paternal ABCC8 or KCNJ11 germline mutation, whereas the second, maternal allele carries a deletion only in the adenomatous focus of the pancreas (4). However, in about 50% of all cases of CHI, a known mutation cannot be detected so far (2).
Clinically, neonates and infants show signs of hypoglycemia, such as lethargy, loss of consciousness, seizures, and poor feeding. Without prompt diagnosis and treatment, brain damage and severe mental retardation occur in approximately 40% of affected children (5).
Treatment requires a high glucose intake (>10 mg/kg·min), which usually results in an immense gain of body weight.
Medical therapeutic options are diazoxide, glucagon, octreotide, and nifedipine (2). Diazoxide action requires a reconstitutionable sulfonylurea receptor and, therefore, is frequently ineffective in the focal form. Glucagon and the somatostatin analog octreotide are effective in the majority of cases to prevent hypoglycemia. However, both drugs have to be administered parenterally.
Patients who do not respond to medical treatment or who suffer from serious side effects have been subjected to surgery. In the case of diffuse CHI, a subtotal (80–94% removal of the gland) or near-total (>95%) pancreatic resection has been proposed to prevent brain damage (6, 7). But only very few infants with the diffuse form have been cured surgically (8, 9), and the risk to cause iatrogenic diabetes mellitus or exocrine insufficiency is quite high (8, 9, 10).
However, the focal form can be completely cured by enucleation of the focal lesion or a limited pancreatic resection, thereby minimizing the risk of diabetes mellitus or exocrine insufficiency (11).
Therefore, it is of utmost importance to distinguish between the diffuse and focal form, and, in case of a focal form, to localize the focus exactly within the gland.
Until recently, the only diagnostic tools addressing these questions have been pancreatic venous sampling (12, 13) and the pancreatic arterial calcium stimulation test (11, 14, 15). However, both tests are technically very demanding and highly invasive, and even in the most experienced hands, never reached an accuracy above 80%. Computed tomography (CT) and magnetic resonance imaging scans alone cannot localize focal lesions because they lack morphological abnormalities.
Recently, the noninvasive positron emission tomography (PET) has been applied in CHI (16, 17, 18). Because islet cells metabolize L-DOPA to dopamine by decarboxylase action, the increased islet cell activity can be identified with (18F)fluoro-L-DOPA ([18F]FDOPA) (19). A diffuse accumulation of the radiotracer all over the gland indicates diffuse CHI and a "hot spot" the focal form. The synchronous acquisition of metabolic data by PET and anatomical data by CT allows the correlation with easily identifiable anatomical landmarks and can thus guide the surgeon to find a focal lesion, which as a rule, is hardly visible, even with magnifying glasses (3).
In this report we describe our single institution experience with [18F]FDOPA PET-CT guided surgery for focal CHI in 10 infants.
| Patients and Methods |
|---|
|
|
|---|
The diagnostic criteria included a hyperinsulinemic (>3 mU/liter plasma concentration), hypoketotic, hypofatty acidemic hypoglycemia (<2 mmol/liter), and a positive response to glucagon (1).
Patients
From February 2005 to September 2007, a total of 30 children fulfilling the diagnostic criteria for CHI had been investigated by [18F]FDOPA PET-CT at our institution. In 20 patients the accumulation of the radiotracer was demonstrated over the whole pancreatic gland, indicating diffuse CHI. To alleviate medical treatment, we performed a subtotal pancreatic resection in one of these children, which confirmed diffuse CHI. Because the others had not been subjected to surgery so far, there was no histological confirmation of the [18F]FDOPA PET-CT diagnosis in the majority of diffuse cases. These patients were manageable medically and were not the subject of this report.
However, in 10 other children, a defined "hot spot" could be detected. They underwent surgery, two of them twice.
Imaging procedure
The [18F]FDOPA PET-CT investigation followed a standard protocol as described previously (20). Until December 2005, all drugs were suspended at least 2 d before the investigation. Since 2006, only glucagon (and all other drugs affecting the glucagon or dopamine metabolism) was still suspended, but diazoxide, octreotide, or nifedipine was continued. Intravenous glucose was administered to prevent hypoglycemia. [18F]FDOPA was injected iv in a dosage of 4 MBq/kg, and a dynamic data acquisition was recorded from 5–60 min after injection in 5- to 10-min intervals. The upper and middle abdomen was examined in the axial section in slices of 3 mm. The images were reconstructed in the axial, coronal, and sagittal plane to obtain a three-dimensional impression.
For quantification, a standardized uptake value (SUV) was obtained. A SUV ratio was calculated using the formula: SUV focal divided by SUV mean of the other pancreatic tissue. A value more than 1.5 indicated focal disease (20). By this ratio, a focal lesion can be distinguished from diffuse accumulation and from normal uptake. In the simultaneously recorded CT scan, the splenic, superior, and inferior mesenteric vessels, the venous confluens, and the portal vein were visualized by classical angiography using iodine-contrasting agent.
Surgical procedure
At surgery, anesthesia was induced by inhalation, and maintained with desflurane and remifentanil (0.1–0.3 µg/kg·min). The glucose level was measured every hour.
After surgical exposition of the pancreas, three frozen sections were taken from the head, body, and tail to exclude diffuse disease (21, 22). Next, the gland was scrutinized carefully according to [18F]FDOPA PET-CT imaging. The focal lesion was removed with continuous frozen section control until exclusively normal islet cells in the remaining pancreatic tissue were found (21, 23). In case of involvement of the pancreatic duct, a Roux-en-Y loop was established for drainage.
Pathology
Tissue specimens of all cases were fixed with formalin and embedded in paraffin for definitive histological analysis. Three-micrometer thick hematoxylin/eosin-stained sections were examined to evaluate the following histological features: dimension of collected pancreatic tissue, architectural patterns and disorders of islets, cytomorphology, nuclear atypia, and fibrosis. Immunohistochemical studies were performed with antibodies against insulin and proinsulin. The Iview Ventana (Ventana Medical Systems, Inc., Tucson, AZ) with 3c3-diaminobenzidine-chromogen detection system was used for all antibodies (Anti-Insulin 1:100; BioGenex, San Ramon, CA; and Anti-Proinsulin 1:8000; Novocastra Vision BioSystems, Newcastle Upon Tyne, UK).
Mutation analysis
Genomic DNA was extracted from blood samples using standard procedures. The single exon of the KCNJ11 gene and the 39 exons of the ABCC8 gene were amplified with intronic primers and directly sequenced using a Big-Dye Terminator Cycle Sequencing Kit (Applied Biosystems, Foster City, CA) on an ABI PRISM 3730 Genetic Analyzer (Applied Biosystems).
Genomic DNA (
50 ng) was subjected to PCR amplification performed in a 25 µl volume containing 1 x PCR MasterMix (Promega Corp., Madison, WI), 0.25 µl each forward and reverse primer under the following cycle conditions:
Primer sequences are available on request. Each PCR amplicon was directly sequenced with forward and reverse primers using standard protocols.
Informed consent
Informed consent for participation in this observation study was obtained from the parents of each patient. The study protocol was reviewed and approved by the Ethical Committee and Review Board at Medical Faculty of the Otto von Guericke University Magdeburg, Germany.
Follow up
For follow up the patients were seen at least every 6 months in our endocrinological outpatient clinic. Regular glucose and glycosylated hemoglobin measurements and a neurological examination were performed.
| Results |
|---|
|
|
|---|
|
|
|
|
|
|
Eight children became euglycemic instantly after surgery. Today, with a follow-up between 1 and 32 months, all these children are on a normal diet. There is no medication, no hypoglycemia, no diabetes, and no exocrine pancreatic insufficiency.
However, in two patients hypoglycemia persisted after the first surgery.
In our patient no. 1, the [18F]FDOPA PET-CT showed increased uptake appearing late (after 45 min) in the processus uncinatus of the pancreatic head. This was interpreted as focal disease. The processus uncinatus was removed, and histopathological examination confirmed a focal lesion in the specimen. However, hypoglycemia and hyperinsulinism persisted. A second [18F]FDOPA PET-CT shortly after surgery could not be evaluated because of bile retention artifacts, and a third [18F]FDOPA PET-CT surprisingly now showed a diffuse pattern. At the second surgery, the frozen section biopsies from various locations of the remaining pancreas revealed the diagnosis of atypical intermediate CHI. This form combines hyperfunctional β-cells distributed within a pancreas containing normal (suppressed) islets as well. In our case the hyperactive cells were unevenly distributed within the gland with a high density of affected cells in the head and only few hyperactive cells in the tail, building a gradient of hyperactivity from the proximal (head) to the distal gland (tail), mimicking a focal lesion in the head. To alleviate medical treatment, a volume reduction of the pancreatic tissue was performed during the second surgery by resecting the remaining head and body. The child responds to diazoxide treatment and is well today without serious side effects except hypertrichosis.
The second child without an instant successful operation was our patient no. 6. The [18F]FDOPA PET-CT indicated focal disease in the pancreatic body adjacent to the head. At surgery a confined resistance could be palpated in this area, and the frozen section biopsy identified this tissue to be suspicious. However, interpretation of the islet cells was difficult because of mechanical artifacts. Because the presumed focal lesion laid deeply inside the pancreatic gland, an extended left resection of the pancreatic tail and body was performed to avoid Roux-en-Y drainage. However, after surgery the child continued to be hypoglycemic, necessitating hyperalimentation, octreotide, and glucagon again. Immunohistochemical-stained sections failed to demonstrate increased insulin production in the presumed focal lesion in the surgical specimen. A second [18F]FDOPA PET-CT 5 months later demonstrated the focal lesion at the same localization as before. At second surgery, the focal lesion was found a few millimeters away from the resection margin inside the gland and resected. The child got euglycemic a few days later.
| Discussion |
|---|
|
|
|---|
Because the normal pancreatic lobular structure is preserved in focal CHI (23), the focal lesions are seldom identifiable macroscopically, and visual detection rates of over 60% (6, 11) seem to be extraordinary. The previously used diagnostic tools, pancreatic venous sampling and the pancreatic arterial calcium stimulation test, have been abandoned in most centers because of the technical difficulty, their side effects, and their low sensitivity and specificity (22, 26).
Therefore, the development of the PET technique has been a great step forward. In a recent study, [18F]FDOPA PET was able to distinguish correctly between nine children with diffuse and five children with focal CHI (16). This high accuracy was confirmed by the French group (18). All 15 focal cases and nine of 34 children with diffuse accumulation underwent surgery, and in 21 of 24 operated children, the histopathological results confirmed the [18F]FDOPA PET findings.
Furthermore, the sensitivity of the PET technique in CHI is illustrated by case reports demonstrating ectopic focal lesions in the duodenal wall (17) and the jejunum (27).
[18F]FDOPA PET scanning produces a bright spot in the focal form, but the exact anatomical localization for the surgeon is poor (16, 28). Therefore, the additional acquisition of the anatomical data by CT has been of great value. In the only study published so far comparing the [18F]FDOPA PET-CT findings in CHI with the histological report, the [18F]FDOPA PET-CT was accurate to distinguish between focal and diffuse disease in 44 of 50 children (88%) and was 100% accurate in localizing the focal lesion (29).
However, there are some pitfalls. In our patient no. 1, there was a gradient of adenomatosis declining in activity from the pancreatic head to the body and tail. Because the density of the affected cells in the head was much higher than in the tail, the picture mimicked a suppression of β-cell activity as typical for the focal form of CHI. After resection of the accumulation in the uncinate process, the diffuse involvement of the remaining pancreas became apparent. There are only a few cases described so far that really showed two or more independent focal lesions within an otherwise normal pancreas (27, 28, 30). Most cases of "newly " detected focal lesions after surgery are probably either remaining adenomatosis after an incomplete resection or unevenly distributed hyperactivities within the intermediate form of CHI.
In addition, there is another possible source of error in interpretation of [18F]FDOPA PET-CT imaging. It cannot be totally excluded that the late appearing activity (>45 min) in our patient no. 1 was at least partially related to radiotracer elimination in the bile and accumulation in the choledochal duct. Most of the radioactivity injected is excreted by the kidneys, but variable uptakes have been observed in the liver, gallbladder, biliary tract, and duodenum as well (18). This possible pitfall of [18F]FDOPA PET-CT imaging has been described before (31). After our patient no. 1, the interpretation of [18F]FDOPA PET-CT has been changed in our institution to exclude bile-related artifacts. Since then, late-appearing focal radiotracer accumulations are only interpreted as focal lesions if they have been proven at the identical position on a second scan with at least a 15-min time difference.
In our second problem patient (no. 6), the interpretation of the frozen section was difficult. Enlarged islets with nuclear atypia suspicious of a focal lesion could not be confirmed after paraffin embedding. Clinically hyperinsulinism persisted. This patient was treated in the early stage of our study, too. Comparable situations are reported from the Childrens Hospital of Philadelphia (CHOP) (21). In one patient 13 intraoperative frozen section biopsies were interpreted as focal disease. However, routinely processed sections revealed diffuse CHI and hypoglycemia persisted even after a 75% pancreatic resection (21). In two other reported patients, who remained hypoglycemic after initial surgery with extensive frozen section biopsies, the resection specimen of the second surgery still contained residual focal lesions.
The fact that the frozen section report can differ profoundly from the definitive histology is a well-recognized problem in CHI (21, 22, 23). Smith et al. (22) demonstrated an interobserver discrepancy in CHI on permanent sections of 12% among experienced pathologists. In the CHOP study (21), in 17 of 18 diffuse but in only 26 of 30 focal cases, the intraoperative frozen section corresponded with the final report. Furthermore, there have been some doubtful cases showing the presence of large islet cell nuclei in areas nonadjacent to the focal lesion or a typical diffuse pattern in only confined areas of the pancreas (21).
In conclusion, results from frozen sections in CHI must be cautiously interpreted. Immunohistochemical staining confirms the diagnosis easily but takes its time and is not available during surgery.
Many focal lesions have ill-defined borders with tentacle-like offshoots. Therefore, we take multiple frozen sections from the remaining pancreatic tissue as recommended by other groups (11, 32, 33). Furthermore, the rare but existing possibility of a second or third focus within the pancreas must be considered to be a possible reason for failure of surgery (27, 30).
The intermediate form of CHI, which was the final diagnosis in our patient no. 1, has been diagnosed frequently in Australia (34): 24 of 31 patients were allocated to have diffuse CHI, but in nine of these 24 diffuse cases (38%), additionally defined areas of focal adenomatosis have been described. However, because the histopathological analysis is not standardized yet, and the range of individual interpretation is wide (21, 22, 23), there is the possibility of an observer bias. The fact that only two of 31 (6%) patients in the Australian study (34) exhibited the classical focal pattern underlines this probability. The value of a limited volume reduction of pancreatic tissue in intermediate CHI to alleviate medical treatment must be further evaluated.
All the children described here have been operated upon conventionally. In single cases children reportedly had surgery by laparoscopy (35). Because of the well-known advantages, i.e. less surgical trauma, small scars, and better visualization, it seems worthwhile to consider the laparoscopic option if the localization of a focal lesion is in the distal pancreas (36). Promising experiences have been made in a few centers and by us.
So far, we did not observe either persistent exocrine pancreatic insufficiency or diabetes mellitus. However, our follow-up time is relatively short, and diabetes can develop many years after resection (8, 9, 37).
Comparison of our results with previous series underlines the advantage of a concise diagnostic workup in CHI before and during surgery. The London group (26) reports about a high rate (88%) of extensive (>95%) pancreatic resections with subsequent problems like choledochal injury (17%) and diabetes mellitus (42%). In the German group (37), 60 of 63 children have been subtotally or near totally resected, resulting in persistent hypoglycemia in 40% and diabetes mellitus in 27% of cases. In the Paris group, the success rate after partial resection for focal CHI was 98% (33), but only 7% after near-total pancreatectomy for diffuse disease (32).
The CHOP group (11) reports about a complete response to surgery in 92% of children with focal CHI, but in only 11% of cases with diffuse CHI (21). Interestingly, in many cases the focal lesion extended into the portion of the pancreatic head, close to the duodenum, which is normally not removed even during a 95% resection.
Conclusions
[18F]FDOPA PET-CT is a valuable tool to distinguish between focal and diffuse forms of CHI, and in the latter case, to localize the focus for limited pancreatic surgery. However, the interpretation of the [18F]FDOPA PET-CT imaging as a dynamic process is sophisticated, and the presence of an experienced pathologist during surgery is mandatory.
| Footnotes |
|---|
First Published Online December 11, 2007
1 W.B. and O.B. contributed equally to this work. ![]()
Abbreviations: CHI, Congenital hyperinsulinism; CHOP, Childrens Hospital of Philadelphia; CT, computed tomography; [18F]FDOPA, [18F]fluoro-L-DOPA; PET, positron emission tomography; SUV, standardized uptake value.
Received September 10, 2007.
Accepted December 5, 2007.
| References |
|---|
|
|
|---|
This article has been cited by other articles:
![]() |
N. Khen-Dunlop, C. Capito, V. Valayannopoulos, C. Elie, M.-J. Ribeiro, J. Rahier, F. Jaubert, J.-J. Robert, Y. Aigrain, P. de Lonlay, et al. Predictive Value of Postoperative Glycosuria After Partial Elective Pancreatectomy in Focal Congenital Hyperinsulinism Diabetes Care, October 1, 2008; 31(10): e71 - e71. [Full Text] [PDF] |
||||
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |