PURPOSE: To retrospectively compare
fluorine 18 ((18)F) fluoro-L-dopa
positron emission tomography (PET) and pancreatic
venous sampling (PVS) in the preoperative differentiation of diffuse from focal
congenital hyperinsulinism (CHI) and localization of focal
lesions. MATERIALS AND METHODS: This study was approved by the institutional ethical committee, and informed consent for the research study was obtained from the parents of all subjects. Fifty-one patients evaluated for focal CHI between January 1, 1995, and January 31, 2008, were included. Thirty five underwent PVS evaluation alone, and 16 underwent a PET evaluation alone. The sensitivity values of each technique for the diagnosis and localization of focal
lesions were compared in regard to results of
surgery and pathologic analyses. In each patient, perioperative treatment was reviewed, and the presence of postoperative
hypoglycemia was assessed as evidence of incomplete resection. Comparisons of the sensitivity values and recurrence rates were performed by using the
Fisher exact test in regard to the number of patients. Comparisons of
median age, weight, or number of
biopsies were performed with a two-tailed unpaired
Mann-Whitney U test. A difference with P < .05 was considered significant. RESULTS: For PVS and PET groups, there was no error in differentiating focal from diffuse forms. PVS was not completed in four of 35 patients. In 27 (87%) of 31 patients in whom PVS was completed and 13 (81%) of 16 patients in whom PET was completed, preoperative localization of the focal
lesion was in accordance with the
surgical findings (P = .7). Although not significant, the number of
biopsies performed before discovering the focal
lesion was higher in the PET group compared with the PVS group (P = .06). Inadequate localization occurred in two (6%) patients in the PVS group and five (31%) patients in the PET group at initial preoperative imaging study; these patients underwent repeat
surgery for residual CHI (P = .03). CONCLUSION: (18)F-fluoro-L-dopa PET is equivalent to PVS in the characterization of CHI but does not provide localization of the
lesion as precisely as does PVS.
