Selective Arterial Calcium Stimulation With Hepatic Venous Sampling Differentiates Insulinoma From Nesidioblastosis
Scott M Thompson1, Adrian Vella1, Geoffrey B Thompson1
1Department of Radiology (S.M.T., J.C.A.), Department of Surgery (S.M.T., G.B.T., C.S.G.), Division of Endocrinology, Diabetes, and Metabolism (A.V., F.J.S.), and Department of Laboratory Medicine and Pathology (K.M.R.), College of Medicine, Mayo Clinic, Rochester, Minnesota 55905.
Context:
In adult patients with endogenous hyperinsulinemic hypoglycemia and negative or inconclusive noninvasive imaging, insulinoma and non-insulinoma pancreatogenous hypoglycemic syndrome (NIPHS) resulting from diffuse nesidioblastosis must be considered in the differential diagnosis. It is not known whether the biochemical results of selective arterial calcium stimulation (SACST) with hepatic venous sampling can differentiate insulinoma from diffuse nesidioblastosis.
Objective:
To determine the specificity of SACST with hepatic venous sampling in differentiating insulinoma from diffuse nesidioblastosis.
Design:
Retrospective review (January 1996 to March 2014).
Setting:
Tertiary referral center.
Patients Or Other Participants:
A total of 116 patients with biochemical evidence of endogenous hyperinsulinemic hypoglycemia and negative or inconclusive noninvasive imaging who were subsequently shown at surgery to have insulinoma (n = 42) or nesidioblastosis (n = 74) after undergoing SACST with hepatic venous sampling.
Intervention(S):
SACST with hepatic venous sampling before pancreatic exploration.
Main Outcome Measure(S):
Receiver operating characteristic curves were generated from the biochemical results of SACST to determine the specificity of the maximum hepatic venous insulin concentration (mHVI) and the relative-fold increase in hepatic venous insulin concentration (rHVI) over baseline after calcium injection from the dominant artery in differentiating insulinoma from nesidioblastosis.
Results:
The mHVI (21.5-fold; P < .001) and rHVI (3.9-fold; P < .001) were significantly higher in the insulinoma group compared to the nesidioblastosis group. The areas under the receiver operating characteristic curve for mHVI and rHVI were excellent (0.94; P < .0001) and good (0.83; P < .0001), respectively, for differentiating insulinoma from nesidioblastosis. mHVI cutoffs of > 91.5 and > 263.5 μIU/mL were 95 and 100% specific for insulinoma, respectively. A 19-fold increase in rHVI over baseline was 99% specific for insulinoma.
Conclusions:
These data suggest that the mHVI and rHVI at SACST may be useful in differentiating insulinoma from nesidioblastosis with high specificity in patients with hyperinsulinemic hypoglycemia and negative or inconclusive noninvasive imaging.
