Approach Considerations
The classic hallmark of insulinomas is the Whipple triad, which consists of the following: [18]
-
Symptoms of hypoglycemia, especially with fasting or exercise
-
Documented low blood sugar at the time symptoms are present
-
Reversal of symptoms by glucose administration
Common diagnostic criteria include documentation of blood glucose level < 50 mg/dL (2.775 mmol/L) with hypoglycemic symptoms, with relief of symptoms after eating, and absence of sulfonylurea on plasma assays. The classic diagnostic criteria also include demonstration of the following during a supervised fast:
-
Increased plasma insulin level (≥6 μU/ml)
-
Increased C peptide level (≥0.2 nmol/l)
-
Increased proinsulin level (≥5 pmol/l)
Once the diagnosis of insulinoma has been established biochemically, imaging studies are used to localize the tumor.
Laboratory Studies
Failure of endogenous insulin secretion to be suppressed by hypoglycemia is the hallmark of an insulinoma. Thus, the finding of inappropriately elevated levels of insulin in the face of hypoglycemia is the key to diagnosis. Considering the reference range, the fasting plasma levels of insulin, C-peptide, and, to a lesser degree, proinsulin need not be elevated in insulinoma patients in absolute terms.
A prolonged supervised fast in hospitalized patients provides the most reliable results. During this test, levels of plasma glucose, insulin, C peptide, and proinsulin are measured every 6 hours until the plasma glucose level falls to 60 mg/dl or less, at which point the testing interval is reduced to every 1–2 h. The fast is terminated when the plasma glucose level is 45 mg/dl or less and the patient has symptoms and signs of hypoglycemia
Traditionally, a 72-hour fast has been used. However, with current assays, 90–95% of insulinomas can be diagnosed with a supervised fast of 48 hours. [18]
The calculation of ratios of insulin (µU/mL) to plasma glucose (mg/dL) during the fast is diagnostic: Healthy patients maintain a ratio of less than 0.25 (obese patients may have a slightly higher rate), whereas in patients with insulinoma, the ratio rises during fasting.
In a study from the Netherlands, a positive Whipple triad on a prolonged fasting test, in combination with an insulin/C-peptide ratio < 1, had a sensitivity of 88.9% and a specificity of 100% for the diagnosis of insulinoma. [34]
The biochemical diagnosis is established in 95% of patients during prolonged fasting when the following parameters are found:
-
Serum insulin levels of 10 µU/mL or more (normal < 6 µU/mL)
-
Glucose levels of less than 40 mg/dL
-
C-peptide levels exceeding 2.5 ng/mL (normal < 2 ng/mL)
-
Proinsulin levels ≥22 pmol/L, [35] or greater than 25% (or up to 90%) that of immunoreactive insulin
If these studies confirm elevated endogenous insulin, insulin antibody levels should also be checked to rule out autoimmune disease.
Stimulation tests are no longer recommended. The intravenous administration of tolbutamide, glucagon, or calcium can be hazardous, as it may induce prolonged and refractory hypoglycemia.
The presence of multiple endocrine neoplasia type 1 (MEN 1) must be evaluated by excluding the following:
-
Hyperprolactinemia due to a pituitary adenoma
-
Hyperparathyroidism due to parathyroid hyperplasia
-
Hypergastrinemia due to a gastrinoma
Imaging Studies
Start imaging studies only after the diagnosis has been confirmed biochemically, because 80% of insulinomas are less than 2 cm in size and may not be visible by computed tomography (CT) scanning or transabdominal ultrasonography.
Successful preoperative tumor localization is achieved in about 60% of patients. [36] Some experienced surgeons perform only transabdominal ultrasound preoperatively. Other surgeons argue that the preoperative localization of insulinomas is not necessary at all because surgical exploration and intraoperative ultrasonography identify more than 90% of tumors. [37] Thus, the extent to which one attempts to define the anatomy of the beta cell lesion before surgery is a matter of judgment.
Computed tomography and magnetic resonance imaging
Helical or multislice CT has 82-94% sensitivity. In one study, dual-phase helical CT proved more sensitive than single-phase for detection of insulinoma; in addition, image acquisition in the arterial phase proved more helpful than acquisition during the portal-venous phase. [38] See the image below.
CT scan image with oral and intravenous contrast in a patient with biochemical evidence of insulinoma. The 3-cm contrast-enhancing neoplasm (arrow) is seen in the tail of the pancreas (P) posterior to the stomach (S) (Yeo, 1993).
Zhu et al reported that the addition of whole-pancreas perfusion to biphasic contrast-enhanced CT (CECT) may improve the detection of insulinomas. In their study of 79 consecutive cases, mean blood flow values of both insulinomas and insulinoma-harboring regions were statistically significantly higher (P < 0.01, for both) than those of tumor-free pancreatic parenchyma. The addition of perfusion allowed correct diagnosis of nine of 46 tumors (19.6%) in which biphasic CT was negative or indeterminate. [39]
In a subsequent prospective study, however, Zhu et al reported that volume perfusion CT (VPCT) is significantly more accurate than CECT for insulinoma detection, and 3 Tesla magnetic resonance imaging (3T MRI) with diffusion-weighted imaging (DWI) is better still. Lesions were more conspicuous on MRI than on VPCT, and conspicuity with both was better than with CECT. Tumor-to–pancreatic duct distance was better appreciated on MRI than on VPCT and CECT. [6]
Arteriography
The accuracy of selective arteriography varies but is around 60% in the best hands. Arteriography with catheterization of small arterial branches of the celiac system combined with calcium injections (which stimulate insulin release from neoplastic tissue but not from normal islets), and simultaneous measurements of hepatic vein insulin during each selective calcium injection localizes tumors in 90% of patients. [7]
Nuclear medicine
Insulinomas have been shown to have a very high density of glucagon-like peptide–1 (GLP-1) receptors, and radiolabeling GLP-1 analogs (eg, exendin-4) with appropriate radioisotopes can provide noninvasive localization of benign insulinomas. [40, 41] For example, indium-111 (111In)–labeled exendin-4 (111In-DOTA-exendin-4) has successfully been used to localize small insulinomas both preoperatively and, with the use of a gamma probe, intraoperatively. [42] Luo et al reported that PET/CT scanning with gallium-68 (68Ga)-NOTA–exendin-4 correctly detected insulinomas in 42 of 43 patients. [43]
Malignant insulinomas can be missed by GLP-1 target imaging because these tumors often lack GLP-1 receptors. However, these tumors often express somatostatin receptor subtype 2 (SSTR2), which can be detected with indium-111 (111In)-pentetreotide single-photon CT (SPECT). [44, 45, 46] Nockel et al studied the accuracy of 68Ga–DOTA-(Tyr3)-octreotate (Ga-DOTATATE) PET/CT for localizing insulinoma and suggested that this technique may be considered as an adjunct imaging study when all other imaging studies are negative and when a minimally invasive surgical approach is planned. [8]
Ultrasonography
Endoscopic ultrasonography (EUS) detects 77% of insulinomas in the pancreas. [3, 4, 5] Combining EUS with CT can increase the yield, resulting in the detection and localization of the majority of sporadic insulinomas.
See the image below.
Endoscopic ultrasonography in a patient with an insulinoma. The hypoechoic neoplasm (arrows) is seen in the body of the pancreas anterior to the splenic vein (SV) (Rosch, 1992).
Other Tests
Endoscopic ultrasound–guided fine-needle aspiration biopsy of insulinoma has been described. However, biopsy cannot distinguish benign from malignant insulinoma, and preoperative tissue sampling is usually not necessary.
Histologic Findings
Insulinomas are solitary tumors in 90% of patients. In MEN 1 syndrome, multiple microinsulinomas and macroinsulinomas are found, although hypoglycemia may be caused by a single tumor. The tumors are distributed evenly throughout the pancreas. Tumor size does not relate to the severity of clinical symptoms. Ectopic insulinomas may be found in the ligament of Treitz.
No histologic criteria are available to distinguish benign from malignant insulinomas. Malignant tumors are usually larger (average size 6.2 cm), and a third of them have metastasized to the liver. Insulinoma tumor cells contain less insulin and secretory granules than normal B cells but higher levels of proinsulin. Atypical granules, or even agranular cells, are frequent. The clinical response to diazoxide and somatostatin correlates with the frequency and type of granules.
No histologic criteria are available to distinguish benign from malignant insulinomas. Malignant tumors are usually larger than benign ones (approximately 4.2 versus 1.8 cm, respectively). Insulinoma tumor cells contain less insulin and secretory granules than normal B cells but higher levels of proinsulin. Atypical granules, or even agranular cells, are frequent. The clinical response to diazoxide and somatostatin correlates with the frequency and type of granules. [47]
Staging
A specific tumor-node-metastasis (TNM) staging system has not been validated yet for insulinomas. The American Joint Committee on Cancer (AJCC) staging for well-differentiated neuroendocrine tumors arising in the pancreas is used. [48, 49]
Primary tumor (T):
-
TX - Tumor cannot be assessed
-
T1 - Tumor limited to the pancreas and < 2 cm in size
-
T2 - Tumor limited to the pancreas and 2−4 cm in size
-
T3 - Tumor limited to the pancreas and > 4 cm in size; or tumor invading the duodenum or common bile duct
-
T4 - Tumor invading adjacent organs (stomach, spleen, colon, adrenal gland) or the wall of large vessels (celiac axis or the superior mesenteric artery)
Regional lymph Nodes (N):
-
NX - Regional lymph nodes cannot be assessed
-
N0 - No regional lymph node involvement
-
N1 - Regional lymph node involvement
Distant metastasis (M):
-
M0 - No distant metastasis
-
M1 - Distant metastases
-
M1a - Metastasis confined to liver
-
M1b - Metastases in at least one extrahepatic site (eg, lung, ovary, nonregional lymph node, peritoneum, bone)
-
M1c - Both hepatic and extrahepatic metastases
AJCC prognostic stage groups are shown in the table below.
Table. AJCC Prognostic Stage Groups (Open Table in a new window)
Stage |
T |
N |
M |
I |
T1 |
N0 |
M0 |
II |
T2 |
N0 |
M0 |
T3 |
N0 |
M0 |
|
III |
T4 |
N0 |
M0 |
Any T |
N1 |
M0 |
|
IV |
Any T |
Any N |
M1 |
-
CT scan image with oral and intravenous contrast in a patient with biochemical evidence of insulinoma. The 3-cm contrast-enhancing neoplasm (arrow) is seen in the tail of the pancreas (P) posterior to the stomach (S) (Yeo, 1993).
-
Endoscopic ultrasonography in a patient with an insulinoma. The hypoechoic neoplasm (arrows) is seen in the body of the pancreas anterior to the splenic vein (SV) (Rosch, 1992).
Tables
Stage |
T |
N |
M |
I |
T1 |
N0 |
M0 |
II |
T2 |
N0 |
M0 |
T3 |
N0 |
M0 |
|
III |
T4 |
N0 |
M0 |
Any T |
N1 |
M0 |
|
IV |
Any T |
Any N |
M1 |
