How is an Insulinoma diagnosed?

After the clinical suspicion of an insulinoma is raised, tests are employed to:

• confirm the presence of excessive insulin levels to make the biochemical diagnosis
• locate the tumour producing the insulin.

Confirming the biochemical diagnosis

Perhaps the most practically important tests are: a three day fast, a prolonged oral glucose tolerance test and calcium stimulation. Non-stimulated insulin and C-peptide levels are only interpretable in the presence of hypoglycaemia (glucose < 2.2mmol/l).

Measuring glucose and insulin levels
Measures of insulin levels and glucose levels by simple blood tests should show an excess insulin level (>30mU/ml) and a decreased glucose level (normal range being between 65 and 95 mg/dl). This is best represented as a ratio of insulin to glucose (I-G ratio). In affected individuals this ratio is generally ³ 1.0 where normally it is less than 0.4. Ideally, the sample should be taken during a hypoglycaemic episode, but due to the fact that most patients have intermittent occurrences, this is usually not possible.

Fasting test
In hospital, the patient is fasted for 72 hours to induce controlled hypoglycaemia in a safe setting until blood glucose falls below 50mg/dL (or until symptoms begin). In a normal individual, glucose and insulin levels would both fall in this scenario. In insulinoma the insulin alone remains abnormally elevated.

Pro-insulin and C peptide levels
When insulin is produced, proinsulin and C peptide molecules (that make up part of the precursor molecule of insulin) are released into the bloodstream. Measurements of both are increased in insulinoma.

C peptide suppression test
Administration of insulin normally causes a decrease in the production and release of C peptide into the blood, by a process of negative feedback. In patients with an insulinoma, because the gland is not responding normally to physiological signals, this suppression is not seen.

Tolbutamide test
This drug causes production and release of insulin from the pancreas. Normally, when 1g is infused over 2 minutes, it causes a decrease in glucose concentration of around 50%, followed by a return to normal of both insulin and glucose concentration after three hours. In 80% of patients with insulinoma, glucose levels remain lower and insulin levels a lot higher for the three hours.

Glucagon test
Normally, glucagon causes a raised glucose concentration with a return to normal after about three hours. In 72% of patients with insulinoma, there is an initial rise in glucose followed by a major drop. This is coupled with persistently elevated insulin levels.

Diagram of glucagon test - click to enlarge

Glucose Tolerance Test
Glucose is administered and levels measured over the next three hours. In normal individuals, the response is for glucose to rise transiently, fall to slightly below normal in response to insulin released from the pancreas and recover to normal values within three hours. In 60% of patients with insulinoma there is a much greater fall below normal and a slower recovery of glucose levels.

Calcium Infusion Test
An infusion of calcium gluconate causes a release of insulin, C peptide and proinsulin in patients with an insulinoma. This is not seen in normal patients.

Location of the tumour

CT, MRI and Ultrasound
Due to their small size, these imaging methods do not accurately detect insulinomas. Ultrasound scans pick up 50% of tumours (a sensitivity of 50%) but show a 25% false positive rate (i.e. one quarter of lesions picked up on ultrasound and identified as insulinomas will in fact be something else). CT and MRI both have similar sensitivity (about 35%), too low to rely on as localisation tools.

Angiography
Insulinomas are highly vascular tumours and tend to show a characteristic 'blush' (a region of diffuse vasculature) on the angiogram. This is approximately 80% accurate in localising the tumour.

Percutaneous trans-hepatic portal venous sampling
Blood samples are taken from several of the main veins draining the different parts of the pancreas (i.e. the head, neck, body and tail), measuring for insulin, C-peptide and pro-insulin. The highest concentrations indicate the veins that drain the part of the pancreas secreting the most insulin (i.e. the likely site of the insulinoma) giving a guide to where the tumour is found. This is 82% sensitive.

Venous sampling after arterial stimulation
Calcium is administered locally into arteries feeding specific regions of the pancreas. This should produce a fall in insulin and C-peptide.

Only if there is an insulinoma present in the area of pancreas supplied by that artery will insulin, C-peptide and proinsulin be produced, detected by blood samples taken from veins draining the pancreas. The main parts of the pancreas (e.g. the head, neck, body and tail) are stimulated separately, and the response measured, to allow some pre-operative localisation of the tumour. This technique is 88% sensitive.

Summary

None of these methods of localisation are ideal, the more accurate ones being more invasive and hazardous to the patient, involving many trained staff and great expense. Precise localisation is actually done during surgery, where, once the pancreas has been exposed and mobilised, intraoperative ultrasound and manual palpation by the surgeon are almost 100% sensitive at picking up the tumour.