Tumor localization is important in gastrinoma patients, as it may have a direct bearing on patient management. A variety of non-invasive and invasive techniques have been used, with varying levels of success (see table 2). Non-invasive techniques include ultrasound, CT scan, MRI, and most recently somatostatin receptor scintigraphy. Transabdominal ultrasound[1, 35] and CT[35] have proven to be of limited value, as these tumors are frequently small and below the resolution of the techniques. In addition, the density of these tumors is similar to that of surrounding pancreatic tissue, making them difficult to detect. MRI has not been shown to be superior to the previous techniques [36, 37]. Perhaps the greatest value of CT scan is that it identifies most patients with liver metastases.
Table 2. Methods of Tumor Localization in Gastrinoma
|
PTHVS |
||||||||
|---|---|---|---|---|---|---|---|---|
|
Factor |
Ultrasound |
Infusional CT |
Computed tomography |
SRS |
Selective angiography |
Secretin angiography |
Local |
Regional |
|
PTHVS = percutaneous transhepatic portal, pancreatic, and hepatic venous gastrin sampling. Data from Norton and colleagues,[40] Jensen[41], as well as Vinik and colleagues [42-44]. |
||||||||
|
Sensitivity |
21 |
40 |
31 |
71 |
60, 29 |
89 |
35 |
94 |
|
Specificity |
92 |
100 |
66 |
86 |
100, 100 |
100 |
89 |
97 |
|
Positive predictive value |
80 |
100 |
83 |
85 |
100, 100 |
--- |
--- |
94 |
|
Negative predictive value |
44 |
50 |
15 |
52 |
60, 100 |
--- |
89 |
--- |
Some have advocated endoscopic ultrasound. In experienced hands, localization rates of 80-100% in patients with pancreatic gastrinomas have been reported [38, 39]. However, the ability to detect small lesions less than 5 millimeters in size, or occult duodenal lesions, is uncertain. The pancreatic head is examined with the scanner positioned in the duodenum, and the body and tail of the pancreas are studied with the scanner in the stomach. Some have used a special saline filled balloon at the tip of the instrument with installation of approximately 400 cc of saline into the stomach to provide an interface between the ultrasonic unit and the stomach wall [38, 39].
The development of somatostatin receptor scintigraphy has been an important advance in the imaging of pancreatic neuroendocrine tumors, and specifically gastrinoma. Somatostatin receptor scintigraphy should be obtained in all patients with a diagnosis of gastrinoma. The results with this technique approach those obtained with more invasive localization methods (see table 2). A recent study of 146 patients by Gibril and colleagues from the National Institutes of Health revealed a sensitivity of 71%, specificity of 86%, and a positive predictive value of 85%, higher than any other non-invasive modality [45]. Somatostatin receptor scintigraphy altered management in 47% of patients overall. Like most imaging modalities, somatostatin receptor scintigraphy tends to miss small duodenal tumors less than 1cm in diameter [46]. False positives do occur in about 10% of patients, thus the clinical context must be considered and the results interpreted carefully. Our current approach is to combine somatostatin receptor scintigraphy, along with either CT scan or MRI, to provide anatomic information and to help exclude liver metastases.
With the advent of somatostatin receptor scintigraphy, there is currently less need for the use of invasive techniques. The invasive techniques are highly accurate, however they involve discomfort, considerable expense, and require the expertise of an experienced angiographer. Selective visceral angiography has been shown to be useful for localizing tumors in patients with gastrinoma, as well as identify occult metastatic disease. The addition of intra-arterial injection of secretin to selective angiography greatly improves the accuracy of the technique [3, 47]. This requires the insertion of a catheter in a hepatic vein and a peripheral vein to sample gastrin levels. When secretin is injected into a vessel that supplies the gastrinoma, the hepatic vein levels of gastrin greatly exceed the peripheral levels. It may also increase the number of patients who show a tumor blush on angiography. This technique may be more accurate than percutaneous transhepatic venous sampling (PTHVS) alone [47].
Percutaneous transhepatic venous sampling (PTHVS) is a highly accurate technique, however it is both costly and time consuming [31]. This technique has been shown to be of value in those with a primary gastrinoma not detected by conventional imaging methods [37, 48-51]. It has clearly been shown that PTHVS is a useful method to regionalize a tumor. However, the success of this technique is highly operator dependent, and requires a detailed understanding of the variable venous anatomy present in the region [24, 50, 51]. A variety of technical issues, including placement of cannulae to avoid vessel obstruction, and streaming of portal blood flow, can affect the results of the study. Because of potentially rapid changes in the secretory rate of gastrinomas, it is important to simultaneously measure gastrin in pancreatic veins and central arteries, to determine gradients.