Abstract

Article Series: Dual Energy CT – Scientific Evidence and Clinical Application (4/7) – Abdominal Imaging

posted by Thorsten R. C. Johnson, M.D. | Dec 10, 2010

This article is part of the seven-article series on “Dual Energy CT – Scientific Evidence and Clinical Application”. T. Johnson, MD, talks about Dual Energy applications in abdominal imaging comprising imaging of the liver, the biliary system, the kidneys, the adrenal glands, pancreas, colon, aorta, and plaque imaging.

Abdominal Imaging

Liver Imaging

In abdominal imaging, iodine enhancement is the main diagnostic feature in the characterization of organ lesions, especially in the liver. Although a specific identification and quantification of the enhancement would be desirable, standard CT examination protocols mostly rely on one single venous phase to avoid the radiation exposure of multi-phase exams. With Dual Energy CT, it is feasible to identify iodine by its photo effect and to quantify the iodine-related density [72]. With this information, it is possible to subtract the iodine-related density from the CT image to generate a “virtual” non-contrast image from the same dataset. As the iodine related density is derived from the difference of two image datasets, the noise is doubled in this virtual non-contrast image. Although the image quality is therefore not exactly equivalent to a true pre-contrast acquisition, several studies confirmed that this virtual non-contrast image is sufficient to assess and quantify the enhancement of focal liver lesions [73-75].
The detection and quantification of iodine can also be used to assess vitality and monitor therapy response. For example, a study showed good results in evaluating liver lesions after radiofrequency ablation [76]. Limitations with the first generation DSCT were the limited field of view and the high noise in the 80 kVp image deteriorating image quality in large patients [8]. These issues are largely resolved with the second generation system and its spectral filter and larger field of view, as previously mentioned.

Apart from providing virtual non-contrast images and quantifying the iodine content of lesions, Dual Energy CT can also increase the conspicuity of lesions. Two studies showed that the contrast of hypervascular lesions is increased in low energy images and that an optimal compromise between contrast and image noise can be found by altering the weighting factor for average image generation [77-78]. Sigmoidal blending algorithms may be the best option for this purpose, combining the high contrast of low energy acquisitions with the low noise of high energy acquisitions [16].

Biliary system

Regarding gall stones, a clinically relevant difference is the composition which mostly contains calcium, cholesterol or pigment. As cholesterol stones can be dissolved medically, this would imply a non-invasive treatment option, while other types of stones would have to be removed mechanically if symptomatic. In vitro experiments have confirmed that Dual Energy CT can reliably differentiate cholesterol from other stones [79-80]. However, there are previous single energy CT studies that showed that this differentiation is also feasible based normal single energy CT density [81-82]. Therefore, the specific advantage of Dual Energy CT in this application is not evident.

Other researchers describe the use of iodinated biliary contrast material for cholangiography. Here, optimized blending or specific depiction of iodine can improve the visualization of the biliary system [83]. Still, the poor tolerance of biliary contrast agents [84] and the availability of MRCP as alternative reserves this technique to special indications such as living donor examinations in the preparation of liver transplantation.

Kidney Imaging

Figure 4 Multiplanar reformat with color coded superimposition of iodine signal confirming the absence of enhancement in cysts of variable density.

In kidney imaging, the assessment of contrast enhancement in small lesions is the most important diagnostic feature in the detection of cancer. If there are multiple cysts, it can be quite difficult or cumbersome to evaluate each lesion individually. However, especially patients with autosomal dominant polycystic kidney disease bear an increased risk of developing cancer. Therefore, most kidney CT protocols comprise an unenhanced, a contrast-enhanced, and a late excretory phase. Thus, it is attractive to use Dual Energy CT to specifically assess the iodine content of kidney lesions and to reduce radiation dose and postprocessing effort at the same time. Similar to the algorithms quantifying enhancement in liver lesions, iodine can be color-coded and quantified in the kidneys to differentiate hyperattenuating structures, e.g. hemorrhagic cysts, from enhancing tumors.
In an experimental study, the feasibility of differentiation of lesions containing contrast, protein, or blood was verified [85-86]. Clinical studies then confirmed the ability to distinguish hyperattenuating cysts from enhancing renal masses [87-88]. Further studies showed the advantages of the second generation DSCT with its tin filter and the large field of view. The former is necessary to obtain sufficient transmission and contrast to noise ratios at thin collimation in the abdomen; the latter is required in some patients to cover both kidneys entirely [89]. Recently, a study with histopathological correlation confirmed the ability of Dual Energy CT to predict the dignity of renal lesions [90].

Kidney Stone Differentiation

Figure 5 Color coding of kidney stones shows that the calculus in the right ureter consists of uric acid.

As iodine provides a very strong photoelectric effect, most clinical applications of Dual Energy CT are based on iodinated contrast material. However, the differentiation of kidney stones is a clinical application that is based on the spectral properties of the renal calculi themselves [91]. As Dual Energy CT requires a minimal dose that is higher than a low-dose single energy scan, quite a few institutions use a single energy low dose scan to detect any calculi and a second very short Dual Energy acquisition to cover just the stone [92]. This requires a physician to immediately screen the dataset for calculi while the patient is in the scanner, but works with least dose. Other institutions adopted an intermediate-dose protocol which uses less than a standard abdominal CT scan but more than a true single energy low dose scan [93].

The detailed knowledge of kidney stone composition can help to recognize the underlying disease and may in some instances guide therapy. Also, calculi consisting of uric acid can be dissolved by medication with Allopurinol and urine alkalization. Therefore, the postprocessing algorithm is tailored to differentiate uric acid from other types of stones. Another component of interest is struvite, because it indicates a bacterial infection which may require treatment. However, there is a spectral overlap between calcium and struvite so that this differentiation is not always feasible [94], while the differentiation of uric acid is reliable both in vitro [95-97] and in vivo [98-99]. The tin filter of the second generation DSCT improves the results in tiny calculi or large patients [100].

Apart from differentiating renal stones, Dual Energy CT is also used to identify calculi in contrast-enhanced urinary systems [101-102]. The detection of small stones below 3 mm diameter is limited due to volume averaging [103]. However, acute renal congestion represents a contraindication for the administration of iodinated contrast material as rupture of renal calices can be the consequence. Therefore, an unenhanced scan is acquired first at most institutions, and calculi can be differentiated or excluded before contrast material is administered.

Adrenal glands

Dual energy CT can also be used to characterize adrenal nodules [88]. A clinical study showed 50% sensitivity and 100% specificity in the differentiation of adrenal adenomas from metastases based on the differences in density in an unenhanced sequential Dual Energy acquisition [104]. Thus, MRI with opposed phase imaging and CT protocols including an early and delayed phase after contrast enhancement remain the reference standard in the differentiation of adrenal lesions. An algorithm relying on the identification of fat content in Dual Energy CT may be a further option but has not yet been investigated.

Pancreas

Figure 6 Color coded iodine image of the pancreas confirms absence of enhancement in a cystic lesion.

Similar to liver and kidney imaging, contrast enhancement plays the most important role in the detection and evaluation of pancreatic masses. By specifically color-coding iodine, Dual Energy CT can help to differentiate between normal and abnormal parenchyma [88, 105] or to confirm the absence of contrast enhancement in cystic lesions. A study in 15 patients with adenocarcinoma showed improved lesion conspicuity in the low energy images in portal venous phase.

Colon

Virtual colonoscopy is performed for screening for colorectal cancer in large trials. In this setting, the CT scan is performed at comparatively low dose and without intravenous contrast material. However, it is generally required to acquire two scans, one in supine and one in prone position, to evaluate the bowel wall in the area of remaining fluid levels. Also, virtual colonoscopy still requires complete cleansing prior to the examination, partially combined with ‘fecal tagging’, i.e. the administration of iodinated contrast material to opacify the residual fluid.
An approach with Dual Energy CT would be to eliminate iodine-containing voxels from the dataset based on the spectral information [10, 106]. Then, residual fluid is removed, so that a supine acquisition would be sufficient. If this ‘spectral cleansing’ is sufficient, it may be enough to have the patient drink iodinated contrast material with his meals but not actually cleanse the intestine. However, detailed large clinical studies with this technique will be required to confirm an equivalent diagnostic value as with ‘conventional’ virtual colonoscopy.

Aorta

Figure 7 Color coding of iodine and calcium confirms an endoleak in the thrombosed lumen of the aorta after endovascular repair.

Aortic imaging is generally performed with high injection rates of contrast material to achieve a strong opacification. With Dual Energy CT, the influence of the low energy acquisition on the image contrast can be increased, either by sigmoidal blending, by changing the weighting factor of average images, or by reading the low energy images. Then, the amount and injection rate of contrast can be decreased, still achieving an equivalent contrast to noise ratio [107].

Another application is the differentiation of calcium and contrast material, which can occasionally be quite difficult in the thrombosed lumen after endovascular aortic repair [108]. Several studies have shown that Dual Energy CT can help to differentiate calcifications from endoleaks [109-113]. However, most of these clinical studies applied a quantitative algorithm to color-code the degree of enhancement. Instead, there are other algorithms which aim to differentiate between calcium and iodine by their photoelectric effect and assign a different color to both (e.g. the ‘hardplaques’ algorithm), and these should be preferred for this purpose.

Plaque Imaging

Figure 8 Plaque differentiation based on Dual Energy information in a human aortic specimen shows cholesterol signal in some voxels in the center of large plaques.

Plaque imaging has been a challenging area of CT research, and it seems conceivable that Dual Energy CT may improve the differentiation of atherosclerotic plaque components. However, even aortic plaques are generally only represented by a few voxels in the dataset, and these voxels rarely consist of one single component. On the other hand, lipid and calcium as the main components have such a high and low CT density that Dual Energy techniques are not required to evaluate these plaque components. Thus, the aim would be to differentiate mid-density components such as fibrous tissue or thrombus.

So far, there is little scientific evidence on Dual Energy plaque imaging. Initial trials showed potential for an improved plaque differentiation [114-115] but note that translation into clinical application requires further developments. However, there are algorithms that aim to differentiate between contrast-opacified lumen of the vessel and hard plaques based on the spectral properties of iodine and calcium. These algorithms work well and identify the interface between calcified plaque and lumen quite precisely. This technique is especially helpful after endovascular aortic repair as it color-codes calcium and iodine differently so that calcifications in the thrombosed part of the lumen can be differentiated from endoleaks [108]. Studies aiming to apply Dual Energy techniques to differentiate soft plaque components of human atherosclerotic plaque did not yield convincing advantages so far [116-117].