Article Series: Dual Energy CT – Scientific Evidence and Clinical Application (6/7) – Current Application in Routine Clinical Practice
This article is part of the seven-article series on “Dual Energy CT – Scientific Evidence and Clinical Application” and gives an overview about the current Dual Energy applications in routine clinical practice.
Current Application in Routine Clinical Practice
Many Dual Energy CT protocols provide important diagnostic information and have not only been validated in scientific studies but also found widespread clinical application. These protocols have been implemented as standard settings in clinical routine and replaced single energy protocols for many applications. On the other hand, a few protocols lack evidence or do not provide sufficient clinical benefits to warrant general routine use.
Among the ones that have fundamentally changed the clinical diagnostic workup is the Dual Energy CT pulmonary angiography protocol. The simultaneous evaluation of pulmonary arteries and perfusion of the parenchyma increases the sensitivity for pulmonary embolism and provides additional pathophysiological information. Therefore, at our site, the exam is routinely acquired in Dual Energy technique and color-coded MPRs are always generated to visualize lung perfusion.
Another exam that is routinely performed in Dual Energy technique is the angiography of the runoff arteries. After Dual Energy bone removal, the MIPs provide an excellent overview of the entire vasculature and can save precious loading and reading time.
Regarding kidney imaging, there are two routine protocols in Dual Energy technique: Unenhanced low dose scans for the detection of kidney stones provide expedite stone characterization. The protocol for kidney tumor assessment contains a dual energy acquisition in venous phase and a low dose low kV acquisition in excretory phase. Thus, a pre-contrast scan can be omitted, reducing the overall radiation dose by about a third. Especially in the evaluation of polycystic kidneys, we appreciate the direct visualization of contrast enhancement and the easy differentiation from hemorrhagic cysts.
Metal artifact reduction also achieved such a good clinical performance that it was immediately implemented into routine protocols. Instead of high kV-imaging, we routinely reconstruct monoenergetic high energy images in which metal artifacts are significantly reduced.
Among the applications that have not achieved a role in routine clinical care is the differentiation of tendons and ligaments. Although the simultaneous evaluation for fractures and torn ligaments would make sense from a clinical perspective, the contrast to noise ratio is just too weak to make reliable diagnoses.
Also, cardiac imaging is routinely still performed as single energy exam, mostly as Flash Dual Spiral acquisition, because this protocol offers coronary CTA at unbeatably low dose. The simultaneous evaluation for coronary artery stenoses and myocardial ischemia would be desirable but imply a relevant dose penalty, and regarding the fact that coronary CTA should generally be used to rule out significant disease, myocardial perfusion is not of primary interest.
Lung ventilation imaging is also not routinely performed in daily practice, because Xenon gas administration still represents an off-label use. There are large trials confirming the safety of Xenon gas application at higher concentrations and for longer times in brain perfusion imaging, so this is not a major concern. As soon as sufficient study data is available to confirm the diagnostic value, ventilation imaging with a xenon respirator may be implemented as routine exam.
If you see a perfusion defect without identifiable PE on DE CTPA, might it not reflect gas trapping related vasoconstriction rather than an occult Subsegmental PE? How does do you distinguish between the two?
This differentiation can be quite difficult. However, if the perfusion defect is caused by air trapping, you would expect to see a corresponding decrease in density in the respective lobule in the normal average image in lung window and in both dual energy acquisitions. If the perfusion defect is caused by subsegmental PE, you would expect a normal “background” density in lung window.
So in doubt, you may get a clearer impression by measuring density values in “general viewing” and in the “lung PBV”, so you get the density in both acquitisions, the average “background” density and the “iodine related density”. The “iodine related density” will be decreased both in PE and air trapping with reference to surrounding normal lung parenchyma, but the “background density” should be normal in PE but decreased in case of air trapping. Also, lobular air trapping usually has causes like bronchiolitis that affect both lungs or whole lobes, so that lobular air trapping is quite unlikely to occur in a singular location, while this does occur with PE.
I hope these considerations are helpful.
thank you very much. I am wondering whether you ever make a diagnosis of PE based on a perfusion defect, in the absence of an identifiable arterial filing defect, or whether the perfusion scan just increased your diagnostic confidence if there is an equivocal filling defect.
I had imagined that the advantages of using FLASH mode to reduce repiratory and cardiac induced blurring of small vessels in the middle lobe. left lower lobe and lingula would be greater than any additional information provided by DE CTPA. DO you use FLASH for CTPA at all, and if so in what clinical setting?
Finally, how much of a limitation is the 33cm FoV for FLASH / DE modes in your experience?
- Yes, it does happen that you can’t identify an intravascular clot but do see a wedge-shaped perfusion defect, especially in small subsegmental locations and disseminated peripheral PE. And it does increase diagnostic confidence if there is an small but occlusive clot and a corresponding perfusion defect. On the other hand (and that’s what I like about it), if you see that at least 90% of the lung is normally perfused, you feel very comfortable at calling an exam negative, because even if you should have missed a tiny non-occlusive PE, it will not be relevant for the patient.
- Yes, our standard PE protocol is a DE protocol. This does allow to analyze lung perfusion but does not reduce motion artifacts. If this is your main concern, use the dual spiral (“flash”) mode. Then, the whole chest is scanned in 0.7 seconds, and there are absolutely no motion artifacts – but then there’s no dual energy information, because both single energy spiral paths fit into each other, so there’s no spectral data.
- The 33 cm diameter of the inner FoV does not represent a relevant limitation any more, in clear contrast to the 27.5 cm of the previous generation DSCT. Now with the 33 cm, virtually all patients’ bony thorax fits in, so the entire lung is included also with the basolateral parts. Previously, there had mostly been a 4-5 cm basolateral edge which had not been covered, but now this is resolved. You should still routinely use an a.p. and a lateral topogram to ensure correct positioning.
Thank you very much indeed. Just so I am absolutely clear, how do you report a scan in which you can’t identify an intravascular clot but do see on or more wedge-shaped perfusion defects? Will you call this as PE, or will you say “might be subsegmental PE”, or will you say “cannot exlude subsegmental PE”?
I assume you run into problems in the abdomen with 33cmFoV more oftern than in the chest?
If there’s a single wedge-shaped perfusion defect in an otherwise homogeneously perfused lung in a patient with clinical suspicion of PE, I’d clearly call this a single sub-segmental PE. If the perfusion of the remaining lung parenchyma is inhomogeneous, e.g. due to emphysema, COPD, fibrosis etc., I’d call it likely / suspicious of / potential PE, depending on the level of certainty.
Yes, 33 cm diameter is virtually always sufficient for the chest (at least inside the rib cage), and in the abdomen it always covers kidneys and also the ureters and bladder, so the evaluation of kidney tumors and renal stones, which represent the most frequent applications, is warranted in the vast majority of patients. Regarding the liver, e.g. for virtual non-contrast images and quantification of iodine content of focal lesions, there are patients in whom careful positioning would be essential to cover the entire organ and those who just don’t fit in. But in mid-European size patients this is rather uncommon (I’d guess 10-15%).