Optimizing the Bolus Trigger Threshold for Dual-Energy CT Angiography

See also the article by Noda et al in this issue.

Dr Malayeri is a staff clinician and chief of the body imaging section in the Department of radiology and imaging sciences at the National Institutes of Health. Dr Malayeri is the recipient of grant funding from the RSNA Research and Education Foundation and the NIH Research Award for Staff Clinicians Program (RASCL). His research interests focus on optimizing MRI and CT techniques for clinical problem-solving.

In this issue of Radiology, Noda and colleagues (1) present the results of a phantom study and subsequent randomized clinical trial to optimize the bolus trigger threshold (BTT) for dual-energy CT angiography (DECTA) in patients with various aortic abnormalities. The authors found that the optimal threshold was 30 HU for 40-keV virtual monochromatic (VM) imaging of the aorta and its branches. They also demonstrated that a more than 50% reduced dose of injected iodinated contrast material with an optimized threshold and 40-keV VM images did not lower the diagnostic acceptability of DECTA compared with either standard single-energy CT angiography (SECTA) or DECTA with conventional BTT and 50% reduced contrast material dose.

Although dual-energy scanners are not novel to our field, there is a general lack of consensus regarding how to best use this technology in clinical practice. A specific advantage of dual-energy CT is its ability to help determine the atomic number of the imaged material, and hence the ability to perform material decomposition and provide more detailed tissue characterization. One can take advantage of this capability to determine the amount of iodine-based contrast material in a given volume of the tissue or to improve the detection of small masses (2).

The other advantage of dual-energy CT is the use of VM images that allow image reconstruction under the assumption that the x-ray source produced x-ray photons at only a single energy level. The closer the energy level gets to the K edge of iodine (33 keV), the more photons undergo photoelectric absorption, resulting in an increase in iodine attenuation. Contrast material attenuation exponentially increases with the decreasing kiloelectric volt levels such that aortic enhancement on 40-keV VM images is nearly twofold that on 60-keV VM images (3). This phenomenon can be used to improve visualization of the vascular structures at CT angiography to evaluate aneurysm, stenosis, surgical planning, or complications from endovascular or surgical repair of aneurysms.

Because reconstructed lower-energy VM images lead to greater vascular enhancement, it is important to understand if the amount of intravenous contrast material could be decreased for proper evaluation of the vasculature. Others have investigated the use of lower contrast material volume in lower-energy VM images while preserving the diagnostic quality of the CT angiograms (4,5), but what if the amount of contrast material can be optimized based on desired enhancement level on monochromatic images? Do we need to adhere to the conventional 100-HU trigger threshold, or can this be adjusted based on the desired VM reconstruction setup?

To answer these questions, Noda et al (1) set up an elegant study design to determine the optimal trigger threshold for DECTA when VM images are reconstructed at 40 keV. The current standard practice is to set the trigger threshold at 100 HU. This threshold is achieved by monitoring the contrast enhancement within the aorta. In SECTA, the monitoring is performed by using a single tube at 120 kVp, whereas in DECTA with rapid kilovolt switching, the tube is run in single-energy mode at 140 kVp. To identify the equivalent enhancement level to 100 HU at 120 kV, the authors performed a phantom study in which they scanned different iodine concentrations in dual-energy mode. They then performed several VM reconstructions at 5-keV increments between 40 keV and 70 keV. The optimized trigger threshold was determined to be 30 HU using a 140-kVp tube for BTT monitoring that results in 100 HU when images are reconstructed in 40-keV VM reconstruction, which also corresponds to 100 HU at 120 kVp single-energy CT.

Based on the results from the phantom study, a total of 96 clinical trial participants were randomly divided into three groups. In the first group, 32 participants received the full dose of contrast material (600 mg of iodine per kilogram) and underwent conventional SECTA (BTT monitoring performed at 120 kVp) with a threshold set at 100 HU. The second cohort of 32 patients received half of the dose of contrast material (300 mg of iodine per kilogram) and underwent DECTA with BTT at 100 HU (BTT monitoring performed at 140 kVp). The third cohort of 32 patients underwent DECTA with half of the dose of contrast material (300 mg of iodine per kilogram) and an optimized trigger threshold of 30 HU (BTT monitoring performed at 140 kVp). Noda et al were able to demonstrate significantly less contrast material volume and total iodine dose in the optimized DECTA group (43 mL and 13 g, respectively) compared with the conventional DECTA (63 mL and 19 g) and SECTA (87 mL and 26 g) groups. The total iodine dose in the optimized DECTA group was 55% and 32% lower than that in the SECTA and conventional DECTA groups, respectively.

As expected, despite the lower contrast material volume and dose using optimized DECTA, the enhancement levels for different aortic regions were significantly higher in the conventional DECTA group than in the optimized DECTA group, and were greater in the optimized DECTA group than in the SECTA group.

More importantly, when it comes to radiologist interpretation of these cases, the depiction of the major arteries and smaller branches was superior with both DECTA techniques compared with the SECTA technique. For smaller arteries in the abdomen and pelvis, optimized DECTA outperformed conventional DECTA and SECTA. This superior capability of low VM images derived from DECTA in depicting the arteries can be attributed to higher contrast-to-noise ratio in the DECTA technique (6). The overall diagnostic acceptability of all three methods was similar when the axial or multiplanar reformatted images were reviewed.

It appears that the DECTA technique with optimized trigger threshold for BTT is a reasonable alternative to conventional CT angiography because it reduced the contrast material dose and maintained diagnostic image quality. This is particularly true in patients with aortic diseases given the older age and underlying comorbidities that often lead to compromised renal function. Although the data to support the risk of contrast-induced acute kidney injury for patients with severe kidney disease remain scant, the most recent joint statement by the American College of Radiology and the National Kidney Foundation recommends prophylaxis with intravenous normal saline for patients with an estimated glomerular filtration rate less than 30 mL/min/1.73 m² or high-risk individuals with an estimated glomerular filtration rate of 30–44 mL/min/1.73 m² (7). Therefore, lowering the dose of contrast material in these groups of patients is certainly a piece of welcome news. In addition to potentially limiting risks to patients who undergo several CT angiographies in their lifetime, a lower dose of iodinated contrast material would decrease medical costs and improve the overall value of angiography.

Although Noda et al address some of the technical and clinical advantages of DECTA with an optimized threshold for BTT, many questions remain unanswered. Where is the balance between lowering the contrast material dose and obtaining a diagnostic CT scan for detection of extravascular findings? What is the diagnostic performance of higher-kilovoltage-peak VM images that are obtained with a lower dose of iodinated contrast material? Given the fact that the studied population had a normal body mass index, does the performance of DECTA with an optimized threshold for BTT remain similar in patients who are overweight?

There are several recent revolutionary developments in spectral CT. Two of the most promising advances that relate to the current work are the development of the new photon-counting detectors with better dose efficiency, special resolution, and energy-discriminating capabilities compared with energy integrated detectors (8). The second major development is the ever-increasing role of artificial intelligence in denoising, lesion detection, image reconstruction, and segmentation (9). Our collective goal as radiologists at the forefront of technologic innovation is to improve on and develop imaging studies that maximize value and clinically meaningful output while simultaneously minimizing risks to our patients. The study by Noda et al demonstrating improved visualization of the aorta and its branches with the use of less iodinated contrast material in dual-energy CT is a promising step in that direction.