Doing More With Less: Researchers Discuss Techniques for Reducing CT Contrast

Reed Miller

December 07, 2009

December 7, 2009 (Chicago, Illinois) — New research shows that a patient's exposure to iodine contrast medium during coronary CT angiography (CCTA) can be dramatically reduced with some careful planning and attention to detail.

Reducing the volume of contrast medium is critical for patients with renal dysfunction susceptible to contrast-induced nephropathy. "We need to seriously think about options and [how to] do CCTA in these patients with hardly any contrast," Dr Andre Duerinckx (Howard University, Washington, DC) said at the Radiological Society of North America (RSNA) 2009 Scientific Assembly. Duerinckx was a keynote speaker at a session featuring seven research abstracts and expert commentary on minimizing contrast volume in CCTA while maintaining its diagnostic utility.

"We in medicine are constantly making compromises to get away from the most perfect beautiful images to less beautiful images and yet still provide a valuable clinical service," he explained. "Beautiful images are not what we should be aiming for. We should be aiming for helping the patients. So the question is, how much radiation, how much contrast do we really need to help the patient?"

To illustrate the importance of precise technique, Duerinckx provided examples of CT scans performed with less than 50 mL of contrast medium that provided diagnostically valuable images and examples of scans with over 100 mL of contrast that produced "total garbage." The standard contrast dose is about 70 mL.

He added that contrast-injection techniques must be refined to keep up with the new faster scanners. "We're all going to eventually going to be using 320-slice scanners or faster, so we're all going to have to deal with the fact that our attention to detail and injection technique is going to become much more important," he pointed out. "When you do a 12-second injection, almost nothing matters, but with new technology, the details of the execution become much more important."

"We also must understand the relationship of the arterial enhancement to our ability to interpret the study effectively," Dr Geoff Rubin (Stanford University, CA) said. "More specifically, what is the target contrast-to-noise ratio that is needed? That is going to ultimately determine what our target is for the most efficient use of iodine."

Rubin explained that the best techniques for low-contrast CCTA are based on the operator's understanding of the relationship between contrast flow rate and image enhancement. Because the image enhancement increases as the contrast accumulates, the scan should be triggered at the ideal moment for best enhancement of the target coronary vessels as the bolus of contrast medium flows through. Also, the contrast injection should be "tailored" to the patient's specific physiology, he said.

More Contrast Doesn't Always Mean Better Pictures

Dr Alexander Lembcke (Charité Universitätsmedizin, Berlin, Germany) presented results of a 60-patient study that randomized patients undergoing CCTA to either a dedicated low-volume (35 mL injected at 7.5 mL/s) or a standard-volume (70 mL injected at 5.0 mL/s) contrast-media injection protocol [1]. All of the patients were scanned with a 320-detector row system with a standardized, single-beat, nonspiral "snapshotlike" technique with prospective ECG-triggering, meaning the scanner delivers radiation only during specific parts of the cardiac cycle. The scans were also triggered by an automated bolus tracker, so the scan was instantly performed when the contrast reached a density of 300 HU (Hounsfield units) in the left ventricle.

The mean enhancement values were greater for the standard-volume group in the main pulmonary artery, left ventricle, and left atrium, but there was no difference between the groups in enhancement of the aorta or the coronaries.

"The combination of the snapshotlike data acquisition and a dedicated contrast-media injection protocol offers the possibility for a substantial reduction of contrast-media volume without significant loss of vascular enhancement in the coronary arteries," Lembcke concluded. "Optimized injection protocols with a low volume of contrast media might be routinely used for coronary CT angiography, especially when scanning patients with reduced renal function."

Dr Patrick Hein (Charité Universitätsmedizin) presented results from 40 consecutive patients scanned with prospectively ECG-triggered, single-beat, nonspiral CT imaging with randomly assigned volumes of contrast material--40 mL, 50 mL, 60 mL, or 70 mL [2]. The image enhancement of the aortic root was slightly lower in the 40-mL group than the 50-mL group, but the differences between the other groups did not reach statistical significance and the qualitative analysis of the images resulted in comparable scorings for all four groups, with diagnostic image quality achieved in all 40 patients.

Map Out the Patient First

A 25-patient study led by Dr Uma Numburi (Cleveland State University, Cleveland, OH) evaluated a protocol relying on injection of a "test bolus" of contrast before the scan to map out how contrast will flow in individual patients [3].

Each patient was injected with 20 mL of contrast at 4.5 mL/s followed by saline. Then Numburi and colleagues measured the resulting test enhancement curve in the ascending aorta and applied a transfer-function model to predict the contrast-injection profiles required for each patient to achieve and maintain 300 HU in the heart for two, four, or six seconds.

Multiple regression analysis correlating patient-specific injection volumes and flow rates with patient and test enhancement curve characteristics were used to design "biphasic injections" requiring reduced total contrast volumes and flow rates. The researchers concluded that biphasic injections requiring average total contrast volumes of 49, 55, and 64 mL for two-, four-, and six-second coronary scans, respectively, were “appropriate” for the patients in the trial. Future patient-specific protocols can be determined using the empirical equations.

Rubin commented that "improved contrast medium efficiency should result from tailoring the injection to the patient's size and physiology, and in particular the predictive modeling of contrast dynamics using a test bolus is very promising and an intriguing solution; we should look forward to seeing more [research] from that."

Another approach to the test-bolus approach was presented by Dr Dominik Fleischmann (Stanford University, CA). He presented data from a 16-patient trial of a complex mathematical model that predicts a patient’s coronary CT enhancement from a test bolus injection and then computes the injection volume and rate necessary to achieve the desired enhancement [4].

The calculations take less than 30 seconds on the computer and predict levels of contrast enhancement in "good agreement" with measured levels. By improving the predictability and individualization of contrast-media administration, this technique may reduce excessive opacification and therefore reduce overall contrast volumes in CCTA, Fleischmann explained.

In the study, the average arterial opacification achieved in the target vasculature was 386 HU, but the range was wide--from 273 to 520 HU--and while the average difference between the predicted target enhancement and measured enhancement was 50 HU, the range was between 18 and 119 HU.

Rubin pointed out that this illustrates the wide variability of patients' physiology and the need for patient-specific contrast-administration plans instead of simple standard protocols. "Standard protocols or even potentially [patient] weight-based protocols are simply not going to be good enough to optimize and characterize the injection and be able to achieve minimized dose with minimal intra- and interindividual variability."