In phantom, MDCT edges EBT for calcium detection

Oct 10, 2005

Both 16-slice MDCT and electron beam tomography (EBT) performed well in a phantom face-off, but MDCT had a significant advantage in detecting small amounts of coronary artery calcium, according to researchers from Hiroshima University Hospital in Japan.

The researchers concluded that MDCT with retrospective reconstruction yields low variability in coronary artery calcium (CAC) measurements, while cautioning that in vivo studies will be needed to confirm the results.

"Monitoring CAC is suggested to assess the progression and regression of coronary atherosclerosis, thereby documenting risk factors and the need for medical intervention," the authors wrote in the October American Journal of Roentgenology. "For this purpose, low interscan variability of CAC measurements is mandatory. The normal progression of the CAC score per year is reported to be 14-27% (average, 24%) and can climb to 33-48% in patients with significant coronary disease. However, in previous studies, the variability using electron beam CT yields 20-37%, which jeopardizes the detection of any changes in this range" (AJR, October 2005, Vol. 185:4, pp. 995-1000).

Could thin-section imaging on a 16-slice scanner reduce the interscan variability that plagues patient management, thus improving the predictive value of calcium assessment? To find out, the team built a moving heart phantom and scanned it repeatedly in multiple MDCT and EBT protocols.

The phantom consisted of a controller-driven balloon filled with a mixture of water and contrast media (and submerged in corn oil) representing the coronary arteries.

"A controller with an ECG-synchronizer drove the balloon," Dr. Jun Horiguchi, Dr. Yun Shen, Dr. Yuji Akiyama, and colleagues wrote of their phantom. "The motion was achieved by setting four driver sequences -- that is, two speeds of fast emptying for the systolic phase and fast and slow filling for the diastolic phase." The calcium models used three materials -- silicon (305 HU), putty (501 HU), and polytetrafluorethylene (929 HU) -- that were packed inside rubber tubes and attached to the balloon surface.

The study was conducted with three sequential CT scans that were repeated on an electron beam CT scanner (C-150 XL, Imatron, South San Francisco, CA) and a 16-row MDCT scanner (LightSpeed Ultrafast 16, GE Healthcare, Chalfont St. Giles, U.K.).

The EBT protocol used 100-msec acquisition time, 35-40 gapless 3-mm slices, 130 kVp and 625 mAs, and ECG triggering at 80% of the R-R interval. Images were reconstructed using a 512 x 512-pixel matrix and a sharp reconstruction kernel.
The retrospective ECG-gated MDCT protocol used volumetric phantom data that were obtained at 16 x 0.625-mm collimation, 0.5-sec rotation speed, pitch of 0.275, 120 kVp, and 100 mAs at 80% of the R-R interval. Multisector reconstruction was used to improve temporal resolution by combining two to four cardiac cycles, and reconstructions were created at 0.625 mm with a 0.625-mm increment, 1.25 mm/1.25 mm, 2.5 mm/1.25 mm, and 2.5 mm/2.5 mm.
Prospective ECG-triggered axial CT at 80% of the R-R interval was acquired at slice thicknesses of 0.625 mm, 1.25 mm, and 2.5 mm at 0.5-sec rotation speed, 120 kVp, and 100 mAs, according to the authors.

An Advantage Windows workstation version 4.1 (GE Healthcare) was used to measure calcium scores according to the Agatston method (threshold 130 HU, 2-pixel area of at least 0.52 mm²), according to the authors. The study compared the detection of small amounts of calcium, Agatston score and variability, and heart rate sequences, and assessed whether the inclusion of mass or volume calcium measurement methods reduced the variability.

According to the results, all 1-mm calcium models were detected on 16-slice MDCT at 0.625-mm and 1.25-mm slice thicknesses, with both helical and axial scanning. Some of these were missed on EBT, which detected 59% (16/27). Detection of the 1-mm models in thicker collimation spiral CT acquisitions were 96% (26/27), 93% (25/27), and 89% (24/27) on 2.5 mm/1.25 mm MDCT, 2.5 mm/2.5 mm MDCT, and 2.5 mm axial MDCT, respectively.

In Agatston scoring, EBT showed higher variability than MDCT, and the authors found statistically significant differences between EBT and 0.625-mm MDCT (p = 0.03) and between EBT and 1.25-mm MDCT (p = 0.4). The Agatston scores were comparable among all CT protocols, however, showing higher values on 0.625-mm axial protocols.

The use of thin-slice imaging (0.625 mm and 1.25 mm) reduced variability on both helical and axial protocols, while 0.625-mm scans showed less variability than 1.25-mm scans on spiral scanning alone (p = 0.4), the authors wrote. Overlapping reconstruction also reduced variability (p = 0.04:2.5 mm/1.25mm spiral versus 2.5 mm/2.5 mm spiral). Spiral CT was less variable than axial CT, with statistically significant differences seen in 0.625-mm (p < 0.01) and 1.25-mm (p < 0.01) scans.

Heart rate differences did not produce statistically significant variations in Agatston scores. Using volume or mass measurements was less effective in reducing score variability in 0.625-mm and 1.25-mm scans, and there was almost no variability between EBT and MDCT protocols, the authors wrote.

"The detection of small amounts of calcium is considered important because the presence or absence of calcium is suggested as one clear-cut point in a clinical setting," Horiguchi and colleagues wrote. "This is because the presence of CAC predicts an increase in the risk of new cardiac events in asymptomatic adults without a clinical history of cardiovascular disease. "

In an EBT study, Vliegenthart and colleagues showed that almost half of the small calcifications detected on 1.5-mm-slice scans were missed on 3.0-mm-slice scans, the Japanese team noted. "In our current study also," they wrote, "detection of small calcium on 3.0-mm electron beam CT was poor. Thin-slice or overlapping helical CT is considered to have a definite advantage in this respect" (Radiology, November 2003, Vol. 229:2, pp. 520-525).

"Thin-slice or overlapping helical CT, showing high reproducibility using Agatston, volume, or mass scoring, is considered advantageous regardless of which scoring algorithm becomes mainstream in the future," they wrote.

The finding that heart rate had little effect on CAC score variability suggests that although retrospectively gated spiral CT has inferior temporal resolution compared to EBT, it can measure calcium in a variety of heart rates, the team wrote.

As for study limitations, tube current was not assessed, although reduced radiation doses are important in the screening environment, and noise from reduced tube currents is known to increase variability in calcium scores. Moreover, the use of smooth, homogeneous materials such as putty for the calcium models did not necessarily reflect actual plaques optimally, the authors wrote.

"(Sixteen-detector-row) MDCT with retrospective reconstruction yields low variability in CAC measurements and has the potential to be a useful tool in monitoring the progression of coronary atherosclerosis," Horiguchi and colleagues concluded.

By Eric Barnes
AuntMinnie.com staff writer
October 11, 2005

Low-dose CT calcium scores near-equivalent of higher dose, September 20, 2005

CT coronary calcium results vary by scanner, body type, May 31, 2005

Most MDCT scanners fall short in cardiac imaging, September 23, 2004

Mass-based CT calcium scoring leaves room for dose reduction, April 30, 2004