Step-and-shoot mode minimizes dose in dual-source cardiac CT

The use of a step-and-shoot mode with dual-source CT (DSCT) produced radiation doses under 2 mSv for coronary CT angiography studies of patients with normal heart rates, according to a study in the October issue of Radiology.

Researchers from University Hospital Zurich in Switzerland and Siemens Healthcare of Erlangen, Germany, produced diagnostically robust results in a cohort of patients with heart rates of 70 bpm or less. The studies were performed in prospectively gated step-and-shoot (SAS) mode, a low-dose cardiac imaging technique that limits x-ray beam exposure to a small portion of the cardiac cycle, when data acquisition is considered relevant.

With the SAS mode, radiation exposure to the patient is reduced by 80% or more compared to traditional retrospective gating techniques, rendering coronary CT angiography potentially safer and easier to justify for patients with suspected coronary artery disease. The authors, led by principal investigator Dr. Paul Stolzmann, believe their paper is the first to demonstrate the feasibility, image quality, and radiation exposure of dual-source CT coronary angiography performed in SAS mode in a patient population (Radiology, October 2008, Vol. 249:1, pp. 71-80).

The dual-source CTA studies were performed in 90 patients (55 men, 35 women; mean age 63 years ± 10), divided into several groups. The first part of the study, involving 40 patients, was aimed at assessing image noise and vessel attenuation in SAS mode on the DSCT scanner (Somatom Definition, Siemens Healthcare).

Normal-weight patients with a BMI less than 25 (n = 22) were assigned to a scanning protocol with a tube voltage of 100 kV (protocol A); heavier patients (BMI 25-30, n = 18) were assigned a protocol with a tube voltage of 120 kV (protocol B). Both protocols used attenuation-based tube current and 1 mL/kg contrast.

Part two of the study aimed to test the effects of contrast dose reduction, scanning an additional 50 consecutive patients with either a protocol of 120 kV and 1 mL of iodixanol (Visipaque 320, GE Healthcare, Chalfont St. Giles, U.K.) iodinated contrast per kilogram of body weight (n = 21) (protocol C), or 100 kV with a contrast dose reduced by 20% (0.8 mL/kg, n = 29) (protocol D). The kV settings were based on the BMI measurements used in the preliminary portion of the study.

Based on the CTA findings, nine of the 50 patients (10%) underwent conventional coronary angiography, the authors noted.

The patients were referred for atypical chest pain with a low or intermediate suspicion of coronary artery disease. Patients were excluded for high serum creatinine levels (n = 9), hypersensitivity to iodinated contrast materials (n = 3), coronary stents (n = 19), or bypass grafts (n = 18).

Results in the first part of the study showed similar image noise for both protocols, evidenced by a significant positive correlation between image noise and BMI for protocols A (r = 0.63, p < 0.01) and B (r = 0.87, p < 0.001). However, mean attenuation in the coronary arteries for the slimmer patients in protocol A (444 HU) was significantly higher than that of the higher BMI patients in protocol B (358 HU, p < 0.001).

The 20% reduction in contrast dose (combined with slightly higher kV) in protocol C produced similar attenuation to the 1 mL/kg dose in protocol B.

A total of 176 (12.2%) of 1,440 coronary artery segments were smaller than 1 mm or missing because of anatomic variants, leaving 1,264 coronary artery segments for image quality assessment. With excellent interobserver agreement (k = 0.84), subjectively evaluated image quality was considered diagnostic for 1,237 (97.9%) of the 1,264 coronary artery segments in 93% of the 90 patients scanned. Image quality was deemed nondiagnostic in 27 segments (2.1%) in six (7%) patients.

Image quality was impaired in 104 (8.2%) of the segments. Of these, impaired quality was due to image noise in three segments (2.9%), stair-step artifacts in 41 segments (39.4%), and motion artifacts in 60 (57.7%), the authors noted.

Motion artifacts occurred most often in the right coronary artery, which has the fastest motion and often requires reconstructions in late systole, they noted. Patients with high heart rate variability more often produced stair-step artifacts.

"Mean heart rate had a significant effect on stair-step artifacts (area under the curve [AUC] = 0.79l, 95% confidence interval [CI]: 0.723, 0.892; p < 0.001), whereas heart rate variability had a significant effect on stair-step artifacts (AUC = 0.79l, 95% CI: 0.687, 0.865; p < 0.001)," Stolzmann and colleagues wrote.

With use of the exclusion criteria, 45% (73 of 163) patients had to be excluded from SAS-mode images, including 29% (25 of 85) of the patients with a high heart rate.

"Theoretically, this number could have been lowered by administering beta-blockers before CT," they wrote.

The mean estimated effective dose, averaged between male and female models in Monte Carlo simulations, was 1.2 mSv ± 1.2 for protocols A and C, and 2.6 mSv ± 0.5 for protocol B.

"Our results indicate that (low-dose prospectively gated DSCT angiography) is feasible in patients with a heart rate lower than 70 beats per minutes and a BMI lower than 30 kg/m2, and that it depicted 97.9% of the coronary segments with diagnostic image quality," the team wrote.

On the other hand, retrospectively gated studies rely on faster gantry speeds for improved temporal resolution but slower pitch to avoid discontinuities in anatomic coverage when images are reconstructed from consecutive cardiac cycles. Imaging the entire cardiac cycle generally produces a higher effective radiation dose.

Hausleiter et al reported mean effective doses of 9.4 mSv for 64-slice CTA using tube current reduction and 14.8 mSv without it (Circulation, March 14, 2006, Vol. 113:10, pp. 1305-1310). Sigal-Cinqualbre and colleagues used another method, tube voltage reduction, to lower the mean dose; however, increased image noise with this method can impair image quality (Radiology, April 2004, Vol. 231:1, pp. 169-174).

Though widely used for coronary calcium measurement, SAS mode "requires the use of strict exclusion criteria," the authors wrote. "At lower heart rates, this examination generally is sufficient for reconstructions during only one phase in mid-diastole for the acquisition of diagnostic-quality images of the entire coronary artery tree. At higher heart rates, however, reconstructions often must be performed at several phases of the RR interval to obtain diagnostic-quality images of different segments," precluding use of the SAS technique.

A limitation of the study was that higher heart rates and significant variability in heart rates were not evaluated for SAS imaging feasibility, they noted. "Second, semiquantitative image quality grading may have been influenced by subjectivity bias," they wrote. Finally, DSCT's sensitivity for stenosis detection in SAS mode was not evaluated.

"Use of the SAS mode combined with body weight-adapted tube voltage and current resulted in mean estimated effective doses of 1.2 mSv in patients of normal weight, and 2.6 mSv in overweight patients at cardiac CT," the group wrote. And a 20% reduction in contrast dose did not compromise image quality. "Further studies must be aimed at assessment of the diagnostic performance of SAS-mode coronary CT angiography for the diagnosis and exclusion of CAD."

By Eric Barnes staff writer
September 29, 2008

Related Reading

Dual-source CT edges into cardiac SPECT turf, March 6, 2008

DSCT matches angiography for stenosis detection -- even in fast hearts, December 4, 2007

Michigan heart center aims DSCT guns at image quality, October 8, 2007

CTA predicts functional recovery of the myocardium, September 7, 2007

Dual-source coronary CTA images the calcium-burdened, April 13, 2007

Copyright © 2008

Page 1 of 1247
Next Page