Evidence for prospective ECG-triggering coronary CT angiography in routine practice

Abstract

Over the past decade, cardiac multidetector row CT (MDCT) has evolved, and the technology is poised to potentially alter the indications for diagnostic invasive coronary angiography (ICA) [1–10]. However, ionizing radiation concerns as related to increased cancer risks, initially applied to other organ systems [11, 12], have been translated to coronary CT angiography (CTA) and compared to ICA [13–17], while the dose dependent relation between low-radiation-dose exposure from cardiac procedures and subsequent risk of cancer has been reported [18]. Several strategies including ECG-triggered current modulation [19], lower tube voltage settings [13, 17, 20], and prospective ECG-triggering [21, 22] have progressively lowered patient radiation doses. One of the most applied approaches is prospective ECGtriggering [23, 24], where the exposure is confined to a predefined fraction of the time points of the cardiac cycle, rather than over the entire cycle. Depending on the details of the acquisition, the reduction in radiation dose can be as high as 80% [21, 22]. The effective dose of coronary CTA using prospective ECG-triggering (0.9–3.3 mSv) [2, 5] is comparable to, or lower than, diagnostic ICA (typical value of 7 mSv) [25]. However, there are unresolved concerns regarding diagnostic performance related to the limited phases of acquisition, and to date these methods have been most widely applied to patients with little risk of variations in cardiac cycle length and heart rates less than 70 beats per minute during the CT acquisition. In the current issue of the International Journal of Cardiovascular Imaging, Sun and Ng [26] performed a systematic review and meta-analysis of diagnostic value for coronary artery disease (CAD) using CTA with prospective ECG-triggering in comparison to ICA, and the authors also investigated the effective radiation dose. The 19 studies selected on primary selection criteria were reduced to fourteen that underwent the meta-analysis. Seven studies used singlesource 64-slice CT, four studies were performed with dual source 64-slice CT, two studies with 320-slice CT, and one with second generation of dual-source CT (128-slice). In 14 studies included in patient level (n = 910) analysis, pooled sensitivity for detection of significant CAD was 99% (95% CI: 98%, 100%) with specificity of 91% (95% CI: 88%, 94%) and the median negative predictive value (NPV) and positive predictive value (PPV) across studies were 94% (95% CI: 91%, 96%) and 99% (95% CI: 97%, 100%), respectively. Pooled positive (?LR) and negative (-LR) likelihood ratios were not reported at any of patient, vessel, or segment levels. Because observed heterogeneities were significant between 11 studies included in the vessel level (n = 3,531) and the segments level (n = 12,056) A. Salavati F. J. Rybicki (&) Applied Imaging Science Laboratory, Department of Radiology, Brigham and Women’s Hospital, Harvard Medical School, RA024, 75 Francis Street, Boston, MA 02115, USA e-mail: frybicki@partners.org