Reproducibility of semi-automatic coronary plaque quantification in coronary CT angiography with sub-mSv radiation dose

Abstract

IntroductionCoronary computed tomographic angiography (CTA) can characterize coronary atherosclerotic plaque components as calcified and non-calcified. Quantitative measurements of coronary plaque burden by coronary CTA may play a role in serial studies to determine disease progression or response to medical therapies. The reproducibility from repeated assessment of such quantitative measurements from low-radiation dose coronary CTA has not been previously assessed. PurposeTo evaluate the interscan, interobserver and intraobserver reproducibility for coronary plaque volume assessment using semi-automatic plaque analyses algorithm in low radiation dose coronary CTA. MethodsIn 50 consecutive patients undergoing two 128-slice dual source CT scans within 12 days with a mean radiation dose of 0.7 mSv per coronary CTA, the interscan, interobserver and intraobserver reproducibility of coronary plaque assessment using validated software (AutoPlaq) were evaluated. ResultsInterscan, interobserver and intraobserver agreement for non-calcified and calcified plaque volumes were excellent (Spearman rho 0.87–0.99). Interscan mean percentage difference in non-calcified and calcified plaque volumes were 0.1% (p = 0.8) and 1.9% (p = 0.19) with limits of agreement of ±11% and ±48.5%; per inter- and intraobserver mean percentage differences were 0.1% (p = 0.25) and 0.3% (p = 0.001), and 0.3% (p = 0.33) and 0.4% (p = 0.59) with limits of agreement of ±7% and ±32.9%, and ±6.6% and ±32.1%, respectively. ConclusionA semi-automatic plaque assessment algorithm in repeated low radiation dose coronary CTA allows for high reproducibility of coronary plaque characterization and quantification measures.