Optimizing Image Contrast Display Improves Quantitative Stenosis Measurement in Heavily Calcified Coronary Arterial Segments on Coronary CT Angiography: A Proof-of-Concept and Comparison to Quantitative Invasive Coronary Angiography

Abstract

Blooming artifact from calcified plaques often renders measurement of stenosis impossible on coronary computed tomographic angiography (CTA). We sought to evaluate the impact of modifying window level on reducing blooming artifact, and its impact on stenosis quantification. We analyzed 125 calcified segments from 53 patients who underwent CTA and invasive coronary angiography (ICA). Segmental stenosis on CTA was measured using three window settings: width of 1000 Hounsfield units (HU) and level of 200 HU ("default"), 1500/200 HU ("widened"), and width and level based on the mean HU of the calcified plaque and pericoronary fat ("calcium-specific"). Segmental stenosis on ICA was quantified by a blinded experienced reader. ICA found ≥50% stenosis in 30 segments. Displaying segments with widened and calcium-specific settings improved overall accuracy of detecting ≥50% stenosis (P's < 0.001) by increasing the rate of accurately quantifying <50% stenosis (P's < 0.001), and improved correlation of stenosis quantification to ICA (P's < 0.05). There was no difference in stenosis quantification accuracy between widened and calcium-specific window settings. Limits of agreement between CTA stenosis quantification and ICA narrowed with widened and calcium-specific settings. We showed for the first time that in calcified segments, widening display window width significantly improved CTA quantification of stenosis compared to ICA.