Additional diagnostic value of new CT imaging techniques for the functional assessment of coronary artery disease: a meta-analysis.

Abstract

To determine the diagnostic performance of cardiac computed tomography (CT)-based modalities including coronary CT angiography (CTA), stress myocardial CT perfusion (stress CTP), computer simulation of fractional flow reserve by CT (FFR_CT), and transluminal attenuation gradients (TAG), for the diagnosis of hemodynamic significant coronary artery disease (CAD), using invasive fractional flow reserve as the reference standard. PubMed and Cochrane databases were searched for original articles until July 2018. Diagnostic accuracy results were pooled at per-patient and per-vessel level using random effect models. Fifty articles were included in the meta-analysis (3024 subjects). The per-patient analysis per imaging modality demonstrated a pooled positive likelihood ratio (PLR) of 1.78 (95% confidence interval CI 1.49-2.11), 4.58 (95% CI 3.54-5.91), and 3.45 (95% CI 2.38-5.00) for CTA, stress CTP, and FFR_CT respectively. Per-patient specificity of stress CTP (82%, 95% CI 76-86) and FFR_CT (72%, 95% CI 68-76) were higher than for CTA (48%, 95% CI 44-51). At the vessel level, PLR was 2.42 (95% CI 1.93-3.02), 7.72 (95% CI 5.50-10.83), 3.50 (95% CI 2.73-4.78), 1.97 (95% CI 1.32-2.93) for CTA, stress CTP, FFR_CT, and TAG respectively. With improved PLR and specificity, stress CTP and FFR_CT have incremental value over CTA for the detection of functionally significant CAD. • New functional CT imaging techniques, such as stress CTP and FFR_CT, improve diagnostic accuracy of coronary CTA to predict hemodynamically relevant stenosis. • TAG yields poor diagnostic performance. • Combination of CTA and some functional CT techniques (stress CTP and FFR_CT) might become a "must" to improve diagnostic accuracy of CAD and to reduce unnecessary invasive coronary angiography.