Whole-heart coronary magnetic resonance angiography: value for the detection of coronary artery stenoses in comparison to multislice computed tomography angiography

Abstract

Background: Coronary magnetic resonance imaging and computed tomography are being discussed as alternatives to catheter angiography in the detection of coronary artery disease. Yet, only few comparative validations have been performed. Purpose: To compare steady-state free precession whole heart coronary magnetic resonance imaging (MRI) with multidetector coronary computed tomography angiography (CTA) for the detection of coronary artery disease using catheter angiography as the standard of reference. Material and Methods: Twenty patients with known CAD were examined with navigator (NAV) gated and corrected free-breathing 3D balanced gradient echo whole heart coronary MRI and coronary CTA. Subjective overall image quality (4 point scale, 1 = excellent), visibility of vessel segments and accuracy for the detection of significant coronary stenoses (>50%) were compared to coronary x-ray angiography by two blinded readers. Results: Median of subjective image quality was 3 for coronary MRI and 2 for coronary CTA. Of a total of 209 segments, 67 segments (32%) had to be excluded from the evaluation by coronary MRI (61 due to insufficient image quality and 6 due to stent artifacts). For coronary CTA, 31 segments (15%) had to be excluded from the evaluation (12 due to insufficient image quality, 15 due to severe calcifications superimposing the vessel lumen and 4 due to stent artifacts. Segment based values for the detection of ⩾50% diameter coronary x-ray angiographic stenoses were: specificity: MRI 88%, CTA 95%; sensitivity: MRI 82%, CTA 84%; diagnostic accuracy: MRI 87%, CTA 93%; positive predictive value: MRI 68%, CTA 77% and negative predictive value: MRI 94%, CTA 95%. Conclusion: Coronary WH-MRI was inferior to coronary CTA regarding image quality and number of evaluable segments but both had similar diagnostic value for the detection and exclusion of CAD when only evaluable segments were included.