Diagnostic accuracy of myocardial perfusion imaging and stress echocardiography for the diagnosis of left main and triple vessel coronary artery disease: a comparative meta-analysis

Objectives Compare the diagnostic performance of stress echocardiography (SE) and myocardial perfusion imaging (MPI) for the diagnosis of left main disease (LM) and triple vessel disease (TVD). Background Limited comparative data on MPI and SE for the detection of LM and TVD (high-risk coronary artery disease) exist in the literature. Methods Quantitative meta-analysis was performed using studies identified by systematic electronic literature search. Articles were included if they reported data on exercise, dobutamine SE or exercise, adenosine, dipyridamole, thallium201, technetium 99m sestamibi MPI and used coronary angiography as the reference test. Summary receiver-operating characteristic (SROC) curves were constructed for each imaging modality. Additionally, pooled sensitivity, specificity and likelihood ratios were calculated per modality. Meta-regression was performed to adjust for patient and study characteristics. Results Thirty-two studies met inclusion criteria; 23 (MPI-15; SE-14; Common studies-6) provided sufficient data for analysis. In a SROC model comparing the two imaging modalities, SE was associated with higher area under curve (0.82 (0.03) vs 0.73 (0.02), p = 0.01) and index Q* value (0.75 (0.02) vs 0.67 (0.02), p = 0.01). Using pooled summary point estimates, SE had higher pooled sensitivity (94% vs 75%, p<0.001) and lower negative likelihood ratio (0.21 vs 0.47, p<0.001) compared to MPI. No evidence of a difference in the pooled specificity (40% vs 48%, p = 0.16) and positive likelihood ratio (1.52 vs 1.58, p = 0.36) was seen between the two stress modalities. Pooled diagnostic OR on meta-regression (9.78 vs 4.06, p = 0.02) remained significantly higher for SE compared to MPI for identification of LM and TVD even after adjustment for study characteristics. Conclusions Since LM alone or in combination with TVD are categorised as representing potentially life-threatening variants of CAD, a screening test with high sensitivity, low negative likelihood ratio or higher discriminatory capacity would be desirable for risk stratification. In the absence of a direct head-to-head comparison of the diagnostic accuracies of SE and MPI, our findings indicate that SE appears to be the preferred screening modality for high-risk coronary artery disease.