Abstract

A lack of specificity for myocardial perfusion imaging has been widely reported, mostly related to false-positive defects on the inferior wall. The application of depth-dependent resolution recovery (RR), attenuation correction (AC) using external source devices, and scatter correction has been proposed to resolve this pitfall.

Methods: We studied the clinical benefit of depth-dependent RR, nonuniform AC using a scanning line source, and scatter correction (photon energy recovery [PER]) compared with filtered backprojection alone. Eighty-two patients were included: 40 healthy volunteers with a low likelihood of coronary artery disease (control group) and 42 patients with proven right or circumflex coronary artery disease but without involvement of the left anterior descending artery. Among these 82 patients, the images of 33 were also processed with PER.

Results: RR did not alter the performance of filtered backprojection alone. AC + RR greatly improved specificity and the rate of normal (201)Tl SPECT findings in the control population (from 56% to 95% and from 53% to 100%, respectively) but significantly decreased sensitivity (from 92% to 54%). AC + RR generated a false anteroapical defect in 21% of patients and reverse redistribution of the apex in 23%. AC + RR significantly decreased the extent of the stress defect (from 4.09 to 3.21 segments, P < 0.003) and increased the perfusion score of the stress defect (from 0.78 +/- 0.72 to 1.47 +/- 1.11, P < 0.00061). Moreover, AC + RR generated overcorrection on the inferior wall, leading to false estimation of viability for 11 of 15 patients with an old inferior myocardial scar without evidence of residual viability. PER decreased overcorrection on the inferior wall, but without improving sensitivity. PER did not significantly reduce the number of anteroapical false-positives or the number of apical reverse distribution cases.

Conclusion: AC + RR improved the specificity and normalcy rate of (201)Tl SPECT myocardial perfusion imaging but generated overcorrection on the inferior wall, leading to low sensitivity and to false evaluation of myocardial viability in 73% of the patients with inferior infarction. AC + RR also generated anteroapical artifacts. The addition of scatter correction did not significantly reduce these drawbacks.