Background: Myocardial perfusion imaging has demonstrated a limited sensitivity as a means of accurately identifying left main (LM) coronary disease. Because regional quantitative perfusion biases are eliminated with attenuation corrected (AC) single photon emission computed tomography (SPECT), as compared with uncorrected (NC) SPECT, we hypothesized that AC SPECT would demonstrate increased diagnostic accuracy for the detection of significant LM coronary stenosis.
Methods and results: We studied 28 patients (23 men, 5 women; mean age, 66+/-9 years) with significant LM stenoses (> or =50%) and 34 control patients (27 men, 7 women; mean age, 65+/-11 years) with 2-vessel coronary disease. Rest thallium-201 and stress technetium 99m sestamibi SPECT imaging with and without AC were performed, as described earlier. Both AC and NC images were analyzed visually and quantitatively in comparison with corresponding normal databases. A greater sensitivity for detection of an LM defect pattern (64% vs. 7%, P = .0009) with equivalent specificity (94% vs. 100%, P = not significant) was demonstrated by means of visual analysis of AC SPECT images. More disease was demonstrated in a greater number of territories with AC SPECT images than with NC images (2.14+/-0.97 for AC images vs. 1.43+/-0.84 for NC images, P = .0001). Similar improvement in the detection of LM disease was shown by means of automated quantitative analysis (57% for AC SPECT vs 14% for NC SPECT, P = .0005), again with no loss in specificity.
Conclusions: AC SPECT with the University of Michigan method in consecutive patients with LM stenoses and a select control population with severity matched multivessel coronary disease significantly improved the diagnostic accuracy of myocardial perfusion imaging for the identification of LM coronary disease, compared with uncorrected SPECT.