A total of thirty-nine studies (982 patients) were included: Twelve studies (332 patients) evaluated F-18 FDG PET; seven studies (209 patients) evaluated T1-201 stress-redistribution-reinjection; eight studies (145 patients) evaluated T1-201 rest-redistibution; ten studies (207 patients) evaluated Tc-99m MIBI; Sixteen studies (448 patients) evaluated LDDE.
The pooled weighted mean values of sensitivity and specificity for the different imaging techniques were as follows:
for Tc-99m MIBI, the sensitivity was 83% (95% CI: 78, 87; 99% CI: 77, 89) and the specificity was 69% (95% CI: 63, 74; 99% CI: 61, 76);
for LDDE, the sensitivity was 84% (95% CI: 82, 86; 99% CI: 81, 87) and the specificity was 81% (95% CI: 79, 84; 99% CI: 79, 84);
for Tl-201 reinjection, the sensitivity was 86% (95% CI: 83, 89; 99% CI: 82, 90) and the specificity was 47% (95% CI: 43, 51; 99% CI: 42, 52);
for F-18 FDG-PET, the sensitivity was 88% (95% CI: 84, 91; 99% CI: 83, 92) and the specificity was 73% (95% CI: 69, 77; 99% CI: 69, 77);
for Tl-201 rest-redistribution, the sensitivity was 90% (95% CI: 87, 93; 99% CI: 86, 94) and the specificity was 54% (95% CI: 49, 60; 99% CI: 48, 61).
The sensitivity for predicting regional functional recovery after revascularisation was high for all techniques. The specificity of both Tl-201 protocols was significantly lower (p<0.05) than that of the other techniques, whilst the specificity of LDDE was significantly higher (p<0.01).