Thirteen studies, seven of FDG-PET (270 patients, range 17 to 85) and six of PIB-PET (222 patients, range 21-65), were included in the review. No study assessed both FDG- and PIB-PET. The median methodological quality score was 12.7 (range 11.5 to 13.5); potential test and diagnostic review bias was the most commonly reported methodological quality issue in included studies.
FDG-PET
The sensitivity estimates from included studies ranged from 37.5% to 100% and specificity from 47.4% to 100%. The pooled estimate of sensitivity was 78.7% (95% CI, 68.7 to 86.6; Ι²=53.4%) and the pooled estimate of specificity was 74.0% (95% CI, 67.0 to 80.3%; Ι²=84.7%). Ι² values indicated moderate heterogeneity in sensitivity estimates and high heterogeneity for specificity. There was no evidence of a threshold effect. With the exception of follow-up time, meta-regression analyses found no significant variables. There was evidence of publication bias.
PIB-PET
The sensitivity estimates from included studies ranged from 83.3% to 100% and specificity from 41.7% to 76.5%. The pooled estimate of sensitivity was 93.5% (95% CI, 71.3 to 99.9%; Ι²=0) and the pooled estimate of specificity was 56.2% (95% CI, 47.2 to 64.8%; Ι²=28.4%). Ι² values estimated low heterogeneity in both sensitivity and specificity estimates. There was no evidence of a threshold effect. There was no evidence of publication bias.
Pooled estimates of positive and negative likelihood ratios and diagnostic odds ratio were also reported.