In Reply: Drs Carome and Wolfe question the interreader reliability in the interpretation of the florbetapir-PET images in our study. Florbetapir-PET imaging was performed on 29 participants who subsequently died and had autopsies. The prespecified, FDA-endorsed, primary end point was the median of 3 readers' visual ratings on a 5-point scale. The median value was used to focus on the ability of the images to reflect β-amyloid pathology rather than on individual skills in image interpretation. The correlation between the median rating and histopathology was good, with Spearman ρ of 0.71 and 0.78 (Bonferroni corrected P < .001), depending on pathologic criteria.
Clark CM, Mintun MA, Pontecorvo MJ. Florbetapir-PET Imaging and Postmortem β-Amyloid Pathology—Reply. JAMA. 2011;305(18):1857-1858. doi:10.1001/jama.2011.580