Evaluation of Web-based Training to Educate Physicians in the Methods of Interpreting Florbetapir-PET Scans

Avid has previously developed a florbetapir-PET scan binary read methodology and training program, which was successfully applied in studies 18F-AV-45-A08(NCT01565369), 18F-AV-45-A09(NCT01565382) and 18F-AV-45-A16(NCT01447719). In these previous studies, training was conducted in-person. Study 18F-AV-45-PT01 will evaluate an automated version of this training program suitable for web-based distribution. Nuclear medicine physicians will complete the automated training program then read the florbetapir-PET images from 151 subjects with repeat readings of 33 randomly selected images. Readers will rate each case as either positive for significant tracer accumulation in cortical gray matter or negative for significant tracer accumulation in cortical gray matter. The inter-rater reliability, sensitivity and specificity of the readers will be evaluated.

The primary image set for determination of inter-reader agreement is comprised of images from 119 subjects from two previous clinical studies (A07[NCT00857415]/A16[NCT01447719] and A05[NCT00702143]). This included all 59 subjects who came to autopsy in Study A07(NCT00857415) and it's follow-up study, A16(NCT01447719), and a set of 60 cases randomly selected from Study A05(NCT00702143) subjects, including 20 cognitively-normal controls, 20 with mild cognitive impairment) MCI, 20 with AD.

Subsequently, the protocol was amended to include all remaining A05(NCT00702143) MCI patients that were not used in the training program (n=32); 13 of these 32 scans presented for repeat reading (to increase the number of A05 MCI cases with repeated reads to 20). This increased the number of unique cases in the validation dataset to 151 scans with 33 scans repeated yielding a total of 184 scans reviewed by the readers.