Question Posted 03/30/17:
For part of our project, we are trying to validate measurements with the UCBERKELEYAV45 analysis uploaded on the shared data website. From using the provided documentation, it appears that ALL of the posted PET values are SUVR values with a composite reference region (whole cerebellum, brainstem/pons, and eroded WM regions) weighted by each region\'s contribution to the volume. Can you please confirm?

Also, what AV45 images are actually used? Is the 4-image dynamic series compiled into 1 static Mean IP (time-weighted) image? I could not find this information in the documentation.

Third, while performing my post-processing analysis with MIM in attempt to measure the SUVR values in the UCBERKELEY regions, I had issues retrieving SUV values for several of the cases.

I just got off the phone with MIM, and they mentioned that the \"calibration factor\" is missing in the dicom header for these cases, which prevents the software from calculating the SUVR. According to them this information refers to the calibration factor on Philips scanners. Can you please advice how I can get access to the information so I can either manually enter this dicom information or get the non \"raw\" images that have this data.
Response posted 03/30/17 by Susan Landau:
Hello, the answers to the first two questions can be found in our methods document ("UC Berkeley - AV45 Analysis Methods (PDF)" located on the Study Data page) -- briefly, the posted UC Berkeley AV45 data are not normalized by the composite reference region. There are several reference regions (including the whole cerebellum and the composite reference region) that can be used to calculate SUVRs. We use the fully pre-processed images as described here ( and in the UC Berkeley methods document .

For your third question you may want to contact Bob Koeppe at to discuss how dicom header information is handled during pre-processing of the ADNI PET images.