Centiloid Project


Amyloid imaging with positron emission tomography (PET), has produced remarkably consistent qualitative findings across a large number of centers, however there has been considerable variability in the exact numbers reported as quantitative outcome measures of tracer retention. The need for standardization in amyloid PET has prompted the creation of 100-point scale termed “Centiloid” which is an average value of zero in “high certainty” amyloid negative subjects and an average of 100 in “typical” Alzheimer’s disease (AD) patients.

All data used in the Centiloid Project is available for download below.

This is the basic dataset that most users will employ when doing Level 2 or Level 3 Centiloid scaling analyses. It consists of all of the AD-100 and YC-0 PiB PET 50-70 min data and the corresponding MR data in both dicom and nifti formats.

Full Dynamic Dataset Advanced Users File Names and descriptions

This dataset is intended for users who are seeking to scale PiB PET data that is not 50-70 SUVr to the Centiloid scale. This would include dynamic PiB data (e.g., BPND) and SUVr data from a time period other than 50-70 min. This dataset may allow users to do this without acquiring any new PiB PET data. Users who seek to perform a standard Centiloid scaling analysis should use the 50-70 min dataset.

Disclaimer: This data is intended for users already experienced in the use of dynamic PET data. These users should refer to the “CentiloidPETFraming.xls” file as an aid to determining the framing of each subject dataset.

Supplementary Data Advanced Users File Names and descriptions

This is the data described in section 2.1.2 “The standard VOIs” in the Centiloid Project paper (Klunk et al. Alzheimers Dement 2015;11:1-15). This data was used to generate the average AD-CTX and average OC-CTX images - the difference between which at 1.05 SUVr units produced the standard cortical target (CTX) VOI.

Please email here for questions or additional support regarding the Centiloid Project.

Klunk, W. E., Koeppe, R. A., Price, J. C., Benzinger, T. L., Devous, M. D., Jagust, W. J., … & Mintun, M. A. (2014). The centiloid project: Standardizing quantitative amyloid plaque estimation by PET. Alzheimer’s & Dementia.

Calibration of F-18 Amyloid Tracers


A common quantitative output value for beta-amyloid (Ab) imaging that is applicable across tracers and methods, will allow better comparison of data across sites, ultimately resulting in the application of universal diagnostic and prognostic values. A method has recently been developed that generates a unit of measurement called the Centiloid (Klunk et al, Alzheimers Dement, 2015).

50-70 Minute PIB/NAV Dataset (dicom data) File Names and descriptions

The Centiloid method was applied on 18F-NAV4694 (NAV4694) and 11C-PiB (PiB) Ab images to derive the scaling factor required to express NAV4594 tracer binding in Centiloids (CL). This is the basic dataset that users will use to convert NAV PET SUVR into Centiloid scale. It consists of 50-70 min PIB and NAV scans from 45 elderly subjects and 10 young controls and the corresponding MR data in dicom.

More details can be found in the following paper:

Rowe CC, Jones G, Dore V, Pejoska S, Margison L, Mulligan RS, Chan JG, Young K, Villemagne VL. Standardized Expression of 18F-NAV4694 and 11C-PiB beta-Amyloid PET Results with the Centiloid Scale. J Nuclear Medicine. 2016 57:1233-7. doi 10.2967/jnucmed.115.171595.

PLEASE NOTE: The correct conversion equation for NAV4694 SUVr to PIB SUVr is:

SUVRNAV4694 = 1.09 x SUVRPiB – 0.07

And the correct equation for NAV4694

SUVr into centiloid is:CL NAV4694 = 100 x (SUVRNAV4694 – 1.031)/1.172

Please email here for questions or additional support regarding the 50-70 minute PIB/NAV Dataset.