TauIQ: A Canonical Image Based Algorithm to Quantify Tau PET Scans

Abstract

Recently, Amyloid^IQ was introduced as a new canonical image-based algorithm to quantify amyloid PET scans and demonstrated increased power over traditional SUV ratio (SUVR) approaches when assessed in cross-sectional and longitudinal analyses. We build further on this mathematical framework to develop a Tau^IQ algorithm for the quantitative analysis of the more complex spatial distribution displayed by tau PET radiotracers. Methods: Cross-sectional (n = 615) and longitudinal (n = 149) ¹⁸F-flortaucipir data were obtained from the Alzheimer's Disease Neuroimaging Initiative along with necessary adjunct amyloid PET and T1-weighted structural MRI data. A subset of these data were used to derive a chronological tau dataset, using Amyloid^IQ analysis of associated amyloid PET data to calculate the subject's temporal position in the canonical AD disease process, from which canonical images for the nonspecific and specific binding components of ¹⁸F-flortaucipir in AD were calculated. These 2 canonical images were incorporated into the Tau^IQ algorithm that enables the quantification of both global and local tau outcome measures using an image-based regression and statistical parametric analysis of the initial residual image. Performance of the Tau^IQ algorithm was compared with SUVR approaches for cross-sectional analyses, longitudinal analyses, and correlation with clinical measures (Alzheimer Disease Assessment Scale-Cognitive Subscale [ADAS-Cog], Clinical Dementia Rating scale-sum of boxes [CDR-SB], and Mini-Mental State Examination [MMSE]). Results: Tau^IQ successfully calculated global tau load (Tau_L) in all 791 scans analyzed (range, -3.5% to 185.2%; mean ± SD, 23% ± 20.5%) with a nonzero additional local tau component being required in 31% of all scans (cognitively normal [CN], 22%; mild cognitive impairment [MCI], 35%; dementia, 72%). Tau^IQ was compared with the best SUVR approach in the cross-sectional analysis (Tau_L increase in effect size: CN- vs. CN+, +45%; CN- vs. MCI+, -5.6%; CN- vs. dementia+, +2.3%) (+/- indicates amyloid-positive or -negative) and correlation with clinical scores (Tau_L increase in r²: CDR-SB+, 7%; MMSE+, 38%; ADAS-Cog+, 0%). Tau^IQ substantially outperformed SUVR approaches in the longitudinal analysis (Tau^IQ increase in power: CN+, >3.2-fold; MCI+, >2.2-fold; dementia+, >2.9-fold). Conclusion: Tau_L as calculated by Tau^IQ provides a superior approach for the quantification of tau PET data. In particular, it provides a substantial improvement in power for longitudinal analyses and the early detection of tau deposition and thus should have significant value for clinical imaging trials in AD that are investigating the attenuation of tau deposition with novel therapies.

Keywords: IQ platform; canonical images; flortaucipir; power in clinical trials; statistical analysis; tau PET.

Graphical abstract

FIGURE 1.

Methodology for creating ¹⁸F-flortaucipir canonical images. (A) Creation of chronological ¹⁸F-flortaucipir dataset using associated ¹⁸F-florbetapir data. (B) Generation of ¹⁸F-flortaucipir K and NS canonical images from voxelwise modeling of chronological dataset.

FIGURE 2.

Canonical images for ¹⁸F-flortaucipir PET derived from spatiotemporal modeling of chronological AD data in MNI152 space: tau canonical image (K) (A), nonspecific canonical image (NS) (B), and reference structural T1-weighted MR image (C).

FIGURE 3.

Tau^IQ algorithm that estimates global and local tau outcome measures.

FIGURE 4.

Examples of Tau^IQ decomposition of ¹⁸F-flortaucipir data into nonspecific, global, and local tau signals.

FIGURE 5.

Cross-sectional analysis shown as box plots of distributions of ¹⁸F-flortaucipir Tau_L and SUVR outcome measures by clinical group.

FIGURE 6.

Longitudinal analysis shown as box plots of distributions of change in ¹⁸F-flortaucipir Tau_L and SUVR outcome measures.

FIGURE 7.

Correlations between clinical scores (CDR-SB, MMSE, and ADAS-Cog) and tau imaging outcome measures.