Sepulcre J, Grothe MJ, Sabuncu M, Chhatwal J, Schultz AP, Hanseeuw B, El Fakhri G, Sperling R, Johnson KA. Hierarchical Organization of Tau and Amyloid Deposits in the Cerebral Cortex. JAMA Neurol. 2017 Jul 1;74(7):813-820. PubMed.

Recommends

Please login to recommend the paper.

Comments

  1. Jorge Sepulcre and colleagues have performed an elegant voxel-wise hierarchical clustering approach to [11C]PIB (Aβ) and [18F]flortaucipir (tau) PET data, in order to investigate the cortical distribution of the two AD hallmark pathologies in 88 cognitively normal individuals. Using silhouette analysis (i.e., assessment of the number of voxels within and outside the hierarchical cluster and of cluster stability) the optimum number of clusters turned out to be four for both PET tracers. There were notable differences in the distribution, intensity, and size of Aβ and tau clusters, providing important information about the spatial relationships between Aβ and tau pathology in a clinically normal population. There are many interesting observations made in this dense paper, of which I will highlight a few. 

    1. To date, most studies employing tau PET tracers (Cho et al., 2016; Johnson et al., 2016; Schöll et al., 2016; Schwarz et al., 2016) have selected regions of interest (ROIs) or derivatives based on neuropathological staging of tau pathology [i.e., Braak NFT staging, (Braak and Braak, 1991)]. Computationally extracting detailed regional information with full brain coverage is an advantage of neuroimaging over neuropathologic data, however, and provides the unique opportunity to capture tau pathology in the living human brain. An advantage of clustering (or any data-driven approach) is that it has the potential to identify distinct sources in the signal (i.e., target and noise). It is encouraging that the identified hierarchical clusters (specifically no. 1 and no. 2) seem biologically meaningful as they include brain regions that are vulnerable in both early (cluster no. 1) and more advanced (cluster no. 2) stages of AD. Logical next steps would be to investigate whether the identified clusters are stable across cohorts (i.e., external validation) and in different clinical populations (e.g., MCI or AD dementia patients).

    2. Related to the above, there was some overlap between the hierarchical tau clusters and traditional neurofibrillary tangles (NFTs) Braak stage areas. Cluster no. 1 partially mapped onto areas affected in Braak stages I-IV (medial temporal lobe, inferior temporal cortex and orbitofrontal cortex), while cluster no. 2 showed some overlap with Braak stage V/VI (isocortical regions). The correspondence is imperfect, however, because there are substantial differences between the data-driven clusters in this study and neuropathologically-derived ROIs. In light of ongoing efforts to define tau PET-positivity, it is important to test whether the identified hierarchical clusters yield greater predictive accuracy for cognitive status and brain atrophy compared to traditional, neuropathologically derived tau clusters. It might very well be that agnostic, data-driven approaches outperform theory-driven ROIs due to the abundance of information available in neuroimaging data.

    3. Figure 3 nicely highlights different features of Aβ and tau clusters. For tau PET, the highest signal intensity was found in a rather constrained cluster (no. 1) that typically shows early signs of neurodegeneration (e.g., medial temporal lobe structures), while the greatest total signal was found in the larger cluster (no. 2) comprising frontoparietal and lateral occipital regions. For Aβ PET, however, both intensity and total signal was highest in the same large cluster (no. 1) of heteromodal association regions. This suggests a separation in space (and possibly time) between Aβ and tau pathology, and that the location of high-intensity tau pathology might dictate where neurodegeneration will occur. This needs to be tested in future longitudinal studies.

    References:

    Braak H, Braak E. Neuropathological stageing of Alzheimer-related changes. Acta Neuropathol. 1991;82(4):239-59. PubMed.

    Cho H, Choi JY, Hwang MS, Kim YJ, Lee HM, Lee HS, Lee JH, Ryu YH, Lee MS, Lyoo CH. In vivo cortical spreading pattern of tau and amyloid in the Alzheimer disease spectrum. Ann Neurol. 2016 Aug;80(2):247-58. Epub 2016 Jul 8 PubMed.

    Johnson KA, Schultz A, Betensky RA, Becker JA, Sepulcre J, Rentz D, Mormino E, Chhatwal J, Amariglio R, Papp K, Marshall G, Albers M, Mauro S, Pepin L, Alverio J, Judge K, Philiossaint M, Shoup T, Yokell D, Dickerson B, Gomez-Isla T, Hyman B, Vasdev N, Sperling R. Tau positron emission tomographic imaging in aging and early Alzheimer disease. Ann Neurol. 2016 Jan;79(1):110-9. Epub 2015 Dec 15 PubMed.

    Schöll M, Lockhart SN, Schonhaut DR, O'Neil JP, Janabi M, Ossenkoppele R, Baker SL, Vogel JW, Faria J, Schwimmer HD, Rabinovici GD, Jagust WJ. PET Imaging of Tau Deposition in the Aging Human Brain. Neuron. 2016 Mar 2;89(5):971-82. PubMed.

    Schwarz AJ, Yu P, Miller BB, Shcherbinin S, Dickson J, Navitsky M, Joshi AD, Devous MD Sr, Mintun MS. Regional profiles of the candidate tau PET ligand 18F-AV-1451 recapitulate key features of Braak histopathological stages. Brain. 2016 May;139(Pt 5):1539-50. Epub 2016 Mar 2 PubMed.

    View all comments by Rik Ossenkoppele

  2. While the early results are very encouraging, it would be great to see the hierarchical clustering approach implemented in a larger sample, as well as with the inclusion of cognitively impaired individuals to ensure that the complete range of outcome values for amyloid and tau are included.

    View all comments by Prashanthi Vemuri

Make a Comment

To make a comment you must login or register.

This paper appears in the following:

News

  1. Analysis of PET Scans Suggests Link Between Amyloid and Tau