JNNP Online First, published on January 5, 2016 as 10.1136/jnnp-2015-312782

PostScript J Neurol Neurosurg Psychiatry: first published as 10.1136/jnnp-2015-312782 on 5 January 2016. Downloaded from

LETTER across the cortex.2 Arterial spin labelling suppressed 3D GRASE imaging readout5 (ASL) is an MRI methodology that uses (acquisition time 6 min 40 s, TI1/2=800/ endogenous arterial blood water as a 2000 ms, voxel size 3.8×3.8×4.0 mm, A novel use of arterial spin tracer to quantify cerebral blood flow refocusing pulse flip angle 130°, five repe- labelling MRI to demonstrate (CBF).3 We aimed to assess the ability of titions). A set of three saturation recovery focal hypoperfusion in ASL to detect patterns of reduced CBF in images (TR=1,2,5 s) with the same PCA, and to compare these results with readout module was also acquired to gen- individuals with posterior those from other imaging modalities. erate tissue M0 and T1 maps for CBF cortical atrophy: a multimodal quantification. On day 2, each participant imaging study METHODS underwent a 10 min PET scan 50 min Five patients fulfilling clinical diagnostic after an intravenous bolus of 300 MBq criteria for PCA,4 and five controls 18F-florbetapir. On day 3, a 20 min PET attended for three scanning visits usually scan was acquired 30–35 min postinjec- INTRODUCTION on consecutive days. On day 1, MRI scans tion of 185 MBq 18F-FDG. PET scans Posterior cortical atrophy (PCA) is a rare were acquired on a 3 T Siemens TIM Trio were performed on a GE Discovery ST neurodegenerative syndrome, typically scanner with a 32-channel phased array PET/CT scanner, with CT scans acquired due to Alzheimer pathology, characterised head-coil. Sequences included a sagittal immediately before each scan for attenu- by impairments in higher-order visual three-dimensional (3D) MPRAGE T1- ation correction. function and other parieto-occipital weighted volumetric scan (acquisition T1-weighted images were processed skills.1 MRI measures of atrophy and time 9 min 23 s, TE/TR/TI=2.9/2200/ using a Gaussian mixture model optimised 18F-labelled fluorodeoxyglucose (FDG) 900 ms, dimensions 256×256×208, through expectation maximisation to positron emission tomography (PET) mea- voxel size 1.1×1.1×1.1 mm), and coronal produce probabilistic segmentations of sures of glucose metabolism typically T2 fluid-attenuated inversion recovery grey matter, white matter and cerebro- show posterior cortical deficits broadly (TE/TR/TI=87/9000/2500 ms, voxel size spinal fluid. Regional labels were pro- mirroring the focal cognitive deficits.1 By 0.9375×0.9375×5 mm). Perfusion data duced simultaneously from the integration contrast, amyloid PET studies demonstrate were acquired using pulsed ASL (FAIR of the template parcellations and the seg- that fibrillar amyloid is widely deposited Q2TIPS) with an 8-segment, background- mentation6 to define (1) a reference copyright. http://jnnp.bmj.com/ on September 26, 2021 by guest. Protected

Figure 1 Single-participant axial images for one control participant and five patients with PCA showing cerebral blood flow (ASL), glucose metabolism (FDG-PET), atrophy (structural MRI), and amyloid deposition (florbetapir-PET). For clinical purposes, 18F-florbetapir images should be read on a grey scale. ASL, arterial spin labelling; CBF, cerebral blood flow; FDG-PET,18 F-labelled fluorodeoxyglucose positron emission tomography; PCA, posterior cortical atrophy; SUVR, standard uptake value ratio.

J Neurol Neurosurg Psychiatry Month 2016 Vol 0 No 0 1 Copyright Article author (or their employer) 2016. Produced by BMJ Publishing Group Ltd under licence. PostScript J Neurol Neurosurg Psychiatry: first published as 10.1136/jnnp-2015-312782 on 5 January 2016. Downloaded from region (whole cerebellum) for the PET the cortex, including the frontal lobes. 5Neuroradiological Academic Unit, Brain Repair & normalisation; and (2) frontal, medial Regression analyses showed no significant Rehabilitation, UCL Institute of Neurology, London, UK temporal, lateral temporal, parietal, pos- relationships between amyloid load and Correspondence to Dr Jonathan M Schott, Dementia terior cingulate and occipital grey matter any other biomarker. There was no evi- Research Centre, UCL Institute of Neurology, lobar regions which were subsequently denceforarelationshipbetweengrey 8-11 Queen Square, London WC1N 3BG, UK; [email protected] corrected for total intracranial volume. matter volume and CBF (p=0.4), border- ASL images were aligned to the line evidence for a relationship between Acknowledgements The authors acknowledge the T1-weighted image using a symmetric grey matter volume and FDG-PET metab- support of the Leonard Wolfson Experimental affine registration algorithm to reduce the olism (p=0.05), but a significant correl- Neurology Centre, the NIHR Queen Square Dementia 7 Biomedical Research Unit and UCL/UCLH Biomedical effects of any residual motion. CBF maps ation between regional CBF and FDG-PET Research Centre. The Dementia Research Centre is an 2 were generated by fitting a derivative of metabolism (p=0.004, R =0.26). Alzheimer’s Research UK Co-ordinating Centre. the general kinetic model to the ASL Contributors All the authors were responsible for 8 data. Florbetapir scans were rated drafting and revising the manuscript for content. ML, amyloid-positive/negative on grey-scaled DISCUSSION JCD, XG, SO, JMS were responsible for study concept images according to standard guidelines; All individuals in this study had clinical and statistical analysis; ML, AM, JCD, MM, MJC, DLT, FDG scans were rated for regional hypo- and cognitive phenotypes consistent with JB, BFH, JMS for analysis of the data; RMA, SJC, JDW, fl CJM, NCF, and JMS were responsible for patient metabolism. Standard uptake value ratios PCA, and unequivocally positive orbeta- recruitment; JCD, DLT, EDV for scan acquisition; and were calculated for each grey matter lobar pir PET scans consistent with underlying JMS for study supervision. region and as a composite, using whole AD pathology. In keeping with previous Funding AVID Radiopharmaceuticals (a wholly owned cerebellum as the reference region. studies, amyloid deposition was widely dis- subsidiary of Eli Lilly), and the UK FTD Support Group. 2 Neuroimaging data from each patient tributed across the cortex while atrophy, Competing interests ML is supported by funding were compared against the controls using hypometabolism and CBF deficits were all from Alzheimer’s Research UK (ART-TRFUS2011-2). a modified t test developed to conduct more marked in posterior cortical regions. SC is supported by ESRC/NIHR (ES/K006711/1), EPSRC ’ single participant comparisons, by treating Previous studies have shown corres- (EP/M006093/1) and an Alzheimer s Research UK Senior Research Fellowship. SO is supported by the the control sample as sample statistics pondence between the pattern of cerebral EPSRC (EP/H046410/1, EP/J020990/1, EP/K005278), 9 rather than as population parameters. To hypometabolism measured using the MRC (MR/J01107X/1), the EU-FP7 (FP7-ICT-2011- explore the regional relationships between FDG-PET and ASL measures of CBF in 9-601055), the NIHR BRC UCLH/UCL HII (BW.mn. the different techniques, z-scores for each typical late-onset AD10 and semantic BRC10269). MM and DLT are supported by the UCL modality and each lobe were calculated dementia.11 Here, we extend these find- Leonard Wolfson Experimental Neurology Centre (PR/ ylr/18575). JMS acknowledges the support of ARUK, relative to controls, and linear regression ings to PCA, but importantly show that the NIHR Biomedical Research Centre, Queen Square copyright. was used across all brain regions to these group level changes extend to indi- Dementia Biomedical Research Unit, and Medical compare imaging modalities. vidual patients. Posterior cortical CBF Research Council. reduction was detectable in each of the Patient consent Obtained RESULTS individual patients with PCA and broadly Ethics approval Queen Square NRES Committee. Participant demographics and neuroimaging mirrored by the observed focal patterns of — Provenance and peer review Not commissioned; data are shown in online supplementary FDG-PET glucose hypometabolism the externally peer reviewed. table S1. At the time of scanning, the gold standard for detecting regional meta- ▸ Additional material is published online only. To view patients (3 female, 2 male) had a mean±SD bolic changes. Both modalities exhibited please visit the journal online (http://dx.doi.org/10. age of 59.4±1.8 years. Single-participant more extensive changes than expected 1136/jnnp-2015-312782) images are shown in the figure 1,alongside purely due to the posterior atrophy seen the results from a control participant (male, on structural imaging. These data show http://jnnp.bmj.com/ 59 years old) as a reference. All patients that ASL, a non-invasive MR sequence were amyloid-positive, with fibrillar which can implemented on standard clin- amyloid distributed widely across the ical 3 T MR scanners, can usefully dem- cortex. Much more focal, posterior cortical onstrate CBF changes in individual fl Open Access This is an Open Access article reductions in cerebral blood ow, glucose patients with PCA. ASL may be a useful distributed in accordance with the terms of the Creative metabolism and atrophy were seen. While diagnostic adjunct in suspected PCA—and Commons Attribution (CC BY 4.0) license, which — permits others to distribute, remix, adapt and build atrophy in posterior regions was present in potentially other focal dementias where on September 26, 2021 by guest. Protected all patients, on visual assessment it was less an FDG-PET scan is unavailable. upon this work, for commercial use, provided the fi original work is properly cited. See: http:// pronounced than the metabolic de cits creativecommons.org/licenses/by/4.0/ assessed using either FDG-PET or ASL. 1 2 Manja Lehmann, Andrew Melbourne, To cite Lehmann M, Melbourne A, Dickson J C, et al. fi 3 1 Quanti cation of the regional results for John C Dickson, Rebekah M Ahmed, J Neurol Neurosurg Psychiatry Published Online First: 2 2 each modality is shown in the online sup- Marc Modat, M Jorge Cardoso, [please include Day Month Year] doi:10.1136/jnnp- 1,4 4,5 David L Thomas, Enrico De Vita, 2015-312782 plementary table S1. Compared with con- 1 1 trols, there were consistent and significant Sebastian J Crutch, Jason D Warren, Colin J Mahoney,1 Jamshed Bomanji,3 Received 25 November 2015 reductions in grey matter volume and Brian F Hutton,3 Nick C Fox,1 Xavier Golay,4 Accepted 27 November 2015 glucose metabolism in the parietal lobe in Sebastien Ourselin,2 Jonathan M Schott1 J Neurol Neurosurg Psychiatry 2016;0:1–3. all patients, and in the parietal, occipital 1Dementia Research Centre, UCL Institute of Neurology, doi:10.1136/jnnp-2015-312782 and lateral temporal lobes in the majority. London, UK Cerebral blood flow reductions were seen 2Translational Imaging Group, Centre for Medical Image in the parietal and occipital lobes and pos- Computing, University College London, London, UK REFERENCES 3Institute of Nuclear Medicine, University College 1 Crutch SJ, Lehmann M, Schott JM, et al. Posterior terior cingulate for all patients, reaching London Hospitals, London, UK cortical atrophy. Lancet Neurol 2012;11:170–8. fi 4 signi cance in three. In all cases amyloid Lysholm Department of Neuroradiology, National 2 Lehmann M, Ghosh PM, Madison C, et al. Diverging deposition was distributed widely across Hospital for Neurology and Neurosurgery, London, UK patterns of amyloid deposition and hypometabolism

2 J Neurol Neurosurg Psychiatry Month 2016 Vol 0 No 0 PostScript J Neurol Neurosurg Psychiatry: first published as 10.1136/jnnp-2015-312782 on 5 January 2016. Downloaded from

in clinical variants of probable Alzheimer’s disease. the 17th Annual Meeting of ISMRM. Honolulu, HI, for ASL in dementia. Magn Reson Med 2015;73: Brain 2013;136:844–58. USA; 2009:622. 102–16. 3 Haller S, Zaharchuk G, Thomas D, et al. Arterial spin 6 Cardoso MJ, Modat M, Wolz R, et al. Geodesic 9 Crawford JR, Garthwaite PH. Investigation of the label perfusion of the brain: emerging clinical information flows: spatially-variant graphs single case in neuropsychology: confidence limits on applications. Radiology 2015;In Press. and their application to segmentation and the abnormality of test scores and test score 4 Tang-Wai DF, Graff-Radford NR, Boeve BF, et al. fusion. IEEE Trans Med Imaging 2015;34: differences. Neuropsychologia 2002;40:1196–208. Clinical, genetic, and neuropathologic characteristics 1976–88. 10 Chen Y, Wolk DA, Reddin JS, et al. Voxel-level of posterior cortical atrophy. Neurology 2004;63: 7 Modat M, Cash DM Daga P, et al. Global image comparison of arterial spin-labeled perfusion MRI 1168–74. registration using a symmetric block-matching and FDG-PET in Alzheimer disease. Neurology 5 Feinberg DA, Ramanna S, Guenther M. Evaluation of approach. J Med Imaging 2014;1:024003. 2011;77:1977–85. new ASL 3D GRASE sequences using parallel 8 Alsop DC, Detre JA, Golay X, et al. Recommended 11 Newberg AB, Wang J, Rao H, et al. Concurrent CBF imaging, segmented and interleaved k-space at 3T implementation of arterial spin-labeled perfusion MRI and CMRGlc changes during human brain activation with 12- and 32-channel coils. In: Proceedings of for clinical applications: a consensus of the ISMRM by combined fMRI-PET scanning. Neuroimage perfusion study group and the European consortium 2005;28:500–6. copyright. http://jnnp.bmj.com/ on September 26, 2021 by guest. Protected

J Neurol Neurosurg Psychiatry Month 2016 Vol 0 No 0 3