HAI Book 2020 - Dec 18B

01

Keywords: amyloid-PET,tau-PET,Neurofilament Ligh
40

ht Chain,YKL40,CSF
02

Kang, Min Su

P139: Activated microglia and amyloi
leading to cognitive dysfunction in Alz

Min Su Kang1,3,4, Julie Ottoy2, Melissa Savard1,
Mathotaarachchi1, Andréa Benedet1,4, Mira Cha
Cécile Tissot1,4, Emilie Thomas1, Jenna Stevens
Wang1,4, Gassan Massarweh3, Jean-Paul Soucy3

1Translational Neuroimaging Laboratory, The McGill
Montréal, QC, Canada
2Molecular Imaging Center Antwerp, University of An
3Montreal Neurological Institute, Montréal, QC, Cana
4Douglas Mental Health Research Institute, Verdun, Q

Background: Growing evidence from in vitro studies
(Aβ) and neuroinflammation leads to neurofibrillary t
multimodal PET study, we aimed to reveal that the Aβ
([11C]PBR28) converges to potentiate the NFTs ([18F]
the NFTs deposition driven by Aβ and neuroinflamma
individuals.

Methods: A total of 95 participants (67 CN, 18 MCI,
binding for TSPO from the TRIAD cohort underwent
Static 40–60min [18F]NAV4694, 90-110min [18F]MK
generated. All PET images were normalized to the AD
region, and adjusted for age, sex, education, and APO

Results: We showed a significant synergistic effect b
Precuneus/PCC, entorhinal cortex, basolateral tempor
was mainly driven by the Aβ+ group as there was no s
the greater NFTs driven by the interactive effect in Aβ
cognitive dysfunction based on MOCA (F(1)=7.04, p=
with CDR-SoB (F(1)=3.09, p=0.084).

Conclusion: Our study demonstrated that the Aβ and
deposition in AD vulnerable regions. Consequently, N
cognitive dysfunction in AD. This supports the idea th
combination therapy may have greater therapeutic eff

Keywords: Alzheimer’s pathophysiology, cognitive dy

40

id load potentiate tau deposition
zheimer’s disease

,4, Tharick Pascoal1,4, Sulantha
amoun1,4, Joseph Therriault1,4, Firoza Lussier1,4,
son1,4, Nesrine Rahmouni1,4, Jaime Arias1,4, Tina
3, Serge Gauthier1,3,4, Pedro Rosa-Neto1,3,4

l University Research Centre for Studies in Aging,

ntwerp, Antwerp, Belgium
ada
QC, Canada

s suggest that the interaction between amyloid-beta
tangles (NFT) in Alzheimer’s disease (AD). In this
β ([18F]NAV4694) and activated microglia
]MK6240) deposition in AD. We hypothesized that
ation leads to cognitive dysfunction in Aβ+

and 10 AD; 60 Aβ- and 35 Aβ+) with high-affinity
3 PET scans and MMSE, CDR-SoB, and MOCA.
K6240, and 60-90min [11C]PBR28 SUVR images were
DNI template, used cerebellar grey as a reference
OE.
between Aβ and neuroinflammation on NFTs in
ral cortex, and medial frontal cortex (Figure). This
significant interaction in Aβ- group. Consequently,
β+ group was associated with significantly greater
=0.0097) and MMSE(F(1)=3.86, p=0.053) and a trend

neuroinflammation converge to greater NFTs
NFTs load in these regions was associated with greater
hat targeting Aβ and neuroinflammation as a
ficacy in AD.

ysfunction, amyloid, tau, activated microglia

03

Kang, Min Su

P140: Tauopathy in females is more vu
neuroinflammation in Alzheimer’s dis

Min Su Kang1,3, Julie Ottoy2, Mira Chamoun1, S
Andréa Benedet1, Tharick Pascoal1, Joseph The
Thomas1, Jenna Stevenson1, Nesrine Rahmouni
Massarweh3, Jean-Paul Soucy3, Serge Gauthier1

1Translational Neuroimaging Laboratory, The McGill
Montréal, QC, Canada
2Molecular Imaging Center Antwerp, University of An
3Montreal Neurological Institute, Montréal, QC, Cana

Background: Understanding sex dimorphism in neur
for clinical trial enrichment and personalized medicin
Alzheimer’s disease (AD) pathophysiology is still poo
sex modulates the association among the major targets
tau ([18F]MK6240), and neuroinflammation ([11C]PB
hypothesized that amyloid or neuroinflammation show
to males.

Methods: A total of 95 participants (67 CN, 18 MCI,
from the TRIAD cohort underwent 3 PET scans. Stati
[18F]MK6240, and 60-90min [11C]PBR28 SUVR im
normalized to the ADNI template, used cerebellar gre
education, and APOE.

Results: Our results showed that the females had a gr
Precuneus/PCC, basolateral temporal cortex, and med
demonstrated that the females had a greater positive a
Precuneus/PCC and basolateral temporal cortex comp

Conclusion: Our study demonstrated that the sex inte
where the greater positive association between tau and
females compared to males. This study supports the fr
to males in the AD pathophysiology.

40

ulnerable to amyloid or
sease

Sulantha Mathotaarachchi1, Melissa Savard1,
erriault1, Firoza Lussier1, Cécile Tissot1, Emilie
i1, Jaime Fernandez-Arias1, Tina Wang1, Gassan
1, Pedro Rosa-Neto1

l University Research Centre for Studies in Aging,
ntwerp, Antwerp, Belgium
ada
rodevelopment and neurodegeneration is imperative
ne. However, the manifestation of sex dimorphism in
orly understood. Here, we aimed to investigate how
s of AD clinical trials – amyloid ([18F]NAV4694),
BR28) – through a multimodal PET study. We
w a greater positive effect on tau in females compared

and 10 AD) with a high-affinity binder for TSPO
ic 40–60min [18F]NAV4694, 90-110min
mages were generated. All PET images were
ey as a reference region, and adjusted for age, sex,

reater positive association between tau and amyloid in
dial frontal cortex (Figure). In addition, our results
association between tau and neuroinflammation in
pared to males (Figure).
ersects amidst the complex AD pathophysiology
d amyloid or neuroinflammation was observed in
ramework that females are more vulnerable compared

04

Keywords: Sex, amyloid, tau, neuroinflammation, cog
40

gnitive dysfunction
05

Kim, Jaeho

P141: Prediction of brain tau accumul
impairment patients using multimoda
approach

Jaeho Kim1,2,3, Seongbeom Park1, Yuhyun Park
Hyemin Jang1,2,3, Hee Jin Kim1,2,3, Duk L. Na1,2

1Department of Neurology, Samsung Medical Center,
Korea
2Neuroscience Center, Samsung Medical Center, Seou
3Samsung Alzheimer Research Center, Samsung Medi
4Stem Cell & Regenerative Medicine Institute, Samsun
5Department of Clinical Research Design & Evaluatio
6Department of Health Sciences and Technology, SAIH

Background: Brain tau accumulation is an important
prodromal Alzheimer’s disease (AD), brain tau burden
PET imaging is costly, and access is restricted to spec
prediction model to detect prodromal AD who are liab
aimed to develop a classifier of tau positivity in prodr

Methods: We used the Alzheimer’s Disease Neuroim
patients who underwent 3.0T MRI scanning, 18F-AV
at baseline. We defined participants as having abnorm
III/IV by a conditional inference tree approach. Then,
develop a statistical classifier to classify tau positivity

Results: Our machine learning algorithm of GBM and
patients with the AUC of 0.865 for GBM and 0.792 fo
neuropsychological test to baseline patient informatio
from 0.681 to 0.815 in the GBM model and further in
added. Through analysis of impact order of variables i
occipital lobe, neuropsychological tests of memory do
features for each classifier.

Conclusion: We have developed a novel algorithm fo
using baseline information, neuropsychological test re
algorithm showed fair to good accuracy to predict bra
population targeting tau and predict disease severity a

40

lation in amyloid positive cognitive
al biomarkers with machine learning

k1,5, Sung Hoon Kang1,2,3, Soo Jong Kim1,
2,3,4,6, Hyejoo Lee1,2,3, Sang Won Seo1,2,3,5,6

, Sungkyunkwan University School of Medicine, Seoul,

ul, Korea
ical Center, Seoul, Korea
ng Medical Center, Seoul, Korea
on, SAIHST, Sungkyunkwan University, Seoul, Korea
HST, Sungkyunkwan University, Seoul, Korea

marker to predict cognitive function and decline. In
n appears in various spectrum. However, brain tau
cialized centers. Therefore, there is a need to develop a
ble to be tau positive prior to tau PET imaging. We
romal AD using statistical learning algorithms.
maging Initiative (ADNI) dataset, and included MCI
V45(florbetapir) PET, and AV1451(flortaucipir) PET
mal “Tau” (T+) if their in-vivo Braak stage was ≥
, we performed machine learning algorithms to
y.

d RF predicted tau positivity in prodromal AD
or RF. We found that additional information of
on (age, gender, education) increased the AUC value
ncreased to 0.865 when the brain MRI information was
in each classifier, cortical thickness of parietal,
omain, and sex were found to be more important

or predicting the brain tau burden in prodromal AD
esults, and brain structural changes. Since our novel
ain tau burden, it can be useful to screen study
and prognosis.

06

Keywords: brain tau, machine learning, prodromal A
40

AD, classifier, tau PET
07

Klein, Gregory

P142: Comparison of longitudinal cha
among cognitively unimpaired and pa
the BioFINDER2 study

Gregory Klein1, Sandra Sanabria4, Antoine Leuz
Sebastian Palmqvist2,3, Erik Stomrud2,3, Ruben S

1Roche Pharma Research and Early Development, Ba
2Clinical Memory Research Unit, Department of Clini
3Memory Clinic, Skåne University Hospital, Malmö, S
4Genentech, Inc., San Francisco, CA, US

Objectives: Compare tau PET metrics of longitudinal
cognitively unimpaired (CU) individuals, patients wit
disease (AD) dementia, recruited to date in the Swedis

Methods: Study participants (N=51) are scanned at ba
obtained for 20 minutes, 70 min post injection. A MPR
Two processing pipelines are compared: 1) FreeSurfer
computing SUVR, tau extent and tau load [2]. Both pi
approximating the anatomical definitions of entorhina
and neocortical (V/VI) stages. Subjects were analysed
(Aβ+ or Aβ-) and baseline clinical diagnosis: cognitiv
Cohen’s effect size of annualized longitudinal change

Results: Both pipelines showed highest effect size in
and Braak V/VI for AD_Aβ+ groups respectively. Fre
to the Hammer’s atlas method for SUVR-based comp
MCI groups was obtained with the Hammers tau load

Conclusions: Tau burden change seen with 18F-RO9
tau, with the highest effect size and percent change se
Braak regions III/IV and V/VI for Aβ+ MCI and AD s
may offer superior effect size in detecting change for

References

[1] Cho et al., Ann Neurol ; 2016 ; 80 :247-258

[2] Sanabria Bohorquez, et al Eur J Nucl Med Mol Im

40

ange metrics for 18F-RO948 PET
atients with MCI or AD dementia in

zy2, Edilio Borroni1, Niklas Mattsson2,
Smith2,3, Oskar Hansson2,3

asel, Switzerland
ical Sciences, Malmö, Sweden
Sweden

l change using 18F-RO948 PET in a population of
th mild cognitive impairment (MCI) or Alzheimer’s
sh BioFINDER2 study.
aseline and 12 months with 18F-RO948 tau PET,
RAGE T1-MRI sequence is used for PET processing.
er SUVR [1], and 2) A Hammer’s atlas-based method
ipelines used a Braak staging model, defining ROIs
al/hippocampal (stage I/II), temporal/limbic (III/IV),
d separately based on CSF Aβ42/Aβ40 amyloid status
vely unimpaired (CU), MCI and dementia due to AD.
e is used to compare processing pipelines.
Braak I/II for CU_Aβ+, Braak III/IV for MCI_Aβ+
eesurfer methods showed higher effect size compared
parisons. Highest effect size overall in the CU and
d metric.
948 replicates the post-mortem spreading pattern of
een in Braak regions I/II for Aβ+ CU subjects, and
subjects, respectively. A tau extent and load metric
early AD.

maging. 2019

08

Keywords: Tau PET, RO948, Image analysis
40

09

Mormino, Elizabeth

P143: Validation of clinical protocols f
tau PET/MRI images

Mary Ellen Koran1, Sara Shams1, Patrick Adam
Corso1, Madison Hunt1, Jessa Castillo1, Jacob H
Michael Greicius1, Anthony Wagner1, Greg Zah
Elizabeth Mormino1

1Stanford University, Stanford, CA, US

Diagnosing tau pathology in vivo is crucial in disease
study was to create and validate a clinical visual rating
with regional values and disease.
A protocol was created following Braak staging of tau
tracer uptake greater than inferior cerebellum--0: none
temporal lobe (MTL), 2: inferior/lateral temporal lobe
radiologists were blinded to clinical data such as diagn
PET/MRI for 48 subjects along the non-AD and AD s
and AD dementia, as well as semantic variant primary
progressive supranuclear palsy, and dementia with Le
MRI system at Stanford.
Using this protocol, there was significant agreement f
a score of 2 and higher were consistently rated this wa
reliability for the protocol that reflected disagreement
1m scores; K = 0.67; p <0.0001). Despite differences
temporal lobe uptake, group level differences were ob
value ratios (SUVRs, inferior cerebellum normalized)
cognitive impairment and AD dementia (p<0.0001). C
12.7, p<0.0001).
As tau tracers are being implemented in clinical trials,
PET/MRI is promising for characterizing disease burd
protocols and quantitative SUVRs.

41

for clinicians analyzing 18F-PI-2620

ms1, Emily Cazevedo1, Tyler N. Toueg1, Nicole
Hall1, Sharon Sha1, Carolyn Fredericks1,
harchuk1, Guido Davidzon1, Frederick Chin1,

diagnosis and therapy development. The goal of this
g protocol for 18F-PI2620 PET/MRI and correlate

u accumulation with the following scoring system of
e, 1e: entorhinal cortex, 1h: hippocampus, 1m: medial
e, 3: in neocortex beyond the temporal lobe. Two
nosis and age, and independently evaluated fused
spectrum (normal controls, mild cognitive impairment
y progressive aphasia, corticobasal syndrome,
ewy bodies. Data were collected on a GE Signa PET-

for cases with cortical binding, such that all cases with
ay across readers. Overall, there was interrater
t within the medial temporal lobe (stages 1e, 1h, and
in qualitative reads for individuals with focal medial
bserved for medial temporal lobe standardized uptake
) between normal controls and patients with mild
Cortical SUVRs correlated with radiologist rating (F=

, clinical evaluation is necessary. 18F-PI2620
den in vivo clinically, using modified Braak staging

10

Keywords: tau PET, PI2620, clinical, visual
41

11

Benedet, Andrea

P144: Cerebrospinal fluid tau phospho
biomarker comparison in the detection
in the Alzheimer’s disease spectrum

Andrea Benedet1, Nicholas Ashton2,3,4,5, Tharick
Eugeen Vanmechelen6, Thomas Karikari2, Sulan
Therriault1, Mira Chamoun1, Henrik Zetterberg2

1Translational Neuroimaging Laboratory, McGill Cen
Montreal, QC, Canada
2Department of Psychiatry and Neurochemistry, Instit
Academy at the University of Gothenburg, Möndal, Sw
3Wallenberg Centre for Molecular and Translational
Sweden
4King’s College London, Institute of Psychiatry, Psych
Neuroscience Institute, London, UK
5NIHR Biomedical Research Centre for Mental Health
London & Maudsley NHS Foundation, London, UK
6ADx NeuroSciences, Ghent, Belgium
7Clinical Neurochemistry Laboratory, Sahlgrenska Un
8Department of Neurodegenerative Disease, UCL Inst
9UK Dementia Research Institute at UCL, London, UK
10Montreal Neurological Institute, Montreal, QC, Can
11Department of Neurology and Neurosurgery, McGil

Introduction: Neurofibrillary tangles (NFT) are key h
and they are constituted of paired helical filaments. Th
of tau protein (p-tau), which can occur on several ami
181 (p-tau181) is an established and specific biomarke
merit further evaluation as they can potentially better
in AD.

Aims: To compare the cross-sectional association bet
amyloid PET in the AD spectrum using data from the
(TRIAD) cohort.

Methods: Were evaluated111 participants (17 young
impairment (MCI) and 12 AD) with cross-sectional C
[18F]MK6240 and [18F]AZD4694 PET.For PET, stand
using the cerebellar cortex as reference tissue. LUMIP
p-tau181 whereas CSF p-tau231 was quantified using
regression was implemented at the voxel level, in orde
based measures.

Results: Overall, as compared to p-tau181 and t-Tau/
MMSE, global amyloid and tau load(Figure 1). ROC
diagnosis (Figure 2). The voxel-wise analysis demons
in the temporal, posterior cingulate and medial prefron
strongly associated with tau PET than p-tau181(Figur
associations with amyloid PET, with p-tau231 showin

Conclusion: These findings suggest that CSF p-tau23
41

orylated at amino acid 181 or 231: a
n of neurofibrillary tangle pathology

k Pascoal1, Erik Stoops6, Cindy Francois6,
ntha Mathotaarachchi1, Melissa Savard1, Joseph
2,7,8,9, Kaj Blennow2,7, Pedro Rosa-Neto1,10,11

ntre for Studies in Aging, McGill University,

tute of Neuroscience & Physiology, the Sahlgrenska
weden
Medicine, University of Gothenburg, Gothenburg,

hology & Neuroscience, Maurice Wohl Clinical

h & Biomedical Research Unit for Dementia at South

niversity Hospital, Möndal, Sweden
titute of Neurology, London, UK
K
nada
ll University, Montreal, QC, Canada

hallmarks of the Alzheimer’s disease (AD) pathology
hese are promoted by the abnormal phosphorylation
ino acids. P-tau phosphorylated at threonine residue
er for AD. However, alternative phosphorylation sites
index NFT pathology and track disease progression

tween CSF p-tau231 and p-tau181 with tau and
Translational Biomarkers of Aging and Dementia

controls, 57 normal controls (CN), 25 mild cognitive
CSF p-tau181 and 231, t-Tau/p-Tau181 ratio and
dard uptake value ratios (SUVR) were determined
PULSE® G1200 (Fujirebio) was used to measure CSF
g a custom ELISA assay (ADx Neuroscience). Linear
er to examine the association between CSF and PET-

/p-Tau181 ratio, p-tau231 was better correlated with
curves revealed p-tau231 as the best predictor of
strated an association between CSF p-tau and tau PET
ntal cortices, in which p-tau231 seemed to be more
re 3). Similar findings were observed in the
ng high correlations with amyloid load.

31 may be a superior fluid biomarker that reflects NFT
12

pathology in the AD spectrum.
41

13

Keywords: phosphorylated tau, MK6240, NFT, PET,
41

Alzheimer’s disease,
14

Leuzy, Antoine

P145: Diagnostic performance of [18F
tomography in the differentiation of A
disorders

Antoine Leuzy1, Ruben Smith1,2, Rik Ossenkopp
Gregory Klein4, Tomas Olsson5, Jonas Jögi6, Se
Strandberg1, Erik Stomrud1,8, Oskar Hansson1,8

1Clinical Memory Research Unit, Department of Clin
2Department of Neurology, Skåne University Hospital
3VU University Medical Center, Neuroscience Campu
4Roche Pharma Research and Early Development, Ba
5Department of Radiation Physics, Skåne University H
6Skåne University Hospital, Department of Clinical P
7Wallenberg Centre for Molecular Medicine, Lund Un
8Memory Clinic, Skåne University Hospital, Lund, Sw

Objectives: To examine the diagnostic performance o
discriminating AD from other neurodegenerative diso

Methods: We included 613 subjects with [18F]RO948
including 257 cognitively unimpaired (CU) controls, 1
102 with various non-AD disorders. [18F]RO948 PET
svPPA patients underwent an additional [18F]flortauci
FreeSurfer-based ROIs were created: Braak I-II (entor
ROI) and V-VI (widespread neocortical areas). For th
comparison in svPPA, subject-specific ROIs were dra
including subcortical WM). The diagnostic performan
(hippocampal volume, temporal meta-ROI and whole
Aβ42/P-tau) biomarkers. Area-under-the-curve (AUC)
analyses were compared using bootstrap (n=1000) pro

Results: [18F]RO948 SUVR within Braak-stages were
diagnostic groups (Figure 1). [18F]RO948 showed hig
and non-AD subjects, with the highest AUC seen whe
non-AD disorders), which outperformed MRI (highes
AD disorders, temporal thickness, 0.798) and CSF (hi
disorders, Aβ42/Aβ40, 0.927) (Figure 2). Generally, [18
Aβ+ cases/MAPT R406W mutation-carriers. [18F]RO9
showed lower temporal lobe uptake compared to [18F]

Conclusions: [18F]RO948 Tau-PET positivity was on
was lower compared to [18F]flortaucipir in svPPA (pre
This supports that [18F]RO948 has high specificity for
diagnostic marker in the differential diagnosis of AD.

41

F]RO948 tau positron emission
AD from other neurodegenerative

pele1,3, Alexander Santillo1, Edilio Borroni4,
ebastian Palmqvist1,2, Niklas Mattsson1,2,7, Olof

ical Sciences, Lund University, Malmö, Sweden
l, Lund, Sweden
us Amsterdam, Amsterdam, Sweden
asel, Switzerland
Hospital, Lund, Sweden
Physiology and Nuclear Medicine, Lund, Sweden
niversity, Lund, Sweden
weden

of the novel tau-PET tracer [18F]RO948 in
orders.

8 PET from the Swedish BioFINDER-2 study,
154 patients with MCI, 100 with AD dementia, and
T was performed 70-90 min post-injection. Three
ipir scan (80-100 min post-injection). For tau-PET,
rhinal), III-IV (temporal/limbic), I-IV (temporal meta-
he head-to-head [18F]RO948/[18F]flortaucipir
awn (temporal voxels showing elevated SUVR values,
nce of [18F]RO948 SUVR was compared against MRI
e-brain cortical-thickness) and CSF (Aβ42/Aβ40,
) values from receiver operating characteristic
ocedures.

e higher in AD dementia compared to other
gh accuracy for distinguishing AD dementia from CU
en using Braak I-IV (0.977, versus CU, and 0.972
st AUC versus CU, whole-brain thickness, 0.910; non-
ighest AUC versus CU, Aβ42/P-tau, 0.931; non-AD
8F]RO948 cortical positivity was only observed in
948 retention was not increased in svPPA, and
]flortaucipir.

nly found among Aβ-positive cases and the retention
edominantly caused by TDP-43 type-C pathology).
r AD-type tau and highlight its potential as a
.

15

41

16

Keywords: Tau, PET, RO948, Flortaucipir, Alzheimer
41

r
17

Leuzy, Antoine

P146: Longitudinal changes in tau pat
PET are associated with elevated CSF
Swedish BioFINDER-2 study

Antoine Leuzy1, Gregory Klein2, Rik Ossenkop
Sebastian Palmqvist1,4, Olof Strandberg1, Precio
Ruben Smith1,4, Oskar Hansson1,6

1Clinical Memory Research Unit, Department of Clini
2Roche Pharma Research and Early Development, Ba
3VU University Medical Center, Neuroscience Campu
4Department of Neurology, Skåne University Hospital
5Wallenberg Centre for Molecular Medicine, Lund, Sw
6Memory Clinic, Skåne University Hospital, Lund, Sw

Objectives: 1) To describe the longitudinal change in
using [18F]RO948-PET in a cohort of cognitively unim
the Swedish BioFINDER-2 study; 2) To examine the
[18F]RO948 SUVR and CSF phosphorylated-tau (P-ta

Methods: 51 subjects from the ongoing BioFINDER-
Aβ-,14 Aβ+), 21 MCI (3 Aβ-,18 Aβ+) and 7 AD dem
Aβ42/Aβ40. All participants underwent [18F]RO948-PE
and at a targeted follow-up of approximately 12-mont
to calculate mean-SUVR values (inferior cerebellar gr
entorhinal/hippocampus; III/IV, temporal/limbic; V/V
annualized percentage-change and effect-size (Cohen’
assay (INNOTEST, Fujirebio). The relationship betwe
examined ROI-wise using linear-regression, adjusting
interscan-interval.

Results: The largest longitudinal change and effect-si
Braak I/II region (+3.14%, effect-size 0.72). For Aβ+M
and effect-sizes were seen in the Braak III/IV (+3.62%
size 0.58), respectively. No significant increases were
stages, -0.65%; CU and MCI combined due small sam
with increases in [18F]RO948 SUVR across Braak-sta

Conclusions: Initial longitudinal results with the nove
capture the progression of early tau pathology in Aβ+
with cognitive impairment. Importantly, increases in [
baseline P-tau, supporting its use as a biomarker predi

41

thology measured by [18F]RO948 tau-
F P-tau: Preliminary findings from the

ppele1,3, Niklas Mattsson1,4,5, Shorena Janelidze1,
osa Coloma2, Edilio Borroni2, Erik Stomrud1,6,

ical Sciences, Lund University, Malmö, Sweden
asel, Switzerland
us Amsterdam, Amsterdam, The Netherlands
l, Lund, Sweden
weden
weden

n tau pathology seen after approximately one-year
mpared (CU), MCI, and AD dementia subjects from
association between longitudinal increases in
au).

-2 longitudinal sub-study were included: 23 CU (9
mentia (all Aβ+), with Aβ-status determined using CSF
ET (list-mode, 70-90 min post-injection) at baseline

ths (mean=13.2;SD=2.6). FreeSurfer (v.6.0) was used
rey reference) across Braak-stages (I/II,
VI, neocortical). Longitudinal measures included
’s d). CSF P-tau was measured using a commercial
een P-tau and change in [18F]RO948 SUVR was
g for age, baseline-SUVR, diagnostic-group and

ize (Figure 1) for Aβ+CU subjects was seen in the
MCI and AD dementia subjects, the largest changes
%, effect-size 0.61) and V/VI regions (+3.98%, effect-
e seen across Aβ- subjects (average-change across
mple-size). CSF P-tau was significantly associated
ages (Figure 2).

el tau-PET tracer [18F]RO948 indicate that it is able to
+CU subjects, as well as cortical increases in subjects
[18F]RO948 SUVR were significantly associated to
ictive of tau PET.

18

Keywords: Tau, PET, RO948, CSF, P-tau
41

19

Lin, Kun-Ju

P147: A comparison of ischemic strok
(18F-PMPBB3) and 18F-THK-5351 up

Kun-Ju Lin1,3, Ing-Tsung Hsiao1,3, Kuo-Lun Hu

1Department of Nuclear Medicine and Molecular Ima
Taoyuan, Taiwan
2Department of Neurology, Chang Gung Memorial Ho
Medicine, Taoyuan, Taiwan
3Department of Medical Imaging and Radiological Sc
Taiwan

Background: Post-stroke cognitive impairment (PSC
however, the mechanism underlying it remains unkno
1607 (18F-PM-PBB3) and 18F-THK-5351, are availabl
in patients with tauopathy. However, alternative targe
explore differences, we compared the 18F-APN-1607 a
with chronic ischemic stroke.

Method: Three patients with PSCI were recruited for
The cerebellum cortex was used as the reference regio
calculation. Another 12 healthy controls were recruite
extent of 18F-APN-1607 uptake in ischemic stroke pat
Results: 18F-APN-1607 and 18F-THK-5351 were perf
and one patient with MCA infarction. Increased 18F-T
in the areas around ischemic lesions but also in basal g
off-target binding in basal ganglia, thalamus and midb
Conversely, 18F-APN-1607 uptake was not significant
the background values of healthy controls.
Conclusion: Different binding patterns between 18F-A
suggest that the increased uptake of 18F-THK-5351 in
binding such as MAO-B rather than tau. However, the
Longitudinal observation, e.g., by pathological and cli

Keywords: Post-stroke cognitive impairment, Tau, Am

42

ke-induced changes on 18F-APN-1607
ptake patterns

uang2, Chin-Chang Huang2

aging Center, Linkou Chang Gung Memorial Hospital,
ospital Linkou Medical Center and College of
ciences and Healthy Aging Research Center, Taoyuan,

CI) may affect up to one third of stroke survivors,
own. Several novel radiotracers, such as 18F-APN-
le for visualization of in vivo tau protein distribution
eting to MAO-B was reported for 18F-THK-5351. To
and 18F-THK-5351 imaging presentations in patients

18F-APN-1607 and 18F-THK-5351 imaging studies.
on for standardized uptake value ratio (SUVR)
ed for 18F-APN-1607 PET studies to determine the
tients.
formed in two patients with PCA territorial infarction
THK-5351 uptake was not only observed prominently
ganglia, thalamus and midbrain. On the other hand,
brain was not observed on 18F-APN-1607 imaging.
tly increased in the peri-lesional areas as compared to

APN-1607 and 18F-THK-5351 on PET scans may
n the peri-lesional area may result from off target
e pathogenesis is complex and changes with time.
inical correlation is warranted.
myloid, 18F-APN-1607, 18F-THK-5351

20

Manser, Paul

P148: Longitudinal change in [18F]GT
on baseline SUVR intensity and spatia
cognitive decline

Paul Manser1, Sandra Sanabria Bohorquez1, Edm
Marik1, Robby Weimer1

1Genentech, South San Francisco, CA, US
Objective: To characterize the longitudinal change in
relationship with cognitive decline and baseline [18F]G
Methods: [18F]GTP1 PET scans were performed at ba
and positive cognitive normal subjects (CN; n=2 and 7
(Prod; n=24; MMSE 24-30, CDR = 0.5) and mild/mod
subjects. Measurements were made within whole cort
in vivo Braak ROIs (Schöll et al., Neuron 2016). Rela
cognitive scores were summarized by timepoint with S
change for each subject were calculated using simple
SUVR using Spearman correlations.
Results: Local baseline [18F]GTP1 was prognostic of
subjects: baseline SUVR in Braak 1&2 had a negative
0.23, p=0.069; rs= -0.23, p=0.069), while later Braak R
p=0.013; rs= 0.43, p=<0.001, rs= 0.42, p=0.001). Cha
correlations with change in ADAS-Cog13 and CDR-S
significance (rs= 0.14, p=0.323; rs= 0.18, p=0.210).
Conclusions: We observed that baseline [18F]GTP1 S
an in vivo Braak ROI-dependent manner. We observe
[18F]GTP1 SUVR and CDR-SB and ADAS-Cog13. T
and its use in AD clinical trials.

42

TP1 SUVR over 18 months depends
al distribution and shows trends with

mond Teng1, Suzanne Baker1, Balazs Toth1, Jan

n tau burden measured by [18F]GTP1 and its
GTP1 signal.
aseline 6, 12 and 18 months in amyloid PET negative
7, respectively), and amyloid PET positive prodromal
derate (MM; n=30; MMSE 22-30, CDR = 0.5-2) AD
tical gray matter (WCG), at the regional level, within
ationships between change in SUVR and change in
Spearman correlations. Annualized rates of SUVR
linear regression and compared against baseline

annualized rate of local SUVR change in AD
e correlation with its annualized SUVR change (rs= -
ROIs (4, 5, and 6) had positive correlations (rs= 0.31,
ange in [18F]GTP1 WCG SUVR showed weak positive
SB at 18 months that did not meet statistical

SUVR was prognostic of 18 month SUVR change in
ed weak positive correlations between changes in
These results support the clinical utility of [18F]GTP1

21

Keywords: [18F]GTP1, tau PET, cognitive decline
42

22

Margolin, Richard

P149: 18F-APN-1607: a promising PE

Richard Margolin1,2, Kun-ju Lin1,3,4, Paul Temp
Russell5, Christine Sandiego5, Chin-Chang Hua
David Alagille7, Yihui Guan8, Jiaying Lu8, Chua

1APRINOIA Therapeutics, Taipei, Taiwan
2CNS Research Solutions LLC, Cambridge, MA, US
3Department of Nuclear Medicine and Molecular Med
Linkou, Taoyuan City, Taiwan
4Department of Medical Imaging and Radiological Sc
of Medicine, Chang-Gung University, Taoyuan City, T
5Invicro, LLC, New Haven, CT, US
6Department of Neurology, Chang Gung Memorial Ho
7Xingimaging LLC, Beijing, China
8Huashan Hospital/Fudan University, Shanghai, Chin
9Quantum Sciences and Technology Institute, Chiba, J

Background: Tau aggregation/deposition is a key fea
including PSP, CBD, and MAPT mutation-induced FT
predominant (PiD), 4R-predominant (PSP, CBD), and
characterization in-vivo; one PET tracer for all tauopa
useful. While several tracers perform well in AD, the
unclear. An initial PBB-class tracer, [11]C-PBB3, de
radiopharmaceutical properties. A second-generation
lacked off-target binding on in-vitro screen; competiti
model (rTg4510) signal increased over time, mirroring
uptake; initial human scans supported further evaluati

Methods: APN-1607 has been evaluated in investigat
studies and 2 Phase 1 studies investigating dosimetry,
SUVR analyses were performed, referencing cerebella

Results: APN-1607 has favorable dosimetry (6.4 mSv
between 60-150 min post-injection in HC and AD (R2
consistently <5%. Approximately 265 individuals hav
including 75 HC, 40 AD, 45 PSP, 17 MAPT FTD, and
plexus uptake unrelated to known biochemical/clinica
HC; aMCI/AD patterns match known tau pathology d
correlated. In PSP, retention is observed in relevant s
the latter (Figure). Importantly, signal intensity correl
unreported finding. Uptake in a CBS case correlated

42

ET tracer for multiple tauopathies

pest1, Poe-Jou Chen1, Kenneth Marek5, David
ang4,6, Ing-Tsung Hsiao3,4, Gilles Tamagnan7,
antao Zuo8, Makoto Higuchi9, Ming-Kuei Jang1

dicine Center, Chang Gung Memorial Hospital

ciences, and Healthy Aging Research Center, College
Taiwan

ospital Linkou, Taoyuan City, Taiwan

na
Japan

ature of AD and non-AD tauopathies (NADTs),
TD. Tauopathies differ by repeat type: 3R-
d mixed (AD). PET tracers are enabling tau burden
athies could illuminate pathology and be clinically
eir 4R-predominant tauopathy performance is
emonstrated appropriate signal but had problematic
n PBB, [18F]APN-1607 (originally [18]F-PM-PBB3),
ion showed no MAO A/B binding. In a 4R mouse
g histopathology. NHP imaging revealed good brain
ion.

tor-initiated studies, 2 sponsored proof-of-concept
, kinetics, and test/retest. Dynamic DVR and static
ar cortex.

v/185 mBq), excellent SUVR/DVR agreement
2 >0.95), and test/retest DVR/SUVR %CVs
ve been scanned in Taiwan, Japan, US, and China,
d 9 CBS. No meningeal binding was noted; choroid
al factors is sometimes seen. Little uptake occurs in
distribution, with uptake and clinical severity generally
subcortical regions and brainstem, often prominent in
lates with clinical severity (PSPRS), a previously
with biopsy-confirmed tau.

23

Conclusion: APN-1607 merits clinical development;
results. Further study in diverse tauopathies/clinical s
Keywords: APN-1607, PET, PSP, tau, tauopathy

42

preliminary NADT findings extend 4R-tau model
settings is planned.

24

Mecca, Adam

P150: Entorhinal cortical tau accumul
hippocampal synaptic density in older
and early Alzheimer’s disease

Adam Mecca1, Ming-Kai Chen1, Mika Naganaw
Harris1, Hugh Bartlett1, Wenzhen Zhao1, Jean-D
Huang1, Amy Arnsten1, Richard Carson1, Christ

1Yale University School of Medicine, New Haven, CT,

Background: Synaptic loss is the major structural cor
disease (AD). Using [11C]UCB-J–PET we have shown
specific binding as a marker of synaptic density in par
degeneration of entorhinal cortical (ERC) cells that pr
performed PET imaging with [18F]flortaucipir and [11C
ERC would be inversely associated with synaptic den
Methods: [11C]UCB-J binding to SV2A and [18F]flor
10 cognitively normal (CN) participants. [11C]UCB-J
using SRTM2 and a whole cerebellum reference regio
(SUVRs) were calculated using an inferior cerebellum

Results: AD participants (68.8±6.6 years, CDR=0.5-1
([11C]PiB) PET and spanned the disease stages from a
mild dementia (n=5). CN participants (72.1±7.9 years
found to be Aβ– by [11C]PiB PET, but 2 were Aβ+. A
significant reductions in hippocampal [11C]UCB-J bin
p=0.002) and significant increases in ERC [18F]flortau
SUVR=1.68±0.16, p<0.00001). In the overall sample,
synaptic density (r=-0.63, p=0.005, Figure 1).

Conclusions: Our results revealed higher ERC tau an
compared to CN participants. Consistent with our hyp
was associated with lower hippocampal synaptic dens
failure due to tau pathology in ERC neurons projectin
elucidate the relationship between tau accumulation a

42

lation is inversely associated with
r individuals with normal cognition

wa1, Takuya Toyonaga1, Tyler Godek1, Joanna
Dominique Gallezot1, Nabeel Nabulsi1, Yiyun
topher van Dyck1

, US
rrelate of cognitive impairment in Alzheimer’s
n significant reductions in hippocampal SV2A
rticipants with AD, consistent with the early
roject to hippocampus via the perforant path. We
C]UCB-J and hypothesized that tau deposition in
nsity in hippocampus.
rtaucipir binding to tau were measured in 10 AD and
distribution volume ratios (DVRs) were calculated
on. [18F]Flortaucipir standardized uptake value ratios
m reference region.
1.0) were all Aβ+ by [11C]Pittsburgh Compound B
amnestic Mild Cognitive Impairment (aMCI, n=5) to
s) were free of clinical symptoms (CDR=0); 8 were
AD compared to CN participants demonstrated
nding (AD DVR=0.88±0.13, CN DVR=1.06±0.08,
ucipir binding (CN SUVR=1.11±0.12, AD
, ERC-tau was inversely associated with hippocampal

nd lower hippocampal synaptic density in AD
pothesis, in the overall sample ERC tau accumulation
sity. This inverse association may reflect synaptic
ng to the hippocampus. Further studies are needed to
and synaptic loss in AD.

25

Keywords: syanptic imaging, [11C]UCB-J, Alzheimer
42