HAI Book 2020 - Dec 18B

III

uary 17, 2020

CHAIRS:
Elizabeth Mormino, PhD, Stanford University
Susan Landau, PhD, University of California,
Berkeley

and Amelia Strom, BS, University of California
San Francisco

der Hyun-Sik Yang, MD, Brigham and Women's
Hospital

nal Isadora Lopes Alves, PhD, VUmc

01

Strom, Amelia

Glucose metabolism reflects local atro
symptomatic stages of Alzheimer's dis

Amelia Strom1, Leonardo Iaccarino1, Lauren Ed
Meigooni1, William Jagust2,3, Bruce Miller1, Gi

1University of California, San Francisco, San Francis
2University of California, Berkeley, Berkeley, CA, US
3Lawrence Berkeley National Laboratory, Berkeley, C

Background. We assessed the relative contributions o
PET hypometabolism in posterior cingulate cortex (PC
regions of hypometabolism in AD: local (amyloid and
connected regions, and deafferentation from the degen

Methods. Eighty-five patients with MCI or AD deme
positive by visual read) and FTP within one year (Tab
reference regions: FDG-SUVR30-60min (pons), PIB-SU
(inferior cerebellar cortex). PCC and IP were defined
or low connectivity to PCC/IP were defined based on
Imaging variables were transformed into age-adjusted
specific control groups (Table 1). Multiple regression
with FDG-SUVR.

Results. Local gray matter volume (GMV) and FTP-S
associated with FDG-SUVR in PCC (R2=.461, βGMV=
βFTP=-0.416, Fig1). Models including PIB/FTP in con
GMV/FTP alone (Fig1b). Results were similar when i
regions (Fig1c). MTL volume was weakly associated
analyses were repeated in subgroups split by median a
PCC-FDG was only seen in the older patient group (F

Conclusion. Hypometabolism reflected local atrophy
on PCC hypometabolism was seen primarily in older
might be region- and group-dependent. Our data did n
pathology in connected regions in the symptomatic sta

50

ophy and tau pathology at
sease

dwards1, Orit Lesman Segev1, David Soleimani-
il D. Rabinovici1,3, Renaud La Joie1

sco, CA, US
S
CA, US

of three potential mechanisms associated with FDG-
CC) and inferior parietal (IP) regions, two primary
d tau) pathology and atrophy, pathology in distant but
nerating medial temporal lobe (MTL).
entia underwent MRI and PET with FDG, PIB (all
ble 1). SUVR were calculated using tracer-specific
UVR50-70 (cerebellar cortex) and FTP-SUVR80-100
in native space using FreeSurfer. Regions with high
task-free fMRI data from neurosynth.org (Fig1a).
d z-scores using demographic-matched modality-
analyses were run to identify variables associated

SUVR, but not PIB-SUVR, were independently
=0.425, βFTP=-0.479) and IP (R2=.652, βGMV=0.476,
nnected regions did not outperform models with local
including measures of pathology in non-connected

with PCC-FDG but not IP-FDG (Fig2a). When
age (62 yo), the detrimental effect of MTL atrophy on
Fig2b).
and tau pathology. A distant effect of MTL atrophy
patients, suggesting that drivers of hypometabolism
not support hypotheses of a detrimental effect of
age of AD.

02

50

03

50

04

Keywords: metabolism, connectivity, Flortaucipir
50

05

Yang, Hyun-Sik

Plasma FLT1 predicts amyloid-β relat
normal older adults

Hyun-Sik Yang1, Becky Carlyle1,3, Bianca Trom
Pruzin1,2,3, Colleen Fitzpatrick1, Dylan Kirn1,2, D
Johnson1,2,3, Reisa Sperling1,2,3, Steven Arnold1,

1Massachusetts General Hospital, Boston, MA, US
2Brigham and Women’s Hospital, Boston, MA, US
3Harvard Medical School, Boston, MA, US

Background: Vascular endothelial growth factor (VE
Alzheimer’s disease (AD) pathophysiology through ce
However, in vivo relationships between plasma VEGF
trajectories remain unclear. Here, we examined plasm
Brain Study (HABS) and their relationship to individu
and systemic inflammatory state.

Methods: Data came from 298 cognitively normal (C
(Pittsburgh Compound B PET), baseline plasma colle
Cognitive Composite 5 (PACC5). First, we screened f
VEGF-C, VEGF-D, PlGF, and FLT1 (also known as s
Discovery V-PLEX platform, to identify plasma mark
Second, we tested the association of the identified pla
characteristics. Finally, using linear mixed effect mod
identified marker with Aβ in predicting longitudinal c

Results: Baseline plasma FLT1 predicted PACC5 dec
education; β=-0.025, nominal p=9.0×10-3, FDR=0.045
not (FDR>0.05). FLT1 was associated with age (β=0.
or baseline hippocampal volume. FLT1 was not assoc
Cardiovascular Risk, but was strongly associated with
systemic inflammation; Figure 1A; p=6.8×10-4) and b
(WMH; Figure 1B; p=0.013). FLT1 robustly interacte
even after adjusting for WMH and CRP (β=-0.12, p=3

Conclusion: Plasma FLT1 (soluble VEGFR1) is asso
vascular brain injury, and with Αβ-related cognitive d

50

ted cognitive decline in cognitively

mbetta1, Can Zhang1,3, Aaron Schultz1,3, Jeremy
Dorene Rentz1,2,3, Rudolph Tanzi1,3, Keith
,3, Jasmeer Chhatwal1,3

EGF) pathway dysregulation has been implicated in
erebrospinal fluid and post-mortem brain studies.
F pathway biomarkers and prospective cognitive
ma VEGF pathway measures in the Harvard Aging
ual cognitive trajectories, amyloid-β (Aβ) pathology,

CN) HABS participants with baseline Aβ measure
ection, and longitudinal Preclinical Alzheimer’s
five plasma VEGF pathway measures, VEGF-A,
soluble VEGFR1), measured on the Meso-Scale
kers that predict cognitive decline on the PACC5.
asma marker with baseline demographic and imaging
dels, we tested the interactive association of the
cognitive decline.
cline in CN older adults (adjusted for age, sex and
5), while VEGF-A, VEGF-C, VEGF-D, and PlGF did
.042, p=3.2×10-8), but not sex, APOE ε4, baseline Aβ,
ciated with baseline Framingham Heart Study
h baseline C-reactive Protein (CRP, marker of
baseline FLAIR white matter hyperintensity volume
ed with Αβ to predict PACC5 decline (Figure 1C),
3.7×10-3).
ociated with markers of systemic inflammation,
decline in CN older adults.

06

Keywords: plasma biomarker; amyloid-β; VEGF; FLT
50

T1; cognitive decline
07

Lopes Alves, Isadora

Reducing sample sizes to detect longit

Isadora Lopes Alves1, Lyduine Collij1, Fiona He
Adriaan Lammertsma1, Frederik Barkhof1,4, Jua
AMYPAD Consortium5

1Department of Radiology and Nuclear Medicine, Am
Netherlands
2BarcelonaBeta Brain Research Center, Barcelona, S
3Janssen Pharmaceutica NV, Beerse, Belgium
4Institute of Neurology and Healthcare Engineering, U
5This work has received support from the EU-EFPIA
(grant No 115952)., Brussels, Belgium

Background: Detecting amyloid-β plaque accumulati
understanding the earliest stages of Alzheimer’s disea
that would be needed to detect statistically significant
value ratios; SUVR) and quantitative (distribution vol
regional levels.

Methods: [11C]PIB scans of 239 cognitively normal (
underwent at least two PET scans, were included. Reg
cerebellar gray matter as reference region. PET positiv
the global SUVR (SUVR>1.16). Annual accumulation
sample sizes to detect a significant longitudinal chang
the sampsizepwr function in Matlab. The percentage o
group) was also computed.

Results: As expected, rates of change were higher in s
carriership increased accumulation rates in both negat
metric, orbitofrontal and precuneus displayed the high
rendered higher accumulation rates than DVR, but als
detect 0.22% early accumulation (mean of negative gr
N=917 for global SUVR (Figure 1). In addition, DVR
compared with 9% for SUVR. Finally, precuneus and
accumulators than global SUVR, i.e. 17 and 25%, resp

50

tudinal amyloid accumulation

eeman1, José Luís Molinuevo2, Mark Schmidt3,
an Domingo Gispert2, On behalf of the

msterdam UMC, location VUmc, Amsterdam, The
Spain
University College London, London, UK
Innovative Medicines Initiatives 2 Joint Undertaking

ion in initially amyloid negative subjects can help
ase (AD). This study aimed to determine sample sizes
t longitudinal changes using semi- (standard uptake
lume ratios; DVR) PET measures at global and

(CN) subjects in the OASIS-3 data-set, who
gional SUVR and DVR were extracted using
vity was defined by Gaussian Mixture Modelling of
n rates were calculated using linear regression, and
ge (alpha=0.05; 1-beta=0.80) were determined with
of accumulators (rates of change >2SD of negative

subjects with positive PET at baseline, and APOE-ε4
tive and positive groups (Table 1). Using either
hest rates of change in the negative group. SUVR
so displayed higher variability. As a consequence, to
roup), global DVR required N=584 compared with
R classified 13% of subjects as accumulators
d orbitofrontal regions classify more subjects as
pectively.

08

Conclusion: As compared to SUVR, DVR can reduce
amyloid changes by approximately 40% in amyloid ne
statistical power can be achieved by focusing on precu
Keywords: sample sizes, amyloid PET, accumulation,

50

e the sample size needed to detect longitudinal
egative populations. Additional improvements in
uneus and orbitofrontal regions.
, trial design

09

Friday, January 17, 2020 - 4:15pm - 5:15pm

SESSION 9: PLASMA

Friday, Janu

4:15 PM SESSION 9:
PLASMA

4:15 PM Plasma P-tau181 as a marker of tau patho
4:30 PM disease: relationship to Tau PET, differen
4:45 PM neuropathology and longitudinal progress

Plasma levels of an N-terminal tau fragm
associated with future cognitive decline a
neurodegeneration in clinically normal el
the Harvard Aging Brain Study

Plasma biomarkers associate with amyloi
binding in cognitively unimpaired older a
history of AD

5:00 PM High-performance plasma phospho-tau18
Alzheimer’s disease

51

uary 17, 2020

CHAIR
Gil Rabinovici, MD, UCSF

ology in Alzheimer’s Oskar Hansson, PhD, Lund
ntial diagnosis, University
sion

ment are highly Jasmeer Chhatwal,
and MD, MGH/Harvard
lderly: Findings from University

id and tau PET Pierre-Francois Meyer,
adults with a parental PhD, McGill University

81 biomarker for Tharick Pascoal, PhD, McGill
University

10

Hansson, Oskar

Plasma P-tau181 as a marker of tau p
relationship to Tau PET, differential d
longitudinal progression

Oskar Hansson1, Shorena Janelidze1, Niklas Ma
Thomas Beach2, Geidy Serrano2, Xiyun Chai3, N
Blennow4, Eric Reiman5, Jeffrey Dage3

1Clinical Memory Research Unit, Lund University, Lu
2Banner Sun Health Research Institute, Sun City, AZ,
3Eli Lilly and Company, Indianapolis, IN, US
4Department of Psychiatry and Neurochemistry, the S
Gothenburg, Gothenburg, Sweden
5Banner Alzheimer’s Institute, Phoenix, AZ, US

Plasma phosphorylated tau181 (P-tau181) might be in
if it can be used for differential diagnosis and prognos
a total of 589 individuals, including cognitively unimp
impairment (MCI), AD dementia, and non-AD neurod
in preclinical AD and further increased at the MCI and
and predicted positive Tau PET scans (AUC=0.87-0.9
differentiated AD dementia from non-AD neurodegen
and CSF P-tau181 (AUC=0.94-0.98) (Figure 2d-f), an
individuals with neuropathological confirmation. High
development of AD dementia in cognitively unimpair
SD unit increment) (Figure 3). In conclusion, plasma
biomarker of AD, which may be useful in clinical pra
be used to screen for individuals with high risk of AD
subsequently confirm presence of tau pathology.

51

pathology in Alzheimer’s disease:
diagnosis, neuropathology and

attsson1, Sebastian Palmqvist1, Ruben Smith1,
Nicholas Proctor3, Henrik Zetterberg4, Kaj

und, Sweden
US
Sahlgrenska Academy at the University of

ncreased in Alzheimer’s disease (AD), but it is unclear
sis. We studied plasma P-tau181 in three cohorts, with
paired participants and patients with mild cognitive
degenerative diseases. Plasma P-tau181 was increased
d dementia stages. It correlated with CSF P-tau181
91) (Figures 1 and 2a-c). Plasma P-tau181
nerative diseases with an accuracy similar to Tau PET
nd it could also detect AD neuropathology in
h plasma P-tau181 was associated with subsequent
red and MCI subjects with a hazard ratio of 3.8 (per
P-tau181 is a non-invasive diagnostic and prognostic
actice. Further, in clinical trials plasma P-tau181 might
D-like tau pathology where Tau PET could be used to

11

Figure 1
51

12

Figure 2

Figure 3
Keywords: Plasma P-tau, blood-based biomarker, Ta

51

au PET, early diagnosis, prognosis
13

Chhatwal, Jasmeer

Plasma levels of an N-terminal tau fra
future cognitive decline and neurodeg
Findings from the Harvard Aging Bra

Jasmeer Chhatwal1,2,3, Aaron Schultz1,3, Yifan D
Yang1,2,3, Keith Johnson1,2,3, Reisa Sperling1,2,3,

1Massachusetts General Hospital, Boston, MA, US
2Brigham & Women’s Hospital, Boston, MA, US
3Harvard Medical School, Boston, MA, US

Introduction: High sensitivity biofluid assays have a
neurodegenerative disease, including for Alzheimer's
that levels of an N-terminal fragment of the tau protei
differentiate clinically-normal people from those with
we asked whether plasma levels of NT1 in clinically-n
decline and neurodegeneration.

Methods: Baseline plasma samples from participants
assayed for NT1 and neurofilament light (NfL) using
Baseline NT1 and NfL were assessed alone or interac
PET; PiB-PET) as predictors of longitudinal cognitive
Composite; PACC) and neurodegeneration using linea
NT1 measured within 1 year of 18F-Flortaucipir PET
between NT1 and regional tau pathology.

Results: Greater plasma NT1 at baseline was highly p
longitudinal follow-up (median 5.04+/-0.99y; p<0.000
baseline β-amyloid burden (p<0.005). Greater NT1 w
gray matter and hippocampal volumes measured by lo
significant association between NT1 and FTP-PET sig
inferior temporal cortex. Baseline plasma NfL also int
(p=0.0327). However, NT1 remained a strong predicto
interaction with β-amyloid (NT1*PiB: p <0.0001; NfL

Discussion: Plasma NT1 levels may reflect ongoing n
processes. Plasma NT1 has the potential to be a usefu
AD research, particularly when combined with measu

51

agment are highly associated with
generation in clinically normal elderly:
ain Study

Dang2, Beth Ostaszewski2, Lei Liu2, Hyun-Sik
, Dennis Selkoe2,3

accelerated biomarker development for
disease (AD). Building on recent work demonstrating
in (NT1) in cerebrospinal fluid and plasma can
h mild cognitive impairment and dementia due to AD,
normal elderly are predictive of prospective cognitive

in the Harvard Aging Brain Study (n=236) were
single-molecule array technology (Quanterix Simoa).
ctively with β-amyloid (11C-Pittsburgh Compound-B
e decline (Preclinical Alzheimer's Cognitive
ar mixed effects models. In participants with plasma
(FTP-PET), we assessed cross-sectional associations

predictive of greater cognitive decline during
01; Figure 1). This effect was synergistic with
was also associated with greater decreases in cortical
ongitudinal structural MRI (Figure 2). We observed a
gnal in the entorhinal cortex (p=0.002) but not in the
teracted with β-amyloid to predict PACC decline
or of PACC decline after correcting for NfL and its
L*PiB: p=0.234).
neurodegenerative and tau-related pathobiologic
ul, non-invasive biomarker in clinical and translational
ures of β-amyloid.

14

51

15

Keywords: plasma, biomarker, tau, neurodegeneratio
51

on, amyloid
16

Meyer, Pierre-Francois

Plasma biomarkers associate with amy
cognitively unimpaired older adults w

Pierre-Francois Meyer1,2,4, Nicholas Ashton5, Th
Gonneaud1, Alexa Pichet Binette1,2,4, Hazal Ozl
Simrén5, Anne Labonté1, Pedro Rosa-Neto1,2,3,4,
Zetterberg5,6, Kaj Blennow5, Sylvia Villeneuve1

1Douglas Mental Health University Institute, Centre f
(StoP-AD), Montréal, QC, Canada
2Department of Psychiatry, McGill University, Montré
3Department of Neurology and Neurosurgery, McGill
4McGill Centre for Integrative Neuroscience, McGill
5Department of Psychiatry and Neurochemistry, Instit
Sahlgrenska Academy, University of Gothenburg, Got
6Department of Neurodegenerative Disease, UCL Que
London, London, UK

Objective: Plasma biomarkers may facilitate or accel
pathology in early AD pathogenesis. We therefore inv
and tau PET-binding in pre-symptomatic AD.
Methods: Aβ-[18F-NAV-4694] and tau-PET [18F-AV
unimpaired older individuals with a parental history o
Plasma Aβ (Aβ40 and Aβ42) biomarkers were measure
(SiMoA, Quanterix, Billerica, MA, USA) and immun
MS). Plasma tau (181P-tau and total-tau) biomarkers w
analyses investigated associations of plasma and PET
ROI for Aβ and temporal metaROI for tau) and voxel
plasma biomarker performance for detection of Aβ-PE

51

yloid and tau PET binding in
with a parental history of AD

homas Karikari5, Theresa Köbe1, Julie
len1, Anne Labonté1, Josef Pannee5, Joel
, John Breitner1,2,4, Judes Poirier1,2, Henrik

1,2,3,4

for Studies on the Prevention of Alzheimer’s Disease
éal, QC, Canada
l University, Montréal, QC, Canada
University, Montréal, QC, Canada
tute of Neuroscience and Physiology, The
thenburg, Sweden
een Square Institute of Neurology, University College

lerate the detection of amyloid-beta (Aβ) and tau
vestigated the relation of plasma biomarkers with Ab

V-1451] were obtained from 129 cognitively
of AD from the PREVENT-AD cohort (Table 1).
ed with with Single Molecule Array technology
noprecipitation coupled with mass spectrometry (IP-
were measured with SiMoA only. Linear regression
T measures using region of interest (global cortical
-wise approaches. ROC analyses further determined
ET positivity.

17

Results: Plasma Aβ42/Aβ40 ratio measured by either m
4694 SUVRs (R2 range=0.07-0.12 and R2=0.07, all P<
increased 18F-AV1451 binding (R2=0.11, P<0.001, Fi
plasma Aβ42/Aβ40 measures were associated with Ab-
and medial pre-frontal cortex, Figure 2D-E). Only pla
regions recognized for ‘early’ tau accumulation (Figu

51

method and 181P-tau were associated with 18F-NAV-
<0.001). Only plasma 181P-tau was associated with
igure 1A-C). Voxel-wise analyses indicated that both
-PET binding in ‘typical’ Ab-regions (e.g, precuneus
asma 181P-tau correlated with 18F-AV-1451 binding in
ure 1F).

18

51

19

In ROC analyses, combining Aβ42/Aβ40 and 181P-tau y
status (area under the curve range=0.808-0.836; Figur
Aβ42/Aβ40 were more sensitive than SiMoA to Aβ-PE

Conclusions: In cognitively unimpaired older adults a
the disease, plasma biomarkers capture PET evidence
pathology. Thus, plasma biomarkers show promise fo
stage of AD pathological progression.
Keywords: Plasma, PET, biomarkers, amyloid, tau

52

yielded highest accuracy to detect PET Aβ-positivity
re 2). In all analyses, mass spectrometry measures of
ET binding.

at risk for AD dementia owing to a family history of
e of established Aβ pathology and nascent tau
or the detection of AD pathology even at the earliest
20

Pascoal, Tharick

High-performance plasma phospho-ta
disease

Tharick Pascoal2, Thomas Karikari1, Nicholas A
Min Su Kang2, Joseph Therriault2, Michael Sch
Serge Gauthier2, Henrik Zetterberg1, Pedro Rosa

1University of Gothenburg, Gothenburg, Sweden
2McGill University, Montreal, Canada

Objective: To present the discovery and validation of
biomarker to depict brain neurofibrillary tangles accum

Methods: We developed and validated an automated
quantifying phospho-tau181in plasma and serum. Dis
performed in three independent cohorts in Sweden and
tau and amyloid-β (Aβ), cerebrospinal fluid, and both
structural brain imaging measurements.

Results: Serum and plasma p-tau181 accurately discr
discovery sample (AUCs≥90%) (Fig. 1A, B). In the v
tau181showed a progressive increase from young adu
negative, cognitively unimpaired (CU) elderly Aβ pos
(Fig. 1C). Plasma phospho-tau181discriminated AD d
frontotemporal dementia patients (AUC=100%), and y
phospho-tau181 was highly correlated (Fig. 2) and pre
pathologies measured by PET (Fig. 3). Plasma phosph
Aβ positive from Aβ negative (Fig. 3D). Remarkably,
1-year cognitive impairment and reduction in hippoca
cohort that consisted of individuals assessed in a clini
accurately discriminated AD from young adults (AUC

Conclusions: To conclude, we have presented a high-
a simple, cost-effective, and scalable test for the diagn
applications for diagnosis and recruitment for disease-
incorporated in clinical practice as a rapid screening te
and clinical management of dementia patients.

52

au181 biomarker for Alzheimer’s

Ashton1, Andréa Benedet2, Melissa Savard2,
höll1, Gassan Massarweh2, Jean-Paul Soucy2,
a-Neto2, Kaj Blennow1

f a novel high-performance plasma phospho-tau181
mulation in Alzheimer’s disease (AD).
ultrasensitive Single Molecule Array assay for
scovery and validation of the novel assay were
d Canada using positron emission tomography (PET)
h cross-sectional and longitudinal cognitive and

riminated AD from age-matched controls in the
validation sample (n=226), plasma phospho-
ults to cognitively unimpaired (CU) elderly Aβ
sitive, MCI Aβ positive, and AD dementia (P<0.0001)
dementia from CU elderly (AUC=98.24%),
young adults (AUC=99.44%) (Fig. 1D). Plasma
edicted brain tau (AUC=93%) and Aβ (AUC=88%)
ho-tau181 differentiated individuals tau PET negative
, plasma phospho-tau181 was highly associated with
ampal volume (P<0.0001). In the second validation
ical setting (n = 105), plasmaphospho-tau181
C=100%) and CU elderly (AUC=84.44%).
-performance blood p-tau181 assay that may represent
nosis of AD. This technology has immediate
-modifying trials. This assay has the potential to be
est to rule out AD pathophysiology and guide therapy

21

52

22

52

23

Keywords: p-tau, neurofibrillary tangle, blood biomar
52

rker, Alzheimer’s disease, amyloid
24

1