The Early Diagnosis of Alzheimer’s disease

Imaging and CSF

Biomarkers

May 9, 2010

Nice France

Mony de Leon Professor of Psychiatry

NYU School of Medicine

Alzheimer’s disease

Familial < 5% Early Onset < 60 years

Inherited abnormalities:

presenilin 1 (PS1)- ch14

presenilin 2 (PS2)- ch1

amyloid precursor protein (APP)- ch21

Sporadic 95%

Late-Onset > 65 years

Age

High BP

High cholesterol

Diabetes

Apolipoprotein-E e4

Head Trauma

Risk factors

DIAGNOSTIC CRITERIA FOR AD

Clinical Diagnosis a. Deficit in two or more areas of cognition.

b. Progressive and gradual impairment.

c. Interference with ADL.

d. Rule out other brain or systemic disease.

Neuroimaging Diagnosis- Not available

Biomarker Diagnosis- Not available

DIAGNOSTIC CRITERIA FOR MCI

Clinical Diagnosis

a. Complaints corroborated by informant. b. Not demented and no interference with ADL.

c. Below normal performance on neuropsychological testing

d. Rule out other brain or systemic disease.

Biomarker Objectives

• Early detection (sensitivity)

• Valid for AD pathology (specificity)

• Sensitive to disease progression

• Probe disease mechanisms

MRI structure and FDG-PET metabolism

Oxidative stress and inflammation

Cellular mechanism

Primary etiology

Aβ and Tau Sensitive

Diagnostic

Targets

MCI

AD

Preclinical

Age

Objective of Early Diagnosis

Time

Bra

in H

ea

lth

1976-1980 1981-1988

Ventricle Enlargement in AD

de Leon et al Lancet 1979

Cortical Atrophy in AD

1976-1980 1981-1988

de Leon et al Lancet 1979

CM Plane

1984

A P

Canthomeatal Plane Negative Angle

MRI Negative Angle- Hippocampal Atrophy

391-2

de Leon MJ, et al. Lancet. 1988

.

Perc

en

tag

e C

ases

Age

AD (n = 203)

MCI (n = 72)

NL (n = 130)

0 1 2 3

Coronal

Axial

Hippocampal Fissures

Bobinski, de Leon, Wegiel and Wisniewski 2000

MCI

AD

Time

Bra

in H

ea

lth

Preclinical

Age

Objective of Early Diagnosis

de Leon MJ, et al. Lancet. 1989;2:672-673.

de Leon MJ, et al. AJNR Am J Neuroradiol. 1993;14:897-906.

MRI Hippocampal Atrophy 4-year advance prediction of MCI

to AD

Accuracy

SS/SP (n=23,9)

90%

91/89%

MCI

AD

Preclinical

Age

Objective of Early Diagnosis

Time

Bra

in H

ea

lth

Automated MRI Regional Boundary

Shift

6-year prediction of Normal to MCI

Prediction of MCI

Accuracy

SS/SP (n=13,32)

89%

91/85% Rusinek H, de Leon, MJ et al. Radiology. 2003;229:691-696.

MESH MODELING

Thompson et al, Neuroimage 2004

3-year MRI hippocampal atrophy rate

superior inferior

Ap

os

tolo

va

, M

osc

on

i, d

e L

eo

n e

t a

l, N

eu

ro B

iol A

gin

g 2

00

9

NL-NL

n=10

NL-decline

n=7

ANATOMY

Annual tissue loss %

0

0.2

0.4

0.6

0.8

1

0 20 40 60 80 100 120

Cyclotron Radiotracer PET camera

0.01

0.1

1

0 20 40 60 80 100 120

THE PET SCAN

Input function

1. Blood

2. Simplified

Tissue Uptake Image

Acquisition

Modeling

Synthesis

Analysis

18F FDG 110 min 11C-PIB 20 min

Labeling

Simplification

MCI

AD

Preclinical

Age

Objective of Early Diagnosis

Time

Bra

in H

ea

lth

Ferris, de Leon, Wolf et al Neuro Biol Aging 1980

Normal AD

FDG-PET

Temporo-parietal and posterior cingulate

• R. Frackowiak 1981 15O2 PET

• R. Friedland 1983 18FDG-PET

• S. Minoshima 1997 18FDG-PET

Minoshima et al Ann Neurol 1997

FDG-PET

NYU-BNL

MCI

AD

Preclinical

Age

Objective of Early Diagnosis

Time

Bra

in H

ea

lth

PET Prediction of Decline from MCI

Minoshima Ann of Neurol 1997

Herholz Dem Geri Cog Dis. 1999

Drzezga et al EJNM 2003

Chetelat et al Neurology 2003

Mosconi et al Neurology 2004

MCI

AD

Time

Bra

in H

ea

lth

Preclinical

Age

Objective of Early Diagnosis

NL MCI AD 1993 1997 2003

E E E

Atrophy and Metabolism 10-Year decline from NL to AD

x H x x H H 4 V

4

4

V V

de Leon et al PNAS 2001

Entorhinal Cortex Glucose Metabolism

4-year prediction of Normal to MCI

Accuracy

SS/SP (n=12,13)

84%

83/85% de Leon MJ, et al. PNAS, U S A. 2001;98:10966-10971.

Automated Hippocampus Sampling 7-year advanced prediction of Normal to MCI and AD

Prediction of MCI

Accuracy

SS/SP (n=25,47)

76%

82/79% Mosconi L, de Leon MJ et al. Neurobiol Aging. 2008;29:676-692.

26%

(47)

(19)

4% /year

X

X

Estimated Earliest

Detection -11 yrs

X

X

X

LONGITUDINAL

PREDICTION

74%

BASELINE

PREDICTION

76%

2% /year

1% /year

(6) X -- NL-AD (6)

Years from Baseline

-- NL-NL (47)

-- NL-MCI (19)

0 5 -5

FDG-PET for Prediction and Progression: NL Aging

10 -10

NL-AD (1993-2003) Post-mortem validated AD

NL

MCI

Mild AD

Mod AD

2003

CERAD definite, Braak stage C-VI

4

0

Z score

Path: AD

Mosconi, de Leon, et al Eur J Nuc Med 2009

Typical DB (n=55, top) and Robust DB (n=55, bottom)

Z

AD n=33 MCI n=37

Eff

ect

Siz

e

Eff

ect

Siz

e

Typical DB Robust DB

Typical vs. Robust Normal Reference Databases

Mosconi, DeSanti, de Leon et al, J Nucl Med 2007

MRI and FDG-PET Imaging

Aβ and Tau

Oxidative stress and inflammation

Cellular response

Primary etiology

Pathology

Specific

Targets

Secretase Cleavage of APP

Senile or Amyloid Plaque

Tangles drawn by Bonfiglio 1908

From Robert Terry

Buerger K, et al. Arch Neurol. 2002;59:1267-1272.

AD vs Other

SS = 93, SP = 67

AD vs FTD

SS = 93, SP = 85

P-tau231 Diagnostic Specificity for AD

AD

82

FTD

26

VD

20

LBD

17

OND

26

HC

21

–20

0

20

40

60

80

100

120

140

160

CS

F P

-tau

231,

µL

CS

F E

qu

ivale

nts

Groups

Glodzik-Sobanska L, de Leon, MJ. et al. Neurobiol Aging. 2007

CSF Analytes in Normal Aging n=78

Z-S

co

re

Age

T-tau

P-tau231

IP

Aβ42/40

MCI

AD

Preclinical

Age

Objective of Early Diagnosis

Time

Bra

in H

ea

lth

2-Year Prediction of MCI to AD

CSF analyte Accuracy

T-tau 83*

P-tau231 83*

A 42/40 69*

Brys M, de Leon, MJ. et al. Neurobiol Aging. 2007.

*p < .01

MCI-MCI (n = 43) and MCI-AD (n = 22)

2-Year Prediction of MCI to AD

CSF analyte Accuracy

T-tau 86*

P-tau 84*

A42 82*

Mattson, N, de Leon, MJ, Blennow, K, et al. JAMA. 2009.

*p < .01

MCI-MCI (n = 258) and MCI-AD (n = 271)

MCI

AD

Time

Bra

in H

ea

lth

Preclinical

Age

Objective of Early Diagnosis

CSF Biomarkers Predict CDR > 0 Biomarker Hazard Ratio (95%

CI)

P Value

Aβ40

1.00 (1.00-1.00) .92

Aβ42 0.99 (0.99-1.00) .29

Aβ40/Aβ42 1.06 (0.99-1.14) .10

Tau 1.00 (1.00-1.01) .09

Ptau181 1.02 (0.99-1.04) .21

T-tau/Aβ42 ratio 2.42 (1.15-5.08) .02

P-tau181/Aβ42 ratio 1.78 (1.00-3.16) .05

Fagan AM, et al. Arch Neurol. 2007;64:343-349

Longitudinal memory by high and low P-tau231 in 57 NL

P

Longitudinal MTL-GM by high and low P-tau231 in 57 NL

Glo

dzik

, d

e L

eo

n e

t al N

eu

rob

iol.

Ag

ing

20

10

P

2-Year Progression Effects in MCI

60

50

40

30

20

10

P-t

au

231 [

pg

/mL

] ±

SE

M

NL-NL MCI-MCI MCI-AD Diagnostic group

900

700

500

400

300

200

T-t

au

[p

g/m

L] ±

SE

M

NL-NL MCI-MCI MCI-AD Diagnostic group

800

600

70

50

40

30

NL-NL MCI-MCI MCI-AD Diagnostic group

60 IP

[p

g/m

L] ±

SE

M

0.14

0.12

0.10

0.09

0.08

0.07

NL-NL MCI-MCI MCI-AD Diagnostic group

0.13

0.11

Aβ

-42/Aβ

-40 ±

SE

M

P-tau231 T-tau

A42/40 IsoP

Baseline Follow-up Longitudinal effect p

Annual Isoprostane Changes NL n = 11, MCI n = 6

de Leon MJ, and Pratico, D. et al. J Neurol. 2007;254:1666-1675

Sensitivity 83 100 100

Specificity 91 91 82

Overall 88* 94* 88*

Subjects

= MCI - AD

NL MCI NL MCI NL MCI 8,1

2-i

so

-iP

F2

-V

I (p

g/m

L) Baseline Year 1 Year 2

20

60

80

40 37

= NL

= MCI

*P < .05

= NL - MCI

PIB

FDG

AD

Li ,

Rin

ne,

de L

eon

Eur.

J.N

uc.M

ed. 2008

NL

11C-PIB DVR Images

Rowe et al Neurology 2007

MCI

AD

Time

Bra

in H

ea

lth

Preclinical

Age

Objective of Early Diagnosis

AD vs NL MCI vs NL

FDG Accuracy (%)

92 85

PIB Accuracy (%)

96 75

%

Agreement 94 54 ?

Dx Accuracy for FDG and PIB

Li , Rinne, de Leon Eur.J.Nuc.Med. 2008

PIB

FDG

Follow-up Baseline

2-year Amyloid Deposition and CMRglc in AD

Engler, Nordberg et al Brain 2006

?

?

MRI structure and FDG-PET metabolism

Oxidative stress and inflammation

Cellular dysfunction

Primary etiology

Aβ and Tau Progression

Mechanism

Targets

?

Lisa Mosconi The maternal history of AD

• Among AD pts, affected mothers to fathers is 3:1

• 25% of AD associated with maternal inheritance

• Maternal AD - lower age of onset

The Maternal Transmission of AD?

Maternal

Paternal

Maternal and Paternal

Neither

Other

-

Parental History of AD in NL n=733

No Hx

59%

Normal cognition, MMSE 28-30 Age 48-80 yrs Education>11yrs

Paternal 10%

Maternal 27%

M+P

4%

FDG-PET in individual cases

umol/100g/min

60

30

0

FH- FHp FHm

Mosconi, de Leon et al, Proc Natl Acad Sci USA 2007

Reduced CMRglc in FHm re: FH- Reduced CMRglc in FHm re: FHp

P

-10 Z

Longitudinal change in NL with AD mothers

FH- (n=37)

FHp (n=9)

FHm (n=20)

+10 Z

30

35

40

45

50

55

baseline 2-yrs

CM

Rglc

-1%/yr -0.5%/yr

-5%/yr

30

35

40

45

50

55

baseline 2-yrs

CM

Rglc

-0.5%

-6.5%

-1%

Mosconi et al, Neurology 2008

CSF Biomarkers and AD Maternal History in NL

Ab42/40 IsoP

P-tau231 T-tau

FH- FHp FHm FH- FHp FHm 22 14 23 22 14 23

*

*

FH- FHp FHm

FH- FHp FHm

Parietal PIB-PET Uptake in NL with Maternal History of AD

Mosconi, de Leon et al PNAS et al 2010

Mosconi, de Leon et al. Unpub. 2010

CMRglc Decreased and PIB Increased in FHm

(ADNI MCI data)

FDG-PET

PIB-PET

Lidia Glodzik Henry Rusinek Vascular pathology in AD

15H2O-PET EPI-ASL True-FISP-ASL

PERFUSION IMAGING PET AND MRI

MRI with rebreathing apparatus

Vasoreactivity to CO2 challenge ~2% CBF change/mmHg change in CO2

*

Plasma HIPPOCAMPUS

A40 VR CO2

* *

NL MCI AD NL MCI AD

Plasma A40 and NL Aging

r =.44, n=40, p

Plasma A40 and VR-CO2 in Elderly NL

rho = -.61, n=11, p

Conclusions • PET and MR imaged hippocampal pathology

appears early and is progression sensitive.

• The pathology specific P-tau231 and Ab also

appear early but are not progression sensitive.

• Imaging and biomarkers may have a role in

identifying mechanisms of AD-progression by

considering energetics and vascular function.

Collaborating Scientists

• NYU

– Lidia Glodzik

– Lisa Mosconi

– Henry Rusinek

– Susan De Santi

– Wai Tsui

– Elizabeth Pirraglia

– Yi Li

• Institute Basic Research

– Pankaj Mehta

• Temple University

– Domenico Pratico

• Einstein College of Medicine

– Peter Davies

• University of California L.A.

– Liana Apostolova

– Paul Thompson

• Applied Neurosolutions

– Ray Zinkowski

• Turku U. Finland

– Juha Rinne

• Sahlgrenska U. Goteborg

– Kaj Blennow

– Henryk Zetterberg

– Anders Wallin

• Cornell University - Shankar Vallabhajosula

- Stanley Goldsmith