Using Imaging Markersto Study Drug Effects in AD

Giovanni B FrisoniDeputyDeputy ScientificScientific DirectorDirector

IRCCS Fatebenefratelli IRCCS Fatebenefratelli –– The National Center for The National Center for AlzheimerAlzheimer’’ss DiseaseDisease. Brescia. Bresciawww.centroAlzheimer.org

MRI, hippocampal atrophy, and cognition:markers for enrichment in clinical trials of MCI

Why enriching study groupsin clinical trials of diseasemodifiers in MCI?

How to and consequences: there’s no free lunch

Hippocampal volumetry asenrichment marker

Probable Alzheimer’s Disease(NINCDS-ADRDA)

Mild cognitive impairment

Am

yloi

d/ta

u de

posi

tion

No cognitive symptom

Age

Disease modification in AD

AD non ADpathology

AD non ADpathology

AD non ADpathology

Enriched MCI

Biomarkers in the IWG and NIA-AA diagnostic criteria

Fris

oni H

ampe

lO’B

rien

Ritc

hie

Win

blad

, Lan

cet N

euro

l201

1

�

1.0 1.5 2.0

70th 85th 95th 99th

99 th

95 th 8 5th 70 th

1000

2000

3000 4 000

5 000

6000

44 th �

PIB uptake

Hippocampal volume

MCI non converters

MCI converters to dementia

AllMCIs

Distribution of AD markers in MCIs

Lore

nzi e

t al.,

Neu

robi

ol A

ging

201

0

THE PAYOFF OF ENRICHMENT IN MCI CLINICAL TRIALS

Mild

ly a

bnor

mal

bio

mar

ker

Mod

erat

ely

abno

rmal

bio

mar

ker

Sev

erel

y ab

norm

al b

iom

arke

r

Ver

y se

vere

ly a

bnor

mal

bio

mar

ker

(Nat Reviews Neurol 2010)

40:60

� enrolled into clinical trial

� excluded from clinical trial

likely not to have AD

likely to have AD

THE PAYOFF OF ENRICHMENT IN MCI CLINICAL TRIALS

Mod

erat

ely

abno

rmal

bio

mar

ker

Sev

erel

y ab

norm

al b

iom

arke

r

Ver

y se

vere

ly a

bnor

mal

bio

mar

ker

(Nat Reviews Neurol 2010)

40:60

32:32

THE PAYOFF OF ENRICHMENT IN MCI CLINICAL TRIALS

Sev

erel

y ab

norm

al b

iom

arke

r

Ver

y se

vere

ly a

bnor

mal

bio

mar

ker

(Nat Reviews Neurol 2010)

40:60

32:32

25:20

THE PAYOFF OF ENRICHMENT IN MCI CLINICAL TRIALS

Ver

y se

vere

ly a

bnor

mal

bio

mar

ker

(Nat Reviews Neurol 2010)

40:60

32:32

25:20 20:1

5

THE PAYOFF OF ENRICHMENT IN MCI CLINICAL TRIALS (Nat Reviews Neurol 2010)

40:60

32:32

25:20 20:1

5 18:10

0.7

1.0

1.251.33

1.8

72%

0%

35%

55%

68%

“Low hippocampal volume … dichotomized … appears to help enriching recruitment into clinical trials aimed at studying drugs potentially slowing the progress/conversion to AD dementia [and] … might be considered a marker of progression to dementia in subjects with … predementiastage of AD (Dubois 2007), for the purposes of enriching a clinical trial population .”

EMA conclusions

“Collection, handling and measurements of Low hippocampal volume, as measured by MRI should be performed according to Good Clinical Practice and to the specific highest international standards .”

AUTOMATED SEGMENTATION:Valid versus what?

Morey et al., NeuroImage 2009

MANUAL SEGMENTATION OF THE HIPPOCAMPUS

PATHOLOGY MR

Geuze et al., Mol Psychiatry 2005;10:147-59

Ref. Med borderLat border

Inf borderNorm. hippo vol

(cm3)Left Right

Watson et al.

Mesial edge of temporal lobe

Temp horn of lat ventr

Incl subicular complex & uncal cleft w/ border separating subicular complex from parahippo gyrus

4.903 5.264

Zipursky et al.

Regional outline at choroidalfissure

Notmentioned

The interface of hippocampal tissue and parahippocampal gyruswhite matter

1.990 2.070

There’s more than 40 (very!) different ways to manually segment the hippocampus,

resulting in wildly different volume estimates

2.5-fold difference

The EADC-ADNI Working Group on the Harmonized Protocol for Hippocampal Volumetry

GB Frisoni, M Boccardi, M Bocchetta, R Ganzola, A Redolfi, D Tolomeo, G Corbetta, E Cavedo, M Lanfredi Brescia

B Bartzokis, PM Thompson, L

Apostolova, M Tinley

Los Angeles

CR Jack, G Preboske, C WardRochester

S Duchesne, N RobitailleQuebec City

H Soininen, M Pihlajamäki, Y Liu Kuopio

B Dubois, S Lehericy, C BoutetParis

H Hampel Frankfurt

S Teipel, M GrotheRostock

L-O Wahlund, B WinbaldStockholm

F Barkhof, P Scheltens, W Henneman, M Pronk, F van

Dommelen

Amsterdam

N Fox, J Barnes,

M BlairLondon

A SimmonsLondon

C deCarli, S Hollander

Davis

M Weiner, S Mueller

San Francisco

L deToledo-Morrell, D Bennet, T Stoub

Chicago

JG Csernansky, L Wang, A Christensen Evanston

RJ Killiany, C BauerBoston

M AlbertBaltimore M deLeon

New York

J Kaye, L Silbert, T SwihartPortland

J Pruessner, S Pietrantonio

MontrealR Camicioli, N MalykhinEdmonton

J PantelFrankfurt

C Watson Detroit

J O’Brien, M Firbank, E BurtonNewcastle

P Sachdev, JJ Maller(PATH through life

Study)

M Geerlings, L Gerritsen, M Portegies(SMART-Medea Study)

T denHeijer(Rotterdam Scan Study)

P Pasqualetti,Rome

L CollinsMontreal

PJ VisserMaastricht

L Froelich Mannhei

m

G Waldemar Copenhagen

L Launer Bethesda W Jagust

Berkeley

C Hock, H Wolf,Zurich

Frisoni & Jack, Alzh Dement 2011; Boccardi et al., JAD 2011; Boccardi et al., Alzh Dement submitted

Standardization of manual hippocampal volumetry

Acquisition Orientation Segmen-tation

ADNI

EADC-ADNI Hippocampal Harmonization Project

Funded by Alzheimer’s Association Co-funded by unrestricted grant from Wyeth and

Lilly

Extraction of largest possible common units(“LEGO blocks” or Segmentation Units)

Boccardi et al., JAD 2011

METHODS

Assessment of measurement

properties of SUs

Break down into segmentation

units

Bar

tzok

is, 1

998

Con

vit,

1997

Kill

iany

, 199

3S

oini

nen,

199

4Le

heric

y, 1

994

Jack

, 199

4;de

Tol

edo-

Mor

rell,

200

4H

alle

r, 1

997

Pru

essn

er, 2

000

Mal

ykhi

n, 2

007

Pan

tel,

2000

Wat

son,

199

2

Survey and operationalization

Boc

card

i et a

l., J

AD

201

1; B

occa

rdi e

t al.,

Alz

h D

emen

t sub

mitt

ed

METHODS

Assessment of measurement

properties of SUsDelphi panel

Answers are summarized and

anonymized

Disagreement issues are extracted

Disagreement issues are submitted to the

panel

Panel answers

Initial set of questions

5 ROUNDS

Covers 100% of hippo properCaptures 100% of AD atrophyVery high IRR & TRTR: <.96

Boccardi et al., JAD 2011; Boccardi et al., in preparation

1) test-retest variability

2) impact on total hippo volume

3) informativenessfor differences between AD and controls

How is the Harmonized Protocol Working?Agreement among L Apostolova (UCLA), M Bocchetta (IRCCS FBF), and

G Preboske (Mayo)

FUTURE STEPS

1. Develop reference probabilistic masks (ongoing)- 5 “master tracers”

2. Develop a qualification procedure, environment, and thresholds for

- naïve tracers - automated algorithms

3. Validate on:- 1800 ADNI hippocampi segmented by 20 human

tracers- MR-pathological datasets (volume on pathology,

neuronal density, volume on ex vivo MR)

4. Segmentation protocol and benchmark masks:- available to beta-testers based on ad hoc agreements

until the end of validation process (summer 2013), - no restrictions afterwards

CONCLUSIONS

Enrichment of clinical trials of diseasemodifiers in MCI is critical to detectsignal of efficacy

There is no magic enrichment recipe: a matter of earn/pay balance

Hippocampal volumetry isconceptually mature for enrichmentpurposes

Operational procedures might be refined with the harmonized manualsegmentation protocol as benchmark

G BartzokisLos Angeles

M deLeonNew York

L deToledo-MorrellChicago

J CsernanskyChicago

CR JackRochester

R KillianyBoston

S LehericyParis

N MalykhinEdmonton

J PantelFrankfurt

J PruessnerMontreal

H SoininenKuopio

C WatsonDetroit

DELPHI PANELISTS

L ApostolovaLos Angeles

J BarnesLondon

G BartzokisLos Angeles

C deCarli Sacramento

L deToledo-Morrell Chicago

M FirbankNewcastle

L GerritsenStockholm

W HennemanAmsterdam

CR JackRochester

R KillianyBoston

N Malykhin Edmonton

J PruessnerMontreal

H SoininenKuopio

L WangChicago

C WatsonDetroit

H WolfZurich

AUTHORS OF ORIGINAL SEGMENTATION PROTOCOLS

L ApostolovaLos Angeles

M BocchettaBrescia

R GanzolaBrescia

G PreboskeRochester

D WolfMainz

MASTER TRACERS

COWORKERS AT THE LAB OF NEUROIMAGINGThe National Centre for Alzheimer’s Disease, Brescia

Hippo Harmo TeamM Boccardi, M Bocchetta, E Cavedo

A Redolfi, G Corbetta, D Tolomeo

EnrichmentM Lorenzi

Aphasia Apraxia Agnosia

Spared SM and visual functions

Encoding semantic memories

Smell discrim.

Encoding episodic &

spatial memories

AMYLOID PLAQUES

Braak & Braak , Acta Neuropath 1991

Tha

l et a

l., N

euro

logy

200

2

?

TANGLES

Atrophy is an intermediate phenotype between neurodegeneration and clinical symptoms

SYMPTOMSNEURODEGE-

NERATION ATROPHY

Frisoni et al., J Neurol 2009

Courtesy of P. Giannakopoulos

Frisoni et al., J Neurol 2009

HIPPOCAMPAL MAPPINGThe method 2

CA1

Fimbria

Subiculum

CA1

CA2-3

Presubiculum

Subiculum

R L

Frisoni et al. Neuroimage 2006

31

Atrophy in AD maps onto areas known to be affected by tau pathology and is specific to AD

L’atrophie dans la MA mappe sur les régions atteint es par le dépot de tau et elle est spécifique à la MA

Alzheimer’s pattern

Frisoni et al., Brain 2007

CA1 CA2-3Aging pattern

Frisoni et al., Brain 2007

Lewy body pattern

Sabattoli et al., NeuroImage 2008

Antisocial dis. pattern

Boccardi et al., Hum Brain Mapp 2010

Correlation of brain atrophy with rate of MMSE decline in 29 AD patients ( r= 0.80, p < 0.001).

Fox

et a

l., N

euro

logy

1999

;52:

1687

THE TRADEOFF OF ENRICHMENT IN MCI CLINICAL TRIALS

Lorenzi et al., Neurobiol Aging 2010

0.560.980.890.870.891.460.871.14

RATIO CONV./NON CONV.

0%56%38%35%46%77%55%86%

% SCREENED OUT