Accelerated Aging, Decreased White Matter Integrity, and Associated Neuropsychological Dysfunction...

Accelerated Aging, Decreased White Matter Integrity, andAssociated Neuropsychological Dysfunction 25 Years AfterPediatric Lymphoid MalignanciesIlse Schuitema, Sabine Deprez, Wim Van Hecke, Marita Daams, Anne Uyttebroeck, Stefan Sunaert,Frederik Barkhof, Eline van Dulmen-den Broeder, Helena J. van der Pal, Cor van den Bos, Anjo J.P. Veerman,and Leo M.J. de Sonneville

See accompanying editorial on page 3309

Ilse Schuitema and Leo M.J. de Sonn-eville, Leiden University, Leiden; IlseSchuitema, Marita Daams, FrederikBarkhof, Eline van Dulmen-den Broe-der, and Anjo J.P. Veerman, VrijeUniversiteit University Medical Center;Helena J. van der Pal and Cor van denBos, Academic Medical Center,Amsterdam, the Netherlands; SabineDeprez, Anne Uyttebroeck, and StefanSunaert, University Hospitals Leuven;Wim Van Hecke, icoMetrix, Leuven,Belgium.

Published online ahead of print atwww.jco.org on August 19, 2013.

Supported by Grants No. UL 2006-3630from the Dutch Cancer Society(L.M.J.D.S., A.J.P.V., and C.V.D.B.) andNo. G.048010N from Fonds Weten-schappelijk Onderzoek Vlaanderen andby Stichting tegen Kanker.

I.S. and S.D. contributed equally to themanuscript.

Presented at the 44th Congress of theInternational Society of PaediatricOncology, London, United Kingdom,October 5-8, 2012.

The funding sources had no role instudy design; collection, analysis, andinterpretation of data; writing the arti-cle; and the decision to submit forpublication.

Authors’ disclosures of potential con-flicts of interest and author contribu-tions are found at the end of thisarticle.

Corresponding author: Ilse Schuitema,MSc, Department of Clinical Child andAdolescent Studies, Faculty of SocialSciences, Leiden University, Wasse-naarseweg 52, P.O. Box 9555, 2300RB, Leiden, the Netherlands; e-mail:[email protected].

© 2013 by American Society of ClinicalOncology

0732-183X/13/3127w-3378w/$20.00

DOI: 10.1200/JCO.2012.46.7050

A B S T R A C T

PurposeCNS-directed chemotherapy (CT) and cranial radiotherapy (CRT) for childhood acute lymphoblasticleukemia or lymphoma have various neurotoxic properties. This study aimed to assess their impacton the maturing brain 20 to 30 years after diagnosis, providing a much stronger perspective onlong-term quality of life than previous studies.

Patients and MethodsNinety-three patients treated between 1978 and 1990 at various intensities, with and without CRT,and 49 healthy controls were assessed with magnetic resonance diffusion tensor imaging (DTI)and neuropsychological tests. Differences in fractional anisotropy (FA)—a DTI measure describingwhite matter (WM) microstructure—were analyzed by using whole brain voxel-based analysis.

ResultsCRT-treated survivors demonstrated significantly decreased FA compared with controls in frontal,parietal, and temporal WM tracts. Trends for lower FA were seen in the CT-treated survivors.Decreases in FA correlated well with neuropsychological dysfunction. In contrast to the CT groupand controls, the CRT group showed a steep decline of FA with age at assessment. Younger ageat cranial irradiation and higher dosage were associated with worse outcome of WM integrity.

ConclusionCRT-treated survivors show decreased WM integrity reflected by significantly decreased FA andassociated neuropsychological dysfunction 25 years after treatment, although effects of CT aloneseem mild. Accelerated aging of the brain and increased risk of early onset dementia aresuspected after CRT, but not after CT.

J Clin Oncol 31:3378-3388. © 2013 by American Society of Clinical Oncology

INTRODUCTION

Prophylaxis to prevent meningeal relapse afterchildhood acute lymphoblastic leukemia (ALL) orlymphoma used to consist of both intrathecal (IT)chemotherapy (CT) and cranial radiotherapy(CRT). The neurotoxic adverse effects of prophylac-tic CRT became apparent during the 1980s and arewell documented.1 Around that time, CRT wasmostly abolished and replaced with more intensiveIT CT.2-4 Knowledge of late effects is mostly re-stricted to the first decade after treatment,5-8 butthese effects may have an impact on quality of life formany decades because both long-term survival ratesand general life expectancy keep increasing. Thisstudy provides information about the long-term im-pact of CT and CRT on the maturing brain 20 to 30

years after diagnosis by using magnetic resonance(MR) diffusion tensor imaging (DTI) and neuro-psychological assessment. The outcomes will help tobetter understand these survivors’ current needs andcould aid in anticipating late effects of CT and ofCRT that is currently applied for brain tumors and isstill used for ALL in some countries.

There are many reports on late neurocognitivedeficits after CT alone, particularly in the domain ofexecutive functions, although these are relativelysubtle compared with the effects of CRT.9-12 Se-quelae seem more severe after high-risk treatment.13

Advanced neuroimaging techniques allow us to in-vestigate possible neural substrates of these cognitivesequelae. An MR voxel-based morphometry studyby Carey et al7 linked reduction of white matter(WM) volumes within the right frontal lobe

JOURNAL OF CLINICAL ONCOLOGY O R I G I N A L R E P O R T

VOLUME 31 � NUMBER 27 � SEPTEMBER 20 2013

3378 © 2013 by American Society of Clinical Oncology

132.229.26.160Information downloaded from jco.ascopubs.org and provided by at WALAEUS LIBRARY on January 10, 2014 from

Copyright © 2013 American Society of Clinical Oncology. All rights reserved.

to decreased cognitive functioning 10 years after treatment with CTonly. The effects of CT beyond 10 years after treatment are sparselystudied. Porto et al14 studied 10 female ALL survivors, on average 15years post-treatment with CT only, and found distributed reductionsof gray matter and WM concentrations and a trend for decreased WMintegrity. Research by Koppelmans et al15 indicates that cognitiveeffects of adjuvant CT for breast cancer can persist beyond 20 yearsafter treatment, but it remains unclear how this translates to IT CTin childhood.

Late neurotoxic effects of CRT include an increased risk of neu-rologic complications such as vascular malformations, secondary neo-plasms, and focal necrosis.16-18 In the cognitive domain, deficits inmemory, information processing speed, and attention have beenreported.19-24 WM volume reduction is more apparent after CRT thanafter CT and correlates significantly with cognitive impairment.6

On a microstructural level, DTI is able to quantify WM organi-zation by assessing the restriction of randomly moving water mole-cules. The degree of directional preference of diffusion is quantified bythe DTI parameter fractional anisotropy (FA). Damage to WM mi-crostructures will result in lower FA because of relatively more diffu-sion of water perpendicular to the fiber orientation.25 Significantdifferences in FA between cancer survivors (up to10 years after treat-ment) and controls have been reported and linked to cognitiveimpairment.26-30 Dellani et al31 demonstrated WM alterations in 13CRT-treated ALL survivors 16 to 28 years after treatment. Porto et al14

demonstrated reduced FA after CRT in 11 males 15 years after treat-ment for childhood ALL. Associations with cognition remained un-addressed by Dellani and Porto, and they were both unable todemonstrate a relation between FA and age at diagnosis.

Younger age at treatment with CRT has been associated withmore severe WM lesions and worse neurocognitive outcome.32-35

Progression of radiation-induced WM changes has also been re-ported.36,37 Prophylactic CRT (20-30 Gy) for small-cell lung cancer inadults has been associated with early-onset dementia.38-41 In general,theories are emerging that cancer and cancer treatment may causeaccelerated aging of the brain and of cells in general, even after CTwithout CRT.42,43 This study investigated WM changes and associatedneuropsychological dysfunction 25 years after treatment. The hypoth-esis of accelerated WM decay will be addressed by comparing thecorrelations of FA and age between survivors treated with and withoutCRT and controls. In addition, the relation between age at diagnosisand vulnerability to neurotoxicity of treatment will be explored.

PATIENTS AND METHODS

Patients

We identified 285 survivors of ALL or lymphoma from patient records ofthe Vrije Universiteit University Medical Center, the Academic Medical Cen-ter Amsterdam in the Netherlands, and the University Hospitals Leuven inBelgium. Dates of diagnoses after 1978 were included, and time since diagnosishad to be at least 18 years. After exclusion of 42 survivors (Appendix Fig A1,online only), 243 eligible survivors were contacted. Ninety-six survivors werewilling to participate and were asked to recruit a control (sibling, partner, orfriend; n � 49). Assessments took place between 2007 and 2011.

Survivors were treated according to Berlin-Frankfurt-Munster (BFM)–based protocols with a duration of approximately 2 years.44 Between 1979 and1983, standard-risk CRT patients were treated according to the Dutch Child-hood Leukemia Study Group (DCLSG) protocol ALL-5 or the Riehm proto-col, both characterized by CRT (15 to 25 Gy) in addition to five to seven IT

injections of 12 to 12.5 mg methotrexate (MTX). This cohort included 24survivors treated according to a standard-risk CRT protocol. We includedseven high-risk (HR) patients from this period, treated with additional cus-tomized high-dose MTX intravenously (IV). Around 1983, CRT was abol-ished and standard-risk patients were treated according to DCLSG protocolALL-6 (13 � 12 mg MTX IT and 6 g/m2 MTX IV) or European Organisationfor Research and Treatment of Cancer (EORTC) Trial 58831 (6 � 12 mg MTXIT and 2 g/m2 MTX IV), of which 29 patients were included. HR patients weretreated with customized protocols based on either EORTC Trial 58832 (8 � 12mg MTX IT and 10 g/m2 MTX IV), the BACOP [bleomycin, doxorubicin,cyclophosphamide, vincristine, and prednisone] protocol (customized dose ofMTX IV), or ALL-6 (customized additional MTX IV) without CRT (HR CT,n � 20).45 Administration of MTX IV was always followed by leucovorin (12to 15 mg/m2 every 6 hours until serum levels of MTX had dropped below 10�7

mol/L). Thirteen patients were treated for relapse after standard-risk or HRtreatment, seven of whom were irradiated only during relapse and six of whomwere irradiated at both initial occurrence and relapse. Group means of dosageswere calculated excluding missing data. Dosages of CRT were available for allpatients, but dosages of MTX IV were missing for three survivors in the CRTgroup and dosages of MTX IT were missing for one survivor in the CT group.Three irradiated survivors were excluded because meningiomas were discov-ered during assessment. Survivors were grouped into irradiated (n � 44) andnonirradiated (n � 49), and dose-dependent effects of CRT, MTX IV, andMTX IT were studied. The ethical principles of the Helsinki Declaration werefollowed and approval was obtained from the local ethical committees.

Acquisition Details for MR Imaging

Patients were scanned on a 1.5T Sonata system (Siemens, Erlangen,Germany), including a T1-weighted 3D gradient sequence (TR, 2,700 ms; TE,5.17 ms; flip angle, 8 degrees; 160 coronal slices; voxel size, 1 � 1 � 1.5 �L).DTI was measured by using a 9-minute echo planar imaging sequence (TR,8,500 ms; TE, 86 ms; voxel size, 2 mm isotropic; 59 slices; acquisition matrix,128 � 128 mm; FOV, 256 mm; 60 directions [b value, 700 s/mm2]; 10b0 images).46

DTI Processing and Statistical Analysis

DTI preprocessing was done by using ExploreDTI consisting of motionand distortion correction with reorientation of the b-matrix and an iterativeweighted nonlinear tensor estimation process to generate FA maps.47 Individ-ual DTI data sets were nonrigidly registered to a population-based DTI atlasgenerated from DTI images from all patients.48-50 Finally, the resulting imageswere smoothed with a 3D-Gaussian kernel of full width at half maximum of6 mm.

Statistical Parametric Mapping 8 (SPM8) whole-brain voxel-based anal-ysis of variance was performed to assess differences in FA between the differentgroups.51 Age at assessment (AaA) was used as covariate. A WM mask wasapplied to limit the analysis to WM voxels only. The resulting statisticalparametric maps were thresholded at the voxel-level P � .001. Only clusterssignificant at the family-wise error P � .05 level corrected for multiple com-parisons were retained.

Neuropsychological Assessment and Statistical Analysis

The Amsterdam Neuropsychological Tasks (ANT) program was used toassess executive functions.52 The computerized ANT provides standardizedassessments and automated recordings of speed and accuracy of informationprocessing, attention processes, and working memory53-60 (see Appendix,online only). Intelligence quotient was estimated by using a four subtestshort-form of the Wechsler Adult Intelligence Scale Revised (WAIS-R III).61

Differences between groups were tested by using analysis of variance andsimple contrasts with controls as the reference group. AaA was used as acovariate. Task parameters that discriminated between survivors and controlswere selected for correlation analyses with FA.

Correlations Between FA and Discriminative

Neuropsychological Parameters

To study general associations between FA and cognition, a whole-brainvoxel-based correlation analysis was conducted that included all survivors andcontrols, with FA as dependent variable, the selected neuropsychological test

White Matter Integrity 25 Years After Cranial Irradiation For ALL

www.jco.org © 2013 by American Society of Clinical Oncology 3379



Tabl

e1.

Cha

ract

eris

tics

and

Neu

rops

ycho

logi

calT

ask

Per

form

ance

ofth

eIn

clud

edP

atie

nts

Cha

ract

eris

tic

Con

trol

s(n

�49

)

CR

T(n

�44

)

CT

(n�

49)

AN

OV

A(b

etw

een-

patie

nts

effe

cts)

Gro

upFa

ctor

AaA

Fact

or

%

Tota

l

Sta

ndar

dR

isk

(n�

24)

Hig

hR

isk

(n�

7)R

elap

se(n

�13

)

%

Tota

lS

tand

ard

Ris

k(n

�29

)H

igh

Ris

k(n

�20

)

%M

SD

MS

DM

SD

MS

DM

SD

MS

DM

SD

MS

DF

PF

P

Mal

es42

.952

.357

.11.

025

.362

N/A

N/A

AaA

(yea

rs)

26.5

5.9

31.2

�4.

831

.94.

332

.64.

029

.15.

726

.75.

124

.52.

329

.96.

211

.459

�.0

01N

/AN

/AA

aD(y

ears

)N

/A5.

73.

75.

43.

47.

14.

65.

54.

05.

33.

54.

31.

66.

84.

80.

353

.554

N/A

N/A

Tim

esi

nce

diag

nosi

s(y

ears

)N

/A25

.43.

226

.42.

725

.43.

223

.73.

421

.42.

920

.22.

123

.23.

041

.175

�.0

01N

/AN

/AC

umul

ativ

edo

seC

RT

(Gy)

N/A

22.5

916.

991

20.8

33.

4424

.43

3.51

24.8

511

.52

00

00

00

N/A

N/A

N/A

N/A

Cum

ulat

ive

dose

MTX

IV(m

g/m

2)

N/A

15,8

53.7

30,4

98.0

00

15,5

0023

,402

45,2

3139

,062

18,2

24.5

19,8

80.1

5,92

985

834

,619

21,2

99N

/AN

/AN

/AN

/AC

umul

ativ

edo

seM

TXIT

(mg)

N/A

111.

995

.754

1375

8923

863

116.

455

.614

030

8467

N/A

N/A

N/A

N/A

Est

imat

edIQ

107.

419

.496

.9�

20.2

103.

418

.52.

194

.116

0.98

6.3

23V

isuo

mot

orac

cura

cy3.

30.

63.

9�0.

93.

50.

83.

175

.045

5.69

7.0

18V

isuo

mot

orst

abili

ty2.

10.

72.

6�0.

92.

10.

53.

022

.052

7.81

1.0

06S

usta

ined

atte

ntio

nw

ork

pace

8.2

1.4

9.5�

2.3

8.5

1.5

5.41

3.0

050.

053

.818

Vis

uosp

atia

lse

quen

cing

3.2

3.2

8.2�

5.7

4.5

4.8

9.10

1�

.001

4.22

5.0

42

NO

TE.

Larg

erva

lues

for

visu

omot

orac

cura

cy,

visu

omot

orst

abili

ty,

sust

aine

dat

tent

ion

wor

kpa

ce,

and

visu

ospa

tials

eque

ncin

gde

note

wor

sepe

rfor

man

ce.

Abb

revi

atio

ns:A

aA,a

geat

asse

ssm

ent;

AaD

,age

atdi

agno

sis;

AN

OV

A,a

naly

sis

ofva

rianc

e;C

RT,

cran

ialr

adio

ther

apy;

CT,

chem

othe

rapy

;IQ

,int

ellig

ence

quot

ient

;IT,

intr

athe

cally

;IV

,int

rave

nous

ly;M

,mea

n;M

TX,

met

hotr

exat

e;N

/A,

not

appl

icab

le.

�S

igni

fican

tco

ntra

stre

sult

(sim

ple

cont

rast

vco

ntro

ls).

Schuitema et al

3380 © 2013 by American Society of Clinical Oncology JOURNAL OF CLINICAL ONCOLOGY



scores as regressor, and age as covariate. Pearson’s r was calculated withinclusters showing significant correlations.

Age at Diagnosis and Time Since Diagnosis

The separate effects of AaA and age at diagnosis (AaD) on WM organi-zation were studied with a whole-brain voxel-based correlation analysis, con-trolling the correlation between FA and AaA for AaD, and between FA andAaD for AaA. Only participants between age 20 and 40 years were included,because within a healthy population, FA values are minimally affected bynormal maturation and aging in this age range (CRT, n � 30; CT, n � 45;controls, n � 44).63 Relapse patients needed to be excluded because of theirdouble AaD. Subsequently, the relationship between FA and AaA and AaD inCRT-treated survivors was studied with linear regression models within sig-nificant clusters also relevant to cognition.64

Dosage Correlations

The potential influence of therapy dosage on FA was explored with avoxel-based correlation analysis between FA and cumulative doses of CRT anddoses of MTX IV and/or MTX IT within the CRT and CT groups, respectively.AaA and AaD were included as covariates.

RESULTS

Patient Demographics and

Neuropsychological Function

Table 1 provides a summary of participants’ characteristics andcognitive performance. As expected, the groups differed significantlyin AaA and time since diagnosis because the protocols with and with-out CRT were applied consecutively. However, the age of controls(range, 17.3 to 43.4 years) covers the full age range of survivors (range,

18.9 to 43.7 years). Estimated intelligence quotient was significantlydecreased in the CRT group.

Visuomotor accuracy, visuomotor stability, work pace duringsustained attention, and visuospatial sequential working memoryshowed a significant overall group difference. Contrasting each survi-vor group with controls showed that irradiated survivors performedworse than controls. These neuropsychological tasks were selected forcorrelation with FA maps.

Assessment of Differences in FA

CRT-treated survivors demonstrated significantly decreased FA(family-wise error corrected P � .05) compared with controls inorbitofrontal WM, genu, anterior body, and forceps minor of corpuscallosum (CC), cingulum (frontal and parietal), and inferior fronto-occipital fasciculus (IFOF), superior longitudinal fasciculus (SLF) anduncinate fasciculi (Fig 1). After CT, trends for lower FA were seen infrontal WM tracts (Table 2).

Correlation Analysis of Neuropsychological

Performance With FA Values

Voxel-based correlation analysis between FA maps of all patientsand task performance revealed significant correlations in frontal, pa-rietal, and temporal WM tracts with measures of visuomotor control,visuospatial sequencing, and sustained attention work pace (family-wise error corrected P � .05; Table 3 and Fig 2A). Figure 2D illustratesthe correlation between FA in frontal WM and visuomotor accuracy.

A

B

R

R

T

6

3

Fig 1. (A) Sagittal and (B) axial slices ofregions showing significantly decreasedfractional anisotropy in cranial radiotherapy–treated survivors when compared withhealthy controls (thresholded T maps [P �.001]). Color indicates significance.





Tabl

e2.

FAV

oxel

-Wis

eA

naly

sis

for

Bra

inR

egio

nsS

how

ing

Sig

nific

antly

Red

uced

FA�

inS

urvi

vors

Com

pare

dW

ithH

ealth

yC

ontr

ols

Con

tras

tR

egio

nC

lust

erFa

mily

-Wis

eE

rror

Cor

rect

edP

Clu

ster

Siz

e(in

num

ber

ofvo

xels

)A

nato

mic

Ext

ent

ofC

lust

erT

Mea

nFA

Con

trol

sM

ean

FAS

urvi

vors

Con

trol

sv

CR

T-tr

eate

dsu

rviv

ors

R�

Lfr

onta

l�

.001

2,51

4C

lust

erco

verin

gor

bito

fron

talW

M,

genu

,an

terio

rbo

dy,

and

forc

eps

min

orof

CC

,ci

ngul

um,

IFO

F,un

cina

tefa

scic

ulus

,A

LIC

,S

LF

7.58

0.37

40.

332

Rpa

rieta

l.0

215

4C

lust

erco

verin

gpa

rtof

cing

ulum

and

CC

4.12

0.41

50.

375

Con

trol

sv

CT-

trea

ted

surv

ivor

sL

fron

tal

.031

474

Clu

ster

cove

ring

forc

eps

min

orof

CC

,ci

ngul

um,

coro

nara

diat

a4.

730.

388

0.37

5

NO

TE.T

hres

hold

set

atP

�.0

01(c

lust

ers

sign

ifica

ntat

fam

ily-w

ise

erro

rco

rrec

ted

P�

.05

wer

ere

tain

edfo

rth

eC

RT

grou

p).F

orth

eC

Tgr

oup,

thre

shol

dw

asse

tat

P�

.01

(clu

ster

ssi

gnifi

cant

atfa

mily

-wis

eer

ror

corr

ecte

dP

�.0

5w

ere

reta

ined

).M

ean

frac

tiona

lani

sotr

opy

(FA

)va

lues

are

repo

rted

for

the

iden

tified

clus

ters

.A

bbre

viat

ions

:A

LIC

,an

terio

rlim

bof

inte

rnal

caps

ula;

CC

,co

rpus

callo

sum

;C

RT,

cran

ialr

adia

tion

ther

apy;

CT,

chem

othe

rapy

;FA

,fr

actio

nala

niso

trop

y;IF

OF,

infe

rior

fron

to-o

ccip

italf

asci

culu

s;L,

left

;R

,rig

ht;

SLF

,su

perio

rlo

ngitu

dina

lfas

cicu

lus;

WM

,w

hite

mat

ter.

�Tr

ends

for

sign

ifica

ntly

redu

ced

FAar

epr

ovid

edfo

rC

T-tr

eate

dsu

rviv

ors

vco

ntro

ls.

Schuitema et al




Tabl

e3.

Bra

inR

egio

nsS

how

ing

Sig

nific

antly

Neg

ativ

eC

orre

latio

ns(F

amily

-Wis

eE

rror

Cor

rect

edP

�.0

5)B

etw

een

FAan

dC

ogni

tive

Sco

res

and

FAan

dC

RT

Dos

age

AN

TTa

skS

ide

Reg

ion

Clu

ster

Fam

ily-W

ise

Err

orC

orre

cted

PC

lust

erS

ize

Ana

tom

ical

Ext

ent

ofC

lust

erT

Pea

rson

r�

AN

Tpu

rsui

tV

isuo

mot

orac

cura

cyR

�L

Fron

tal�

parie

tal

�.0

016,

231

Hug

ecl

uste

rco

verin

gbi

late

ralt

hala

mic

radi

atio

nan

dsu

bcor

tical

orbi

tofr

onta

lW

M,

PLI

C,

ante

rior

coro

nalr

adia

ta(L

�R

)in

clud

ing

ante

rior

part

ofS

LF,

CC

,m

ajor

and

min

orfo

rcep

s,ci

ngul

um(L

�R

),an

dW

Mof

the

left

ante

rior

tem

pora

lpol

ein

clud

ing

the

unci

nate

fasc

icul

us

6.51

–0.4

85

RFr

onta

l.0

4511

8C

lust

erin

clud

ing

WM

unde

rth

ean

terio

rsu

perio

rfr

onta

lgyr

us4.

63–0

.374

RTe

mpo

ral

.002

263

Clu

ster

incl

udin

gsu

bcor

tical

WM

,in

ferio

rlo

ngitu

dina

lfas

cicu

lus,

and

infe

rior

fron

to-o

ccip

italf

asci

culu

s

4.41

–0.3

38

Vis

uom

otor

stab

ility

RTe

mpo

ral

�.0

0156

6C

lust

erin

clud

ing

subc

ortic

alW

M,

infe

rior

and

supe

rior

long

itudi

nalf

asci

culi,

and

infe

rior

fron

to-o

ccip

italf

asci

culu

s

6.18

–0.4

59

RFr

onta

l�

.001

626

Bila

tera

lant

erio

rth

alam

icra

diat

ion

and

subc

ortic

alor

bito

fron

talW

M5.

25–0

.382

LLy

mbi

c.0

0130

8C

ingu

lum

4.99

–0.3

79R

�L

Fron

tal

�.0

0197

3C

lust

erco

verin

gle

ftan

drig

htci

ngul

um,

right

genu

ofC

C4.

96–0

.433

RP

arie

tal

.042

122

Forc

eps

maj

orC

C3.

99–0

.323

AN

Tvi

suos

patia

lse

quen

cing

Acc

urac

yof

sequ

entia

lw

orki

ngm

emor

ypr

oces

ses

R�

LFr

onta

l�

.001

1,22

4C

lust

erco

verin

gbi

late

rala

nter

ior

thal

amic

radi

atio

nan

dsu

bcor

tical

orbi

tofr

onta

lW

M

5.37

–0.4

15

RFr

onta

l.0

0222

6C

lust

erin

clud

ing

ante

rior

part

ofS

LF4.

82–0

.354

R�

LP

arie

tal

�.0

0152

0C

lust

erco

verin

gci

ngul

uman

dfo

rcep

sm

ajor

CC

4.76

–0.3

10

RP

arie

tal

.007

200

Clu

ster

incl

udin

gS

LF4.

55–0

.355

LFr

onta

l.0

0224

7C

lust

erin

clud

ing

ante

rior

part

ofS

LF4.

52–0

.410

AN

Tsu

stai

ned

atte

ntio

nW

ork

pace

Fron

tal

.026

142

Clu

ster

cove

ring

body

ofC

Can

dci

ngul

um4.

28–0

.340

CR

Tdo

sage

LFr

onta

l.0

510

2C

lust

erin

clud

ing

part

ofIF

OF

and

unci

nate

fasc

icul

us6.

08–0

.673

RFr

onta

l.0

0127

2C

lust

erin

clud

ing

part

ofco

rona

radi

ata

and

CC

5.26

–0.6

60

RFr

onta

l.0

2213

5C

lust

erin

clud

ing

part

ofIF

OF

and

unci

nate

fasc

icul

us4.

84–0

.631

Abb

revi

atio

ns:A

NT,

Am

ster

dam

Neu

rops

ycho

logi

calT

asks

;CC

,cor

pus

callo

sum

;FA

,fra

ctio

nala

niso

trop

y;IF

OF,

infe

rior

fron

to-o

ccip

italf

asci

culu

s;L,

left

;PLI

C,p

oste

rior

limb

ofin

tern

alca

psul

a;R

,rig

ht;S

LF,

supe

rior

long

itudi

nalf

asci

culu

s;W

M,

whi

tem

atte

r.�P

ears

on’s

corr

elat

ion

coef

ficie

ntca

lcul

ated

betw

een

AN

Tva

riabl

ean

dav

erag

eFA

ofsp

here

of3

mm

arou

ndpe

akvo

xelo

fcl

uste

r.





Correlations With AaD and AaA

Within the CRT group, significant positive correlations (family-wise error P � .05) between FA and AaD (corrected for AaA) werefound in frontal and parietal WM, including parts of the forcepsmajor, forceps minor, and body of CC, uncinate fasciculus, IFOF,posterior limb of internal capsula (PLIC), SLF, and orbitofrontal WM(Fig 2B).

Significant negative correlations (family-wise error correctedP � .05) between FA and AaA, controlled for AaD, were found infrontal, parietal, and temporal WM, including parts of the forcepsminor, forceps major, and body of CC; PLIC and anterior limb ofinternal capsula (ALIC), SLF, thalamic radiation, corona radiata,IFOF, and the uncinate fasciculus (Fig 2B). Within controls, correla-tions with AaA were not significant. In the CRT group, FA declined

T

5

3

5

T

3

A

B

C

FED

R

R

R

Visu

omot

or A

ccur

acy

Mean FA of Cluster

6.0

5.5

5.0

4.5

4.0

3.5

3.0

2.5

2.00.24 0.26 0.28 0.30 0.32 0.34 0.36 0.38

M - 2 SD M - 1 SD

M + 2 SD

M + 1 SD

M

M

0.40 0.42

Mea

n FA

of C

lust

er

Age at Assessment (years)

0.68

0.66

0.64

0.62

0.60

0.58

0.56

0.54

0.52

0.5015 20 25 30 35 40 45

TC1NOC CRT

Mea

n FA

of C

lust

er

Cumulative Dose CRT (Gy)

0.38

0.36

0.34

0.32

0.30

0.28

0.26

0.240 10 20 30 40

M

M - 1 SD

M - 2 SD

50

CRT Linear (CRT)CON CRT Linear(CON)

Linear(CRT)

Fig 2. Thresholded T maps* (P � .001) showing regions with significant negative correlation between fractional anisotropy (FA) and (A) visuomotor accuracy and (B)age at assessment (AaA; blue) and significant positive correlation between FA and age at diagnosis (AaD; red). Overlapping regions show mixed colors (pink). Theanalysis was performed for the cranial radiotherapy (CRT) group (n � 30) with relapse patients being excluded because of double AaD. Cluster (white arrow) selectedon the basis of overlap between an area of significant correlation (family-wise error corrected P � .05) between FA and AaA, and an area of significant correlationbetween FA and cognitive variables. (C) Thresholded T maps* (P � .001) showing regions with significant negative correlation between FA and cumulative doses ofCRT. Cluster (white arrow) selected on the basis of overlap between an area of significant correlation (family-wise error corrected P � .05) between FA and dosageof CRT and an area of significant correlation between FA and cognitive variables. (D) Scatterplot of visuomotor accuracy and mean FA in a cluster in the frontal WM(sphere of 3 mm around peak voxel) indicated by the white arrow in (A). Horizontal lines indicate the normal range of the mean (M) visuomotor accuracy score of controls(CONs) (M, �1 standard deviation [SD], and �2 SD). Note that higher values of visuomotor accuracy indicate larger distance to target (ie, worse performance). Verticallines indicate the normal range of FA of controls (M, �1 SD, and �2 SD). The colored lines on the x axes of this scatterplot indicate the mean FA values of the differentgroups. There is a significant correlation between FA and visuomotor accuracy (r � �0.485; P � .001) calculated for all participants together. Note that a small subgroupof chemotherapy-treated survivors falls outside the normal range. (E) Scatterplot of AaA and FA in the cluster indicated by the white arrow in (B). Since AaD isdisregarded here, all CRT-treated survivors (n � 43) and controls (n � 44) between age 20 and 40 years were included. The correlation between FA and AaA is absentin controls (blue line; r � 0.004; P � .980) but is significantly negative in the group of CRT-treated survivors (gold line; r � �0.414; P � .006). The significant interactioneffect between these correlations is indicative of accelerated aging (interaction beta � �.282; P � .008). (F) Scatterplot of all CRT-treated survivors (n � 44), displayingthe correlation between FA and cumulative doses of CRT in the cluster indicated by the white arrow in (C) (r � �0.632; P � .001. Horizontal lines indicate the normalrange of FA of controls (M, �1 SD, and �2 SD). (*) T-maps of A, B, and C are overlaid on axial slices of the population-based atlas.

Schuitema et al




with age, as demonstrated by interacting regressions of FA on AaAbetween survivors and controls. This significant interaction indicatesaccelerated aging of WM (Fig 2E). This effect was shown to be largerwhen the effect of AaD was taken into account (Table 4). Note that ascatterplot of FA in cluster splenium CC from Table 4 is displayed inFig 2B and a scatterplot of FA in cluster IFOF, uncinate fasciculus inFig 2C. These effects are also visible on the neuropsychological vari-ables (Table 4). No significant correlations with AaA or AaD werefound within the CT group. Correlations between AaA and the neu-ropsychological variables are described in Table 4.

Dosage Correlations

Significantly negative correlations between FA and cumulativedoses of CRT were found in clusters covering the CC, corona radiata,IFOF, and uncinate fasciculus (Fig 2C; Table 3). Interactions or cor-relations with doses of MTX IT or IV could not be established. Withinthe CT group, no correlations with doses of MTX IT or IV were found.No effects of doses of CRT or MTX IT or IV on neuropsychologicaloutcome were observed.

DISCUSSION

This study demonstrated decreased WM integrity, as determined byFA, in a large group of leukemia and lymphoma survivors 25 yearsafter treatment. Younger age at cranial irradiation and higher dosagewere associated with lower FA. Accelerated aging of the irradiatedbrain was suggested. In addition, decreased FA was significantly asso-ciated with neuropsychological dysfunction.

For irradiated survivors, both FA and neuropsychological func-tion were significantly below average. The dependence of these long-term outcomes on AaA and dosage is consistent with literature

describing short-term outcomes of CRT.65 The steep decline of FAwith AaA compared with controls, most importantly within the fron-tal and parietal WM, is a strong indication of accelerated aging. Ingeneral, the risk of developing dementia increases with age. There arealso anatomic similarities between our survivors and patients withAlzheimer’s disease. Parente et al66 reported decreased FA in the CC,cingulum, and SLF in patients with Alzheimer’s disease and those withmild cognitive impairment, similar to what we found. Furthermore,our own magnetoencephalography findings from the same cohortdisplayed an oscillatory activity pattern resembling the pattern foundin patients with Alzheimer’s disease.67 Together, these findings suggestthat the irradiated survivors could be at increased risk of developingearly-onset dementia.

In general, FA values correlate well with cognition, in particularwith tests of executive functions.68,69 In this cohort, neuropsycholog-ical dysfunction correlated significantly with lower FA in the CC,cingulum, SLF, inferior longitudinal fasciculus, and IFOF. These ma-jor WM tracts are crucial roadways for functional networks, which iswhy small focal damage in these tracts can have widespread impact onbrain functioning.68,69

The observed decreases in FA might be related to decreasedmyelin and/or axonal injury. There is evidence that suggests thatboth CT and CRT can cause early apoptosis of oligodendrocytes—essential for the myelinization of axons—and vasculopathyleading to ischemia.6,14,70-73 Both CT and CRT can limit neuralrepair by damaging periventricular progenitor cells that wouldotherwise maintain WM integrity and stimulate hippocampalneurogenesis.74-77 Evidence is accumulating that these processesare fundamental to understanding late cognitive effects andWM decay and are likely to provide targets for future therapeu-tic interventions.74,77,78

Table 4. Linear Regression Models

Dependent Variable

Independent Variables

Step 1

Step 2 Step 3

AaA r P R2AaA

Partial r PAaD

Partial r P R2AaA

Partial r PAaD

Partial r PDosage CRT

Partial r P R2

FA� splenium CC �0.409 .025 0.168 �0.766 � .001 0.713 � .001 0.591 �0.765 � .001 0.728 � .001 �0.378 .047 0.649FA cingulum �0.462 .010 0.213 �0.592 .001 0.429 .020 0.358 �0.581 .001 0.425 .024 �0.053 .789 0.360FA subcortical orbitofrontal WM �0.579 .001 0.335 �0.641 � .001 0.401 .031 0.442 �0.624 � .001 0.396 .037 �0.267 .169 0.482FA genu CC �0.507 .004 0.257 �0.631 � .001 0.453 .014 0.409 �0.614 .001 0.447 .017 �0.166 .399 0.425FA body CC �0.474 .008 0.224 �0.633 � .001 0.483 .008 0.405 �0.616 � .001 0.482 .009 �0.288 .137 0.455TIQ (normalized for age) 0.056 .778 0.003 �0.233 .243 0.341 .081 0.119 �0.153 .454 0.287 .155 �0.177 .387 0.147Visuomotor accuracy† 0.270 .149 0.073 0.425 .021 �0.342 .070 0.181 0.389 .041 �0.340 .077 0.409 .031 0.318Visuomotor stability† 0.287 .125 0.082 0.410 .027 �0.307 .105 0.169 0.371 .052 �0.304 .116 0.433 .021 0.325Sustained attention work pace† 0.265 .157 0.070 0.146 .448 0.038 .843 0.072 0.158 .421 0.032 .871 �0.077 .697 0.077Visuospatial sequencing† 0.316 .089 0.100 0.322 .088 �0.158 .412 0.123 0.287 .139 �0.144 .465 0.229 .241 0.169FA IFOF, uncinate fasciculus �0.228 .227 0.052 �0.362 .053 0.290 .127 0.131 �0.318 .099 0.300 .121 �0.581 .001 0.425

NOTE. Bold correlation coefficients are indicative of accelerated aging. Linear regression models assessing the effects of age at assessment (AaA), age at diagnosis(AaD), and dosage of cranial radiotherapy (CRT) on fractional anisotropy (FA) in cognitively relevant regions and on the neuropsychological variables. The analysis wasperformed for the CRT group (n � 30), with relapse patients being excluded because of double AaD. The first five dependent variables were selected on the basisof both a significant correlation (family-wise error corrected P � .05) between FA and a neuropsychological deficiency and a significant correlation between FA andAaA. The correlations with AaA are supposedly zero in the normal population, and therefore, the large negative correlations in these regions indicate acceleratedaging. The correlation between FA (or cognitive variables) and AaA is suppressed by the correlation with AaD, as shown in step 2 of the linear regression models.Without controlling for AaD, the accelerated aging effect is underestimated.

Abbreviations: CC, corpus callosum; IFOF, inferior fronto-occipital fasciculus; TIQ, total IQ (estimated based on four subtests); WM, white matter.�FA, mean FA of cluster.†Higher scores mean worse performance.





Furthermore, younger age at treatment with CRT was associatedwith lower FA, mostly within frontal and parietal WM tracts. Thesetracts are known to myelinate at a later age than the rest of the brain.79

This could suggest that CRT affects the cells that create myelin morethan it damages existing myelin. This might leave survivors with alower peak level of WM density in young adulthood. Concurrently,treatment could injure the axons unprotected by myelin.

For nonirradiated survivors, both FA values and neuropsycho-logical performance were lower on average, but not more than onestandard deviation below the mean of controls. Trends for decreasedFA can be shown, confirming findings by Porto et al,14 but impact oncognition seems to be limited. The CT group displays no signs ofaccelerated aging. This suggests that, although acute leukoencepha-lopathy is frequently reported, long-term effects after 20 years aremild.80,81 No dose-effect relationship could be established for MTX.Although outcome of nonirradiated survivors was on average withinthe normal range, there is a small subgroup with below-average FAand related neuropsychological deficiencies. This group should beacknowledged by clinicians. Future research should focus on the iden-tification of this subgroup, identification of the risk factors, and devel-opment of preventive measures.

With older AaD, the risk of relapse and therefore treatmentintensity increase. However, evidence indicates that CRT is moredetrimental at younger AaD. This interaction is difficult to quantifybut important to acknowledge.

In line with previous studies, we observed similar impairments incognitive functioning19-24 and WM integrity in frontal, parietal, andtemporal WM tracts. Porto et al14 reported that WM around thefrontal horns of the lateral ventricles and subcortical frontal WM werethe areas most affected. Dellani et al31 reported decreased FA in thetemporal lobes, the hippocampi, and thalami, in which we founddecreased FA in WM surrounding these structures. We also founddecreased FA in the cingulum, CC, and SLF. Our analysis might havebeen more sensitive because we used a population-based atlas, whichallows for more accurate image registration, and a region of interest–independent analysis. A major strength of this study is the strongerperspective on long-term quality of life than previous studies. We wereable to demonstrate age dependence by disentangling the effects ofAaD and AaA, a method not applied by either Porto14 or Dellani.31 Wealso used larger patient samples and more homogeneous patientgroups, and we excluded patients with pre-existing CNS disorders.The established association between decreased FA and neuropsycho-logical dysfunction in this population is a major contribution to theexisting literature. Evidently, longitudinal research with even larger

groups is necessary to confirm our accelerated aging hypothesis, whichis now based only on cross-sectional data. Alternatively, data fromolder controls, acquired in the same experimental setting, could beused to investigate similarity of FA levels and cognitive status betweenolder adults and cancer survivors to further support the acceleratedaging hypothesis. More research is needed to elucidate the clinicalrelevance of the observed trends of decreased FA after CT only.

Thevariabilityintreatmentregimens,especiallyintermsofchem-otherapeutic agents, might be a limitation of this study. However, thisis inevitable when aiming for larger samples. Dose-effect relationshipsof agents other than MTX should be studied.

In conclusion, this study stresses the importance of followingcohorts many decades after neurotoxic treatment in childhood, pref-erably throughout life. The growing support for the concept of accel-erated aging after CRT implicates screening for early-onset dementia.Recommending lifestyle modifications that are implicated in slowingthe progression of dementia, such as not smoking and getting regularphysical exercise, could be considered.82 Although detrimental effectsof CT on WM and neuropsychological function are not completelyabsent, they are mild compared with CRT, although CT is equallyeffective in terms of survival and recurrence rates after ALL.3 Thiswarrants a recommendation to use CRT only as a last resort.

AUTHORS’ DISCLOSURES OF POTENTIAL CONFLICTSOF INTEREST

The author(s) indicated no potential conflicts of interest.

AUTHOR CONTRIBUTIONS

Conception and design: Ilse Schuitema, Frederik Barkhof, Eline vanDulmen-den Broeder, Cor van den Bos, Anjo J.P. Veerman, Leo M.J.de SonnevilleAdministrative support: Ilse Schuitema, Marita Daams, AnneUyttebroeck, Frederik BarkhofProvision of study materials or patients: Anne Uyttebroeck, FrederikBarkhof, Cor van den Bos, Anjo J.P. Veerman, Leo M.J. de SonnevilleCollection and assembly of data: Ilse Schuitema, Marita Daams, AnneUyttebroeck, Frederik Barkhof, Helena J. van der Pal, Cor van den BosData analysis and interpretation: Ilse Schuitema, Sabine Deprez, WimVan Hecke, Marita Daams, Stefan Sunaert, Frederik Barkhof, Leo M.J.de SonnevilleManuscript writing: All authorsFinal approval of manuscript: All authors

REFERENCES

1. Cole PD, Kamen BA: Delayed neurotoxicityassociated with therapy for children with acutelymphoblastic leukemia. Ment Retard Dev DisabilRes Rev 12:174-183, 2006

2. Veerman AJ, Hahlen K, Kamps WA, et al: Highcure rate with a moderately intensive treatmentregimen in non-high-risk childhood acute lympho-blastic leukemia: Results of protocol ALL VI from theDutch Childhood Leukemia Study Group. J ClinOncol 14:911-918, 1996

3. Veerman AJ, Kamps WA, van den Berg H, et al:Dexamethasone-based therapy for childhood acute lym-phoblastic leukaemia: Results of the prospective Dutch

Childhood Oncology Group (DCOG) protocol ALL-9(1997-2004). Lancet Oncol 10:957-966, 2009

4. Pui CH, Campana D, Pei D, et al: Treatingchildhood acute lymphoblastic leukemia withoutcranial irradiation. N Engl J Med 360:2730-2741,2009

5. Jenney ME, Levitt GA: The quality of survivalafter childhood cancer. Eur J Cancer 38:1241-1250,2002

6. Reddick WE, Shan ZY, Glass JO, et al: Smallerwhite-matter volumes are associated with largerdeficits in attention and learning among long-termsurvivors of acute lymphoblastic leukemia. Cancer106:941-949, 2006

7. Carey ME, Haut MW, Reminger SL, et al: Re-duced frontal white matter volume in long-term child-

hood leukemia survivors: A voxel-based morphometrystudy. Am J Neuroradiol 29:792-797, 2008

8. Brown RT, Madan-Swain A, Walco GA, et al:Cognitive and academic late effects among childrenpreviously treated for acute lymphocytic leukemiareceiving chemotherapy as CNS prophylaxis. J Pe-diatr Psychol 23:333-340, 1998

9. Buizer AI, de Sonneville LM, Veerman AJ:Effects of chemotherapy on neurocognitive functionin children with acute lymphoblastic leukemia: Acritical review of the literature. Pediatr Blood Cancer52:447-454, 2009

10. Kingma A, Van Dommelen RI, Mooyaart EL,et al: No major cognitive impairment in youngchildren with acute lymphoblastic leukemia usingchemotherapy only: A prospective longitudinal

Schuitema et al




study. J Pediatr Hematol Oncol 24:106-114,2002

11. Copeland DR, Moore BD 3rd, Francis DJ, et al:Neuropsychologic effects of chemotherapy on chil-dren with cancer: A longitudinal study. J Clin Oncol14:2826-2835, 1996

12. Mennes M, Stiers P, Vandenbussche E, et al:Attention and information processing in survivors ofchildhood acute lymphoblastic leukemia treatedwith chemotherapy only. Pediatr Blood Cancer 44:478-486k 2005

13. Buizer AI, de Sonneville LM, van den Heuvel-Eibrink MM, et al: Chemotherapy and attentionaldysfunction in survivors of childhood acute lympho-blastic leukemia: Effect of treatment intensity. Pe-diatr Blood Cancer 45:281-290, 2005

14. Porto L, Preibisch C, Hattingen E, et al: Voxel-based morphometry and diffusion-tensor MR imag-ing of the brain in long-term survivors of childhoodleukemia. Eur Radiol 18:2691-2700, 2008

15. Koppelmans V, Breteler MM, Boogerd W, etal: Neuropsychological performance in survivors ofbreast cancer more than 20 years after adjuvantchemotherapy. J Clin Oncol 30:1080-1086, 2012

16. Ball WS Jr, Prenger EC, Ballard ET: Neurotox-icity of radio/chemotherapy in children: Pathologicand MR correlation. Am J Neuroradiol 13:761-776,1992

17. Jain R, Robertson PL, Gandhi D, et al:Radiation-induced cavernomas of the brain. Am JNeuroradiol 26:1158-1162, 2005

18. Hijiya N, Hudson MM, Lensing S, et al: Cumu-lative incidence of secondary neoplasms as a firstevent after childhood acute lymphoblastic leukemia.JAMA 297:1207-1215, 2007

19. Brouwers P, Poplack D: Memory and learningsequelae in long-term survivors of acute lympho-blastic leukemia: Association with attention deficits.Am J Pediatr Hematol Oncol 12:174-181, 1990

20. Butler RW, Copeland DR: Neuropsychologicaleffects of central nervous system prophylactic treat-ment in childhood leukemia: Methodological consid-erations. J Pediatr Psychol 18:319-338, 1993

21. Butler RW, Copeland DR: Attentional pro-cesses and their remediation in children treated forcancer: A literature review and the development of atherapeutic approach. J Int Neuropsychol Soc 8:115-124, 2002

22. Precourt S, Robaey P, Lamothe I, et al: Verbalcognitive functioning and learning in girls treated foracute lymphoblastic leukemia by chemotherapywith or without cranial irradiation. Dev Neuropsychol21:173-195, 2002

23. Mulhern RK, Palmer SL: Neurocognitive lateeffects in pediatric cancer. Curr Probl Cancer 27:177-197, 2003

24. Butler JM, Rapp SR, Shaw EG: Managing thecognitive effects of brain tumor radiation therapy.Curr Treat Options Oncol 7:517-523, 2006

25. Pierpaoli C, Basser PJ: Toward a quantitativeassessment of diffusion anisotropy. Magn ResonMed 36:893-906, 1996

26. Leung LH, Ooi GC, Kwong DL, et al: White-matter diffusion anisotropy after chemo-irradiation:A statistical parametric mapping study and histo-gram analysis. Neuroimage 21:261-268, 2004

27. Khong PL, Leung LH, Fung AS, et al: Whitematter anisotropy in post-treatment childhood can-cer survivors: Preliminary evidence of associationwith neurocognitive function. J Clin Oncol 24:884-890, 2006

28. de Ruiter MB, Reneman L, Boogerd W, et al:Late effects of high-dose chemotherapy on whiteand gray matter in breast cancer survivors: Converg-

ing results from multimodal magnetic resonanceimaging. Hum Brain Mapp 33:2971-2983, 2012

29. Deprez S, Amant F, Yigit R, et al: Chemother-apy-induced structural changes in cerebral whitematter and its correlation with impaired cognitivefunctioning in breast cancer patients. Hum BrainMapp 32:480-493, 2011

30. Deprez S, Amant F, Smeets A, et al: Longitu-dinal assessment of chemotherapy-induced struc-tural changes in cerebral white matter and itscorrelation with impaired cognitive functioning.J Clin Oncol 30:274-281, 2012

31. Dellani PR, Eder S, Gawehn J, et al: Latestructural alterations of cerebral white matter inlong-term survivors of childhood leukemia. J MagnReson Imaging 27:1250-1255, 2008

32. Davis PC, Hoffman JC Jr, Pearl GS, et al: CTevaluation of effects of cranial radiation therapy inchildren. AJR Am J Roentgenol 147:587-592, 1986

33. Suc E, Kalifa C, Brauner R, et al: Brain tumoursunder the age of three: The price of survival—Aretrospective study of 20 long-term survivors. ActaNeurochir (Wien) 106:93-98, 1990

34. Mulhern RK, Palmer SL, Reddick WE, et al:Risks of young age for selected neurocognitivedeficits in medulloblastoma are associated withwhite matter loss. J Clin Oncol 19:472-479, 2001

35. Spiegler BJ, Bouffet E, Greenberg ML, et al:Change in neurocognitive functioning after treat-ment with cranial radiation in childhood. J Clin Oncol22:706-713, 2004

36. Chan YL, Roebuck DJ, Yuen MP, et al: Long-term cerebral metabolite changes on proton mag-netic resonance spectroscopy in patients cured ofacute lymphoblastic leukemia with previous intra-thecal methotrexate and cranial irradiation prophy-laxis. Int J Radiat Oncol Biol Phys 50:759-763, 2001

37. Khong PL, Kwong DL, Chan GC, et al:Diffusion-tensor imaging for the detection and quan-tification of treatment-induced white matter injury inchildren with medulloblastoma: A pilot study. AJNRAm J Neuroradiol 24:734-740, 2003

38. Bleyer WA: Neurologic sequelae of metho-trexate and ionizing radiation: A new classification.Cancer Treat Rep 65:89-98, 1981

39. D’Ambrosio DJ, Cohen RB, Glass J, et al:Unexpected dementia following prophylactic cranialirradiation for small cell lung cancer: Case report.J Neurooncol 85:77-79, 2007

40. Einhorn LH 3rd: The case against prophylacticcranial irradiation in limited small cell lung cancer.Semin Radiat Oncol 5:57-60, 1995

41. Pedersen AG, Kristjansen PE, Hansen HH:Prophylactic cranial irradiation and small cell lungcancer. Cancer Treat Rev 15:85-103, 1988

42. Edelstein K, Spiegler BJ, Fung S, et al: Earlyaging in adult survivors of childhood medulloblas-toma: Long-term neurocognitive, functional, andphysical outcomes. Neuro Oncol 13:536-545, 2011

43. Maccormick RE: Possible acceleration of ag-ing by adjuvant chemotherapy: A cause of earlyonset frailty? Med Hypotheses 67:212-215, 2006

44. Riehm H, Gadner H, Henze G, et al: Resultsand significance of six randomized trials in fourconsecutive ALL-BFM studies. Haematol BloodTransfus 33:439-450, 1990

45. Vilmer E, Suciu S, Ferster A, et al: Long-termresults of three randomized trials (58831, 58832,58881) in childhood lymphoblastic leukemia: ACLCG-EORTC report—Children Leukemia Coopera-tive Group. Leukemia 14:2257-2266, 2000

46. Reese TG, Heid O, Weisskoff RM, et al:Reduction of eddy-current-induced distortion in dif-

fusion MRI using a twice-refocused spin echo.Magn Reson Med 49:177-182, 2003

47. Leemans A, Jones DK: The B-matrix must berotated when correcting for subject motion in DTIdata. Magn Reson Med 61:1336-1349, 2009

48. Van Hecke W, Sijbers J, D’Agostino E, et al:On the construction of an inter-subject diffusiontensor magnetic resonance atlas of the healthyhuman brain. Neuroimage 43:69-80, 2008

49. Van Hecke W, Leemans A, Sage CA, et al: Theeffect of template selection on diffusion tensorvoxel-based analysis results. Neuroimage 55:566-573, 2011

50. Van Hecke W, Leemans A, De Backer S, et al:Comparing isotropic and anisotropic smoothing forvoxel-based DTI analyses: A simulation study. HumBrain Mapp 31:98-114, 2010

51. Ashburner J, Friston KJ: Statistical ParametricMapping (SPM). http://www.fil.ion.ucl.ac.uk/spm/software/spm8

52. De Sonneville LMJ: Amsterdam Neuropsy-chological Tasks: A computer-aided assessmentprogram, in Den Brinker BPLM, Beek PJ, Brand AN,et al (eds): Cognitive Ergonomics, Clinical Assess-ment, and Computer-Assisted Learning. Lisse, theNetherlands, Swets & Zeitlinger, 1999, pp 187-203

53. Burgard P, Rey F, Rupp A, et al: Neuropsy-chologic functions of early treated patients withphenylketonuria, on and off diet: Results of across-national and cross-sectional study. PediatrRes 41:368-374, 1997

54. Huijbregts S, de Sonneville L, Licht R, et al:Inhibition of prepotent responding and attentionalflexibility in treated phenylketonuria. Dev Neuropsy-chol 22:481-499, 2002

55. Huijbregts SC, de Sonneville LM, Licht R, etal: Sustained attention and inhibition of cognitiveinterference in treated phenylketonuria: Associa-tions with concurrent and lifetime phenylalanineconcentrations. Neuropsychologia 40:7-15, 2002

56. De Sonneville LM, Boringa JB, Reuling IE, etal: Information processing characteristics in sub-types of multiple sclerosis. Neuropsychologia 40:1751-1765, 2002

57. Rowbotham I, Pit-ten Cate IM, Sonuga-BarkeEJ, et al: Cognitive control in adolescents withneurofibromatosis type 1. Neuropsychology 23:50-60, 2009

58. Huijbregts S, Swaab H, de Sonneville L:Cognitive and motor control in neurofibromatosistype I: Influence of maturation and hyperactivity-inattention. Dev Neuropsychol 35:737-751, 2010

59. Gunther T, Herpertz-Dahlmann B, Konrad K:[Reliability of attention and verbal memory testswith normal children and adolescents: Clinical impli-cations]. [In German]. Z Kinder JugendpsychiatrdPsychother 33:169-179, 2005

60. De Sonneville LM: [Amsterdam Neuropsycho-logical tasks: Scientific and clinical applications]. [InDutch]. Tijdschr Neuropsychol 0:27-41, 2005

61. Silverstein AB: Two-subtest and four-subtestshort forms of the Wechsler adult intelligence scale-revised. J Consult Clin Psychol 50:415-418, 1982

62. Reference deleted63. Westlye LT, Walhovd KB, Dale AM, et al:

Life-span changes of the human brain white matter:Diffusion tensor imaging (DTI) and volumetry. CerebCortex 20:2055-2068, 2010

64. Neter J, Kutner MH, Wasserman W, et al:Applied Linear Statistical Models (ed 4). New York,NY, WCB McGraw-Hill, 1996





http://www.fil.ion.ucl.ac.uk/spm/software/spm8

http://www.fil.ion.ucl.ac.uk/spm/software/spm8

65. Mulhern RK, Kepner JL, Thomas PR, et al:Neuropsychologic functioning of survivors of child-hood medulloblastoma randomized to receive con-ventional or reduced dose cranospinal irradiation: APediatric Oncology Group study. J Clin Oncol 16:1723-1728, 1998

66. Parente DB, Gasparetto EL, da Cruz LC Jr, etal: Potential role of diffusion tensor MRI in thedifferential diagnosis of mild cognitive impairmentand Alzheimer’s disease. AJR Am J Roentgenol190:1369-1374, 2008

67. Daams M, Schuitema I, van Dijk BW, et al: Long-term effects of cranial irradiation and intrathecal chemo-therapy in treatment of childhood leukemia: A MEGstudy of power spectrum and correlated cognitive dys-function. BMC Neurol 12:84, 2012

68. O’Sullivan M, Morris RG, Huckstep B, et al:Diffusion tensor MRI correlates with executive dys-function in patients with ischaemic leukoaraiosis.J Neurol Neurosurg Psychiatry 75:441-447, 2004

69. Turken A, Whitfield-Gabrieli S, Bammer R, etal: Cognitive processing speed and the structure ofwhite matter pathways: Convergent evidence from

normal variation and lesion studies. Neuroimage42:1032-1044, 2008

70. Kurita H, Kawahara N, Asai A, et al: Radiation-induced apoptosis of oligodendrocytes in the adultrat brain. Neurol Res 23:869-874, 2001

71. Smith B: Brain damage after intrathecal metho-trexate. J Neurol Neurosurg Psychiatr 38:810-815, 1975

72. Schultheiss TE, Kun LE, Ang KK, et al: Radia-tion response of the central nervous system. Int JRadiat Oncol Biol Phys 31:1093-1112, 1995

73. Tofilon PJ, Fike JR: The radioresponse of thecentral nervous system: A dynamic process. RadiatRes 153:357-370, 2000

74. Monje ML, Mizumatsu S, Fike JR, et al:Irradiation induces neural precursor-cell dysfunction.Nat Med 8:955-962, 2002

75. Fukuda A, Fukuda H, Swanpalmer J, et al: Age-dependent sensitivity of the developing brain to irradia-tion is correlated with the number and vulnerability ofprogenitor cells. J Neurochem 92:569-584, 2005

76. Dietrich J, Monje M, Wefel J, et al: Clinicalpatterns and biological correlates of cognitive dys-function associated with cancer therapy. Oncologist13:1285-1295, 2008

77. Han R, Yang YM, Dietrich J, et al: Systemic5-fluorouracil treatment causes a syndrome of de-layed myelin destruction in the central nervoussystem. J Biol 7:12, 2008

78. Dietrich J, Han R, Yang Y, et al: CNS progen-itor cells and oligodendrocytes are targets of chem-otherapeutic agents in vitro and in vivo. J Biol 5:22,2006

79. Tau GZ, Peterson BS: Normal development ofbrain circuits. Neuropsychopharmacology 35:147-168, 2010

80. Shuper A, Stark B, Kornreich L, et al:Methotrexate-related neurotoxicity in the treatmentof childhood acute lymphoblastic leukemia. Isr MedAssoc J 4:1050-1053, 2002

81. Reddick WE, Glass JO, Johnson DP, et al:Voxel-based analysis of T2 hyperintensities in whitematter during treatment of childhood leukemia.Am J Neuroradiol 30:1947-1954, 2009

82. Mangialasche F, Kivipelto M, Solomon A, etal: Dementia prevention: Current epidemiologicalevidence and future perspective. Alzheimers ResTher 4:6, 2012

■ ■ ■

Schuitema et al




Acknowledgment

We thank T. Schweigmann and G. de Vos (Vrije Universiteit University Medical Center); H.J.A. Smaling and P.M. Kroonenberg (LeidenUniversity); S. van Gool, D. Lemmens, and A. Michiels (Katholieke Universiteit Leuven); and R. Heinen and M. van Overveld (Academic

Medical Center) for their valuable support in collecting and interpreting the data.

Appendix

The Amsterdam Neuropsychological Tasks Program: Descriptions of Subtests

The computerized Amsterdam Neuropsychological Tasks (ANT) provides standardized assessments and automated recordings ofspeed and accuracy of information processing, attention processes, and working memory.52 The program has proved to be helpful indefining neuropsychological deficit profiles in various clinical domains associated with generally diffuse impact on the brain andparticularly in middle-late effects of childhood acute lymphoblastic leukemia (ALL).12,13,53-58 On the basis of these studies, tasksevaluating baseline response speed, pattern recognition, sustained attention (work pace and attentional fluctuations), cognitive flexibility(set shifting and inhibition), visuomotor skills, and visuospatial sequential working memory were selected for assessment of this study’spopulation. The reliability and validity of these tasks are excellent.59,60 The following are task descriptions.

Baseline speed. A simple reaction time task in which cognitive demands are restricted to the mere detection of a stimulus.Provides a reference level for response speed. Parameters: T_bs (mean response time of left and right hand), S_bs (standarddeviation of response times).

Feature identification. To test speed and accuracy of processing complex abstract visuospatial patterns. Task demands includemaintenance and manipulation of working memory representations. Parameters: Ts_fi (mean response time in similar condition, inwhich target is surrounded by similar patterns and distinction is based on detailed information processing), Td_fi (mean response time indissimilar condition, in which target is surrounded by dissimilar patterns and distinction is based on more global and simple informationprocessing), PES_fi (percentage of errors in similar condition), PED_fi (percentage of errors in dissimilar condition).

Memory search objects. Patients have to detect a predefined target set in a signal of four two-dimensional symbols (red, green, blue,yellow/circle, triangle, cross, square). Memory load is increased across two task parts. Part 2 requires continuous monitoring and updatingof the contents of the working memory. Parameters: T1_2d � (th1_2d � tc1_2d)/2 (mean of response times hits and correct rejectionsin part 1), T2_2d � (th2_2d � tc2_2d)/2 (mean of response times hits and correct rejections in part 2), Ne1_2d (mean number of errorspart 1), Ne2_2d (mean number of errors part 2).

Sustained attention. To evaluate changes and fluctuations in speed and accuracy of processing over time. The paradigm induces aresponse bias providing indices for response inhibition and behavioral adaptation to feedback. Parameters: T_sa (mean response time),S_sa (standard deviation of response time, a measure for fluctuations of attention), PM_sa (percentage of misses; answer is no althoughit should be yes), PF_sa (percentage of false alarms; answer is yes although it should be no). Note: T_sa is referred to in the main text assustained attention work pace.

Shifting attentional set. To evaluate inhibition of prepotent responses and attentional flexibility. In the signal—a horizontal bar—acolored square may jump from left to right or vice versa. Depending on the color of the square, the patient should execute a compatibleresponse (part 1), that is, press right (left) key when square jumped to the right (left), or is required to execute an incompatible response(press opposite keys, part 2), In part 3, trials of part 1 and 2 are randomly mixed, which requires a switch between the two types of responsesets. Parameters: T_inhib (mean response time inhibition: difference between condition 2 and 1), T_flex (mean response time flexibility:difference between condition 3 and 1), P_inhib (percentage of errors on inhibition), P_flex (percentage of errors on flexibility).

Pursuit. This task evaluates the quality of visuomotor control. The patient has to track a small star, which continuously moves acrossthe screen in random directions. The task requires concurrent planning and execution of unpredictable movements. Parameters: D_pu(mean distance to target), S_pu (standard deviation of D_pu). Note: D_pu is referred to in the main text as visuomotor accuracy; S_pu isreferred to as visuomotor stability.

Tracking. This task serves the same purpose as the pursuit task, but now the patient is asked to execute planned, more automatedmovements. The patient has to draw a circle by moving the mouse cursor in between two large concentric circles on the screen. This taskrequires less controlled processing than the pursuit task. Parameters: Da_tr (mean absolute distance to target), S_tr (standard deviation ofDa_tr). Diff_D_pu_tr (difference of D_pu and Da_tr) and Diff_S_pu_tr (difference of S_pu and S_tr) are measures of working memory.

Visuospatial sequencing. This task evaluates memory of visuospatial temporal patterns. In each trial, several circles are pointed out inan array of nine circles, arranged in a 3 � 3 matrix on the computer screen. The patient has to point out the same circles in the same orderby moving the mouse cursor and must press a button when the cursor is positioned at the right location(s). The test consists of 24 trials inwhich the number of target circles varies from 3 to 7 and in which the spatial sequential patterns increase gradually in complexity.Parameters: Nit_vs (number of correctly identified circles), Nitco_vs (number of correctly identified circles in the correct order). Diff_vs(Nit_vs – Nitco_vs) is a measure of the sequential working memory component. Note: where visuospatial sequencing is mentioned in themain text, it refers to the parameter Diff_vs.


www.jco.org © 2013 by American Society of Clinical Oncology



Standard-risk CRT(n = 78)

High-risk CRT(n = 28)

Relapsed CRT(n = 29)

Participated(n = 26)

Participated(n = 8)

Included(n = 7) Excluded

(n = 0)Included(n = 13)

Invited by letter(n = 58)


Excluded for medical reasons Brain tumor Trisomy 21 Retardation NOS Sotos' syndrome Antidepressants Additional type of cancer Meningeomas Cavernous hemangeoma Head trauma Epilepsy Pregnant

(n = 20)(n = 1)(n = 1)(n = 4)(n = 1)(n = 1)(n = 2)(n = 6)(n = 1)(n = 1)(n = 1)(n = 1)

Excluded for medical reasons Retardation NOS Cavernous hemangeoma Epilepsy

(n = 3)(n = 1)(n = 1)(n = 1)


Excluded for medical reasons Meningeomas Additional types of cancer Pregnant

(n = 5)(n = 2)(n = 2)(n = 1)

Included Meningeoma not pressing on the cerebrum

(n = 24)(n = 1)

Did not participate Untraceable Declined Only neuropsy. tests

(n = 32)(n = 14)(n = 14)(n = 4)


(n = 17)(n = 6)

(n = 10)(n = 1)



(n = 11)(n = 2)(n = 5)(n = 4)

Excluded Meningeomas

(n = 2)(n = 2) Excluded

Meningeomas(n = 1)(n = 1)

Standard-risk CT(n = 104)

High-risk CT(n = 46)

Relapsed CT, not analyzed(n = 6)



Included(n = 20)

Excluded(n = 0)

Excluded(n = 0)

Included(n = 29)

Included(n = 2)



Excluded for medical reasons Von Recklinghausen Trisomy 21 Retardation NOS Turner's syndrome Antidepressants Anxiolytics Bardet-Biedl's syndrome Schizophrenia History of bacterial meningitis Pregnant

(n = 11)(n = 1)(n = 1)(n = 2)(n = 1)(n = 1)(n = 1)(n = 1)(n = 1)(n = 1)(n = 1)

Excluded for medical reasons Pregnant Developmental disorder NOS Antidepressants

(n = 3)(n = 1)(n = 1)(n = 1)


Excluded for medical reasons Trisomy 21

(n = 1)(n = 1)

Did not participate Untraceable Declined Only neuropsy. tests Moved abroad

(n = 64)(n = 27)(n = 26)(n = 8)(n = 3)

Did not participate Untraceable Declined Only neuropsy. tests Moved abroad

(n = 23)(n = 10)(n = 7)(n = 1)(n = 5)

Participated(n = 3)

Did not participate Untraceable

(n = 2)(n = 2)

Excluded Cortical infarct

(n = 1)(n = 1)

Fig A1. Flow diagram of survivor selection. CRT, cranial radiotherapy; neuropsy., neuropsychological; NOS, not otherwise specified.

Schuitema et al

© 2013 by American Society of Clinical Oncology JOURNAL OF CLINICAL ONCOLOGY



Accelerated Aging, Decreased White Matter Integrity, and Associated Neuropsychological Dysfunction...

Documents

Transcript of Accelerated Aging, Decreased White Matter Integrity, and Associated Neuropsychological Dysfunction...