R53
Jou nal o Radiological P o ec ion
Re iew
Childhood leukaemia isks: om unexplained
indings nea nuclea ins alla ions o
ecommenda ions o u u e esea ch
DLau ie 1,12, B G osche2, A Au inen3, JCla el4, CCobaleda5,
ADehos2, SHo nha d 2, SJacob1, PKaa sch6, OKos i7,
CKuehni8, TLigh oo 9, BSpyche 8, AVan Nieuwenhuyse10,
RWake o d11 and GZiegelbe ge 2,*
1 Ins i u de Radiop o ec ion e de Sû e é Nucléai e (IRSN), BP 17, F-92262
Fon enay-aux-Roses Cedex, F ance
2 Fede al O ice o Radia ion P o ec ion (B S), Ingols aed e Lands . 1, D-85764
Neuhe be g, Ge many
3 Radia ion and Nuclea Sa e y Au ho i y (STUK), Helsinki/Uni e si y o Tampe e,
School o Heal h Sciences, FI-33014 Tampe e, Finland
4 CESP-Equipe d’Epidémiologie en i onnemen ale des cance s INSERM UMR-S 1018,
Uni e si é Pa is-Sud, 16 a enue Paul Vaillan -Cou u ie , 94807 Villejui Cedex, F ance
5 Cen o de Biología Molecula Se e o Ochoa, CSIC/UAM, Nicolas Cab e a 1, 28049
Mad id, Spain
6 Ge man Childhood Cance Regis y (GCCR), Ins i u e o Medical Bios a is ics, Epi-
demiology, and In o ma ics, Uni e si y Medical Cen e Mainz, 55101 Mainz, Ge many
7 Na ional Academy o Sciences, 500 Fi h S ee , NW, Washing on, DC 20001, USA
8 Ins i u e o Social and P e en i e Medicine, Uni e si y o Be n, Finkenhubelweg
11-3012 Be ne, Swi ze land
9 ECSG, Depa men o Heal h Sciences, Uni e si y o Yo k, Seebohm Rown ee
Building, Hesling on, Yo k YO10 5DD, UK
10 Scien i ic Ins i u e o Public Heal h WIV-ISP, Julie e Wy smans aa 14, 1050
B ussels, Belgium
11 Dal on Nuclea Ins i u e, The Uni e si y o Manches e , Pa ise Building—G Floo ,
Sack ille S ee , Manches e M13 9PL, UK
E-mail: [email p o ec ed], [email p o ec ed], [email p o ec ed],
jacqueline.cla[email p o ec ed], [email p o ec ed], pe e .kaa sch@unimedizin-mainz.
de, [email p o ec ed], [email p o ec ed], T acy[email p o ec ed],
an. annieuwenhuyse@wi -isp.be and Richa d.W[email p o ec ed]
Recei ed 19 Decembe 2013, e ised 20 Feb ua y 2014
Accep ed o publica ion 3 Ap il 2014
Published 18 June 2014
Re iew
P in ed in he UK
R53
jRP
© 2014 IOP Publishing L d
2014
34
j. Radiol. P o .
jRP
10.1088/0952-4746/34/3/R53
REV
jou nal o Radiological P o ec ion
Socie y o Radiological P o ec ion
JRP-100042.R1
12 Au ho o whom any co espondence should be add essed.
* On behal o all he Wo kshop pa icipan s (lis in appendix).
0952-4746/14/030R53+16$33.00 © 2014 IOP Publishing L d P in ed in he UK
J. Radiol. P o . 34 (2014) R53–R68 doi:10.1088/0952-4746/34/3/R53
Con en om his wo k may be used unde he e ms o he C ea i e Commons A ibu ion 3.0
licence. Any u he dis ibu ion o his wo k mus main ain a ibu ion o he au ho (s) and he i le
o he wo k, jou nal ci a ion and DOI.
Re iew
R54
J. Radiol. P o . 34 (2014) R53
Abs ac
Recen indings ela ed o childhood leukaemia incidence nea nuclea
ins alla ions ha e aised ques ions which can be answe ed nei he by cu en
knowledge on adia ion isk no by o he es ablished isk ac o s. In 2012,
a wo kshop was o ganised on his opic wi h wo objec i es: (a) e iew o
esul s and discussion o me hodological limi a ions o s udies nea nuclea
ins alla ions; (b) iden i ica ion o di ec ions o u u e esea ch in o he causes and
pa hogenesis o childhood leukaemia. The wo kshop ga he ed 42pa icipan s
om di e en disciplines, ex ending widely ou side o he adia ion p o ec ion
ield. Rega ding he p oximi y o nuclea ins alla ions, he need o con inuous
su eillance o childhood leukaemia incidence was highligh ed, including a
be e cha ac e isa ion o he local popula ion. Thec ea ion o collabo a i e
wo king g oups was ecommended o consis ency in me hodologies and he
possibili y o combining da a o u u e analyses. Rega ding he causes o
childhood leukaemia, majo ields o esea ch we e discussed (en i onmen al
isk ac o s, gene ics, in ec ions, immuni y, s em cells, expe imen al esea ch).
The need o mul idisciplina y collabo a ion in de eloping esea ch ac i i ies
was unde lined, including he p e alence o po en ial p edisposi ion ma ke s
and in es iga ing u he he in ec ious ae iology hypo hesis. Animal s udies
and gene ic/epigene ic app oaches appea o g ea in e es . Rou es o u u e
esea ch we e poin ed ou .
Keywo ds: childhood, leukaemia, isk, nuclea ins alla ions, ae iology,
epidemiology
1. In oduc ion
Leukaemia is he commones ype o childhood malignancy, accoun ing o ~30% o all
diagnoses in child en less han 15 yea s o age in economically de eloped egions o he
wo ld. The he e ogeneous na u e o leukaemia is well es ablished. In childhood, he majo
sub ypes a e acu e lymphoblas ic leukaemia (ALL) and acu e myeloid leukaemia (AML) wi h
ALL occu ing app oxima ely i e imes mo e equen ly han AML [1, 2]; ch onic myeloid
leukaemia (CML) also occu s a ely. Acu e leukaemia can be u he classi ied in o dis inc
sub ypes acco ding o cell lineage, s age o immunopheno ypic ma u a ion and speci ic molec-
ula gene ic abno mali ies [3]. O hese, B-cell-p ecu so ALL is he commones sub ype. In
con as o o he sub ypes like T-cell ALL o AML, i s incidence has a dis inc age peak in
child en aged wo o i e yea s [4].
Many ac o s ha e been p oposed o be causa i e o childhood leukaemia [5–7]. Howe e ,
o da e, he only es ablished isk ac o s o childhood leukaemia a e Down synd ome, sex
(wi h boys mo e o en a ec ed han gi ls, wi h a a io ~1.2 : 1), chemo he apeu ic d ugs and
acu e exposu e o ionising adia ion a mode a e and high doses (abo e 100 mS ). Se e al
s udies sugges ed ha isk may exis a lowe adia ion doses and dose a es, and ha his
isk is compa ible wi h cu en p edic ions [8–11]. In 1990, he hypo hesis o he edi a y
adia ion e ec s on childhood leukaemia isk was p oposed by Ga dne e al o explain
he Sella ield clus e [12], bu consecu i e nega i e s udies lead o he ejec ion o his
hypo hesis [13, 14]. Exposu e o 50 Hz elec ic and magne ic ields (ELF-EMF) is also
sugges ed as a isk ac o o childhood leukaemia [15], bu he e is as ye no biological
e idence suppo ing he epidemiological e idence [16]. The e is consis en e idence om
Re iew
R55
J. Radiol. P o . 34 (2014) R53
epidemiological s udies ha high bi h weigh is a de e minan o disease isk [17–19]. The
ole o exposu e o in ec ious agen s and immune unc ion in ela ion o isk o childhood
ALL is also unde in es iga ion [20, 21], bu he mechanisms by which hese ac o s could
play a ole a e s ill no clea .
Li ing nea a nuclea acili y became a suspec ed isk ac o o childhood leukaemia
a e he 1984 epo on an inc eased leukaemia isk in he illage o Seascale close o he
B i ish nuclea uel ep ocessing plan o Sella ield [22]. Since hen, a numbe o s udies
ha e in es iga ed he leukaemia isk nea nuclea acili ies, including nuclea powe plan s
(NPPs), and se e al e iews we e published [23–25]. Be ween 2008 and 2011, se e al mee -
ings we e held and se e al commi ees we e con ened o e iew exis ing esul s o cance
isks a ound nuclea acili ies and p o ide ecommenda ions, e.g. in Ge many [7, 26], in
F ance [27], in Sweden [28] and in he UK [29]. Based on he ecommenda ions om hese
mee ings, he Ge man Fede al O ice o Radia ion P o ec ion (B S) and he F ench Ins i u e
o Radiological P o ec ion and Nuclea Sa e y (IRSN) decided o o ganise a ocused wo k-
shop on he opic o u u e esea ch on childhood leukaemia, b inging oge he esea che s
om a wide ange o ela ed disciplines. This wo kshop was o ganised unde he auspices o
he Mul idisciplina y Eu opean Low Dose Ini ia i e (MELODI, www.melodi-online.eu) and
was held in F ance in June 2012. The Wo kshop had wo aims: (a) o e iew he la es esul s
and o discuss me hodological limi a ions o s udies o childhood leukaemia incidence close
o nuclea ins alla ions; and (b) o iden i y new di ec ions o u u e esea ch in o he causes
and pa hogenesis o leukaemia in child en. Thus, he Wo kshop did no jus co e he e ec
o adia ion- ela ed leukaemia isk, bu also o he majo known o suspec ed isk ac o s o
childhood leukaemia.
This pape summa ises he esul s o he Wo kshop. I s i s pa deals wi h he issue o
childhood leukaemia isk nea nuclea ins alla ions and ocuses on a e iew o a ailable ind-
ings, he design o epidemiological s udies and sugges ed imp o emen s o he u u e. The
second pa conce ns he ae iology o childhood leukaemia and he unde lying disease biology
(B-cell de elopmen and haema opoie ic s em cells (HSCs)), en i onmen al, in ec ious and
gene ic isk ac o s and he ele ance o animal models. Recommenda ions a e hen de ailed.
2. S udies o childhood leukaemia isk nea nuclea ins alla ions
A majo aspec o he Wo kshop was o discuss he cu en hypo hesis ega ding an excess
isk o childhood leukaemia nea nuclea ins alla ions. Pa icipan s p o ided an o e iew o
s udies unde aken in Eu ope in ecen yea s and o hose planned in he Uni ed S a es. Special
a en ion was paid o leukaemia isk in he younges age g oup. Due o he di e si y o epi-
demiological designs and analysis me hods, me hodological aspec s o ecen s udies we e
conside ed.
2.1. Re iew o cu en knowledge abou he isk o childhood leukaemia nea
nuclea ins alla ions
Based on a combina ion o se e al c i e ia (s a is ical signi icance, alidi y o he me hodol-
ogy, con i ma ion by di e en au ho s using di e en app oaches and pe sis ence o e ime),
he Wo kshop pa icipan s ag eed ha he e is con incing e idence o h ee con i med clus-
e s o childhood leukaemia nea nuclea ins alla ions: hese a e nea he Sella ield ep ocess-
ing plan in Cumb ia, England; he Doun eay ep ocessing plan in No he n Sco land; and
he K uemmel NPP in No he n Ge many [23]. Howe e , leukaemia clus e s also ha e been
Re iew
R56
J. Radiol. P o . 34 (2014) R53
epo ed elsewhe e and a e no speci ic o nuclea ins alla ions. Fo example, he Fallon clus e
in Chu chill Coun y, Ne ada, is a s iking example o a leukaemia clus e [30]: 14 child en
we e diagnosed wi h ALL du ing he pe iod 1997–2003 while only h ee o ou cases we e
expec ed. Despi e he comp ehensi e public heal h in es iga ion launched by he US Cen e s
o Disease Con ol and P e en ion (CDC) he cause o he clus e emains unknown.
Se e al e iews o childhood leukaemia isks nea nuclea ins alla ions al eady exis
[23–25,31, 32], as well as a me a-analysis [33], which has been c i icised due o doub s o e
he alidi y o he selec ion p ocess and po en ial publica ion bias [29, 34, 35]. The e iew o
esul s om mul isi e s udies was upda ed du ing he Wo kshop by conside ing s udies abou
childhood leukaemia isk nea NPPs published a e 2008 (Ge many [36, 37], Swi ze land[38],
F ance [39, 40], Finland[41], G ea B i ain[29] and Belgium[42]). Consis en wi h o he
au ho s [43], he Wo kshop pa icipan s no ed ha , e en when he associa ions a e no s a-
is ically signi ican , he esul s om hese ecen s udies demons a ed some ele a ed isk o
childhood leukaemia when conside ing he 0–4-yea age ca ego y wi hin 5 km om a NPP.
Howe e , esul s a e based on small numbe s, and conside able me hodological di e ences
be ween s udies (design, s a is ical analyses, pe iods co e ed, diagnosis quali y and adminis-
a i e uni s analysed) do no allow he combina ion o hei esul s by simply adding numbe s.
Mo eo e , a compila ion o esul s a ailable o he complemen a y 5–14 yea age ca ego y
pe o med du ing he Wo kshop did no indica e an ele a ed isk o childhood leukaemia in
his age ca ego y.
Based on he a ailable li e a u e, he Wo kshop pa icipan s concluded ha he e was no
ele a ed isk o childhood leukaemia globally nea NPPs in child en less han 15 yea s old.
The a he consis en pa e n o inc eased leukaemia isk in he 0–4-yea olds needs o be
e i ied in he u u e and should no be in e p e ed as a causal associa ion, bu i may p o ide
clues abou a possible link be ween childhood leukaemia and li ing in he close p oximi y o
a nuclea acili y (see[44]).
Two new s udies ha e been published since he Wo kshop. The i s one conside ed can-
ce incidence wi hin 25 km o he h ee On a io NPPs using an ecological s udy design[45].
Thesecond is a case–con ol s udy including all leukaemia and non- Hodgkin lymphoma
cases diagnosed below age i e yea s in G ea B i ain, and analysing isk ela ed o dis-
ance om 13 NPPs[46]. Bo h s udies showed no excess isks nea NPPs. Ne e heless,
as in p e ious s udies, he numbe s o cases we e small when ocusing on he 0–4 yea
age ca ego y wi hin 5 km o he NPPs. The US Na ional Resea ch Council (NRC) ecen ly
published a epo on he assessmen o cance isks in popula ions nea US nuclea acil-
i ies[47]. The NRC commi ee ecommended conside ing wo di e en s udy designs:
(1) an ecologic s udy o mul iple cance ypes o popula ions li ing nea nuclea acili-
ies and (2) a eco d-linkage-based case–con ol s udy o cance s in child en bo n nea
nuclea acili ies. Tes ing he easibili y o a dose assessmen was also ecommended. The
launch o a pilo s udy, based on se en nuclea ins alla ions loca ed in six US s a es, was
announced in Oc obe 2012.
2.2. Me hodological aspec s o s udies pe o med
2.2.1. Epidemiological design. Mos o he s udies pe o med since he ea ly 1980s ha e been
o ecological design (i.e. a e age a es in geog aphical a eas). This ype o analysis can be
subjec o well-known ypes o bias (ecological allacy, mig a ion, sensi i i y o he a bi a y
choice o he s udy pe ime e , e c). Thus, mo e ecen s udies used indi idual based designs,
i.e. case–con ol[36, 40] and coho design[38, 41]. Whils hese app oaches can a oid some
o he limi a ions associa ed wi h ecologic s udy design, o he limi a ions exis (selec ion bias,
Re iew
R57
J. Radiol. P o . 34 (2014) R53
pa icipa ion bias, mig a ion bias, lack o con ol o con ounding ac o s, small sample size,
e c). Fu he mo e, esponse bias speci ic o case–con ol s udies is well documen ed[48, 49].
Indeed, a he Ge man KiKK s udy da a om a de ailed ques ionnai e we e no used due o he
po en ial esponse bias[36]. Ne e heless, i he geog aphic uni used in an ecologic s udy is small
enough ( o example municipali ies in F ance), esul s om case–con ol and ecologic s udies
a e in ag eemen [36, 37, 40]. Inconclusion, o he speci ic ques ion o childhood leukaemia
incidence nea NPPs, no clea p e e ence is e iden ; each design p esen s ad an ages and limi-
a ions. Fu u e s udy p o ocols ha e o be de eloped acco ding o he a ailabili y o da a, he
inpu equi ed e sus he scien i ic e u n and he hypo heses o be es ed, wi h speci ic a en-
ion paid o minimise po en ial biases. No wi hs anding his, e o s o e alua e he easibili y
o combining da a om di e en exis ing s udies o inc ease s a is ical powe migh be use ul.
2.2.2. Powe and p io hypo hesis. The Wo kshop pa icipan s unde lined he necessi y o a
con inuing su eillance o childhood leukaemia incidence nea NPPs, bu no limi ed o NPPs.
Ne e heless, hey ecommended ha no new s udies a e se up be o e an explici hypo hesis
is o mula ed. The main analysis o es his hypo hesis mus be speci ied, and he associa ed
s a is ical powe mus be calcula ed ap io i (which is no sys ema ically done in he usual
p ac ice). Mul isi e s udies a e p e e ed o single-si e s udies, as hey ha e a g ea e s a is i-
cal powe and p o ide a b oade con ex o he in e p e a ion o esul s, i.e. compa ing isks
be ween si es o simila cha ac e is ics.
La ge-scale s udies, no ocused on po en ial sou ces o isk, a e use ul wi h espec o he
ques ion o whe he childhood leukaemia has he endency o clus e . Hi he o, esul s a e no
consis en (e.g. [50–53]). The issue o clus e ing is impo an o he in e p e a ion o local
clus e s (in pa icula wi h espec o he in ec ion hypo hesis) and should be in es iga ed
u he , bu independen ly om he exis ence o nuclea si es. I was also concluded ha new
s udies on childhood leukaemia nea nuclea ins alla ions should aim o look a o he po en ial
causes o leukaemia.
2.2.3. Pe inen ou comes. The Wo kshop pa icipan s ag eed ha i was easonable o
s udies nea nuclea si es o ha e conside ed childhood leukaemia as he main po en ial
ou come; bu gi en he he e ogenei y o he diseases collec i ely desc ibed as ‘leukaemia’,
u u e s udies should dis inguish be ween leukaemia sub ypes. Combined analyses o da a
om se e al coun ies in Eu ope may allow he accumula ion o a su icien ly la ge numbe
o cases o such sub ype-speci ic analyses. Fu he , he unclea and somewha a bi a y (a
leas , his o ically) dis inc ion be ween some o ms o childhood NHL and some o ms o
childhood ALL mus be bo ne in mind.
O he ou comes may be conside ed, bu only i hey a e clea ly hypo hesis-d i en. Fo exam-
ple, he Belgian NUCABEL s udy conside ed hy oid cance in addi ion o childhood leukae-
mia, because he s udy was o de ed ollowing he Fleu us nuclea inciden in ol ing eleases o
iodine-131 [42]. Among pa hologies ha we e hypo hesised o be po en ially associa ed o ei he
p e-concep ional [54] o in u e o exposu e, congeni al mal o ma ions migh also be conside ed
as a possible heal h indica o , as was done o example in Ge many in pa allel o he KiKK
s udy[55]. I was no ed ha in o ma ion on o he diseases (especially childhood cance s o he
han leukaemia) migh help wi h he in e p e a ion o esul s in ha , o example, a common
bias could be mani es o e all ypes o childhood cance a he han jus childhood leukaemia.
2.2.4. Exposu e indica o s. Dis ance can be easily and eliably de e mined, bu i cons i u es
only a c ude p oxy o adia ion exposu e om nuclea ins alla ions. Be e exposu e indica-
o s a e needed o in es iga ing heal h e ec s a ound nuclea si es. This ecommenda ion is

Re iew
R58
J. Radiol. P o . 34 (2014) R53
in acco dance wi h ha o he NRC epo which calls o abso bed doses o indi idual o gans
o be es ima ed[47].
Based on he esul s om exposu e moni o ing o dosime ic es ima es, he o de s o
magni ude o he doses a ibu able o cu en ou ine discha ges appea o be small. The
so-called ‘ adioecological s udies’ conduc ed in he UK[56–58], in Ge many[59] and in
F ance[60, 61] indica ed ha he le els o doses a ibu able o eleases om he plan s we e
well below hose due o na u al sou ces o adia ion (100 o 1000 imes, o mo e) and could
no explain he obse ed numbe o cases[62, 63]. The easibili y and pe inence o measu e-
men s o adionuclides ( i ium and ca bon-14 and possibly o he adionuclides) in en i on-
men al media o di ec ly in child en (e.g. u ine analysis), o o he dosime ic app oaches o
e alua e indi idual doses should be in es iga ed, as he di ec measu emen o he body con-
en s o adionuclides such as caesium-137 and plu onium-239/240 ha e p o ided aluable
in o ma ion in he pas (e.g. [58, 64, 65]).
I is also impo an ha he cha ac e is ics o he local popula ion (demog aphy, sociology,
socio-economic s a us (SES), li e habi s, day-ca e a endance, e c) a e de e mined in as much
de ail as possible. Also, a en ion should be gi en o suspec ed po en ial en i onmen al isk
ac o s o leukaemia, such as ELF-EMF, pes icides, ai pollu ion, and p oximi y o a pe ol
s a ion o indus ial si es.
2.3. Rou es o imp o emen
The Wo kshop pa icipan s concluded ha sys ema ic egis a ion o childhood leukaemia
cases is needed based on exhaus i e and s anda dised eco ding sys ems, including popula-
ion-based egis ies. Con inua ion and de elopmen o ongoing egis a ion ac i i ies ha e
o be suppo ed; his migh include addi ional in o ma ion o be collec ed in o de o cap-
u e ea ly li e exposu es. Recons uc ing he exposu e his o y o he child en (place o bi h,
esiden ial his o y, an ena al exposu es, e c) is desi able, whe e e possible wi h easonable
e o . Also, e o s should be made o s udies o include collec ion o biological samples o
suppo esea ch in o he pa hogenesis o childhood leukaemia.
The Wo kshop pa icipan s encou age collabo a ion be ween eams in ol ed in his esea ch
opic o b ing oge he in es iga o s om di e en s udies, and con ol o me hodological di -
e ences, in o de o be e in e p e and e alua e he cohe ence and di e ences o he esul s
ob ained in a ious coun ies. Pa icipa ion in c oss-bo de ini ia i es would also acili a e in es-
iga ing possible heal h e ec s in he icini y o he many nuclea si es ha a e loca ed close o
na ional bo de s, as is in he case o F ance and Belgium (Chooz NPP). Also, he conside a ion o
se e al po en ial con ounde s ha may lead o di e ences in indings among coun ies—e.g.SES,
u ban– u al s a us, popula ion densi y—should be conside ed; bu he s ong o e lapping o hese
ac o s should be kep in mind in he in e p e a ion o he esul s. Finally, collabo a ion should
p o ide he amewo k o an in e na ional combined analysis, which is he bes app oach o he
de i a ion o pooled isk es ima es. A wo king g oup a he Eu opean le el should e i y he ea-
sibili y and pe inence o such pooling e o , wi h a special a en ion o minimising biases.
3. Ae iology and mechanisms o childhood leukaemia ini ia ion
andde elopmen
The second pa o he Wo kshop ocused on he ae iology and mechanisms o childhood
leukaemia ini ia ion and de elopmen , aking in o accoun di e en en i onmen al expo-
su es and gene ic isk ac o s[5]. I is he aim o in e na ional conso ia like he Childhood
Re iew
R59
J. Radiol. P o . 34 (2014) R53
Leukaemia In e na ional Conso ium (CLIC) [66] and he In e na ional Childhood Cance
Coho Conso ium (I4C) [67], o pool da a o u he unde s and he causes o childhood
leukaemia. Wo kshop pa icipan s p o ided an o e iew o cu en knowledge and iden i ied
possible s eps o wa d. Speci ic a en ion was gi en o B-cell-p ecu so ALL, which is he
commones sub ype o childhood leukaemia.
3.1. The mul is ep de elopmen o childhood leukaemia
Childhood leukaemia is a he e ogeneous disease, and i s de elopmen is a mul is ep p ocess[7, 68].
Cu en esul s indica e ha he i s s ep occu s in u e o con e ing a haema opoie ic p ecu -
so o s em-cell o a p eleukaemic clone[69, 70]. Speci ic ch omosomal ansloca ions, as a
esul o DNA double s and b eak o high hype ploidy, a ising p obably om a single-s ep
mechanism[71] a e he commones cy ogene ic abno mali ies in B-cell-p ecu so ALL. Fo
example, he ansloca ion RUNX1-ETV6 occu s wi h a equency o app oxima ely 20%
in ALL, and high hype ploidy is ound in 25–30% o ALL cases. I has been shown ha
app oxima ely one in 100 new-bo ns ca y he RUNX1-ETV6 ansloca ion, which is a a e
signi ican ly highe han he o e all incidence a e o ALL in he popula ion [72]. This would
imply ha p eleukaemic clones a e equen , and no mally ex inguished (o a leas kep a
bay) by na u al p ocesses. Howe e , such high equency o RUNX1-ETV6 ansloca ion was
no con i med in ano he ecen s udy [73].
Beyond he known gene ic o epigene ic e en (s), o he elemen s mus in e ac o lead o
he de elopmen o acu e leukaemia [74, 75], bu so a , epidemiological s udies ha e no been
able o dis inguish be ween he co ela ion o a ious po en ial isk ac o s wi h he ini ial and
subsequen gene ic e en s.
E o s need o concen a e on de e mining he ime o o igin o he p eleukaemic clone
and i s p e alence a bi h, as his is impo an o he unde s anding o he ole and pe inen
ime window a which isk ac o s unde conside a ion would be in ol ed [76]. The na u e o
he cell-o -o igin canno be iden i ied by s udying ALL samples a he ime o diagnosis [77].
Thus, p ospec i e designs ha e o be conside ed, and bi h coho s may p o e o g ea in e es .
Gi en he a i y o he disease, a wo ldwide coo dina ed app oach is needed o allow compa i-
son o he p e alence o he p eleukaemic clone in di e en coun ies and e hnic g oups. So
a , li le is known abou ALL incidence and ends in de eloping coun ies, bu he incidence
o p e-B ALL gene ally appea s o be low in low-income and middle-income coun ies and
inc eases subs an ially as coun ies unde go socio-economic de elopmen [78, 79].
3.2. Inhe i ed suscep ibili ies
Childhood ALL a ises as a consequence o a limi ed numbe o gene ic al e a ions [80].
Se e al independen genome-wide associa ion s udies (GWAS) ha e con i med a numbe
o gene ic a ian s ha a ec gene ic suscep ibili y o ALL, al hough hey a e indi idually
modes in hei e ec s. In he mos ecen GWAS on ETV6-RUNX1-posi i e ALL, a ew
(ETV6-RUNX1-speci ic) suscep ibili y loci, p edominan ly ela ed o pa hways con olling
emb yonic and B-cell de elopmen and di e en ia ion, we e iden i ied [81]. Based on copy
numbe al e a ions, cases could be ca ego ised in o ou dis inc subg oups e ealing gene ic
di e si y wi hin ETV6-RUNX1-posi i e ALL which is o impo ance o imp o ing ea men
s a egies [82, 83]. Polymo phisms in many o he pa hways ha e been in es iga ed in ela ion
o isk o ALL such as DNA epai and cell cycle con ol genes [84, 85]. A me a-analysis on
candida e gene ic a ia ions, howe e , saw signi ican associa ions in only eigh ou o 25poly-
mo phic a ian s [86]. GWAS e ealed ha common gene ic a ia ions con ibu e o he isk
Re iew
R60
J. Radiol. P o . 34 (2014) R53
o B-cell-p ecu so ALL [87]. On he o he hand, a high conco dance o sub ypes o ALL
wi hin amilies has also been epo ed [88], indica ing s ong gene ic and/o en i onmen al
isk ac o s a e es ic ed o speci ic ALL sub ypes. This implies ha u u e s udies on isk
ac o s should dis inguish be ween ALL sub ypes. Simila ly, leukaemias in in an s and in
young child en a e di e en gene ic and clinical en i ies ha should be examined sepa a ely
in u u e epidemiological s udy.
Nex -gene a ion sequencing (whole genome o ansc ip ome sequences, exome cap u e
and sequencing, analyses o he me hylome) o ALL cases migh unco e a common signa-
u e. In con as o GWAS, deep sequencing ocuses on mu a ions and o he de ailed di e -
ences e.g. in he epigenome, which could, a leas pa ly, e lec he e ec s o he en i onmen .
As o o he cance s [89, 90], his could help o cha ac e ise leukaemia subg oups ha migh
no show up in an o e all app oach, o de ec common pa e ns/ oo p in s possibly co ela ed
o ex e nal isk ac o s [91–93].
3.3. De elopmen al aspec s
Haema opoiesis in he emb yo occu s in s ages [94]. The i s s age occu s ea ly in in a-
u e ine li e and aims a p oducing ed blood cells, whils he second gene a es all lineages o
blood cells and HSCs. Du ing oe al li e, HSCs a e cha ac e ised by a high le el o cell cycle
ac i i y (app oxima ely 100% pe 24 h) and become quiescen upon seeding he bone ma -
ow, a p ocess inalised in humans a an age o app oxima ely wo yea s. The p ope ies and
unc ionali y o he de eloping haema opoie ic sys em may p o e impo an o he onse , bu
also o he ou come o he disease. Indeed, di e en age g oups a e associa ed wi h di e en
leukaemia sub ypes: MLL- ela ed B-cell ALL is he main sub ype in he age g oup 0–1, while
B-cell-p ecu so childhood ALL peaks a ound 2–5 yea s.
Cell cycle ac i i y o HSCs is ca e ully modula ed by a complex ne wo k o cell-in in-
sic and cell-ex insic mechanisms [94] and he unde s anding o he age-dependen cell
cycle egula ion migh , he e o e, p o ide clues o he de elopmen o leukaemia sub ypes.
Epidemiologic s udies should aim a dis inguishing he leukaemia sub ypes associa ed o di -
e en age ca ego ies. Animal models may gi e u he insigh in o haema opoiesis in di e en
age g oups and on ex e nal in luences on hese p ocesses.
3.4. Mic oen i onmen
In iew o he high cell di ision a e du ing oe al and in an haema opoiesis and he mu a ion
po en ial o lymphoid cells, any geno oxic compound as well as he weakening o cell cycle
checkpoin s ingency and immunological su eillance can inc ease malignan ans o ma ion.
The s omal mic oen i onmen is known o be impo an o HSCs’ p oli e a ion, ac i i y
o quiescence, espec i ely, and o B-lineage de elopmen and di e en ia ion. Leukaemia-
speci ic gene ic abe a ions ha e been ound in he mesenchymal s em cells in some ALL
cases [95]. O e all, he ole o he mic oen i onmen on he pe sis ence o p eleukaemic
clones and he plas ici y o leukaemia cells needs o be unco e ed. In a mouse model he
lineage a e o leukaemia cells (B-cell ALL e sus AML) was de e mined by he hos mic o-
en i onmen [96].
3.5. Immune sys em and he in ec ious hypo hesis
The hypo hesis ha childhood leukaemia is a a e esponse o an in ec ion is pe sis en . Kinlen
sugges ed ha childhood leukaemia could a ise as a a e consequence o exposu e o a speci ic
Re iew
R61
J. Radiol. P o . 34 (2014) R53
uniden i ied common in ec ion, he a e ou come being pa icula ly e iden a imes o unusual
popula ion mixing, when a ela i ely la ge numbe o suscep ible child en encoun e a ela i ely
la ge numbe o in ec ed indi iduals leading o (la gely subclinical) epidemics [21, 97–99].
I has been sugges ed ha he cons uc ion o a nuclea acili y in a u al a ea migh p omo e
unusual social con ac s and popula ion mixing ha could ep esen a leas a co ac o o he
isk obse ed [98]. This in ec ion hypo hesis could also p o ide an explana ion o he leu-
kaemia clus e in Fallon, USA [30, 100]. G ea es hypo hesised ha a pauci y o exposu e o
in ec ious agen s in he i s yea o li e and subsequen ‘delayed’ in ec ious challenge may be
causal in he de elopmen o B-cell-p ecu so ALL in he peak ages wo o i e yea s [20, 101].
This hypo hesis is suppo ed by s udies ha ha e looked a sel - epo ed day-ca e a endance
as p oxy o social con ac s [102], bu o he s udies obse ed ha child en who de eloped
B-ALL aged 2–5 yea s had signi ican ly mo e clinically diagnosed in ec ious illness episodes
in he i s yea o li e compa ed o con ols [103]. This sugges s ha immune de egula ion
in child en who de elop ALL may be de ec able se e al yea s be o e diagnosis [104, 105].
Fu he suppo o his obse a ion is p o ided by demons a ing ha In e leukin 10 le els in
neona es a e educed in new-bo ns who de elop leukaemia [106].
The in luence o in ec ions on HSC p oli e a ion has been shown [107], hus con i ming
a ela ion wi h in ec ions, immune sys em and in lamma ion on a molecula basis. Fu he
s udies should y o include he de elopmen and s a us o he immune sys em in he oe us
compa ed o in an s, young child en and adul s.
I a speci ic in ec ious agen is esponsible o childhood leukaemia hen i should be
de ec able. Iden i ica ion o gene a ian s should be possible by e isi ing al eady a ailable
single-nucleo ide polymo phism (SNP) da a. Deep sequencing should help inding he p es-
ence o pe inen i uses o epigene ic ma ks o hei ac ion i he e was a i us-based hi -and-
un mechanism, bu i should be bo ne in mind ha i childhood leukaemia is a a e esponse
o a common in ec ion hen he esponsible agen is likely o ha e in ec ed mos child en, and
only hose in emo e u al a eas may ha e escaped in ec ion.
3.6. Animal models
As an animal model, a ‘mul i-hi ’ model is highly desi able o s udy he mechanisms o ini-
ia ing e en s (gene ic, epigene ic o en i onmen al hi s) as well as he na u e and ole o
he u he hi s, hei kine ics and age-dependence and po en ial isk ac o s (in insic e sus
ex insic ac o s: chemicals, adia ion, in ec ions/in lamma ion, immune con ol o malignan
p og ession). Fo many yea s he exis ing animal models we e only o limi ed e sa ili y, and
he e o e he gene a ion o new, mo e adequa e animal models o childhood leukaemia is
necessa y [108, 109]. Mouse models can con ibu e in suppo ing hypo heses de i ed om
human molecula s udies, and ice e sa, as o ins ance he con ibu ion o gene a ian s o
he iden i ica ion o new (epi)gene ic suscep ibili y genes.
The ul ima e goal is o be able o mimic in he mouse he ea u es o human B-cell ALL, a each
le el (molecule, cell, issue, whole o ganism), including ini ia ion, p og ession, e olu ion, esponse
o he apy and e en ual cu e o elapse. Mouse models o B-cell ALL a e use ul o explo ing he
abno mal p ocesses in haema opoiesis leading o leukaemia and he ole o he immune sys em
and o iden i ying inge p in s o exposu e [77]. O he models, like he NOD/SCID mice, o e he
possibili y o s udying p emalignan human co d HSCs and ollow up hei de elopmen .
Addi ionally, he analysis o hos gene ic backg ound in leukaemia de elopmen would
gi e u he bene i s. In a well-de ined mouse model he animals should be exposed o possi-
ble isk ac o s ( adia ion, chemicals, in ec ions/in lamma ion, eplica i e s ess, e c) ollowed
by comple e pheno yping wi h s anda dised p o ocols. A mouse model o con olled gene ic
Re iew
R68
J. Radiol. P o . 34 (2014) R53
[100] S einmaus C, Lu M, Todd R L and Smi h A H 2004 P obabili y es ima es o he unique
childhood leukemia clus e in Fallon, Ne ada, and isks nea o he US Mili a y a ia ion
acili ies En i on. Heal h Pe spec . 112 766–71
[101] G ea es M F 1988 Specula ions on he cause o childhood acu e lymphoblas ic leukemia
Leukemia 2 120–5
[102] U ayama K Y, Ma X and Bu le P A 2008 Exposu e o in ec ions h ough day-ca e a endance
and isk o childhood leukaemia Radia . P o . Dosim. 132 259–66
[103] C ouch S, Ligh oo T, Simpson J, Smi h A, Ansell P and Roman E 2012 In ec ious illness in
child en subsequen ly diagnosed wi h acu e lymphoblas ic leukemia: modeling he ends
om bi h o diagnosis Am. J. Epidemiol. 176 402–8
[104] Roman E, Simpson J, Ansell P, Kinsey S, Mi chell C D, McKinney P A, Bi ch J M, G ea es M
and Eden T 2007 Childhood acu e lymphoblas ic leukemia and in ec ions in he i s yea
o li e: a epo om he Uni ed Kingdom Childhood Cance S udy Am. J. Epidemiol.
165 496–504
[105] Roman E, Simpson J, Ansell P, Ligh oo T and Smi h A 2009 In ec ious p oxies and childhood
leukaemia: indings om he Uni ed Kingdom Childhood Cance S udy (UKCCS) Blood
Cells Mol. Dis. 42 126–8
[106] Chang J S, Zhou M, Bu le P A, Chokkalingam A P, Me aye C and Wiemels J L 2011 P o ound
de ici o IL10 a bi h in child en who de elop childhood acu e lymphoblas ic leukemia
Cance Epidemiol. Bioma ke s P e . 20 1736–40
[107] Bald idge M T, King K Y, Boles N C, Weksbe g D C and Goodell M A 2010 Quiescen
haema opoie ic s em cells a e ac i a ed by IFN-gamma in esponse o ch onic in ec ion
Na u e 465793–7
[108] McCo mick D L and Ka e R 2004 Animal models o he s udy o childhood leukemia:
conside a ions o model iden i ica ion and op imiza ion o iden i y po en ial isk ac o s
In .J.Toxicol. 23 149–61
[109] Vicen e-Duenas C, Cobaleda C, Pe ez-Losada J and Sanchez-Ga cia I 2010 The e olu ion o
cance modeling: he shadow o s em cells Dis. Models Mech. 3 149–55