Summer Conference on Neonatology in Provence€¦ · • Erik van Tilborg • Elise Turk • Kim Annink Image science institute: • Prof. Max Viergever • Alexander Leemans •

Predicting outcome using neonatal MRI

in preterm infantsManon Benders, neonatologist

Summer Conference on Neonatology in Provence

Avignon, France, 6th - 9th September 2017

the human brain

30 wks MRI

the third trimester is important for brain development

In utero

Ex utero

40 wks MRI

Affected Neurodevelopment

Longterm outcome of very preterm infants

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

UK Zweden UK Zweden

Ou

tco

me

(%

)

24 wks 25 wks

Outcome EPICURE-II en EXPRESSof NICU admissions

mild/no abn

moderate/severe

died

• Who is at risk?

• How can we identify those infants?

• Who is at risk?

• How can we identify those infants?

• Reasons:

• Prediction and information to the parents

• Selection for early neuroprotective intervention

B. Larroque et al, Lancet 2008

5 yrs

decreased IQ

Longterm consequences of prematurity

MRI ?

• MRI and predicting neuro-motor outcome

• MRI and adverse neurodevelopmental outcome

– White matter injury

– Cerebellar injury

• MRI and quantitative analysis and prediction

Predicting outcome using neonatal MRI

in preterm infants

Conventional MRI: Prediction of motor outcome

Peri-ventricular hemorrhagic infarction:Evaluation of myelin in the PLIC – T1

NO PLIC

Girl, GA 28+5 wks, IVH gr IV, > hemiplegia - cerebral palsy

De Vries LS, et al , Neuropediatrics 1999De Vries LS, et al. Eur J of Pediatr Neuro 2001

T2 T2 T130wks TEA TEA

Decreasing Incidence and Severity of CP

van Haastert I. et al. J of Pediatr 2011

N=3287, GA <34 wks1990-2005

Who is performing routine MRI in preterm infants?

Decrease of CP, however 50% abnormal development> Does routine MRI help to predict outcome?

B. Larroque et al, Lancet 2008

All babies<28 wks

• MRI and predicting neuro-motor outcome

• MRI and adverse neurodevelopment outcome

– White matter injury

– Cerebellar injury

• MRI and quantitative analysis and prediction

Predicting outcome using neonatal MRI

in preterm infants

Conventional neuroimaging in the preterm infant:

more subtle lesions – adverse cognitive outcome

M. Hinojosa-Rodríguez et al. NeuroImage: Clinical 16 (2017) 355–368

• Very preterm infants with low-grade GMH-IVH (n=69)

on cUS have a similar early neurodevelopmental

outcome compared with controls (n=96)

• However, several infants developed neurodevelopmental

impairment

• Additional MRI showed subtle abnormalities that were

missed with cUS, this partly explained CP

Pauline Reubsaet et al, Neonatology 2017

The Impact of Low-Grade IVH on Neurodevelopmental Outcome

of Very Preterm Infants

The Impact of Low-Grade IVH on Neurodevelopmental Outcome

of Very Preterm Infantsadditional injury on MRI

Pauline Reubsaet et al, Neonatology 2017

Omizzolo et al. Memory. 2015

Neonatal Brain Abnormalities and Memory and LearningOutcomes at 7 Years in Children Born Very Preterm

n=198 preterm infants vs n=70 term

• MRI and predicting neuro-motor outcome

• MRI and adverse neurodevelopmental outcome

– White matter injury

– Cerebellar injury

• MRI and quantitative analysis and prediction

Predicting outcome using neonatal MRI

in preterm infants

White matter injury

Compare to CUS: MRI is more sensitive for detecting

punctate white matter lesions

de Vries, L. S., Benders, M. J. N. L., Groenendaal, F. (2013) Neuroradiology

Identifying PWML– preterm infantsSusceptibility weighted imaging (SWI)

• Clearly detection of hemorrhagic lesions

• More hemorrhagiclesions were identifiedusing SWI than T2

T1

SWIT2

SWI

Identifying PWML– preterm infantsDiffusion weighted imaging (DWI)

Hypoxic-ischemic detection of lesions= very clear on DWI and T1/IR

DWICUS IR/T1

Punctate white matter lesions

Kersbergen PLoS One 2014

112 preterminfants, 2 MRIs

linear patternwith low SI on SWI (n=67)

cluster patternwith high SI on

DWI (n=26)

mixed pattern(n=19)

PWML n=112

TEA scan: 70%Linear pattern

TEA scan: 12%Cluster pattern

TEA scan: 18%mixed

More florid on T2Often associated withadditional lesions/IVH

More florid on T1 mixed

SWI low signal intensity

DWI restricted diffusion

Follow-up data were available for 91 infants:• 9 infants developed cerebral palsy in cluster group

• 6 with additional lesions• 3 with ‘only’ PWML in the cortical spinal tracts

Kersbergen PLoS One 2014

Presence of any punctate white matter lesion in corticospinal tracts

GMFCS 2-5 Motor compositescore <85

Odds Ratio

7.859 (2.56–24.17)

2.192 (1.12–4.31)

AUC 0.728 (0.6-0.9)

0.574 (0.5-0.7)

Tusor N, Benders et al. 2017 submitted

Consequence of white matter injury: PWML

Location is important

Quantitative assessment of white matter

injury in preterm neonates: Association with outcomes

Guo, et al. Neurology 2017

Location motor cognition language

• MRI and predicting neuro-motor outcome

• MRI and adverse neurodevelopmental outcome

– White matter injury

– Cerebellar injury

• MRI and quantitative analysis and prediction

Predicting outcome using neonatal MRI

in preterm infants

Preterm girl of 27+3 wks of GA.

1st wk of life 3st wk of life 30wk corrected age

CUS term equivalent age MRI term equivalent age

Unilateral lesion

Compare to CUS: MRI is more sensitive for detecting cerebellar hemorrhages

Baby boy, GA 27 wks

Bilateral lesion

term equivalent age

30wk corrected age30wk corrected age

1st wk of life 1st wk of life

MASTOID window

CUS CUS

T1 MRI T2 MRI T2 MRI

CB hemorrhage: cerebello-cerebral diaschisis> altered trophic trans-synaptic interactions

Limperopoulous et al, cerebral cortex 2012

• isolated CB injury

• n = 40

• at 34 mnd MRI and FU

Conclusions:• isolated CB injury is related with secondary impairment of

regional volumetric growth in the contralateral cerebral cortex• remote regional impairments of cerebral volumetric growth

due CB injury related to domain-specific functional deficits

VolpeLimperopoulous et al, Pediatrics 2005, cerebral cortex 2012

• MRI and predicting neuro-motor outcome

• MRI and adverse neurodevelopmental outcome

– White matter injury

– Cerebellar injury

• MRI and quantitative analysis and prediction

Predicting outcome using neonatal MRI

in preterm infants

brain development in preterm period

remote injury – white matter injury

Volpe Lancet Neurol 2009

Cortical folding – sulcus level and prediction of

outcome

Kersbergen K, et al, J of Pediatr 2017

brain volumes and neurodevelopment

Keunen et al J Pediatr 2016, Anbeek et al PLoS One 2013

neonatal 2 years 5.5 years

Neonatal network is related to performance IQ at

early school age

Effect size is resp 25% and 27%

Keunen et al, DMCN 2017

Prediction of cognitive and motor outcome of preterm

infants based on automatic quantitative descriptors

from neonatal MR brain images

Prems <28 wksN=71 serial scans2 yrs Bayley outcome

Moeskops P, et al, scientific report 2017

Prediction of cognitive and motor outcome of preterm

infants based on automatic quantitative descriptors

from neonatal MR brain images: 2 yrs of age

Moeskops P, et al, submitted

Cognitive score30 wks

Motor score30 wks

40 wks lower AUC

Machine learning – for diagnosis and prognosis

> individualized care

• Selection for intervention trial?

Parikh Na, Seminars of perinatology 2016

Why would we like to perform an MRI in

preterm infants?

• To detect brain injury

• To assist in the prediction of high-risk children

• Early neuroprotective intervention?

Why would we like to perform an early MRI in preterm infants?

XVI Update in Neonatologia – Florence

Erytropoietine

n=24 vs n=34

n=77 vs n=88

Frac

tio

nal

anis

otr

op

y

• GA 28-33 wks

• Intervention: 1st 72h until 2 wks CA

• Outcome: ABIP and Prechtl (2 wks – 9mnths CA)

electrophysiology (2 wks CA)

MRI (2 wks CA)

Bayley II (24 mths)

Als, H.et al (2004). Pediatrics

Early brain activity relates to subsequent brain development

NIDCAP alters Brain Structure and Function

Higher relative anisotropy in left internal capsulewith a trend for right internal

capsule and frontal white matter

intervention

No intervention

• n=10, GA 30-33wks

• day 10

• 3 times/day

• two blocks of 5 d

• separated by a 2 d

- higher levels of blood IGF-1

-

- accelerates the maturation of brain activity

Massage Accelerates Brain Development and

the Maturation of Visual Function Early brain activity relates to subsequent brain development

Shortening in IBI Shortening VEP

Guzzetta A et al. J Neurosci 2009,

DMCN 2011

Conclusion

predicting outcome from MRI in preterm infants

• MRI predicts neurodevelopmental outcome

– More sensitive for smaller/additional lesions – related to outcome

• White matter injury

• Cerebellar injury

– Useful to use quantify brain parameters

• Prediction of neurodevelopmental outcome

– Selection for intervention?

– Evaluation of clinical intervention trials?

Veel dank Thank youNICU, especially:• Prof. Linda de Vries• Floris Groenendaal• Dr. Niek van der Aa• Dr. Elise Roze

PhD - / medical students• Kristin Keunen• Nienke Wagenaar• Nathalie Claessens• Lisanne Stolwijk• Erik van Tilborg• Elise Turk• Kim Annink

Image science institute:• Prof. Max Viergever• Alexander Leemans• Ivana Isgum• Pim Moeskops

MRI technicians

International collaborators:LONDON:• Prof. Serena CounsellPARIS: • Jessica Dubois

Physiotherapist• Ingelot van Haastert

Physician assistents

RMI / psychiatry• Martijn van de Heuvel

Questions?