MODELING OF TEMPERATURE WITHIN THE NEONATAL HEAD SUBJECTED TO TRANSCRANIAL CONDUCTIVE COOLING

1
MODELING OF TEMPERATURE WITHIN THE NEONATAL HEAD
SUBJECTED TO TRANSCRANIAL CONDUCTIVE COOLING
G Van Leeuwen1, J Hand1 D Edwards2 and D
Azzopardi2 1 Radiological Sciences Unit, Dept of
Imaging, 2 Department of Paediatrics Neonatal
Medicine Imperial College School of Medicine
Hammersmith Hospital, London
2
Hypothermal neural rescue therapy

• studies of cerebral injury have shown that
• promoting moderate brain cooling may reduce
  damage and improve functional outcome
• preliminary clinical trials of hypothermic neural
  protection in newborn infants suffering perinatal
  asphyxia are underway

3
Hypothermal neural rescue therapy

• injury to deep brain structures predicts severe
  neurological impairment
• cortical injury is relatively benign
• MR studies have shown basal ganglia are
  significantly warmer than superficial cerebral
  tissues
• few data available regarding
• temperature distribution within the newborn brain
• effect of local or systemic cooling
• such data are difficult to acquire
• measurements by MR or MW radiometry?

4
Hypothermal neural rescue therapy

• how to cool?
• localised head cooling?
• whole body cooling?
• In this work we have
• used a thermal model to predict 3D temperature
  distributions in an anatomically realistic model
• tested sensitivity of results to
• geometry and description of heat transfer
• brain perfusion
• thermal conductivity

5
Baby head model
Orthogonal cross-sections within the 3
dimensional MR data set on which the temperature
distributions are contoured
6
Assumptions made in the model
water cap temperature Twater 10 oC (isothermal
boundary)
air temperature Tair 32 oC
infants core temperature Tcore 37/34 oC
7
Bioheat transfer equation (Pennes, 1948)
rate of increase in energy per unit volume due
to metabolism
rate of loss of energy per unit volume due to
perfusion heat sink B - w cblood (T -Tcore)
rate of change of energy per unit volume due to
thermal conduction
8
Tissue properties
9
Head cooling only
37 oC
37 oC
27 oC
Twater_cap 10 oC Tcore 37 oC contours at 2
oC intervals
37 oC
10
Head whole body cooling
34 oC
34 oC
24 oC
Twater_cap 10 oC Tcore 34 oC contours at 2
oC intervals
34 oC
11
Reduced adult head models

• 11/15 scale model from T1 weighted MRI adult head
  data set
• heat sink
• discrete vessels
• additional data regarding cerebral vasculature
  from MR angiography
• visible vessels traced and small vessels added
  using a vessel generation
• algorithm (vessels down to 0.2 mm diameter
  included)

12
Discrete arterial and venous vessel determined
by MRA
13
Discrete arterial and venous vessel networks
14
Reduced adult head - heatsink model
37 oC
37 oC
Twater_cap 10 oC Tcore 37 oC contours at 2
oC intervals
27 oC
37 oC
15
Reduced adult head - heatsink model
34 oC
34 oC
Twater_cap 10 oC Tcore 34 oC contours at 2
oC intervals
24 oC
34 oC
16
Reduced adult head - discrete vessels
37 oC
37 oC
Twater_cap 10 oC Tcore 37 oC contours at 2
oC intervals
27 oC
37 oC
17
Reduced adult head - discrete vessels
34 oC
34 oC
Twater_cap 10 oC Tcore 34 oC contours at 2
oC intervals
24 oC
34 oC
18
Baby head
Reduced adult head
discrete vessels
heatsink
heatsink
Tcore 34 oC
Tcore 37 oC
19
Temperature profiles
20
Temperature profiles - perfusion and conductivity
changes
low w 0.75 x normal w for brain
high k 3 x normal k for skin/skull
21
Summary (1)

• the general predictions of the thermal modelling
  are robust
• transcranial conductive cooling is effective in
  lowering the temperature significantly only in
  tissues up to approximately 20 mm deep
• cooling of superficial brain tissues
• minimal effect in diencephalon
• must be combined with whole body cooling
• to reduce temperature in diencephalon
• Reiterate key goals
• Thanks

22
A 2nd application of the thermal model

• in the development of multi-frequency microwave
  radiometry to monitor deep brain temperature
  during mild hypothermal neural rescue therapy
• Reiterate key goals
• Thanks

23
Microwave radiometry
Radiometer 1GHz lt fi (i 1, 5) lt 4GHz Dfi
0.4MHz between 1 and 4 GHz
Antenna
Pant,i k Dfi
TB,i i 1,,5
TB,i are known as brightness temperatures
24
TB,i ?antenna view field Wi (r)T(r)dv
1 GHz
2 GHz
Weighting functions Wi (r)
3 GHz
4 GHz
25
TB,i ?antenna view field Wi (r)T(r)dv
Equi-temperature shell model
T(z) T0 dT exp(-z/a) exp(-z/b)
26
Temperature profile retrieval
measure set of brightness temperatures
assume initial temperature profile (a, b, dT)
random fluctuations
calculate corresponding brightness temperatures
compare predicted and measured brightness
temperatures
adjust parameters a ,b, dT for best fit
temperature profile
predicted temperature profile and 2 s estimate
of precision
27
Summary (2)

• thermal model suggests equi-temperature shell
  model is a good approximation to T(z) within the
  baby head
• a good approximation of the predicted temperature
  profile can be obtained using 3 parameters
• T(z) T0 dT exp(-z/a) exp(-z/b)
• used in iterative solution to inverse problem to
  retrieve the temperature profile beneath the
  radiometer antenna
• Reiterate key goals
• Thanks