WM Proposal - Transcranial MR guided Focused Ultrasound for Non-Invasive Treatment of Brain DiseasesDr. Silke Lechner-Greite, Lead Scientist, GE Global Research, Munich, Germany

Thermal Treatments

2

High Intensity FUS

- kills tissue directly (ablation)

- >>55°C

- ~ 15 minutes treatment

- heated regions:

- ~ 0.1 to 1cm in diameter

- ~ 1 to 5cm in length

- Temporal resolution: ~3s

Deep Region RF

Hyperthermia

- enhances radiation & chemo therapy

- 40-43°C

- ~60mins

- heated region:

- ~5 to 10 cm diameter

- Temporal resolution: ~6-10 min

Other thermal therapies

- MR-guided laser-induced thermal therapy (LITT)

- Radio frequency ablation (RFA) (kHz)

- Microwave ablation (MWA) (GHz)

- Cryotherapy

- Nanoparticles heating

- Transcranial MR guided FUS (clinical setup and procedures)- Treatment planning & monitoring with MR thermometry- Problems, challenges, and potential goals of WM

Visualase, Medtronic CryoCath Technologies Inc

BSD Medical

Why MR guided Focused Ultrasound

� Non-invasive

� Local therapy

� Immediate result

� No late effects, e.g. delayed necrosis

� No dose accumulation effects

� No long term toxicity

� Repeatable

Key objective: … at any phase of the treatment, the surgeons need reliable guidance through the procedures and the confidence that the focal spot of the treatment is identical to the predefined target and that no other healthy tissue is getting harmed by the procedure …

� Long exam duration (~6h)

� Restricted to certain area in the

brain

� Treatment area limited by size of hot

spot

� Some steps of the therapy need a

more confident guidance (treatment

area localization, hot spot tracking,

etc.)

� New treatment studies are difficult

to be introduced – regulatory issues

But:

System Details @ University Children’s Hospital

4

Transducer

3 axis mechanical positioner

Flexible membrane

Stereotactic Frame

Martin et al., Ann Neurol, 66:858–861, 2009.

- InSightec ExAblate 4000 Neuro system

- Hemispherical 1024-element phased-array transducer operating at 650 kHz

- Discovery MR750w 3.0T, GE Healthcare

- Stereotactic frame to immobilize the patient

- Degassed water used for acoustic coupling & cooling of skull surface

- CT-based acoustic modeling

- PRFS- based MR thermometry

CT-based Modeling

Tra

nsd

uce

r

RF signals

Skull deflects the ultrasound beam

Phase correction

CT – based modeling

5

Courtesy of Beat Werner, Childrens Hospital Zurich

Treatment Planning - Drawbacks

� CT based modeling for phase correction:

� exposure to ionizing radiation

� ~ 30% of the patients are not treatable:

� porosity of bone

� Geometry

� Target area is selected based on a brain

atlas

� no individual adaptation is possible

6

Images - Courtesy of Beat Werner, Childrens Hospital Zurich

Treatment Monitoring – PRFS based MR Thermometry

7

The Phase of the water signal is temperature dependent.

Lower Proton Resonant Frequency

Higher Temperature – “free” H2O molecules

Hydrogen nucleus is more de-shielded

Higher Proton Resonant Frequency

Lower Temperature – “ice-like” H-bonded state

∆T = T2

−T1

=φ(t

2,T

2)−φ(t

1,T

1)

αγBoTE

Cppmo

00909.0−=αx

z

before heating

after heatingy

Hydrogen bonded state is less prevalent

Treatment Planning - Drawbacks

� Only 2D MR thermometry clinically available

Fast MR thermometry protocols needed:

single shot, parallel imaging, compressed sensing, etc.

BUT: system interactions of transducer <-> MR scanner

� Accurate temperature mapping:

� Robust MR thermometry protocols : corrected PRFS, fat-referenced PRFS, hybrid methods, etc.

� BUT: it is difficult to introduce new methods

� Track the heat in bone � consider other MR contrasts for temperature mapping

8

Bernstein et. al, Handbook of MRI Pulse

Sequences, 2004, Elsevier Inc.

Surface coil array for PI: SNR↑,

acquisition speed ↑

EPI

Minimizing eddy currents induced in the ground plane of a large phased-array ultrasound applicator for EPI-based MR thermometry.

9

Lechner-Greite et al., 23rd ISMRM, Toronto, 2015, #2239

Robust MR thermometry techniques

Fat-Referenced MRT• Automatic fat-water separation using 3-echo Dixon method.

• Uses spatial fitting of fat phase difference signals (B0 drift).

• Does not require fat and water to be in the same voxel

10

...)(ˆ 2

4

2

3210+++++=∆ yaxayaxaarf

vϕ

Hofstetter L, et. al. JMRI 2012; 36: 722-32.

water

fat

fat+water

IDEAL acquisition Fat image phase difference

map

fitted polynomial background

phase difference

),( yxfφ∆ ),(ˆ yxfφ∆

Fat-Referenced MRT

11

RMSE = 2.8 ºC

Conventional PRFS

Motion Map

Fat-Referenced

RMSE = 4.5 ºC

Conventional

Fat-Referenced

Temperature Probe

Water Image

ROI

Mixed fat-water phantom (cream)Temperature probe

Fat Image SPGR

Hofstetter L, et. al. JMRI 2012; 36: 722-32.

Clinical Application:

• RF hyperthermia as

sensitizer for

radiotherapy and

chemotherapy

• BSD-2000/3D

applicator for

treatment in pelvis

BSD-2000/3D

Hyperthermia System,

BSD Medical

Joint T1 and PRFS using balanced SSFP

Technical Goal:

• Accurate MR

temperature

mapping for

accurate dose

delivery

• Motion insensitive

• Cover muscle &

adipose tissue

within 3D volume

• Increase patient

comfort

Wu et al., 23rd ISMRM, Toronto, 2015, #3579 13

Technical Description:

• T1 temperature mapping for MRT in adipose

tissue

• combined with balanced SSFP: high SNR,

high accuracy

• combined with inversion recovery : fast T1

mapping (~20s)

• extracting phase information from bSSFP:

PRFS in muscle

Automated Research MRT Platform

14

Description:- Easy-to-use graphical user interface- Flexible programming environment (MATLAB) to allow rapid

integration of new thermometry algorithms - Technique comparison mode available- First prototype comes with PRFS & fat-referenced methods

Thermal

Applicator

MR scanner consoleMR scanner console

automated MR thermometry platformautomated MR thermometry platform

network linknetwork link

Silent MRI

15Wiesinger et al., MRM 2015: DOI 10.1002/mrm.25545

log(1/image) scaling results in CT-like contrast

log

(1/i

ma

ge

)

ima

ge

BW=±62.5kHz, FA=1.2deg, FOV=22cm, res=0.86mm, ~6min- Zero TE imaging

- Based on RUFIS (Madio 1995)

- non-selective hard pulse excitation

- 3D center-out radial k-space sampling

- min. grad. ramping:

silent, robust

- fast: TR~0.6ms

- low FA~1…4deg �

PD weighting

16

Zero TE head

BW=±125kHz, FA=1.0deg, FOV=32cm, res=1.0mm, ~8min

Wiesinger et al., MRM 2015: DOI 10.1002/mrm.25545

- Very good bone contrast –useful for treatment monitoring?

- 3D coverage –for dose planning?

Zero TE head: threshold-based segmentation

18Wiesinger et al., MRM 2015: DOI 10.1002/mrm.25545

− Implement a single modality planning workflow

� shorter treatment times, replace CT scan?

− Investigate in fast but accurate MR thermometry techniques to

− monitor 3D thermal dose during treatment

− achieve high-res anatomical imaging for treatment planning and response monitoring.

− Investigate in other MR signal contrasts � monitor temperature not only in tissue but also in bone.

− Calibrate the transducer (pre-emphasis, higher order eddy current compensated image reconstruction (magnetic field camera) )

− Simplify sequence testing (easy-to-use toolbox)

− Increase predictability of treatability: tissue property simulations

− Support treatment area identification: probabilistic atlas, functional probing

WM Proposal

19

Innovation @ GE Global Research

Global Research Center

Niskayuna, NY

India Technology Center

Bangalore, India

China Technology Center

Shanghai, China

Global Research Europe

Munich, Germany

Advanced Manufacturing &

Software Technology Center

Ann Arbor, MI

Global Software Center

Silicon Valley, CA

Brazil Technology Center

Rio de Janeiro, Brazil

• ~2000 scientists/engineers, nearly two-thirds PhDs.

• 3,615 US patents filed by GE in 2011

• One of the world's most diversified industrial research organizations, providing innovative technology for all of GE’s businesses

Acknowledgements

21

Thank you very much for your attention.

Team - GE Global Research Munich

Mingming Wu (PhD students of Technical

University Munich)

Nicolas Hehn (PhD students of Technical University Munich)

Team - GE Global Research Niskayuna:

Matthew Tarasek, PhD

Desmond Yeo, PhD

Bob Darrow, PhD

TU Munich

Axel Haase

UCH - Zurich

Beat Werner

Ernst Martin

InSightec

Eyal Zadarico

Erasmus MC

Gerard van Rhoon Maarten Paulides

References

22

[1] McDannold et al., Neurosurgery, 66:323–332, 2010.[2] Martin et al., Ann Neurol, 66:858–861, 2009.[3] Jeanmonod et al., Neurosurgery Focus, 32:1–11, 2012.[4] Coluccia et al., Journal of Therapeutic Ultrasound, 2:17, 2014.[5] Hindman et al., Journal of Chemical Physics, 44:4582–4592, 1966.[6] Ishihara et al., Magnetic Resonance in Medicine, 34:814–823, 1995.[7] De Poorter et al., Magnetic Resonance in Medicine, 33:74–81, 1995.[8] De Poorter et al., Magnetic Resonance in Medicine, 34:359–367, 1995.[9] Rieke et al., Journal of Magnetic Resonance Imaging, 27:376–390, 2008.[10] Peters et al., Magnetic Resonance in Medicine, 40:454–469, 1998.[11] Guo et al., Magnetic Resonance in Medicine, 24:903–915, 2006.[12] Bankson et al., Magnetic Resonance in Medicine, 53:658–665, 2005.[13] Weidensteiner et al., Journal of Magnetic Resonance Imaging, 19:438–46, 2004.[14] Mei et al., Magnetic Resonance in Medicine, 66:112–122, 2011.[15] Marx et al., In ISMRM Workshop on Data Sampling & Image Reconstruction, Sedona, Arizona, USA, 2013.[16] Nachman et al., In Proceedings of the 20th Annual Meeting of ISMRM, Melbourne, Australia, page 2918, 2012.[17] Gaur et al., Magnetic Resonance in Medicine, 00:00–00, 2014.[18] Köhler et al., Medical Physics, 36:3521, 2009.[19] Mougenot et al., Medical Physics, 38:272–282, 2011.[20] Stafford et al., Journal of Magnetic Resonance Imaging, 20:706–714, 2004.[21] Weidensteiner et al., Magnetic Resonance in Medicine, 50:322–330, 2003.[22] Holbrook et al., 8th International Symposium on Therapeutic Ultrasound, 1113:223–227, 2009.[23] Wiesinger et al., Magnetic Resonance in Medicine, 0:0-0, 2015.[24] Lechner-Greite et al., 23rd ISMRM, Toronto, 2015, #2239.