Spring 2014

medicalphysicsweb
review
In association with the journal Physics in Medicine & Biology
Spring 2014
MRI guides radiation treatments
Clinical treatment has now started on the ViewRay MR-guided radiation therapy system.
The ViewRay MR-guided radiotherapy system has been used for the
first time to perform real-time continuous MR imaging during a patient
treatment. This breakthrough treatment took place in January of this
year at the Siteman Cancer Center at
Barnes-Jewish Hospital and Washington University School of Medicine
(St Louis, MO), where the ViewRay
system is now in regular clinical use.
MR guidance enables the clinician
to visualize a patient’s tumour and
surrounding anatomy in real-time
and ensure that the radiation beam
remains on target as the tumour
moves. “This has never been done
before,” said Michael Saracen, ViewRay’s senior director of marketing.
“Previously, it had not been possible
to visualize what was going on during the whole treatment. Now, we
can make movies of it.” Tami Freeman caught up with Saracen to find
out more.
TF: How does ViewRay’s device
differ from other MR-guided
radiotherapy systems?
MS: ViewRay’s system uses cobalt-60
sources, rather than integrating the
MR scanner with a linac. Cobalt-60
was somewhat of an arcane technology that delivered low doses, but
ViewRay uses highly activated cobalt
and three heads that fire simultaneously to provide an equivalent
output to a 4 MV linac. The system
also uses doubly focused multileaf
collimators that provide the sharpest beam penumbra possible and
deliver equivalent performance to a
standard linac.
Welcome to
medicalphysicsweb
review, a special
supplement
brought to you
by the editors of
medicalphysicsweb.
This issue, distributed at ESTRO
33 in Vienna, Austria, and the
Joint Annual Meeting ISMRMESMRMB in Milan, Italy, brings
you a taster of our recent online
content. If you like what you
see, why not register for free
as a member – simply go to
medicalphysicsweb.org or visit
us at booth #1900 (ESTRO) or
booth #112 (ISMRM-ESMRMB).
Tami Freeman
Editor, medicalphysicsweb
Real-time monitoring: the ViewRay system includes a rotating gantry assembly with three Cobalt-60
teletherapy heads and three multileaf collimators, as well as a split-magnet MRI system for soft-tissue imaging.
ers superior imaging with no dose.
Even if you couldn’t image during
treatment, this improves patient
alignment. But because it’s MRI,
you are able to image in real time
throughout the treatment. The doctors can see the target, check it is in
range and make changes or adapt
the treatment if required. It’s not
like other systems that use surrogates to track respiratory motion.
What treatments have been
Here, you’re looking at the anatomy
performed to date?
One of the first treatments was ste- and treating the anatomy – it’s really
reotactic body radiation therapy game changing.
(SBRT) for lung cancer; 3D conformal and intensity-modulated radio- Are adaptive treatments being
therapies have also been performed. performed?
The centre has already treated a wide Knowing that this has never been
variety of disease sites including done before, the Washington Unilung, stomach, colon, bladder, ilium, versity team is moving slowly
abdomen, breast and mediansti- towards the adaptive stage. They
num, to name but a few. The system are collecting data with the goal of
is really being used for everything, defining protocols. For example, if
from palliative to curative therapies the targets are within a percentage
at all sites. Five patients have already threshold during a treatment then
completed their treatment, but most it’s good to go; if not, then maybe the
of the others are undergoing stand- doctors need to replan. With the Vieard fractionated courses of six to wRay system, you can do that at the
eight weeks.
console while the patient is on the
table. I believe in the next month or
How does the MR imaging help? two they plan to start incorporating
Replacing cone-beam CT with that as well, they’re very eager to start
high-contrast soft-tissue MRI deliv- using this.
Sign up as a member at medicalphysicsweb.org
EDITORIAL
So it’s possible to perform online
changes during treatment?
Yes the ViewRay system is capable
of doing that. When James Dempsey
designed this almost 10 years ago,
that’s what he had in mind. The system includes autocontouring tools
and ultrafast Monte Carlo dose calculation that can, for example, create
a 9-beam prostate plan in just 20 s.
about treating 20 patients a day.
Alongside the clinical treatments,
the researchers are also performing
imaging studies. They are working
on getting a bunch of abstracts out
for forthcoming meetings – there’s a
lot of activity on that front.
Are any other ViewRay systems
in use yet?
A total of six systems have been purchased, with three currently in the
Is the Siteman Cancer Center
solely using the device clinically? installation phase. One system is now
It’s a research centre, so they’re tak- being used for imaging studies at the
ing the approach that it’s not just University of Wisconsin in Madison,
and UCLA has finished the RF shielding and is now installing the system.
Typically, once a system is delivered,
it’ll be used for imaging studies for a
month or two and then the team will
install the sources and perform the
dosimetry. Interestingly, because it is
cobalt-based, and cobalt is the standard by which all radiation is measured, some commissioning elements
of the ViewRay system can actually
be much faster than for a standard
linac. The therapists have been really
keen on using the system and have
got up to speed really quickly.
“Previously, it had
not been possible to
visualize what was
going on during the
whole treatment.
Now, we can make
movies of it.”
Tami Freeman is editor of
medicalphysicsweb.
PMB UPDATE
Physics in Medicine & Biology
focuses on the application of physics
to medicine and biology and has
experienced outstanding growth in
recent years.
The journal continues to build on its
reputation for publishing excellent
research rapidly. Our 2012 impact
factor stands at 2.701*.
Editor-in-Chief:
S R Cherry
University of California, Davis, USA
iopscience.org/pmb
*As listed in the 2012 ISI Journal Citation Reports®
medicalphysicsweb review Spring 2014
Tomorrow’s Technology Today
See what is possible with these advanced treatment planning system features:
 Monte Carlo Electrons  Atlas Based Auto Segmentation  Fallback Planning
 Multi Criteria Optimization for IMRT & VMAT - Real Time Dose Adjustments with Navigational Sliders
 Protocol Based Planning  Dose Tracking / Adaptive Replanning
 Deformable Registration  Proton Planning... and much more
www.raysearchlabs.com
[email protected]
EUROPE +46 (0)8 545 061 30
NORTH AMERICA +1-877-778-3849
ASIA +46 (0)8 545 061 30
focus on: MR-guided therapy 3
How to optimize MRI-linac gradient coils
Hybrid MRI-linacs offer a promising new approach to image-guided
radiotherapy, delivering soft-tissuebased position verification during
treatment. Such systems are often
designed with split gradient coils to
accommodate the accelerator and
the patient in the gap. The rapid pulsing of these coils, however, induces
electric fields and currents within the
patient, a phenomenon that needs to
be evaluated for patient safety.
Potential health risks of electric
field induction by pulsed gradient coils include stimulation of the
peripheral nerves and muscles and/
or induction of retinal light flashes.
If the patient is positioned between
split coils at right-angles to the magnet bore, the field interactions are
very different to those seen in conventional MR scanners. In addition,
the gradient field produced outside
the imaging region by a split coil
has a different spatial distribution
to that produced by a standard nonsplit coil.
“Further research is needed to
work out the optimal design of the
gradient coils and patient positioning,” explained Stuart Crozier,
Director of Biomedical Engineering
at the University of Queensland in
Australia.
With this aim, Crozier and colleagues performed a detailed simulation study to evaluate the electric
fields and associated current densities induced in a patient being
scanned in an MRI-linac system.
Hybrid strategy
ups target dose
Hybrid adaptive radiotherapy using
MRI provides higher doses to treatment targets compared with conventional strategies in patients with
cervical cancer, according to a retrospective treatment planning study by
researchers in Canada. The strategy
exploits the soft-tissue detail provided by MRI and uses scans over the
course of therapy to both guide the
positioning of the patient’s tumour
and replan the treatment, correcting for target motion and deformation (Radiother. Oncol. doi: 10.1016/j.
radonc.2013.11.006).
“We found that the integrated
strategy is remarkably successful
even for the most difficult cases of
cervix cancer,” said Young-Bin Cho,
who conducted the study with colleagues at the Princess Margaret
Cancer Centre in Toronto.
Traditionally, cervical cancer
patients are positioned for treatment
using X-ray-based imaging of bony
landmarks in the pelvis, with a single treatment plan prepared for the
duration of therapy. However, the
treatment target can move relative
to the bony anatomy and pelvic softtissue anatomy can change significantly over the course of treatment,
Sign up as a member at medicalphysicsweb.org
Patient orientations: the human model positioned inside the coils along the radial (left) and axial (right) directions.
They studied human voxel models
in six different treatment positions,
using a range of split-geometry gradient coils (Phys. Med. Biol. 59 233).
The models were sited in the
MRI-linac along the axial or radial
direction, with the targeted regionof-interest (ROI) in the body centre
(position 1), the chest (position 2) or
the head (position 3). The researchers examined a series of activelyshielded gradient coils, including
three split transverse (x-) coils: classic connected coil (CC), classic separated coil (CS) and flanged coil (FC),
a conventional non-split x-coil (CV),
and a split longitudinal (z-) coil (ZC).
The researchers first simulated
a voxel model of the body with 4
mm resolution, to compare the
x-coil geometries. Within the ROI,
the amplitudes and distributions of
electric fields and current densities
induced by the different coils were
similar. Outside the ROI, however,
the fields and currents induced by
the split gradient coils were higher
than those due to the non-split
conventional coil. This effect was
especially pronounced in position
3, where most of the body was far
from the ROI.
With the model in the axial orientation, the largest peak current density (0.9463 A/m 2 ) occurred using
the CC in position 2, while the maximal electric field (6.655 V/m) was
seen using the CS in position 1. In the
radial position, the largest peak current density (0.2284 A/m2 ) and highest electric field (8.8244 V/m) were
both induced by the CS in position 2.
According to ICNIRP guidelines,
the 99th percentile value of the
induced electric field (E99%) should
remain below 0.8 V/m for tissues
exposed to a time-varying magnetic
field between 400 Hz and 3 kHz. The
highest value of E99% (0.9021 V/m)
was seen for the axially-oriented
model in position 1, induced by the
CC. This value was 22.6% higher than
that calculated for a conventional
non-split coil in the same position.
The researchers also simulated
voxel models with 2 mm resolution,
to study the distribution of peak
induced current densities inside the
body. For the flanged coil (FC), peak
current densities occurred in different organs for different body position, orientation and gender, while
peak electric fields mostly occurred
reducing a plan’s effectiveness.
The researchers investigated the
impact of soft-tissue MRI-based
positioning and replanning over
conventional approaches, simulating
treatments in 15 patients. Each patient
had already been treated and received
CT and MRI scans for treatment planning, plus weekly MRI scans over
their five-week treatment course. The
weekly scans were carried out using
an MRI simulator with the patient
positioned as if for treatment.
Simulated soft-tissue and bony
positioning approaches – both using
MRI – were combined with either
zero, one or four (weekly) replans.
The updated plans, generated from
the weekly scans, were applied after a
week’s delay to mimic clinical workflow. Online replans based on the
weekly MRIs were also simulated,
but these were introduced without
delay. Patient anatomy at each treatment fraction was assumed to be the
same as that seen on the preceding
weekly scan.
In general, a hybrid strategy that
used soft-tissue image guidance
and off line replanning provided
superior treatment volume coverage. The improvement was marked
in five cases identified as difficult
and increased with the frequency of
replanning. In these cases, weekly
replans resulted in D98 values – the
dose to 98% of the target volume –
that exceeded 95% of the prescribed
dose in 100% of the treatment volumes. In the non-adapted treatments that relied on bony anatomy,
the equivalent figure was only 60%.
Online replanning was the only
approach to match the hybrid strategy in dosimetry, but Cho argues this
is not the solution for routine clinical
treatments given its labour intensive
nature and associated higher costs.
“More importantly, the technology
is not yet available in clinics” said
Cho. “On the other hand, hybrid
adaptation strategy, which performs
similarly, is feasible with current
technology and requires only moderate overall expenses.”
In ongoing work managed by coauthor David Jaffray, head of Radiation Physics at Princess Margaret
Hospital, the MRI system used in
the study is being adapted to image
patients immediately before treatment while they are on the linac
couch. The mobile system sits on
rails that enable it to be removed
from the room during treatment.
The team is currently testing the
system and expects to have it up and
running later this year.
MR guidance for
cardiac ablation
Jude Dineley is a freelance science
writer and former medical physicist
based in Sydney, Australia.
MRI Interventions and Siemens
Healthcare announced an agreement to co-develop and commercialize a next-generation software
platform that will enable minimally
invasive catheter-based procedures
to be performed under real-time
MRI guidance. Today, virtually all
catheter-based interventions are
performed using fluoroscopy. The
software, to be used in combination
with companion MRI-guided catheters, will enable procedures to be performed under MRI guidance instead.
This shift is significant because MRI
provides superior visualization of
soft tissue, continuous 3D visualization and eliminates radiation exposure for the patient and physician.
The platform will serve as the
software component of MRI Interventions’ ClearTrace system (in
development and currently limited to investigational use). The full
ClearTrace system is an integrated
platform of software, reusable
hardware and disposable catheters
designed to enable real-time, MRIguided catheter interventions. MRI
Interventions’ software will be a
commercial successor to an innovative research software platform
in the skin and fat tissues.
The highest current densities
(0.7658 A/m2 ) occurred in cerebrospinal f luid (CSF) in the axiallyoriented model in position 2. The
highest electric field (4.7102 V/m)
was in the skin of the radially-oriented model in position 2, while the
largest E99% (0.9743 V/m) occurred
in the fat tissue of the radially-oriented model in position 2.
Repeating these simulations for
the split z-gradient coil (ZC) revealed
that most peak current densities
occurred in the CSF, while peak electric fields occurred mostly in skin
and fat tissues. The maximum current density (0.6006 A/m2 ) was in the
CSF of the axially-oriented model
in position 2, while the maximum
electric field (5.6157 V/m) was in the
skin of the axially-oriented model in
position 3. The largest E99% (0.6661
V/m) was in the fat tissue of the radially-oriented model in position 2.
The authors concluded that the
amplitudes and distributions of the
current density depended both on
the body position and orientation. In
addition, they noted that both electric fields and current densities were
higher for the axially-oriented than
the radially-oriented models. “The
findings in this paper can help when
designing and implementing split
gradient coils,” Crozier told medicalphysicsweb. “They could also be used
to assist clinicians in the best placement of the patient within the MRIlinac unit during treatment.”
created by Siemens. Under a 2009
agreement, the companies worked
together on the development of the
research platform, specifically for
use in MRI-guided cardiac ablation
procedures with MRI Interventions’
catheters. Under this new agreement,
they plan to develop a commercial
version of the research platform,
which MRI Interventions will then
sell as its own product.
“We are pleased to continue our
strong working relationship with
MRI Interventions,” said Robert
Krieg, vice-president of MR product
innovation and definition at Siemens
Healthcare. “We see tremendous
potential to improve patient care by
further expanding the therapeutic
uses of MRI.”
MRI Interventions’ first product,
the ClearPoint Neuro Intervention
System, is in commercial use for
MRI-guided, minimally invasive
brain surgery. ClearPoint offers
neurosurgeons real-time 3D visualization of the target neuro-anatomy and surgical instruments with
no radiation exposure. MRI Interventions seeks to bring these same
capabilities to catheter-based procedures outside of the brain, through
its development of the ClearTrace
system, with an initial focus on cardiac ablation procedures to treat
arrhythmias.
medicalphysicsweb review Spring 2014
Physics in Medicine & Biology
The leading international journal of biomedical physics
Fast publication • Worldwide visibility • High impact
Editor-in-Chief: S R Cherry, University of California, Davis, USA
If you are working in any of the following areas then we
would like to invite your submissions:
• all areas of radiotherapy physics
• radiation dosimetry (ionizing and non-ionizing radiation)
• biomedical imaging (e.g. x-ray, MR, ultrasound, optical, nuclear medicine)
• image reconstruction and kinetic modelling
• image analysis and computer-aided detection
• other radiation medicine applications
IMPACT
• therapies (including non-ionizing radiation)
FACTOR
• biomedical optics
2.701*
• radiation protection
• radiobiology
See a Live Demo at ESTRO!
* As listed in ISI®’s 2012
Science Citation Index
Journal citation reports
Image: Reconstructed image for a human dynamic Tc-99m-sestamibi
cardiac SPECT study. G T Gullberg et al 2010 Phys. Med. Biol. 55
R111
Sun Nuclear | Booth #5500
For more information, visit iopscience.org/pmb orVisit
e-mail
us at [email protected].
See a special presentation at our Satellite Symposium and In-Booth T
Plan Quality Distribution
Metric Scoresheet
Plan Summary
Quality Reports [EMR]®
See a Live Demo at ESTRO!
Plan Quality Management
Visit Sun Nuclear | Booth #5500
Filling the critical gap in plan quality assessment and electronic
medical records (EMR) for radiation oncology.
Attend our ESTRO Satellite Symposium
and In-Booth Talks
Vienna, Austria | April 4-8
• Plan quality metrics enable quality management and continual improvement
• Facilitates powerful and accessible electronic plan review and chart rounds
View Program Agenda and Register
Online www.sunnuclear.com/ESTRO
• Automates comprehensive and customized reports vital for billing and audits
• Compatible with all TPS systems, delivery modalities, and R&V systems
Your Most Valuable QA and Dosimetry Tools
focus on: MRI 5
PET/MR shows promise for brain studies
MRI addresses
lung screening
Screening high-risk individuals for
lung cancer can identify malignant
nodules sooner and at earlier stages
than they might otherwise have been
detected, saving lives in the process.
But every CT scan performed, even
at a low dose, confers a small risk of
causing cancer as a result of cumulative radiation dose exposure. Could
a non-ionizing MRI scan provide a
viable diagnostic option?
German researchers have conducted a study to investigate the
potential of MRI for detecting
cancer-suspicious lung lesions.
Gregor Sommer, from the department of radiology at the German
Cancer Research Center (DKFZ) in
Heidelberg, and co-researchers prospectively recruited 49 participants
of the ongoing German Lung Cancer Screening and Intervention trial
(LUSI), who had either a pulmonary
lesion larger than 10 mm or a lesion
with calculated doubling time less
than 400 days at follow-up. All had
been immediately recalled for another
CT exam according to LUSI protocol
and, once enrolled in the study, an
MRI exam (Eur. J. Radiol. 83 600).
T hir ty chest M R I datasets,
recorded using identical sequences
to those used for the study particiSign up as a member at medicalphysicsweb.org
Whole-body PET/MRI: the Biograph mMR from Siemens Healthcare.
such studies are missing so far.”
Schmidt and colleagues first evaluated the quantification accuracy of
small lesions measured using two
APD-based PET/MRI systems: the
BrainPET brain scanner and Biograph mMR whole-body scanner.
For comparison, they also assessed
the Biograph mCT, a state-of-the-art
PET/CT scanner (Invest. Radiol. doi:
10.1097/RLI.0000000000000021).
The researchers used the three systems to acquire PET images of a cylindrical phantom containing eight
spheres (with diameters of 4–8 mm)
filled with [18F]-fluoride solution at
different positions inside the PET
field-of-view (FoV). They calculated
“recovery values” for all images,
defined as the measured activity concentration divided by the decay-cor-
rected true activity concentration.
The BrainPET system exhibited
the highest recovery values – up to
99.0% for 8 mm spheres – and could
better detect the smaller spheres. The
whole-body PET/MRI, meanwhile,
showed similar performance to the
PET/CT scanner, with recovery values up to approximately 60% for the
8 mm spheres. However, the variability of recovery values for different
sphere positions was higher for the
BrainPET (up to 7.4%) than the PET/
MRI (up to 4.1%) and PET/CT (up to
4.3%) systems.
The use of MR-derived (rather than
CT-based) attenuation correction to
reconstruct images from the wholebody scanner led to an underestimation in PET activity of up to 7.2%. The
researchers also evaluated the effect
pants, were selected from healthy
individuals. Two board-certified
radiologists then interpreted the
anonymized combined exams. The
radiologists indicated the presence
of suspicious lesions and their level
of diagnostic confidence. For positive findings, they provided the size
and position of each detected lesion
and indicated the sequence that was
most useful for diagnosis. A third
radiologist unblinded the data and
compared the positions of the MRdetected pulmonary nodules to the
corresponding CT data sets.
The CT exams showed one suspicious lesion (with an average
maximum diameter of 15 mm) in
46 studies, and two lesions in the
remaining four datasets. On MRI, one
radiologist identified half of these (27)
lesions; the other radiologist identified 25. Two malignant nodules were
missed by both radiologists. The false
negative ratio was 52%. The overall
sensitivity of MRI was 48% (26/54)
and the specificity was 88% (29/33).
The sensitivity of the MR images was
78% for malignant nodules and 36%
for benign ones.
Based on the findings of their study,
the researchers believe that MRI has
the potential to replace low-dose CT
as the primary tool for lung cancer
screening, provided that its sensitivity and specificity for detection
of malignant lung nodules can be
confirmed in further clinical trials
as being greater than 75% and 90%,
respectively. They pointed out that
MRI offers the possibility of shortening screening intervals for individuals
at very high risk. They also suggest
that MRI may be used to supplement a
low-dose CT study when clinical scenarios merit, due to its sensitivity for
identifying malignant nodules.
“The expense of an MRI exam
might avoid futile surgery or bronchoscopy by increasing the specificity and predictive value of the
screening test,” Sommer told medicalphysicsweb. “As these invasive procedures are much more expensive than
any imaging exam, I see a potential
benefit of using MRI instead of CT
from an economic point of view.”
The authors also suggested that
simultaneous evaluation of MRI
along with the large prospective CT
screening trials to assess the benefit
of cancer-specific mortality is warranted. However, Sommer does not
anticipate that this will happen soon.
“As the prevalence of malignancy even
in a high-risk population is very low,
a prospective validation of MRI for
lung cancer screening would require
inclusion of thousands of individuals,”
he said.
Cynthia E Keen is a freelance
journalist specializing in medicine and
healthcare-related innovations.
of activity outside the PET FoV on
PET quantification. For the wholebody PET/CT and PET/MRI systems,
high activity concentrations outside
the PET FoV had little or no influence
on sphere activity quantification.
The BrainPET, however, exhibited
an activity-dependent recovery gradient, resulting in activity underestimations of up to 80% near the
outside source. This is most likely
due to erroneous scatter correction.
To investigate the long-term PET
stability of the whole-body PET/MRI
system, the researchers recorded a
5 min scan of a [68Ge]-cylinder phantom 14 times over eight days. The PET
detector showed good long-term stability, with standard deviations (SD)
over the 14 measurements of 0.10%,
0.10%, 0.11% and 0.03%, for prompt,
random and true counts and single
count rate, respectively. The SD of
the recovery values in the resulting
PET images was 0.3%.
Energy spectra of each crystal and
position profiles of each detector
block revealed no distinct deviations
between the different measurements. This long-term stability
implies that reliable PET measurements can be performed, with no
problems due to temperature variations of the APDs causing gain variation. They also checked the influence
of MR scanning during PET acquisition, finding no influence of simultaneous MRI on counts, count rates or
recovery values.
For functional MRI (fMRI) studies,
the MR stability of the scanner must
be tested according to the functional
Bioinformatics Research Network
(fBIRN) protocol. The researchers
tested the stability of the wholebody PET/MRI system by scanning
a stability phantom using an echo
planar imaging sequence for bloodoxygen-level-dependent measurements. They performed fMRI with
an idle PET system, with [68Ge]-rod
sources inside the PET FoV and with
the PET system turned off and saw no
PET-related changes. Measurements
showed that the percentage fluctuation (0.07% for all PET states) and drift
(between 0.16 and 0.28) were well
below the fBIRN recommendations.
Schmidt and colleagues concluded
that the homogeneity and accuracy
of APD-based PET detectors were
comparable with those of bulky
PMT-based detectors. The stability
of the whole-body PET/MRI system
was also comparable with that of
stand-alone systems, allowing for
quantitative PET and fMRI measurements, assuming that suitable attenuation correction is applied.
As the BrainPET system showed
problems with stability, the researchers plan to use the whole-body PET/
MR system for future functional brain
studies. “We are performing a clinical
trial on brain tumours and are planning several trials for investigating
different diseases using diverse short
half-life radiotracers,” said Schmidt.
Phantoms eye ultralow-field MRI
NIST
Simultaneous PET/MRI combines
the sensitive molecular imaging of
PET with soft-tissue contrast and
functional information from MRI.
One potentially important application for PET/MRI is quantitative
brain studies, which demand high
performance and stability from
both modalities. With this in mind,
researchers from the University of
Tübingen in Germany have evaluated the quantification accuracy,
homogeneity and stability of two
PET/MRI systems.
To prevent adverse interaction
between the two instruments, simultaneous PET/MRI systems employ
avalanche photodiodes (APDs)
instead of photomultiplier tubes
(PMTs) in the PET detector. APDs,
however, are highly sensitivity to
temperature changes or bias voltage
shifts. PET detectors based on APDs
must therefore be systematically
evaluated before use in demanding
investigations.
“High stability of the PET/MR system is a prerequisite for functional
brain studies,” explained lead author
Holger Schmidt. “Our main motivation was to test whether existing
APD-based PET/MR systems are
capable of performing dynamic PET
scans using short half-life radiotracers, such as cerebral blood flow studies with 15O-labelled water. While
standard PET and MR quality checks
are already carried out, more detailed
investigations of the feasibility of
Calibration tool: NIST physicist Michael Boss positions the prototype
phantom in an experimental ultralow-field MRI scanner at UC Berkeley.
The National Institute of Standards
and Technology (NIST) in the US
has developed two prototype phantoms for calibrating ultralow-field
(ULF) MRI systems, which operate at
microtesla magnetic field strengths.
“Tissues that may look the same in
clinical MRI can look very different
in ULF-MRI, which provides new
contrast mechanisms,” explained
NIST physicist Michael Boss. “Our
hope is that we can move this technique along to attract more interest
from [industry] vendors.” ULF-MRI
scanners are also simpler in design,
lighter and less expensive than regular MRI systems.
The NIST prototypes offer a quan-
titative means to assess ULF-MRI
performance, validate the technique, and directly compare different experimental and clinical MRI
scanners. The phantoms comprise
short plastic cylinders containing
six or 10 jars filled with various salt
solutions that become magnetized
in a magnetic field. Tests using conventional MRI at 1.5 and 3 T, and an
experimental ULF-MRI scanner at
between 107 and 128 µT, showed that
the prototype phantoms were wellmatched to ULF-MRI applications
and allowed direct comparison of
ULF and clinical MRI performance
(Magn. Reson. Med. doi: 10.1002/
mrm.25060).
medicalphysicsweb review Spring 2014
6 focus on: MRI
Scans identify
cancer spread
Metabolic MRI detects heart disease
Whole-body, diffusion-weighted
MRI can reveal the spread of cancer
in patients with myeloma, reducing
the reliance on painful bone marrow biopsies. That’s the conclusion
of a study from the UK’s Institute
of Cancer Research (ICR) and Royal
Marsden NHS Foundation Trust.
Twenty-six myeloma patients were
scanned before and after treatment.
In 86% of cases, experienced doctors
correctly identified whether patients
responded to treatment; accurately
identifying non-responders in
80% of cases. The researchers also
assessed visible changes on the scans
using the apparent diffusion coefficient (ADC). Changes in ADC correctly identified treatment response
for 24 of 25 myeloma patients (Radiology doi: 10.1148/radiol.13131529).
“This is the first time we’ve been
able to obtain information from all
the bones in the entire body for myeloma in one scan without having to
rely on individual bone X-rays,” said
ICR’s Nandita deSouza. “We can look
on the screen and see straight away
where the cancer is and measure
how severe it is. The scan is better
than blood tests, which don’t tell us
in which bones the cancer is located.
It also reduces the need for uncomfortable biopsies, which don’t reveal
the extent or severity of the disease.”
A metabolic MRI technique has
the potential to detect the ischaemic myocardium associated with
early-stage heart disease with
unprecedented sensitivity, according to a proof-of-concept study
by researchers at the University of
Philadelphia in Pennsylvania. The
researchers used creatine chemical exchange saturation transfer
(CrEST) MRI to map deficiencies in
the metabolite produced by necrotic
myocardium. Unlike other cardiac
imaging techniques, CrEST does not
need a contrast agent or expose the
patient to ionizing radiation (Nature
Medicine 20 209).
“Measuring creatine with CrEST
is a promising technique that has the
potential to improve clinical decision
making while treating patients with
heart disorders and other diseases,
as well as spotting problems sooner,”
said senior author Ravinder Reddy.
In clinical practice, ischaemic
myocardium can be identified with
nuclear medicine stress tests using
radiopharmaceuticals, but they
expose the patient to ionizing radiation. A second clinical approach,
anatomical MRI using a gadolinium
contrast agent GdDTPA, can distinguish between healthy and necrotic
myocardium, but cannot identify
ischaemia in tissue that is still viable.
Nature Medicine
CrEST contrast (%)
14
0
Creatine maps: healthy myocardium (left) shows noticeably more creatine
than infarcted heart tissue (right). Arrow indicates infarcted tissue region.
The technique is also contraindicated
in patients with severe kidney disease.
Measurements of local changes in
metabolism that occur with earlystage heart disease, before myocardium suffers permanent damage,
offer an alternative approach.
CrEST works on a standard 3 tesla
scanner and exploits the naturally
occurring exchange of protons
between the amine groups of the
creatine and surrounding water
molecules, which happens around
1000 times each second. During the
scan, the nuclear magnetizations
of the amine protons are saturated
using a radiofrequency pulse from
the MRI scanner specially tuned to
their resonant frequency, making
their net signal zero. As the protons
are exchanged, more saturated protons accumulate in the water, reducing the signal from the water by an
amount proportional to the concentration of creatine in the tissue.
The CrEST image is formed by subtracting the saturated image from
a second non-saturated image and
normalizing the difference using the
non-saturated image.
T he researchers first compared the CrEST against 1H MR
spectroscopy(MRS) using excised
samples of partially necrotic myocardium taken from a pig. CrEST
was significantly more sensitive,
producing MR signals over 70 times
greater than MRS measurements on
the same samples.
In a second study, scans of two
healthy pigs in vivo showed uniform distributions of creatine. In
two sheep and one pig that had
partially necrotic myocardium,
CrEST-measured creatine levels
were significantly lower in these
regions compared to surrounding
healthy tissue. Finally, in an analogy
of a cardiac stress test, the researchers scanned the calf muscles of three
volunteers after exercise. CrEST
detected the expected rise and subsequent fall of creatine generated
by the muscle activity and the variations were validated by 31P MRS
measurements made under identical
conditions.
In ongoing work, the researchers
plan to demonstrate the differentiation of ischaemic but viable myocardium from necrotic tissue in animal
models using CrEST, and are optimizing the technique for the first in
vivo cardiac scans in humans. Clinically, Reddy says CrEST could eventually be used for both diagnosis and
post-treatment follow-up.
Other potential applications
include the diagnosis and further
study of other heart pathology,
including cardiac hypertrophy and
neurological conditions associated
with creatine deficiency.
ESTRO
VISIT US AT ESTRO
BOOTH 4700
ON YOUR MARKERS, GET SET, GO!
The new CARINAnav laser control
makes patient marking a breeze.
Move to your preferred working location
é it’s portable.
Do it your way
é it’s flexible.
Just get it going
é it’s intuitive and easy to handle.
CARINAnav – SMART LASER CONTROL.
www.LAP-LASER.com
medicalphysicsweb review Spring 2014
Sign up as a member at medicalphysicsweb.org
focus on: MRI 7
A PET insert that enables simultaneous PET/MRI scans overcomes
previous hurdles relating to the combined use of PET and functional MRI
(fMRI) to analyse brain function. The
insert enables the combination of
images from both modalities, allowing researchers to directly compare
PET and fMRI measurements showing the brain during activation and at
rest (Nature Medicine 19 1184).
The feasibility of combined PET/
MRI in small animals and humans
has been previously demonstrated.
However, the technique has not been
used before to compare brain function between these two modalities by
acquiring highly spatially and temporally correlated multiparametric data,
according to the study authors.
The researchers, from the Werner Siemens Imaging Center at the
Eberhard Karls University Tübingen in Germany, have begun to use
simultaneous PET/MRI to explore
functional processes in the brains
of laboratory rats. Bernd Pichler,
chair of the centre’s Department of
Preclinical Imaging and Radiopharmacy is overseeing the project
with lead author and post-doctoral
research fellow Hans Wehrl. The PET
insert tool was developed and built at
the University of Tübingen.
Functional MRI uses blood oxygen
level-dependent (BOLD) contrasts
and allows researchers to depict
changes in blood oxygenation that
are associated with brain function.
The BOLD effect displays the complex interactions of blood oxygenation and cerebral metabolic rate of
oxygen (CMRO2 ), cerebral blood
flow (CBF) and changes in cerebral
blood volume (CBV).
PET imaging, meanwhile, can
help explain the metabolic basis of
the fMRI signal and provide complementary information. The PET tracer
can quantify changes in brain activity that are reflected by increases and
decreases in the cerebral metabolic
rate of glucose, as well as the baseline
brain activity resulting from excitatory synaptic activity and metabolic
plasticity.
The researchers explained that
simultaneous [18 F]FDG-PET and
BOLD-fMRI acquisitions allowed
the study of neuronal activity using
both PET, which provided glucosemetabolic level data over a period of
minutes using a sustained stimulus,
and fMRI, which provided vascularlevel data and information about
oxygen metabolism in a few seconds
using block-design stimulus. The
stimulus consisted of stimulating the
whisker pads of eight laboratory rats.
When the researchers stimulated
a whisker pad in 30 s time blocks,
BOLD-fMRI images showed that
the contralateral whisker barrel field
of the brain was strongly activated.
They also observed strong BOLD
activation in several other areas of
the brain.
The PET stimulus was permanently
applied for 45 minutes. The data
acquired during this time revealed
several additional activated areas
that were not seen on the BOLDfMRI images, and which are known
to be involved in pain processing,
emotion and attention. In total, the
fMRI images identified a total of
nine known neural networks. The
PET images identified seven glucose
metabolism-related networks.
In certain brain regions, the
researchers identified mismatches
between glucose metabolismrelated brain activation measured
with PET and oxygenation-related
signals measured with MRI. They
attributed this to the fact that PET
and fMRI methods are sensitive to
different components of the metabolic-haemodynamic coupling.
They suggested that further PET/
MRI studies using the capability to
observe multiple metabolic scales
could help clarify these issues.
“PET/MRI is an excellent tool
that can be used to explore and
potentially help to explain the enigmatic nature of the BOLD signal in
activated and resting states. Our
study clearly shows that an interplay between multiple techniques
and disciplines is needed to further
enhance knowledge about how the
brain works,” the authors concluded.
Sodium MRI aids breast imaging
Mark Philbrick/BYU Photo
A close look at
neural activity
Increasing the accuracy: Brigham Young University’s Neal Bangerter and
colleagues are developing new MR methods for breast cancer screening.
Investigators from Brigham Young
University and the University of
Utah have created a breast cancer
screening device based on sodium
MRI, which acquires MR data from
sodium nuclei (rather than the hydrogen nuclei used in the vast majority
of MR images). As sodium concentrations are thought to increase in
malignant tumours, the method
could potentially improve assessment of breast lesions and reduce
false positives. The team developed
a novel composite array – comprising a 7-channel sodium receive array,
a large sodium transmit coil, and a
4-channel proton transceive array
– for combined proton/sodium 3T
breast MRI (Magn. Reson. Med. doi:
10.1002/mrm.24860).
The device produced images with
up to five times better signal-to-noise
ratio than previous sodium imaging
efforts. High-quality images were
obtained in 20 minutes, suggesting
that sodium MRI breast scans could
be implemented clinically. “The
images we’re obtaining show a substantial improvement over anything
that we’ve seen using this particular MRI technique for breast cancer
imaging,” said senior author Neal
Bangerter. “We believe this can aid
in early breast cancer detection and
characterization while also improving cancer treatment and monitoring.” The team’s goal is to produce
a device capable of obtaining both
sodium and proton images without
repositioning the patient.
™
STEREOTACTIC END-TO-END VERIFICATION
Sign up as a member at medicalphysicsweb.org
medicalphysicsweb review Spring 2014
Confidence
Consistency
Efficiency
RapidPlan™ – Knowledge-Based Planning
RapidPlan™ knowledge-based planning opens the door to the next generation
of individualized treatment planning by giving clinicians the confidence to treat
a wide range of cancer types using knowledge-based planning.
By providing access to pre-configured plan models, RapidPlan may help clinics
reduce variability in treatment planning to achieve greater consistency, efficiency
and quality in patient care.
Streamline your planning process!
Varian Medical Systems International AG
Cham, Switzerland
Phone +41 - 41 -749 88 44
www.varian.com/rapidplan
[email protected]
focus on: radiotherapy 9
First patient treated with DMLC tracking
A dynamic multileaf collimator has
been used to track a moving tumour
with a therapeutic X-ray beam for the
first time in Sydney, Australia.
Last November, Tony Saunders, a
78-year-old prostate-cancer patient
from the city received radiotherapy
using the adaptive technology that
corrects for tumour motion during
treatment. His treatment at Royal
North Shore Hospital is part of a
clinical trial supported by Varian
Medical Systems that will quantify
the dosimetric and clinical benefits
of DMLC tracking over conventional
non-tracking treatments.
“This trial is the first clinical
implementation of MLC tracking,
and with Calypso guidance represents the most accurate and precise
treatment of prostate cancer,” said
Paul Keall, director of the Radiation
Physics Laboratory at the University
of Sydney, who started to develop
the technique in 1999 while an assistant professor at Virginia Commonwealth University in Richmond.
“It does feel really wonderful to see
the technology that so many people
have worked so hard on for so long
come to fruition and be realised in a
patient treatment,” said Keall.
Implemented on a Varian Trilogy
linear accelerator at the Northern
Sydney Cancer Centre, the technique adapts the treatment aperture
to track the moving tumour during
treatment. Real-time data on the
tumour location is provided by the
Calypso 4D localization system that
uses electromagnetic transponders
implanted in the prostate. Unlike
How to view the
treatment beam
Air scintillation – the weak emission
of UV light due to excitation of nitrogen gas by ionizing radiation – has
previously been used to study cosmic ray showers entering the atmosphere and as a means to count alpha
emitters. Previously, however, the
technique has not been considered in
the context of clinical radiotherapy.
Now, researchers from Stanford
University School of Medicine have
assessed whether air scintillation
produced during radiotherapy can
be visualized and used to monitor
the treatment beam. They found
that an electron-multiplication CCD
camera could successfully image air
scintillation induced by megavoltage electron beams and kilovoltage
X-ray beams (Med. Phys. 41 010702).
“We believe that air scintillation
gives us a unique opportunity to
look at patient radiation treatment
over many fractions, without any
disruption in clinical workf low,”
lead author Guillem Pratx told
medicalphysicsweb. “We often assume
that treatment machines always
deliver the exact same radiation
dose during fractionated treatment
Sign up as a member at medicalphysicsweb.org
passed to the DMLC controller for
actuation, adapting the treatment
aperture within 220 ms of detecting
a change in tumour position. Future
versions will adapt to tumour rotation and deformation.
The trial is the next step in the
development of the technology that
has previously been tested with clinical treatment plans delivered to a
moving phantom and in vivo in pigs.
Thirty prostate-cancer patients will
receive 40 fractions of volumetricmodulated arc therapy (VMAT)
using the tracking technology. Data
collection is estimated to finish by
mid-2015.
The recorded motions of the
tumour and the DMLC will be used
to reconstruct the delivered dose
distributions, as well as the distribution that would have been delivered
had DMLC tracking not been used.
A comparison of the two, along with
the original treatment plan, will enable improvements in tumour coverage and bladder and rectum sparing
to be quantified.
“We hypothesise that Calypsoguided MLC tracking will reduce
tumour underdose and associated normal tissue overdose from
30% , observed in some cases, to
less than 3% in all cases,” Keall told
medicalphysicsweb.
Keall and his co-workers plan to
publish preliminary results after
the accrual of data from the first 10
patients. Beyond the trial, they hope
to exploit the accuracy of the technique to provide stereotactic body
radiotherapy for prostate patients
for the first time, condensing treatment courses from 40 fractions
down to five. They also plan to apply
the technique to other tumour sites
in the abdomen and thorax where
more extreme intrafraction motion
occurs, including the lungs and
pancreas.
and then averaged the remaining
image data.
Air scintillation images of a 6 MeV
electron beam at dose rates of 100,
60 0 and 10 0 0 MU/min clearly
showed the beam outline, even at 100
MU/min. The scintillation intensity
was proportional to the dose rate.
The researchers then fixed the
dose rate at 1000 MU/min and
imaged beam energies of 6, 12 and
20 MeV. Line profiles through the
scintillation images (perpendicular
to the beam) showed that as energy
increased, the divergence of the
beam decreased and the penumbra improved. The general appearance of the beams was comparable
to that seen in Monte Carlo dose
simulations.
The team also imaged the electron beam with a transparent head
phantom placed in its path, to assess
the impact of Cerenkov radiation –
another luminescent phenomenon
that arises from interaction with
ionizing radiation. “Because Cerenkov radiation is brighter that air
scintillation, we were worried that it
might overshadow weaker air scintillation,” Pratx explained.
They observed intense Cerenkov
radiation in the region of the head
phantom, but the air scintillation
could still be observed. “Our results
showed that Cerenkov and air scintillation are localized within different
parts of the image and thus do not
interfere with each other,” said Pratx.
Finally, the team visualized air
scintillation produced during a total
skin irradiation treatment. Using the
same approach as previously (averaging 20 five-second frames), they
were able to image air scintillation
from a treatment comprising two
angled 9 MeV electron fields. Each
field was imaged in real-time and the
total irradiation was visualized by
superposing the images. Overall, the
combined electron fluence appeared
spatially uniform throughout the
field-of-view.
“Air scintillation may be useful
for total body and skin irradiation
treatments because it can provide
dosimetry over a very large area in a
single image,” Pratx explained. “We
are moving this technology towards
the clinic. There are only a few technical issues left be resolved, mainly
having to do with room lights. We
are implementing a time-gating system to allow for room lights during
air scintillation imaging. Once this is
completed, we will acquire the first
images of air scintillation during
patient treatment.”
Treatment dry run: Paul Keall, developer of the DMLC tracking technology, with the patient Tony Saunders.
other previously implemented
tumour tracking approaches, DMLC
tracking does not rely on a surrogate for tumour motion, such as
the patient’s external contour, and
therefore also does not require a correlation model to relate tumour and
surrogate motion.
as measured during QA; with air
scintillation, we have a non-intrusive
way of verifying this assumption.”
“The main application we envision
is beam monitoring during treatment. We think that there would be
an advantage in having a camera in
the treatment room that monitors
radiation delivery in real-time. This
could be used to document every
treatment fraction, or to discover
hardware malfunctions and other
errors before they can lead to patient
harm.”
Pratx and colleagues first investigated air scintillation produced by
a 50 kVp/30 mA X-ray beam. They
prepared the room by eliminating
background light sources and placed
the camera 3.4 m from the source.
Images showing the conical shape
of the X-ray beam could be produced with an exposure time of just
10 s. The intensity of the air scintillation – which is produced mostly in
the 300–430 nm range – decreased
with distance from the source as the
beam diverged.
Next, the researchers imaged a
megavoltage electron beam inside
a linac vault. Here, they could not
remove all sources of ambient light
and instead used a short-pass filter
to block photons above 440 nm.
The positional data are fed into the
DMLC tracking software and used
to translate the treatment aperture
according to the tumour motion.
An updated set of leaf positions are
then calculated, within system constraints, to deliver the new aperture.
The modified leaf sequences are
Air scintillation: visualization of a
20 MeV electron beam.
The camera was placed 4 m from
the beam, which was aimed at the
ground. To address noise in the CCD
sensor, they acquired 20 images with
5 s exposure time each, removed any
hot pixels (due to energetic photons
interacting directly on the CCD)
medicalphysicsweb review Spring 2014
10 focus on: radiotherapy
Tabletop source eyes microbeam therapy
Microbeam radiation therapy (MRT),
in which a tumour is irradiated with
an array of high-dose, microscopically thin X-ray beams, shows great
promise for brain tumour treatments. Studies in animal models
demonstrate that MRT can selectively eradicate tumour cells while
sparing normal tissue. However,
the technique’s current reliance on
a synchrotron radiation source is a
major roadblock for potential clinical translation.
In a quest for an alternative X-ray
source for microbeam radiation,
researchers from the University of
North Carolina (UNC) at Chapel
Hill have developed a tabletop MRT
system based on a carbon nanotube
(CNT) field emission source array
technology (invented in the laboratory of Jianping Lu and Otto Zhou at
UNC and commercialized by startup XinRay Systems). Using a hybrid
image-guided MRT protocol, they
have determined the microbeam
targeting accuracy in mice bearing
small brain tumours (Phys. Med. Biol.
59 1283).
“Instead of a single small focal spot,
a distributed CNT X-ray source array
spreads the thermal power over a
long narrow focal track on the anode.
It’s equivalent to having a large number of high-power X-ray tubes irradiating the target at the same time,”
explained first author Lei Zhang. “In
this way, a CNT field emission cathode can generate higher microbeam
flux than conventional X-ray tubes.”
Stained images: mouse brain slice irradiated with two microbeams. The
pink strips correspond to the microbeam pattern and show the presence
of double-strand breaks; the circled area indicates the tumour site.
The prototype MRT system
includes the CNT-based irradiator
and a high-resolution CNT-based
micro-CT scanner, also developed by
the team. The irradiator comprises a
linear CNT field emission cathode
array, an electrostatic focusing lens
and a tungsten anode. For this study,
the X-rays were collimated into a 280
μm wide microbeam.
Zhang and colleagues studied 14
tumour bearing mice. To identify the
tumour location with respect to the
microbeam position, they proposed
an image guidance protocol based
on registration of X-ray images with
MR images. One day before treat-
ment, the mice were imaged with
a 9.4 T MR scanner. The 2D shape,
size and location of the tumour were
clearly visualized in the MR image
and the average tumour size was
1.4 × 2.2 mm.
On the day of irradiation, the mice
were anaesthetized and immobilized in customized holders using
hardware including two ear bars.
Two-dimensional X-ray projections
recorded by the on-board microCT scanner showed bone structures
such as the skull, jaw and spinal cord,
as well as the ear bars. These images
were scaled, aligned manually with
the sagittal MR images and used
to calculate the distance from the
tumour centre to the ear bars.
The researchers irradiated the
first two mice with a single microbeam of 138 Gy entrance dose. The
second pair was irradiated by two
parallel microbeams with centreto-centre distance of 900 μm and an
entrance dose of 108 Gy per beam.
The remaining mice were irradiated
by three microbeams, with a 900 μm
pitch, delivering 48 Gy each. The
average dose rate at the mouse head
entrance plane was 1.16 Gy/min.
To verify beam delivery, the
researchers placed Gafchromic
films at the entrance and exit planes
of the mouse head during irradiation. Selected films were scanned
and analysed to create dose profiles,
which were then compared with the
prescribed treatment plans.
An example analysis of films from
a mouse treated with three microbeams showed an entrance beam width
of 280 μm, consistent with tube calibration results. The measured peak
dose of each beam was also consistent with the prescribed value. The
peak-to-valley-dose-ratio (PVDR,
the average peak dose divided by the
average valley dose) for this animal
was roughly 16 at the entrance plane,
and 15 at the exit plane - values well
within the range of synchrotronbased MRT.
The researchers also used γ-H2AX
immunofluorescence staining as a
quantitative biomarker of radiationinduced DNA double-strand breaks
(DSB). Stained images of irradiated
mouse brain tissue slices showed
DSBs as pink strips corresponding
to the microbeam path. Of 13 mice
analysed, 11 received the prescribed
number of microbeams on the
tumour, while two received one of
three planned microbeams.
They used the stained sections
to measure the displacement from
the centre of the microbeam to the
centre of the tumour. The average
localization accuracy for the 13 mice
was 454 μm. They also manually registered the stained image back to the
corresponding pre-treatment MR
image and measured an average targeting error of 537 μm. Such values
are comparable with the targeting
accuracy reported for other (nonmicrobeam) image-guided smallanimal irradiators.
The authors conclude that their
first-generation image-guided MRT
system can generate microbeams
with the necessary energy, beam
width and PVDR to produce similar therapeutic effects to those seen
in synchrotron MRT studies. The
researchers are now constructing a
second-generation CNT irradiator
that will increase the average dose
rate by a factor of 20, enabling targeted delivery of microbeams with
peak dose matching the lower end
of doses used in synchrotron MRT.
“The second-generation CNT-MRT
system is under construction. We
expect it to be coming on line soon,”
said Zhang.
Tracking ramps IOERT precision Cerenkov light emission tracks radiotherapy dose
Image-guidance: operating theatre with multi-camera tracking system.
Researchers in Spain have developed
a multi-camera optical tracking
system for image guidance of intraoperative electron radiation therapy
(IOERT). The system, installed at
Gregorio Marañón Hospital, uses
three video monitoring cameras and
eight IR cameras to record the position of the radiotherapy applicator
in relation to the patient’s anatomy
throughout the surgical procedure.
This information is sent to the planning system, which appends the
applicator position onto the patient
CT and displays it on high-definition
medicalphysicsweb review Spring 2014
screens in the operating room.
This navigation system enables
the radiation oncologist to compare
the current position and orientation
of the radiation applicator to the
planned position, and better estimate
the dose delivered to the patient. The
researchers evaluated the precision
of the system using a phantom study
of representative IOERT scenarios.
Results revealed positioning and
orientation errors of less than 2 mm,
below the acceptance threshold for
external radiotherapy practice (Phys.
Med. Biol. 58 8769).
Scientists in the US have demonstrated that Cerenkov light emitted
during radiation therapy on a live
animal can reveal the radiation dose
being administered. The technique
could enable doctors to make sure
that the correct radiation doses are
administered to humans undergoing radiation therapy (J. Biomed. Opt.
18 110504).
It is important to carefully regulate
the amount of radiation that a patient
receives when undergoing radiation
therapy, to minimize the risk of side
effects. Unfortunately, knowing
exactly how much radiation enters the
body is tricky. Normally, treatment is
planned according to computer simulations of the body, generated with
CT scans. One or two times during a
treatment course, a patient can also
have a detector inserted into them to
estimate the amount of incident radiation. But these methods aren’t foolproof: a patient could lose weight, for
instance, leading to a relatively high
dose.
One new approach to measuring
radiation dose is to make use of Cerenkov radiation, which is emitted
when electrons and other charged
particles travel through a dielectric
medium faster than light would
travel. Cerenkov radiation is poten-
tially visible as flashes of light during
radiation therapy, as the radiation
penetrates the tissue, and is in most
cases proportional to the dose. The
trick is to see it above the ambient
light, which is much, much brighter.
Brian Pogue, Professor of Engineering at Dartmouth College in
Hanover, NH, and colleagues have
now managed to measure the Cerenkov radiation emitted during radiotherapy on a dog’s oral tumour – the
first time the technique, known as
Cerenkoscopy, has been tested successfully on a live animal. They use
a gated camera that is attuned to
bursts of light, and can remove the
background of room lights. “A very
low light signal, which is not even
visible, can be captured in real time
and displayed on a computer monitor, as the beam dose hitting the tissue,” says Pogue.
In tests on the dog during its radiation therapy, Pogue and colleagues
found that they could not only measure the Cerenkov radiation, but that
the frame rate was fast enough to
capture changes in the beam shape
over time. “The linear accelerator is
programmed to give a dynamically
changing beam, and the imaging was
able to verify that the beam shape
changed as it was programmed to
do,” says Pogue.
The main advantage for patients
would be receiving exactly the right
dose of radiation, in exactly the right
place. But Pogue believes there may
be other benefits to come. “As the
system becomes established and
more widely used, it will likely lead to
other innovations in radiation therapy delivery, because no one has ever
been able to image radiation delivery
before with real time visualization of
the deposited dose and the mapping
to the patient anatomy,” he says.
With the animal tests already successful, the researchers are moving on to human clinical trials with
12 subjects. They expect the trial to
finish early next year, and will decide
then what the next stage will be.
“The unique part about this system is it could be implemented with
essentially no extra work, because
it is simply imaging of the patient
during their normal treatment,”
commented Pogue. “The cameras
would need to be permanently
positioned in the room, but this is
already done for video imaging of the
patient and monitoring for breathing
movement.”
Jon Cartwright is a freelance
journalist based in Bristol, UK.
Sign up as a member at medicalphysicsweb.
focus on: radiotherapy 11
The impact of cell signalling
Cell response to ionizing radiation depends not only on physical
parameters such as dose, but also on
biological and genetic factors. Recent
research has revealed intercellular
signalling to be one such factor that
makes a significant contribution
to cell killing in vitro, both in irradiated cells and their non-irradiated
neighbours. Based on in vitro data,
researchers in Northern Ireland have
developed a model that incorporates
this effect and used it to assess the
radiobiological impact of intercellular signalling on clinical radiotherapy plans (Int. J. Radiat. Oncol. Biol.
Phys. 87 1148).
“One common argument brought
up against the data from the in vitro
signalling experiments is that they’re
so different to what’s normally
assumed clinically that they can’t
possibly translate into a clinical setting,” said Stephen McMahon, first
author and physicist at the Centre for
Cancer Research and Cell Biology at
Queen’s University Belfast. “As far
as we could see nobody had actually
modelled that, which was part of the
motivation for doing this work.”
I ntercellula r signa lling, or
bystander effects, occur when irradiated cells secrete a signalling
molecule that diffuses through the
surrounding cell population, initiating damage in neighbouring cells.
The group noticed the phenomenon
when they observed deaths in cells
that could not solely be attributed to
the dose they had received.
“[Potential implications] include
an effective broadening of beams,
which may reduce the benefit of
extremely tight dose conformation,
and a dependence of target dose on
the doses deposited in surrounding
Signalling-adjusted dose: comparisons of physical dose distributions
(top) and distributions adjusted for cell signalling (bottom) were
similar, though signalling resulted in higher adjusted doses outside
the treatment fields and slightly lower adjusted doses inside the fields.
tissues,” said McMahon.
The model assumes cell damage
occurs from direct “hits” of radiation and indirectly from signal
propagation between irradiated
cells and their neighbours. A simplified version of the phenomenon,
it also assumes homogeneous diffusion of cellular signals as a starting
point for investigation. The diffusion rate is a function of λ, a constant that describes the exponential
decay in signal secretion following
irradiation and R, the spatial range
of the signal.
The model was applied to calculate
IBA introduces OmniPro I’mRT+
IBA Dosimetry of Belgium released
its OmniPro I’mRT+, a new software system for intensity-modulated radiotherapy (IMRT) and
rotational pre-treatment plan
verification. “OmniPro I’mRT+,
provides additional value for our
existing MatriXX users as well
as for all physicists that require a
sophisticated, jet-fast solution for
pre-treatment IMRT and rotational
plan verification,” said Ralf Schira,
vice-president of marketing for
IBA Dosimetry. “The software was
entirely redeveloped based on input
from our vast user group, and with
a focus on workflow efficiency. The
result is a tool that fulfils the needs
of a modern and busy radiation
therapy department.”
According to IBA, the software’s
clear workflow, as well as intuitive
and easy-to-use controls, enable
faster operations. The new Central
Database access via the hospital network makes the entire plan verification process easier and much more
efficient compared to previous soluSign up as a member at medicalphysicsweb.org
tions. The different plan verification
steps - Import, Measure, Verify and
Report (including approvals) - can
now be managed through the network. This allows physicists and
dosimetrists to choose the most efficient QA workflow for not only their
department but also connected satellite hospitals. All patients, projects,
and equipment data are stored on the
central database providing increased
data safety and integrity.
“The entire software package has
been completely redesigned and
re-engineered from the ground up
to maximize efficiency and streamline the IMRT QA process,” said says
James P. Nunn, senior medical physicist at Lewis Gale Regional Cancer
Center in Pulaski, VA. “The elimination of repetitive steps allows me to
greatly reduce the time required for
patient plan verification. For example, I save linac time by preparing all
patient data beforehand in my office.
During measurements I simply load
the patient data and start right away
using my MatriXX detector.”
the survival of a population of human
fibroblast cells in 3D conformal,
intensity-modulated radiotherapy
(IMRT) and volumetric-modulated
arc therapy (VMAT) plans in 10
prostate cancer patients. Physical
dose distributions were compared
with distributions adjusted to the
uniform physical dose required to
produce the same cell death as when
signalling effects were accounted
for. A value for λ was derived from
in vitro measurements and a range
of biologically feasible values were
chosen for R.
Signalling had the greatest impact
on tissue outside of the treatment
fields. Contributions from nearby
in-field tissue resulted in greater cell
killing and signalling-adjusted doses.
For example, in the IMRT plans, the
physical median mean bladder dose
over the 10 patients was 23.5 Gy,
compared to 33.4 Gy when signalling
was accounted for.
The reverse was seen in cells in the
treatment beam’s path, though the
effect was less marked. In the IMRT
plans, the physical mean planning
target volume dose of 74.5 Gy compared to a signalling-adjusted dose
of 71.3 Gy. Here, a lack of signalling from cells outside the beams
meant the model predicted less cell
damage compared to the scenario
assumed clinically, where damage is
independent of the dose received by
neighbouring cells.
Nevertheless, adjusting for signalling resulted in no dramatic
departures from the physical dose
distributions. It also made no significant difference to comparisons of the
different treatment techniques, leading the researchers to conclude that
the in vitro observations on which
the model is based can be reconciled
with clinical experience.
“If, for example, we had predicted
massive treatment failures in IMRT
compared to 3DCRT that would have
suggested something was wrong
with the model as we know that’s not
the case clinically,” said McMahon.
“Fortunately, everything has come
out looking fairly sensible thus far,
suggesting that this is an area worth
more investigation.”
Future work by the researchers will
include comparisons of model predictions with clinical data, in vivo testing and further in vitro investigations.
Plan incorporates
tumour shrinkage
Radiation doses delivered to lung
tumours may be efficiently escalated
over a course of radiotherapy using a
method that predicts tumour shrinkage, according to a preliminary planning study by US researchers. The
technique, called robust predictive
treatment planning (RPTP), uses a
geometric model and is significantly
less labour-intensive than existing
adaptive approaches that require replanning treatments mid-course (Int.
J. Radiat. Onc. Biol. Phys. 88 446).
“In principle, RPTP should work at
least as well as adaptive radiotherapy,
but with a much reduced workload,”
said Perry Zhang, physicist at the
Memorial Sloan-Kettering Cancer
Centre, New York. Shrinking tumour
volumes are commonly observed
over the course of radiotherapy in
patients with non-small cell lung
cancer (NSCLC). Now, predicting
tumour regression before the patient
sets foot in the treatment bunker
may only require the preparation of
a single treatment plan. “We can usefully take advantage of regression as
typically seen in many lung tumours,
even though we don’t know the exact
form of shrinkage,” said Zhang.
The researchers developed the
shrinkage prediction model based
on in vivo tumour data, obtained from
one planning CT scan and weekly
cone-beam CT scans, from five
patients already treated for NSCLC.
The shrinkage of the tumour from
the volume observed in the pre-treatment CT (GTVorig ) in each patient
was then predicted using transformations derived from the other four.
Each GTVorig was then morphed
onto the predicted residual GTV
(GTVpred ) incrementally, resulting
in a total of five GTV volumes that
represented the progressive shrinkage of the tumour over the course
of treatment. Intensity-modulated
treatment plans were calculated for
planning target volumes (PTV) generated from the GTVs. PTVorig was
prescribed a standard clinical dose
that increased with each PTV shell,
where PTVpred received the maximum dose achievable that restricted
doses to surrounding organs-at-risk
(OAR) to acceptable levels.
The dosimetric performance of
RPTP was compared with three
other planning approaches, including adaptive radiotherapy. The
researchers found that PTVpred overlapped with the true PTV (PTVresid )
derived from the GTV on the final
CBCT scan by 76.0–88.3%. Averaged
over all patients, the RPTP approach
delivered the highest dose to PTVresid,
with a moderate improvement
seen over the adaptive, replanned
approach. RPTP resulted in a mean
PTVresid dose that was 5.7 Gy higher
than the adaptive approach, while
the minimum PTVresid was 0.4 Gy
higher. There were no significant differences in doses to the lungs and spinal cord between RPTP and the three
other planning approaches.
medicalphysicsweb review Spring 2014
12 focus on: particle therapy
A simpler way to image with protons
medicalphysicsweb review Spring 2014
y (cm)
6
4
100
2
80
0
60
–2
30
–4
20
–6
–6 –4
–2
2
0
x (cm)
4
WEPL (mm)
Proton radiography and proton CT
employ high-energy proton beams
to create planar or tomographic
images of patients. Using protons
rather than X-rays to image a patient
enables more accurate treatment
planning calculations. But the cost
and technical complexity of traditional proton detection systems currently limit their clinical use.
“Conventional proton radiography based on individual
tracking of protons requires fast dataacquisition systems in order to
acquire radiographies in a tolerable
time,” explained Mauro Testa, from
Massachusetts General Hospital and
Harvard University Medical School.
To address these limitations, Testa
and colleagues propose a new proton radiography technique based on
time-resolved dose measurements.
The method requires only a 2D
dosimeter array and the clinical proton beam. What’s more, it can create a
radiographic image in about 100 ms,
using a dose of just 0.7 cGy. The team
has performed a proof-of-principle
study using a prototype diode-array
detector (Phys. Med. Biol. 58 8215).
Proton therapy delivered via passive scattering uses a spinning range
modulatorMMDI_MPW_Review_HP_Ad.pdf
wheel to construct spread1
out Bragg peaks (SOBP). The dose
6
Left: proton radiography of a human skull. Right: X-ray radiography of
the skull placed over the sensitive area of the diode-array detector.
rate produced by a beam passing
through such a wheel is periodic,
with a unique pattern in time at each
point along the beam path, encoding the water equivalent path length
(WEPL) of the protons. Time-dose
patterns measured behind a patient
can therefore be used to create an
image of WEPL distribution.
To determine the WEPL of protons
traversing phantoms with arbitrary
thicknesses, the researchers first
created a database of dose-rate functions at known WEPLs from 5.5 to
20 cm in a water-phantom. Once the
database was acquired, the WEPL
through an unknown phantom
thickness
could
be determined using
2014-02-25
10:04 AM
a single dose-rate measurement at a
point distal to the phantom.
For this study, the researchers used
a prototype detector comprising a
2D octagonal array of 249 semiconductor diodes. Although commercial
ion-chamber/diode-array detectors
can acquire radiographic images,
their dose sampling time is in the
order of 20–200 ms. “Our prototype
detector allows a sampling time of
2 ms,” said Testa.
One important potential application of proton radiography is pretreatment range verification, using a
low dose “scout-beam” to image the
patient in the treatment position.
Comparing measured WEPL values
to those computed by the treatment
planning system could improve
calculation of the exact beam range
needed to cover the target volume.
When treating medulloblastoma
patients, for example, this could significantly reduce the skin dose.
With this in mind, the researchers compared proton radiography
of a human skull with conventional
X-ray imaging. The lower resolution
seen for the proton radiography was
partly due to the inherent multiple
Coulomb scattering of protons in
the patient, which leads to image
blurring. However, the detector spatial resolution played a key role, with
sub-millimetre resolution easily
achievable with a commercial X-ray
panel, but limited by the diode pitch
in the proton detector.
The team also examined the potential of real-time imaging of a Lucite
cube moving with 1D and 2D sinusoidal motion, creating movies using
a single radiographic image for each
frame. For both movies, the cube’s
shape and trajectory were distinctly
recognizable, with the centre of the
cube and the motion table position
within millimetre agreement.
Proton CT (pCT) is particularly
advantageous for treatment planning as it removes the need to convert
X-ray Hounsfield units into relative
stopping power (RSP) distributions.
Testa and colleagues performed pCT
on a cylindrical Lucite phantom
containing various rod inserts. They
recorded radiographic projections at
angles from 0 to 178.2°, using interleaved measurements to increase
spatial resolution.
Reconstructed pCT images clearly
showed the lung, air and bone inserts.
The RSP value for Lucite obtained
from the images was close to the
actual value, while RSP values calculated for the inserts were somewhat
biased towards to that of Lucite. The
accuracy of the obtained RSP values
was impacted by the detector’s limited spatial resolution. A full-scale
detector with higher spatial and time
resolution should increase calculation accuracy, as would accounting
for multiple Coulomb scattering in
the reconstruction algorithms.
“We are currently working on a
new detector design that utilizes a
scintillating screen coupled with
a fast CCD camera,” Testa told
medicalphysics web. “ This detector allows sub-millimetre spatial
resolution and a sampling time of
0.1–1 ms. We plan to use it to acquire
higher resolution CT images and to
perform an end-to-end validation
study of the range tuning method
for medulloblastoma cranial fields.”
Sign up as a member at medicalphysicsweb.org
focus on: particle therapy 13
Proton rescanning: which works best?
Active spot scanning is the delivery method of choice for most new
particle therapy facilities. Treating
mobile targets with spot scanning,
however, is limited by interplay
effects between tumour and beam
motion, which lead to target dose
inhomogeneities. One way to mitigate such effects is to use rescanning
– either volumetric scanning, where
dose is delivered across a series of
full-volume rescans, or layered rescanning, which delivers all scans in
one energy plane before moving to
the next.
To compare the two schemes,
researchers from ETH Zurich and the
Paul Scherrer Institute (PSI) in Switzerland have performed a detailed
simulation study modelling the
interplay effects for a range of system
and beam delivery scenarios. In particular, they examined the impact of
the beam position adjustment time
(BPAT), the speed at which a particular delivery system can move the
beam from one spot to the next (Phys.
Med. Biol. 58 7905).
“Active scanning is particularly
sensitive to organ motion, and has
thus been used sparingly up to now
to treat mobile tumours,” said ETH’s
Kinga Bernatowicz. “Rescanning is
one promising motion mitigation
Scanning study: the PSI team with the Gantry 2 proton therapy system.
technique, and in this study we investigated its effectiveness for different
delivery scenarios under the assumption of different delivery speeds and
characteristics typical of currently
available commercial systems.”
Bernatowicz and colleagues performed 4D dose calculations for two
clinical liver cases, with large and
small tumour volumes, simulated
for up to seven rescans. For each case,
they simulated four BPAT scenarios.
Motion was simulated with a typical
breathing period of 5 s. They assessed
the resulting plans to determine the
impact of rescanning on treatment
delivery time and dose homogeneity.
The researchers focused first on
single-field plans. They examined
four BPAT scenarios: two with a
fast lateral BPAT of 4 ms, and depth
(energy) BPATs of 80 ms and 1 s (scenarios 1 and 2, respectively); and two
with a slower lateral BPAT of 10 ms
and energy BPATs of 80 ms and 1 s
(scenarios 3 and 4, respectively).
For systems with fast energy
BPATs, dose homogeneity improved
as the number of rescans increased.
No significant difference in performance was seen between the two rescanning methods, although layered
rescanning showed a higher dependence on the selected phase.
For both patients, layered rescanning led to a smoother change in dose
homogeneity with increasing rescan
number, whilst volumetric rescanning showed more fluctuations.
For scenario 2 (fast lateral changes
and slow energy changes), results
were similar to those described
above. However, if both BPAT values
were slow (scenario 4), clear differences in the two scanning methods
appeared: layered rescanning performed best after 2–4 rescans, while
for a higher number of rescans, both
methods gave similar results.
For scenarios with fast energy
changes, layered and volumetric rescanning required similar treatment
delivery times to achieve satisfactory dose homogeneity. For systems
with longer energy adjustment
times, however, layered rescanning
achieved dose homogeneity in a
noticeably shorter treatment time.
“Our study shows a preference for
layered rescanning for systems with
slower energy changes, whereas for
faster systems, both approaches
result in similar dose homogeneity to the target in acceptable delivery time,” said Bernatowicz. “On
the other hand, the effectiveness of
volumetric rescanning appears to be
somewhat less sensitive to the starting phase of the motion.”
The researchers then examined
multiple-field plans, comparing
three rescans per field for twofield plans, two scans per field for
three-field plans and six rescans for
single-field plans. For treatment plans
coming from the anterior-superior
direction, no differences were seen
between any types of rescanning or
BPAT scenario. For treatments from
the right or superior direction, layered rescanning showed similar dose
homogeneity no matter how many
fields were applied. For volumetric
rescanning, multiple-field plans
had slightly worse dose homogeneity than single-field plans. Overall,
layered rescanning exhibited better
dose homogeneity for both fast and
slow BPAT scenarios, but was more
sensitive to the starting phase. Layered rescanning also reduced the
treatment time for slow BPATs, by
up to 300 s for scenario 4.
“Our study indicates that using
multiple SFUD [single-field uniform
dose] plans per field leads to a similar improvement in dose homogeneity as rescanning, and that smaller
amounts of rescanning per field are
necessary when using multiple field
plans,” Bernatowicz explained. The
team is now working on experimental validation of the simulated
results.
Delta --At
-At
AtTreatment
Treatment
Delta
Delta
Treatment
4AT
4AT
4AT
Your
insurance
for
patient
safety
Simple
Efficient
Accurate
Your
Your
insurance
insurance
forfor
patient
patient
safety
safety
Simple
Simple
Efficient
Efficient
Accurate
Accurate
A truly
independent
QAallowing
system
allowing
A truly
A truly
independent
independent
QA system
QA system
allowing
youperform
toyour
perform
your
in-vivo
dosimetry
you to
youperform
to
your
in-vivo
in-vivo
dosimetry
dosimetry
the background
you
focus
entirely
in theinbackground
theinbackground
whilewhile
youwhile
focus
you focus
entirely
entirely
onpatient.
the patient.
on the
onpatient.
the
4
4
4
– Confidence
on measurements
real measurements
Delta
Delta
–Delta
Confidence
– Confidence
based
based
onbased
real
on real
measurements
www.scandidos.com
www.scandidos.com
www.scandidos.com
varian.com
A New wAy to do
whAt you do best:
Fight CANCer.
Please
visit us at
ESTRO 33
Booth 2800
Proton Therapy System
By introducing proton therapy into your clinic, you can maximize your potential to deliver
treatment above and beyond conventional radiation therapy, while continuing to thrive as
you create an innovative future. Whether it’s fully integrated technology, or a robust network
of partners, we believe working together creates mutual success.
So join us with ProBeam® — the latest chapter in our mission to simplify, innovate and
support cancer-fighting solutions.
© 2014 Varian Medical Systems, Inc. All rights reserved. Varian, Varian Medical Systems and ProBeam are registered trademarks of Varian Medical Systems, Inc.
focus on: particle therapy 15
LET painting improves outcome Beam targeting:
acoustic options
LET
160
160
180
Jakob Toftegaard
200
220
104
88
180
72
200
220
240
260 280
x (mm)
300
320
220
220
56
240
260 280
x (mm)
300
320
LET maps: the non-painted plan (left) has elevated LET at the tumour
edge due to the distal portion of the Bragg peak; the painted plan (right)
also has elevated LET in the hypoxic volume, due to the ramped fields.
Heavy ion therapy produces highlinear-energy-transfer (LET) Bragg
peaks that offer one way to tackle
hypoxic – or oxygen deficient –
tumours that respond poorly to
radiotherapy. However, if delivered
across an entire tumour, high LET
can also result in the excessive exposure of adjacent healthy tissue and
unacceptable side-effects.
Instead, researchers in Denmark
and Germany are investigating the
targeted approach of LET-painting,
where high-LET radiation is steered
into the hypoxic portion of a tumour
only. In an initial modelling study,
they predict superior tumour control compared to the current clinical
standard of non-painted treatments
that assume normoxic – or oxygenated – tumour composition (Acta
Oncologica 53 25).
“LET-painting could quite dramatically increase tumour control, without increasing the side effects that
are normally associated with highLET radiation,” said Niels Bassler,
first author and physicist at Aarhus
University.
The researchers compared the
tumour control probabilities (TCP)
calculated for LET-painted and nonpainted carbon-12 and oxygen-16
treatment plans in a patient who had
already been treated for oropharyngeal cancer.
Non-painted plans mimicked
existing clinical approaches, using
two coincident pairs of opposed,
scanned beams, each delivering a
spread out Bragg peak (SOBP) that
gave a uniform dose and largely
uniform dose-averaged LET across
the entire tumour volume. TCPs for
these plans were calculated for two
scenarios: the first assumed a normoxic tumour, as is the current practice in centres delivering carbon-12
treatments, while the second took
tumour hypoxia and the associated
radioresistance into account.
Painted plans were calculated for
a range of hypoxic volume sizes
up to 10 cm3, centred on a hypoxic
volume identified in the patient’s
tumour that was delineated using
F-18 f luoroazomycin arabinoside
(FAZA) PET imaging. The same field
arrangement was used as in the nonpainted plans, but the parts of each
field incident on the hypoxic volume
were ramped, the SOBPs delivering
100% of the dose at the proximal
edge of the volume and 0% dose at
Sign up as a member at medicalphysicsweb.org
the distal edge.
The opposed ramped fields elevated dose-averaged LET across the
hypoxic volume compared to the
surrounding non-hypoxic tumour.
Total energy f luence was fixed
across all the plans, pinning it to the
value that gave a TCP of 95% in the
clinical planning standard of the
non-painted plans that assumed a
normoxic tumour. This ensured fair
comparisons of the painted and nonpainted approaches.
LET painting proved feasible,
demonstrating consistently superior TCP values to the clinical reality of a non-painted plan delivered
to a hypoxic tumour; this returned
a TCP of 0%. However, this advantage dropped off sharply as hypoxic
volume increased, with TCPs of the
order of 50% or more only observed
in volumes less than 1 cm3 in the oxygen-16 plans and less than 0.5 cm3 in
the carbon-12 plans.
The trend arises from the differences in SOBP composition required
to cover the range of hypoxic volumes. As the width of the hypoxic
volume increases, so too does the
energy range of the SOBP needed
to ensure dose coverage. At a given
depth, this translates to increasingly
large contributions to the doseaveraged LET from the plateaus of
the individual Bragg curves. These
exhibit significantly lower LET than
the peaks of the curves.
To negate the volume effects, the
researchers boosted dose to the
hypoxic volume in the oxygen-16
plans, while still limiting the total
energy fluence used to irradiate the
tumour. A boost of 5% produced
a TCP of around 50% for a 2 cm3
hypoxic volume, around double that
achieved for the non-boosted case.
Overall, carbon-12 ions resulted in
significantly lower tumour control
than the oxygen-16 ions. “We might
need ions heavier than carbon-12
ions to achieve significant increases
in tumour control,” Bassler told medicalphysicsweb. “Scanning beams of
oxygen-16 ions, which are available
at the heavy ion facility in Heidelberg,
seem to be a good candidate here.”
Bassler adds that while the results
are encouraging, much more
research is needed before LET-painting can be introduced clinically,
including similar studies with other
radiobiological models and extensive in vitro and in vivo testing.
Accurate targeting is imperative to
fully exploit the potential benefits of
ion beam therapies. Current efforts
in this area include development of
in-room imaging techniques that
use X-rays or transmitted ions for
anatomic confirmation of tumour
location. Meanwhile, range monitoring techniques that assess delivered dose via detection of emerging
prompt gammas or charged secondary particles are the focus of much
research effort.
Many of these current approaches,
however, involve complex instrumentation and bulky detectors;
what’s needed is a compact and costeffective alternative. Speaking at the
ICTR-PHE meeting, held in February
in Geneva, Switzerland, Katia Parodi
described novel investigations on
a technique that may just meet
these requirements: “ionoacoustic”
imaging.
Parodi, Chair for Medical Physics at Ludwig Maximilian University (LMU) in Munich, Germany,
explained that ionoacoustic imaging is very similar to photoacoustic imaging, in which light pulses
absorbed by tissue cause localized
thermal expansion and generate
acoustic waves. The difference is that
in this case, the therapeutic ion beam
is used to induce the heating effect.
“The idea is that most of the energy
deposition will be concentrated in
the area of the Bragg peak,” she said.
“This results in thermal expansion
and generation of an acoustic wave,
which we can then detect with an
35
x-axis (mm)
140
120
LET (keV/um)
140
y (mm)
y (mm)
LET
34
36
37
38
22
23
24
25
y-axis (mm)
26
Ionoacoustics: the 2D Bragg peak.
acoustic transducer.”
For clinical application, Parodi and
colleagues are aiming to develop an
ionoacoustic imaging system with
a penetration depth of more than
10 mm and sub-millimetre resolution. Ultimately, they hope to combine anatomic ultrasound imaging
prior (or even in parallel) to treatment with ionoacoustic imaging
of the Bragg peak during ion beam
therapy.
To test the potential of this
approach, the LMU team performed proton beam experiments
at the Maier-Leibnitz-Laboratory’s
Tandem accelerator. They used
20 MeV protons to irradiate a water
phantom with pulsed beams of
varying pulse length (1.5 ns – 4 µs).
The ensuing acoustic signals were
detected using a 3.5 MHz cylindrically focused ultrasound transducer
and a spherically focused 10 MHz
transducer.
For beam intensities above 105
protons/pulse, a clear acoustic signal
could be acquired. The signal showed
three spikes, attributed to the Bragg
peak, the entrance window and a
ref lection. The distance between
the first and second of these spikes
is proportional to the ion beam penetration depth and can be used to
assess beam range. The measured
range resolution agreed with that
seen in simulations. “Sub-millimetre
resolution of the Bragg peak position
seems to be possible, quantification
is ongoing,” said Parodi.
Next, the researchers moved onto
2D Bragg peak imaging, achieved by
x–y scanning the 10 MHz transducer.
In parallel, they imaged the Bragg
peak using Gafchromic film and performed Geant4 Monte Carlo simulations. The ionoacoustic images of the
Bragg peak agreed reasonably well
with both the film and the simulated
images, both with and without the
aluminium absorber.
Very recently, the LMU team has
begun to investigate ionoacoustic tomography, using a 64-channel ultrasound transducer array.
Parodi noted that the team has now
recorded its first 3D images. She concluded that ionoacoustic detection
offers a direct way to observe the
energy deposition of an ion beam in
tissue, with the potential for 2D and
3D imaging with sub-millimetre
range resolution. The minimum
detectable signal was so far found to
be 10 4 protons/pulse, representing a
1012 eV energy threshold.
The next steps will involve testing
the ionoacoustic effect using more
realistic targets and higher-energy
ion beams from conventional and
novel high-dose-rate sources, as well
as simulating the entire imaging process and exploring the technique’s
applicability to various clinical treatment sites.
OnQ rts®
A vendor neutral clinical software suite to
complement your treatment planning system
Automate the Complex with atlas-based
auto-contouring, deformable multi-modality
image registration, 4D contour propagation
and SUV based GTV delineation for PET.
Adapt your Workflow with DIR based
re- contouring, adaptive dose plan review
and off-line task scheduler.
Analyse your Performance with the
comparison metrics in Contour Analysis to
improve accuracy and precision.
With Task Scheduler let the software work
24/7 so you don’t have to.
ESTRO 33 - Booth 5150
www.osl.uk.com [email protected] +44 (0)1743 462694
medicalphysicsweb review Spring 2014
16 focus on: particle therapy
CERN unifies medical physics research
CERN has transferred a great deal
of technology into the field of medical physics over the years, from
adaptation of its high-energy particle detectors for PET scanning
to the design and development of
dedicated accelerators for particle
therapy. Now, CERN is consolidating all of its diverse activities in the
medical physics arena, with the
creation of a new office for medical
applications. “Since the start of this
year, we are trying to combine all
of our research on medical applications at CERN into one coordinating
office,” explained CERN’s DirectorGeneral Rolf Heuer, speaking at the
recent ICTR-PHE meeting in Geneva,
Switzerland.
Heading up the office for medical
applications is Steve Myers, formerly
CERN’s director of accelerators and
technology. Myers’ first task in his
new role was to set up a brainstorming workshop to help guide the direction of the fledgling programme. The
workshop, which took place immediately after the ICTR-PHE meeting,
saw around 80 leading experts from
around the globe come together to
discuss matters such as accelerators
and gantries, clinical perspectives,
biomedical research and radioisotope production, detectors and
dosimetry, and the application of
large-scale computing.
One of the first projects that Myers
will oversee is the transformation of
CERN’s Low Energy Ion Ring (LEIR)
into a biomedical facility. LEIR is a
At the LEIR facility: (left to right) Steve Myers, with researchers Daniel
Abler and Adriano Garrona, and CERN Senior Scientist Ugo Amaldi.
small accelerator currently used to
pre-accelerate lead ions for injection into the Large Hadron Collider
(LHC). But it only does this for several weeks each year, leaving a lot of
spare beam time. Importantly, the
ring also has an energy range that
matches that of medical accelerators
(440 MeV/u for carbon ions). CERN
has now confirmed that LEIR can be
converted into a dedicated facility for
biomedical research. This “BioLEIR”
facility will provide particle beams of
different types at various energies for
use by external researchers.
Myers explained the rationale for
developing such a facility. He pointed
out that although protons and carbon ions are already in extensive
clinical use for treating tumours,
and other species such as oxygen and
helium ions are under investigation,
there’s still a lack of controlled exper-
iments that directly compare the
effect of different ions on cancer cells
under identical conditions. “The big
advantage here is that we don’t treat
patients,” he said. “Our aim is to provide a service, so researchers don’t
have to do experiments at a clinical
site, they can come here instead.”
As well as radiobiology experiments, LEIR is lined up for use in
detector development, dosimetry
studies and basic physics investigations such as ion beam fragmentation. The facility would work in the
same way as particle physics experiments are carried out at CERN:
researchers propose experiments,
which are peer reviewed by a panel of
experts who select suitable projects,
and CERN controls the beam time
allocation.
Before this can happen, however,
LEIR needs some hardware modifi-
cations. CERN researcher Adriano
Garonna presented a status update
on the ongoing design studies. He
emphasized that LEIR will continue
to be used as a lead ion accumulator for the LHC and other heavy ion
experiments. The challenge, therefore, is to redesign LEIR for biomedical experiments, using the simplest
and most cost-effective configuration, but while maintaining its current operational performance for the
LHC.
Firstly, Garonna told the ICTRPHE delegates, the current front-end
is tailored for use with heavy ion
sources and requires at least three
weeks to swap between heavy and
light ions. What’s needed is a dedicated front-end, comprising a light
ion source and a radiofrequency
quadrupole (RFQ) optimized for
light ions, that will offer fast switching between species. The beams will
be injected into and accelerated in
LEIR up to a kinetic energy of 440
MeV/u for carbon ions. Garonna and
colleagues have identified a suitable
commercial light ion source (the
Supernanogan Electron Cyclotron
Resonance source) that can provide
hydrogen, helium, carbon, nitrogen,
oxygen and neon ions.
The next task is modification
of the extraction system, which is
currently designed to transfer highbrightness, 200 ns pulses to the Proton Synchrotron. Here, the goal is to
create a dedicated resonant extraction system that delivers long spills
(1–10 s) with uniform intensity for
biomedical experiments. Such a
slow extraction scheme could be
implemented using minimal new
hardware, explained Garonna, by
installing an electrostatic septum
and two magnetic septa. The CERN
team has also demonstrated the feasibility of two potential resonance
driving mechanisms: quadrupole
driven extraction, which is easy
to implement; and RF knock-out,
which delivers better beam quality but requires installation of new
hardware.
Finally, Garonna detailed a first
proposal for two experimental
beamlines for LEIR: a horizontal
beamline running at up to the maximum energy (440 MeV/u) and a vertical beamline with a reduced energy
of up to 75 MeV/u.
The adjacent hall next to the ring,
currently used for storage, will also
need to be fitted out as an experimental area suitable for various biomedical experiments. The total cost of the
infrastructure developments will
likely come in at around €15 million.
And while CERN has provided seed
funding for this project, the majority of the money needs to be sourced
from elsewhere. So when can we
expect to see the BioLEIR facility
opening its doors to the international research community? Myers
predicts that the full funding should
be sourced during this year. Once
this is achieved, the site could be up
and running roughly two years later.
Deformable Image Registration tools for RT planning and assessment
Multi-modal fusion
Dose warping and summation
Auto-contouring
Visit Mirada at ESTRO 2014, booth #5400
www.mirada-medical.com
medicalphysicsweb review Spring 2014
Sign up as a member at medicalphysicsweb.
focus on: nuclear medicine 17
Gelatin phantoms calibrate PET
ImSimQA™
TM
Virtual Phantoms for IGRT QA
Atlas-based auto-contouring, deformable
multi-modality image registration (DIR) and
adaptive RT are on the increase.
How can DIR software be tested and validated
with limited access to clinical scanners?
Improved quantification: a mould containing one of the gelatin spheres.
Accurate volume estimation is
imperative when using PET for
tumour staging and treatment planning. Currently, PET systems are calibrated using phantoms made from
hollow spheres filled with radioactive liquids. But the presence of a
non-radioactive wall at the edge of
such a sphere can cause inaccuracies.
Instead, researchers in Sweden suggest that a wall-less gelatin phantom
will improve PET activity quantification and volume delineation (Phys.
Med. Biol. 59 1097).
Previous studies on PET quantification phantoms have shown that
the presence of a non-active wall
decreases the lesion-to-background
contrast. This creates inaccuracies
in quantification, particularly when
imaging a small sphere where the
wall represents a large portion of its
overall volume. The presence of cold
walls has also been shown to affect
the measured standard uptake values, which reduce as wall thickness
increases.
To address these shortcomings,
Marie Sydoff and colleagues at Lund
University’s Department of Medical
Radiation Physics in Malmö have
developed a simple method for creating radionuclide-doped gelatin
phantoms without non-active walls.
To compare these with conventional
phantoms, they used a Jaszczak
phantom containing three hollow
plastic spheres (with inner diameters of 15.6, 22.5 and 27.9 mm) in
a polymethylmethacrylate (PMMA)
cylinder.
The researchers dissolved 18F-FDG
in water to an activity concentration
of approximately 80 kBq/ml, simulating a clinical situation. They added
gelatin granules to half of the solution and water to the other half. The
gelatin solution was heated, poured
into spherical aluminium moulds
(with diameters of 15.6, 22.5 and 27.9
mm) and then frozen to create solid
spheres. The other half of the solution was used to fill the hollow plastic
spheres. Finally, all six spheres were
mounted in the Jaszczak phantom.
The spheres were imaged with a
Gemini TF PET/CT system (from
Philips Healthcare) using a matrix
size of 144 × 144, a voxel size of 4 ×
4 × 4 mm, and a scan time of 8 min/
bed position. Measurements were
made with the spheres surrounded
Sign up as a member at medicalphysicsweb.org
by water (zero background), and
with background fractions of 0.08,
0.1, 0.13 and 0.2.
After PET image reconstruction
and attenuation correction, the data
were processed to calculate the background-corrected relative threshold - Tvol - for all of the spheres. Tvol
provides an intensity threshold with
which to delineate the volumetric
edges of a tumour or spherical phantom. For the largest spheres and the
lowest background fraction (0.08),
Tvol was lower for the hollow plastic
sphere (39.7%) than for the gelatin
sphere (41.4%). With the highest
background (0.2), Tvol was 33.7% for
the large plastic sphere and 41.1% for
the large gelatin sphere.
The researchers plotted Tvol as
a function of the five background
fractions. With zero background
activity, the Tvol values for the gelatin
and hollow spheres were practically
the same. As the background activity increased, however, Tvol for all of
the hollow spheres decreased while
Tvol for the gelatin spheres remained
practically constant.
“When we used the Tvol values
for volume calculations, the values
obtained from the hollow spheres
showed increasing overestimation of
volume with increasing background
activity levels,” explained Sydoff.
“This makes the Tvol values for the
gelatin spheres more accurate for
structures in an active background,
which is always the case in patient
measurements.”
The authors conclude that threshold values estimated using spheres
with non-active walls should not
be used for tumour delineation in
patients, especially for small volumes within a high-activity background. With the gelatin spheres,
however, the background corrected
threshold method can be employed
to define tumour volume from PET
images.
The researchers now plan to study
the effects of plastic walls in microPET systems, where phantom walls
are a much bigger problem, since the
volumes are far smaller. “I am also
looking at the possibilities of using
this method of volume estimation
on structures other than spheres, to
be able to implement it in real patient
measurements such as within oncology, for example,” said Sydoff.
ImSimQA is an essential software toolkit that
generates ground-truth DICOM test images to
validate clinical software systems (OnQ rts®,
MIM Maestro™, Velocity, ABAS, SPICE and
Smart Segmentation®), and provides
quantitative analysis of DIR algorithms.
Save time and close the loop using ImSimQA,
the only commercially available standalone
software designed for DIR + RIR QA.
ESTRO 33 - Booth 5150
www.osl.uk.com [email protected] +44 (0)1743 462694
To find out more about advertising opportunities on medicalphysicsweb and in the medicalphysicsweb review,
simply contact our sales representative Paul Rucci.
e-mail [email protected], tel +1 (215) 627 0880 ext. 268
medicalphysicsweb review Spring 2014
18 focus on: nuclear medicine
CD Well eases small-animal studies
Performing PET functional measurements in very small animals can be
difficult, due to the limited blood volume that can be sampled and rapid
changes in blood radioactivity after
the radiotracer is injected. To ease this
process, Japanese researchers have
developed the CD Well, which uses
microlitre-sized sample volumes to
measure whole blood and plasma
for dynamic PET studies in mice and
rats. The aim was to develop a system
that measures time-activity curves
for arterial whole blood (wTAC) and
plasma (pTAC) separately (Phys. Med.
Biol. 58 7889).
A quantitative PET study requires
concurrent acquisition of pTAC and
wTAC data to estimate the behaviour
of an administered radiopharmaceutical in target tissue using compartment model analysis. For a small
rodent, the volume per sample for
TAC analysis is limited to 4 or 5 µl.
This is constrained by the number
of samples needed for metabolite
analysis and pTAC and wTAC measurements, and the restriction that no
CD Well researchers: the team includes collaborators from NIRS, Chuba
University, Fujita Health University, Kinki University and Shimadzu.
more than 10% of the animal’s total
blood volume should be sampled
to prevent adverse physiological
effects.
The small volume of these microsamples is a limiting factor when
they are centrifuged to separate
plasma from blood cells, and then
used to perform volume and radio-
activity measurements, according to
lead author Yuichi Kimura, a professor at Kinki University. The CD Well
overcomes this difficulty because it
uses only a minute volume of blood.
T he apparatus contains 36
U-shaped channels. Each holds a
maximum of 4 µl of blood. The
hydrophilic-lined channels have
tapered inlet openings at each end,
one used to sample a drop of blood
and the other serving as an air vent.
Each channel has a precise, uniform
rectangular cross-sectional area of
200 ± 10 µl × 500 ± 10 µl to enable
a volume of sampled blood to be
obtained from the length of the sample in the channel.
“The small size of the CD Well enables it to be placed next to a rodent
during the blood sampling. This
shortens the catheter length, reducing distortion of the TAC due to dispersion and minimizing the total
volume of blood loss,” the authors
wrote. Its size also reduces the delay
between the true and measured TAC,
as well as distortion of the TAC shape
caused by dispersion due to the transit of sampled blood through the
catheter; delay and dispersion being
the major factors that cause inaccurate TAC.
After serial sampling is completed, the CD Well is centrifuged to
separate plasma from blood cells and
then scanned with a flatbed scanner
A new take on
PET registration
The advantages
of microSPECT
A new way of validating PET imaging and the segmentation of tumour
images with histopathological data
has been developed by researchers
in the US. The technique synthesizes
PET images from autoradiography
images of excised tumours and
could provide a more informed
basis for customized PET-guided
radiotherapy that selectively boosts
dose to certain tumour subvolumes
(Radiother. Oncol. doi: 10.1016/j.
radonc.2013.12.017).
“While the scientific community
is excited about PET-guided fully
personalized radiotherapy, we cannot ignore the fact that we are yet
to demonstrate spatial correlation
between even the most common of
PET tracers and any of their assumed
‘identities’ such as hypoxia, cell proliferation, cell metabolism state and
cell density,” said first author Marian
Axente, who carried out the research
at Virginia Commonwealth University Medical Centre in Richmond,
VA, and now works at Stanford University in California.
A common approach to histopathological PET validation is to
register in vivo PET scans of a tumour
with histopathological images
acquired after its excision. However,
registration techniques have limited
accuracy as the two image sets can
lack common identifying landmarks
and the excision and subsequent processing of the tumour can deform it
from its in vivo, scanned state.
Instead, the researchers’ approach
uses ex vivo autoradiography to
image slices of tissue sampled from
throughout an excised tumour perfused with radiotracer. The images
State- of-the-ar t microCT and
microSPECT scanners can produce
images with high enough spatial
and temporal resolution to quantitatively evaluate the cardiac function of laboratory mice. According
to researchers at Duke University
both of these imaging technologies
offer advantages over MRI, the “gold
standard” for in vivo murine cardiac imaging (Mol. Imaging Biol. doi:
10.1007/s11307-013-0686-z).
In a direct comparison study, the
researchers also discovered that
microSPECT had several advantages
over microCT. This finding may be
particularly valuable, because the
microSPECT scanner utilized is in
commercial production, whereas
the microCT is unique to the Duke
research laboratory.
In vivo cardiac imaging of mice is a
useful tool for researchers of human
cardiac disease and for developers working on new clinical treatments. However, the small size of a
mouse’s heart and its rapid heart rate
make the acquisition of high-quality
images challenging.
Duke’s dual-source microCT
system produces what lead author
Nicolas Befera and colleagues
believe is the highest spatial resolution currently available for 4D cardiac microCT imaging. The lab also
owns a U-SPECTII/CT system (from
MILabs of the Netherlands) fitted
with a 0.35 mm multi-pinhole collimator. The research team used both
systems to acquire images of a mouse
with myocardial infarction, as well
as six control mice, all of which were
anaesthetized for the procedures. For
the 4D microCT, mice were injected
medicalphysicsweb review Spring 2014
Increased accuracy: first author Marian Axente and colleagues have
come up with a new technique for segmentation of PET images.
are used to build a 3D map of radiotracer distribution. Then, the map
is combined with known PET scanner characteristics to synthesize
PET images that can be compared
with histopathological images
also acquired from throughout the
tumour.
For their proof-of-concept study,
they tested two human head-andneck tumour xenografts excised
from two mice that had been injected
with carbon-14 FDG. The radiotracer
is chemically identical to the fluorine-18 FDG tracer used clinically
to image glucose metabolism. The
researchers cut 8 µm thick slices of
tumour, in stacks of four every 120
µm throughout the volume. They
used autoradiography to image one
slice from each stack and carried out
histopathological testing in neighbouring slices. Mayer’s haematoxylin
and eosin (H&E) staining was used to
identify viable tissue subvolumes in
one slice and, while unused in this
study, the other two slices can be
used to visualize other characteristics such as hypoxia or proliferation.
A 3D autoradiography set that
mapped the carbon-14 FDG distribution was reconstructed based on
a series of image registrations that
included the use of needle tracks left
by slice preparation procedures as
registration landmarks. Overall registration accuracy was confirmed by
a consistency analysis that demonstrated smooth variations in voxel
intensity between slices in the 3D
volume and high normalized correlation coefficients of 0.84±0.05 and
0.94±0.02 for the two tumours.
Synthetic PET images were generated by convolving the 3D autoradiography data with the point
spread function of a small-animal
PET scanner. In an illustration of the
application of their technique, the
researchers showed that thresholdbased segmentation could identify
viable tumour. Using thresholds of
14% and 22%, the extent of viable
tumour could be detected with high
sensitivities of 0.99 and 0.94 in the
two tumours, respectively. Specificity, however, was low.
Potential technique improvements
include the use of Monte Carlo techniques to increase PET simulation
accuracy and automation to increase
efficiency. “Automation of the process would bring about an exquisite
amount of preclinical and clinical
data,” Axente told medicalphysicsweb.
However, due to a lack of funding,
there are no immediate plans for further testing and development.
to define the regions of plasma and
blood cells. The measured length is
converted to volume. The apparatus
is then exposed to an imaging plate
for 20 minutes to measure positrons
emitted from the blood in the channels, after which it is scanned by a
phosphor imaging plate scanner
with a spatial resolutions of 100 µm/
pixel and a 16-bit output resolution.
The two images are processed using
a dedicated software program. The
scanned image, which depicts locations of whole blood and plasma, is
superimposed on the imaging-plate
derived image to acquire the radioactivity concentrations (in Bq/µl) in
whole blood and plasma separately.
Kimura and colleagues reported
that in in vitro validation testing, the
radioactivity concentrations measured by the CD Well were not significantly different from those obtained
by conventional manual methods.
And because the CD Well measures
blood volume directly, individual
differences in blood specific gravity
do not affect the result.
with a liposomal blood pool contrast
agent, immediately after which sampling was performed using retrospective cardiorespiratory gating. Image
acquisition of 2250 projections per
mouse took 5–10 minutes, with a
delivered radiation dose of 360 mGy.
For microSPECT, mice were injected
with technetium tetrofosmin, and
imaged for two hours. They received
a radiation dose ranging from 400 to
800 mGy.
Measures of global left ventricle
functions between the two modalities were comparable. Regional
functional indices – specifically wall
motion, wall thickening and regional
ejection fraction – were also similar. One of the relative advantages
of microSPECT over microCT was
evident when both modalities were
used to image a mouse with a surgically induced myocardial infarction. “MicroSPECT readily identified
a large perfusion defect consistent
with myocardial infarction involving the cardiac apex and anterolateral
left ventricle wall, which could not be
seen by microCT,” the authors wrote.
MicroSPECT also offered an advantage in imaging myocardial infarction
in that both function and radiotracer
distribution data could be obtained
from a single acquisition without
the need for additional for additional
specialized contrast agents. Another
advantage is its nanomolar sensitivity in detection of molecular probes,
with a 103 greater sensitivity than
currently available MR techniques.
The primary disadvantage compared
to MRI was the use of ionizing radiation. “While this level of radiation
exposure would have no effect on
cardiac imaging, it could potentially
become an issue in cancer studies
by affecting the progression of some
tumours,” they observed.
Sign up as a member at medicalphysicsweb.
h eALth PhySicS
nUcLe A r med i c i n e
d i A g n o S ti c rA d i o L o g y
R a D I at Io n t h E R a p y
Nearly as good as water.
true
Innovation
First commercially available
synthetic single crystal diamond
detector for clinical dosimetry
}
Nearly water equivalent for all beam energies
}
Extremely small sensitive volume (0.004 mm³),
ideal for small field dosimetry
}
One single detector for all field sizes up to 40 cm x 40 cm
}
Precise, accurate measurements in photon and electron fields
}
Minimal energy, temperature and directional dependence
R A D I AT I O N T H E R A P Y
When small things matter.
Small Field Dosimetry
Application Guide
More information on small field dosimetry?
Contact your local PTW representative for a free
copy of our application guide Small Field Dosimetry
or download it from our website.
www.ptw.DE
USA | LAtin AmericA | chinA | ASiA PAcific | indiA | UK | frAnce | iberiA | germAny
Knowing what
responsibility means
Deliver Lung SBRT
in 43% Less Time
Sophisticated respiratory motion management
without compromised treatment times
4513 371 1229 11:13
The new standard in the Power of Care. Visit us at ESTRO booth #6200.
One Solution. Unlimited Possibilities.
www.VersaHD.com
As compared to previous generation Elekta digital linear accelerators. Stieler F, Steil V, Wenz F, Lohr F, Department
of Radiation Oncology, University Medical Center Mannheim, University of Heidelberg, Germany.
Versa HD is not available for sale or distribution in all markets. Please contact your Elekta representative for details.