The UMBRELLA Study - MRI protocol development for MRI for MR-guided radiotherapy on the MR-Linac

The MR-Linac (MRL) is a hybrid machine which enables true image guided
radiotherapy. Image guidance of radiation treatment minimizes the uncertainties
about the location, motion and shape of the tumor and surrounding organs at
risk. Therefore, margins to ensure tumor coverage can be reduced. This gives
the opportunity for dose escalation to the tumor and/or a decrease in the
number of fractions, which can result in improved tumor control, reduction of
the organs at risk dose, decreased radiation-induced toxicity and better
quality of life.
MR imaging in combination with radiation delivery requires technical
adjustments to both the MRI and the linac hardware to prevent interference of
the magnet on the linac and vice versa.
The hardware of the MRI scanner installed in de MR-Linac is different from a
stand-alone (diagnostic) MRI itself. Therefore, scans have to be optimized and
imaging strategies have to be validated for the MRL. Technically, this means
that we have to deal with reduced gradient strength, which demands different
optimization of the acquisition protocols when compared to diagnostic scanners.
In addition, MRI guided radiotherapy on the MRLinac has additional requirements
on the images acquired for treatment planning on the diagnostic MRI compared to
the scans already used within the radiotherapy department, such as a large FOV
fully 3D acquisition large volume with a high geometric fidelity.
For the development of optimal MRI guided radiotherapy strategies, images have
to be acquired both on the MR-Linac and diagnostic MRI. During a treatment on
the MR-Linac MR images will be required in different stages of the treatment:
the pre-beam, beam-on and post-beam stage. In the pre-beam phase, target
definition, i.e. contouring (segmentation) of the tumor and organs at risk,
dose calculation and motion characterization is required. In the beam-on phase
motion target tracking and dose accumulation is performed, while in the
post-beam phase evaluation of the radiation treatment can be tested.
Automatic segmentation of tumor and organs at risk needs images with a proper
contrast and signal to noise ratio. In motion tracking, the challenge is to
image 3D volumes at a rate that is sufficient to track the tumor or gate the
dose delivery. Additionally, the images need a high geometric accuracy.
At this stage, several target regions have been defined for treatment
development on the MR-Linac [6-8]. The tumor sites consist among others of:
oligo lymph node metastases, rectum, prostate, esophagus, bladder, breast,
pancreas, cervix, head and neck, lung, liver metastases and primary liver
tumors, adrenal gland oligo metastases, kidney, endometrium carcinoma, anal
canal tumors. For all these
regions it is essential to optimize the dedicated scanning and image processing
protocols for the MR-Linac.

Goal of the UMBRELLA Study is the optimization of existing MRI protocols and
processing techniques required for MRI guided radiation treatment on the Marlin
system (MRI of the MR-Linac) and diagnostic MRI scanners in the radiology
department needed for treatment preparation.
The initial scientific focus of the MRI protocol development project is to
optimize MRI techniques and processing techniques that will enable MR guided
radiation treatment. The current project will encompass the optimization of
several techniques such as: methods for improved target definition and
treatment planning, and evaluation for MR guided radiotherapy. Since the
underlying technical problems are comparable but not equal for all tumor sites,
the development of the techniques will be performed parallel to each other.
From the technical viewpoint, the goals of this study are to acquire
geometrically accurate MR imaging with a good contrast to noise ratio to be
able to contour targets and organs at risk accurately within a limited amount
of time.

This is an observational study in which available clinical MRI acquisition
software and protocols for radiotherapy are tested, in which MRI acquisition
software and protocols are optimized or new protocols are generated for MRI*s
at the Department of Radiology to prepare for new radiotherapy indications and
in which processing techniques required for MRI guided radiation treatment are
evaluated.

Adjustment of parameters of the acquisition software on the Marlin system will
be rather limited. The available acquisition software can be tested and the
preferred contrast can be chosen for visualization of tumor and organs at risk
at the different tumor regions.
Planning MRI protocols for the visualization of tumor tissue and organs at risk
on the MR scanners of the Department of Radiology have to be optimized or even
generated. Differences in SNR, resolution, geometric accuracy and image
artefacts will be explored and reviewed in order to assess whether the
protocols are adequate for application for treatment planning on an MR-Linac.
Assessed MRI data will also be used to develop and test processing techniques
which are needed to improve the use of the MR-Linac as a safe treatment option
for future cancer patients. This includes contour and dose plan generation i.e.
by transfer of contours from MR images acquired in the pre-treatment simulation
phase to MR images acquired on the Marlin.

The number of subjects per study required to finalize a particular protocol may
vary depending on the technical challenges that have to be solved for that
particular acquisition software or protocol. Patients will only be included if
tumor visualization is required. Both patients and volunteers can be asked to
undergo an MRI on the Marlin system, on the diagnostic MR scanner at the
Department of Radiology for treatment simulation, or both.
Image quality and/or the performance of quantitative imaging methods is
assessed by clinicians (radiation oncologists, radiologists and others) and MR
physicists. Special attention will be paid to geometric accuracy. If necessary,
an equivalent MR image will be acquired on the normal diagnostic MR scanner at
the Department of Medical Imaging for image quality assessment. This in order
to detect the origin of artifacts or limited signal-to-noise ratio and to
further optimize the protocols.
Most artifacts in MRI can be resolved by means of tuning of acquisition
parameters. Optimization of acquisition parameters is performed in-vivo by
systematically changing the MR parameters involved. Such optimization
procedures should be performed in several subjects to obtain robust settings
that are subject independent and preserve geometric accuracy.

Although not the aim of the study, medical imaging of this study may reveal
unexpected or incidental findings. Beforehand, subjects are informed that this
may happen and informed about the procedure following on these incidental
findings. The aim of this procedure is to inform subjects if they have
potentially clinically relevant findings so that further investigations can be
initiated. A statement of this procedure is part of the ICF. If the potential
subject does not agree with this procedure, he/she cannot participate in this
study. A Standard Operating Procedure is in place (see SOP Nevenbevindingen
medische beeldvorming

For some participating patients a contrast agent will be injected. This will
only be done after a medical doctor has estimated the risk and burden as low
for the patient based on the available kidneyfunction and information about
allergies. Contrast agents will not be used for healthy volunteers.