MRI-linac: a progress update
Image guidance is becoming an integral part of radiation therapy, offering the potential to incorporate patient-based information into the treatment process. Such data enable patient repositioning or even replanning immediately prior to irradiation. Ultimately, image-guided radiotherapy (IGRT) could offer real-time imaging during treatment, enabling treatment of moving targets with motion compensation during beam delivery.
MRI could provide the ultimate means to perform such real-time guidance – offering soft-tissue-based, online position verification and monitoring during radiation treatment. A team at the University Medical Center Utrecht in the Netherlands is developing just such an IGRT system by integrating an MRI scanner with a linear accelerator. The system comprises a modified 1.5 T Achieva MRI scanner (Philips Healthcare, the Netherlands) with a 6 MV accelerator (Elekta, UK) rotating around it on a ring gantry.
The Utrecht researchers have already demonstrated simultaneous operation of the MR scanner and linac, using a prototype system with a static accelerator (see: MR-guided radiotherapy: proof of concept). With installation of a gantry-based prototype now imminent, Tami Freeman spoke to project leaders Jan Lagendijk, professor of radiation oncology physics, and Bas Raaymakers, associate professor of radiation oncology, to find out how the project is progressing.
TF: What is the main driver for implementing image guidance during radiotherapy?
JL/BR: Enormous progress is being made in the dose painting capabilities of radiotherapy systems. As a consequence, the question of "where to paint" is becoming more and more important. Shaping the dose according to our radiobiology knowledge is frustrated by positioning uncertainties. Also, as a result of these uncertainties, certain tumours can't be treated well: think of kidneys, liver, pancreas, oesophagus, rectum, etc.
What are the benefits of using MRI for such guidance?
MRI is excellent for soft tissue visualization and can provide cine imaging at a rate sufficient to track all motion – even breathing-related movements – in real time. By integrating high-quality MRI with modern accelerator technology, tissue can be tracked online and beams can be guided to their (moving and deforming) targets with (sub)millimetre precision.
This enables boosting of the dose to the gross tumour volume (GTV) and providing intermediate doses to the clinical target volume, avoiding visible non-involved structures. Almost all radiotherapy applications will benefit from reduced NTCP [normal tissue complication probability] and higher GTV doses. Such MRI-guided treatments can also provide a breakthrough for difficult locations, such as the abovementioned pancreas, kidneys, liver, oesophagus and rectum.
What are the key challenges in achieving simultaneous MRI and irradiation?
The interactions between the two systems were the main challenge. We needed to bring the systems together while maintaining their individual performance levels.
How did the Utrecht team achieve this?
Those interactions are dealt with using active magnetic shielding and smart RF design. We also stayed with a diagnostic-quality closed-bore system, which implied that we had to modify the gradient coils and cryostat to allow passage of the beam. At present, a major challenge is the online Monte Carlo-based treatment planning. Secondary electrons interact with the magnetic field, and this interaction, including the electron return effect (ERE), has to be taken into account in the treatment-planning software. A second challenge is fast online image registration and tracking. A lot of progress is needed in that field.
The US company ViewRay makes a low-field system with three60Co sources for irradiation. This system may be the first commercial one. The company has really helped to open thinking on MRI-guided radiotherapy. Their advertisement campaign, including the brilliant "seeing helps" cards, really helps in getting the rest of the industry going.
Could the MR images recorded during treatment be used for real-time repositioning?
Remember that the space inside an MRI is limited, which implies that table movements are also restricted. We have chosen a solution that we call "virtual couch shift", which means that every position adjustment is done within the treatment-planning environment. In practice, this implies that the whole system is prepared for online treatment planning. This enables us to account for not only translations, but also rotations and deformations.
Could the MRI-linac system enable new types of radiation treatments?
We are sure of that. At this moment, we are building in Utrecht the Centre for Image Guided Oncological Interventions. This new centre will have three treatment rooms dedicated entirely to new treatments. We are opening research lines on kidney, liver, pancreas, oesophageal and rectal cancer. Within this new centre, we will also investigate competitive technologies such as MRI-guided brachytherapy and MRI-guided high-intensity focused ultrasound.
What stage is the development of the Utrecht MRI-linac at?
The prototype gantry will be installed this summer, and the full system will be tested this autumn. If everything works as expected, then we will start working on the next steps to complete the system. We have begun preliminary testing, which involves a lot of MR imaging and treatment planning studies on all types of tumours. This helps us to understand the capabilities and limitations of the system and prepares us for the real thing.
What do you see as the first clinical application?
Our first clinical application will be very simple: the palliative treatment of spinal bone metastases. This will allow us to test the geometrical accuracy of the system on real patients. The vertebra is easily visualized on the integrated megavoltage imaging system and allows for accurate comparison of field co-ordinates with MRI co-ordinates.
When do you predict that first patient treatments using the MRI-linac may begin?
This is a hard question. The prototype that we will install this summer is still non-clinical; patient safety is not guaranteed, the system has no covers, etc. The first clinical system will hopefully be somewhere around the end of 2012, but this will depend on many other factors. We may decide to make the present prototype clinical if the clinical system takes too long.