Aug 15, 2012
MRI tracks motion during radiotherapy
Real-time tumour tracking could prove invaluable for radiotherapy of moving targets such as lung tumours, enabling minimization of treatment margins. The use of MRI for such tracking provides soft-tissue-based position verification, without exposing the patient to additional ionizing radiation. At the recent AAPM annual meeting in Charlotte, NC, Gino Fallone shared the latest results from the hybrid linac-MRI being developed at the Cross Cancer Institute in Edmonton, Canada.
The hybrid system comprises a 6 MV linac mounted on the open end of a biplanar, low-field MRI magnet, with both the linac and magnet sited on a gantry that rotates around the patient. In this "rotating-biplanar" geometry, the magnetic field vector is fixed with respect to the beam direction. The team at the Cross Cancer Institute has already built and tested a fully functioning head system, and is now constructing and installing a whole-body version using a 0.6 T superconducting MRI.
Fallone, director of the Cross Cancer Institute's department of medical physics, described the various steps required to perform real-time tumour tracking. The MR system images the tumour during irradiation; the tumour is automatically contoured and its position is determined (and future positions predicted); and then the multileaf collimator (MLC) leaves are adjusted to account for any motion during the treatment. This process repeats throughout radiation delivery.
Prior to treatment, the physician defines a region-of-interest on an MR image of the tumour. During irradiation, an auto-contouring algorithm combines this region-of-interest with information on the range of motion (determined pre-treatment over a few breathing cycles) to establish the tumour position.
Fallone and colleagues evaluated their auto-contouring algorithm using a lung tumour motion phantom. They scanned the moving phantom using 3T MRI with images obtained at 4 frame/s (as recommended for lung tracking). Two tumour shapes and four motion patterns were studied, with 600 images recorded for each. The researchers then processed the 3T images to approximate those recorded at 0.2T and 0.5T.
The algorithm successfully contoured the moving tumours, with Dice coefficients (a measure of similarity) of above 0.96 and above 0.93 for the 0.5T and 0.2T images, respectively. The auto-contoured centroid positions agreed with the actual positions (provided by an optical encoder on the phantom) to within root-mean-squared-errors of 0.92 mm at 0.2T, and 0.55 mm at 0.5T.
The researchers also performed an in vivo study, in which a lung cancer patient was scanned with 3T MRI (at 4 frame/s) to create 650 MR images. As before, the images were then degraded to simulate low-field images and the auto-contoured shape and its centroid were compared with reference images. In this study, the tumour was contoured with a Dice coefficient of 0.84 for the 0.2T image, and 0.86 for the 0.5T image.
Fallone pointed out that the auto-contouring process takes less than 5 ms per image, and noted that conventional dynamic lung imaging sequences provide sufficient tumour-tissue contrast-to-noise ratio and temporal resolution for real-time tumour tracking at 0.2 and 0.5T. "Even though these images are more noisy at 0.2T, the software can still pick the image out," he said.
The process of MR signal acquisition, image reconstruction and processing, and MLC motion takes approximately 220–280 ms. Thus a means of motion prediction is required to compensate for tumour motion during this system delay. To do this, Fallone's team is applying artificial neural networks (ANN), using seven (separately trained) ANNs to predict tumour positions at 40 ms intervals.
To assess the entire tumour tracking process, the researchers performed an evaluation experiment on the prototype linac-MRI system using a phantom undergoing sinusoidal motion. The phantom was imaged using 0.2T MRI (at 4 frame/s), the auto-contouring algorithm determined the centroid position, and the resulting data were then used to control the motor on a 10-leaf MLC. The radiation beam was incident upon the phantom for 2 min, with radiographic film placed on the phantom revealing the pattern of irradiation.
When no tracking was applied, the image recorded on the film was, as expected, highly blurred. Tracking without motion prediction improved the image, but some blurring occurred due to the dead time. Adding motion prediction to the tracking resulted in a sharp image with well defined edges.
The researchers repeated the experiment using a phantom that moved with a modified cosine pattern (which roughly approximates breathing). The 20–80% profile of the dose penumbra – a measure that indicates the sharpness of dose deposition – was 6.6–mm for the stationary phantom and 34.1–mm for a moving phantom with no tracking. Tracking without motion prediction led to a 20–80% penumbra of 15.8 mm, while tracking with motion prediction improved this further to 8.6 mm.
"The system really requires the seven-ANN motion-prediction algorithm to ensure that we know exactly where the target is," concluded Fallone.
About the author
Tami Freeman is editor of medicalphysicsweb.