NDI Polaris Provides Patient Motion Compensation In PET Procedures
The Problem
The movement of the head during neurological Positron Emission Tomograph (PET) procedures causes reduced resolution in the reconstructed image. Short of rigidly securing the
subject's 
head in a
fixation frame, it is not possible to eliminate all head movement during the procedure, which may take two hours. It is the slow and gradual movement of the subject's head that
occurs continuously during the procedure that causes the reduced resolution. Fast sudden movements, for example, a cough or sneeze, only last a few seconds and have little or no
effect on the final resolution (as long as the subject's head returns to its original position, or if, after coughing or sneezing, the head remains in the the new position).
If the end result is data that cannot be used, the PET procedure is an expensive failure. The problem requires a solution that is:
- Cost effective
- Simple to implement
- Adds no additional time to the procedure
- Causes minimal discomfort to the subject
The Solution
Peter M. Bloomfield has developed a motion tracking system that incorporates the NDI Polaris system (Figure 1). The motion tracking system compensates for the slow, gradual
movement of the subject's head, maintaining the resolution
in the reconstructed image.
The Polaris system accurately tracked the location of four passive reflective markers. The markers were attached securely to the subject's head, using a neoprene cap and a Plexiglas mounting plate (Figure 2). The neoprene cap caused minimum discomfort or stress to the subject.
The Polaris system enabled monitoring of the location of the subject's head in space. No electrical connection was required to the subject and no magnetic fields were present, that would cause interference.
The Polaris system data was synchronized with the PET data to within 1mS. The synchronization was achieved by inserting pseudo random numbers into the Polaris data and PET data
stream. These random numbers were matched, post acquisition, providing synchronization of Polaris and PET data.
The results achieved with the Polaris system are illustrated in Figures 3 and 4. Figure 3 illustrates reconstructed transverse images of a multi-line source phantom. Images (a),
(b) and (c) are not corrected for motion and clearly
show blurring.
Image (a) is a stationary phantom, while images (b) and (c) are images taken during discrete and gradual ±5° and ±15° rotation respectively. Images (d), (e)
and (f) show the same images after motion compensation data has been supplied with the Polaris system.
Figure 4 illustrates similar correction applied to a Hoffman brain scan. Image (a) is static, image (b) is uncorrected and image (c) shows the scan with motion correction applied.
Acknowledgement
NDI would like to acknowledge the assistance given by Peter M. Bloomfield in the production of this case study, based on his work at Imaging Research Solutions Ltd, Hammersmith Hosptial, London, UK. Peter is currently working at the Centre for Addiction and Mental Health, Toronto Canada.
