Head motion of anesthetized patients during high-resolution PET brain imaging

Objectives Patient motion during brain scans can compromise the high spatial resolution capabilities of PET scanners like the Siemens ECAT HRRT. One may assume that there is no observable head motion for patients under anesthesia, or that head restraints are the best solution to prevent motion. However, for high-resolution imaging, even small movements can degrade the imaging of small brain structures. Head restraints may also hinder maintenance or emergency access to patients during scanning, access that may introduce head movements. Here we retrospectively investigated the measured head motion of anesthetized patients without head restraints during HRRT brain scans.

Methods The HRRT is optimized for human brain imaging and has a spatial resolution of 2.4 mm. A Polaris infrared retro-reflector system (Northern Digital, Inc.) provides real-time position and orientation of the patient’s head to within 0.35 mm. The reference position is defined as the average location of the patient during transmission scanning and tracking is continued through the emission scan. The position information is transformed into scanner coordinates and motion is corrected event-by-event during list-mode reconstruction. In this study we evaluated the motion of eight patients having 90-minute emission scans for trends and displacement magnitude.

Results All patients demonstrated gradual drifts in position over the duration of the 90-minute scan period. On average, translations of 4.8 millimeters were observed due to this gradual motion. A maximum translation of six millimeters was observed for drift motion. Rotations of 1-2 degrees were common. Additional rapid translations of several millimeters were typical when anesthesiologist intervention occurred. The greatest observed sudden translation was 12 millimeters.

Conclusions Motion correction may be desired for high resolution PET scans of anesthetized patients. This is particularly true for scans of one hour or more, or where anesthesiologist intervention is likely.

Research Support This research was supported by the Intramural Research Program of the NIH, CC