Evaluation of electromagnetic motion tracking and event-by-event motion correction in clinical brain PET imaging

Introduction: A high resolution electromagnetic motion tracking (EMMT) system with event-by-event motion correction has been developed for PET-CT scanners to monitor head motion in real time. Our prior studies of the EMMT system focused on phantom experiments, demonstrating promising results. Nevertheless, the more crucial clinical application of the EMMT system in brain PET imaging has not been explored. In this work, we conducted a pilot clinical study with human participants to evaluate the performance of the EMMT system when integrated with a commercial Siemens Biograph mCT PET-CT scanner.

Methods: Figure 1 (a) illustrates the integration of the EMMT with a brain PET scanner. The EMMT captures the head motion in real time by using a wearable electromagnetic (EM) sensor array inside a 3D EM dipole field generated by an EM source. The sensor array is affixed to human and mannequin heads using a head strap, depicted in Figure 1 (b). The EM source is custom-designed to accommodate three Na-22 point sources. Coordinate calibration between the EMMT and PET scanner is achieved by locating the positions of three Na-22 point sources in both coordinates, as shown in Figure 1 (c). Time synchronization between EMMT motion and PET list-mode data is achieved by starting their data acquisition simultaneously. This synchronization is validated by aligning the characteristic motions (e.g., abrupt changes in motion signal) existing in both the EMMT and a PET data-driven motion (centroid of distribution (COD)). According to the motion captured by the EMMT, event-by-event motion correction is implemented by repositioning each line-of-response (LOR) of the PET data, resulting in a motion-corrected data set. The image of the motion-corrected data set is reconstructed using the Customized and Advanced Software for Tomographic Reconstruction (CASToR) with an TOF-OSEM algorithm (4 iterations, 21 subsets, and 1×1×2 mm³ voxels).

Two participants were rerolled in an institutional review board-approved study at Weill Cornell Medicine, for evaluation of the proposed EMMT and event-by-event motion correction in clinical brain PET imaging. The participants, administered approximately 5 mCi [¹⁸F]MK6240 tracer, underwent a 60-minute dynamic for a separate Alzheimer’s disease study. Following the dynamic scan, it took less than 5 minutes to set up our EMMT device on the participants’ heads. Subsequently, the participants underwent 15-minute scan (65-80 minutes post-injection (pi)) with head motion. During this 15-minute scan, participants were instructed to move their heads to random positions and orientations every 5 minutes. Finally, a 2-minute mannequin study, using a set of point sources attached to a mannequin head, was conducted after the human study to access the accuracy of the proposed motion tracking and correction method.

Results: Figure 2 shows the reconstructed [¹⁸F]MK6240 PET images of two human participants with and without head motion correction. The images highlighted two structures, namely the eyeball and ventricle, with a noticeable reduction in motion-induced blurring, as indicated by white arrows. Figure 3 (a) displays the measured motion along all 6 degrees-of-freedom (3 translations and 3 rotations) of the mannequin in the scanner coordinates as well as the time synchronization of one-dimensional EMMT and COD motion. Figure 3 (b) shows the reconstructed static and EMMT motion-compensated images of three point sources on the mannequin. Furthermore, the full width at half maximum (FWHM) and center shift of these three point sources are shown in Table 1.

Conclusions: We evaluated an electromagnetic motion tracking system using two participants in brain PET imaging and radioactive point sources on a mannequin head. Initial findings demonstrated the successful integration of the EMMT technology with the mCT scanner. Ongoing work includes performing region of interest analysis of clinical data and conducting additional studies to obtain statistically significant results.

• Download figure
• Open in new tab
• Download powerpoint