Target audience

Clinicians and medical researchers who are interested in advanced applications of next generation PET/MR system

Purpose

The integrated PET/MR combines the advantages of functional imaging device positron emission tomography and high resolution high contrast magnetic resonance imaging, simultaneously acquiring PET and MR images at the same position, improving image fusion accuracy. However, respiratory motion during abdominal imaging causes notorious motion artifact in the MRI images and blurring the PET images. To improve diagnosis accuracy, motion blurring needs to be eliminated or reduced in PET/MR scan. In this work, a PET/MR motion correction method based on real-time 2D excitation navigator has been implemented and assessed on phantom and clinical data.

Methods

A motion correction strategy contains two steps: (1) extracting precise motion vector during MRI acquisition; (2) correcting PET list mode data according to the motion vector. All experiments were performed on a simultaneous TOF PET/MR scanner (uPMR790, United Imaging Healthcare, Shanghai, China). A pencil beam navigator excited by a 2D RF pulse was positioned along the motion direction and the projection profile was obtained by the Fourier transform of the measured signal. By utilizing an edge detection algorithm, the respiratory motion vector can be extracted. Then the estimated motion is deployed to prospectively gating the MR data acquisition and retrospectively re-binning PET data. To verify this approach, three point-sources was positioned in a shaddock. Bed drives shaddock moving back and forth to imitate a respiratory motion. In the human scans, navigator was placed across the right hemi-diaphragm along the superior-inferior direction and respiratory belt was placed around the upper abdomen to acquire respiratory signals for contrast. Simultaneous PET data acquiring and navigator interleaved with FSE sequence are implemented in uPMR 790 PET/MR system (United Imaging Healthcare, Shanghai, China).

Results

Navigator accurately tracked the abdominal motion and the motion vector extracted by edge detection algorithm well delineated the motion (Figure2). Motion blurring in the PET image was significantly reduced after correction (Figure 3). Signal to noise ratio (SNR), ghost to image ratio (GIR) of MR images and FWHM values of lesions in PET images has been evaluated quantitatively in phantom. SNR increased from 163.89 to 662.23 whereas GIR reduced from 24.87% to 2.94%, suggesting a significant improvement compared to the one without motion correction. In human scan, there is no significant difference (P>0.8) of the image quality that obtained with navigator (SNR = 89.35, GIR= 7.93%) and respiratory belt (SNR = 91.05, GIR = 8.15%).

Discussion

Respiratory motion is the main cause of artifacts in abdominal imaging. A motion correction strategy based on navigator for simultaneous PET/MR has been developed and evaluated. Phantom and human imaging result implies that this technique can precisely acquire object motion and effectively eliminate motion blurring.

Conclusions

We have successfully implemented and evaluated a respiratory gating sequence for abdominal imaging in a simultaneous PET/MR. Based on the qualitative and quantitative analysis of phantom and human data, we can conclude that navigator provides a respiratory monitor which is suitable for PET/MR motion correction in abdomen imaging. Reducing motion artifacts and improving diagnosis accuracy, without additional operation and device, it offers a simple and cost-down way for clinical use.

Acknowledgements

No acknowledgement found.

References

No reference found.