MENU

Click here to

×

Are you sure ?

Yes, do it No, cancel

Impact of Low Contrast Objects On a PET Continuous Bed Motion Data Driven Gating Algorithm

J Meier*, O Mawlawi, MD Anderson Cancer Ctr., Houston, TX

Presentations

(Sunday, 7/12/2020)   [Eastern Time (GMT-4)]

Room: AAPM ePoster Library

Purpose:

Data driven gating(DDG) in PET imaging determines the patient respiratory waveform from periodic changes in PET data, overcoming the need for external hardware. However, the ability to extract a DDG waveform from small low contrast moving objects could be limited. In this work we evaluate the performance of a recently developed DDG continuous bed motion(CBM_DDG) algorithm under such conditions using a phantom study.

Methods:

A phantom with 5 spheres (inner diameter 10-28 mm) was scanned with motion amplitudes of 0, 1, 2 and 3 cm with two sphere to background ratio(SBR) preparations: 5:1_SBR and a low contrast 2:1_SBR on a Siemens mCT PET/CT with 1 mm/s continuous table speed. For each sphere amplitude and contrast the CBM_DDG waveforms were compared to the gold standard Anzai hardware belt(ANZ_EMDB) waveforms using Pearson correlation coefficient(PCC). The ratios of sphere SUVmax(RSmax) and SUVpeak(RSpeak) for motion corrected images using DDG and Anzai waveforms were averaged across all spheres for each acquisition amplitude respectively.

Results:

At 2:1_SBR, CBM_DDG waveforms had low correlations, while at 5:1_SBR, correlations improved with increasing amplitude. Absolute values of PCCs for 0, 1, 2, and 3 cm acquisitions were 0.27, 0.10, 0.03, and 0.13 for 2:1_SBR, and 0.23, 0.53, 0.75, and 0.89 for 5:1_SBR. RSmax and RSpeak for 2:1_SBR were lower than 5:1_SBR values as amplitudes increased. RSmax values for 0, 1, 2, and 3 cm acquisitions were 1.01, 1.01, 0.88, and 0.83 for 2:1_SBR, and 0.96, 1.04, 0.92, and 0.94 for 5:1 SBR. RSpeak values for 0, 1, 2, and 3 cm acquisitions were 1.00, 1.00, 0.93, and 0.86 for 2:1_SBR, and 0.99, 1.02, 0.93, and 0.93 for 5:1 SBR.

Conclusion:

This study showed that performance of the CBM_DDG algorithm degrades with low contrast objects; however, the CBM_DDG algorithm is optimized for patient data necessitating future verification in patient studies.

Funding Support, Disclosures, and Conflict of Interest: This research was supported in part by a grant from Siemens Healthineers.

Keywords

PET, Motion Artifacts, Respiration

Taxonomy

IM- PET : Motion management

Contact Email