Introduction
Myocardial blood flow (MBF) has proven additive value in rubidium-82 cardiac PET. However, non-periodic patient motion, as a result of the side-effects of pharmacological stressing, can have substantial impact on MBF data. This work evaluated frame-by-frame motion correction both with and without corresponding adjustment of the PET attenuation map.
Method
Frame-by-frame translational motion correction was performed by three experienced users on 30 rubidium-82 studies. Data were divided equally into 3 groups of motion severity [mild (M1), moderate (M2) and severe (M3)]. MBF data were generated in syngo.PET MBF (Siemens Healthineers) for non-motion-corrected (NC), motion-corrected only (MC) and motion-corrected with corresponding adjustment of the attenuation map (MCAC). MCAC were generated with prototype reconstruction software (e7 tools, Siemens Healthineers). Percentage differences of MBF were calculated in the three coronary territories and also in 17-segment polar plots. 17-segment polar plots of spill-over values, obtained from Syngo.PET MBF, were also generated for NC, MC and MCAC data.
Results
The table shows the median and inter-quartile range of absolute percentage change of MBF for the three motion categories for MC compared with NC data. In the 17-segment polar maps, median MBF differences of up to 50% were observed in the mid-anterior segment and up to 30% in the mid-inferior segment for MC vs NC. Median segmental differences for MCAC compared with MC were considerably smaller with a maximum of 15%. Segmental plots of median spill-over values were notably more uniform for MC and MCAC data compared with NC data.
Conclusion
Motion correction results in substantial adjustments to MBF data, highlighting the need for readily available robust motion correction techniques. Frame-by-frame adjustment of the attenuation map results in marginal differences and therefore is unlikely to be an essential requirement. This is fortunate as it involves close integration into the PET reconstruction and is not currently available commercially. Assessing the uniformity of spill-over plots post-correction is a useful visual aid for verifying both manual and automated motion correction techniques.
Myocardial blood flow (MBF) has proven additive value in rubidium-82 cardiac PET. However, non-periodic patient motion, as a result of the side-effects of pharmacological stressing, can have substantial impact on MBF data. This work evaluated frame-by-frame motion correction both with and without corresponding adjustment of the PET attenuation map.
Method
Frame-by-frame translational motion correction was performed by three experienced users on 30 rubidium-82 studies. Data were divided equally into 3 groups of motion severity [mild (M1), moderate (M2) and severe (M3)]. MBF data were generated in syngo.PET MBF (Siemens Healthineers) for non-motion-corrected (NC), motion-corrected only (MC) and motion-corrected with corresponding adjustment of the attenuation map (MCAC). MCAC were generated with prototype reconstruction software (e7 tools, Siemens Healthineers). Percentage differences of MBF were calculated in the three coronary territories and also in 17-segment polar plots. 17-segment polar plots of spill-over values, obtained from Syngo.PET MBF, were also generated for NC, MC and MCAC data.
Results
The table shows the median and inter-quartile range of absolute percentage change of MBF for the three motion categories for MC compared with NC data. In the 17-segment polar maps, median MBF differences of up to 50% were observed in the mid-anterior segment and up to 30% in the mid-inferior segment for MC vs NC. Median segmental differences for MCAC compared with MC were considerably smaller with a maximum of 15%. Segmental plots of median spill-over values were notably more uniform for MC and MCAC data compared with NC data.
Conclusion
Motion correction results in substantial adjustments to MBF data, highlighting the need for readily available robust motion correction techniques. Frame-by-frame adjustment of the attenuation map results in marginal differences and therefore is unlikely to be an essential requirement. This is fortunate as it involves close integration into the PET reconstruction and is not currently available commercially. Assessing the uniformity of spill-over plots post-correction is a useful visual aid for verifying both manual and automated motion correction techniques.
