Determination of phase differences in hardware- and data-derived respiratory signals in whole body F-18-FDG PET/CT scans

Ziel/Aim PET raw data-derived respiratory signals have been demonstrated to result in motion-corrected whole-body PET images comparable to those based on conventional hardware-derived signals [1]. We here analyze hardware-based signals and data-driven gating (DDG) signals in a clinical cohort with respect to phase differences, as these may affect motion resolution and thus motion correction.

Methodik/Methods 103 whole-body F-18-FDG PET/CT datasets of patients with suspected malignancies in thorax or abdomen acquired in continuous bed motion (1.1 mm/s) with respiratory signals based on the Anzai belt (belt gating, BG) and a dedicated DDG algorithm [1] were investigated. The interval between scanning the bladder and the lung apex was divided into ten regions of equal length, and regional phase differences between BG and DDG signals were determined using cross correlation analysis. Repeated continuous fast MR scans in 4 healthy volunteers were additionally performed to corroborate the findings.

Ergebnisse/Results Overall, a temporal difference of 52 ± 110 ms (mean±SD) was found between DDG and BG signals, reflecting either physiological differences between internal and external motion or synchronization uncertainties between BG and DDG. Moreover, the region around the kidney demonstrated the highest differences (81 ± 103 ms), while DDG signals from regions close to the bladder and lung apex were more in phase with BG (-22 ± 374 ms, 47 ± 390 ms, respectively). These results were in line with the MR data, with temporal shifts between the kidney region and lower abdominal/upper thoracic regions of 52 ± 37 ms and 67 ± 58 ms, respectively.

Schlussfolgerungen/Conclusions The results demonstrate small phase differences between hardware-based and data-driven respiratory signals. DDG signals from different body regions also typically exhibit phase differences (respiratory hysteresis). However, the magnitude of these effects seem too small to influence PET motion correction approaches in most cases.

• Literatur/References

• 1 Büther F, Jones J, Seifert R. , et al. J Nucl Med 2020; 61: 1520-1527.

Publication History

Article published online:
08 April 2021