Clinical evaluation of a data-driven gating algorithm for whole-body PET/CT scans in continuous bed mode

Ziel/Aim Methods for correcting respiratory motion of patients during PET scans are widely available for routine PET scans. Data-driven gating (DDG) algorithms are of great interest, since they do not rely on additional equipment for measuring respiratory information. Instead, gating signals are determined from the measured PET data themselves. The clinical performance of a novel DDG algorithm specifically designed to handle PET data acquired in continuous bed motion (CBM) is evaluated in this study.

Methodik/Methods PET/CT acquisitions of 56 patients with suspected lesions in the thorax or the abdomen were included into this study. All underwent whole-body CBM PET/CT scans (Siemens mCT, 4 MBq/kg [¹⁸F]FDG, 1 h p.i., CBM speed: 1.1 mm/s) with conventional respiration measurement using the Anzai belt. A dedicated CBM-capable algorithm was additionally used to derive gating signals from the PET raw data. The respiratory signals were compared by correlation analysis and used for subsequent optimal gating and full motion correction. Lesion SUV_max, SUV_mean, and metabolic volumes were analyzed for differences between static, gated, and fully-corrected reconstructions.

Ergebnisse/Results On average, correlation coefficients between belt and DDG signals were highest when scanning regions close to kidneys and liver (r = 0.89 ± 0.07), decreasing to values around 0 around the bladder and the lung apex. In total, 196 lesions were identified. Gated and motion-corrected images demonstrated significant increases in SUV_max and SUV_mean, and decreases in volumes as compared to the static reconstructions. No significant differences were observed between both gated and corrected images based on either the Anzai or the DDG signal.

Schlussfolgerungen/Conclusions The investigated DDG algorithm for CBM scans resulted in images comparable in quality to conventional hardware-based gated and corrected images.