Portal de Periódicos da CAPES

Performance of a Continuous Bed Motion Data Driven Respiratory Gating algorithm with low count PET acquisitions

2020; Society of Nuclear Medicine and Molecular Imaging; Volume: 61; Linguagem: Inglês

1535-5667

Joseph Meier, Harrison Vickers, Osama Mawlawi,

MRI in cancer diagnosis

1467 Objectives: Data driven respiratory gating (DDG) determines respiratory waveforms (WF) from the acquired PET data alone, overcoming the need for external devices to record the respiratory WF. Previous studies showed that the performance of current DDG techniques degrade with decreasing number of detected coincidence events. However, this has not been determined for a recently introduced DDG method developed for PET data acquired in continuous bed motion (CBM_DDG)[1]. Our aim is to determine the impact of reducing coincident count density on the DDG signal determination and corresponding radiotracer quantification in PET/CT imaging. Methods: 41 patients (Mean Age: 62.2, BMI: 28, injected activity: 320.3 MBq of F18-FDG, imaged post injection: 69.1 min) were imaged on a SIEMENS mCT scanner in CBM mode. The Anzai device was used to record the breathing WFs and considered as the gold standard. To simulate acquisitions with reduced counts, detected events were randomly removed from the original list dataset to create list datasets with 75, 50, 25, 12.5, and 6% of the original counts respectively. The CBM_DDG WF was then determined for each patient and list dataset retrospectively. The two WFs (Anzai and CBM_DDG) were then compared by calculating the correlation coefficients over the zone between the aortic arch to the center of the right kidney for each list dataset. Additionally, motion compensated PET images, using each of the WFs (Anzai and CBM_DDG) and list datasets, were then reconstructed using an elastic motion correction via deblurring (EMDB) algorithm that was recently introduced (ONCOFREEZE) and the resultant images(EMDB_DDG and EMBD_ANZ) were compared. The EMDB reconstructions were performed with 2 iterations 21 subsets, TOF, PSF, 200 x 200 matrix, a 35% duty cycle for the reference image and 5 mm post filtration. Lesion SUVmax and SUVpeak were then measured (1 lesion per patient) and the ratio of EMDB_DDG to EMDB_ANZ for SUVmax and SUVpeak was calculated. Due to setup or hardware malfunction, 10/41 patients had abnormal Anzai respiratory WFs and a second analysis of the WF correlation coefficients and SUV ratios was performed when leaving out the cases with abnormal Anzai WFs. Results: When including all patients, the average(standard deviation) correlation coefficient for the reduced count acquisition was: 0.72(0.24), 0.70(0.24), 0.65(0.28), 0.55(0.30), 0.46(0.28), 0.25(0.30) for the 100, 75, 50, 25, 12.5, and 6% list datasets respectively. Reanalysis of this data using only the normal Anzai WFs showed 0.80(0.17), 0.78(0.18), 0.75(0.20), 0.65(0.25), 0.57(0.22), 0.29(0.33) respectively. When including all patients, the average(standard deviation) of the SUVmax ratios was 0.98(0.06), 0.98(0.05), 0.99(0.06), 0.97(0.08), 0.95(0.14), 0.97(0.18) and for SUVpeak was 0.99(0.03), 1.00(0.03), 1.00(0.04), 0.98(0.06), 0.97(0.09), 0.97(0.12) for the 100, 75, 50, 25, 12.5, and 6% list datasets respectively. Reanalysis of this data using only the normal Anzai WFs showed SUVmax ratios of 0.98(0.05), 0.98(0.06), 0.99(0.06), 0.97(0.07), 0.93(0.15), 0.95(0.18) and SUVpeak ratios of 0.99(0.03), 1.00(0.03), 1.00(0.04), 0.99(0.06), 0.96(0.09), 0.97(0.12) respectively. Conclusions: This study shows that as coincidence count density decreases, the ability of the CBM_DDG algorithm to determine the respiratory WF degrades as indicated by the decreasing correlation coefficients. However, the effects of these degradations in the CBM_DDG WF have little impact on the SUV measurements. Performing the analysis when removing the cases with abnormal Anzai WFs did not alter these conclusions. [1] Schleyer P, Hong I, Jones J, Hamill J, Panin V, Fuerst S. Data-Driven Respiratory Gating Whole Body PET Using Continuous Bed Motion. In: 2018 IEEE Nuclear Science Symposium and Medical Imaging Conference (NSS/MIC). IEEE; 2019:1-5.