Can a penalized-likelihood estimation algorithm be used to reduce the injected dose or the acquisition time in 68Ga-DOTATATE PET/CT studies?

Abstract

Background: Both image quality and quantitative accuracy of PET depend on several factors such as uptake time, scanner characteristics and image reconstruction methods. Ordered subset expectation maximization (OSEM) is considered today the gold standard for image reconstruction. Penalized-likelihood estimation (PL) algorithms have been recently developed for PET reconstruction to improve quantitation accuracy while maintaining or even improving image quality. In the present study, we aim to compare the performance of a PL algorithm (Q.Clear, GE Healthcare) and 3D OSEM for 68Ga-DOTATATE (68Ga-DOTA) PET studies, both visually and quantitatively. Thirty consecutive whole-body 68Ga-DOTA studies were included. The data were acquired in list mode and reconstructed using 3D OSEM and Q.Clear with various values of the regularization parameter β, and various acquisition times per bed position (bp), thus generating images with reduced injected dose (1.5 min/bp: β=300-1100; 1.0 min/bp: β=600-1300 and 0.5 min/bp: β=800-2200). Evaluation was performed using a phantom and clinical data. Finally, two experienced nuclear medicine physicians blinded to the variables assessed the image quality visually. Results: Clinical images reconstructed with Q.Clear , set at 1.5 and 1.0 min/bp using β = 1100 and 1300 respectively, resulted in images with noise equivalence to 3D OSEM (1.5 min/bp) with a mean increase in SUVmax of 14 % and 11%, in SNR of 18% and 10%, and in SBR of 14% and 12%, respectively. Reconstruction using 0.5 min/bp and β = 2200 resulted in SUVmax, SNR and SBR with a relative difference < 1%. Visual assessment yielded similar results with mean scores for Q.Clear (1.5, 1.0 and 0.5 min/bp) vs 3D OSEM (1.5 min/bp) of 3.58 vs 3.38, 3.64 vs 3.47 and 3.60 vs 3.61, respectively. Conclusion: 68Ga-DOTA reconstructions with Q.Clear, 1.5 and 1.0 min/bp resulted in increased tumor SUVmax and in improved SNR and SBR at a similar level of noise compared to 3D OSEM. Q.Clear with β =1300 enabled a one-third reduction of acquisition time or injected dose, with similar image quality compared to 3D OSEM.

Phantom and clinical evaluation of a new Bayesian penalized likelihood reconstruction algorithm HYPER Iterative on the image quality of 68Ga-DOTA-NOC PET/CT

“…For 68 Ga tracer, one preliminary study suggested BPL algorithm with a penalization factor between 350 and 450 could be used in clinical practice for all studied radiopharmaceuticals [15] , while a 68 Gallium-DOTA0-1NaI3-octreotide( 68 Ga-DOTA-TOC) PET study recommended that a factor equal to 400 or higher than 400 could be optimal [16] . However, another 68 Gallium-DOTA0-Tyr3-octreotate ( 68 Ga-DOTA-TATE) PET/CT study concluded that BPL algorithm with the penalization factor within 1100 to 1400 with 1.5 m/b resulted in better image quality than OSEM, while choosing from 1300 to 1600 for 1 m/b also had better image quality [17] . In line with those studies, our study suggested a penalization factor between 0.03 and 0.5 for HYPER Iterative reconstruction that provided higher lesion contrast compare to OSME, while lower image noise was acquired between 0.1 and 0.5.…”

“…In clinical practice, the penalization factor should be xed to maintain the consistency of SUV. In a previous study of 68 Ga-DOAT-TATE PET, the optimal penalization factor was determined from the equivalent noise group which had increased tumor SUV max and improved signal-to-background ratio [17] , while it also depended on lesion detection rate and a high patient throughput in oncologic whole-body 18 F-FDG examinations [14] . Since the radiologists remain the intention to diagnosis as center task, it must be appropriate to consider the raters' preference on the choosing of optimal noise penalization factor.…”

Digitale SPECT und PET: Klinische Konsequenzen

“…Die Bayesian penalized Likelihood (BPL)-Bildrekonstruktion ist aktuell als Q.Clear der Firma General Electric (GE) kommerziell verfügbar. Mehrere klinische Studien konnten zeigen, dass die Bildqualität und das Bildrauschen unter Verwendung des BPL-Algorithmus für verschiedene Fluor-18- [49][50][51][52] und Gallium-68-markierte Radiopharmaka [53][54][55] den OSEM-basierten Rekonstruktionsverfahren überlegen sind. Auch konnte bereits mehrfach eine höhere Läsionsdetektionsrate als mit OSEM-basierten Verfahren beobachtet werden, sogar wenn Letztere eine TOF-und/oder PSF-Rekonstruk-FA ZIT Take Home Message Zusammenfassend zeigt die Literatur, dass Verbesserungen der Bildqualität, der zu injizierenden Aktivität und der Läsionsdetektion mit der SiPM-Technologie wahrscheinlich sind, auch wenn es in Patientenstudien schwer ist, den klinischen Mehrwert gegenüber PMT-basierten Scannern einwandfrei nachzuweisen.…”