Can Q.Clear reconstruction be used to improve [68 Ga]Ga-DOTANOC PET/CT image quality in overweight NEN patients?

ABSTRACT

AIM/INTRODUCTION: Digital PET/CT allows Q.Clear image reconstruction with different Beta (ß) levels. However, no definitive standard ß level for [68 Ga]Ga-DOTANOC PET/CT has been established yet. As patient's body mass index (BMI) can affect image quality, the aim of the study was to visually and semi-quantitatively assess different ß levels compared to standard OSEM in overweight patients. MATERIALS AND METHODS: Inclusion criteria (1) patients with NEN included in a prospective CE-approved electronic archive; (2) [68 Ga]Ga-DOTANOC PET/CT performed on a digital tomograph between September2019/March2021; (3) BMI ≥ 25. Images were acquired following EANM guidelines and reconstructed with OSEM and Q.Clear with three ß levels (800, 1000, 1600). Scans were independently reviewed by three expert readers, unaware of clinical data, who independently chose the preferred ß level reconstruction for visual overall image quality. Semi-quantitative analysis was performed on each scan SUVmax of the highest uptake lesion (SUVmax-T), liver background SUVmean (SUVmean-L), SUVmax-T/SUVmean-L, Signal-to-noise ratio for both liver (LSNR) and the highest uptake lesion (SNR-T), Contrast-to-noise ratio (CNR). RESULTS: Overall, 75 patients (median age 63 years old [23-87]) were included pre-obesity sub-group (25 ≤ BMI < 30, n = 50) and obesity sub-group (BMI ≥ 30, n = 25). PET/CT was positive for disease in 45/75 (60.0%) cases (14 obese and 31 pre-obese patients). Agreement among readers' visual rating was high (Fleiss κ = 0.88) and the ß1600 was preferred in most cases (in 96% of obese patients and in 53.3% of pre-obese cases). OSEM was considered visually equal to ß1600 in 44.7% of pre-obese cases and in 4% of obese patients. In a minority of pre-obese cases, OSEM was preferred (2%). In the whole population, CNR, SNR-T and LSNR were significantly different (p < 0.001) between OSEM and ß1600, conversely to SUVmean-L (not significant). These results were also confirmed when calculated separately for the pre-obesity and obesity sub-groups ß800 and ß1000 were always rated inferior. CONCLUSIONS: Q.Clear is a new technology for PET/CT image reconstruction that can be used to increase CNR and SNR-T, to subsequently optimise overall image quality as compared to standard OSEM. Our preliminary data on [68 Ga]Ga-DOTANOC PET/CT demonstrate that in overweight NEN patients, ß1600 is preferable over ß800 and ß1000. Further studies are warranted to validate these results in lesions of different anatomical region and size; moreover, currently employed interpretative PET positivity criteria should be adjusted to the new reconstruction method.