A New Vision for PET Patients: Clinical Application of Scan Performance Gains from Digital PET-CT
- Programme
- STP
- Specialty
- Imaging with Ionising Radiation
- Author
- Kate van Elteren
- Training location
- Liverpool University Hospitals NHS Foundation Trust/The Clatterbridge Cancer Centre
This work has explored the potential of ‘digital’ (SiPM) PET-CT systems in terms of imaging protocol adjustments with respect to anticipated improvements in performance compared with ‘analogue’ (APD) PET systems. Two PET scanners – the analogue mCT-Flow™ and the digital Vision™ – were compared in terms of image quality performance, using a range of standard phantom scans performed at varying bed speeds between 0.6 – 4.4 mm/s. SNR (COV) and comparisons between expected to measured concentration ratios were the predominant quantitative metrics of exploration in this work.
A lower limit on feasible bed speed adjustments was subsequently established by modelling equations to ‘COV-match’ anticipated image quality between the two scanners. The proposed bed speed adjustments from this work were subsequently implemented into the clinical protocol for the Vision™ scanner, leading to an approximately 29% reduction in mean acquisition time. An audit of patients pre- (n = 60) and post- (n = 33) protocol implementation (with similar distributions in weight and injected activity) revealed only modest changes to the SNR distributions between both patient cohorts, with a shift of −8% in the distribution mean for the UHD reconstructions and no change in distribution detected in the Iterative case. Moreover, there was no evidence found for any changes in liver SUV distribution between the two cohorts, in line with predictions from phantom scans with respect to recovered concentration.
Exploring further protocol adjustments, it was found that matrix size could be increased from 128 × 128 to 220 × 220 at patient-like activity concentrations with little effect on image COV. Finally, a comparison was included between current protocol and a neighbouring imaging centre also in possession of a Vision™ scanner. Due to practical capacity in terms of patient throughput on the Vision™ presently at the author’s centre, two potential dose regimens were offered for exploration in place of further bed speed increases: one based on achieving a mean liver SNR of ∼20 (as in the neighbouring imaging centre) and one based on achieving image quality overlap with the mCT-Flow™ (mean liver SNR ∼ 12.6 in a cohort of 50 patients).
In short, even with the increases in bed speed resulting from this work, it should theoretically be possible to reduce injected activity between 33-73%, in line with clinician image quality requirements. Further exploratory work is needed to establish the feasibility of implementing a suitable activity regimen, particularly in larger patients.