Benchmarking 18F and 68Ga Phantom Performance on a Digital Time-of-Flight PET/CT System: Reference Values for Spatial Resolution, Contrast Recovery, and Background Variability.
TL;DR
Under matched reconstruction conditions, 18F retained superior intrinsic resolution and contrast recovery, whereas longer acquisitions mainly reduced noise and could not overcome the physical resolution limitations of 68Ga.
Abstract
Our purpose was to establish reference performance benchmarks for spatial resolution, contrast recovery coefficient (CRC), and background variability (BV) for 68Ga and 18F under matched conditions and to determine the extent to which acquisition duration can compensate for isotope-specific resolution differences. Methods: Phantom studies were performed on a digital PET/CT platform using capillary sources in air (minimal scatter) and a triple-line insert in water (scatter conditions). A National Electrical Manufacturers Association International Electrotechnical Commission body phantom (spheres 10-37 mm; 4:1 activity ratio) was used to assess CRC (International Atomic Energy Agency definition) and BV (coefficient of variation). Data were analyzed using Kruskal-Wallis tests for resolution and 2-way ANOVA for CRC/BV, evaluating the interaction between isotope and acquisition duration. Results: 18F consistently yielded resolution limits (full width at half maximum) superior to those of 68Ga in both air and water (Kruskal-Wallis, P ∼ 0.008). Notably, extending the acquisition duration from 2 to 5 min provided no significant improvement in full width at half maximum for either tracer (P = 0.70). In National Electrical Manufacturers Association testing, 18F maintained a significant CRC advantage across all sphere sizes and durations. At 10 min, the CRC for the 37-mm sphere reached about 97.5% for 18F versus about 81.3% for 68Ga. For small spheres (10-13 mm), 18F exceeded 68Ga recovery by about 14%-22%. Although increasing scan duration from 2 to 10 min significantly reduced BV (noise) for both tracers (18F: 1.92%→1.37%; 68Ga: 2.47%→1.29%; P = 0.026), it did not bridge the CRC gap. Conclusion: This work provides practical benchmark values for 18F and 68Ga on a digital PET/CT platform. Under matched reconstruction conditions, 18F retained superior intrinsic resolution and contrast recovery, whereas longer acquisitions mainly reduced noise and could not overcome the physical resolution limitations of 68Ga. These data support quality assurance, protocol optimization, and harmonization of prostate-specific membrane antigen PET imaging across both tracers.