Measurement accuracy and practicality of ¹⁵O-gas PET using a high-sensitivity whole-body PET/CT scanner

¹⁵O-gas PET allows for the quantitative measurement of cerebral circulation and metabolism. However, the administration of high radioactivity doses raises concerns about count losses, particularly when using high-sensitivity whole-body PET/CT scanners equipped with BGO detectors. Therefore, optimizing the administration dose is crucial. This study aims to evaluate the validity of ¹⁵O-gas PET performed with such scanners.

A 3D-PET/CT scanner, Discovery IQ, which has a 5-ring configuration with a 26 cm axial field of view, was used. A 100 mL saline pack containing ¹¹C at 460 MBq was placed on a 3D Hoffman brain phantom (HBP) filled with ¹⁸F at 5.0 kBq/mL, and dynamic scans were performed. The quantitative capability of measured radioactivity in the HBP was evaluated. Subsequently, the HBP filled with ¹⁸F at 5.0 kBq/mL was equipped with a gas mask system, and ¹⁵O-gas was administered at flow rates ranging from 3000 MBq/min, decreasing in 500 MBq/min increments, while dynamic scans were performed. The quantitative accuracy of radioactivity in HBP and count rate characteristics were evaluated. To confirm the feasibility of the ¹⁵O-gas PET method in the clinical situations both practically and quantitatively, cerebral blood-flow (CBF) values obtained via PET and SPECT were compared in 10 patients with cerebrovascular diseases. CBF was calculated by the autoradiography method. Analyses included standard brain transformation and alignment, followed by averaging CBF values from 26 brain regions for comparison.

The measured radioactivity in the HBP approached the true value as the radioactivity of ¹¹C in the saline pack decreased. To achieve a quantitative error within ± 5%, the radioactivity of ¹¹C needed to be below 40 MBq. Similarly, for ¹⁵O-gas PET, an administration rate below 1000 MBq/min was needed to maintain quantitative accuracy within ± 5%. When exceeding 1500 MBq/min, random events surpassed 50%, compromising reliability. The correlation coefficient between PET and SPECT-derived CBF was 0.39 (p < 0.001). After excluding two outliers, it improved to 0.71 (p < 0.001), indicating a strong correlation. PET-measured CBF tended to be higher than SPECT.

These findings suggest that an administration rate of 800 MBq/min for ¹⁵O-gas provides acceptable quantitative parameters for cerebral circulation and metabolism. Clinical component of this study was registered at the UMIN Clinical Trials Registry ( www.umin.ac.jp/ctr/ ) as # UMIN000046761.