Quantization accuracy of short-duration respiratory-gated PET/CT acquisitions.

Purpose: PET/CT acquisitions are affected by physiological motion, which lowers the quantization accuracy. Respiratory-gated PET/CT methods require a long acquisition time, which may not be compatible with the clinical schedule. The objective of the present study was to assess the quantization accuracy of short-duration, respiratory-gated PET acquisitions and processing with the "CT-based" methodology developed in our laboratory.

Methods: Quantization accuracy was first assessed in a phantom study. A standard ("Ungated") PET/CT acquisition was followed by a 10-minute list-mode acquisition with simultaneous respiratory signal recording and a short breath-hold CT scan (BH-CT). These acquisitions were repeated 10 times. For the CT-based images, we reconstructed (i) 10 full-duration (FD-CT-based) volumes that took account of all events recorded in the position defined by BH-CT and (ii) 10 short-duration (SD-CT-based) volumes based on only 30 seconds of selected events. Using these volumes, we performed a bias-variance analysis to assess the effects of respiration-motion reduction and the counting statistics on the quantization accuracy. We also applied Ungated, FD- and SD-CT-based methods to 16 patients (21 pulmonary lesions) and measured the maximum standardized uptake (SUVmax) values.

Results: The bias values were 71%, 40% and 44% for Ungated, FD- and SD-CT-based images, respectively. In the clinical study, there was a statistically significant difference in SUVmax between Ungated images and both the CT-Based images (p < 0.02) but not between the FD-CT-Based and SD-CT-Based images (p = 0.42).

Conclusion: Our findings demonstrated that the additional acquisition time required by the CT-based method can be reduced without altering quantitative accuracy.