Abstract
149
Objectives 4D PET/CT can be used to suppress motion artifacts by correlating the PET raw data with the breathing phase. The accuracy of each PET phase is dependent on the reproducibility and consistency of the breathing cycle during acquisition. The objective of this study is to evaluate the impact of breathing amplitude and phase irregularities on the quantitative accuracy of 4D PET. We demonstrate a simple technique of using 4D CT to derive quantitatively accurate SUVs from a non-gated PET.
Methods Phantom studies were performed using simulated lesions ranging in diameter from 9 mm to 47 mm. Motion was simulated using patient breathing data. We compared the accuracy of SUVs derived from gated PET (4 bins) for breathing patterns ranging from ideal to highly irregular. The combined effects of motion and partial volume errors were also simulated in a digital phantom to derive a set of correction factors that can be applied to a non-gated PET based on the tumor size and motion amplitude.
Results For an ideal breather, gated PET produced SUVs that were within (3.4±6.8)% of the true value. As breathing pattern becomes more irregular, the quantitative accuracy of gated PET decreased. Gated PET underestimated SUVs by (7.3±12.2)% for an average breather and (13.6±14.7)% for a highly irregular breather. Using the simulated correction factors on a non-gated PET, corrected SUVs were within (0.8±7.6) % of the true value.
Conclusions Breathing irregularities can significantly impact the quantitative accuracy of gated PET/CT. In practice, gated PET still underestimates the true lesion activity due to motion. Applying correction factors to a non gated PET is potentially more robust against breathing irregularities and can correct for the combined effects of both motion and partial volume errors