Author: Chunhui Han, An Liu, William T. Watkins, Qiuyun Xu 👨🔬
Affiliation: Department of Radiation Oncology, City of Hope National Medical Center 🌍
Purpose: To evaluate the RefleXion X1 imaging system as a standalone positron emission tomography and computed tomography (PET/CT) simulator for radiotherapy treatment planning and adaptive re-planning.
Methods: RefleXion PET/CT (rfxPET/rfxCT) and diagnostic PET (dxPET) were registered to planning CT (dxCT) for ten patients. Seven Patients received dxPT 12-43 days prior to rfxCT; 3 patients were imaged in a same-day workflow (rfxPET/rfxCT followed by dxPET) with a single 15 mCi injection. The image quality of rfxCT including spatial resolution, signal and contrast to noise ratios (SNR and CNR) were measured in cheese phantom and in patients using mean and standard deviation (sd) on low contrast structures fat (HU=-150,50), water (HU=-50,50), muscle (HU=50-150), and cortical bone (HU>150). CT noise is reported as a percentage of the attenuation coefficient of water. The CNR of each ROI is calculated as CNRROI=(meanROI-meanwater)/sdwater. PETs were compared by SNR and CNR in GTV and rings 1-2 cm, 2-3 cm, and (background) 3-4 cm. The CNR of GTV with the 1-cm ring (GTV-CNR) is critical to visualization of tumors.
Results: The rfxCT scanner demonstrated excellent spatial resolution (≤1mm), noise <3.5%, and CT-number integrity. The average CNR for fat/water, muscle/water, and bone/water in rfxCT vs. dxCT was (34.0,23.4,161.4) vs. (-32.3,20.8,146.1) confirming this scanner can be used for adaptive re-planning. The rfxPET SNR dropped by a factor of 1.75±0.42 in 7/10 patients and increased in 3 others. The PET GTV-CNR dropped by an average factor of 1.8±0.7 in 8/10 patients on rfxPET, with two patients showing significantly higher loss of GTV-CNR. For one patient, GTV-CNR dropped to <1, and was not visible on rfxPET, all other patients successfully completed PET-guided treatment.
Conclusion: The RefleXion X1 can operate as a standalone PET/CT simulator, but the PET system demonstrates a significant drop in tumor SNR and CNR compared to dxPET.