Room: AAPM ePoster Library
Purpose: Clinical SPECT/CT can offer quantitative SPECT images (voxel value unit: bq/ml), providing for measurement of tracer activity in volumes-of-interest (VOIs) within patients. Validation of the accuracy of methodology implemented is critical.
Method: An IEC Image Quality phantom cover was customized by 3D printing to hold 1-2-3-4-5-6 cm diameter hollow spheres with a minimum separation of 2.5 cm between surfaces. Spheres and phantom chamber were filled with known Tc-99m activity concentrations. Acquisition with (8:1 hot to background activity ratio) and without background activity were performed. The GE Discovery 670 with Q.metrix, and the Philips Brightview with SUV SPECT by Hermes were tested. Clinical acquisition and reconstruction protocols were used.
Results: 3D printing offered design flexibility, but had difficulty with water leakage through printed materials requiring additional sealing. In the experiment without background activity, the ratio of the measured to true activity (TF) within VOIs matching sphere dimensions were 0.33-0.85 for the 1-6 cm spheres, respectively. When the VOI diameters were expanded by 2 cm, TF were 0.81-0.97. For acquisition with background activity, TF were 0.47-0.89 for 2-6 cm spheres (1-cm sphere had no activity) for matching VOIs. For VOIs expanded by 2 cm in diameter, the TF was 1.86-1.09. Applying background correction to subtract activity only within the expanded 2 cm VOI shell, TF were 0.79-0.97. The TF of background activity concentrations were 0.96-1.02.
Conclusions: Expanded sphere VOI's in acquisitions with no background activity to include spill-out of activity from partial-volume effects can be used to assess quantitative accuracy. Expansion of VOI's to include spill-out activity in the presence of background activity is possible with proper correction. The background concentrations were closely matched. Further evaluation of these systems may be able to improve the overall accuracy.
Not Applicable / None Entered.