Method for measuring PET/CT fusion accuracy

Objectives PET/CT has been widely used in clinical application. One of its vital advantages is that fused PET functional information and CT anatomical structure can provide more accurate diagnosis than images from individual modality. Hence, the accuracy for image fusion plays very important role in the PET/CT imaging display. Since there is no measurement standard can be referred in standards by the National Electrical Manufacturers Association (NEMA) and the International Electrotechnical Commission (IEC), this article demonstrated a simple method to measure the performance of PET and CT system fusion with very low error.

Methods The size and position of point sources can result in inaccurate estimation of deviation between PET and CT system. Four capillary tubes with inner diameter 0.6mm and outer diameter 1mm were used as the line sources for measurement by absorbing the mixed solution of 0.15ml 20uCi 18F-FDG and 0.2ml iohexol. The tubes labeled A, B, C and D were respectively bound on each edge of a cuboid and kept parallel to each other. CT and 1 min PET scanning of the cuboid were individually performed on uMI-510 PET system (United Imaging Healthcare). During the scanning, the capillary were placed parallel to the axial to obtain the deviation in axis Z and placed vertical to the patient bed to obtain the deviation in axis X and axis Y. Above scans were repeated after a 78Kg phantom was put on the patient bed. The coordinates of the capillary tubes’ center were located independently by using the uIdeal-MI Fusion software (United Imaging Healthcare) in all the enlarged PET and CT images. The fusion deviations between PET and CT system in axis X and Y were obtained with the images from data when capillary parallel to the axial. Similarly, the fusion deviation in axis Z were obtained with the images from data when capillary vertical to the patient bed.

Results The measured deviations from PET to CT respectively for capillary A, B, C and D are -0.5mm, 0.1mm, 0.1mm, 0.1mm in X axis, 0.2mm, 0.7mm, 0.2mm, -0.1mm in Y axis and 0.1mm, 0.1mm, 0.3mm, 0.0mm in Z axis when no burden was on the patient bed, and -0.7mm, -0.2mm, 0.0mm, -0.2mm in X axis, -0.2mm, 0.1mm, 0.2mm, 0.4mm in Y axis, and 0.7mm, 0.1mm, 0.3mm, 0.0mm in Z axis while the 78Kg phantom was on the patient bed.

Conclusions A new method for fusion deviation measurement was provided. By applying four line sources and measuring in different directions, the accuracy of deviation measurement in each direction was optimized to less than 1mm, and the accuracy was also measured when there was weight on patient bed. Test with more PET/CT systems will be performed to evaluate this method.