Feasibility study on clinical application of 18F-FDG PET/CT imaging “uWS - MI” medical image processing software

Objective：By comparing with “GE AWS46 workstation” developed by GE Company in the United States, the effectiveness of “uWS-MI” medical imaging software developed by Shanghai United Imaging Healthcare Co., Ltd. was evaluated, and whether its stability, reliability, and data management met the clinical requirements was assessed. Methods：A total of 30 patients with malignant tumor were analyzed by the 18F-FDG PET/CT imaging (Discovery VCT 64, GE, U.S.) before and after the treatment. On “uWS-MI” and “GE AWS46” workstations, a typical lesion was segmented by semi-automatic one-click method. Two nuclear medicine physicians, who both had the qualifications of nuclear medicine and CT certification, conducted a blind evaluation using the 3-point method (satisfaction, general, and unsatisfactory) for the evaluation of the tumor completeness and contouring accuracy. The volume, standard uptake value (SUV) maximum, SUV average, SUV threshold, SUV peak, total lesion glycolysis (TLG) value of a typical tumor lesion in primary PET/CT imaging were calculated based on tumor segmentation. At the same time, the two physicians evaluated the interaction and workflow of “uWS-MI” medical image processing software, as well as whether the stability, reliability, and data management of the system met the clinical requirements. Results：There was no significant difference in the evaluation of lesion segmentation between the two physicians using “uWS-MI” and “GE AWS46” independently. The primary PET/CT images from “uWS-MI” medical imaging software and “GE AWS46” workstation based on tumor segmentation calculation showed that a typical tumor’s volume, SUV maximum, SUV average, SUV threshold, and TLG were not significantly different, but SUV peak from “uWS-MI” medical imaging software was greater than that from “GE AWS46” workstation (t=3.36, P=0.002). The two physicians respectively evaluated the interaction and workflow of “uWS-MI” medical image processing software, and the stability, reliability and data management of the system. The satisfaction rates of the evaluation were all 100%. Conclusions：The segmentation effect of “uWS-MI” medical image processing software on lesions, as well as the acquired lesion volume and SUV data could be used for clinical diagnosis, and the interaction and workflow of the software are satisfactory, which can be popularized and applied in clinical practice.