A prospective study to determine the need for physical simulation following virtual simulation

Purpose: Virtual simulation is CT based planning utilizing computed digitally reconstructed radiographs (DRRs) in a manner similar to conventional fluoroscopic simulation. However, conventional or physical simulation is still widely used to assure precise implementation of the devised plan. To evaluate the need for performing physical simulation, we prospectively studied patients undergoing virtual simulation who either had or did not have a subsequent physical simulation. Materials and Methods: From July, 1995 to February, 1996, 48 patients underwent conformal 4-field radiation therapy for prostate cancer using a commercial grade spiral CT simulator. All patients were immobilized in a foam body cast and positioned by using a fiducial laser marking system. Following prostate and seminal vesicle definition on a slice-by-slice basis, virtual simulation was performed. The isocenter defined by this process was marked on both the patient and the immobilization device before leaving the CT simulator room. The isocenter position of the devised plan was evaluated by three verification methods: physical simulation, first day treatment port filming, and port filming immediately following physical simulation. Simulator radiographs and port films were compared against DRRs for x, y, and z deviations of the isocenter. These deviations were used as a measure of the implementation precision achieved by each verification method. Results: Thirty-seven patients underwent physical simulation and first day port filming. Eleven had first day treatment verification films only and never had a physical simulation. A total of 79 simulator radiographs and 126 first day treatment port films were reviewed. The tabulation of all deviations is as follows: There was significantly more setup error (≥ 5 mm) observed when the devised treatment was implemented in the treatment room as opposed to the physical simulator. The physical simulator did not lead to a significant reduction in setup error when the subsequent on-treatment verification was considered. Conclusions: Following virtual simulation, physical simulation was more precise in reproducing patient setup than first day treatment verification but did not translate into better execution of the treatment plan. Since physical simulation did not appear to improve on-treatment setup, we are currently exploring whether it can be eliminated from the process of virtual simulation