Room: Exhibit Hall | Forum 6
Purpose: To investigate geometric and dosimetric differences between two image registration workflows for frameless patient treatments on the Leksell GammaKnifeÂ® Iconâ„¢: registering the planning MRI to a diagnostic CT, which is then in turn registered to the stereotactic reference CBCT, or registering the planning MRI directly to the stereotactic reference CBCT.
Methods: Nine targets were treated in eight patients using the current clinical registration chain of planning MRI (3D T1-weighted scan) to diagnostic CT to stereotactic CBCT. For each case, the planning MRI was used for target, OAR, and skull definition, and registrations were performed using the GammaPlanÂ® automatic registration software using the full available MRI as the volume of interest. Alternate treatment plans were created using the same shots, but removing the diagnostic CT from the registration chain. Geometric differences between the target center locations and the resulting changes in target coverage and selectivity were recorded and analyzed for dependence on target volume and location.
Results: The mean 3D displacement of target centers between the clinical workflow and the CBCT-MRI workflow was 0.98Â±0.56mm (max=2.3mm). The mean decrease in target coverage was 2.6%Â±1.5% (max=5%), and the mean decrease in selectivity was 1.7%Â±1.3% (max=4%). The superior-inferior direction had the largest mean shift of 0.6mm and the largest maximum shift of 1.9mm. There was no observed dependence of displacement or coverage loss on target size or location.
Conclusion: Geometric and dosimetric differences exist between the two registration workflows, which have the potential to affect patient outcomes. Additional patient data is necessary to determine the target locations and sizes which are most susceptible to these differences.