2288 - Proton Range-Based Patient Setup
Presenter(s)
W. Yao1, B. Y. Yi1, M. Huang2, and B. Zhang3; 1Maryland Proton Treatment Center, Baltimore, MD, 2Department of Radiation Oncology, Medical College of Wisconsin, Milwaukee, WI, 3Department of Radiation Oncology, University of Maryland School of Medicine, Baltimore, MD
Purpose/Objective(s): For proton therapy, image-guided patient setup may not guarantee the desired dose delivery. This is because the delivered dose is sensitive to the proton range, but the image guidance, which is practiced currently is based on patient geometry, not based on the proton range. This study investigated the delivered dose to patients under image-guided and range-based setups.
Materials/Methods: 3 representative proton head-and-neck patients with QACTs and CBCTs were studied. Treatment lesions in these cases were located at the right neck, on the skin to the skull and at the base of the tongue in case 1, case 2 and case 3, respectively. A range-based registration algorithm was developed to register the QACTs and CBCTs to the planning CT. The algorithm rigidly searched the registration parameters (shifts and rotations) to minimize the deviation of water equivalent thickness (WET) and meanwhile balanced the WETs between their corresponding D95 and V98. The WET was the surrogate for the proton beam range from the beam source to the surface of the target of interest. The WETs in the planning CT were used as the reference for the QACTs, and the WETs in the 1st day CBCT were used as the reference for the following CBCTs. The WET deviations corresponding to D95=95 and V98=90 were determined by analyzing the robustness of the treatment plan. With the obtained range-based registration parameters, QA-plan evaluation procedures were performed to deform the target contours from the planning CT to QACTs and then to calculate dose from the treatment plan to the QACTs in a treatment planning system. The image-guided registration between the QACTs and planning CT was performed by the dosimetrist during the QA-plan evaluation. For CBCTs, the image-guided registration between the CBCT and planning CT was done by the therapist during patient setup. Range-based and image-guided registration parameters were applied to the planning CT to calculate the dose. Ideally, if there is no difference between these two registrations, the dose will be the same as the planned dose, and otherwise the dose will be different.
Results: For case 1, D95=87.7% in a QACT (a replan was triggered due to low D95) registered to the planning CT by image guidance, but D95=98.9% in the QACT with extra shifts (4, 1, -1) mm in the lateral, AP and SI directions from image-guided registration. The extra shifts were obtained by range-based registration. For case 2, the results from CBCTs registered to the planning CT by image guidance showed that D95=92.8% on average. For the 3rd case, V100 was improved in a QACT from 91.8% to 95.0% by extra shifts (0, 1, 1) mm. From an image match point of view, both registrations were acceptable.
Conclusion: Image-guided patient setups for all cases used in this study were acceptable, but delivered doses were not acceptable unless extra shifts were applied. Range-based registration significantly improved proton dose delivery to the intended targets, especially in the 1st and 2nd cases where beams tangential to the skin were used.