E. A. Simiele1, I. O. Romero2, J. Wang3, Y. Chen4, Y. Lozko3, Y. Severyn3, L. Skinner4, Y. Yang5, L. Xing5, I. C. Gibbs4, S. M. Hiniker4, and N. Kovalchuk5; 1University of Alabama Birmingham School of Medicine, Birmingham, AL, United States, 2Stanford University School of Medicine, Palo Alto, CA, 3Stanford University, Stanford, CA, 4Department of Radiation Oncology, Stanford University School of Medicine, Stanford, CA, 5Department of Radiation Oncology, Stanford University, Stanford, CA
Purpose/Objective(s): Create a comprehensive automated solution for pediatric and adult VMAT CSI including contouring, planning, and plan check to reduce planning time and improve plan quality. Materials/
Methods: Seventy-seven previously treated CSI patients (age, 2-67 y.o) were used for creation of an auto-contouring model to segment 25 organs at risk (OARs). The auto-contoured OARs were evaluated using the Dice Similarity Coefficient (DSC), 95% Hausdorff Distance (HD95), and a qualitative ranking by two experts (scale: 1-acceptable, 2-minor edits, 3-major edits). The auto-planning scripts were developed using the Eclipse Scripting Application Programming Interface and tested with 20 previously treated patients treated with either low-dose VMAT-CSI (12 Gy) or high-dose VMAT CSI (36 Gy + 18 Gy boost). Clinically relevant metrics, planning time, and blinded physician review were evaluated for significant differences between the auto and manual plans. Finally, the plan preparation for treatment and plan check processes were automated to improve efficiency and safety of VMAT CSI. Results: The auto-contours achieved an average DSC of 0.71±0.15, HD95 of 4.81±4.68, and reviewers ranking of 1.22±0.39, indicating "acceptable-as-is" contours. Compared to the manual CSI plans, the auto-plans for both dose regimens achieved significant reductions in body V50% and Dmean for parotids, submandibular, and thyroid glands. The variance in the dosimetric parameters decreased for the auto-plans compared to the manual plans indicating better plan consistency. From the blinded review, the auto-plans were marked as equivalent or superior to the manual-plans 88.3% of the time. The required time for the auto-contouring and planning was consistently between 1-2 hours compared to an estimated 5-6 hours for manual contouring and planning. Automated checks for plan quality were developed and implemented into a separate script, Automated Plan Checker (APC), that is used during the physics second check process. In addition to the routine plan quality checks, e.g., prescription, dose, energy, etc., the APC tool checks the geometry of the CSI beams, the location of the isocenters, and verifies the entered isocenter shifts in the document used by the therapists during treatment delivery. Conclusion: Compared to manual planning, the developed auto-planning process achieved superior sparing of clinically relevant OARs without compromising target coverage. Significant reductions in planning time were observed with the proposed auto-planning process, potentially decreasing the required time between simulation and treatment for these patients. To the authors’ knowledge, this work represents the first automated planning solution for both low-risk (initial plan only) and medium/high risk (initial + boost plans) VMAT CSI. All the developed software has been made open source (https://github.com/esimiele/VMAT-TBI-CSI) so other institutions can adopt the developed auto-planning process into their own practice.
