Room: AAPM ePoster Library
Purpose: Dynamic conformal arc (DCA) offers low plan complexity and high delivery accuracy, and thus is preferred for linac-based single lesion brain stereotactic radiosurgery (SRS). But it often requires a large amount of planner intervention. This work investigates whether RapidPlan knowledge-based planning and the VMAT photon optimization (PO) algorithm could provide similar plan quality with a more clinically efficient and consistent solution.
Methods: A RapidPlan model was trained using 80 SRS plans with lesions ranging from 5-69mm and treated with doses from 12-27Gy. The model was validated on 26 additional patients with lesions from 14-30mm and prescriptions between 14-21Gy. For these 26 lesions, clinical plans were generated using 3-4 non-coplanar DCA arcs. Validation plans were created from the RapidPlan model using VMAT with a preset beam arrangement (3 non-coplanar arcs with gantry at 185°CW175°/179°CCW10°/181°CW350°, collimator at 315°/30°/80°, and couch at 0°/300°/65°, respectively). Optimization was performed using the PO algorithm with high modulation control to minimize the MLC complexity. All plans were normalized to 99.5% PTV coverage. Plan quality was evaluated by modulation complexity score (MCS), Nakamura’s conformity index (NCI), MU, and normal tissue volume encompassed by 12Gy and 6Gy (V12 and V6).
Results: No significant changes were observed for V12 and V6 between DCA and VMAT plans (p>0.1). The mean difference in V12 and V6 is (0.4+/-1.1)cc (-1.4~2.9) and (0.3+/-3.2)cc (-5.7~7.7), respectively. The average MU is (3416+/-601) for DCA and (3577+/-562) for VMAT. The average MCS is (0.55+/-0.04) for DCA and (0.61+/-0.05) for VMAT. The average NCI is (1.23+/-0.07) for DCA and (1.20+/-0.05) for VMAT.
Conclusion: RapidPlan knowledge-based planning and the VMAT PO algorithm with high modulation control offers comparable plan quality as the DCA technique and improves clinical efficiency and consistency for single lesion brain SRS. Further efficiency could be achieved when combined with contouring and planning automation.