4πVMAT: A Novel Method to Efficiently Deliver Non-Coplanar Treatment

Abstract

Existing Volumetric Modulated Arc Therapy (VMAT) optimization using one or more arcs, each being coplanar by itself, is a highly efficient way to deliver intensity modulated radiation therapy (IMRT). Static beam IMRT utilizing optimized non-coplanar beam orientation has resulted in superior dosimetry to VMAT but is substantially less efficient. In this study, we developed a novel optimization framework for non-coplanar VMAT, incorporating simultaneous couch and gantry rotation during the arc motion, to maximally utilize the 4π solid angle (4πVMAT) with high efficiency. The 4πVMAT optimization framework includes a Direct Aperture Optimization (DAO) module and a beam trajectory selection (BTS) module. DAO was achieved by adding to the dose fidelity objective a total variation term for regulating the fluence smoothness, a single segment term for forming deliverable fluence map, and a group sparsity term for selecting beam angles. The BTS was formulated as a travelling salesman problem on a hierarchical graph, where the edge and node costs were determined by the maximal gantry rotation speed and the estimated fluence map at the current iteration, respectively. Continuous gantry/couch angle trajectories were selected on the graph using the Dijkstra's algorithm. Candidate beams in the couch-gantry-patient collision space, evaluated on actual machine geometry and a human subject 3D surface, were excluded in both the DAO and BTS modules. The optimization framework was solved by alternating between the DAO module and the BTS. Efficacy of 4πVMAT using one full-arc or two full arcs was tested on a brain, a lung and a prostate cancer patient. The plan was compared against a coplanar VMAT (2πVMAT) plan using more arcs with collimator rotation. The 4πVMAT consisted of two efficient arcs, each taking 160 seconds in estimate to deliver. In the first arc, the couch rotated in either the clockwise or counterclockwise direction, and then reversed the rotation in the second arc. The gantry rotations were coordinated with the couch rotation for optimal dose distribution while avoiding couch-gantry-patient collision. With the same target coverage, 4πVMAT remarkably reduced the max and mean organs-at-risk (OARs) doses compared to 2πVMAT, as shown in Table 1. The dose compactness measured using R50 was reduced by 19.7%. By optimizing feasible and efficient non-coplanar arc trajectories, the novel 4πVMAT optimization framework simultaneously achieves the desirable dosimetric quality of 4π therapy and the delivery efficiency of coplanar 2πVMAT.Abstract TU_35_3241: Table 1OAR dose: 2πVMAT - 4πVMAT (% of the prescription dose)patientDmaxDmeanLargestAverageLargestAveragebrain12.847.529.324.68lung32.6012.7810.583.62prostate10.502.764.300.35OAR mean and maximum dose differences between πVMAT and 2πVMAT for all patients. “Largest” and “Average” represent the largest and average values among all OARs. Open table in a new tab