Lateral head flexion as a noncoplanar solution for ring gantry stereotactic radiosurgery.

Abstract

Ring gantry radiotherapy devices are often limited to deliver beams in the axial plane, severely limiting beam entrance angles and rendering noncoplanar beam delivery impossible. However, a ring gantry geometry greatly simplifies delivery machines and increases the efficiency of treatment with the potential to decrease the overall costs of radiotherapy. This study explores the use of lateral head flexion in order to increase beam entrance angles and extend the available solid angle space for a ring gantry stereotactic radiosurgery (SRS) application. A 1.5T magnetic resonance imaging scanner was used to scan seven healthy volunteers at three different head positions: a neutral position, a left lateral flexion position and a right lateral flexion position. The lateral flexion scans were co-registered to the neutral head position scan using rigid registration and extracting the rotational transformation. The head pitch, roll, and yaw were computed for each registration to evaluate the natural range of motion for all volunteers. A ring gantry plan geometry was used to generate two sets of single fraction SRS plans (21Gy): one coplanar set for head neutral scans, and a three-arc plan set using the head neutral and lateral head flexion scans. The conformity index (CI), intermediate dose fall-off (R50), low dose spillage (R10), and gradient measure (GM) were used to evaluate both sets of plans. The treatment plans were generated for a ring-gantry linear accelerator (linac) (Varian Halcyon 2.0) as well as radiosurgery linac (Varian Edge) for comparison. The average pitch, yaw, and roll for the lateral head flexion scans were 4.1°±4.7°, 16.9°±3.7°, and 2.5°±4.9° for the right flexion and 4.9°±4.3°, 14.0°±3.7° and 2.8°±5.4° for left flexion. When comparing the head flexion technique with a fully coplanar geometry, the ring gantry plans showed an average improvement in CI of 7.3% (1.46±0.25 vs 1.36±0.28), a decrease of 13% in R50 (5.46±1.14 vs 4.78±1.12), a decrease of 32% in R10 (85.7±20.3 vs 58.2±15.1), and a decrease of 7.8% in GM (0.53±0.05 vs 0.49±0.04). The Edge plans showed an average improvement in CI of 3.0% (1.49±0.26 vs 1.45±0.25), a decrease of 6.8% in R50 (5.19±1.03 vs 4.82±0.83), a decrease of 29% in R10 (84.1±16.3 vs 59.9±12.5), and a decrease of 5.0% in GM (0.50±0.04 vs 0.47±0.03). Lateral head flexion was shown to increase beam entrance angles considerably improving plan conformity and normal tissue sparing in this pilot study of seven sets of plans. Rigid registrations demonstrated each lateral flexion to be analogous to a 15° couch kick. The head flexion technique outlined here was shown to be a feasible solution for SRS treatments being delivered on ring gantry devices.