Treatment Plan Comparison Between Self-Shielding Gyroscopic Radiosurgery and Robotic Radiosurgery

Stereotactic radiosurgery with established systems like the Gamma Knife and CyberKnife (Accuray Inc., Madison, WI, USA) is a well-characterized treatment concept. The novel ZAP-X^® platform (ZAP Surgical Systems Inc., San Carlos, CA, USA) for vault-free, self-shielding gyroscopic radiosurgery (GRS) promises high plan quality due to advantageous beam properties. However, the clinically usable workspace in GRS is reduced due to potential collisions with a spacious headrest. A novel "conformal" headrest was introduced to GRS in December 2023 to remedy this, using narrower masks to minimize collision zones and maximize the usable solid angle. This study analyzes the GRS plan quality for 30 simple and complex cases, comparing GRS plans with the old and new headrests to robotic radiosurgery (RRS) as an established reference platform. The GRS system consists of a 3 MV linear accelerator mounted on coupled gimbals for non-coplanar beam delivery, a collimator wheel for circular beam shaping, and a kV image guidance system. The RRS system is a full-body treatment platform with a 6 MV linear accelerator on a robotic arm for non-coplanar, non-isocentric beam delivery. A total of 30 clinical single-fraction plans treated with the GRS system prior to the headrest update is selected. Clinical GRS treatment plans are created by manually placing isocenters within the target volume and using an inverse optimization algorithm. GRS plans are reoptimized using the new software and headrest (further referred to as GRS*) for comparison. RRS plans are generated using circular apertures and the VOLO™ optimization technique. Treatment plans from the GRS, GRS*, and RRS platforms are compared with respect to quality metrics, number of beams, total monitor units (MU), and expected treatment time. The updated GRS* plans show a significantly better new conformity index (nCI) and gradient index (GI) than the clinical GRS plans. The volume of the brainstem receiving 8 Gy or more is significantly reduced with the GRS* platform. The number of beams, total MU, and expected treatment time increase significantly with the new GRS* treatment planning system. Compared to GRS* plans, the nCI of RRS plans is better, but the GI is worse. The total number of beams and MU were significantly lower with the RRS platform, while the expected treatment times were equivalent. The introduction of the new headrest design in the GRS* system has led to a notable improvement in the treatment plans of GRS. As a trade-off for the overall improvement in dosimetric quality, the number of beams and the expected treatment time increase. RRS and GRS* systems now exhibit equivalent plan quality, with a trend of the GRS* toward sharper dose gradients but lower conformity, attributed to the specialized delivery design.