2448 - Evaluation of a Real-Time Positional Verification System for High-Dose Dorsal Spinal Nerve Radiosurgery

S. Xing¹, L. Zhang¹, L. Zhu², K. Macek², D. M. Lovelock³, L. I. Cervino¹, Y. Yamada⁴, and S. B. Lim¹; ¹Department of Medical Physics, Memorial Sloan Kettering Cancer Center, New York, NY, ²Varian Medical Systems, Palo Alto, CA, ³Department of Radiation Oncology, Mount Sinai Hospital, New York, NY, ⁴Department of Radiation Oncology, Memorial Sloan Kettering Cancer Center, New York, NY

Purpose/Objective(s): Intrafraction motion management is crucial for accurate and safe delivery during spine SBRT. Currently, quantitative 3D motion tracking techniques using the on-board imager on a standard equipped linac are limited. In this pilot study, we evaluate the efficacy of a pre-clinical real-time motion tracking system (RTTS) to quantitatively monitor the vertebral body during high dose (90 Gy) dorsal spinal nerve ablation SRS for patients with intractable back pain. Materials/

Methods: This study is part of an IRB-approved phase-I protocol to deliver 90 Gy in a single fraction to ablate the involved dorsal nerve root ganglion. RTTS was evaluated using an anthropomorphic spine phantom placed on a programable one-dimensional motion platform at 30^o from the longitudinal axis of the couch to generate a two-dimensional motion. The phantom was shifted 6 cm in 1 cm increments during the arc delivery, simulating a total lateral and longitudinal motion of 3 cm and 5.2 cm respectively. Fluoroscopic kilovoltage images were continuously acquired at 2 frames/s. A set of digitally reconstructed radiographs of the target vertebra, which served as a template for triangulating the target positions in three dimensions, was generated from the planning CT at 1° gantry angle resolution. After each shift, clinical standards, CBCT and stereoscopic kV-images (SKV), were acquired to compare with the RTTS results. Tracking accuracy was further evaluated in a pilot study that included six spine SBRT patients. CBCT projections were used to triangulate the target positions instead of fluoroscopic images to avoid deviation from the current clinical protocol. The average positional offset and standard deviation (SD) were evaluated against the measured clinical shifts after CBCT in all three directions.
Results: In our phantom study, the mean offset measured with RTTS was within 0.2±0.1 mm of the simulated shifts, with a maximum deviation of 0.3mm and 0.4mm in the lateral and longitudinal directions respectively. The positional offsets determined by RTTS agreed with the CBCT-based shifts, within 0.1±0.1 mm laterally and 0.2±0.2 mm longitudinally. The tracking accuracy of RTTS is also similar to that of the SKV system, where an average difference of 0.1±0.1mm from the simulated shifts were observed in both directions. Six SBRT patients with two C-spine, T-spine and L-spine lesions were included in our pilot study. The average lateral, longitudinal and vertical shifts measured by RTTS were within 0.2±0.1 mm, 0.2±0.1 mm and 0.1±0.1 mm difference from the CBCT-based shifts, with a maximum deviation of 0.3mm in the lateral/longitudinal directions and 0.2mm in the vertical direction.
Conclusion: Our preliminary results demonstrate that RTTS quantitatively monitors positional change of the vertebral body during beam delivery. The target position measured by RTTS is consistent with the CBCT-and SKV-based positions, suggesting RTTS as a potential alternative for intrafraction motion monitoring during spine radiosurgery.