Objectives:

Stereotactic spine radiosurgery (SRS) requires submillimeter accuracy in the treatment planning and delivery process to achieve optimal treatment efficacy and minimize unacceptable toxicities. The purpose of this study was to investigate user detectability of known intrafraction motion in a phantom, using triggered kilovoltage (kV) imaging with and without spinal fixation hardware.

Methods:

An anthropomorphic phantom with spinal inserts (native spine, titanium implants, carbon fiber reinforced PEEK implants) was planned for SRS on a C-arm linac. After the phantom was CBCT-localized, translational (1–1.5 mm) and rotational (1°–2° pitch) errors were applied. Triggered kVs were acquired every 20° with contours of the vertebral body (VB) or hardware (HW) plus 1 mm. Images were reviewed by 8 clinicians blinded to the errors. Reviewers identified when intervention/relocalization was required.

Results:

Users intervened in the absence of shifts in 7 of 32 (22%) cases. Detectability of errors at 1 mm was 59% (vert), 78% (lat), 84% (lng); at 1.5 mm was 84% (vert), 88% (lat), 100% (lng). For 1° and 2° pitch, detectability was 41% and 53%, respectively. Isolating VB from HW, detectability of 1 mm translational errors was 71% (VB) and 75% (HW), and 1.5 mm translational errors was 79% (VB) and 94% (HW). Conversely for pitch, detectability of 1° was 75% (VB) and 29% (HW) and 2° was 63% (VB) and 50% (HW).

Conclusion(s):

Based on this phantom study, triggered kV imaging proved useful as a tool for monitoring intrafraction motion. As the magnitude of error increased, detectability increased. The majority of users were able to identify translational errors ≥ 1 mm. Tracking HW increased detectability of translational errors ≥ 1.5 mm, compared to VB alone. Meanwhile, VB tracking was superior to HW in detecting pitch errors ≥ 1°. Consideration of HW plus VB tracking may increase detectability of translations and rotations for patients with spinal fixation hardware.