Main Session
Sep 29
PQA 05 - Physics

2983 - Triggered Spectral kV Imaging for Real-Time Intrafraction Bone Tracking Using a Novel Dual-Layer kV Imager

12:30pm - 01:45pm ET
Poster Hall - Exhibit Hall A
Screen: 17
POSTER

Presenter(s)

Francois De Kermenguy, PhD Headshot
Francois De Kermenguy, PhD - Brigham and Women's Hospital, Cambridge, MA

F. De Kermenguy1, M. Jacobson1, Y. H. Hu1, N. Lowther1, R. Etemadpour1, T. C. Harris1, G. C. Sharp2, B. Vecchione1, M. Myronakis1, M. Lehmann3, V. Birrer3, R. Fueglistaller3, P. Corral Arroyo3, R. Bruegger3, R. I. Berbeco1, and D. Ferguson1; 1Department of Radiation Oncology, Brigham and Women’s Hospital, Dana-Farber Cancer Institute, Harvard Medical School, Boston, MA, 2Department of Radiation Oncology, Massachusetts General Hospital, Boston, MA, 3Varian Medical Systems, Baden-Dattwil, Switzerland

Purpose/Objective(s):

Intrafraction motion during ablative bone SBRT can compromise target coverage and organ-at-risk sparing due to steep dose gradients and millimeter-scale PTV margins. Conventional triggered kV imaging may be limited by suboptimal image quality, reducing alignment confidence during beam delivery. A novel dual-layer spectral kV imager enables rapid, energy-resolved triggered acquisition with material decomposition to generate bone-only projections. We evaluated the geometric accuracy and registration performance of triggered spectral kV imaging for direct intrafraction bone motion monitoring.

Materials/Methods:

A phantom study was performed to benchmark bone-tracking performance using triggered images acquired with the dual-layer imager (DLI). Following initial setup of an anthropomorphic pelvis phantom, triggered projections were acquired at 10-degree gantry intervals. From each DLI acquisition, two image sets were generated: (1) a top-layer image equivalent to a conventional clinical kV panel and (2) a bone-only image derived from combined top- and bottom-layer spectral decomposition using a bespoke algorithm developed for the DLI. Each image set was registered to digitally reconstructed radiographs and residual translational and rotational differences relative to known applied shifts were quantified.

Results:

Spectral bone-only imaging enabled high-confidence registration from single triggered projections. In preliminary phantom evaluation, bone-only images demonstrated reduced the mean translational offset from 8.0 mm to 2.8 mm and rotational offset from 1.5° to 0.7° compared with conventional triggered kV images.

Conclusion:

Triggered dual-layer spectral kV imaging provided high-contrast visualization of treated bony anatomy during beam delivery and improved positional accuracy relative to standard triggered kV imaging in phantom testing. These findings support dual-layer spectral imaging as a promising approach for intrafraction geometric verification in high-precision bone SBRT.

X Registration Offset

Y Registration Offset

Pitch Registration Offset

Phantom

µ ± s [mm]

µ ± s [mm]

µ ± s [deg]

Top-layer kV (clinical equivalent)

-2.6 ± 32

-8.4 ± 48

1.5 ± 3.5

Bone Decomposed

2.5 ± 8.1

-1.4 ± 8.0

0.7 ± 1.4