Click here to

Session: Novel Technologies [Return to Session]

Demonstration of a Novel Metal Artifact Reduction Technique with KV-MV CBCT

M Jacobson1*, T Harris1, M Lehmann2, P Huber2, M Myronakis1, I Ozoemelam1, R Fueglistaller2, D Morf2, R Berbeco1, (1) Brigham and Women's Hospital, Dana-Farber Cancer Institute, and Harvard Medical School, Boston, MA (2) Varian Medical Systems, Baden-Dattwil, Switzerland

Presentations

MO-C930-IePD-F6-1 (Monday, 7/11/2022) 9:30 AM - 10:00 AM [Eastern Time (GMT-4)]

Exhibit Hall | Forum 6

Purpose: Cone beam CTs at typical kV energies exhibit metal implant artifacts that may obscure anatomy in IGRT. We propose an imaging technique to suppress artifacts by combining a noise-reducing, poly-energetic correction algorithm with MV-CBCT from a high DQE, 4-layer MV-imager. Previous simulation tests of the method were promising. In this submission, it is tested further in a physical, anthropomorphic phantom with bi-lateral metal hip inserts.

Methods: In the proposed method, the role of the MV-CBCT is to: (1) provide additional data for the final reconstruction (2) obtain a prior 3D model of the implants. The prior model is used by the poly-energetic correction algorithm to eliminate the contribution of metal from the projections. Long, photon-starved kV x-ray paths through metal are also pre-detected using the prior model and over-painted with MV measurements. The corrected MV and kV data are FDK-reconstructed and consolidated into a single image via weighted averaging. The phantom was a customization of the CIRS-801-P-F with 2.5” diameter aluminum spheres in each femoral head, emulating hip replacements. Artifact reduction was measured in terms of structural similarity (SSIM) to a metal-free diagnostic CT in the region of the bladder. Contrast-to-noise (CNR) between adipose and muscle near the femoral head was also quantified. Our technique was compared both to standard clinical kV-CBCT (140kVp) and single-layer MV-CBCT (2.5MV), all at CTDIw=40mGy.

Results: The kV-MV method maintained comparable CNR to standard kV-CBCT, while significantly reducing metal artifacts. This was evident both visually and in terms of SSIM, which improved 25%. The method also demonstrated a 60% improvement in CNR relative to single-layer MV-CBCT.

Conclusion: Supplementing kV-imager projections with measurements from a high efficiency, multi-layer MV-imager promises to improve image guidance in patients with metal implants at normal set-up CBCT doses.

Funding Support, Disclosures, and Conflict of Interest: Funding Support: 1. NIH/NCI R01CA188446 2. Varian Medical Systems, Inc.

Keywords

Image-guided Therapy, Reconstruction, Megavoltage Imaging

Taxonomy

IM- Cone Beam CT: Image Reconstruction

Contact Email

Share: