Click here to

Session: Emerging Imaging, Therapy, and Dosimetry Solutions I [Return to Session]

A Raw Data Correction and Iterative Reconstruction Pipeline for Quantitative CBCT-Guided Radiation Therapy

F Bayat*, M Eldib, M Miften, C Altunbas, University of Colorado School of Medicine, Aurora, CO


SU-J-202-3 (Sunday, 7/10/2022) 4:00 PM - 5:00 PM [Eastern Time (GMT-4)]

Room 202

Purpose: Raw data inconsistency is one of the major barriers in achieving quantitatively accurate CBCT images needed for adaptive radiation therapy. To achieve high quantitative accuracy, a raw data correction pipeline coupled with iterative reconstruction was developed and used in conjunction with a novel scatter rejection hardware in a clinical linac for quantitative CBCT-guided radiation therapy.

Methods: To improve CBCT raw data fidelity, a data correction pipeline was developed with image lag correction, scatter correction and beam hardening correction modules. To reduce the effects of scatter, a 2D antiscatter collimator was developed and integrated into the CBCT system. To reduce image noise amplification after scatter correction, an iterative reconstruction method with total variation regularization was implemented. CBCT imaging experiments were performed using a Varian TrueBeam linac and clinical CBCT protocols. Images generated by our data correction pipeline were compared to TrueBeam iCBCT images to benchmark performance, where scatter was corrected using a 3D object model and images were iteratively reconstructed. CT number accuracy was evaluated both in head and torso sized phantoms, and in anthropomorphic phantoms to study the effect of phantom size and structure on HU accuracy.

Results: After residual scatter, image lag, and beam hardening correction with our approach, phantom size dependent median HU inaccuracy was reduced to 13 HU, whereas HU inaccuracy under the same imaging conditions was 21 HU with iCBCT. HU nonuniformity in water equivalent sections was reduced down to 12HU with our approach, and HU nonuniformity in iCBCT was 54 HU. The use of iterative reconstruction reduced HU standard deviation by a factor of two.

Conclusion: The proposed data correction pipeline provided a substantial improvement in quantitative accuracy of CBCT images. This work indicates that our approach may enable fast CBCT-based dose calculations and CBCT-based treatment plan adaptations in radiation therapy.

Funding Support, Disclosures, and Conflict of Interest: This work was funded in part by grants from NIH/NCI R21CA198462 and R01CA245270.


Quantitative Imaging, Cone-beam CT, Beam Hardening


IM/TH- Cone Beam CT: Quantitative Imaging

Contact Email