Evaluation of a method for correction of scatter radiation in thorax cone beam CT

Purpose: Cone beam CT (CBCT) enables three-dimensional imaging with isotropic resolution. X-ray scatter estimation is a big challenge for quantitative CBCT imaging of thorax: scatter level is significantly higher on cone beam systems compared to collimated fan beam systems. The effects of this scattered radiation are cupping artefacts, streaks, and quantification inaccuracies. The beam stops conventional scatter estimation approach can be used for CBCT but leads to a significant increase in terms of dose and acquisition time. At CEA-LETI has been developed an original scatter management process without supplementary acquisition. Methods and Materials: This Analytical Plus Indexing-based method (API) of scatter correction in CBCT is based on scatter calibration through offline acquisitions with beam stops on lucite plates, combined to an analytical transformation issued from physical equations. This approach has been applied with success in bone densitometry and mammography. To evaluate this method in CBCT, acquisitions from a thorax phantom with and without beam stops have been performed. To compare different scatter correction approaches, Feldkamp algorithm has been applied on rough data corrected from scatter by API and by beam stops approaches. Results: The API method provides results in good agreement with the beam stops array approach, suppressing cupping artefact. Otherwise influence of the scatter correction method on the noise in the reconstructed images has been evaluated. Conclusion: The results indicate that the API method is effective for quantitative CBCT imaging of thorax. Compared to a beam stops array method it needs a lower x-ray dose and shortens acquisition time. (authors)