A beam-hardening correction using dual-energy computed tomography

Abstract

The problems associated with beam hardening in X-ray transmission computed tomography (CT) have been described by Brooks and DiChiro (1976). These include the introduction of non-linear artefacts in CT images with corresponding errors of as high as 10% in CT measurements of the linear attenuation coefficient, thus severely limiting the use of CT in quantitative applications. Many of the techniques proposed to eliminate or reduce this artefact have been summarised by Stonestrom et al. (1981). A practical technique for correcting this artefact which has been implemented on an EMI CT5005 whole-body scanner is presented. It is proposed that by taking dual-energy measurements and obtaining two polyenergetic ray sums (p₁ and p₂) it is possible to obtain a better estimate of the monoenergetic ray sum m than can be obtained from a single polyenergetic measurement of the ray sum p (Herman 1979). An expression of the form given can be used to estimate m. By using dual-energy data collected in a single scan the correction can be carried out quickly with no additional patient dose.