Statistical Image Analysis in Cone-Beam Computed Tomography

• Activity Number: 59
• Type: Topic Contributed
• Date/Time: Sunday, August 3, 2014 : 4:00 PM to 5:50 PM
• Sponsor: Section on Statistics in Imaging
• Author(s): Susan Doshi*+ and Chris Jennison
• Companies: Cardiff University and University of Bath
• Keywords: CBCT ; Image analysis ; Fiducial markers

Abstract

Cone-beam computed tomography (CBCT) is used to verify the position of the prostate prior to radiotherapy. Gold markers may be implanted, allowing the prostate to be located more accurately. These cause artefacts in the 3D reconstructions. I will discuss the application of statistical image analysis techniques to CBCT data with two purposes: estimating the marker locations with an assessment of uncertainty, and creating reconstructions with fewer artefacts.

I will describe two approaches. Firstly, we model the markers as they appear in the projection images. By combining information from many projections, we accurately estimate the marker locations in 3D space; however, we do not have an accurate assessment of uncertainty. Secondly, we use a model for the markers in 3D space, modelling the tissue separately. In phantom experiments, we obtain accurate estimates of the tissue properties and marker locations (with uncertainty). With patient data, reconstruction artefacts are present in the estimates of tissue properties, but we are able to accurately estimate the marker locations (with uncertainty). Using either approach, we are able to generate reconstructions with fewer artefacts.