Evaluation of factors that affect contrast-detail in digital X-Ray and computed tomography

The central aim of this project was to develop a new methodology of evaluation and optimisation of image quality based on low contrast-detail (LCD) detectability performance of computed tomography (CT). This method is well established in digital radiography however similar tool of image evaluation and quality optimisation for CT images are not available. In comparison with other image evaluation methods in digital radiography, the tool of LCD detectability performance—particularly the automated approach—is a good choice for image quality optimisation. This method helps to determine appropriate exposure factors to provide optimum image quality while maintaining a lower radiation dose to patients. This method is a straightforward and direct way to assess image quality as it provides quantitative evaluations of low contrast and small detail measurements of medical images. The subjectivity of image evaluation methods based on human observers is avoided via automated scoring software that is utilised in automated approach of LCD detectability performance. The trade-offs between perceived image quality, diagnosis efficacy and exposure dose can be determined by LCD detectability measurements. A newly designed LCD CT (CDCT) phantom was manufactured and dedicated software was developed with the cooperation of Artinis Medical Systems (Zetten, The Netherlands) for the new evaluation method of LCD detectability.<br><br>The specifications of the phantom design were optimised based on the standard recommendations of phantom manufacturing and the requirements of the proposed new evaluation methodology. The CT inverse image quality figure (CTIQFinv) was determined as a measure of LCD detectability performance of CT images. An equation was developed and implemented in the software to calculate and objectively measure CTIQFinv values. The new proposed method of LCD detectability performance was validated by evaluating the influences of exposure factors kVp and mAs, slice thicknesses and object location on image quality in terms of CTIQFinv values based on software and radiographers’ scoring results. The results showed that the new evaluation methodology-based CDCT phantom, along with the automated measurement of CTIQFinv value, had generally shown to be consistent with a prior knowledge of image quality in relation to change of mAs, kVp and slice thickness settings. This work showed that the CDCT phantom and the measurement of CTIQFinv values can provide a measure of CT image quality in terms of LCD detectability performance. This method has a promising role for CT image evaluation and optimisation, and has the potential to effectively evaluate the effects of protocol parameters on image quality of different CT scanners and systems. Future changes to the phantom design and/or software is required to overcome some of the current limitations