Study on the effects of patient miscentring on radiation dose and image quality in CT imaging

• Main Author: Teo, Siti Humairah Mohd Razlan
• Format: Thesis
• Language: English
• Published: 2020
• Subjects: R Medicine
• Online Access: http://eprints.usm.my/54238/1/Siti%20Humairah-24%20pages.pdf; http://eprints.usm.my/54238/

Incorrect setting of patient positioning on the treatment table always happened in diagnostic imaging which leads to misinterpretation of data and degrades the image quality. This study focuses to evaluate the effects of patient miscentring on patient dose and image quality in CT imaging. The CT scan parameter, patient positioning information was taken from the head and thorax-abdomen-pelvic protocols acquired in 128-slice Siemens SOMATOM Definition AS + CT system. This study started with the retrospective survey on patient miscentring shifting data by using data registration (DoseWatch, version 1.2, GE Healthcare) to collect information about patient isocenter shifting. The range of miscentring for minimum and maximum had been obtained and will be used as guidelines in phantom studies. CTDI phantom and water phantom were used in dose measurement and image quality assessment respectively. CTDIw is used to calculate the dose across FOV of the phantom. In this study, isocentre shifting of both axes (positive or negative axis) can cause an increase or a decrease of CTDIw across the FOV. This research showed that the shift x- and y- axes and the phantom dose (CTDIw) have weak correlation which is 0.1796 and -0.2127 respectively for the 16cm CTDI phantom. While for 32 cm, the correlation is -0.3555 (weak) and 0.5582 (strong) between both variables. In image quality assessment, SNR is observed. Miscentring of both axes also associated with the quality of the image produced. There is no correlation between x-axis shifting and SNR (r = -0.0267). However, the correlation between mean HU and mean SD with the shift x-axis is weak. There is a strong positive association between the y-axis shifting and the SNR (r = 0.8156). The correlation between mean HU and mean SD also good with the y-axis.