1. Data collection
After appropriate authorization, a six-week prospective study conducted at a medium-sized hospital in the centre region of Portugal collected information on multiphase hepatic-CT examinations of adult patients. The patients’ confidentiality, as well as of all actors involved in health care, was ensured. The work took place under the initiative of continuous improvement of professional practice in the medical imaging department.
A questionnaire was used to collect pertinent ICM injection protocol data and anthropometric measurements (body weight and body height). Also, relevant DICOM metadata were collected and analysed. Most relevant collected data are resumed in Table 1.
Table 1: Resume of collected data
References: School of Health Sciences, University of Aveiro (ESSUA) - Aveiro/PT
2. CT examination protocol
Hepatic CT studies used institution technical protocol. Most relevant parameters were resumed in Table 2.
Table 2: Resume of the technical protocol
References: School of Health Sciences, University of Aveiro (ESSUA) - Aveiro/PT
3. Structures enhancement
Contrast enhancement was evaluated in CT images for three structures – Aorta artery, portal vein and liver parenchyma. The Hounsfield Units (HU) evaluation was performed in axial images at the level of the celiac trunk artery using circular ROI’s (100 mm2) placed in the aorta (centre of the aorta), portal vein (hepatic hilar zone) and hepatic parenchyma (segment III with minimal vascular or biliary structures contribution). Measurements were performed to non-enhanced and contrast-enhanced phases (arterial and portal acquisitions) using institution PACS viewer ( Fig. 1).
HU average of each measurement was collected, and structures enhancement calculated to each contrast-enhanced phase by the difference between enhanced and non-enhanced structures HU average.
Fig. 1: Example of the ROI's arrangement to perform structures HU measurement.
References: School of Health Sciences, University of Aveiro (ESSUA) - Aveiro/PT
3. Statistical analysis
Collected data were evaluated using IBM© SPSS© Statistics 25. Scale variables were summarized as the average value ± standard deviation (SD) and categorical variables were calculated as the count (percentage). Pearson correlations used to bivariate correlations analysis. Multivariate analysis of variance was performed and when statistical significance was observed, Tests of Between-Subjects Effects used.
Statistical significance was considered when p<0.05.