Discovery ct750

CT examinations were performed using 2 scanners with intravenous contrast media. The volume of the contrast media was determined by multiplying the body weight (in kilograms) by 2 to a maximum of 100 mL. The concentration of the iodinated contrast media was 350 mg/mL with an injection rate of 2 mL/s. The scanning parameters of the 2 scanners were as follows: (1) The 128-channel CT scanner (Discovery CT750, GE Healthcare, US): field of view, 25 cm; matrix, 512 × 512; tube voltage, 120 kVp; tube current, 200–400 mA; reconstructed thickness, 5 mm; (2) The 128-channel CT scanners (Somatom Definition or Definition AS + , Siemens Healthcare, US): field of view, 35 cm; matrix, 512 × 512; tube voltage, 80–120 kVp; tube current, 248–578 mA; reconstructed thickness, 5 mm. Finally, the arterial phase images of the CT examination were anonymized and assigned a research code for the assessment of general imaging features and the extraction of texture features.

Thirty-eight patients underwent CT from the diaphragm to the iliac crest. Unenhanced CT and dual-phase contrast enhancement spectral spiral CT scans were performed using a spectral CT scanner (Discovery CT 750, GE Healthcare, Waukesha, WI, USA). Contrast medium containing 350 mg of iodine per ml (Omnipaque™; GE Healthcare, Cork, Ireland) was injected at a flow rate of 3 ml/s via the elbow vein. The dose of the contrast medium was calculated as 1.5 ml per kg body weight. Scanning was triggered when the CT value of the aortic arch reached 100 HU. Contrast-enhanced CT images were acquired with a scanning delay of 30 s in the arterial phase and 70 s in the portal venous phase after the start of intravenous contrast medium injection [11 (link)].

This is a secondary use of existing clinical CT scans. Based on the REB approval, 200 patient CT images were obtained: The first non-contrast whole-head CT scans of patients who were newly identified with a primary diagnosis of acute hemorrhagic stroke between January 1st 2011 and January 1st 2018 across Fraser Health, British Columbia. The samples were retrieved using non-probability sampling based on accessibility at the time of data preparation^{44 (link)}. Patients with history of hematomas or with hemorrhage due to tissue plasminogen activator administration for treatment of ischemic stroke were excluded. Out of the retrieved data, we randomly selected 55 CT image sets (52.7% male; mean age = 64.1 ± 14.9 years) applying a random data generator. The images were then annotated and pixel-wise labeled for use in the study (i.e., there were 780,894,208 labeled voxels per CT scan). The CT dataset contained images that were acquired using several models of CT scanners, including GE Discovery CT750, GE LightSpeed VCT, and Siemens SOMATOM Definition Flash.

All patients underwent abdominal and pelvic CT examinations (Discovery CT 750, GE Medical Systems, USA; or Aquilion One, Toshiba, Japan). The imaging parameters were as follows: slice thickness, 0.625 mm; pitch, 0.984; gantry rotation time, 0.5 s; tube voltage, 120 kV; and automatic tube current modulation, 100–200 mA. A 5-mm interval was used for CT image reconstruction.
All patients' CT Digital Imaging and Communications in Medicine (DICOM) images were transferred to a dedicated image analysis workstation equipped with an open source software (Fire Voxel, New York University, NY, USA), and then processed by an abdominal radiologist in a double-blinded manner. First, renal parenchyma of two sides were drawn on the superior and inferior layers of the axial kidney image. The software automatically filled the entire kidney according to the Hounsfield unit (HU) threshold to obtain a volume of interest (VOI), including the renal cortex and medulla. Second, the VOI was magnified, and the edges were manually modified to ensure that all functional renal parenchyma was contained while hydronephrosis, calculi, and cysts were avoided (Figure 1). Third, RCMs (morphological and CTTA parameters) were automatically calculated based on the final VOI, including RPV, HU, parenchymal voxel, skewness, kurtosis, and entropy.