Magnetom verio

Prior to prostate biopsies, all men underwent MRI using a 3 Tesla Siemens Magnetom Verio or a 1.5 Tesla Philips Ambition in the event that a 3 Tesla was contraindicated. The European Society of Urogenital Radiology MRI protocol was used containing axial fast spin-echo T1-weighted images of the pelvis [15] (link). T2-weighted fast recovery turbo spin-echo images of the prostate were acquired in the axial, sagittal, and coronal planes (slice thickness 3 mm). Axial diffusion-weighted imaging was performed using a spin-echo echo-planar imaging pulse sequence with slice thickness 5 mm (b values: b-0, b-1000 or b-1400, b-2000 s/mm²); apparent diffusion coefficient maps were calculated automatically.
Images were prospectively interpreted by seven experienced radiologists, all of whom have >3 yr of experience in reading prostate MRI. The standardized five-point PI-RADS classification was used, in concordance with the PI-RADS version 2.1 guidelines [16] (link). If multiple unilateral MRI-positive lesions were observed, the lesion with the highest PI-RADS assessment category was designated as the index lesion and used for analysis. If multiple MRI-positive lesions had the highest PI-RADS assessment category, the largest lesion was designated as the index lesion and used for analysis.

All enrolled subjects underwent MRI of the head prior to surgery, including several different acquisition protocols [axial T1WI and T2WI, axial, coronal, and sagittal contrast-enhanced T1WI (CE-T1WI)]. The contrast agent, dimeglumine gadopentetate, was injected at a dose of.2 ml/kg, following pre-contrast T1 scanning. MRI images were obtained from four different MRI scanners with 3 T (GE Discovery MR 750, n = 59; Siemens MAGNETOM Trio TimSystem, n = 43; Siemens MAGNETOM Verio, n = 22; Philips Ingenia, n = 9), and a 1.5 T scanner (GE Medical System Genesis Signa, n = 35). Supplementary Material 2 shows the type of the contrast medium and the parameters used for axial CE-T1WI for five MRI modalities.
Whole tumors, based on preoperative axial CE-T1WI images, were identified as the region of interest (ROI). The manual delineation of each ROI was conducted by a neuroradiologist with 5 years of experience, using 3D-Slicer software (version 4.10.2 r28257, National Institutes of Health). Prior to segmentation, we applied three steps to standardize different MRI images: N4ITK bias correction, resampling with resampled voxel sizes of 1, 1, and 1, and Laplacian of Gaussian (LOG) with LOG kernel sizes being 1.5, 2, and 2.5 by 3D-Slicer.