Computational protocol: Whole-organ and segmental stiffness measured with liver magnetic resonance elastography in healthy adults: significance of the region of interest

Similar protocols

Protocol publication

[…] The study was approved by a local bioethical committee. Sample size was calculated according to Kelley K. [], assuming for 90 % Confidence Interval (CI) for the calculated coefficient of variation (CV) of the measured liver stiffness, a 15 % population CV, a 0.8 desired degree of assurance for achieving a CI no wider than 15 %, and a desired full CI width of 10 %.Inclusion criteria were informed consent for the participation and a normal liver image on ultrasound. Sonographic examination was performed by a radiologist with 13 years of experience in abdominal imaging using Toshiba Xario unit with a convex probe (2–5 MHz).Exclusion criteria were as follows: serum level of alanine transaminase over 40 IU/L, liver pathologies in anamnesis, any known risk factors of liver disease, any previous hospital stay, any previous surgical procedures, a regular diet for at least 6 months, no alcohol abuse (less than 20 g of pure alcohol per day), overweight or obesity, and common contraindications to MRI (uncontrolled claustrophobia, an implanted pacemaker/ICD, any ferromagnetic foreign bodies, etc.).Twenty-four subjects were screened using ultrasound for the potential participation in the study. Four of them were excluded due to detection of hemangiomas (2 subjects) and steatosis (2). Thus, twenty healthy adults aged 24–45 years (mean age 39.1 years) were recruited for this prospective pilot study. Examinations were performed in a supine position after a 4-hour fasting. A 1.5 T scanner (Optima 450w, GE Healthcare, Waukesha, WI) with a 16-channel abdominal phased array coil and an MR elastography system (MR-Touch, GE Healthcare) was used. Before MRE, standard axial T2WI imaging was performed to rule out any morphological pathologies of the liver. None of subjects were excluded at this stage.A passive driver of 19 cm in diameter was placed on the lower chest wall and the upper abdomen, on the right side, with the driver’s central point located at the level of the xiphisternum—over the right lobe of the liver. Acoustic vibrations of 60 Hz generated and transmitted by the system produced shear waves in the epigastrium and the liver. The output power was manually adjusted (30–50 % of the peak capacity of the driver) to ensure sufficient penetration of shear waves within the liver and was within the limits of safety, according to the European Union directive on occupational exposure to whole-body and extremity vibrations.The propagation of the shear waves within the liver was imaged with a two-dimensional GRE MR elastography sequence (TE 50 ms, TR 19.2 ms, flip angle 30°, BW 31.25 kHz, matrix 64 × 64, slice thickness 10 mm, spacing 10 mm) on a breath hold at end-expiration. MRE images (stiffness maps) were fused with axial T2WI breath-hold images of the same slice thickness. No intravenous contrast was given. Images were evaluated by two radiologists with 3 years of experience in liver MR imaging, and all discrepancies were solved by a consensus.The stiffness was measured on 8 axial slices. The middle (5th) slice was placed at the bifurcation of the portal vein and served as a reference slice in each patient. Whole-liver stiffness was measured using consecutive ROIs, which covered slices of the whole liver, excluding the inferior vena cava and the gallbladder. Therefore, stiffness was an average of 8 cross sections of the liver (Fig. ). Cross-sectional stiffness measured according to the method by Lee et al. []. It was calculated in a ROI that was placed on a single largest cross section of the liver and was excluding great hepatic vessels; it accounted for approximately 70 % of the entire cross section of the liver. Right lobe stiffness was measured on the largest single cross section, excluding hepatic great vessels. Additionally, segmental stiffness was measured in 5 small ROIs (50 pix.—approx. 150 mm2) that were placed in the center of the left lobe (segments 2/3), segments 5/6, 7, 8, and the parahilar region (Fig. ). ROIs were placed centrally within the liver parenchyma avoiding large hepatic vessels. Fig. 1Fig. 2The results were expressed as mean stiffness (kPa) ± standard deviation. Normality of data was tested with Shapiro–Wilk test. The relation between values was determined using Pearson correlation coefficient (r). The significance of differences in the stiffness between the tested regions was tested with ANOVA and with t test for a direct comparison between regions. Relative variability of stiffness measurement using a particular method (within-method variability) was calculated separately for each method as a CV that was a ratio between the mean value across examined subjects and the standard deviation of the mean. Within-subject relative variability of measurements using small segmental ROIs (left lobe, hilum, segment 5/6, segment 7, and segment 8) was calculated as CV, i.e., it was a ratio between the mean value of stiffness across segments in a single subject and the standard deviation of the mean. Agreement between results of whole-liver measurement and those of particular analyzed regions was tested using intraclass correlation coefficient (ICC, two-way model, absolute agreement for average measures). A P-value of <0.05 was considered significant. Statistical analyses were performed using Statistica 10 (StatSoft Inc., Tulsa, OK) and MedCalc Statistical Software version 13.3 (MedCalc Software bvba, Ostend, Belgium) […]

Pipeline specifications