Research Presentation Session

RPS 601c - Diffuse liver disease

Lectures

1
RPS 601c - Non-invasive assessment of liver cirrhosis with multiphasic dual energy CT using iodine maps: correlation with model of end-stage liver disease score

RPS 601c - Non-invasive assessment of liver cirrhosis with multiphasic dual energy CT using iodine maps: correlation with model of end-stage liver disease score

04:56D. Mastrodicasa, Stanford / US

Purpose:

To determine whether multiphasic dual ­energy (DE) CT iodine quantitation correlates with severity of chronic liver disease.

Methods and materials:

28 cirrhotic and 22 non-cirrhotic patients who underwent a multiphasic liver protocol DECT and had MELD scores available were included. All three phases (arterial, portal venous (PVP), and delayed) were performed in DE mode. ROIs were placed in the caudate, left and right hepatic lobe, in the aorta, common hepatic artery (CHA), and portal vein (PV) to measure iodine (I) values (mg I/ml). (λ) were calculated as follows: Idelayed­-arterial/ time and Idelayed­-PVP/ time. λ were correlated with MELD scores and the area under the curve of the receiver operating characteristic (AUROC) was calculated to distinguish cirrhotics and non­cirrhotics.

Results:

Cirrhotic and non-cirrhotic patients had significantly different λdelayed-PVP for caudate (λ = 1.350 vs. 2.350, P<.0001), left (λ = 1.383 vs. 2.200, P<.004), and right (λ = 1.063 vs. 1.913 , P<.0001) lobe. λdelayed-arterial were significantly different for CHA (λ = 2.450 vs. 11.250, P<.023) and PV (λ = 2.750 vs. 3.750, P<.013). A significant correlation was found between MELD scores and λdelayed-PVP of caudate, left and right lobes (rho=0.340, P=.034; rho=0.393, P=.005; rho=0.368, P=.034, respectively). AUROC for caudate, left, and right lobe λ delayed-PVP in differentiating cirrhotics from non-cirrhotics were 0.794, 0.739, and 0.908, respectively.

Conclusion:

Multiphasic DECT iodine quantitation over time is significantly different between cirrhotics and non-cirrhotics, correlates with MELD score, and it could serve as a non-invasive measure of cirrhosis and disease severity with high diagnostic accuracy.

Limitations:

Retrospective study.

Ethics committee approval

IRB and HIPAA-compliant study.

Funding:

Institutional grant by Siemens Healthineers.

2
RPS 601c - Quantification of local 3D texture maps for detection of fibrosis in the liver

RPS 601c - Quantification of local 3D texture maps for detection of fibrosis in the liver

08:33W. Hamilton, Toronto / CA

Purpose:

To develop a method for detection and staging of liver fibrosis from low-dose single-energy CT by quantifying spatial distributions of local 3D texture maps.

Methods and materials:

33 patients who underwent a four-phase liver CT were categorised as having healthy (F0 and F1) or fibrotic livers (F3 and F4). Spherical ROIs, radius 20mm, were extracted from each liver. The ROIs were selected from regions close to the liver surface, avoiding major vessels and lesions. To simplify patterns in tissue structure, a Gaussian filter was applied to each ROI before binning gray levels using k-means clustering with k=20. Local 3D texture maps were produced from each binned ROI using the PyRadiomics library. The maps were computed voxel-wise for spherical subsets of 5-voxel radius. As a proof of concept, features were extracted from local 3D variance maps only. To quantify the spatial distributions of the features as they varied across each map, entropy, variance, gray level run length, and size zone matrices were computed from each local variance map.

Results:

The mean entropy, variance, run length non-uniformity, and zone size variance for healthy (SD)/fibrotic (SD) livers are 0.200(0.247)/0.392(0.440), 0.071(0.087)/0.121(0.160), 1700.68(1410.26)/548.071(186.141), 0.267(0.366)/0.027(0.029), respectively. Kruskal-Wallis tests demonstrate significant differences between distributions of healthy and fibrotic livers for run length non-uniformity (U=7.456, p=0.006), however distributions of variance (U=0.938, p=0.333), entropy (U=0.926, p=0.336), and zone size variance (U=2.370, p=0.123) did not demonstrate significant differences.

Conclusion:

Our novel approach shows promise for quantification of fibrosis in the liver and circumvents the need for objective comparison of parameter-sensitive texture features, which suffer from poor interpatient standardisation. Run length non-uniformity may represent the leading feature for separating healthy from fibrotic livers.

Limitations:

Number of livers analysed.

Ethics committee approval

Research ethic board approval.

Funding:

MSH-UHN AMO Innovation Fund.

3
RPS 601c - Quantitative measurement of hepatic fibrosis on gadoxetic acid–enhanced magnetic resonance imaging in patients with chronic liver disease: multicentre study

RPS 601c - Quantitative measurement of hepatic fibrosis on gadoxetic acid–enhanced magnetic resonance imaging in patients with chronic liver disease: multicentre study

06:23Y. Kim, Iksan-Si / KR

Purpose:

The aims of this study were to compare coefficient of variation (CV) in MR hepatobiliary image and serum biomarkers such as aspartate aminotransferase to platelet ratio index (APRI) and fibrosis-4 index (FIB-4) values according to the histopathologic fibrosis score, to identify the diagnostic performance of CV map in diagnosing hepatic fibrosis.

Methods and materials:

This study was a prospective multicenter study which included 71 patients who underwent liver MRI using gadolinium EOB-DTPA and liver biopsy or surgery. Patients were divided into 4 groups according to the liver fibrosis score; Group 1 (F0, 1), Group 2 (F2), Group 3 (F3), Group 4 (F4). To quantitatively measure the hepatic fibrosis, a hepatobiliary image was analysed to identify inhomogeneous signal intensities calculated from CV map in the liver parenchyma. We also evaluated the comparison study of among CV, APRI, and FIB-4. The diagnostic performance of the CV map for fibrosis and cirrhosis was evaluated using ROC curve.

Results:

Mean CV values in each group were Group 1 (n=18): 3.50±0.36, Group 2 (n=8): 4.09±0.50, Group 3 (n=21): 4.45±0.70, and Group 4 (n=30): 5.49±1.12, respectively (p<0.001). Mean FIB-4 values were Group 1: 1.19±1.20, Group 2: 3.36±2.74, Group 3: 3.38±3.12, and Group 4: 4.86±3.39, respectively (p=0.001). Area under curves of CV values on ROC analysis were 0.930 for significant fibrosis, 0.874 for cirrhosis (p<0.001, respectively).

Conclusion:

CV value based on hepatobiliary MR image provides accurate discrimination of hepatic fibrosis with reliable measurements and demonstrates high diagnostic performance.

Limitations:

Small study populations. Various MR vendors/protocols.

Ethics committee approval

Approved by the institutional review board.

Funding:

Supported by the grants of the National Research Foundation of Korea (NRF) (2016M3A9A7918501) and Bayer Korea. Ltd.

4
RPS 601c - Functional liver imaging score derived from gadoxetic acid-enhanced MRI predicts outcomes in patients with chronic liver disease

RPS 601c - Functional liver imaging score derived from gadoxetic acid-enhanced MRI predicts outcomes in patients with chronic liver disease

04:27L. Beer, Vienna / AT

Purpose:

The aim of the study was to investigate the accuracy of the gadoxetic acid-enhanced MRI based functional liver imaging score (FLIS) in predicting hepatic decompensation and transplant-free survival (TFS) in patients with chronic liver disease (CLD).

Methods and materials:

265 with CLD met the inclusion criteria of this retrospective study. Those patients were retrospectively assigned a FLIS based on the sum of three hepatobiliary-phase features, each scored on an ordinal 0-2 scale: enhancement, excretion, and portal-vein-sign. FLIS scores of 0-3 and 4-6 indicate impaired and preserved liver function, respectively. Patients were further stratified into three groups: non-advanced CLD (non-ACLD); compensated advanced CLD (cACLD); and decompensated advanced CLD (dACLD). The predictive value of FLIS for first and/or further hepatic decompensation and for TFS was investigated using Kaplan-Meier analysis, log-rank tests, and Cox-regression.

Results:

Intraobserver (correlation-coefficient κ=0.983; 95% confidence-interval [CI]:0.971-0.991) and interobserver (κ=0.931; 95%CI: 0.898-0.954) agreement for FLIS was excellent. In patients with cACLD, the FLIS was independently predictive of a first hepatic decompensation (adjusted-hazard-ratio, [aHR]: 3.72,95%CI: 1.1-12.64; P=.04), but not for further hepatic decompensations in patients with dACLD (aHR:1.43, 95%CI:0.92-1.94, P=.17). The FLIS was an independent risk factor for mortality in both patients with cACLD (aHR:7.44, 95%CI:2.74-20.17, P<.001) and those with dACLD (aHR: 3.84, 95%CI:1.16-9.48; P=.004).

Conclusion:

A simple gadoxetic acid-enhanced MRI-derived FLIS can identify patients with advance chronic liver disease at increased risk for hepatic decompensation and mortality.

Limitations:

Single-centre study.

Ethics committee approval

Approved by the Ethics Committee of the Medical University of Vienna.

Funding:

No funding was received for this work.

5
RPS 601c - Evaluation of liver fibrosis and cirrhosis on the basis of T1 mapping considering acute inflammation, age, and liver volume as confounding factors

RPS 601c - Evaluation of liver fibrosis and cirrhosis on the basis of T1 mapping considering acute inflammation, age, and liver volume as confounding factors

03:35C. Breit, Basel / CH

Purpose:

To evaluate confounding factors for the assessment of liver fibrosis by T1 mapping.

Methods and materials:

200 patients who underwent routine abdominal MRI at 1.5T and 3T scanners were retrospective included. 93 patients were defined as healthy based on the information of the hospital information system, 40 patients showed acute elevation of liver and bile parameters, and 67 subjects had a clinically or biopsy proven liver fibrosis or cirrhosis. A ROI based analysis of the T1 maps of the liver was performed. Additionally fat fraction, R2*, liver volume, laboratory parameters, sex, and age were regarded as potential confounding factors. Fibrosis was staged by using the Child-Pugh score and the METAVIR score.

Results:

There was a significant difference at 1.5T between T1 values in patients without known fibrotic changes and normal laboratory parameters (574.8+/-56ms) and those with elevated liver enzymes and bile levels (657.4+/-73ms) or known fibrotic liver disease (643.8+/-83 ms). At 3T the T1 values for patients with liver fibrosis or cirrhosis (995+/-150.3ms) were significantly higher than for the healthy group (870+/-128ms) and not significantly higher than for patients with laboratory abnormalities (952.4+/-137ms). There was a significant, moderate positive correlation between T1 values and the Child-Pugh stage at 1.5T (ρ=0.38, p=0.0112) and a non-significant, moderate positive correlation at 3T (ρ=0.3, p=0.26). There was no correlation between age and T1 values at 1.5T and 3T.

Conclusion:

T1 mapping is a reasonable method for the detection of liver fibrosis and cirrhosis. Age is no confounding factor hence age independent thresholds can be defined. Acute liver diseases are a confounding factor that need to be regarded.

Limitations:

Retrospective study with potential selection bias.

Ethics committee approval

Informed consent waived.

Funding:

No funding was received for this work.

PEP Subscription Required

This course is only accessible for ESR Premium Education Package subscribers.