Variability of noninvasive MRI and biological markers in compensated cirrhosis: insights for assessing disease progression

Abstract

Background: We annually monitored stable compensated cirrhosis (CC) patients to evaluate serial variation in blood serum, liver stiffness, and multiparametric magnetic resonance imaging (mpMRI) measures to provide reference change values (RCV) and sample size measures for future studies.

Methods: Patients were recruited from a prospectively followed CC cohort, with assessments at baseline and annually over three years. We report on blood markers, transient elastography liver stiffness measures (LSM) and noninvasive mpMRI (volume, T1 mapping, blood flow, perfusion) of the liver, spleen, kidneys, and heart in a stable CC group and a healthy volunteer (HV) group. Coefficient of variation over time (CoVT) and RCV are reported, along with hazard ratio to assess disease progression. Sample size estimates to power future trials of cirrhosis regression on mpMRI are presented.

Results: Of 60 CC patients enrolled, 28 with stable CC were followed longitudinally and compared to 10 HVs. CoVT in mpMRI measures was comparable between CC and HV groups. CoVT of Enhanced Liver Fibrosis score was low (< 5%) compared to Fibrosis-4 index (17.9%) and Aspartate Aminotransferase-to-Platelet-Ratio Index (19.4%). A large CoVT (20.7%) and RCV (48.3%) were observed for LSM. CoVT and RCV were low for liver, spleen, and renal T1 values (CoVT < 5%, RCV < 8%) and volume (CoVT < 10%, RCV < 16%); haemodynamic measures were high (CoVT 12-25%, RCV 16-47%).

Conclusions: Evidence of low CoVT and RCV in multiorgan T1 values. RCV and sample size estimates are provided for future longitudinal multiorgan monitoring in CC patients.

Trial registration: ClinicalTrials.gov identifier: NCT02037867 , Registered: 05/01/2013.

Keywords: Biomarkers; Disease progression; Liver cirrhosis; Multiparametric magnetic resonance imaging; Sample size.

Conflict of interest statement

The authors declare that they have no competing interests.

© 2022. The Author(s) under exclusive licence to European Society of Radiology.

Figures

Fig. 1

Schematic of the study and consort diagram. Schematic of blood markers (MELD, UKELD, APRI, FIB4, ELF), Fibroscan® LSM, and multiparametric magnetic resonance imaging (volume, T1-mapping, blood flow, perfusion) of the liver, spleen and kidneys, and cardiac index. Illustration of healthy volunteers (HV) and compensated cirrhosis (CC) patients studied longitudinally as indicated by consort diagram. MELD Model for end-stage liver disease, UKELD United Kingdom model for end-stage liver disease, APRI Aspartate aminotransferase to platelet ratio index, FIB4 Fibrosis-4, ELF Enhanced liver fibrosis, LSM Liver stiffness measure, TE Transient elastography

Fig. 2

Baseline magnetic resonance imaging parameters for 40 healthy volunteers (HV), 60 compensated cirrhosis (CC) patients, and 7 decompensated cirrhosis (DC) patients. Baseline measures (mean and standard error of the mean) of the liver (volume, portal vein area, total hepatic blood flow, liver perfusion, liver T1), spleen (volume, splenic artery flow and superior mesenteric artery flow, spleen perfusion, spleen T1), kidney (renal cortex T1), and heart (cardiac index and left ventricle [LV] wall mass index) are shown, with the percentage change between the HV and CC groups, and CC and DC groups shown by arrows. Asterisk indicates measures which are significantly different (p < 0.05, independent samples t-test) between the CC and HV group [10]

Fig. 3

a Year-to-year percentage change in clinical measures in the stable compensated cirrhosis control group. b Year-to-year coefficient of variation (CoVT) in clinical measures in the stable compensated cirrhosis control group. Clinical measures of MELD, UKELD, APRI, FIB4, ELF scores, and Fibroscan® LSM are shown. Bars indicate the interquartile range and the horizontal bold line shows the median, dots represent outliers. MELD Model for end-stage liver disease, UKELD United Kingdom model for end-stage liver disease, APRI Aspartate aminotransferase to platelet ratio index, FIB4 Fibrosis-4, ELF Enhanced liver fibrosis, LSM Liver stiffness measure

Fig. 4

Year-to-year percentage change in magnetic resonance imaging measures in the stable compensated cirrhosis control group. Measures liver (volume, portal vein area, total hepatic blood flow, liver perfusion, liver T1), spleen (volume, splenic and superior mesenteric artery flow, spleen perfusion, spleen T1), kidney (renal cortex T1), and heart (cardiac index and left ventricle [LV] wall mass index) are shown. Bars indicate interquartile range and horizontal bold line shows the median percentage change, dots represent outliers

Fig. 5

Year-to-year coefficient of variation (CoVT) in magnetic resonance imaging measures in the stable compensated cirrhosis control group. Bars indicate the interquartile range and the horizontal bold line shows the median CoVT at each time point, dots represent outliers. The technical variation termed the analytical CoV (CoVA) measured in healthy volunteers from triplicate scans collected 1 week apart [10] is shown by the grey dashed line

Fig. 6

ELF score, Fibroscan® LSM and liver T1 for the stable compensated cirrhosis control patients who completed all three annual follow-up scans: a individual subject percentage change from baseline values at year 1, year 2, and year 3; b group percentage change from baseline values at year 1, year 2, and year 3. Bars indicate the interquartile range and bold line shows the median percentage change. There was no significant difference from baseline (p > 0.05 Bonferroni-corrected); c Year-to-year coefficient of variation (CoVT) in magnetic resonance imaging measures. Bars indicate the interquartile range and the horizontal bold line shows the median CoV, dots represent outliers. ELF Enhanced liver fibrosis, LSM Liver stiffness measure

Fig. 7

A schematic of the hazard ratio (HR) for disease progression. Panel a (i) shows the progression from healthy volunteer (HV) to compensated cirrhosis (CC), whilst panel a (ii) shows that the hazard ratio for progression from CC to decompensated cirrhosis (DC). Positive values indicate an increase in measure and negative values a decrease in measures. If the absolute value of the HR is >1 this indicates the reference change value (RCV) is less than the clinical change. Panel b shows sample size estimation for the number of CC patients required in clinical trial to detect a change from stage F4 (compensated cirrhosis, CC) to F3 (advanced cirrhosis) liver disease which has clinical significance; data points are shown for a 25%, 50%, 75%, and 100% regression from F4 to F3. BSA Body surface area