Date Published: July 21, 2018
Publisher: Springer US
Author(s): A. Mojtahed, C. J. Kelly, A. H. Herlihy, S. Kin, H. R. Wilman, A. McKay, M. Kelly, M. Milanesi, S. Neubauer, E. L. Thomas, J. D. Bell, R. Banerjee, M. Harisinghani.
Corrected T1 (cT1) value is a novel MRI-based quantitative metric for assessing a composite of liver inflammation and fibrosis. It has been shown to distinguish between non-alcoholic fatty liver disease (NAFL) and non-alcoholic steatohepatitis. However, these studies were conducted in patients at high risk for liver disease. This study establishes the normal reference range of cT1 values for a large UK population, and assesses interactions of age and gender.
MR data were acquired on a 1.5 T system as part of the UK Biobank Imaging Enhancement study. Measures for Proton Density Fat Fraction and cT1 were calculated from the MRI data using a multiparametric MRI software application. Data that did not meet quality criteria were excluded from further analysis. Inter and intra-reader variability was estimated in a set of data. A cohort at low risk for NAFL was identified by excluding individuals with BMI ≥ 25 kg/m2 and PDFF ≥ 5%. Of the 2816 participants with data of suitable quality, 1037 (37%) were classified as at low risk.
The cT1 values in the low-risk population ranged from 573 to 852 ms with a median of 666 ms and interquartile range from 643 to 694 ms. Iron correction of T1 was necessary in 36.5% of this reference population. Age and gender had minimal effect on cT1 values.
The majority of cT1 values are tightly clustered in a population at low risk for NAFL, suggesting it has the potential to serve as a new quantitative imaging biomarker for studies of liver health and disease.
Liver disease is a growing clinical problem being fuelled by the obesity epidemic. Obesity is a risk factor for fatty liver disease, which has been defined as hepatic fat content greater than 5% . Importantly, liver disease, unlike chronic pulmonary, cardiac, or neurodegenerative diseases, is usually asymptomatic, with few physical manifestations until the development of cirrhosis; thus, detection and case-finding cannot be based on clinical history and examination alone. Scaglione et al. estimated the prevalence of cirrhosis in the United States to be approximately 0.27% from NHANES and death registry data 1999–2010, corresponding to 633,323 adults. Mortality was 26.4% per 2-year interval in cirrhosis compared with 8.4% in propensity-matched controls . Identifying those individuals at risk of cirrhosis before it develops would be of enormous utility.
The pilot phase of the UK Biobank imaging enhancement study has shown that high-throughput phenotyping is feasible. However, as MR methods develop (e.g., multispectral measurement of PDFF), further refinement of these techniques will need testing and implementing. The current ROI-based analysis from a single slice may have limitations in very heterogeneous disease distributions. In such cases, a histogram-based analysis may be more appropriate. A whole liver evaluation would allow better characterization of heterogeneity and reduce sampling errors; however, given the extremely high throughput of the study, there was insufficient time in the protocol to scan the whole liver. Given the UK Biobank subjects were drawn from a nominally healthy population, this is unlikely to significantly impact the conclusions of this study. The imaging data used in this study were acquired on a single scanner and therefore does not include any potential variability that may occur between different MRI scanners; however, computation of cT1 explicitly accounts for differences in field strength and T1 mapping implementation (LiverMultiScan version 2.0, Perspectum Diagnostics Ltd, UK). Approximately 4% of scans in this normal population were incomplete or of insufficient quantity, and this may increase when applied to an actual patient population.
This article describes, for the first time, the reference range of cT1 values as defined in a large population at low risk for NAFL. The range presented here has the potential, where desirable, to serve as a benchmark of normality for future studies assessing NAFL with PDFF and cT1.