Quantitative magnetic resonance imaging to aid clinical decision making in autoimmune hepatitis

Abstract

Background: In autoimmune hepatitis (AIH), clinical practice and treatment guidelines frequently diverge as a reflection of disease heterogeneity and challenges in achieving standardised care. We sought to explore the utility of multiparametric (mp) MR in patients with AIH, and the impact of this technology on physicians' decision making and intended patient management.

Methods: 82 AIH patients, recruited from two sites between June and November 2019 as part of an observational cohort study, underwent non-contrast MRI alongside their standard clinical investigations. Correlations between iron-corrected T1 (cT1) and other markers of disease were investigated alongside the utility of imaging markers to risk stratify patients in biochemical remission. The impact of mpMR on clinical decision making was evaluated using pairwise t-tests. The discriminatory ability of the imaging markers was assessed using area under the receiver operating characteristic curves (AUCs).

Findings: cT1 had a significant impact on clinician intended patient management (p<0.0001). cT1 correlated with ALT (p = 0.0005), AST (p<0.001), IgG (p = 0.0005), and liver stiffness (p<0.0001). Patients in deep biochemical remission (N = 11; AST/ALT <50% upper limit of normal [ULN] and IgG <12 g/L) had low cT1, while 7/34 in normal biochemical remission (AST/ALT between 50 and 100% of ULN) had high cT1 and were at risk of disease flare. cT1 measures of disease heterogeneity, ALP and bilirubin made the best predictor of those not in biochemical remission (AUC:0.85).

Interpretation: This study investigates the impact of mpMR results on intended clinical management in a real world setting. Findings showed that mpMR demonstrated a significant impact on clinical management of AIH and has the potential to inform patient risk stratification.

Funding: This paper presents independent research supported by the Innovate UK grant (104,915).

Keywords: Fibro-inflammation; Non-invasive imaging; Remission; Stratification; cT1.

Conflict of interest statement

ES, AD, MK and RB are employees of Perspectum. Perspectum Ltd is a privately funded commercial enterprise that develops medical devices to address unmet clinical needs, including LiverMultiScan®. MAH, RZA, MNR, MN, LS and ELC have no conflicts of interest to declare.

© 2022 The Authors.

Figures

Figure 1

The identification, recruitment, and active study procedure followed in this project.

Figure 2

(A) Changes in cT1 (ms) associated with increasing fibro-inflammatory burden and differences in cT1 between remission groups. In the images, lower values (cooler colours in maps and colour bar) represent areas with lower cT1 values and therefore lower fibro-inflammation, while higher cT1 values (warmer colours) represent areas of the liver with higher fibro-inflammation. (B) Classification of patients using biochemistry and cT1 to identify the spread of patients with resolved biochemistry that still have active fibro-inflammation.

Figure 3

Markers that were significantly different between those with mild active disease (ALT p = 0.004; pcT1: p<0.0001) and liver stiffness (p = 0.04) compared to those with active disease (ALT > x2 ULN) and high cT1 (cT1>800 ms). The whiskers of the boxplots represent the minimum and maximum values, the box covers the first and third quartiles with a line indicating the median. All outliers falling outside the area covered by the whiskers are indicated. cT1 IQR (ms): cT1 interquartile range a measure of disease heterogeneity, liver stiffness (kPa) measured of fibrosis by transient elastography, pcT1 (%): the percentage of the pixels in the cT1 map above 800 ms, measure of disease burden and heterogeneity.

Figure 4

Markers that were significantly different between those with active disease compared to those in biochemical remission (AST≤40IU/L and ALT≤40IU/L). Those with active disease had significantly higher cT1 (p = 0.002), pcT1 (p = 0.002), liver stiffness (p = 0.003) and IgG (p = 0.01) compared to those in biochemical remission. The whiskers of the boxplots represent the minimum and maximum values, the box covers the first and third quartiles with a line indicating the median. All outliers falling outside the area covered by the whiskers are indicated.

Figure 5

Markers that were significantly different between those in biochemical remission (AST≤40IU/L and ALT≤40IU/L) with (cT1800 ms) sub-clinically active disease on imaging. Those with clinically active disease on imaging had higher AST (p = 0.03), GGT (p = 0.0008), cT1 (p = 0.04) and pcT1 (p<0.0001) compared to those without clinically active disease on imaging. The whiskers of the boxplots represent the minimum and maximum values, the box covers the first and third quartiles with a line indicating the median. All outliers falling outside the area covered by the whiskers are indicated.