This retrospective, single-center study was approved by the institutional review board (Wonkwang University Hospital, 2021-07-031-001) of the authors’ affiliated hospital, which waived the need for written informed consent.

Between June 2014 and June 2021, 69 patients diagnosed with AIH were enrolled. The inclusion criteria were as follows: patients diagnosed with AIH according to the International Autoimmune Hepatitis Group score [2], patients who underwent 2D SWE and liver biopsy within 1 week, and patients older than 18 years of age.

The exclusion criteria were as follows: patients who were diagnosed with AIH in combination with other liver-related diseases; patients who did not undergo 2D SWE and/or liver biopsy; patients who had inadequate histologic results; patients who were younger than 18 years; and patients who underwent 2D SWE and liver biopsy with an interval of more than 1 week (Fig. 1).

Patients diagnosed with AIH were managed with or without steroid therapy according to the physician’s decision under international practice guidelines [1].

Baseline characteristics, laboratory test results, histologic results, and 2D SWE results were collected from electronic medical records and picture archiving and communication systems.

The 2D SWE measurements were performed using two US machines: the Aixplorer US system (SuperSonic Imagine, Aix-en-Provence, France) with a convex broadband probe (SC6-1, 1-6 MHz) (from June 2014 to January 23, 2018) and the Aplio i800 system (Canon Medical Systems Corporation, Tokyo, Japan) with a convex broadband probe (PVI-475BX, 1-8 MHz) (from January 24, 2018 to June 2021). All measurements were performed by one of the two radiologists. All patients were asked to fast for at least 6 hours prior to the examination. The patients raised their right arm above their head in the supine position to obtain the proper intercostal space and were asked to hold their breath during the measurement. The liver parenchymal echotexture was initially evaluated using B-mode imaging. Thereafter, 2D SWE was performed, placing a 2.5×2.5 cm sample box within the liver parenchyma on grayscale imaging. The sample box was placed at the parenchyma of the right hepatic lobe, 2 cm under the liver capsule, to avoid reverberation artifacts. The 2D SWE images were obtained according to the European Federation of Societies for Ultrasound in Medicine and Biology [11] and the World Federation of Societies for Ultrasound in Medicine and Biology [12] practice guidelines. The shear wave propagation data were acquired after the sample box was placed. Using the Aplio i800, the US system automatically displays the twin view of the grayscale images and shear wave propagation maps. The display was then switched to the quad-view mode, showing four maps simultaneously (the elasticity map, propagation map to visualize the confidence of the measurements, grayscale image as a reference, and the DS map) after a 5-s cooling time. Three 1-cm circular regions of interest (ROIs) were placed on each map in the sample box to avoid large hepatic vessels. Under guidance of the propagation map, ROIs were placed in the area exhibiting smooth and parallel lines on the propagation map, indicating stable measurements without reverberation or motion artifacts [13]. Using an elasticity map, LS was measured in units of kilopascals (kPa). Using a dispersion map, the DS was measured, with units of (m/s)/kHz. Activation of shear wave propagation and data filling of the sample box were performed three or four times for each patient, and one shear wave propagation data point had three ROIs. Therefore, 9-12 measurements of the LS and DS were made for each patient (Fig. 2). Measurements were considered reliable only when they were less than 30% of the interquartile range/median value. The median values of valid LS and DS measurements were chosen for further analysis.

Most patients visited every 6 months after their initial diagnosis of AIH, regardless of the type of medication. Follow-up SWE was generally performed 1 year after the initial examination and the second visit after the diagnosis; however, the interval was variable according to the physician’s discretion (15.9±11.3 months). Follow-up examinations were performed in the same manner as the initial examinations.

After 2D SWE measurements, a percutaneous liver biopsy was performed on the same day. Biopsy was performed twice with an 18-gauge semi-automatic biopsy gun (STARCUT, TSK Laboratory, Tochigi, Japan) in the right hepatic lobe. All specimens obtained from each patient were fixed in formalin, embedded in paraffin, stained with hematoxylin-eosin, and counterstained with one of two stains (reticulin or Masson). The samples were evaluated by a pathologist with more than 10 years of experience. The fibrosis stage was defined according to the METAVIR classification as normal (F0), mild fibrosis (F1), significant fibrosis (F2), advanced fibrosis (F3), and cirrhosis (F4) [14].

The clinical data and laboratory findings of AIH patients, including AST, ALT, alkaline phosphatase (ALP), gamma glutamyl transferase (GGT), albumin, total bilirubin, platelet count, and IgG, were collected on the same day as the 2D SWE measurements. Laboratory data were collected and analyzed to evaluate liver fibrosis, including the aspartate aminotransferase to platelet ratio index (APRI) [15] and fibrosis-4 index (FIB-4) [16].

The LS, DS, and laboratory test results are expressed as mean±standard deviation with range and normal distribution. The paired t test with the Bonferroni correction was used to compare variables before and after treatment. One-way analysis of variance was used to compare the LS and DS values among the histological liver fibrosis stages. The correlation between LS values and liver fibrosis was computed using the Spearman correlation coefficient. Correlation coefficient values were characterized as follows: negligible, ρ≤0.3; low, 0.3<ρ≤0.5; moderate, 0.5<ρ≤0.7; high, 0.7<ρ≤0.9; and very high, ρ≥0.9 [17]. The diagnostic performance of LS for liver fibrosis staging was evaluated using the receiver operating characteristic (ROC) curve and the area under the ROC curve (AUC). The optimal cutoff value showing the best diagnostic performance was calculated using the Youden index, maximizing sensitivity plus specificity. Statistical significance was set at P<0.05. All statistical analyses were performed using commercial statistical software (IBM SPSS Statistics for Windows version 26.0, IBM Corp., Armonk, NY, USA and MedCalc Software version 12.7.0, Ostend, Belgium).

A total of 69 patients (60 women and 9 men) were analyzed, with a mean age of 59.7±12.9 years. At the time of diagnosis, the mean modified AIH score was 13.5, with no statistically significant difference between men and women. The baseline characteristics of the 69 patients are listed in Table 1. The 69 patients were divided into four groups according to the METAVIR score (F1, n=17; F2, n=21; F3, n=16; and F4, n=15).

The mean values of LS were 7.3±2.1 kPa in the F1 group, 12.9±5.5 kPa in the F2 group, 13.0±5.1 kPa in the F3 group, and 15.1±4.2 kPa in the F4 group. The LS values were significantly different according to the stage of liver fibrosis (P<0.001). The mean values of DS were 11.8±1.9 (m/s)/kHz in the F1 group, 17.4±5.0 (m/s)/kHz in the F2 group, 14.9±3.3 (m/s)/kHz in the F3 group, and 14.0±3.7 (m/s)/kHz in the F4 group (P=0.044). The mean values of the laboratory tests, including AST, ALT, ALP, GGT, total bilirubin, albumin, and platelet count according to fibrosis grade at baseline are summarized in Table 2. There were no significant statistical differences in most laboratory findings among patients with different fibrosis stages, except albumin (P=0.047) and platelet count (P<0.001) (Table 2).

The AUCs and optimal cutoff values with the highest combined sensitivities and specificities of the LS value for detecting each stage of liver fibrosis are as follows: the AUC of the LS value was 0.903 (95% confidence interval [CI], 0.807 to 0.961; P<0.001; sensitivity, 90.4%; specificity, 82.4%) for ≥F2, 0.815 (95% CI, 0.703 to 0.898; P=0.001; sensitivity, 71.0%; specificity, 79.0%) for ≥F3, and 0.854 (95% CI, 0.748 to 0.927; P=0.002; sensitivity, 73.3%; specificity, 87.0%) for F4, respectively. The AUCs of APRI and FIB-4 for detecting each stage of liver fibrosis are summarized in Table 3.

The AUC of LS was significantly higher than that of APRI: ≥F2: LS vs. APRI, 0.903 vs. 0.657 (P<0.001); LS vs. FIB-4, 0.903 vs. 0.762 (P=0.020); ≥F3: LS vs. APRI, 0.815 vs. 0.531 (P<0.001); LS vs. FIB-4, 0.815 vs. 0.629 (P=0.008); F4: LS vs. APRI, 0.854 vs. 0.507 (P<0.001); and LS vs. FIB-4, 0.854 vs. 0.741 (P=0.193) (Fig. 3). The AUC values of the APRI and FIB-4 for detecting ≥F2 were 0.657 and 0.762 (P=0.106), respectively. The AUC values of APRI and FIB-4 for detecting ≥F3 were 0.531 and 0.629 (P=0.150), respectively, and the APRI and FIB-4 for detecting F4 were 0.507 and 0.741 (P=0.065), respectively.

Twenty-five of the 69 patients did not receive steroid therapy, according to the physician’s decision. Seventeen of these 25 patients underwent follow-up SWE examinations (mean interval, 13.2 months). There was no significant difference in LS or DS between the baseline and follow-up examinations (P=0.824 and P=0.452, respectively) (Table 4, Fig. 4).

Forty-four of the 69 patients received steroid therapy, and 34 of these 44 patients underwent follow-up SWE examination (mean interval, 17.4 months). There were significant differences in both the LS and DS (P=0.012 and P=0.011, respectively) (Table 4, Fig. 5). Therefore, the LS and DS of the patients receiving steroid therapy exhibited statistically significant improvements.