Patient characteristics
Patient characteristics are shown in Table 1. The causes of liver disease were hepatitis B in 31 patients, hepatitis C in 43 (all after achieving sustained virological response), NAFLD in 93, autoimmune hepatitis in 15, primary biliary cholangitis in 14, alcohol-related liver disease (ALD) in 10, and other causes in 3. Of these, 76 patients underwent liver biopsy (Table S1).
Relationship between 2D-SWE and VCTE and associated factors
A strong positive correlation was observed between LSM using 2D-SWE (LSM-SWE) and LSM using VCTE (LSM-VCTE) (r=0.78, p <0.001) (Figure 1A). A Brandt–Altman analysis revealed that the mean difference between LSM-SWE and LSM-VCTE was 1.86 ± 15.84 kPa, and the 95% upper and lower limits of agreement were 17.70 kPa and -13.98 kPa, respectively (Figure 1B). As shown in the plot, there was a proportional error between LSM-SWE and LSM-VCTE, where the difference in measurements increased as LSM increased.
Univariate analysis revealed that the following factors were significantly associated with LSM-SWE: body mass index (BMI), male sex, platelet count, albumin, asparagine aminotransferase (AST), alanine aminotransferase (ALT), γ-glutamyltransferase (GGT), total bilirubin, and skin to capsular distance (SCD). Factors associated with LSM-VCTE were similar to those associated with LSM-SWE. Multivariate analysis showed that platelet count and GGT level were independently associated with LSM-SWE and LSM-VCTE (Table 2).
Diagnostic ability of LSM-SWE for liver fibrosis staging
In cases where a liver biopsy was performed, LSM-SWE increased significantly, according to the liver fibrosis stage (Figure 1C). Areas under the receiver operating characteristic curves (AUCs) of LSM-SWE for diagnosing fibrosis stages F2, F3, and F4 were 0.74, 0.79, and 0.89, respectively (Figure 1D). Optimal cut-off levels of LSM-SWE for diagnosing liver fibrosis were 8.7 kPa for F2 (sensitivity [se], 68.7%; specificity [sp], 76.5%), 9.1 kPa for F3 (se, 69.4%; sp, 68.8%), and 11.6 kPa for F4 (se, 78.6%; sp, 85.2%) (Table 3).
Relationship between ATI and CAP and associated factors
ATI and CAP showed a strong positive correlation (r=0.70, p <0.001) (Figure 2A). Univariate analysis revealed that BMI, age, male sex, platelet count, AST, ALT, GGT, total bilirubin, and SCD levels were significantly associated with ATI. All these factors were associated with ATI, except for GGT. The multivariate analysis revealed that BMI was an independent factor associated with ATI and CAP, whereas SCD was significantly associated only with CAP (Table 4).
Diagnostic ability of ATI for hepatic steatosis
In cases where liver biopsy was performed, the ATI increased significantly, according to the hepatic steatosis stage (Figure 2B). The AUCs of the ATI for the diagnosis of hepatic steatosis stages S1, S2, and S3 were 0.91, 0.81, and 0.88, respectively (Figure 2C). The optimal cut-off levels of ATI for diagnosing hepatic steatosis were 0.66 dB/cm/MHz for S1 (se, 88.2%; sp, 86.7%), 0.79 dB/cm/MHz for S2 (se, 70.8%; sp, 71.4%), and 0.86 dB/cm/MHz for S3 (se, 85.7%; sp, 81.4%) (Table 5).
Diagnostic performance of LSM-SWE and ATI in patients with NAFLD
A subgroup analysis was performed in patients with NAFLD. The patient characteristics are shown in Table S2. The LSM-SWE and LSM-VCTE showed very strong positive correlations in patients with NAFLD (r=0.83, p <0.001) (Figure 3A). In cases where a liver biopsy was performed, LSM-SWE increased significantly, according to the liver fibrosis stage (Figure 3B). The AUCs of LSM-SWE for diagnosing F2, F3, and F4 were 0.80, 0.82, and 0.93, respectively (Figure 3C, Table S3). Multivariate analysis revealed that platelet count and GGT were independently associated with LSM-SWE and LSM-VCTE in patients with NAFLD, which is consistent with the results of the analysis in all the patients (Table S4).
The ATI and CAP levels showed a strong positive correlation in patients with NAFLD (r=0.62, p <0.001) (Figure 3D). In cases where liver biopsy was performed, the ATI increased significantly according to the hepatic steatosis stage (Figure 3E). The AUCs of ATI for diagnosing S1, S2, and S3 were 0.90, 0.77, and 0.83, respectively (Figure 3F, Table S5). Univariate analysis revealed that levels of BMI, AST and ALT levels, and SCD were significantly associated with the ATI and CAP levels. However, the multivariate analysis revealed no independent factors associated with ATI or CAP in patients with NAFLD (Table S6).