Potential role of two-dimensional shear wave elastography, including liver stiffness measurement and dispersion slope, for management of chronic hepatitis B
Article information
Abstract
Purpose
This study assessed the use of liver stiffness (LS) and dispersion slope (DS) in classifying patients with chronic hepatitis B (CHB) by the necessity of antiviral treatment.
Methods
A retrospective review examined 249 patients with CHB (male:female, 107:142; mean age, 53±14 years) between April 2018 and March 2022. Patients for whom treatment was indicated, termed group 1, exhibited either serum alanine transaminase (ALT) and aspartate aminotransferase levels not exceeding five times the upper limit of normal and LS >13 kPa, or hepatitis B virus DNA >2,000 IU/mL and ALT >40 IU/L. The remaining patients comprised group 2 (treatment not required). In subgroup analysis, patients with LS <13 kPa were further examined.
Results
Overall, group 1 exhibited significantly higher LS (12.64±6.76 vs. 6.31±1.67 kPa, P<0.001) and DS (15.52±4.95 vs. 11.04±1.93 [m/s]/kHz, P<0.001) than group 2. In subgroup analysis, group 1 also demonstrated significantly higher values (LS: 8.05±2.12 vs 6.31±1.67 kPa, P<0.001, DS: 13.06±2.75 vs. 11.04±1.93 [m/s]/kHz, P<0.001). The areas under the curve (AUCs) for LS and DS in identifying group1 were 0.855 (95% confidence interval [CI], 0.80 to 0.90; P<0.001) and 0.810 (95% CI, 0.75 to 0.86; P<0.001), respectively. In subgroup analysis, the AUCs for LS and DS in identifying group 1 were 0.751 (95% CI, 0.69 to 0.81; P<0.001) and 0.711 (95% CI, 0.64 to 0.77; P<0.001), respectively. For identifying group 1, the AUCs for LS and DS did not differ significantly (P>0.05).
Conclusion
LS and DS assist in classifying patients with CHB by their need for antiviral treatment.
Introduction
Patients with hepatitis B virus (HBV) infection face an increased risk of developing cirrhosis, which can lead to life-threatening complications such as portal hypertension, liver failure, and hepatocellular carcinoma (HCC) [1]. Approximately 50% of patients with HCC have chronic HBV infection, and 80% of these patients develop HCC in conjunction with cirrhosis [2,3]. The 2017 European Association for the Study of the Liver (EASL) guidelines indicate that outcomes depend not only on the timing of therapy relative to the natural course of the infection but also on the disease stage and the age of the patient at treatment initiation [4].
Chronic infection and hepatitis with ongoing inflammation are key features used to differentiate liver conditions. An accurate assessment of the stages of liver fibrosis and inflammation is crucial for the effective management of chronic hepatitis B (CHB) [5]. While liver biopsy is considered the reference standard for assessing liver fibrosis, its invasiveness and the risk of potential complications limit its use [6]. Liver stiffness (LS) measurement using two-dimensional shear wave elastography (2D-SWE) has become a widely accepted method for diagnosing liver fibrosis [5,7]. The updated statement of the Society of Radiologists in Ultrasound (SRU) presents LS guidelines for patients with viral hepatitis and alcoholic fatty liver disease. Under these recommendations, an LS measurement below 9 kPa in the absence of clinical symptoms allows for the exclusion of compensated advanced chronic liver disease (cACLD); conversely, an LS value above 13 kPa strongly suggests cACLD [8].
Liver tissue exhibits viscoelastic properties, and the propagation of shear waves within the liver depends on viscosity and elasticity [9]. Liver viscosity can be indicative of conditions such as fibrosis, necroinflammatory activity, and steatosis [10-12]. Viscosity introduces frequency-dependent changes in shear wave velocity, a phenomenon referred to as shear wave dispersion (SWD) [13,14]. The dispersion slope (DS) has been shown to correlate with liver necroinflammation in studies involving rats, liver transplant recipients, and patients with nonalcoholic steatohepatitis (NASH)/ nonalcoholic fatty liver disease, which was recently renamed to metabolic dysfunction-associated steatohepatitis/steatotic liver disease [10,12,15-17].
Nevertheless, the details of DS have not been thoroughly investigated, and it remains unclear whether its association is limited to liver necroinflammation or extends to other pathological changes such as fibrosis and steatosis. In prior research, incorporating DS as an additional mechanical parameter did not significantly improve the diagnostic performance of LS for chronic liver disease (CLD) [18]. Moreover, studies of CHB cohorts are rare. The authors speculated that LS and DS could represent meaningful indicators when the patient sample was limited to CHB alone. The purpose of this study was to assess the utility of LS and DS in classifying patients with CHB based on the need for antiviral therapy.
Materials and Methods
Compliance with Ethical Standards
The Institutional Review Board at Wonkwang University Hospital approved this retrospective study (IRB No. 2023-09-014). The board waived the requirement for written informed consent.
Study Design
This study was conducted between April 2018 and March 2022. Data were analyzed from patients with CHB who had not received prior treatment or were diagnosed for the first time. These patients had undergone liver function tests and ultrasound with 2D-SWE within 1 week of laboratory examinations.
The two groups were modified in accordance with the 2017 EASL guidelines and the 2020 updated consensus statement on liver elastography from the SRU [4,8]. According to the 2017 EASL guidelines, since patients with cirrhosis should receive antiviral treatment, cirrhosis should be evaluated using an LS value of 13 kPa or higher, which represents the rule-in criterion for cACLD. In the present study, patients requiring treatment (termed group 1) were those who met one of the following criteria: (1) the cACLD rule-in criteria as presented [8], in which serum alanine transaminase (ALT) and aspartate aminotransferase (AST) levels do not exceed five times the upper limit of normal and LS is greater than 13 kPa; or (2) HBV DNA levels exceeding 2,000 IU/mL and ALT levels above 40 IU/L, as per the modified 2017 EASL guidelines [4]. Group 2 (those not requiring treatment) consisted of patients who did not meet the criteria for group 1.
The analysis was conducted in two stages. The first stage centered on comparing the groups across the entire patient population, regardless of LS level. The second stage of the analysis was restricted to patients with LS values below 13 kPa, within which groups 1 and 2 were compared.
Two-Dimensional Shear Wave Elastography
Two board-certified abdominal radiologists with 20 and 6 years of experience (Y.H.L. and Y.R.K., respectively), evaluated the patients using 2D-SWE with an Aplio i800 Ultrasound System (Canon Medical Systems Corp., Tokyo, Japan) equipped with a 1-8 MHz convex transducer. For optimal intercostal visualization, patients were placed in the supine position with the right arm fully abducted [10,11]. The liver parenchyma was first examined using B-mode attenuation imaging, followed by 2D-SWE. A sample box measuring approximately 2.5×2.5 cm was positioned in the right liver lobe, 1 cm beneath the liver capsule, to avoid reverberation artifacts. Patients were asked to hold their breath for 4-5 seconds during the procedure. Once the sample box had gathered the necessary data, the ultrasound system switched to quad view mode, simultaneously displaying elasticity, propagation, and dispersion maps, along with a grayscale image [10,11]. The LS values, expressed in kilopascals (kPa), and the DS values, expressed in meters per second per kilohertz ([m/s]/kHz), corresponding to the Young modulus E, were obtained concurrently [10,11]. A 1-cm circular region of interest was then manually placed within the sample box. Shear wave propagation was initiated, and 9 to 12 2D-SWE images were captured for each patient. LS, shear wave speed, and DS, corresponding to the interquartile range-to-median ratio of ≤0.3, were assessed. The median values of LS and DS were used for analysis. In addition, the skin-to-liver capsule distances were retrospectively measured on the ultrasound images for all patients.
Clinical Data
Labolatory findings within 1 week before and after 2D-SWE examination were analyzed. The laboratory results included factors related to HBV, specifically hepatitis B surface antigen, hepatitis B surface antibodies, and HBV DNA titers. Measurements were obtained of serum AST, ALT, alkaline phosphatase, gamma-glutamyl transferase, albumin, total bilirubin, and platelet (PLT) levels. Fibrosis-4 index scores were calculated using the formula: age (y)×AST (U/L)/[PLT (109/L)×ALT1/2 (U/L)].
Medical records were reviewed to calculate body mass index and assess previous treatment status, as well as to gather other medical and social history information.
Statistical Analysis
Continuous variables describing patient characteristics are presented as means±standard deviations. These variables were compared between the two groups using an independent-samples t-test. The relationship between LS and DS was assessed through Pearson correlation analysis, both overall and by subgroup. The overall diagnostic performance of LS and DS for the treatment group was determined based on the area under the curve (AUC) of receiver operating characteristic (ROC) analysis, which were expressed with 95% confidence intervals (CIs). Specificity, sensitivity, and 95% CIs were calculated for the cutoff values of these two parameters, obtained using the Youden index. The AUCs of the ROC curves were compared between LS and DS using the DeLong method.
All data were statistically analyzed using SPSS version 26 (IBM Corp., Armonk, NY, USA) and MedCalc version 12.7.0 (MedCalc Software Ltd., Ostend, Belgium). P-values of less than 0.05 were considered to indicate statistical significance.
Results
Patient Characteristics
This study enrolled 249 patients (male:female ratio, 107:142; mean age, 53±14 years) who underwent ultrasonography assessment using 2D-SWE and had complete laboratory findings available between April 2018 and March 2022.
Overall, 249 patients were included in the analysis, with 54 (21.7%) classified as requiring treatment (group 1) and 195 (78.3%) as not requiring treatment (group 2). The clinical characteristics of participants are detailed in Table 1. Within group 1, 23 patients exhibited LS values of 13 kPa or higher, while the remaining 31 patients were assigned based on the EASL guidelines. Of the 226 total patients with LS values below 13 kPa, 31 (13.7%) were in group 1 and 195 (86.3%) were in group 2. A flowchart illustrating the breakdown of each analysis is presented in Fig. 1.
Patient characteristics
Flowchart of patients included in the study.
AST, aspartate aminotransferase; ALT, alanine aminotransferase; HBV, hepatitis B virus; LS, liver stiffness.
Pearson Correlation between LS and DS
In both overall and subgroup analyses, LS and DS tended to increase concurrently. The Pearson correlation coefficient for the overall analysis was 0.716 (P<0.001), signifying a strong relationship. For the subgroup with LS <13 kPa, the Pearson correlation coefficient was 0.577 (P<0.001), indicative of a moderate relationship. Scatter plots illustrating the relationship between LS and DS for each analysis are presented in Fig. 2.
Scatter plots and Pearson correlation coefficients for LS and DS, as determined by two-dimensional shear wave elastography.
These plots are shown for the entire cohort of patients with CHB (A), and the subset of patients with LS <13 kPa (B). LS, liver stiffness; DS, dispersion slope; CHB, chronic hepatitis B.
Analysis of All Patients
The mean LS values for group 1 and group 2 were 12.64±6.76 kPa and 6.31±1.67 kPa, respectively (P<0.001). The mean DS values were 15.52±4.95 (m/s)/kHz for group 1 and 11.04±1.93 (m/s)/kHz for group 2 (P<0.001) (Table 2).
Intergroup comparisons of LS and DS
Among the patients considered to require treatment (group 1), the AUC of LS was 0.855 (95% CI, 0.80 to 0.90; P<0.001). The optimal cutoff value of 7.6 kPa yielded a sensitivity of 72.9%, a specificity of 83.2%, a positive predictive value (PPV) of 53.4%, a negative predictive value (NPV) of 91.5%, and a diagnostic accuracy of 80.3%. For the same group, the AUC of DS was 0.810 (95% CI, 0.75 to 0.86; P<0.001), with the optimal cutoff of 13.0 (m/s)/kHz resulting in a sensitivity of 60.4%, specificity of 87.3%, PPV of 54.1%, NPV of 88.8%, and diagnostic accuracy of 80.3% (Table 3). No significant difference in AUC was observed between LS and DS in group 1 (P=0.148) (Fig. 3A).
Diagnostic performance of LS and DS in patients with CHB
Receiver operating characteristic curves for LS and DS as determined by two-dimensional shear wave elastography in the entire cohort of patients with CHB (A), and the subset of patients with LS <13 kPa (B).
A. AUC of LS (P<0.001) is 0.855; optimal cutoff, 7.6 kPa; sensitivity, 72.9%; specificity, 83.2%. AUC of DS (P<0.001) is 0.810; optimal cutoff, 13 [m/s]/kHz; sensitivity, 60.4%; specificity, 87.3%. No significant differences were observed between LS and DS (P=0.148). B. AUC of LS (P<0.001) is 0.751; optimal cutoff, 6.1 kPa; sensitivity, 82.1%; specificity, 56.1%. AUC of DS (P<0.001) is 0.711; optimal cutoff, 11.0 [m/ s]/kHz; sensitivity, 78.6%; specificity, 53.2%. No significant differences were observed between LS and DS (P=0.413). LS, liver stiffness; DS, dispersion slope; CHB, chronic hepatitis B; AUC, area under the receiver operating characteristic curve.
All participants were further categorized based on thresholds of 7.6 kPa for LS and 13.0 (m/s)/kHz for DS, and the proportion of patients receiving treatment was calculated for each group. Among those with high LS and high DS, 67.4% (29/43) of patients were treated, whereas among patients with low LS and low DS, 7% (11/158) received treatment (Table 4).
Patient classification and treatment based on LS and DS levels
Analysis of Patients with LS <13 kPa
The mean LS values for group 1 and group 2 were 8.05±2.12 kPa and 6.31±1.67 kPa, respectively (P<0.001). The mean DS values were 13.06±2.75 (m/s)/kHz for group 1 and 11.04±1.93 (m/s)/kHz for group 2 (P<0.001) (Table 2).
Among the patients considered to require treatment (group 1), the AUC of LS was 0.751 (95% CI, 0.69 to 0.81; P<0.001). The optimal cutoff value of 6.1 kPa yielded a sensitivity of 82.1%, a specificity of 56.1%, a PPV of 21.6%, an NPV of 94.5%, and a diagnostic accuracy of 57.1%. For the same treatment group, the AUC of DS was 0.711 (95% CI, 0.64 to 0.77; P<0.001), with the optimal cutoff of 11.0 (m/s)/kHz resulting in a sensitivity of 78.6%, specificity of 53.2%, PPV of 19.5%, NPV of 93.2%, and diagnostic accuracy of 53.1% (Table 3). No significant difference in AUC was observed between LS and DS in group 1 (P=0.413) (Fig. 3B).
All participants were further categorized based on thresholds of 6.1 kPa for LS and 11.0 m/s for DS. Among those with high LS and high DS, 29.0% (24/89) received treatment, while 5.4% (4/74) of patients with low LS and low DS were treated (Table 4). A representative case of a patient who underwent treatment is illustrated in Fig. 4.
A 74-year-old woman with chronic hepatitis B who received antiviral treatment (ALT 118 IU/L, HBV DNA 5,700,000 IU/mL).
Using 2D-SWE, the median LS was measured at 6.7 kPa, while the median DS was 15.6 (m/s)/kHz. ALT, alanine aminotransferase; HBV, hepatitis B virus; 2D-SWE, two-dimensional shear wave elastography; LS, liver stiffness; DS, dispersion slope.
Discussion
Selecting patients with CHB for antiviral therapy is crucial to prevent the progression to fibrosis and inflammation. Clinical guidelines, such as those established by the EASL, employ laboratory parameters to identify patients who require antiviral treatment. Additionally, 2D-SWE can be employed to assess LS and help exclude cACLD. In this study, these methods were combined and modified to establish a group of patients requiring treatment and to analyze LS and DS accordingly. Both parameters were higher in patients requiring treatment and demonstrated good diagnostic capabilities. While LS exhibited greater diagnostic power than DS, the difference was not statistically significant.
LS is a reliable parameter for evaluating the stages of liver fibrosis [18-21]. In a previous study of patients with CHB, LS measured using 2D-SWE demonstrated good diagnostic performance for identifying severe liver fibrosis (at or above stage F2) and cirrhosis, with AUCs of 0.975 and 0.912, respectively [20]. Although the present study did not utilize pathological results as a reference standard, it aimed to adopt a clinical approach for identifying patients in need of antiviral treatment. LS was useful in diagnosing patients categorized as requiring antiviral therapy, both in the overall cohort (AUC, 0.855) and in the subset of participants with LS values below 13 kPa (AUC, 0.751).
Studies of DS vary depending on the disease entity. In the present study, the cutoff value of DS for patients with LS <13 kPa was determined to be 11.0 (m/s)/kHz. This is comparable to the 10.8 (m/s)/kHz reported in liver transplantation recipients [16] and in patients with NASH [12], aligning with the cutoff value for inflammatory activity in these diseases. Additionally, DS demonstrated excellent diagnostic performance in assessing the degree of lobular inflammation in NASH, with AUCs of 0.95 for inflammation grade A1, 0.81 for A2, and 0.85 for A3 [10,11]. These results contrast with previous studies suggesting that DS did not improve the diagnostic performance for fibrosis and inflammation in CLD [18,22-24]. The discrepancy may stem from the fact that LS and DS values vary across liver diseases; thus, a combined evaluation may not be meaningful. In the context of CHB alone, as examined in the present study, DS exhibited good diagnostic power (AUC, 0.810). Relative to DS, LS displayed higher AUC values in both total and subgroup analyses; however, the differences were not statistically significant (P=0.148 and P=0.413, respectively).
Although early studies often assumed that LS is associated with fibrosis and DS with inflammation, LS and DS increase and decrease concurrently when measured using 2D-SWE and SWD. These parameters vary depending on the disease entity, such as fibrosis or inflammation [18,22-24]. Since the present study did not separately analyze the pathology for inflammation and fibrosis, the mechanism by which LS and DS increase in patients with CHB requiring treatment remains unclear. Regardless of the distinction between LS and DS, however, both parameters were relatively high in patients clinically requiring antiviral treatment. Another study examining the roles of both LS and DS in assessing the severity of fibrosis and inflammation focused on autoimmune hepatitis (AIH). The research results indicated that LS and DS, as determined using 2D-SWE and SWD, are reliable indicators of hepatic fibrosis and are useful for evaluating the response of hepatic inflammation to treatment in patients with AIH during follow-up [25]. Similarly, in patients with CHB, both LS and DS displayed statistically significant results in classifying patients by the need for antiviral treatment. Therefore, based on previous research, DS does not appear useful when considering patients with CLD collectively. However, this parameter may hold clinical utility for specific diseases, such as NASH, AIH, and CHB.
Previous research on treatment options for patients with CHB has been limited. One study suggested excluding fibrosis and potentially avoiding treatment for patients with LS measurements of 8.5 kPa or lower. It also recommended confirming fibrosis through liver biopsy in patients with LS values between 8.5 and 11 kPa before making treatment decisions [26]. In contrast, regarding the need for treatment, the present study identified cutoff values of 7.6 kPa for the entire patient cohort and 6.1 kPa for those with LS values under 13 kPa. This is notably lower than the 8.5 kPa threshold suggested by the previous study. Furthermore, the present findings indicate that 8.5% (15 of 176) of patients with LS values below 7.6 kPa received antiviral treatment, suggesting that applying the 8.5 kPa criterion could inappropriately exclude some patients who could benefit from treatment. The authors contend that defining a wider range of patients eligible for antiviral therapy could be instrumental in preventing the progression to fibrosis and cirrhosis in hepatitis B. Further research with a larger patient cohort is necessary to ascertain the optimal LS threshold for excluding treatment.
One limitation of this study was its retrospective nature, which may have introduced discrepancies between the study criteria and those used in clinical practice. Generalization of the findings is also difficult, due to the small sample size, single-center design, and uneven distribution of participant groups. Additionally, most patients did not undergo liver biopsy. However, the established treatment for hepatitis B typically relies on imaging and biochemical markers, as not all patients are candidates for biopsy. To clarify the pathological significance of LS and DS in patients with CHB, these results must be validated by larger, prospective multicenter studies.
In conclusion, LS and DS are useful tools for classifying patients with CHB according to the need for antiviral treatment. Although DS exhibited a lower diagnostic power compared to LS, the difference was not statistically significant. Accordingly, both parameters can be beneficial when assessing CHB alone.
Notes
Author Contributions
Conceptualization: Song SJ, Kim YR, Lee YH. Data acquisition: Kim YR, Lee YH, Cho EY. Data analysis or interpretation: Song SJ, Kim YR. Drafting of the manuscript: Song SJ, Kim YR. Critical revision of the manuscript: Kim YR. Approval of the final version of the manuscript: all authors.
No potential conflict of interest relevant to this article was reported.
Acknowledgements
This paper was supported by Wonkwang University in 2023.
References
Article information Continued
Notes
Key point
Liver stiffness and dispersion slope values, as measured by two-dimensional shear wave elastography, can be instrumental in identifying patients with hepatitis B who require treatment.