Print
facebook
twitter
Linkedin
Introduction
Accurately assessing hepatic fibrosis is crucial in patients with chronic liver diseases, such as metabolic dysfunction-associated steatotic liver disease (MASLD) and hepatitis B or C, as fibrosis is directly linked to disease prognosis [1]. Histopathology is traditionally regarded as the reference standard for liver fibrosis staging [2]. This approach employs a semiquantitative scoring system based on the extent and distribution of fibrotic tissue, which causes architectural distortion of the liver parenchyma [3]. In chronic liver disease, diagnostic accuracy is typically assessed by comparing the sensitivity, specificity, and area under the receiver operating characteristic (AUROC) curve with liver biopsy findings. However, liver biopsies sample only about 1/20,000th of the liver and are susceptible to sampling error, making biopsy a suboptimal reference standard [4].
Imaging-based elastography techniques using ultrasound and magnetic resonance imaging (MRI) have been implemented as noninvasive markers for assessing hepatic fibrosis [5,6]. Among these, ultrasound-based elastography is widely used due to its accessibility and reasonable accuracy, although its performance may be limited by patient body habitus, narrow intercostal spaces, and operator dependency [5,7-9]. In contrast, magnetic resonance elastography (MRE) offers higher reproducibility, a broader field of view, and greater accuracy, particularly in the early stages of fibrosis; accordingly, it is suitable for longitudinal monitoring in both clinical and research settings [10,11].
Vibration-controlled transient elastography (VCTE), introduced in 2002, is the most widely used ultrasound-based elastography technique, with numerous studies supporting its application [7,12]. Point shear wave elastography (pSWE) and two-dimensional SWE allow measurements under B-mode guidance, enabling artifacts to be identified and avoided by using quality maps or factors and thus potentially improving accuracy. However, few studies have directly compared these ultrasound-based techniques using MRE—a highly reproducible reference standard—as a comparator [5]. Therefore, the present study aimed to compare the diagnostic performance of VCTE and pSWE, using liver stiffness (LS) values obtained from MRE as the reference standard.
Materials and Methods
Compliance with Ethical Standards
This study received approval from the relevant institutional review board (approval number 20210422-6) and was conducted in accordance with the Declaration of Helsinki. Written informed consent was obtained from all patients prior to study participation for using their clinical and laboratory data. The trial was registered with the University Hospital Medical Information Network Clinical Trials Registry (UMIN000047411). This prospective study was conducted at Ogaki Municipal Hospital.
Patients
Between May 2021 and December 2024, adult patients at risk of chronic liver disease who underwent MRE, VCTE, and pSWE, with all examinations occurring within a 3-month period, were enrolled in the present study. VCTE and pSWE were conducted on the same day by the same experienced operator. The inclusion criteria were an age of at least 18 years and confirmed or suspected chronic liver disease based on blood test results, imaging findings (including ultrasonography and MRI), and consultations. The exclusion criteria were as follows: inaccurate VCTE or pSWE measurements; contraindications to MRI (e.g., presence of metal in the body or claustrophobia); inability to hold breath; known focal liver lesions in the right lobe; liver iron concentration ≥1.8 mg/g [6]; and clinical, laboratory, or imaging evidence of acute hepatitis or biliary tract tumors. Skin-capsule distance (SCD); body mass index (BMI); alanine aminotransferase, aspartate aminotransferase, albumin, and total bilirubin levels; platelet count; albumin–bilirubin (ALBI) score; and fibrosis-4 (FIB-4) index were measured for all patients on the same days as VCTE and pSWE [13,14].
Some of the study participants (187 of 251) were also included in a previous published study [15]. The previous investigation was conducted to assess the diagnostic performance of the improved algorithm of attenuation measurement (iATT). In contrast, the present study focuses on comparing pSWE and VCTE for the assessment of hepatic fibrosis, using MRE as the reference standard.
Ultrasound Examinations
pSWE examinations were conducted using the ARIETTA 850 ultrasonography system (FUJIFILM Corporation, Tokyo, Japan) with a convex broadband probe and shear wave measurement software. The protocol followed the method described by Yada et al. [8]. pSWE measurements were obtained from the right anterior lobe (liver segment V or VIII) [15,16]. If measurements from these segments were difficult to obtain, the right posterior lobe was used. Patients were included if at least five valid pSWE measurements were obtained, with an interquartile range-to-median ratio (IQR/median) less than 30% and a net amount of effective shear wave velocity exceeding 50%. A maximum of 20 measurements was allowed. VCTE examinations were performed using the FibroScan 430 ultrasonography system (Echosens, Paris, France) with the standard M or XL probe. Probe selection was based on the device’s automatic recommendation, and all measurements were performed using the suggested probe. Measurements followed the recommendations of the World Federation for Ultrasound in Medicine and Biology [17]. pSWE and VCTE were performed by one of three ultrasonography technologists, all of whom had more than 10 years of clinical experience. The technologists were blinded to the results of the other elastography modalities and to the patients’ clinical and laboratory data at the time of examination. In a previous study, intraclass correlation coefficients for intraobserver and interobserver reproducibility showed almost perfect agreement [15,16].
MRE and MRI Proton Density Fat Fraction
MRE and MRI proton density fat fraction (PDFF) measurements were conducted in accordance with previous descriptions [15,16,18], with LS values recorded in kilopascals (kPa). Fibrosis was staged as grade 0 (F0, no fibrosis) for LS values <2.61 kPa, grade 1 (F1, mild fibrosis) for LS values ≥2.61 kPa and <2.97 kPa, grade 2 (F2, significant fibrosis) for values ≥2.97 kPa and <3.62 kPa, grade 3 (F3, advanced fibrosis) for values ≥3.62 kPa and <4.69 kPa, and grade 4 (F4, cirrhosis) for values ≥4.69 kPa [19]. Steatosis was classified as grade 0 (S0) for MRI-PDFF values <5.2%, grade 1 (S1) for those ≥5.2% and <11.3%, grade 2 (S2) for those ≥11.3% and <17.1%, and grade 3 (S3) for those ≥17.1% [20].
Statistical Analyses
Continuous variables are presented as median with IQR. The Kruskal-Wallis test was employed for continuous variables, with post hoc analysis conducted using the Steel–Dwass test if a significant difference was found. A normal probability plot was utilized to evaluate whether VCTE, pSWE, and LS values were normally distributed; if not, values were transformed to achieve normality.
The Pearson correlation coefficient (CC) and Lin concordance correlation coefficient (CCC) were calculated to assess the relationship between LS measurements. While CC quantifies the strength of linear associations, CCC additionally accounts for both precision and bias, providing a more comprehensive evaluation of measurement agreement. Correlations between VCTE, pSWE, and LS values were analyzed using the Pearson CC, with statistical tests for differences in correlations conducted according to methods described by Diedenhofen and Musch [21]. Concordance between pSWE or VCTE and MRE was tested using the Lin CCC [22]. Prior to calculating the Lin CCC, linear regression was used to convert pSWE and VCTE values to the same scale as the LS value. The Pearson CC and Lin CCC were categorized as poor (0.00-0.20), fair (0.21-0.40), moderate (0.41-0.60), good (0.61-0.80), or excellent (0.81-1.00) [23]. Additional analyses of Pearson CC and Lin CCC values were performed in subgroups stratified by BMI, SCD, FIB-4 index, and ALBI score, using cutoff values of 30 kg/m2 [24], 25 mm [25], 2.67 [26], and -2.60 [13], respectively.
The diagnostic performances of VCTE and pSWE were evaluated using AUROC analysis, with differences assessed by the DeLong test [27]. Optimal cutoff values were determined using the Youden index, with additional thresholds identified for 90% sensitivity and specificity. Sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) were calculated for the selected cutoffs. The necessary sample size for AUROC analysis was determined based on a null hypothesis AUROC of 0.70 and an expected AUROC of 0.85, with a two-sided significance level of 0.05 and statistical power of 80%. Assuming an approximately equal distribution of patients with and without significant fibrosis, the minimum total sample size was estimated to be 200 patients. Integrated discrimination improvement (IDI) and net reclassification improvement (NRI) were calculated to assess the clinical utility of pSWE compared to VCTE for each fibrosis grade [28]. The Obuchowski index was applied to account for all pairwise comparisons across stages of liver fibrosis, reducing the spectrum effect and risk of multiple testing [29,30]. Each Obuchowski index value was calculated based on subsamples of the original dataset using the bootstrap resampling method [31]. The original dataset consisted of paired measurements, with LS values matched to corresponding VCTE and pSWE values. The Obuchowski index can be interpreted similarly to the AUROC, as it evaluates whether the model correctly ranks every possible pair of patients [29,30].
P-values less than 0.05 were considered to indicate statistical significance. All statistical analyses were conducted using EZR version 1.68 (Saitama Medical Center, Jichi Medical University, Saitama, Japan), a graphical user interface for R (R Foundation for Statistical Computing, Vienna, Austria) [32].
Results
Patient Characteristics
Between May 2021 and December 2024, this study screened 287 adult patients at risk of chronic liver disease who underwent MRE, VCTE, and pSWE within a 3-month period. After application of the exclusion criteria, 36 patients were excluded. The final analysis included 251 patients (97 female, 154 male) with a median age of 64 years and a median BMI of 24.6 kg/m2, all of whom had pSWE, VCTE, and MRE measurements available (Fig. 1). Representative measurement images are shown in Fig. 2A-D, and the characteristics of the enrolled patients are summarized in Table 1.
Quantitative Performance of pSWE and VCTE in Fibrosis Staging
To assess the quantitative performance of VCTE and pSWE in diagnosing liver fibrosis, boxplot analysis was performed based on fibrosis stage determined by LS (Fig. 3A, B). Patients were categorized into five fibrosis stages according to the severity of liver fibrosis: F0 (n=101), F1 (n=30), F2 (n=54), F3 (n=28), and F4 (n=38). For F0, F1, F2, F3, and F4, the mean VCTE values were 5.608 kPa, 5.973 kPa, 9.980 kPa, 17.889 kPa, and 23.905 kPa, respectively, while the mean pSWE values were 4.640 kPa, 5.329 kPa, 8.530 kPa, 11.458 kPa, and 18.679 kPa. The boxplots revealed statistically significant differences between all staging pairs for both VCTE and pSWE.
Pearson CC and Lin CCC between pSWE, VCTE, and LS Values
pSWE, VCTE, and LS values were log-transformed (Supplementary Fig. 1A-C). To assess the Lin CCC, log pSWE and log VCTE values were converted to corrected LS values using two simple linear regression equations: corrected MREpSWE=0.05627+0.55713×log pSWE (0.02444+0.55713×log pSWE) and corrected MREVCTE=0.15371+0.46698×log VCTE (0.06675+0.46698×log VCTE). Table 2 shows the Pearson CC and Lin CCC values between pSWE, VCTE, and LS measurements. In the overall analysis, both pSWE and VCTE demonstrated strong correlations with LS, with no significant difference between them (Fig. 4A, B). The Lin CCC also displayed good agreement with LS (Table 2). Subgroup analyses revealed that, compared with VCTE, pSWE exhibited closer correlation and concordance with LS in patients with BMI <30 kg/m2 and FIB-4 score <2.67. In patients with SCD <25 mm and those with an S grade of 0-1, pSWE showed a significantly higher correlation with MRE than VCTE, although the differences in Lin CCCs were not statistically significant in either subgroup.
Diagnostic Accuracy Based on AUROC
Table 3 presents the AUROC, cutoff values, sensitivity, specificity, PPV, and NPV for pSWE and VCTE across different fibrosis stages. VCTE exhibited significantly higher specificity than pSWE for diagnosing ≥F3 fibrosis. Fig. 5 shows the AUROC values for diagnosing ≥F2, with no statistically significant difference between pSWE and VCTE. The cutoff values for sensitivities and specificities of ≥0.90 are provided in Supplementary Tables 1 and 2. Compared to VCTE, at a sensitivity threshold of ≥0.90, pSWE demonstrated significantly higher specificity and PPV for ≥F2, as well as higher specificity for ≥F4. At a specificity threshold of ≥0.90, pSWE showed significantly higher sensitivity for ≥F1 than VCTE.
Discrimination Performance of pSWE and VCTE
The diagnostic utility of pSWE and VCTE was further evaluated using non-categorical IDI and NRI analyses to evaluate their reclassification performance across fibrosis stages (Table 4). For ≥F1, pSWE exhibited significantly higher IDI and NRI values than VCTE. Similarly, for ≥F2 and ≥F4, pSWE showed significantly higher IDI and NRI. No significant difference was observed between pSWE and VCTE for ≥F3.
Diagnostic Accuracy Using the Obuchowski Index
The diagnostic performance of pSWE and VCTE across liver fibrosis stages was further evaluated using the Obuchowski index (Table 5). For detecting ≥F1, the Obuchowski index for pSWE was 0.890, which was significantly higher than that for VCTE (0.864). pSWE also demonstrated a significantly higher Obuchowski index than VCTE for detecting ≥F2 and ≥F4. However, no significant difference was found between pSWE and VCTE for detecting ≥F3.
Discussion
This study evaluated the diagnostic performance of pSWE and VCTE for staging liver fibrosis, with MRE used as the reference standard. As the fibrosis stage defined by LS values increased, measurements obtained by both VCTE and pSWE also increased significantly. pSWE demonstrated closer agreement with MRE-based fibrosis staging than VCTE, particularly for grades ≥F1, ≥F2, and ≥F4. This increased concordance was supported by higher Obuchowski index scores and improved reclassification metrics, including IDI and NRI. MRE is generally recognized as the most accurate noninvasive imaging modality for liver fibrosis staging [5]. In addition, it has been reported to overcome the limitations of sampling error and invasiveness associated with liver biopsy, as well as to show high concordance with histological evaluation of liver resection specimens [11]. Therefore, the higher level of agreement observed between pSWE and MRE-based fibrosis classification suggests that pSWE could serve as a reliable noninvasive method in clinical practice.
Since its approval by the U.S. Food and Drug Administration as a diagnostic and monitoring tool for liver fibrosis in 2013, VCTE has been widely used worldwide, demonstrating its utility and consistently high technical success rates [33]. VCTE-based liver fibrosis measurement has been increasingly adopted as a noninvasive point-of-care modality in clinical practice; it serves both as an initial assessment tool for the detection of advanced fibrosis in patients with MASLD and as a means for longitudinal disease monitoring [34]. Previous meta-analyses have reported that VCTE demonstrates high diagnostic accuracy for stages F1 to F4, with sensitivities ranging from 76% to 89% and specificities from 67% to 73% [35]. The reference standard used in these studies was liver biopsy, which remains the gold standard for assessing liver fibrosis. However, the invasiveness of liver biopsy limits its clinical applicability, and biopsy is susceptible to sampling error due to the heterogeneous distribution of fibrosis [4]. Consequently, MRE—a noninvasive tool capable of evaluating the entire liver—has been increasingly utilized as an alternative reference standard for liver fibrosis assessment [5].
Consistent with previous research comparing VCTE and pSWE, the present study revealed no significant difference between the two modalities when evaluated using AUROC analysis [36]. Although AUROC is widely used to assess the diagnostic performance of noninvasive tests, it is inherently limited to binary classification. Liver fibrosis, however, progresses in a stepwise, ordinal fashion from F0 to F4, and accurate assessment requires the ability to distinguish between the extremes (e.g., F0 vs. F4) as well as between adjacent stages (e.g., F1 vs. F2). To address the potential bias introduced by differences in fibrosis stage distribution (i.e., spectrum effect) and the loss of information inherent in dichotomizing ordinal outcomes, this study applied the Obuchowski index—a statistical measure designed to assess diagnostic accuracy for ordinal endpoints. This index incorporates all pairwise comparisons across fibrosis stages, providing a more comprehensive and clinically relevant assessment of staging performance than AUROC alone. By applying the Obuchowski index, it was possible to more precisely characterize how effectively pSWE and VCTE differentiate across the full spectrum of fibrosis severity [30]. In the present findings, a higher Obuchowski index was achieved using pSWE than VCTE, particularly for detecting mild fibrosis, significant fibrosis, and cirrhosis. This observation suggests a closer alignment between pSWE and MRE-based classification and may support more consistent fibrosis staging using pSWE in clinical practice.
The observed difference in diagnostic accuracy may be attributed to the fundamental technical differences between VCTE and pSWE. VCTE uses mechanical vibrations to generate shear waves and lacks two-dimensional image guidance, limiting precise anatomical targeting. However, it provides standardized measurements with minimal operator dependency and has been widely validated in various populations. In contrast, pSWE employs focused ultrasound pulses and offers real-time two-dimensional image guidance, allowing for anatomically targeted region of interest placement. While this improves targeting precision, the flexibility of pSWE introduces variability in site selection and increases operator dependency, potentially affecting reproducibility. Accordingly, the American Association for the Study of Liver Diseases Practice Guidelines note that pSWE requires "technical expertise" and may be influenced by "site selection variability" [5].
The risk of future liver-related events begins to rise at stage F2, which indicates significant fibrosis. It is well known that significant fibrosis predicts adverse liver outcomes [37,38], warranting timely clinical intervention. Under the European Association for the Study of the Liver–European Association for the Study of Diabetes–European Association for the Study of Obesity Clinical Practice Guidelines, early identification and management of fibrosis may prevent progression to cirrhosis; the guidelines recommend screening in high-risk populations (level of evidence 3, strong consensus) [1]. Compared to VCTE, pSWE provided fibrosis staging results more closely aligned with those of MRE. This suggests a potential role for pSWE in risk stratification for liver fibrosis, particularly in identifying patients with significant fibrosis. Given the prognostic importance of significant fibrosis, improved alignment through pSWE may facilitate earlier clinical intervention, support patient management, and potentially improve long-term liver outcomes.
This study had several limitations. First, histological confirmation of liver fibrosis was not available. Therefore, MRE was used as the reference standard. Despite some limitations, MRE is a well-validated noninvasive modality with high reproducibility and accuracy for liver fibrosis assessment [10,11]. Compared to liver biopsy, MRE enables comprehensive liver assessment with reduced sampling error. Nonetheless, LS measured by MRE may be confounded by factors such as hepatic inflammation, ballooning degeneration, hyperemia, and edema, which can result in overestimation of fibrosis [5]. Accordingly, the present findings should be interpreted considering these potential confounders. Despite these limitations, MRE was considered the most appropriate reference standard for this study due to its validated performance and noninvasive nature. Further studies incorporating histological assessment are warranted to corroborate these findings. Second, this study took place at a single institution and included a total of 251 patients. As a result, the generalizability of the findings to other institutions, ethnicities, and populations remains uncertain. Future multicenter, prospective validation studies with larger and more diverse patient populations are warranted. Third, only 40 patients (15.9%) in this study had a BMI ≥30 kg/m2. Obesity is a well-established confounding factor in LS measurements, affecting measurement accuracy [5]. Future studies should include a broader BMI range to improve the generalizability and validity of the results.
When MRE is used as a reference, pSWE values demonstrated a closer approximation to LS values than VCTE, particularly in detecting mild, significant, and advanced fibrosis. These findings suggest that pSWE may function as a reliable tool for noninvasive fibrosis assessment, offering greater alignment with MRE-based evaluations and potentially facilitating the early diagnosis and management of liver disease.
Download Citation
PDF Links
PubReader
ePub Link
XML Download
Full text via DOI
Full text via PMC
