Abstract
Background & Aims
Controlled attenuation parameter (CAP) is a non‐invasive method for evaluating hepatic steatosis. However, larger skin capsular distance (SCD) can affect the accuracy. The aim of this study was to investigate the impact of SCD on the diagnostic performance of CAP and liver stiffness measurement (LSM).
Methods
Of 101 patients with non‐alcoholic fatty liver disease (NAFLD) and 280 patients with chronic hepatitis B (CHB) who underwent liver biopsy were prospectively recruited. CAP, LSM and SCD were performed using FibroScan with M probe. The areas under receiver operating characteristics curves (AUROCs) were calculated to determine the diagnostic efficacy. The optimal thresholds were defined by the maximum Youden index.
Results
SCD (B 30.34, P < 0.001) and hepatic steatosis (B 23.04, P < 0.001) were independently associated with CAP by multivariate analysis. The AUROCs were slightly higher for SCD <25 mm compared to those for SCD ≥25 mm for steatosis ≥5% (0.88 vs. 0.81), >33% (0.90 vs. 0.85) and >66% (0.84 vs. 0.72). For SCD <25 mm, the optimal CAP cut‐offs for differentiating steatosis ≥5%, >33% and >66% were 255.0 dB/m, 283.5 dB/m and 293.5 dB/m. However, cut‐offs were elevated by approximately 60–70 dB/m for SCD ≥25 mm. When stratified by fibrosis grade, LSM was significantly affected by SCD ≥25 mm for advanced fibrosis (≥F3) in NAFLD, but not in CHB.
Conclusion
CAP is a promising tool for detecting and quantifying hepatic steatosis. SCD ≥25 mm may cause overestimation of steatosis. Similarly, SCD ≥25 mm affects the detection of advanced fibrosis by LSM in NAFLD patients.
Keywords: controlled attenuation parameter, liver stiffness, non‐alcoholic fatty liver disease, skin capsular distance, transient elastography
Abbreviations
- BMI
body mass index
- CAP
controlled attenuation parameter
- CHB
chronic hepatitis B
- IQR
interquartile range
- LSM
liver stiffness measurement
- NAFLD
non‐alcoholic fatty liver disease
- ROC curves and AUROCs
receiver operating characteristics curves and areas under the curves
- SCD
skin capsular distance
- TE
transient elastography
Key points.
SCD is independently associated with CAP. The areas under receiver operating characteristics curves (AUROCs) are slightly higher for SCD <25 mm compared to those for SCD ≥25 mm.
For SCD <25 mm, the optimal CAP cut‐offs for differentiating steatosis ≥5%, >33% and >66% are 255.0 dB/m, 283.5 dB/m and 293.5 dB/m. However, cut‐offs were elevated by approximately 60–70 dB/m for SCD ≥25 mm.
SCD ≥25 mm affects the detection of advanced fibrosis by LSM in NAFLD patients.
A CAP algorithm for the obese population and developed specifically for the FibroScan XL probe would be welcome.
Liver stiffness measurement (LSM), using transient elastography (TE) by FibroScan, is a popular non‐invasive method for diagnosing liver fibrosis. LSM has a greater accuracy in predicting bridging fibrosis and cirrhosis compared with other biomarkers 1. Although LSM is widely used in cases of chronic liver disease, high body mass index (BMI) and obesity are independent risk factors for unreliable LSM results as well as LSM failure if performed using an M probe 2. This limitation is particularly relevant in obese patients with a thicker subcutaneous fat layer, such as those with non‐alcoholic fatty liver disease (NAFLD) 3, 4.
Recently, a novel physical parameter, ‘controlled attenuation parameter’ (CAP), has been implemented successfully based on TE and automatically estimates the ultrasonic attenuation of the liver. CAP has been shown to be highly correlated with liver steatosis grade 5. However, the failure rate of CAP measurement is increased in overweight and obese patients and can be as high as 58.4% among patients with BMI >40 kg/m2 6.The optimal cut‐off, for example, to discriminate the presence of significant hepatic steatosis (steatosis involving ≥10% of hepatocytes) range from 222 dB/m 7 to 283 dB/m 5. CAP might have suboptimal accuracy for detection of severe steatosis 8.
The layers of tissues that connect the skin to the liver capsule were defined as the skin capsular distance (SCD), and these layers are mainly comprised of subcutaneous fatty tissue. SCD may contribute to the beam distortion and scattering that is known to attenuate the propagation of shear waves into the liver and ultrasonic signals used to assess the speed of propagation 5, 9, 10. In patients with SCD thicker than 25 mm, the attenuation may be enhanced and lead to an overestimation of LSM, and the failure rate of the M probe was markedly increased to 33% 3. Although an XL probe has been designed for overweight and obese patients, it is not widely available 11.
Along with hepatic steatosis, subcutaneous fat can attenuate the ultrasonic signals, and thus high SCD may be an important potential barrier to the effective use of CAP. Currently, CAP measurement has only been implemented using an M probe and can still produce valid results if the SCD is ≥25 mm 5. However, the effects of SCD on both CAP and LSM have not been evaluated.
The objectives of this study were to assess the impact of SCD on the diagnostic performance of CAP and LSM, and to improve the accuracy of CAP in diagnosing hepatic steatosis in chronic liver diseases.
Patients and methods
Study population
A prospective multicenter study was carried out in five Chinese liver centres (Xinhua Hospital, Shanghai; Zhengxing Hospital, Fujian; The Affiliated Hospital of Hangzhou Normal University, Zhejiang; Tianjin Second People's Hospital, Tianjin; 302 Military Hospital, Beijing) between May 2012 and May 2014. Consecutive patients aged ≥18 years who underwent liver biopsies and met the diagnostic criteria for NAFLD 12 or chronic hepatitis B (CHB) 13 were prospectively recruited. We excluded patients with a history of excessive ethanol consumption (140 g per week for men, 70 g for women) in the past 12 months, as well as patients who refused to undergo liver biopsies and those with secondary causes of hepatic steatosis (such as the use of systemic corticosteroids), anti‐hepatitis C virus antibodies or common contraindications for TE (e.g. pregnancy, ascites, pacemaker implanted). The Ethics Committees of Xinhua Hospital approved this study, and all patients provided informed written consent prior to enrolment.
Clinical and laboratory assessment
Demographic information, including age, sex, history of drinking, hypertension and type 2 diabetes mellitus, was collected. Anthropometric measurements included BMI [weight (kg)/height (m)2], waist circumference (cm) and hip circumference (cm). Laboratory tests were performed within a week of liver biopsy. Hepatitis B virus DNA (for CHB patients only), prothrombin time, platelet counts, fasting plasma glucose, alanine aminotransferase, aspartate aminotransferase, albumin, total bilirubin, total cholesterol, triglycerides, high‐density lipoprotein cholesterol and low‐density lipoprotein cholesterol levels were measured at each hospital's laboratory, using a standard testing kit and techniques.
Controlled attenuation parameter and liver stiffness measurement
TE was performed within 1 week preceding liver biopsy. The operators had performed TE evaluations on at least 300 patients prior to this study. CAP is a new proprietary algorithm that estimates the total ultrasonic attenuation, which is expressed in dB/m, and is currently only implemented using the M probe. In this study, FibroScan‐502 (Echosens, Paris, France) with M probe (3.5 MHz) was used to capture CAP and LSM simultaneously, using the same TE signals. CAP values range from 100 to 400 dB/m, while LSM values are expressed as kilopascals (kPa) and range from 1.5 to 75 kPa. Measurements were performed on the right lobe of the liver, through the intercostal spaces, with the patient lying in the dorsal decubitus position and the right hand on the head to extend the intercostal space.
A reliable measurement of LSM was defined as obtaining 10 or more valid shots, a success rate of at least 60% and an interquartile range (IQR) <30% of the median LSM values. CAP was measured only on validated measurements, according to the same criteria used for LSM. If reliable measurements were not obtained after a minimum of 10 attempts, the results were defined as failed or unreliable and were excluded from our analysis.
Skin capsular distance measurement
The SCD includes the layers of tissues connecting the epidermis of the skin to the liver, and can contain various proportions of fatty tissue. A fixed‐position FibroScan‐502 with M probe was used, and a series of temporal ultrasound radiofrequency lines were acquired to measure the subcutaneous layer thickness (<25 mm and ≥25 mm), which was subsequently separated from the liver tissue using real‐time computation during TE scanning 9, 10. The value was based on the first measurement.
Liver histology
Percutaneous liver biopsy was performed using the 18‐gauge BARD Max‐Core Disposable Core Biopsy Instrument (BARD Biopsy Systems, Tempe, AZ, USA) at the right lobe, under real‐time ultrasound guidance. All samples were ≥16 mm, and included at least six portal tracts. Liver biopsy specimens were fixed in formalin, embedded in paraffin and stained with H&E, Masson's trichrome and reticulin. The liver biopsies were interpreted by three experienced hepatopathologists who were blinded to clinical data; discrepancies were resolved by consensus. For all samples, liver steatosis was categorized by visual assessment, according to the NAFLD activity score as S0 (<5%), S1 (5–33%), S2 (34–66%) or S3 (>66%) 14.
The grades and stages of the liver samples were dependent on the liver disease aetiology. The proposed NAFLD activity score ≥5 was considered indicative of non‐alcoholic steatohepatitis (14). Fibrosis of NAFLD was staged as follows: F0, absence of fibrosis; F1, perisinusoidal or periportal fibrosis; F2, perisinusoidal and portal/periportal fibrosis; F3, septal or bridging fibrosis; and F4, cirrhosis. The METAVIR classification was used for CHB patients 15. Given the different diagnostic criteria, METAVIR grades ≥2 in patients with CHB, and NAFLD activity scores ≥5, were classified as moderate to severe inflammation. Significant fibrosis or advanced fibrosis was defined as ≥F2 or ≥F3 respectively.
Statistical analyses
All continuous variables and patient characteristics were expressed as medians (interquartile range (IQR)) or n (%), as appropriate. Chi‐square or Fisher's exact tests were used to compare categorical data. The Mann‐Whitney U test and Kruskal–Wallis test was used to compare two or more groups. Correlations between CAP and continuous variables were assessed using Spearman correlation coefficients (ρ). Parameters that were significantly associated with CAP were subsequently entered into a multiple linear regression models (Stepwise method).
Receiver operating characteristic (ROC) curves were plotted, and areas under the curves (AUROCs) with 95% confidence interval (CI) were calculated to determine the diagnostic efficacy. The accuracy of CAP at optimal thresholds was defined by the maximum Youden index. For each optimal cut‐off value, the sensitivity, specificity, positive predictive values (PPVs) and negative predictive values (NPVs), positive likelihood ratio (PLR) and negative likelihood ratio (NLR) were calculated. The general coincidence rate was calculated as (true positive + true negative)/overall patients.
All statistical analyses were performed using spss version 16.0 for windows (SPSS, Chicago, IL, USA) and medcalc statistical software version 11.0 (MedCalc, Mariakerke, Belgium). Two‐sided P‐values <0.05 were considered statistically significant.
Results
Patients’ characteristics and histological characteristics
Four hundred and thirty‐four patients who underwent liver biopsy and TE using FibroScan were recruited. Twelve patients were excluded because of liver biopsy length <16 mm or insufficient portal tracts. Forty‐one of the remaining 422 patients (9.7%) were excluded because of failed or unreliable LSM and/or CAP values. Thus, the remaining 381 patients included 332 with SCD <25 mm and 49 with SCD ≥25 mm, who were evaluated in this study (Fig. S1).
The characteristics of this study population are outlined in Table 1 and histology features in Table S1. One hundred and one (26.5%) were NAFLD patients and 280 (73.5%) were CHB patients. The majority (70.3%) were men and the median age was 37.0 (IQR 29.0–46.0) years. In the SCD ≥25 mm group, male gender, the prevalence of hypertension and type 2 diabetes mellitus, BMI, waist circumference, hip circumference, albumin, low‐density lipoprotein cholesterol, measurements times, CAP value and rate of CAP/median were significant higher than that in the SCD <25 mm group (all P < 0.05). The median HBV‐DNA (log10) was 5.2 (IQR 3.0–7.2) IU/ml for CHB patients.
Table 1.
Baseline characteristics of study patients
| Characteristics | All patients (n = 381) | Skin capsular distance | P‐value | |
|---|---|---|---|---|
| <25 mm (n = 332) | ≥25 mm (n = 49) | |||
| Liver disease aetiology | ||||
| NAFLD, n (%) | 101 (26.5) | 67 (20.2) | 34 (69.4) | <0.001 |
| CHB, n (%) | 280 (73.5) | 265 (79.8) | 15 (30.6) | |
| Demographics | ||||
| Male gender, n (%) | 268 (70.3) | 228 (68.7) | 40 (81.6) | 0.042 |
| Age (years) | 37.0 (29.0–46.0) | 37.0 (29.0–45.8) | 38.0 (30.0–48.5) | 0.512 |
| Hypertension, n (%) | 36 (9.4) | 30 (9.0) | 6 (12.2) | 0.046 |
| Type 2 diabetes mellitus, n (%) | 27 (7.1) | 20 (6.0) | 7 (14.4) | 0.041 |
| Anthropometrics | ||||
| Body mass index (kg/m2) | 24.8 (22.3–27.5) | 24.3 (21.9–26.5) | 27.7 (25.1–29.6) | <0.001 |
| <18.5 | 17 (4.5) | 16 (4.8) | 1 (2.0) | |
| 18.5–24.9 | 192 (50.4) | 181 (54.5) | 11 (22.4) | |
| 25.0–29.9 | 137 (36.0) | 110 (33.1) | 27 (55.1) | |
| ≥30 | 35 (9.2) | 25 (7.5) | 10 (20.4) | |
| Waist circumference (cm) | 89.0 (82.0–95.0) | 88.0 (81.0–95.0) | 92.5 (88.3–100.0) | 0.001 |
| Hip circumference (cm) | 96.0 (92.0–102.0) | 95.0 (91.0–101.0) | 101.0 (96.0–105.0) | <0.001 |
| Biochemistry | ||||
| Platelets (×109/L) | 190.0 (156.0–228.0) | 188.5 (154.8–226.3) | 208.5 (164.3–234.5) | 0.092 |
| Fasting blood glucose (mmol/L) | 5.1 (4.7–5.7) | 5.2 (4.7–5.7) | 5.0 (4.6–6.1) | 0.845 |
| Alanine aminotransferase (U/L) | 44.2 (23.1–74.1) | 43.0 (23.1–77.2) | 48.1 (26.9–68.9) | 0.971 |
| Aspartate aminotransferase (U/L) | 33.0 (23.2–58.5) | 34.0 (23.1–58.0) | 33.0 (21.2–47.0) | 0.358 |
| Albumin (g/L) | 44.0 (41.0–46.4) | 44.0 (41.0–46.3) | 44.1 (42.4–47.4) | 0.041 |
| Total bilirubin (μmol/L) | 14.0 (10.6–18.3) | 13.7 (10.6–18.3) | 14.6 (10.8–18.4) | 0.834 |
| Prothrombin time (s) | 11.9 (11.0–12.6) | 11.9 (11.1–12.7) | 11.6 (10.7–12.3) | 0.057 |
| Total cholesterol (mmol/L) | 4.9 (4.2–5.6) | 4.9 (4.1–5.6) | 5.0 (4.3–5.4) | 0.456 |
| Triglycerides (mmol/L) | 1.3 (0.9–2.0) | 1.2 (0.9–2.0) | 1.6 (1.0–2.1) | 0.064 |
| High‐density lipoprotein cholesterol (mmol/L) | 1.3 (1.1–1.5) | 1.3 (1.1–1.5) | 1.3 (1.1–1.5) | 0.846 |
| Low‐density lipoprotein cholesterol (mmol/L) | 2.8 (2.3–3.4) | 2.8 (2.2–3.3) | 2.9 (2.7–3.7) | 0.009 |
| Transient elastography | ||||
| Success rate | 99.0 (98.0–100.0) | 100 (100–100) | 96 (91.5–99.0) | 0.133 |
| Time (s) | 99.0 (80.0–141.5) | 97.0 (79.0–137.3) | 131.0 (91.0–201.5) | 0.001 |
| CAP (dB/m) | 254.0 (210.0–305.0) | 241.0 (206.0–290.8) | 345.0 (308.0–360.5) | <0.001 |
| Interquartile range /medianCAP | 12.3% (8.2–18.4%) | 9.1% (6.0–12.3%) | 13.7% (8.9–18.5%) | <0.001 |
| LSM (kPa) | 7.6 (5.5–11.8) | 7.4 (5.5–11.5) | 8.7 (5.7–11.8) | 0.183 |
| Interquartile range /medianLSM | 13.8% (9.2–18.5%) | 13.8% (8.5–19.2%) | 13.9% (9.9–18.7%) | 0.531 |
All data are expressed as medians (interquartile range), or n (%), as appropriate. BMI, high body mass index; CAP, controlled attenuation parameter; LSM, liver stiffness measurement.
The median specimen length of liver biopsy was 20.0 mm (IQR 17.0–21.5), and the median number of portal tracts was 10 (IQR 8–15).
Factors associated with controlled attenuation parameter values
The median CAP measurements for NAFLD patients and CHB patients in S1, S2 and S3 were 273.0 dB/m and 273.0 dB/m, 327.0 dB/m and 335.5 dB/m and 346.5 dB/m and 309.0 dB/m respectively. There was no significant difference between the NAFLD and CHB group for each hepatic steatosis grade (all P > 0.05) (Fig. S2). CAP values were strongly correlated with steatosis grade (ρ = 0.71, P < 0.001) and moderately correlated with high SCD (ρ = 0.44, P < 0.001). Old age, male sex, high BMI, waist circumference, hip circumference, platelet count, fasting blood glucose, aspartate aminotransferase, albumin, total bilirubin, total cholesterol, triglyceride, low‐density lipoprotein cholesterol levels and moderate to severe inflammation were also significantly associated with CAP values by univariate analysis. Multiple linear regression (Stepwise method) revealed that SCD (B 30.34, 95% CI 21.18–39.49, P < 0.001) and hepatic steatosis (B 23.04, 95% CI 19.74–34.34, P < 0.001) were independently positively associated with CAP values (Table S2).
The impact of skin‐liver capsule on the diagnostic performance of controlled attenuation parameter
Median CAP values significantly increased with steatosis grade for both SCD <25 mm and SCD ≥25 mm (all P < 0.001, Table 2). After adjustment, CAP values were significant higher in patients with SCD ≥25 mm compared to those with SCD <25 mm, for the same steatosis grade (Table 2).
Table 2.
Distribution of CAP values for different steatosis grades following stratification of SCD
| Steatosis grade | CAP (dB/m) | P a | ||
|---|---|---|---|---|
| Total subjects (n = 381) | SCD <25 mm (n = 332) | SCD ≥25 mm (n = 49) | ||
| S0 (<5%) | 216.5 (191.8–246.5) | 215.0 (190.0–241.0) | 303.0 (266.5–334.0) | <0.001 |
| S1 (5–33%) | 273.0 (248.5–304.5) | 269.0 (244.5–298.5) | 318.0 (280.5–341.8)d | <0.001 |
| S2 (34–66%) | 331.0 (301.5–359.0) | 315.0 (293.0–349.0) | 350.5 (341.8–366.3) | 0.001 |
| S3 (>66%)c | 344.0 (308.0–359.0) | 318.0 (298.0–346.5) | 362.5 (342.3–375.8) | 0.009 |
| P b | <0.001 | <0.001 | <0.001 | |
Variables were given as medians (interquartile range).
Comparison between CAP in skin capsular distance <25 mm and ≥25 mm, within the same steatosis grade.
Correlation between CAP and the grade of hepatic steatosis.
P > 0.05, S3 vs. S2.
P > 0.05, S1 vs. S0.
To evaluate the ability of CAP to discriminate the grading of hepatic steatosis, ROC curves and the corresponding AUROCs and 95% CIs were calculated. Among subjects with SCD <25 mm or SCD ≥25 mm, the AUROCs for steatosis grades of ≥5%, >33% and >66% were 0.88 and 0.81 (P = 0.406), 0.90 and 0.85 (P = 0.381) and 0.84 and 0.72 (P = 0.303) respectively (Fig. 1).
Figure 1.
Receiver operating characteristic curves and area under the curves (AUROCs) for the detection of steatosis grades ≥5%, >33% and >66% for skin‐liver capsule (SCD) distance <25 mm (A) and SCD ≥25 mm (B).
The optimal cut‐off values of controlled attenuation parameter for estimation of hepatic steatosis
The optimal CAP cut‐offs for the diagnosis of hepatic steatosis grade, stratified by SCD are outlined in Table 3. At each steatosis grade, the optimal CAP cut‐offs were much higher for SCD >25 mm compared to those for SCD <25 mm. Results for the diagnostic accuracy of CAP are also outlined in Table 3.
Table 3.
Diagnostic performance of controlled attenuation parameter (CAP) for the detection of steatosis grade
| Steatosis | SCD (mm) | Optimal cut‐offa (dB/m) | Sensitivity (95% CI) | Specificity (95% CI) | PPV (95% CI) | NPV (95% CI) | PLR (95% CI) | NLR (95% CI) |
|---|---|---|---|---|---|---|---|---|
| S0 vs. S1–3 | <25 | 255.0 | 0.75 (0.67–0.82) | 0.82 (0.75–0.87) | 0.75 (0.67–0.82) | 0.82 (0.75–0.87) | 4.10 (3.00–5.62) | 0.30 (0.23–0.41) |
| ≥25 | 326.5 | 0.71 (0.55–0.83) | 0.80 (0.30–0.99) | 0.97 (0.82–0.99) | 0.24 (0.09–0.50) | 3.52 (0.60–20.55) | 0.37 (0.21–0.64) | |
| S0–1 vs. S2–3 | <25 | 283.5 | 0.86 (0.74–0.93) | 0.86 (0.81–0.90) | 0.60 (0.49–0.70) | 0.96 (0.93–0.98) | 6.23 (4.54–8.53) | 0.16 (0.09–0.30) |
| ≥25 | 344.0 | 0.79 (0.59–0.91) | 0.86 (0.63–0.96) | 0.88 (0.68–0.97) | 0.75 (0.53–0.89) | 5.50 (1.90–15.96) | 0.25 (0.12–0.52) | |
| S0–2 vs. S3 | <25 | 293.5 | 0.88 (0.62–0.98) | 0.81 (0.76–0.85) | 0.20 (0.12–0.31) | 0.99 (0.97–0.99) | 4.56 (3.43–6.06) | 0.15 (0.04–0.54) |
| ≥25 | 350.5 | 0.80 (0.44–0.97) | 0.72 (0.55–0.85) | 0.42 (0.21–0.66) | 0.93 (0.77–0.99) | 2.84 (1.57–5.11) | 0.28 (0.08–0.98) |
95% CI, 95% confidence interval; PPV and NPV, positive and negative predictive values; PLR and NLR, positive and negative likelihood ratio; SCD, skin capsular distance.
Optimal cut‐offs was defined by the maximal sum of sensitivity and specificity.
The general coincidence rate was used to determine the correct rate. For SCD <25 mm, the general coincidence rate of CAP discriminating S1, S2 and S3 was 81.6%, 86.1% and 81.0% respectively. For SCD ≥25 mm, the general coincidence rate was 63.3%, 81.6% and 73.5% respectively.
The impact of skin capsular distance on the diagnostic performance of liver stiffness measurement
There was no significant correlation between SCD and LSM values in NAFLD subjects (ρ = 0.15, P = 0.144) or CHB subjects (ρ = 0.06, P = 0.343). We also evaluated fibrosis stage by dividing cases into significant fibrosis (≥F2) and advanced fibrosis (≥F3). Among subjects without significant fibrosis (F0–F1) or with significant fibrosis (F2–F4), there was a trend where LSM values were slightly higher for SCD ≥25 mm compared to those for SCD <25 mm among both NAFLD (Fig. 2A) and CHB subjects (Fig. 2B).
Figure 2.
Distribution of liver stiffness measurement (LSM) values for skin‐liver capsule distance (SCD) <25 mm or ≥25 mm, according to (A) non‐alcoholic fatty liver disease (NAFLD) and (B) chronic hepatitis B (CHB) with or without significant fibrosis and (C) NAFLD and (D) CHB with and without advanced fibrosis. ●Outlier: any patient result lying outside the upper or lower whiskers.
Similar results were obtained for subjects without advanced fibrosis (F0–F2) among NAFLD (Fig. 2C) or CHB (Fig. 2D) subjects. However, the only statistically significant difference in LSM was detected among NAFLD patients with advanced fibrosis (F3–F4), where the median LSM value was 18.5 kPa (IQR,12.8–23.8) for SCD ≥25 mm, which was significant higher than that for SCD <25 mm (median 8.8 kPa; IQR, 5.3–10.8; P = 0.045) (Fig. 2C).
For NAFLD subjects, the optimal LSM cut‐off for the diagnosis of advanced fibrosis varied according to SCD. For SCD <25 mm, an optimal cut‐off of 8.7 kPa was appropriate, with an AUROC of 0.62 (95% CI 0.52–0.84), sensitivity of 0.71 (95% CI 0.60–0.95) and specificity of 0.77 (95% CI 0.58–0.82). For SCD ≥25 mm, an optimal cut‐off of 12.9 kPa was appropriate, with an AUROC of 0.88 (95% CI 0.70–0.99), sensitivity of 0.83 (95% CI 0.37–0.99) and specificity of 0.89 (95% CI 0.71–0.97).
Discussion
In this multicenter study, we found that CAP could be used to detect hepatic steatosis with good diagnostic accuracy, regardless of aetiology, as has been previously reported 16. However, SCD was an independently factor influence the CAP values and always overestimated for SCD ≥25 mm. Although LSM was slightly higher for SCD ≥25 mm in CHB subjects after adjusting for the liver fibrosis stage, this result was not statistically significant. In contrast, LSM was significantly affected by SCD in advanced fibrosis (≥F3) among NAFLD subjects.
Liver biopsy is prone to sampling error, higher risk of complications and is difficult to repeat; hence, non‐invasive tests are urgently needed. One such method is FibroScan, which can measure the ultrasound attenuation of hepatic fat, and can simultaneously evaluate fibrosis (by LSM). However, the accuracy of this widely used tool has been challenged. LSM may be limited by various factors, including the cause of chronic hepatitis, acute liver parenchymal injury, central venous pressure or extrahepatic cholestasis 17, 18, 19. Higher BMI and SCD can also lead to unreliable or failed LSM measurements with the M probe 3, 4.
A recent study has reported that the diagnosis accuracy of CAP is higher than that of the Hepatic Steatosis Index and the Fatty Liver Index 5 and is even more accurate than 1H‐magnetic resonance spectroscopy 20. However, in a small sample study, we have reported that BMI was the most relevant factor that influenced CAP 21. In the present study, we observed that SCD ≥25 mm had a more significant impact on CAP than BMI. As reported by Myers et al. 5, who noted that the AUROC for differentiating steatosis ≥10% was higher among patients with an SCD <25 mm (0.82 vs. 0.69). Our study observed that the CAP performance was good to excellent in differentiating steatosis for SCD <25 mm, and poor for SCD ≥25 mm. Among reliable measurements, CAP was grossly overestimated in subjects with SCD ≥25 mm and this highlights the need for accurate CAP cut‐off values. Thus, we recommend considering SCD thickness prior to selecting the cut‐off values for CAP when evaluating hepatic steatosis.
For SCD <25 mm, we selected 255.0 dB/m, 283.5 dB/m and 293.5 dB/m as optimal cut‐offs to identify steatosis of ≥5%, >33% and >66% respectively. Our optimal CAP cut‐off value for the diagnosis of hepatic steatosis ≥5% is similar to previously reported values, including 250 dB/m 8, 22 and 263 dB/m 23. However, Myers et al. 5 have proposed a higher cut‐off values (283 dB/m) for the differentiation of steatosis ≥10%. In that study, patients had an BMI of ≥28 kg/m2 and 16% of subjects with SCD ≥25 mm, which might have resulted in an overestimated cut‐off. For patients with SCD ≥25 mm, we observed that the reference cut‐off values were elevated by approximately 60–70 dB/m. We also note that the M probe used for these patients may be unsuitable, as the reference values were too similar to distinguish the degree of fatty liver.
Different LSM cut‐off values have been reported in previous studies of patients with chronic liver disease of various aetiologies 24 and it remains unclear whether SCD affects LSM. In the present study, we observed that NAFLD patients with SCD ≥25 mm had increased LSM values, especially in advanced fibrosis (≥F3). Unlike CAP, LSM was not significantly influenced by SCD in CHB patients. When evaluating NAFLD patients using FibroScan with M probe, we propose that the optimal LSM cut‐offs for the diagnosis of advanced fibrosis (≥F3) is 8.7 kPa for SCD <25 mm, which is similarly to previously reported values of 8.0 kPa 25 and 8.7 kPa 26. Although higher AUROC was observed for SCD ≥25 mm, the risk of misclassification of advanced fibrosis in NAFLD patients is increased, and thus, the use of the XL probe is recommended.
Although fat‐mimicking materials affect the shear waves 10, SCD seems not to have a significant effect on LSM values. However, SCD does affect ultrasonic attenuation, and can lead to overestimation of CAP values. As the M probe is still the only measurement tool for CAP, the current optimal cut‐off values are unsuitable for patients with SCD ≥25 mm.
Our study has several limitations. First, chronic liver disease patients, including those with chronic hepatitis C (CHC), were not included in our study. However, as TE is validated in CHC, there is limited data available for CHB and NAFLD patients 25. For CHB, hepatic steatosis is typically associated with obesity, rather than CHC.
Second, our study did not analyse the factors that caused the failure of CAP or LSM measurements, as the primary aim of this study was to determine whether SCD affected the reliability of TE. We also note that such analysis would be complicated by the small number of patients that experience failed measurements. Third, we did not compare our results to those obtained using other imaging modalities (e.g. ultrasonography), given their low sensitivity 5, 27. Last, we propose new cut‐off values for CAP, but it is needed more validation cohort.
In conclusion, CAP presents several diagnostic advantages (simplicity, operator‐independence, sensitivity and immediate assessment of steatosis) when used simultaneously with LSM to stage hepatic fibrosis. Given the limitations of the M probe, CAP is appropriate for patients with SCD <25 mm, and not for those with SCD ≥25 mm unless using SCD‐adjusted cut‐off value. Moreover, the SCD‐modified cut‐off value of LSM might be used for diagnosis of advanced fibrosis in NAFLD patients with larger SCD. As NAFLD is strongly associated with obesity, a CAP algorithm for the obese population and developed specifically for the FibroScan XL probe, would be welcome.
Supporting information
Additional Supporting Information may be found at onlinelibrary.wiley.com/doi/10.1111/liv.12809/suppinfo
Fig. S1. Flow chart of the study population.
Fig. S2. Controlled attenuation parameter distribution for different grades of histological steatosis or patients with chronic hepatitis B and nonalcoholic fatty liver disease.
Table S1. Histology features of study patients
Table S2. Univariate and multivariate regression analyses for CAP with different parameters in NAFLD and CHB groups
Acknowledgements
Financial support: The National Key Basic Research Project, No. 2012CB517501; Chinese Foundation for Hepatitis Prevention and Control – ‘WANG Bao‐En’ Liver Fibrosis Research Fund, No. XJS20120501; ‘Tian‐qing‐gan‐bing’ Research Fund, No. 20120027.
Conflict of interest: The author do not have any disclosures to report.
Liver Int. 2015; 35: 2392–2400
References
- 1. Degos F, Perez P, Roche B, et al Diagnostic accuracy of FibroScan and comparison to liver fibrosis biomarkers in chronic viral hepatitis: a multicenter prospective study (the FIBROSTIC study). J Hepatol 2010; 53: 1013–21. [DOI] [PubMed] [Google Scholar]
- 2. Wong GL, Wong VW, Chim AM, et al Factors associated with unreliable liver stiffness measurement and its failure with transient elastography in the Chinese population. J Gastroenterol Hepatol 2011; 26: 300–5. [DOI] [PubMed] [Google Scholar]
- 3. Myers RP, Pomier‐Layrargues G, Kirsch R, et al Feasibility and diagnostic performance of the FibroScan XL probe for liver stiffness measurement in overweight and obese patients. Hepatology 2012; 55: 199–208. [DOI] [PubMed] [Google Scholar]
- 4. de Ledinghen V, Vergniol J, Foucher J, et al Feasibility of liver transient elastography with FibroScan using a new probe for obese patients. Liver Int 2010; 30: 1043–8. [DOI] [PubMed] [Google Scholar]
- 5. Myers RP, Pollett A, Kirsch R, et al Controlled Attenuation Parameter (CAP): a noninvasive method for the detection of hepatic steatosis based on transient elastography. Liver Int 2012; 32: 902–10. [DOI] [PubMed] [Google Scholar]
- 6. de Ledinghen V, Vergniol J, Capdepont M, et al Controlled attenuation parameter (CAP) for the diagnosis of steatosis: a prospective study of 5323 examinations. J Hepatol 2014; 60: 1026–31. [DOI] [PubMed] [Google Scholar]
- 7. Sasso M, Tengher‐Barna I, Ziol M, et al Novel controlled attenuation parameter for noninvasive assessment of steatosis using Fibroscan((R)): validation in chronic hepatitis C. J Viral Hepat 2012; 19: 244–53. [DOI] [PubMed] [Google Scholar]
- 8. Chon YE, Jung KS, Kim SU, et al Controlled attenuation parameter (CAP) for detection of hepatic steatosis in patients with chronic liver diseases: a prospective study of a native Korean population. Liver Int 2013; 34: 102–9. [DOI] [PubMed] [Google Scholar]
- 9. Audiere S, Charbit M, Angelini ED, Oudry J, Sandrin L. Measurement of the skin‐liver capsule distance on ultrasound RF data for 1D transient elastography. Med Image Comput Comput Assist Interv 2010; 13: 34–41. [DOI] [PubMed] [Google Scholar]
- 10. Cournane S, Browne JE, Fagan AJ. The effects of fatty deposits on the accuracy of the Fibroscan(R) liver transient elastography ultrasound system. Phys Med Biol 2012; 57: 3901–14. [DOI] [PubMed] [Google Scholar]
- 11. Wong VW, Vergniol J, Wong GL, et al Liver stiffness measurement using XL probe in patients with nonalcoholic fatty liver disease. Am J Gastroenterol 2012; 107: 1862–71. [DOI] [PubMed] [Google Scholar]
- 12. Fan JG, Jia JD, Li YM, et al Guidelines for the diagnosis and management of nonalcoholic fatty liver disease: update 2010: (published in Chinese on Chinese Journal of Hepatology 2010; 18:163‐166). J Dig Dis 2011; 12: 38–44. [DOI] [PubMed] [Google Scholar]
- 13. Shi JP, Fan JG, Wu R, et al Prevalence and risk factors of hepatic steatosis and its impact on liver injury in Chinese patients with chronic hepatitis B infection. J Gastroenterol Hepatol 2008; 23: 1419–25. [DOI] [PubMed] [Google Scholar]
- 14. Kleiner DE, Brunt EM, Van Natta M, et al Design and validation of a histological scoring system for nonalcoholic fatty liver disease. Hepatology 2005; 41: 1313–21. [DOI] [PubMed] [Google Scholar]
- 15. Bedossa P, Poynard T. An algorithm for the grading of activity in chronic hepatitis C. The METAVIR Cooperative Study Group. Hepatology 1996; 24: 289–93. [DOI] [PubMed] [Google Scholar]
- 16. Kumar M, Rastogi A, Singh T, et al Controlled attenuation parameter for non‐invasive assessment of hepatic steatosis: does aetiology affect performance. J Gastroenterol Hepatol 2013; 28: 1194–201. [DOI] [PubMed] [Google Scholar]
- 17. Vigano M, Massironi S, Lampertico P, et al Transient elastography assessment of the liver stiffness dynamics during acute hepatitis B. Eur J Gastroenterol Hepatol 2010; 22: 180–4. [DOI] [PubMed] [Google Scholar]
- 18. Millonig G, Friedrich S, Adolf S, et al Liver stiffness is directly influenced by central venous pressure. J Hepatol 2010; 52: 206–10. [DOI] [PubMed] [Google Scholar]
- 19. Millonig G, Reimann FM, Friedrich S, et al Extrahepatic cholestasis increases liver stiffness (FibroScan) irrespective of fibrosis. Hepatology 2008; 48: 1718–23. [DOI] [PubMed] [Google Scholar]
- 20. Karlas T, Petroff D, Garnov N, et al Non‐invasive assessment of hepatic steatosis in patients with NAFLD using controlled attenuation parameter and 1H‐MR spectroscopy. PLoS ONE 2014; 9: e91987. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. Shen F, Zheng RD, Mi YQ, et al Controlled attenuation parameter for non‐invasive assessment of hepatic steatosis in Chinese patients. World J Gastroenterol 2014; 20: 4702–11. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22. Jung KS, Kim BK, Kim SU, et al Factors affecting the accuracy of controlled attenuation parameter (CAP) in assessing hepatic steatosis in patients with chronic liver disease. PLoS ONE 2014; 9: e98689. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23. Chan WK, Nik MNR, Mahadeva S. Controlled attenuation parameter for the detection and quantification of hepatic steatosis in nonalcoholic fatty liver disease. J Gastroenterol Hepatol 2014; 29: 1470–6. [DOI] [PubMed] [Google Scholar]
- 24. Verveer C, Zondervan PE, ten Kate FJ, et al Evaluation of transient elastography for fibrosis assessment compared with large biopsies in chronic hepatitis B and C. Liver Int 2012; 32: 622–8. [DOI] [PubMed] [Google Scholar]
- 25. Gaia S, Carenzi S, Barilli AL, et al Reliability of transient elastography for the detection of fibrosis in non‐alcoholic fatty liver disease and chronic viral hepatitis. J Hepatol 2011; 54: 64–71. [DOI] [PubMed] [Google Scholar]
- 26. Wong VW, Vergniol J, Wong GL, et al Diagnosis of fibrosis and cirrhosis using liver stiffness measurement in nonalcoholic fatty liver disease. Hepatology 2010; 51: 454–62. [DOI] [PubMed] [Google Scholar]
- 27. Dasarathy S, Dasarathy J, Khiyami A, et al Validity of real time ultrasound in the diagnosis of hepatic steatosis: a prospective study. J Hepatol 2009; 51: 1061–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Additional Supporting Information may be found at onlinelibrary.wiley.com/doi/10.1111/liv.12809/suppinfo
Fig. S1. Flow chart of the study population.
Fig. S2. Controlled attenuation parameter distribution for different grades of histological steatosis or patients with chronic hepatitis B and nonalcoholic fatty liver disease.
Table S1. Histology features of study patients
Table S2. Univariate and multivariate regression analyses for CAP with different parameters in NAFLD and CHB groups