Comparison between shear wave elastography and serological findings for the evaluation of fibrosis in chronic liver disease

Liver cirrhosis is a major cause of death and disability globally⁽¹⁾. Studies have shown that 6–7% of the adult population without a known liver disease have liver fibrosis, typically associated with non-alcoholic fatty liver disease^(2,3). In this study, the following stages of fibrosis were defined: F0–F1: <7, F2: 7–8.9, F3: 9–11.9, and F4: ≥12 kPa, as these cutoff values have been adopted by the Greek National Insurance Program⁽⁴⁾. Those with no risk factors had only a 0.4% prevalence of significant liver fibrosis.

Traditionally, liver biopsy was considered as the ‘gold standard’ in the identification and examining of liver fibrosis and cirrhosis⁽⁵⁾. Liver histological scoring systems including Ishak, Knodell, Sheuer and METAVIR are used to assess liver architecture and fibrosis⁽⁶⁾. However, the technique is constrained by its invasiveness, risk of complications, and high costs⁽⁷⁾. These confines of liver biopsy have prompted research to identify noninvasive methods of evaluating the stages of liver fibrosis.

Noninvasive liver tests (NILTs) can generally be divided into three categories: simple or indirect serum markers, direct serum markers, and imaging modalities⁽⁸⁾. The most widely used imaging modality is transient elastography (TE) or FibroScan (Echosens, Paris)⁽⁴⁾. Later in the mid-1990s, elastography was introduced for the evaluation of stiffness and elasticity of soft tissues by giving external pressure⁽⁹⁾. It is an alternative technique to biopsy, as it is both safe and noninvasive⁽¹⁰⁾.

Elastography can be done with an ultrasonic transducer in combination with shear wave techniques such as TE, point shear wave elastography (pSWE), and two-dimensional shear wave elastography^(11,12). The main restriction of TE in clinical practice is the high amount of uninterpretable outcomes⁽¹³⁾. Shear wave elastography (SWE) is an innovative method that is based on shear waves applied on a diagnostic ultrasound system⁽¹⁴⁾. SWE has the benefit of being able to image liver stiffness in real time because the shear waves are generated by ultrasonographic pulse pushes. Real-time imaging is used, so that masses and large vessels can be identified and avoided. Moreover, the SWE image is directed by a higher frame-rate B-mode image. The method could provide a more accurate score of fibrosis staging resulting from the SWE and B-mode image direction⁽¹⁵⁾. Several serological tests have been established to identify liver fibrosis. The aspartate aminotransferase to platelet ratio index (APRI) and fibrosis-4 index (FIB-4) scoring are the most extensively used compound substitutes for detecting progressive fibrosis⁽¹⁶⁾. Uroš Karić et al. concluded in their study that FIB-4 was superior to APRI in distinguishing severe fibrosis. FIB-4 has been found to be very useful in identifying patients without advanced liver disease, particularly if other noninvasive methods are unavailable⁽¹⁷⁾.

We conducted this study to thoroughly compare the performance of ultrasound-based SWE with routine serological markers, the APRI and FIB-4, for evaluating liver fibrosis in patients with chronic liver disease.

This descriptive, cross-sectional study was performed using a convenient, non-probability sampling technique in the Radiology Department of Shaukat Khanum Hospital, Lahore. In the study, we planned to evaluate the images and medical records for SWE and serological findings to compare the extent of liver fibrosis, based on these two techniques (SWE and serological findings). According to the study design, we included all patients who underwent ultrasound SWE in the period from 1 June 2019 until 1 June 2020. Patients after liver transplant, technically unfit, i.e. severely obese or unable to lie flat on their back, and those with fluid build-up in the abdomen (ascites) were excluded from the study based on adopted exclusion criteria. Toshiba (Aplio 400) unit with 5 MHz broadband (C5-1) curved array transducer was used. The procedure was performed under the supervision of a radiologist familiar with US-SWE techniques. To examine the liver, the regions of interests (ROIs) were placed at least 1 cm underneath the liver capsule to avoid reverberation artifacts. The ROI sample box was a small area with a fixed stature of 12 mm that could move up to 8 cm deep from the skin surface; the size of the box was modified automatically from 5 mm near the surface of the transducer to 9.3 mm at a depth of 8 cm. The sample box was positioned with care to elude vascular structures. The rate of the generated SW (m/s) is calculated by observing tissue disarticulation over time. The measured velocities could afterwards be transformed to measure stiffness (kPa) using two constants, Young modulus, and tissue density.

The ultrasound examination was done either with light breathing or with a short breath-hold. Individually, a total of 15 distinct velocity measurements were achieved in patients across the liver using sub-costal and/or intercostal approaches, 10 in the right lobe and 5 in the left lobe. Most measurements were taken at a depth of 3–5 cm. The entire US (SWE) examination took 10–15 minutes to perform, and there was no need for sedation. There were dietary restrictions of about 8 to 12 hours prior to each examination.

The patients to be included in the study were identified through the health information system. The FIB-4 score was determined using the following formula: FIB-4 = (age × aspartate aminotransferase (AST)) ÷ (platelet count × (√ alanine transaminase (ALT)), and the APRI score was calculated with the formula: APRI = (AST÷AST upper limit of normal) ÷ platelet × 100. The overall concordance and disagreement between serological markers and SWE were analyzed.

Data analysis was performed with the help of Statistical Package for the Social Sciences version 24.0, and Microsoft Excel 2013. Qualitative data, e.g. gender, SWE (ordinal), was presented in the form of frequencies and their respective percentages. Demographic data, e.g. age, AST, ALT, platelet count, elastographic values, and the APRI/FIB-4 score, were expressed in the form of mean ± standard deviation. A ROC curve was generated to show the connection/trade-off between clinical sensitivity and specificity for every possible cutoff in liver fibrosis between SWE and serological findings.

The guidelines and principles established by the ethics committee were followed while conducting the study, and the rights of the participants were duly respected. All information and data collection were kept confidential. The participants were informed that there were no shortcomings or risks associated with the procedure of the study. The subjects were not exposed to any harm or danger. They were also notified that they were free to withdraw at any time throughout the course of the study.

A total 80 individuals were conveniently selected for the study, of which 50 (62.5%) were men, and 30 (37.5%) were women. The mean age of the subjects was 43.47 SD ± 13.85 years, with the minimum age being 5.0 and the maximum being 70.0 years. The mean values of ALT, AST, and platelet count APRI, FIB-4, elastographic mean and elastographic median value with their maximum and minimum values were also noted, i.e. 61.34 SD ± 86.23 (min. 5.0, max. 485.0), 56.25 SD ± 52.2 (min. 14.0, max. 343.0), 239.0 SD ± 113.88 (min. 12.0, max. 556.0), 0.77 SD ± 0.837 (min. 0.104, max. 5.417), 2.288 SD ± 3.42 (min. 0.06, max. 26.09), 17.8 SD ± 17.89 (min. 4.3, max. 109.7), and 18.1 SD ± 18.0 (min. 4.4, max. 106.9), respectively. A comparison of the descriptive variables of APRI and FIB-4 for 5 stages of liver fibrosis is shown in (Tab. 1, Tab. 2). For the implementation of APRI score in the likelihood of F4 patients consistent with SWE, we assumed F0, F1, F2 and F3 as one group, and F4 as another group; then the area under the receiver operator characteristic curve (AUROC) was 0.74 (95% CI 0.63–0.85; p <0.001). With the optimal APRI cutoff value of >0.46, we found F4 fibrosis having a sensitivity of 0.72 (72%) and specificity of 0.70 (70%) (Fig. 1). For the implementation of FIB-4 score in the likelihood of F4 patients consistent with SWE, we assumed F0, F1, F2 and F3 as one group, and F4 as another group; then AUROC was 0.795 (95% CI 0.698–0.89; p <0.001) With the optimal FIB-4 cutoff value of >1.27, we found F4 fibrosis having a sensitivity of 0.78 (78%) and specificity of 0.73 (73%) (Fig. 2). For the implementation of APRI score in the likelihood of F3–F4 patients consistent with SWE, we assumed F0, F1 and F2 as one group, and F3 and F4 in another group; then AUROC was 0.78 (95% CI 0.67–0.88; p< 0.001). With the optimal APRI cutoff value of >0.46, we found F4 fibrosis having a sensitivity of 0.74 (74%) and specificity of 0.76 (76%) (Fig. 3). For the implementation of FIB-4 score in the likelihood of F3–F4 patients consistent with SWE, we assume F0, F1 and F2 as one group, and F3–F4 in another group; then AUROC was 0.80 (95% CI 0.70–0.90; p <0.001). With the optimal FIB-4 cutoff value of >1.27, we found F4 fibrosis having a sensitivity of 0.77 (77%) and specificity of 0.76 (76%) (Fig. 4). For the implementation of APRI score in the likelihood of F1–F4 patients consistent with SWE, we assumed F0 as one group, and F1 F2, F3 and F4 as another group; then AUROC was 0.70 (95% CI 0.54–0.86; p <0.001). With the optimal APRI cutoff value of >0.34, we found F4 fibrosis having a sensitivity of 0.68 (68%) and specificity of 0.75 (75%) (Fig. 5).

Fig. 1.

Fig. 2.

Fig. 3.

Fig. 4.

Fig. 5.

The implementation of F4 score in the likelihood of F1–F4 patients consistent with SWE, if we assume F0 as one group, and F1 F2, F3 and F4 as another group, then AUROC was 0.72 (95% CI 0.56–0.87; p <0.001). With the selected FIB-4 optimal cutoff value of >0.87, we found F4 fibrosis having a sensitivity of 0.72 (72%) and specificity of 0.75 (75%) (Fig. 6).

Fig. 6.

Although liver biopsies are commonly used for investigative purposes, the method also has a number of limitations, such as being invasive and costly. Also, it may bring about sampling errors and inter-and intra-observer variations in considering hepatic fibrosis. Real-time SWE is an innovative, noninvasive practice to evaluate liver fibrosis by assessing liver stiffness (Fig. 7). These confines of the liver biopsy have encouraged research for noninvasive approaches in the assessment of liver fibrosis. SWE is an innovative practice that is grounded on shear waves implemented on an investigative ultrasound method. This technique could end result in a more precise score of fibrosis stages bring about the SWE and B-mode image direction⁽¹⁸⁾. We evaluated the diagnostic performance of the APRI and FIB-4 scores accompanying SWE in determining the stages of fibrosis (F0–F4). The main benefit of biochemical noninvasive scores (APRI and FIB-4) in considering liver fibrosis is that they are generally available at a low cost, and are very easy to perform. Though, SWE measurement is not far and wide existing owing to technical and practical field together with its unusual cost, on the other hand its use is not widespread in low- and mid-income nations^(18,19), whereas APRI and FIB-4 scores have been shown to be quite reliable for evaluating liver fibrosis⁽²⁰⁾. However, authentication in different patients is still required.

Fig. 7.

Liver biopsy has been extensively regarded as the gold standard for the assessment of liver fibrosis, though it has been nearly completely replaced by noninvasive approaches that measure liver stiffness (LS), such as transient elastography (TE)^(21,22), or biochemical markers and scoring systems^(18,23). In the present study, we compared two noninvasive techniques, SWE and serological findings, for the evaluation of fibrosis grading in chronic liver disease (CLD), and observed an agreement between SWE and serological findings (APRI and FIB-4 scores) for the estimation of fibrosis grading in CLD. A total of 80 individuals were evaluated. The associations of the patient’s individual characteristics at different fibrosis stages were assessed by using the one-way ANOVA or chi-square test, as suitable.

Nikolaos Papadopoulos et al. evaluated APRI/FIB-4 scores compared with TE-liver stiffness in detecting significant fibrosis or cirrhosis (F3 or F4). In that study, the authors retrospectively enrolled 575 patients with CHC who underwent TE-LS, and found that both scores projected F4 patients adequately. This also shows that FIB-4 is a suitable evaluation for ruling out noncirrhotic patients⁽²⁴⁾.

A pilot study was conducted in 2012 by Giovanna Ferraiolion et al. on real-time SWE for considering liver fibrosis in CH-C. The purpose of that study was to assess the diagnostic precision of real-time SWE in the evaluation of liver fibrosis in patients with chronic hepatic cirrhosis, in comparison with transient elastography, by using the histologic METAVIR classification as the reference system. In that study, real-time SWE measurements were compared with TE values in severe fibrosis and cirrhosis. Real-time SWE validated a significant improvement in the detection of significant fibrosis when compared with T-elastography⁽¹⁵⁾. Another study was done by Lun-Gen Lu et al. in 2003 on grading and staging of hepatic fibrosis and its correlation with noninvasive investigative considerations. The goal of that study was to see the sights of the grades and stages of pathology and also their relationship with hepatic fibrosis and noninvasive indicative factors. It was concluded that the categorizing and staging of liver fibrosis are interconnected with serum markers, Doppler ultrasound, computed tomography (CT) scan and/or magnetic resonance imaging (MRI). The combinations of the above-stated noninvasive factors were recognized to be relatively sensitive and specific in determining liver fibrosis. The sensitivity, specificity and accuracy values were 80.36%, 86.67%, and 81.10%, respectively⁽²⁵⁾. In our study we compared two noninvasive techniques, ultrasonographic SWE with two biomarkers i.e. APRI and FIB-4.

In the present study, we found different cutoff values for APRI and FIB-4 in different groups of fibrosis to distinguish their optimal cutoff values according to AUROC, and the diagnostic accuracies (sensitivity and specificity) of APRI and FIB-4 (normal AST level up-to 40 IU/L) for predicting the performance of APRI and FIB-4 accompanying ultrasound SW elastography. In a similar type of study, Yi-Hao Yen et al. in 2018 examined the optimum cutoff values of the two compound surrogates for envisaging cirrhosis by the AST level according to the AUROC analysis results differentiating cirrhotic (F4) from noncirrhotic (F0–F3). They concluded that the ideal cutoff values of both APRI and FIB-4 to predict cirrhosis graded by AST levels could be more practicable as compared with the single cutoff values offered in a foregoing research paper⁽²⁶⁾.

Conferring to former findings, APRI and FIB-4 were associated with the international normalized ratio, albumin level and necroinflammatory score^(27,28). Additionally, the positive correlations of APRI and FIB-4 with necroinflammatory score also kept our theory that the use of APRI and FIB-4 causes a possibility of overrating the fibrosis stage due to the influence of necroinflammatory activity on transaminases^(29,30) and the indicative precision of FIB-4 foreseeing liver fibrosis was found to be equivalent to or superior to that of APRI⁽³¹⁾. Even so, the objective of our study was to estimate the competence of the serological findings for the progressive fibrosis by comparing ultrasonographic SWE.

The systematic accuracy of FIB-4 for predicting liver fibrosis was found to be equivalent to or superior to that of APRI for all stages of liver fibrosis. The study also provided optimal cutoff values for different groups of fibrosis for both serum markers, which could be more practicable to compare with the distinct cutoff values suggested in foregoing studies.