Feasibility study for assessing liver fibrosis in paediatric and adolescent patients using real-time shear wave elastography

Abstract

Introduction:

The objective of this study was to evaluate the feasibility of shear wave elastography (SWE), as a non-invasive means of assessing liver fibrosis stage in paediatric and adolescent patients.

Materials and Methods:

Consecutive paediatric and adolescent subjects scheduled for liver biopsy (LB) evaluation of known or suspected diffuse liver disease were included after informed guardian consent and subject assent in this IRB-approved single institution study. Elastograms were acquired prior to liver biopsy, from the liver under a breath-hold after normal inspiration when possible. Biopsy specimens underwent blinded pathologist review using the METAVIR scoring system.

Results:

Twenty-four patients (M: F = 13:11) with a mean age of 17 years (range: 1–21 years) underwent liver biopsy. The distribution of fibrosis on pathological examination was: F0 = 10, F1 = 9, F2 = 1, F3 = 3, and F4 = 1. Subjects with stages F0 and F1 fibrosis had a mean SWE value of 6.93 kPa (95% CI: 6.33–7.44 kPa) and 8.33 kPa (95% CI: 6.83–10.80 kPa) respectively. The SWE value for the one subject with stage F2 fibrosis was 6.36 kPa, whereas for F3 and F4 were 8.86 (95% CI: 5.70–11.40) and 17.85 kPa respectively. The correlation between SWE values and fibrosis grade was strong (r = 0.58, P = 0.003), and the area under the ROC curve differentiatiang ≥F2 fibrosis was 0.62 (95% CI: 0.26–0.98).

Conclusion:

Estimation of liver stiffness using real-time SWE is feasible using the SC6–1 ultrasound probe in paediatric and adolescent patients and strongly correlates with the stage of fibrosis.

Introduction

Chronic liver disease in children is related to various aetiologies, including congenital, metabolic, toxic, autoimmune/inflammatory and infectious processes.¹ Many liver diseases diagnosed in childhood progress to fibrosis and ultimately cirrhosis, often warranting liver transplantation. Early diagnosis, follow-up and progression of fibrosis in the liver are of great clinical importance. Historically, the tools used to assess the degree of hepatic fibrosis included clinical examination, blood tests, sonography and histology.^2,3 Biopsy is considered the gold standard but has several limitations and risks; it yields a semi-quantitative assessment that is subject to sampling error and interobserver variability and is an invasive technique often requiring general anaesthesia or sedation in children.

Current non-invasive techniques for measuring liver fibrosis in the paediatric population consist of biochemical assays such as the Fibrotest (BioPredictive, Paris, France), ActiTest and APRI (aspartate transaminase-to-platelet ratio index),^4,5 which are limited in their specificity for fibrosis and ability to follow fibrosis progression. Non-invasive imaging techniques that assess fibrosis based on changes in liver stiffness have also been developed, including ultrasound and magnetic resonance elastography. Given the high cost and requirement for sedation/anaesthesia for MR imaging in the paediatric population, it is likely that practical image-based liver stiffness measurement techniques will be based on diagnostic ultrasound.

Current ultrasound elastographic techniques for assessing liver fibrosis include transient elastography (TE), acoustic radiation force impulse imaging (ARFI) and shear wave elastography (SWE). TE is a vibroacoustic non-imaging technology that has shown promising results but has several limitations. Because it is not imaging based, it is subject to anatomic error. In addition, paediatric size transducers have only recently been implemented for TE and have not been validated.^6,7 ARFI and SWE are new methods to assess tissue elasticity. Both techniques have the advantage of being coupled with conventional real-time sonographic imaging, allowing the operator to choose a region of interest (ROI) in the liver for analysis. This permits stiffness estimation to be performed within the hepatic parenchyma itself and avoids inadvertent inclusion of focal lesions or stiff structures, such as large blood vessels. Real-time SWE has advantages over ARFI including higher bandwidth and the ability to create a real-time qualitative map of liver tissue stiffness to guide the ROI selection. The objective of this study was to study the feasibility of fibrosis estimation in the paediatric and adolescent age group using SWE technology.

Methods

Design overview and study population

This was a prospective single institution study approved by the institutional review board (IRB) and compliant with the Health Insurance Portability and Accountability Act. Subjects scheduled for an ultrasound-guided non-focal liver biopsy in the interventional radiology department for the evaluation of known or suspected diffuse liver disease were eligible to participate in the study. All subjects were under the care of a paediatric hepatologist and were recruited for the study in the paediatric hepatology outpatient clinics. Informed parental and subject consent was obtained in all cases.

SWE

A sonographer with 16 years of scanning experience (JC) obtained all the SWE measurements. Estimation of liver stiffness was performed using SWE on the Aixplorer US system (SuperSonic Imagine S.A., Aix-en-Provence, France) with a convex broadband probe (SC6–1). SWE technology measures the propagation of shear waves within tissue. The velocity of these shear waves is dependent on liver stiffness, which in turn correlates with the severity of liver fibrosis. Using inversion algorithms, the shear elasticity of the medium can be mapped quantitatively and qualitatively from dynamically obtained estimates of shear wave propagation velocity.⁸ Relevant Young’s modulus estimates can be obtained in a ROI by placing a circular ROI box within the SWE image.

SWE acquisitions were performed immediately prior to the liver biopsy in the Interventional Radiology ultrasound suite. A series of elastography images were obtained without the patient being sedated and during a breath-hold after normal inspiration when possible. A post-doctoral research fellow (MD) with 1 year of experience in elastography placed ROIs in the SWE images on a dedicated research workstation. ROIs were placed in the homogenous region of the elastogram as can be seen in Figure 1a,b. The median of all elastography measurements was used as the liver Young’s modulus estimate for each subject, and all subsequent statistics such as mean SWE value for a fibrosis stage were calculated using these median values.

Fig. 1.

Example of SWE values obtained from offline quantification by placing a ROI placed in an elastogram obtained in (a) a 20-year-old female with autoimmune hepatitis type II and no fibrosis (F0) on biopsy who had a median SWE value of 6.08 kPa and (b) a 19-year-old female with AIH type I disease that had F4 fibrosis on biopsy and the SWE value of 17.94 kPa.

Liver biopsy (LB)

An interventional radiologist performed all non-focal liver biopsies under ultrasound guidance. Informed consent and assent were obtained prior to the procedure. All procedures were performed under intravenous sedation. All biopsies were obtained from the right lower lobe. 16G BioPince™ (Full Core Biopsy Instrument; Medical Device Technologies Inc. Gainesville, FL, USA) biopsy needles were used for all cases. The quantity of tissue obtained was not routinely recorded at the time of biopsy; however, it is our divisional protocol to obtain at least one 2-cm long core biopsy. All biopsy specimens were fixed in formalin, embedded in paraffin, and cut and stained using standard pathology laboratory protocols for the evaluation of liver fibrosis.

Histological examination

A single hepatopathologist (JM) blinded to the SWE values and clinical information reviewed the biopsy specimens. For uniformity of reporting and comparison with other studies, liver fibrosis was assessed using the METAVIR staging system, even if the pathologic process would require a different staging system for clinical reporting purposes. In the METAVIR system fibrosis is staged on a five-point ordinal scale from 0 to 4: (F0, absent; F1, enlarged fibrotic portal tract; F2, few portal-portal septa but intact architecture; F3, many septa with architectural distortion but no obvious cirrhosis; and F4, cirrhosis).

Statistical analysis

Median values of all SWE measurements obtained for each subject were used for statistical analysis. Box and whisker plot depicting the distribution of median SWE values by fibrosis stage was drawn. All statistical analysis was performed on SPSS software (release 22.0; IBM, Armonk, NY, USA).

Results

Patient characteristics

Twenty-four patients participated in the study between June 2010 and November 2014. There were 11 females with a mean age of 18 ± 2.88 years (range: 12–21 years) and 13 males with a mean age of 16.15 ± 4.96 (range: 1–21 years) that were enrolled in the study. The most common indication for liver biopsy was elevated transaminases (45.8%, n = 11/24), and the most common final diagnosis was non-alcoholic fatty liver disease/non-alcoholic steatohepatitis (20.1%, n = 5). Patient demographics, clinical indication for liver biopsy and final clinical diagnosis are summarised in Table 1.

Table 1.

Summary of patient demographics, lab values, indication for biopsy and final clinical diagnosis in our cohort

S. no.	Age	Gender	Lab values				Indication for liver biopsy	METAVIR Stage	Final clinical diagnosis
			ALT (U/L)	AST (U/L)	ALKP (U/L)	T. Bil. (mg/dL)
1	16	M	68	43	230	1	Elevated transaminases, high BMI	0	NAFLD
2	1	M	323	69	456	0.5	Elevated transaminases, cholestasis with normal GGT	0	PFIC 2
3	21	M	59	31	78	0.8	CPT1 deficiency, elevated transaminases	0	CPT1 deficiency with mild steatosis
4	12	F	353	198	209	0.7	Post OLT for Biliary atresia, elevated transaminases	0	Allo-immune hepatitis of the graft
5	17	M	12	18	63	0.5	Portal hypertension	0	Cavernous transformation of the portal vein of uncertain aetiology
6	17	F	158	78	60	0.3	Elevated transaminases, suspected NASH	0	NASH
7	18	F	61	54	72	0.3	Elevated transaminases	0	Lobular injury, presumably DILI
8	20	F	8	18	70	0.3	History of AIH; monitoring of disease activity and progression	0	AIH, with minimal inflammation
9	20	F	57	45	169	0.3	Elevated transaminases	0	Resolving DILI
10	20	F	77	54	101	0.3	Elevated transaminases	0	No significant abnormality
11	15	M	98	54	150	0.4	Elevated transaminases, high BMI	1	NAFLD
12	16	F	40	27	77	0.4	Elevated transaminases, high BMI	1	NAFLD
13	14	M	163	73	365	1.6	Fatty liver, elevated transaminases	1	NAFLD
14	18	M	133	72	48	0.6	Hepatitis B grading and staging	1	Chronic active Hepatitis B
15	17	M	57	60	418	0.5	Cystic fibrosis, portal hypertension	1	Cystic fibrosis, Heterozygous MZ A1AT, mild steatosis
16	18	M	76	60	138	1.3	Ulcerative colitis, suspected PSC	1	PSC
17	20	F	112	50	47	0.4	Hepatitis C grading and staging	1	Chronic active hepatitis C
18	21	M	103	88	155	0.6	Hepatitis C grading and staging	1	Chronic active hepatitis C
19	21	F	34	31	346	0.3	Post OLT for cystic fibrosis with elevated transaminases and alkaline phosphatase	1	Allograft rejection
20	14	F	435	329	111	0.5	Hepatitis B	2	Chronic active hepatitis B
21	17	M	319	196	251	0.7	Ulcerative colitis, suspected PSC, transaminase elevations	3	PSC/AIH overlap
22	17	M	83	45	298	0.3	Hepatitis C grading and staging	3	Chronic active hepatitis C
23	18	M	78	57	115	0.5	History of post-infantile giant cell hepatitis; Monitoring of disease activity and progression	3	AIH, post-infantile giant cell variant
24	19	F	335	223	107	0.8	Jaundice, elevated transaminases	4	AIH

A1AT, alpha 1 antitrypsin; AIH, autoimmune hepatitis; BMI, body mass index; CPT1, carnitine palmitoyltransferase 1; DILI, drug-induced liver injury; GGT, gamma-glutamyl transferase; NAFLD, non-alcoholic fatty liver disease; NASH, non-alcoholic steatohepatitis; OLT, orthotopic liver transplant.; PFIC, progressive familial intrahepatic cholestasis; PSC, primary sclerosing cholangitis; S. no., serial number; T. Bil, total bilirubin.

Clinical data and elastography measurements

Mean values for alanine aminotransferase, aspartate aminotransferases and alkaline phosphatase were 135 U/L (range: 8–435 U/L), 82.20 U/L (range: 18–329 U/L) and 172.25 U/L (range: 47–456 U/L) respectively. Total bilirubin ranged between 0.3 and 1.6 mg/dL (mean = 0.58 mg/dL). On pathological examination, 19 patients had no or mild fibrosis (10 patients had F0 and 9 patients had F1), whereas five patients had clinically significant fibrosis (stage ≥ F2). One patient had stage F2 fibrosis, three patients had stage 3 fibrosis and one patient had stage 4 fibrosis.

SWE values were successfully obtained in all subjects. The SWE values ranged from 3.04 kPa to 25.04 kPa with a mean value of 8.19 kPa. The median value of the measurements for each subject ranged from 5.20 to 17.85 kPa with a mean value of 8.08 kPa. Table 2 summarises the SWE values for each stage of fibrosis in our patient cohort.

Table 2.

Summary of median SWE values obtained from each subject grouped by fibrosis grade

Fibrosis	F0	F1	F2	F3	F4
Total	10	9	1	3	1
Mean	6.93	8.33	6.12	8.86	17.85
Standard deviation	0.95	1.99	-	2.86	-
95% CI	6.33–7.44	6.83–10.80	-	5.70–11.40	-
Min.	5.20	5.52	-	5.61	-
Max.	8.70	12.00	-	11.68	-
Outliers	5.20,8.70	10.80,12.00	-	-	-

The median SWE measurements in the 10 subjects with stage F0 fibrosis had a mean SWE value of 6.93 kPa (95% CI: 6.33–7.44 kPa) with a range of 5.20–8.70 kPa. For stage F1 fibrosis (n = 9), the mean SWE value was 8.33 kPa (95% CI: 6.83–10.80 kPa) with a range of 5.52–12 kPa. The median SWE value was 6.12 kPa for the subject with stage F2 fibrosis and 17.85 kPa for the patient with stage 4 fibrosis. The three subjects with stage F3 fibrosis had a mean SWE value of 8.86 kPa (95% CI: 5.70–11.40 kPa). Figure 2 presents the data in the form of a box-whisker plot.

Fig. 2.

Box and Whisker plot represents the median values of SWE measurements obtained for each fibrosis stage.

Pearson’s correlation between METAVIR fibrosis stage and median Young’s modulus showed a strong correlation, which was statistically significant (r = 0.58, P = 0.003). Receiver-operating characteristic curves to differentiate significant fibrosis (>=F2) from lesser degrees of fibrosis (F0, F1) was 0.62 (95% CI: 0.26–0.98) (Fig. 3). At a cut-off value of 8.78 kPa, the sensitivity and specificity to diagnose significant fibrosis was 60% and 89.5% respectively.

Fig. 3.

Receiver-operating characteristic curve using median SWE values to differentiate significant fibrosis (F2–F4) from lesser degrees of fibrosis (F0, F1).

Discussion

In this feasibility study, we show that the Supersonic Aixplorer SWE machine can be used for the estimation of liver fibrosis in a cohort of paediatric and adolescent patients. Median SWE values obtained from the liver in 24 subjects showed a statistically significant correlation of SWE values with increasing METAVIR histologic fibrosis stage (r = 0.58, P = 0.003) spanning the range of F0 (6.93 kPa), F1 (8.33 kPa), F2 (6.12 kPa), F3 (8.86 kPa) and F4 (17.85) fibrosis. The one outlier was the single patient with METAVIR F2 fibrosis that exhibited a median SWE value (6.36 kPa) below the F1 range. There are known confounders in SWE estimation and similarly sampling errors are known to exist in liver biopsy procedures given that a very small part of the liver is sampled. Hence without additional cases in the F2 fibrosis category, the reason for this outlier cannot be fully explained. There was no technical difficulty obtaining SWE measurements during the elastography examination, and our study demonstrates feasibility of using SWE to estimate fibrosis in the paediatric and adolescent age group. Larger clinical studies are required to confirm SWE accuracy and establish threshold values for fibrosis grading.

Monitoring fibrosis remains an important component in prognostication and management of patients with chronic liver disease, because if not treated, progressive fibrogenesis culminates in cirrhosis. Currently, liver biopsy remains the best standard for evaluation of chronic liver disease.⁹ This poses a particular challenge in the paediatric and adolescent age group, given the frequent requirement for general anaesthesia to perform the procedure and the subjectivity and sampling risk potential associated with fibrosis assessment on biopsy specimens.^9,10 As a result, a clinically useful non-invasive method for evaluation of liver fibrosis would be of great utility. Figure 1a,b demonstrates the clinical utility particularly in the follow-up of patients with known liver disease and more so in the paediatric and adolescent age group where anxiety to a invasive test such as liver biopsy is significantly high.

Serum biomarkers provide one such option, and there have been as many as 14 validated biomarkers.¹¹ However, no serum biomarker has been established as a substitute for liver biopsy, and some investigators have concluded that serum biomarkers should only be used as an adjunct to liver biopsy for monitoring patients with chronic liver disease.¹¹ The literature on serum biomarkers in the paediatric population is limited. It seems reasonable, however, to postulate that serum markers of fibrogenesis would be subject to similar limitations as those experienced in adults.

MR and ultrasound elastography are non-invasive imaging modalities currently under investigation for liver fibrosis staging. MR elastography particularly has been shown to have high sensitivity and specificity for the estimation of fibrosis.¹² Several important considerations need to be addressed before extending its use to children, including issues of cost, need for sedation/anaesthesia and establishment of threshold values for fibrosis detection.

Ultrasound elastography is an inexpensive and time-efficient procedure that can be readily performed in paediatric patients without any need for sedation. TE is the most studied implementation of ultrasound elastography in adults. More recently, ARFI imaging techniques (e.g., that implemented by Siemens as Virtual Touch™ tissue quantification (Siemens Medical Solutions, Mountain View, CA, USA) and ShearWave™ elastography (Aixplorer® SuperSonic Imagine S.A) have emerged.

TE has been extensively studied in adults,^13–15 and to a lesser degree in children.^16–20 TE allows measurements only at a fixed depth and without real-time ultrasound guidance. In addition, the use of TE is limited in obese patients and cannot be performed in patients with ascites.

ARFI imaging elastography is a non-invasive technology to evaluate hepatic fibrosis that has shown utility for the differentiation of intermediate fibrosis stages in adults^21–28 as well as in the paediatric age group.^29–32 Real-time SWE has two main advantages over ARFI: (i) the use of shear waves with higher bandwidth than TE; and (ii) the creation of a real-time qualitative map of liver tissue stiffness allowing visualisation of liver stiffness heterogeneity, facilitating placement of a representative ROI.^33,34 Several studies in adults have been performed using SWE,^33–36 and some have demonstrated a better sensitivity of SWE in the estimation of liver fibrosis over TE,^34,35 but no studies comparing ARFI and SWE have been currently performed.

In this study, we provide evidence that SWE can be successfully and safely be performed in paediatric and adolescent patients. Further studies are needed to demonstrate reproducibility and establish the precise cut-off values for different liver fibrosis stages.

Study limitations

Our study was performed principally to establish feasibility of SWE in children. Our sample is too small to permit robust analysis of the correlations between SWE estimates of Young’s modulus with fibrosis stage. Larger studies to evaluate the accuracy of SWE for liver fibrosis staging in children are needed.

Conclusion

Estimation of liver stiffness using real-time SWE is feasible using the SC6–1 ultrasound probe in children with liver disease, which has great potential as a non-invasive biomarker of hepatic fibrosis.