Evaluation of liver fibrosis using hepatic extracellular volume fraction by contrast-enhanced computed tomography before and after direct-acting antiviral therapy in patients with chronic hepatitis C infection: comparison with serological liver fibrosis markers

Abstract

Objective:

To evaluate time-dependent changes in hepatic extracellular volume (ECV) fraction using contrast-enhanced CT (CECT) and serological liver fibrosis markers, the fibrosis-4 (FIB-4) index and aspartate aminotransferase to platelet ratio index (APRI), before and after direct-acting antiviral therapy (DAA) for hepatitis C virus (HCV) infection.

Methods:

41 HCV-infected patients who achieved sustained virological response (SVR) after DAA (SVR group) and 10 control patients (untreated or unresponsive to treatment) who underwent CECT and serum biochemical tests before or after the first examination/DAA (T1) and at intervals thereafter (T2:<6 months after T1, T3: at 6–12 months, T4: at 12–24 months, and T5:>24 months) were evaluated.

Results:

In the control group, ECV fractions remained relatively unchanged through the study, and significant differences in FIB-4 index comparisons and APRI comparisons were only seen between the T2 and T4 values (p = 0.046 and p = 0.028, respectively). In the SVR group, ECV fractions were significantly different between T1 and T4 and T1 and T5 (p = 0.046 and 0.022, respectively), and both FIB-4 index and APRI were significantly different between T1 and all other time points (p = 0.017 to p < 0.001 and p = 0.001 to p < 0.001, respectively).

Conclusion:

After DAA, ECV fraction decreased slowly, suggesting an improvement in hepatic fibrosis, while serological liver fibrosis markers decreased immediately, probably due to improvement in hepatic inflammation.

Advances in knowledge:

ECV fraction has the potential to be a non-invasive biomarker for the assessment of liver fibrosis after direct-acting antiviral therapy.

Introduction

Chronic hepatitis C virus (HCV) infection causes inflammation and fibrosis, and is known to lead to cirrhosis in approximately 20–30% of patients.^1–3 Cirrhosis-induced fibrosis is a risk factor for eventual progression to hepatocellular carcinoma (HCC). Worldwide, approximately 20% of all cases of HCC are caused by HCV infection.⁴ Hence, HCV infections are treated using anti viral therapy, the goal being to prevent the development of HCC and liver failure. For many years, interferon (IFN) has been used as antiviral therapy for HCV infection. However, since IFN-based treatment is associated with severe adverse events, the number of patients who can tolerate it is limited.^5–7 In recent years, the advent of oral direct-acting antivirals (DAAs) has improved therapeutic outcomes dramatically, since they enable eradication of the HCV in over 90% of infected patients without significant side-effects.^8,9 However, it is unclear whether sustained virological response (SVR) with DAA-based treatment improves hepatic fibrosis and suppresses the development of HCC.^7,8,10–12 Liver biopsy is currently considered the reference standard method for evaluation of hepatic fibrosis.¹³ However, it is an invasive procedure that is not only painful and can cause bleeding, but is also associated with sampling errors.^14,15 Furthermore, although hepatic fibrosis should be regularly evaluated over time following SVR with DAA-based treatment, the highly invasive nature of liver biopsy makes it unsuitable for performing regular evaluations.¹⁶ A variety of techniques for the non-invasive evaluation of hepatic fibrosis have been described in recent years, including magnetic resonance elastography (MRE) and ultrasound elastography.^17–20 However, these techniques need to be performed in addition to regular clinical laboratory tests, and they require more time to perform, as well as special hardware and software.^21,22 The extravascular extracellular volume (ECV) fraction calculated using multiphasic contrast-enhanced CT (CECT) reportedly correlates with hepatic fibrosis.^21–25 The extravascular ECV fraction can easily be calculated using CT values of the hepatic parenchyma and aorta in the pre-contrast and equilibrium phases and the hematocrit level, and it requires neither more time for additional investigation nor any special hardware or software. However, no previous study has used this method to evaluate hepatic fibrosis after a SVR with DAA-based treatment for chronic HCV infection. The fibrosis-4 (FIB-4) index and aspartate aminotransferase to platelet ratio index (APRI) are serological fibrosis markers for the liver and have been reported to be useful in the evaluation of liver fibrosis.^20,26,27 This study aimed to evaluate time-dependent changes in ECV fraction before and after DAA using multiphasic CECT, and serological liver fibrosis markers assessed using blood tests, and to clarify the correlation between these indices.

Methods and materials

Patients

This retrospective study was approved by our institutional review board. The requirement for patient informed consent was waived. CT images and the results of biochemical examinations of the blood of patients with chronic HCV infection who achieved SVR after DAA-based therapies from February 2015 to March 2016 were reviewed (SVR group). Patients who met the following criteria were included: (a) availability of data on pre- and post-treatment multiphasic CECT and biochemical examinations of blood and (b) followed-up for at least 12 months at various time points (T1(P)): time point 1 (pre-treatment); T2: less than 6 months after DAA; T3: 6 to 12 months after DAA; T4: 12 to 24 months after DAA, and T5:>24 months after DAA. A total of 41 patients (16 men, 25 women) with a mean age of 69 years (range, 39–89 years) were enrolled. The mean follow-up period was 32.8 months (range, 12–48 months). As controls, the CT images and results of biochemical examinations of blood of untreated HCV patients or those who were treated but had an inadequate response to treatment from February 2010 to December 2011 were reviewed. The inclusion criteria and time points of investigations in control group subjects were the same as for SVR group subjects, i.e. at the first examination (T1(F)), and at 6 months (T2), 6-12 months (T3), 12 to 24 months (T4) and >24 months thereafter (T5). A total of 10 patients (6 men, 4 women) with a mean age of 70 years (range, 59–83 years) were enrolled as controls. Their mean follow-up period was 25.6 months (range, 18–32 months).

CT protocol

CT images were obtained using one of three multidetector row helical CT scanners (Aquilion Prime 80 and Aquilion 64, Canon Medical Systems, Tokyo Japan; Lightspeed Ultra 16, General Electric Medical Systems, Milwaukee, WI). Imaging was performed at least 3 h after meals in all patients. All patients received non-ionic contrast material (Iopamidol [Iopamiron 370], Bayer Schering Pharma, Osaka, Japan; Iopamidol [Oyparomin 300/370], iohexol [Iopaque 300], Fuji Pharma, Tokyo, Japan, or iohexol [Omnipaque 300], Daiichi Sankyo, Tokyo, Japan) at a rate of 3.3–5.0 ml s⁻¹ and a dose of 600 mgI/kg of patient body weight using an automated power injector (Nemoto Kyorindo Co. Ltd., Tokyo, Japan). A fixed injection duration of 30 s was used; therefore, the injection rate was automatically decided according to patient weight. The contrast medium was injected through a 20-gauge plastic i.v. catheter placed in an antecubital vein. The section thickness and reconstruction interval were 5 mm and 5 mm, respectively. After obtaining pre-contrast CT images, contrast-enhanced dynamic CT images were obtained during the arterial phase, portal venous phase and equilibrium phase, with delays of 40, 70 and 210 s, respectively, based on our optimized clinical protocol.

Image and data analysis

ECV fraction

Absolute enhancements (in Hounsfield units) of the liver parenchyma (E_liver) and aorta (E_aorta) were measured on pre-contrast and equilibrium phase scans. In the liver, approximately 1 cm² regions of interest (ROIs) were designated, avoiding major vessels, the bile duct and tumorous lesions, and the CT attenuation values of each liver segment (eight segments according to the Couinaud classification) in the pre-contrast and equilibrium phases were measured (Figure 1). CT attenuation values for the aorta were measured using ROIs of as large a size as possible. Two radiologists (AK and AH with 14 and 13 years, respectively, of clinical experience in abdominal CTs), who were blinded to any clinical information of the subjects drew operator-defined ROIs by consensus to measure the CT attenuation values of each liver segment. Each ROI was a circle or oval.

Figure 1.

Approximately, 1 cm² ROIs were designated and the CT attenuation values of each liver segment and aorta in the pre-contrast and equilibrium phases were measured. ROI, region of interest.

ECV was calculated using the following equation: ECV (%) = ΔHU_liver / ΔHU_aorta × (100 – hematocrit (Hct) [%]).

FIB-4 index

The FIB-4 index was simultaneously calculated using age, aspartate transaminase (AST), alanine aminotransferase (ALT), and platelet count (Plt), using the following equation: FIB-4 index = AST(IU/L) / (platelet count (10⁹/L) × $\sqrt{ALT(IU/L)}$).²⁸

Aspartate aminotransferase to platelet ratio index (APRI)

The APRI was simultaneously calculated based on the AST level and Plt, using the following equation: APRI = AST (IU/L) /AST (Upper limit of normal (IU/L)) / (platelet count (10⁹/L)×100.

A study coordinator (KM) reviewed the medical records. ECV fraction and the serological liver fibrosis markers, FIB-4 index and APRI, were measured at the same day in each patient.

Statistical analysis

The non-parametric Wilcoxon signed-rank test was used to compare each parameter (Hct, ALT, AST, Plt, Alb, FIB-4, APRI and ECV fraction) pre-treatment and after SVR. Relationships between ECV fractions and the two serological fibrosis markers (FIB-4 index and APRI) pre-treatment and after SVR were assessed using Spearman’s rank correlation coefficient (ρ). The significance of differences in ECV fractions, FIB-4 index and APRI among the five time points (T1 to T5) was assessed using the Friedman test. If p-values from the Kruskal–Wallis test showed a significant difference (p < 0.05), pairwise comparisons between two groups were performed using the Wilcoxon signed-rank test. SPSS software (SPSS, Chicago, IL) was used for all statistical analyses, and p < 0.05 was considered significant.

Results

The results of blood biochemical tests for all available time points are shown in Table 1.

Table 1.

Biochemical results from blood tests for all available time points

SVR group
	T1(P)	T2	T3		T4	T5
Hct (%)	39.0 ± 3.9	38.4 ± 4.8	40.1 ± 4.8		41.1 ± 4.5	39.4 ± 4.2
AST (IU/L)	47.8 ± 28.2	19.0 ± 4.1	16.8 ± 5.0		15.9 ± 6.4	16.5 ± 5.8
ALT (IU/L)	44.3 ± 27.6	27.9 ± 5.3	25.3 ± 6.8		25.0 ± 6.3	25.9 ± 9.3
Plt (x104/μL)	15.9 ± 6.9	15.9 ± 6.7	16.8 ± 6.0		17.2 ± 6.3	16.7 ± 8.2
Alb (g/dL)	4.0 ± 0.5	4.2 ± 0.6	4.3 ± 0.5		4.3 ± 0.4	4.3 ± 0.4
Control group
	T1 (F)	T2	T3	T4		T5
Hct (%)	37.2 ± 3.9	38.6 ± 4.5	36.3 ± 4.8	35.0 ± 6.2		34.3 ± 5.5
AST (IU/L)	66.4 ± 27.8	60.7 ± 24.2	65.0 ± 24.6	88.5 ± 48.3		72.0 ± 32.1
ALT (IU/L)	51.6 ± 29.2	38.7 ± 21.7	46.1 ± 36.6	69.0 ± 78.3		34.0 ± 9.2
Plt (x104/μL)	12.3 ± 7.3	8.8 ± 3.4	9.7 ± 2.9	11.0 ± 5.6		8.2 ± 4.1
Alb (g/dL)	3.5 ± 0.5	3.8 ± 0.6	3.5 ± 0.5	3.4 ± 0.6		3.3 ± 0.6

ALT, alanine aminotransferase; AST, Aspartate transaminase; Alb, albumin; Hct, Hematocrit; Plt, platelet count; T2, less than 6 months after DAA or first examination; T3, 6 to 12 months; T4, 12 to 24 months; T5, greater than 24 months; T1(F), first examination; T1(P), pre-treatment.

In the SVR group, there was a significant difference in all the blood biochemical parameters evaluated except for Hct, at the end of the study period compared to T1(P) (Hct: 39.0 ± 3.9 vs 40.6 ± 4.5, p = 0.098; ALT: 44.3 ± 25.5 vs 8.4 ± 3.47, p < 0.01; AST: 47.8 ± 28.2 vs 16.7 ± 6.1, p < 0.01; Plt: 15.9 ± 6.9 vs 17.6 ± 6.5, p = 0.001; and Alb: 4.0 ± 0.5 vs 4.3 ± 0.4, p < 0.01, respectively). ECV fraction, FIB-4 index and APRI after DAA showed a significant decrease at the end of the study period compared to their values at T1(P) (ECV: 27.49 ± 3.72 vs 29.45 ± 4.83, p = 0.022; FIB-4 index: 3.07 ± 1.88 vs 4.40 ± 3.47, p = 0.001; and APRI: 0.37 ± 0.25 vs 1.39 ± 1.41, p = 0.001, respectively). ECV fraction showed a significant positive correlation with both FIB-4 index and APRI (r = 0.458, p = 0.003; r = 0.285, p = 0.006) at T1(P), although there was no correlation at the end of the study period (r = 0.170, p = 0.289; and r = 0.185, p = 0.248). Comparisons of ECV fraction between the different time points in the SVR group showed a significant difference between T1(P) and T4 and between T1(P) and T5 (p = 0.046 and 0.022, respectively), although no differences in ECV fraction were seen in all other comparisons (p > 0.05) (Figure 2). In FIB-4 index comparisons, significant differences were seen between T1(P) and all other time points (p = 0.017 to p < 0.001), although no differences in FIB-4 index were seen in all comparisons between T2 and all later time point (p > 0.05) (Figure 3). In APRI comparisons, significant differences were seen between T1(P) and all other time points (p = 0.001 to p < 0.001), although no differences in APRI were seen in all comparisons between T2 and all later time points (p > 0.05) (Figure 4).

Figure 2.

Evaluation of ECV fraction in the control group showed no differences in ECV fraction in all comparisons (p > 0.05). In the SVR group, a significant difference in ECV fraction was seen between T1(P) and T4 and between T1(P) and T5 (p = 0.046 and 0.022, respectively). ECV, extracellular volume; SVR, sustained virological response.

Figure 3.

Evaluation of the FIB-4 index in the control group indicated significant differences between T2 and T4 (p = 0.046), with no differences in FIB-4 index in all other comparisons (p > 0.05). In the SVR group, significant differences in the FIB-4 index were seen between T1(P) and all other time points (p = 0.017 to p < 0.001), with no differences in the FIB-4 index in all comparisons between T2 and all later time points (p > 0.05). SVR, sustained virological response.

Figure 4.

Evaluation of APRI in the control group indicated significant differences between T2 and T4 (p = 0.028), although no differences in APRI were seen in all other comparisons (p > 0.05). In the SVR group, significant differences in APRI were seen between T1(P) and all other time points (p = 0.001 to p < 0.001), although no differences in APRI were seen in all comparisons between T2 and all later time points (p > 0.05). APRI, aspartate aminotransferase to platelet ratio index; SVR, sustained virological response.

In the control group, there was no significant difference at the end of the study period compared to T1(F) for all the biochemical parameters evaluated (Hct: 37.3 ± 3.9 vs 33.3 ± 5.8, p = 0.059; ALT: 51.6 ± 29.3 vs 60.2 ± 78.1, p = 0.878; AST: 66.4 ± 27.8 vs 83.4 ± 45.1, p = 0.202; Plt: 12.3 ± 7.3 vs 10.4 ± 6.1, p = 0.575; Alb: 3.5 ± 0.5 vs 3.2 ± 0.5, p = 0.082, respectively). There were no significant differences in ECV fraction, FIB-4 index and APRI at the end of the study period compared to T1(F) (ECV: 35.51 ± 6.40 vs 38.38 ± 5.15, p = 0.059; FIB-4 index: 10.35 ± 6.23 vs 10.94 ± 9.48, p = 0.878; and APRI: 2.29 ± 1.23 vs 5.05 ± 7.18, p = 0.093, respectively). There was no correlation between ECV and both FIB-4 index and APRI at T1(F) (r = 0.224, p = 0.289; r = 0.285, p = 0.425), although ECV fraction showed a significant positive correlation with both FIB-4 index and APRI (r = 0.648, p = 0.043; and r = 0.636, p = 0.048) at the end of the study period. ECV fraction comparisons between the different time points showed no differences in ECV fraction in all comparisons (p > 0.05) (Figure 2). In FIB-4 index comparisons, significant differences were seen between T2 and T4 (p = 0.046), although no differences in FIB-4 index were seen in all other comparisons (p > 0.05) (Figure 3). In APRI comparisons, significant differences were seen between T2 and T4 (p = 0.028), although no differences in APRI were seen in all other comparisons (p > 0.05) (Figure 4).

Discussion

Several reports have evaluated the utility of assessing liver fibrosis using the ECV fraction. Among them, studies conducted with the delay time for the equilibrium phase set to 10 min or more showed good results,^24,25,29,30 although another report found a good correlation between ECV fraction and histological liver fibrosis even using a delay time of 180 s or 240 s for equilibrium-phase images.^22,23,31 Our institution (Kawasaki Medical School,Japanuses a delay time of 210 s for equilibrium-phase images, and this was considered appropriate for measuring ECV fraction.

Long-term changes in ECV fraction and serological liver fibrosis markers, FIB-4 index and APRI, were observed over time in patients who achieved SVR after DAA therapy. Additionally, in the SVR group, both ECV fraction and serological liver fibrosis markers, FIB-4 index and APRI, after DAA showed a significant decrease at the end of the study period compared to their values at T1(P). This result suggests that SVR with DAA treatment results in an improvement in both inflammation and fibrosis of the liver. Conversely, in the control group, there were no significant differences in ECV fraction at the end of the study period compared with T1(F), and there was no correlation between ECV and both the FIB-4 index and APRI at T1(F), although ECV fraction showed a significant positive correlation with the FIB-4 index and APRI (r = 0.648, p = 0.043; r = 0.636, p = 0.048) at the end of the study period. This is contrary to what was seen in the SVR group, in which a correlation was observed between ECV fraction and the FIB-4 index before, but not after SVR. This result might be due to the fact that the ECV fraction and serological liver fibrosis markers, FIB-4 index and APRI, reflect different changes in the liver as a result of DAA treatment. Several studies have reported that HCV eradication with DAA therapy might produce positive short-term effects on liver function.^32–34 However, it is unclear whether this reflects a rapid improvement in fibrosis or simply an improvement in inflammation. There have been no reports on the long-term improvement in fibrosis after DAA therapy. Long-term improvements in fibrosis after interferon treatment have been reported, but these reportedly require several years.^35–37 Evaluation of the changes over time in the ECV fraction and serological liver fibrosis markers after DAA in the present study showed that the serological fibrosis markers started to exhibit significant improvement compared with pre-treatment levels within 6 months of starting treatment, whereas there was no significant difference in the other post-treatment time periods (T2 to T5). In comparison, there was no significant difference in ECV fraction until 1 year after treatment. A study of hepatic fibrosis after DAA treatment measured by vibration-controlled transient elastography found that changes after the start of DAA treatment might be associated with the early disappearance of inflammation, but that improvements more than 12 months after treatment might be an expression of the combination of an improvement in hepatic fibrosis and continued improvement in inflammation.⁹ Based on the principle of ECV fraction, inflammation might also increase the ECV fraction because inflammation potentially expands the intravascular space, although the effect is inferior compared to that of fibrosis. On the other hand, the serological liver fibrosis markers, FIB-4 index and APRI, are highly influenced by decreases in transaminases. Hence, the improvement in the FIB-4 index and APRI seen in the present study probably reflects an improvement in inflammation rather than in fibrosis, whereas the improvement in the ECV fraction probably reflects an improvement in fibrosis.

The present study had several limitations. First, it was a single-center study with a small sample size. Further studies with larger sample sizes are required to corroborate the results. Second, since this was a retrospective study, the intervals between investigations and the duration of follow-up were unspecified. Therefore, it was not possible to compare ECV fraction and the serological liver fibrosis markers, FIB-4 index and APRI at each time point in all cases. However, third, histological data were missing. Since the degree of fibrosis before treatment varied between patients, further studies are required to ascertain differences in the changes in inflammation and fibrosis over time after DAA treatment by means of comparisons with histological data. Finally, since the optimal scan delay for calculation of ECV fraction is not known, the results might improve with further optimization of the technique.

Conclusion

In the present study, long-term changes over time in the serological liver fibrosis markers, FIB-4 index and APRI, and ECV fraction were observed in patients with HCV infection who achieved SVR after DAA therapy. FIB-4 index and APRI improved immediately after treatment, which probably reflected improvements in inflammation, rather than in fibrosis. However, the ECV fraction only started to show a significant improvement relative to pre-treatment levels more than 1 year after treatment, which probably reflected an improvement in fibrosis. Evaluation of the hepatic ECV fraction by CECT might thus provide a useful non-invasive method of evaluating the course of hepatic fibrosis after DAA therapy.