Doppler Ultrasound and Fibroscan Parameters Versus Liver Biopsy ‎in Evaluation of Hepatic Fibrosis in Egyptian Patients with Chronic ‎Hepatitis C‎

INTRODUCTION

Egypt has the highest incidence of hepatitis C (HCV) infection ‎worldwide with major public and economic health burden [1]. With ‎persistent liver damage; fibrosis can progress to cirrhosis in up to 15% to ‎‎20% of cases within 20 years [2]. An estimation of fibrosis progression ‎is critical for assessing the outcome of HCV infection. The gold standard ‎for assessing the grade of liver fibrosis is liver biopsy, which provides ‎diagnostic information not only on fibrosis but also on many other ‎processes of liver damage, such as necrosis, inflammation, steatosis, ‎copper or iron hepatic deposits [3]. Because of its complications, such as ‎bleeding or pain, sampling error can occur in up to 25% of cases with ‎biopsy perception inter- and intra-observer variability [4] and its ‎limitation in distinguishing between stages F1 and F2 of fibrosis [5] and, ‎in order to assess the severity of liver fibrosis, non-invasive methods ‎have been developed [6]. The non-invasive tools available are either ‎‎"biological" methods based on serum fibrosis biomarker levels or‎ ‎‎"physical" methods based on the measurement of liver stiffness using ‎transient elastography [7]‎.‎

Ultrasound-based elastography has led to a breakthrough in the ‎assessment of fibrosis by different methods: one-dimensional ultrasound ‎elastography-transient elastography (TE); or two-dimensional ultrasound ‎‎(or B-mode) using conventional ultrasound imaging-acoustic radiation ‎force (ARFI) imaging or point shear wave elastography (pSWE), real-‎time 2D shear wave elastography (2D-SWE) and real-time elastography ‎‎(RTE) [8]. ‎

The first ultrasound-based elastography for the diagnosis and ‎staging of liver fibrosis [9] is transient elastography (Fibroscan®, ‎Echosens, Paris). By calculation of low- frequency elastic shear wave ‎velocity distributed through the liver, TE uses a mono-dimensional ‎ultrasound to estimate liver stiffness. TE can be accomplished in brief ‎periods and has outstanding (typically less than 5 minutes) intra- and ‎inter- observer variability [10]. ‎

Ultrasound is the screening tool for initial assessment of the liver. With ‎the appearance of Doppler ultrasound carried non-invasive means to ‎evaluate secondary changes of blood circulation to chronic liver disease.  ‎It could detect the decrease of the portal vein flow velocity and caliber ‎increases, both indicative of portal hypertension [12]. Through the ‎comparison of Doppler ultrasounds of hepatic blood flow and fibroscan ‎with liver biopsy, we aimed to enhance the non-invasive assessment of ‎liver fibrosis in Egyptian patients with chronic hepatitis C. ‎

METHODS

This was a retrospective comparative observational analysis, with a ‎purposive convenience nonrandomized sample. The informed consent ‎form authorizing the collection of this data was signed by all patients. ‎‎250 medical records were reviewed and 78 patients were enrolled with ‎chronic HCV genotype 4 infection. The inclusion criterion included ‎chronic HCV- infected patients aged 18 years or older, male or female, ‎who were assessed for treatment (pegylated interferon and ribavirin) ‎under the national therapeutic regimen. ‎

Exclusion criteria: ‎

‎1-    Any medical history of compensated or decompensated congestive ‎heart failure to prevent influence of retrograde hepatic blood flow caused ‎by cardiac reflux  ‎

‎2-    History of decompensated liver disease ‎

‎3-    Post liver transplantation (to prevent anastomosis affecting blood flow ‎measurement) ‎

‎4-    Pregnancy ‎

‎5-    Age less than 18 years ‎

‎6-    Hepatitis B virus co-infected patients ‎

We used the free sample size calculator for cohort study at ‎‎(http://www.openepi.com/ SampleSize/SSCohort.htm) to calculate the ‎sample size for this study. At test power of 0.8 and confidence interval ‎of 95%, sample size was assumed to be 74 patients

‎Liver biopsy specimens were obtained under complete a septic ‎procedures to retrieve 15mm core or at least 15 portal tracts.The ‎specimen was processed and stained with hematoxline and eosine.  ‎Fibrosis was tagedona 0ñ 4scale:F0, no fibrosis; F1, portal fibrosis ‎without septa; F2, portal fibrosis and few septa; F3, numerous septa ‎without cirrhosis; and F4, cirrhosis according to METAVIR scoring ‎system [11].‎

Liver biopsies were taken from 78 HCV- infected Egyptian patients as a ‎part of the work-up before HCV treatment. Fibroscan analysis, ‎abdominal ultrasonography, and Doppler ultrasound were performed by ‎a qualified operator who has no data of the patient's clinical history, ‎biochemical examination, or histo pathological results. ‎

All records of the patients are documented in the hospital inpatient and ‎outpatient clinics in the Tropical Medicine Department of Tanta ‎University Hospital, Tanta, Egypt‏.‏‎ ‎

All patients were exposed to ‎

‎-    History taking,  complete clinical examination, laboratory tests ‎including the full blood analysis, blood urea , serum creatinine  the liver ‎function tests (liver enzymes, albumin and bilirubin), prothrombin time, ‎activity, and APRI score were calculated in each patient, and patients ‎were then exposed to: ‎

‎-    Abdominal ultrasonography.‎

‎-    Transient Elastography (Fibroscan): The hepatic stiffness ‎measurements were conducted by skilled operators following the ‎manufacturer's recommendations, using the 502M fibroscan (echosens-‎France) probe. It was performed with the patient in supine posture and ‎totally abducted by the right arm from the intercostal transthoracic ‎window on the right liver lobe. ‎

The interquartile range (IQR) was considered accurate with a score of ‎less than 30 % and a success rate of not less than 70 %. The median ‎value was known to be an indication of the elastic state of the liver.  The ‎median value was determined automatically by the software, and the ‎results were expressed in Kilopascals (kPa). All measurements were ‎done with the FibroScan (M) probe after a fasting period of 6 hours. ‎The elastogram score obtained by the device is interpreted through ‎special software to identify the corresponding fibrosis stage in the ‎Metavir score.‎

‎-  Doppler ultrasound was done by using Toshiba Nemio XG apparatus ‎with a convex probe 3.5 MHz to assess:‎

a) ‎        the diameter of the portal vein and the volume of the portal vein ‎blood flow

b) ‎ ‎ the Hepatic venous resistance index (HVRI) for the right hepatic ‎vein

c) ‎ The Hepatic artery resistance index (HARI)‎

d) ‎ The Hepatic artery pulsatility index  (HAPI) ‎

e) ‎ The Splenic artery resistance index  (SARI)‎

The normal diameter of portal vein is highly variable but does not ‎exceed 16mm in a resting state on quiet respiration [43]. Normal hepatic ‎artery RI was reported by McNaughton and Abu-Yousef, to be 0.55–‎‎0.7 [44], while normal hepatic pulsatility index was reported by ‎Schneider et al, to be 0.92±0.1 [45]. The normal value of splenic artery ‎resistive index (SARI) was reported by Loanitescu to be 0.51 +/- ‎‎0.05 [46]. ‎

Statistical analysis:

The data was analyzed using version 2 of Sigma ‎Stat. Quantitative data are presented as mean and SD while qualitative ‎data are expressed as number and percent. Non-normally distributed data ‎are expressed as range (min-max) and median.‎

Difference among groups was performed using independent-sample ‎student t-test between groups when normally distributed or ‎ManneWhitney U test if not normally distributed.‎

The Receiver Operator Characteristic (ROC) curves were developed to ‎assess the accuracy of the diagnosis prediction of fibrosis using version ‎‎13.0 of the Social Sciences Statistical Package (SPSS) (IBM Corp, ‎Armonk, NY, USA). The areas of individual tests with 95% confidence ‎intervals under ROC curves (AUCs) were calculated and compared. The ‎coefficient of Correlation was measured by Spearman’s correlation. The ‎findings were found statistically significant when p

RESULTS

 This retrospective research was conducted on 78 HCV patients with a ‎mean age of 43.17 ± 10.77 years and 37 (47.4%) males. There was no ‎statistical difference in (AST to Platelet Ratio Index) APRI score among ‎patients categorized according to fibrosis degree in liver biopsy (P> ‎‎0.05).‎

Sixty patients had abnormal findings in ultrasound; 20 patients had ‎bright liver, 12 patients had periportal fibrosis, and 28 had ultrasonic ‎findings of liver cirrhosis as coarse echo pattern, irregular outlines, ‎enlarged caudate lobe and shrunken liver in some patients. Regarding ‎splenic examination 16 patients had splenomegaly. The kidney ‎examination in all patients showed normal size and good differentiation ‎between cortex and medulla except for one patient who had grade one ‎nephropathy.‎

According to fibrosis and activity detected by liver biopsy:‎

‎•‎    ‎ According to fibrosis degree according to METAVIR scoring ‎system[11]: 57 patients with mild fibrosis (F1 and F2) and 21 patients ‎with significant fibrosis (F3).‎

‎•‎    According to activity degree: 35 patients with mild activity (A1) ‎and 43 patients who showed moderate to severe activity (A2-A3) ‎

‎In order to evaluate the fibroscan results (figure 1):‎

There was a statistically significant difference in fibroscan results with ‎higher scores detected in higher fibrosis P

‎The analysis of Doppler ultrasonic parameters of the liver:‎

The normal diameter of portal vein is highly variable but does not ‎exceed 16mm in a resting state on quiet respiration [43].‎

There  was  a statistically  significant  decrease  in portal vein flow ‎volume (P<0.001) in  moderate fibrosis  in comparison to mild fibrosis ‎also, Portal  vein  flow  volume  was decreased significantly in the  ‎patients  when compared to the average documented values P <0.001, ‎Also   a  significant negative  correlation  between  fibrosis  degree  in  ‎liver  biopsy  and portal  vein  flow  volume was found  (r= - 0.55 ,  p ‎‎<0.001). No statistical differences in splenic artery resistance index ‎‎(SARI) , hepatic  artery  resistance  index  (HARI), and hepatic artery ‎pulsatility index (HAPI)  (P> 0.05) between the two groups, however, ‎in our patients ,HARI , HAPI , and SARI were significantly higher ‎compared to the average documented values (p=0.03,

As regards ROC  curve  of  portal  vein  blood  flow  volume  for  the ‎detection  of  fibrosis  degree  in  liver biopsy; For F2,  AUC was 0.645 ‎‎(95% CI = 0.424-0.866). At a cut-off point ≤ 0.443  L/min, the portal ‎vein  blood  flow  volume  had  sensitivity  of  76.9%  and  specificity  ‎of  60%  in  detecting  F2  hepatic fibrosis in liver biopsy and For F3, it ‎was not an accurate method (AUC = 0.201, 95% CI=0.61-0.340) ‎figure 2.‎

DISCUSSION

In order to improve the efficacy of noninvasive evaluation of  liver ‎fibrosis , we compared Doppler ultrasound of hepatic blood flow and ‎fibroscan: with liver biopsy in In Egyptian chronic hepatitis C patients‏.‏‎ ‎In all of the patients who were examined, a statistically significant ‎positive correlation was observed between ultrasound findings and ‎fibrosis score, liver biopsy, and fibroscan results as changes in liver echo ‎pattern and texture were accompanied with increased estimated fibrosis ‎levels in liver biopsy and fibroscan. This was similar to that obtained by ‎Nishiura et al. (2005) who found that the increase of the texture of the ‎liver parenchymal obtained by ultrasound has a statistically significant ‎correlation with the degree of liver fibrosis [17]. ‎

Our results were also supported by the results of Abd El Dayem et al, ‎‎(2013) who showed the diagnosis of hepatic fibrosis, by ‎ultrasonography had 87.5% sensitivity, 77.5 % specificity, and 84.0 % ‎accuracy, mainly significant fibrosis (F2-4). But, they concluded that ‎ultrasound imaging cannot identify or precisely diagnose fibrosis in the ‎absence of cirrhosis stigma, like shrunken liver and ascites [18]. It  was  ‎also  in  agreement  with Davoudi et al, 2015 who found a  ‎significant   positive  correlation  between  total their  gray-scale score  ‎and  liver  fibrosis.  They concluded that the liver parenchyma ‎echogenicity may be enough and simple for use in clinical practice to ‎detect fibrosis stage [19] Abd El Maksoud et al. (2015) have also, ‎suggested that echo texture is significantly correlated with the level of ‎fibrosis [20]. Choong et al. (2012), concluded that routine clinical ‎ultrasound is not successful for liver fibrosis staging, giving treatment ‎options , or determining the prognosis of chronic hepatitis secondary to ‎chronic liver disease. They observed, however, that ultrasound is still ‎beneficial for cirrhosis detection [21].‎

Fibroscan results revealed a strong positive association with the ‎grade of fibrosis and activity index in histopathological analysis. ‎Fibroscan was seemed to be more effective for F3 fibrosis diagnosis. To ‎diagnose F3 fibrosis, we registered a 7.95 kPa fibroscan cut-off value. ‎

Similar findings were obtained by several authors reporting ‎distinct but nearby cutoff values such as Castéra et al, (2005) who ‎reported a cutoff value of 9.5 for fibrosis prediction   ≥ F3 with ‎specificity 91% and sensitivity 73% [22]. In addition,  Ziol et al. ‎‎(2005) who  stated   a cutoff   value  of  9.6  for  prediction  of  fibrosis  ‎‎≥ F3  with specificity 85% and sensitivity  of 86% [23]. Arena  et  al. ‎(2008)  who put cut off value of  ≥ 10.8 for prediction of  fibrosis  ≥ F3 ‎with specificity 94% and sensitivity 91% [24], This  was  similar  to  the  ‎results  of  Fahmy  and  Badran, (2011)  who concluded that:  With ‎AUCs 0.92 and 0.95, TE is the most accurate tool for predicting severe ‎fibrosis and cirrhosis and finding that with TE at 7 kPa, 86% were ‎correctly classified for the prediction of significant hepatic fibrosis [25]. ‎

Similar to our results, Lutz et al. (2012) ROC analysis showed that in ‎lower fibrosis stages Fibroscan lacks accuracy; and that there was a ‎significant correlation between fibrosis degree detected by fibroscan and ‎inflammation in the histopathological analysis [26]. The positive ‎correlation between liver biopsy and the fibroscan score was also, ‎concluded by Abd El Maksoud et al. (2015), with  moderate  ‎agreement  (matching)  between liver  biopsy  and  fibroscan,  the  ‎lowest  matching  was  in  F0  and  the highest was in F3 [20]. This was ‎in line with El-Saadany S et al. 2016, who studied fibrosan and biopsy ‎results in 348 CHC patients and found that fibroscan data in moderate ‎fibrosis (p < 0.001) were positively correlated with biopsy, but not mild ‎and no fibrosis (p=0.12) and concluded that fibroscan was correlated ‎with fibrosis degree in liver biopsy and could be applied as a ‎noninvasive tool for diagnosis moderate (F2– F3), but not mild (F1) ‎fibrosis [27].‎

We tested Doppler US parameters as an indirect method for ‎grading of hepatic fibrosis. All  our patients had continuous  ‎hepatopetal blood flow, with a significant negative correlation  ‎between PVBF and fibrosis degree in liver biopsy and fibroscan ‎score. These results may be  attributed to the increased of hepatic ‎parenchymal  resistance that might be caused by fibrosis [28].‎

In 2002, Vyas et al. analyzed PVBF and PVV and observed that PVBF ‎and PVV in cirrhotic patients were considerably lower than controls ‎‎[29]. These changes increase as  the liver disease gets rising, while Shi ‎et al. (2005) who has observed 38 cirrhosis patients (Child grades A to ‎C) and 20 controls showed that no variation in the diameter of the portal ‎vein is obtained  between controls and cirrhotic patients‏.‏

Portal flow for child C cirrhosis was significantly lower than for Child A ‎and Child B cirrhosis, but no variations of PVBF was found among ‎Child A and B cirrhosis patients and the controls [30].‎

On  the  other  hand Walsh et al. (1998)  reported that portal  vein flow  and  total  hepatic  flow  were  similar  in  chronic hepatitis  ‎C  and controls  (The indices of inflammation or degree of hepatic ‎fibrosis were not linked to the state of the liver ). These results may be ‎due to the use of old fashion and inaccurate apparatus [31]. While 36 ‎patients with chronic viral hepatitis,  30 patients without any signs of ‎liver disease as a control group, and 63 patients with cirrhosis with no ‎PVBF difference between three groups were observed by Haktanir et ‎al. (2005), the PVD increased significantly and the mean PVV ‎decreased significantly in cirrhosis compared with hepatitis and control ‎groups [32]. One of the limitations of our study is a few patients with ‎cirrhosis grade 4. The small sample size limits our ability to detect ‎significant fibrosis if any.  Also, the  discrepancy between  our  results  ‎and  the  results  obtained  from  studies that included cirrhotic groups ‎may be due to increased hepatic inflow and decreased  PVV mean  these ‎changes that accompany portal hypertension. ‎

ROC curve analysis showed that PVB was more accurate in ‎detecting lower grades of fibrosis than higher grades. ‎

Normal hepatic artery RI was reported by McNaughton and ‎Abu-Yousef, to be 0.55–0.7 [44], while normal hepatic pulsatility ‎index was reported by Schneider et al, to be 0.92±0.1[45].‎

There was a significant increase in mean values of HARI and ‎HAPI in our patients compared to the normally documented values of ‎HARI and HAPI in healthy subjects. These circulatory changes are ‎considered to be occurring in patients with chronic hepatic disease in ‎hepatic arteries are related to the architectural deterioration that occurs ‎within the liver with greater severity of the disease [33].‎

But there was no difference regarding HARI and HAPI between ‎the two studied groups who had been categorized according to degree of ‎fibrosis by liver biopsy, also there was no difference between HARI and ‎HAPI between the two studied groups who had been categorized ‎according to the degree of fibrosis by fibroscan, and there was no ‎correlation between HARI or HAPI and the degree of inflammation in ‎liver biopsy. ‎

Our Results were  similar to those found by Salvatore et al. ‎‎(2012)  who studied  100 patients who had a liver disease caused by ‎hepatitis C,  compared results of Doppler parameters with results of ‎fibroscan  for  detection  of the degree of liver stiffness  and they found ‎that in these patients HARI rises  progressively compared  to healthy ‎subjects.  They put a cut-off value of 0.64 for HARI  at which Patients ‎with LS values above 13 kPa with high sensitivity (84.4 %) and ‎moderate specificity may be identified (69.1%),  but finally, they ‎concluded that fibroscan  results  were  more  accurate  than Doppler ‎parameters despite the direct correlation between HARI and liver ‎stiffness [34].‎

Although Piscaglia et al. (2001) observed that HARI was higher in ‎patients with severe hepatic fibrosis because the hepatic artery resistance ‎indices seem to be influenced by inflammation and chronic repair that ‎determine hepatic fibrous deposition, and secondly by aging [35]. On ‎the other hand, Lutz et al (2012) observed that HARI did not display ‎any substantial variations in fibrosis scores F1-F3 but was significantly ‎higher in cirrhosis, although no statistically significant differences in ‎HAPI measurements in the various stages of fibrosis (F0-F4/cirrhosis) ‎was found [26]. In comparison, Lim et al. (2005) did not find an ‎association between HARI values and histological scores [36].‎

The normal value of splenic artery resistive index (SARI) was reported ‎by Loanitescu to be 0.51 +/- 0.05 [46]. ‎

A significant increase in the mean value of SARI measured among ‎all patients in our study was found compared to normally documented ‎values, this difference may be explained as the advance of hepatic ‎fibrosis contributes to increased portal resistance, which induces ‎increased resistance to splenic artery outflow[37].‎

But our results did not find any significant differences among the ‎studied groups and there was no correlation between SARI and fibrosis ‎score neither to liver biopsy nor fibroscan results.‎

Similar to our result Liu et al, (2007) documented that SARI and ‎SAPI were correlated with the degree of fibrosis [38].‎

Results in this work were different from that obtained by ‎Salvatore et al. (2012) who found that there was a direct correlation ‎between SARI and liver stiffness taking into consideration that the study ‎was restricted to the TE values for F ≥ ‏‎2. The SARI cutoff value of 0.56 ‎was used to detect patients with LS values >13 kPa with strong ‎sensitivity (81.3%) and moderate specificity (48.5%). However, they ‎concluded that despite the significant correlation between LS and ‎Doppler US parameters, they were not excellent in detecting liver ‎stiffness [34].‎

While Cançado et al. (2007) found that the results of SARI were ‎similar in patients with chronic hepatitis and healthy individuals, but it ‎was established that patients with cirrhosis had greater indices [39].‎

We also measured the hepatic vein resistance index of the right ‎hepatic vein (HVRI) and there was no significant difference between the ‎two studied groups and there was no correlation with fibrosis score in ‎liver biopsy or in fibroscan results (p > 0.05). as far we know, the only ‎study conducted with the estimation of HVRI by Lutz et al. (2012 ) ‎who studied 125 patients (with different grade of fibrosis from F0 to F4‎‎) the results showed that HVRI associated with the stage of fibrosis ‎showed the patients without fibrosis exhibited uppermost scores and no ‎significant difference between the different stages of fibrosis was noticed ‎and that HVRI could detect fibrosis of the F2 or higher stages, but it is ‎of limited value in detecting mild fibrosis and that it has superior ‎sensitivity and specificity than fibroscan and HVRI values were not ‎affected significantly by different stages of inflammation or steatosis. ‎This difference may be due to the difference in the study population as ‎they included patients with mixed causes of hepatic cirrhosis while in ‎our result we included only HCV- infected patients who were not ‎cirrhotic [26].‎

The limitation of the study: this was a retrospective study that ‎leads to losing of some records and data .so numbers of cases in our ‎study were less; and it was a single center study.‎

CONCLUSION

Fibroscan has a good matching with liver biopsy in the detection of ‎higher grades of fibrosis and it lacks accuracy in detecting lower grades ‎of fibrosis. Fibroscan results were affected by grade of inflammation in ‎liver biopsy which may cause overestimation of grade of fibrosis.  ‎However, Doppler US is more sensitive in detecting lower grades of ‎fibrosis than higher grades in HCV infected patients‎.

Funding:  None. ‎

Conflict of interest: none.‎

Ethical consideration ‎

Institutional ethical committee approval was obtained before the start of ‎the study. The study protocol conforms to the ethical guidelines of the ‎‎1975 Declaration of Helsinki as reflected in a prior approval by the ‎institution's human research committee.‎