Liver biopsy remains the reference method in order to determine the grade of fibrosis in chronic liver diseases,
but it is an invasive and painful procedure,
which can study only 1/50,000 of the total volume of the liver [5,10,19].
Liver biopsy has also several complications related to its invasiveness (e.g.
patient’s dis-comfort and bleeding),
and an overall procedure-related mortality of 1/1000–10,000 [6].
A heterogeneous distribution of fibrosis in the liver may become a limitation for biopsy,
since only a small portion of parenchyma is sampled during the procedure.
In more than half of cases,
the specimens obtained with a 17-gauge needle have an average length and number of portal tracts well below the minimum sample size requirements (i.e.
10 or more portal tracts) [7].
In the field of non-invasive assessment of liver fibrosis,
ultrasound-based methods play a pivotal role.
TE has been validated in different chronic liver diseases,
including chronic viral hepatitis (HCV and HBV),
alcoholic steatohepatitis,
non-alcoholicsteatohepatitis (NASH),
autoimmune hepatitis,
primary biliary cirrhosis and hepatopathy related to iron overload [10–13,19,21].
Di Marco et al.
[18] performed TE examinations in a cohort of 56 homozygous-β-thalassemic patients,
finding that liver stiffness increased proportionally to liver fibrosis stages detected by liver biopsy (r = 0.70; P > 0.001),
without lack of interference by iron deposits in the parenchyma (i.e.
LIC).
Adhoute et al.
[11] studied the utility of TE and other non-invasive methods in patients with genetic hemochromatosis (C282Y homozygosity).
They included 57 cases with 46 controls,
obtaining a strong correlation between TE and many biochemical markers,
although ferritin levels did not correlate with TE values.
The prevalence of patients with TE values higher than 7.1 kPa (cut-off level for significant fibrosis),
was22.8% in patients with hemochromatosis and 0% in the controls (P < 0.0001).
Fraquelli et al.
[12] evaluated 115 adult thalassemicpatients (59 with β-thalassemia major and 56 with thalassemiaintermedia) with TE,
confirming the significant positive correlation between liver stiffness and fibrosis stage (r = 0.73,
P = 0.003).
In a cohort of 42 transfusion-independent adult patients with β-thalassemia intermedia,
it has been demonstrated that longitudinal changes in serum ferritin levels correlate with TE values of hepatic stiffness during a 4-year follow-up (r= 0.836,P< 0.001) [25].
To our knowledge,
no study investigated the role of RTE in the non-invasive assessment of liver fibrosis in patients with primary and secondary hemochromatosis.
While TE requires a dedicated equipment,
RTE has been implemented on high-end ultrasound systemsby different manufacturers.
In our study we used the RTE module of the ultrasound system MyLab Twice (Esaote,
Genoa,
Italy).
This RTE module differs from that produced by Hitachi due to a shortest range of pixel’s values (i.e.
from 0 to 100 vs the Hitachi 256 step-wise grading) to represent strain distribution and tissue elasticity.
It uses a conventional ultrasound probe to compare and analyze echosignals from the tissue under examination before and under slight compression [26,27],
generating a color-coded map of strain distribution,
called elastogram.
Since elastic ratio is the ratio of strain distribution in two selected ROIs,
some Authors prefer to use small parenchymal blood vessels (<3 mm) as internal control [30].
In our study,
we selected as internal control a homogeneous red (soft) area in the diaphragm,
considering this perihepatic muscular structure less prone to inter-individual variability.
Using this method of elastogram analysis,
mean elasticity ratios significantly differed from patients with severe fibrosis (METAVIR stages F3,
F4) to those with mild fibrosis (METAVIR stages F0/1,
F2) (P< 0.05).
Elastic ratios above the cut-off value of 2.75 identified patients at risk for severe fibrosis (F3 and F4) with a sensitivity of 70% and a specificity of 97.5%.
From the results of the AUC-ROC curve analysis,
it is evident that the best cut-off value for detectingF= 4 is the same of that for identifying patients withF≥3 (i.e.≥2.75).
This result may be due to different factors,
including the small sample size of F4 patients andthe lack of a significant difference between the mean elastic ratios of F3 and F4 fibrosis stages.
This latter result may be rea-sonably related to the exclusion of patients with decompensated liver cirrhosis from the study.
Different investigators criticized thelack of inter-observer agreement in hepatic RTE [28,31],
becausethe operator’s freehand compression of the probe is a parameter difficult to be standardized.
In order to overcome this problem,
newer RTE modules are able to produce elastograms in responseto the pressure generated by heartbeats [22,23].
However,
some Authors found good inter-observer agreement (intraclass correla-tion coefficient = 0.906,P< 0.001) performing RTE by a free-hand compression approach and calculating the elastic ratio using a method similar to ours [29].
In our work,
the mean value of the 10 elastic ratios obtained from each patient by a single experienced operator (who performed more than 100 RTE evaluations according to [30]) was significantly correlated with TE values (r= 0.645,95% CI 0.468–0.772,P< 0.0001).
With TE,
tissue stiffness is estimated along an ultrasonic A-line,
in a fixed region,
which is neither user adjustable nor image guided [31].
The most important limit of this one dimensional approach is that provides an average elasticity estimated over the measurement depth; for this reason we decided to exclude patients with a inhomogeneous pattern of liver iron overload,
which may lead to a patchy deposition of fibrous tissue within the liver parenchyma.
On the other hand,
RTE allows the study of a wider portion of liver parenchyma,
allowing to precisely detect large vessels and nodular lesions inside the sampling area,
which can influence the strain response [26].
The common limitations of both TE and RTE include obesity,
narrow intercostal spaces and ascites [10,20].
The major drawback of our study is that liver biopsy was not systematically used as reference standard; however,
TE has been proven to be 99% efficient for the detection of patients with cirrhosis and 88% efficient for the detection of patients with fibrosis grade superior or equal to F2 [21].
When performing TE,
the elasticity estimate is averaged over a volume that can be approximated by a cylinder of length 20 mm (between 25 mm and 45 mmbelow skin surface) and diameter 20 mm.
This volume represents 1% of the liver total volume,
which is much more relevant than the biopsy sample size,
which is only of 1/50,000 [19].
TE cut-off values used is our study were formerly obtained by Ziol et al.
[21] from a large cohort of 327 patients with chronic viral hepatitis C and validated using liver biopsy as reference standard.
In a more recent work comparing TE,
RTE and aspartate-to-platelet ratio index in the assessment of liver fibrosis [20],
Ferraioli et al.
found that TE is able to accurately assess significant fibrosis (F≥2) with a high specificity 91.4%,
offering an excellent diagnostic performance in the discrimination of severe fibrosis and cirrhosis with a NPV of 99%.
Focusing on differentiating non-significant from significant fibrosis(F≥2),
TE and aspartate-to-platelet ratio index,
with an overlapping AUC (0.88 and 0.86),
performed better than RTE.
Even thoughit is still considered the reference standard,
liver biopsy may fail in the assessment of the degree of liver fibrosis because it is subject to intra- and interobserver variability and to sampling errors,
even when the biopsy length is adequate [32].
In addition,
the METAVIR scoring system takes into account not only parenchymal fibrosis,
but also architectural changes in the liver,
without reference to quantitative changes in liver collagen [32].
So,
like MRI is rapidly replacing liver biopsy for LIC quantification,
TE is increasingly being used to study liver fibrosis in hematologic disorders leading to iron overload.
The use of TE in this clinical concern has raised the question about the possible interference of iron deposition with liver stiffness measurement.
Fraquelli et al.
[12] examined the cross-sectional association between MRI T2* values of liver iron overload and TE measurements in 73 adult thalassemia patients (47 with thalassemia major and 26 with thalassemia intermedia).
Median LIC values were 4.58 mg/gdw(range 1.02–19.7) in patients with thalassemia major and 5.98 mg/gdw(range 1.11–19.02) in those with thalassemia intermedia.
In both groups,
no correlation was found between LIC and TE results (r=−1.4257 andr= 0.09).
In their cohort of 56 thalassemia major patients,
Di Marco et al.
[18] have found that TE values of liver stiffness are independently correlated only with the stage of fibrosis (P= 0.002) and aspartate aminotransferase(P= 0.006) values,
concluding that TE measurements are not influenced by iron deposition.
Examining a smaller cohort of patients,
other Authors [33] have found a significant correlation of TE values with ferritin (r= 0.34,P= 0.01),
T2* (r=−0.46,P= 0.003) and LIC values retrieved from MRI (r= 0.54,P< 0.0001),
suggesting that it is not possible to completely exclude the interference of iron deposition with liver stiffness measurements.
However,
these Authors did not perform a multivariate analysis in their study.
We suggest that RTE could be used as a complementary imaging method to TE for a preliminary assessment of liver fibrosis.
In particular,
RTE may be performed immediately after a standard B-mode ultrasound examination of the liver,
allowing to discriminate between F0/1–F2 vs F3–F4 with a reasonable diagnostic accuracy.
However,
in its present form,
RTE cannot replace TE for assessing liver fibrosis in patients with iron overload.