Patient inclusion:
This was a prospective monocentric study and patient’s enrollment was performed at the Unit of Microcitemia and Hereditary Anaemias of our Institution,
where subjects with different diseases leading to liver iron overload are routinely evaluated with MRI.
Consecutive patients with MRI T2* detectable hepatic iron (liver T2*value≤6.3 ms) were enrolled into the study.
The study protocol was approved by the institutional review board.
Written informed consent was obtained from all patients before the MRI study.
Exclusion criteria were general contraindications to 1.5 T MRI [14],
decompensated liver cirrhosis with ascites (which can influence both TE and RTE results [10]),
and the presence of an inhomogeneous patchy pattern of iron deposition detected by MRI.
This latter criterion was introduced to avoid sampling bias when performing TE,
since it is likely that a patchy iron deposition may lead to inhomogeneous fibrosis throughout the liver parenchyma [8].
Serum ferritin levels were determined using a standard laboratory method performed within 1 month from the imaging examinations (normal adult blood levels are 12–300 ng/mL for males and 10–150 ng/mL for females [5]).
Eligible patients were also screened for the presence of hepatitis C virus (HCV) antibodies and hepatitis B surface antigen (HBsAg) in serum.
Magnetic resonance imaging:
MRI is considered a reliable method for detecting iron deposits in the liver [15],
and gradient-echo sequences are used to quantify the proton-transverse relaxation through transverse relaxation time (T2*) measurements.
The reciprocal of T2*,
known as transverse relaxation rate (R2*),
increases in the presence of iron and is proportional to LIC over the clinically relevant range [16].
Breath-hold R2*-MRI measurements were performed using an eight-element cardiac/torso coil in a 1.5T Signa HDx scanner (General Electric Medical Systems,
Milwaukee,
WI,
USA) scanning the whole liver of the patients.
To obtain quantitative R2* maps,
a multigradient echo sequence with the following parameters was used: eight echoes and minimum echo spacing,
echo times (TE)1.1–6.7 ms,
repetition time (TR) 200 ms,
flip angle 20◦,
matrix128×96 pixels,
bandwidth 125 kHz,
number of excitations 1,
slicethickness 10 mm,
and spacing 0 mm.
The duration of each sequencewas 20–30 s.
Measurements of R2* were performed using a publicly available software (C-Iron,
Camelot Biomedical Systems,
Genoa,
Italy; website: http://c-iron.camelotbio.com) and the signal decay was fitted to every pixel in the image to an exponential plus a constant function.
A region of interest (ROI) comprising the whole liver and excluding blood vessels and biliary ducts was drawn from a transverse midhepatic slice.
Hepatic iron overload was defined by MRI T2* values less than 6.3 ms,
corresponding to a liver iron concentration in dry tissue (LIC dry weight) of 4.2 mg/g.Hepatic iron overload was categorized as mild (6.3–2.7 ms),
moderate (2.6–1.4 ms) or severe (<1.4 ms) [17].
Transient elastography:
TE is a corroborate method for the assessment of liver fibrosis in patients with hemochromatosis [11–13],
since it has been shown that iron overload does not influence the diagnostic accuracy of this technique [18].
TE was performed with FibroScan (Echosens,
Paris,
France).
In this device an ultrasound probe,
mounted on the axis of a vibrator,
transmit slow-frequency vibrations from the right intercostal space,
creating an elastic shear wave that propagates into the liver.
A pulse-echoultrasound acquisition is then used to detect the wave propagation velocity,
which is proportional to tissue stiffness; faster wave progression occurs through stiffer material.
Liver stiffness measurement is then performed and measured in kiloPascals (kPa)(values between 2.5 kPa and 75 kPa are expected) [19]. Measurements of liver stiffness were performed on the right lobe of the liver through intercostal spaces in correspondence to the mid-axillary line,
while patients are lying in the supine position with the right arm in maximal abduction.
Only patients with10 correct measurements with an interquartile range (IQR) of less than 30% of the median liver stiffness value were included [20].
TE values were expressed in kilopascals (kPa); further they were converted in the corresponding semi-quantitative fibrosis score of METAVIR.
It is based on a semi-quantitative 5-point scale from 0 to 4: F0,
the absence of parenchymal fibrosis; F1,
enlarged fibrotic portal tract; F2,
periportal or initial portal-portal septa but intactarchitecture; F3,
architectural distortion but no obvious cirrhosis; and F4,
cirrhosis.
The conversion of TE values into the corresponding METAVIR stage was made by means of validated cut-off values (i.e.
F0/F1 vs F2–F4 = 8.8 kPa; F0/F1–F2 vs F3–F4 = 9.6 kPa; F0/F1–F3vs F4 = 14.6 kPa),
which were obtained in a previous study by Ziol et al.
using biopsy as reference standard [21].
Real-time sonoelastography:
A radiologist (FP) with more than 5 years of experience in conventional ultrasound examinations and 1 year of experience in RTE,
blinded to TE results,
consecutively performed all RTE examinations.
RTE measures mechanically probe-induced deformation (strain) of structures examined in the B-mode ultrasound image,
generating color-coded maps of the strain distribution (i.e.
elas-tograms),
which reflect tissue elasticity [22,23].
The RTE module displays two images simultaneously: the conventional B-mode image and the color-coded elastography region of interest (ROI),
overlaid on the B-mode image (Fig.
1).The system generates a color map where hard tissue areas are marked with blue,
intermediate tissue areas with green,
and soft tissue areas with red.
In the Esaote elastographic module (Elaxto) numerical values of pixels are from 0 to 100 (100 stepwise grading) according to color mapping from blue (0) to red (100).
It is possible to generate a histogram of pixel distribution derived from the color image by 100 stepwise grading.
The examinations were performed on the right lobe of the liver through the intercostal spaces in correspondence to the mid-axillary line,
applying gentle pulsed compression on the skin.
We used an original multi-frequency linear probe with a range of 3–11 MHz (LA332 apple probe,
Esaote,Genoa,
Italy).
The small array (3 cm) of the transducer guarantees a perfect coupling with the patient’s intercostal space,
while its trapezoidal (“convex-like”) view is more panoramic that the conventional view of a linear probe.
The elastogram ROI was positioned in the most superficial portion of the image,
which is not affected by degradation or distortion.
Patients were instructed to continue breathing as usual,
because each elastography image is obtained in a few milliseconds and breathing did not cause any motion artifacts.
The acquisitions were considered reliable only if a pressure of 3–4 at least on a scale of 0–6 arbitrary units was recorded.
Such indicator is a simple feedback for the operator to indicate that the movement of tissues subjected to compression is more or less suitable to the rate of acquisition of the system.
The elastography ROI positioning was made according to a method formerly proposed by Saftoiu et al.
[24,28],
where the elastogram includes the perihepatic soft-tissues,
such as the diaphragmand intercostal muscles,
in order to clearly compare and distinguish the strain between the liver and these structures.
The abdominal wall layers visualized through the right intercostal spaces include the skin,
subcutaneous fat tissue,
intercostal muscles (externaland internal),
diaphragm,
and liver parenchyma.
In some patients,
even the thin adipose tissue between the Glisson’s capsule andvisceral peritoneum (tela subserosa) can be depicted by the elastog-raphy software [24].
To standardize measurements,
a rectangular ROI of 25 mm×20 mm was positioned to visualize the strain map of the inner perihepatic soft tissues (internal intercostal muscle,
diaphragm,
perihepatic fat tissue) and a superficial portion of liver parenchyma free of large vessels.
Ten elastograms were acquired for each patient.
The qualitative information provided by the elastograms was converted in quantitative data by calculating the elastic ratio,
which is the ratio of strain distribution in two selected regions of interest (ROIs).
A large elliptical ROI of 1 cm2(Z1) was positioned in the liver parenchyma near the center of the image and a smaller elliptical ROI of 2 mm2(Z2) was positioned in a homogeneously soft region of the diaphragm,
which was considered as internal control to calculate the elastic ratioZ2/Z1.
The diaphragm was chosen as internal control since it appeared quite homogeneously soft (when compared to the liver parenchyma) in all patients.
A higher elastic ratio indicates harder hepatic elasticity,
corresponding to a higher stage of fibrosis (Fig.
2).
Our method to calculate the elastic ratio was similar to that of Xie et al.
[29],
who reported excellent values of intra- and inter-observer agreement (intraclass correlation coefficient = 0.906,P< 0.001) using a free-hand compression approach.
ROI positioning and calculation of the elastic ratios were performed on a personal computer by a trainee in radiology,
blinded to TE results,
using the original software MyLabDesk (Esaote,
Genoa,
Italy).
Elastic ratios were calculated on the 10 color-coded images obtained from each patient,
and the mean value was used for further statistical analysis.
Each RTE examination lasted about 5 min with other 10 min to calculate the mean elastic ratio.
The mean time necessary to complete a RTE exami-nation was almost the same across the duration of the study.
On the other hand,
the time needed for the off-line analysis of the elastograms was seen to progressively shorten according to the increasing confidence of the operator with the software interface.
Statistical analysis:
Descriptive statistics were produced for demographic,
clinical,
and laboratory characteristics of patients.
The Shapiro–Wilk test was used to test the normality (i.e.
Gaussian distribution) of T2*,TE and RTE values.
Box plots were used to study the distribution of RTE elastic ratios according to the patient’s stage of fibrosis,
and the presence of significant differences in the mean elastic ratio among the four METAVIR stages was assessed by the one-way analysis of variance (ANOVA).
The correlation between fibrosis stage and elastic ratio was calculated via the Spearman’s rank order correla-tion coefficient.
The Pearson correlation coefficient was used to test associations between variables with a normal distribution.
Nominal statistical significance was defined with aPof 0.05.
The diagnostic performance of RTE was assessed by using receiver operating characteristic (ROC) curves.
For the ROC curve analysis,
the area under curve (AUC),
optimal cut-off values,
sensitivity,
specificity,
positive and negative predictive values were calculated.
Optimal cut-off values of RTE elastic ratio were chosen to maximize the sum of sensitivity and specificity for different fibrosis thresholds: F0–F1vs F2–F4 (F≥2),
F0–F2 vs F3–F4 (F≥3),
and F0–F3 vs F4 (F= 4).