Transcript USE OF MRI IN EVALUATING LIVER IRON LOADING (AND MONITORING

USE OF MRI
IN EVALUATING
LIVER IRON LOADING
(AND MONITORING
THERAPY)
G-EXJ-1030713
May 2012
NOTE: These slides are for use in educational oral presentations only. If any published figures/tables from these slides are to be used for
another purpose (e.g. in printed materials), it is the individual’s responsibility to apply for the relevant permission.
Specific local use requires local approval
Outline
● Introduction to iron and liver iron overload
● Key methods for assessing liver iron
– liver biopsy
– SF
– SQUID
– liver MRI
•
SIR method
•
relaxometry methods (R2 and R2*)
● Clinical recommendations for measuring LIC
● Summary
2
G-EXJ-1030713
May 2012
LIC = liver iron concentration; MRI = magnetic resonance imaging;
SF = serum ferritin; SIR = signal intensity ratio;
SQUID = superconducting quantum interface device.
Introduction
to iron
and iron overload
G-EXJ-1030713
May 2012
Iron overload
● Iron overload is common in patients who require intermittent or regular
blood transfusions to treat anaemia and associated conditions
– it may be exacerbated in some conditions by excess gastrointestinal
absorption of iron
● Iron overload can lead to considerable morbidity and mortality1
● Excess iron is deposited in major organs, resulting in organ damage
– the organs that are at risk of damage due to iron overload include
the liver, heart, pancreas, thyroid, pituitary gland, and other
endocrine organs2,3
4
G-EXJ-1030713
May 2012
1Ladis
V, et al. Ann NY Acad Sci. 2005;1054:445-50. 2Gabutti V, Piga A. Acta Haematol. 1996;95:26-36. 3Olivieri NF.
N Engl J Med. 1999;341:99-100.
Importance of analysing liver iron
● A patient’s LIC is the best measure of total body iron stores
● Knowing the liver iron concentration helps to predict the risk of hepatic
and extra-hepatic complications1–4
5
G-EXJ-1030713
May 2012
1Batts
KP. Mod Pathol. 2007;20:S31-9. 2Jensen PD, et al. Blood. 2003;101:91-6. 3Angelucci E, et al. Blood.
2002;100:17-21. 4Telfer PT, et al. Br J Haematol. 2000;110:971-7.
Mean LIC + SD over previous year prior
to enrolment in EPIC trial
(mg Fe/g dry wt)
Importance of analysing liver iron (cont.)
25
20
15
10
LIC
threshold
of 7 mg
Fe/g dry wt
5
0
All
(n = 1,744)
TM
(n =
937)
TI
(n = 84)
SCD
(n = 80)
All transfusion-dependent patients prior to study enrolment
had moderate-to-severe hepatic iron loading
6
G-EXJ-1030713
May 2012
Cappellini MD, et al. Blood. 2008;112:[abstract 3880].
Overview of LIC correlations with other
measurements
Body iron
stores1
LIC
Cardiac
iron5
7
G-EXJ-1030713
May 2012
Hepatocellular
injury2 and
fibrosis3
Cardiac
DFS4
DFS = disease-free survival.
1Angelucci E, et al. N Engl J Med. 2000;343:327-31. 2Jensen PD, et al. Blood. 2003;101:91-6. 3Angelucci E, et
al. Blood. 2002;100:17-21. 4Telfer PT, et al. Br J Haematol. 2000;110:971-7. 5Noetzli LJ, et al. Blood.
2008;112:2973-8.
LIC prediction of total body iron stores
β-TM2
Hereditary haemochromatosis1
Sample > 1 mg dry wt (n = 25)
Body iron stores (mg/kg)
50,000
LIC (µg/g)
40,000
30,000
20,000
10,000
0
0
5
10
15
Iron removed (g)
20
25
300
250
200
r = 0.98
150
100
50
0
0
5
10
20
LIC (mg Fe/g dry wt)
LIC is a reliable measure of total body iron stores in
hereditary haemochromatosis and β-TM
8
G-EXJ-1030713
May 2012
15
BMT = bone marrow transplantation.
1Olynyk JK, et al. Am J Gastroenterol. 1998;93:346-50. 2Angelucci E, et al. N Engl J Med. 2000;343:327-31.
25
Serum ferritin measurement alone
underestimates the body iron load
10,000
14,000
-TI
-TM
12,000
9,000
8,000
7,000
SF (g/L)
SF (g/L)
10,000
8,000
6,000
6,000
5,000
4,000
3,000
4,000
2,000
2,000
0
-TI
 -TM
1,000
0
5
10 15 20
25 30 35
LIC (mg Fe/g dry wt)
0
0 5 10 15 20 25 30 35 40 45 50
LIC (mg Fe/g dry wt)
SF has almost no sensitivity or specificity for iron
stores in thalassaemia intermedia
9
G-EXJ-1030713
May 2012
Origa R, et al. Haematologica. 2007;92:583-8.
Taher A, et al. Haematologica. 2008;93:1584-6.
Assessing liver
iron overload
G-EXJ-1030713
May 2012
Key methods for assessing liver iron
● Liver biopsy  LIC
– advantages and disadvantages
Direct method
– correlation of LIC with other measurements
● SF concentration over time
– advantages and disadvantages
– correlation of SF levels with other measurements
● SQUID
– advantages and disadvantages
● Liver MRI
– advantages and disadvantages
– relaxometry methods (T2 and T2*)
– SIR method
11
G-EXJ-1030713
May 2012
Olivieri NF, Brittenham GM. Blood. 1997;89:739-61.
Indirect methods
Liver biopsy
G-EXJ-1030713
May 2012
Technique for taking a percutaneous
liver biopsy
Liver biopsy
A tiny incision is made between the ribs, and a needle is
inserted to reach the area of the liver where a tissue sample
is taken. The procedure requires local anaesthesia
Patient preparation: Blood tests are done shortly
before the biopsy to check blood clotting time, to
exclude risk of bleeding following the biopsy. The
biopsy is commonly preceded by an ultrasound
examination of the liver to determine the best and
safest biopsy site
Step 1. The patient lies on his back, or his left side
Area where a tissue
sample is taken from
Step 2. The place for the biopsy is cleaned with
antiseptic and local anaesthesia is provided (s.c. on
the right hand side)
Step 3. A special hollow needle is inserted into the
liver, usually between the 2 lower ribs on the right
hand side
Step 4. The patient must hold breath for 5-10 seconds
when the needle is quickly pushed in and out. As the
needle comes out it brings with it a small sample of
liver tissue
adam.com
13
G-EXJ-1030713
May 2012
Overall: The procedure is carried out by a qualified
physician or surgeon in an outpatient care centre or
hospital. It is fast (not longer than 5 min) and the
patient is discharged shortly after
Processing the liver biopsy sample
● Gross histopathological examination
– reveals presence of abnormal cells
or liver tissue
– used to determine presence and
degree of cirrhosis and fibrosis
● LIC measurement
– by iron staining
– by atomic absorption spectroscopy:
the current gold standard!
● Who does the test?
– preparation of the samples might be by
a trained technician
– the analysis requires a qualified pathologist
14
G-EXJ-1030713
May 2012
Angelucci E, et al. Haematologica. 2008;93:741-52.
Image from: www.pathguy.com/lectures/cirrhosis_trichrome.jpg
Liver biopsy
Liver biopsy with iron measurement by atomic absorption
spectroscopy is the gold standard for measuring LIC1
LIC threshold
(mg Fe/g dry wt)2
LIC threshold
(mol Fe/g dry wt)
Clinical relevance
1.8
32
Upper 95% of normal
 15.0
269
Greatly increased risk of cardiac disease and
early death
15
G-EXJ-1030713
May 2012
1Angelucci
E, et al. Haematologica. 2008;93:741-52. 2St Pierre TG, et al. Blood. 2005;105:855-61.
Liver biopsy: pros and cons
Pros1
Cons
●
●
●
●
Direct measurement of LIC
Validated reference standard
Quantitative, specific, and sensitive
Allows for measurement of
non-haem storage iron
● Invasive and painful procedure with risk of potentially serious
complications1
● May involve sampling errors, especially in patients with cirrhosis1
● Requires skilled physicians1
● Laboratory techniques not standardized1
● Provides information on liver
histology/pathology
● Correlates with morbidity and mortality
– iron measurement by atomic absorption spectroscopy2 or
chemical determination3
● wet or dry weight quoted
● iron concentration varies throughout the liver,4 sample size often
insufficient (requires ≥ 1 mg dry weight, or > 4 mg wet weight)
16
G-EXJ-1030713
May 2012
1TIF.
Guidelines for the Clinical Management of Thalassemia. 2nd rev. ed. Cyprus: TIF; 2008. Available from:
www.thalassaemia.org.cy/pdf/Guidelines_2nd_revised_edition_EN.pdf. Accessed December 2010. 2Angelucci E,
et al. Haematologica. 2008;93:741-52. 3Wood JC. Blood Rev. 2008;22 Suppl 2:S14-21. 4Ambu R, et al. J Hepatol.
1995;23:544-9.
Heterogeneity of iron concentration
throughout the liver
0–20%
20–40%
40–60%
60–80%
80–100%
Iron is unevenly distributed in the liver; therefore, a small
sample may not give an absolutely representative mean LIC
17
G-EXJ-1030713
May 2012
From autopsy of a patient with beta-zero-thalassaemia.
Ambu R, et al. J Hepatol. 1995;23:544-9.
SF Concentration
G-EXJ-1030713
May 2012
Ferritin and SF
●
●
Ferritin is primarily an intracellular
protein that
–
stores iron in a form readily accessible to
cells
–
releases iron in a controlled fashion
The molecule is shaped like a hollow sphere
and it stores ferric (Fe3+) iron in its central
cavity
–
SF > 1,000 µg/L is a marker of
excess body iron
19
G-EXJ-1030713
May 2012
the storage capacity of ferritin
is approximately 4,500 Fe3+ ions
per molecule
●
Ferritin is found in all tissues, though
primarily in the liver, spleen, and
bone marrow
●
A small amount is also found in
the blood as serum ferritin
Harrison PM, Arosio P. Biochim Biophys Acta. 1996;1275:161-203.
SF: pros and cons
● SF levels from a blood sample are measured
Pros
Cons
● Easy to assess
● Inexpensive
● Positive correlation with morbidity
and mortality
● Allows longitudinal follow-up of patients
● Indirect measurement of iron burden
● Fluctuates in response to inflammation,
abnormal liver function, ascorbate
deficiencies
20
G-EXJ-1030713
May 2012
TIF. Guidelines for the Clinical Management of Thalassaemia. 2nd rev. ed. Cyprus: TIF; 2008. Available from:
www.thalassaemia.org.cy/pdf/Guidelines_2nd_revised_edition_EN.pdf. Accessed December 2010.
SQUID
G-EXJ-1030713
May 2012
SQUID: superconducting quantum
interference device
Principle of the technique: Normal tissue is diamagnetic and has a
magnetic susceptibility similar to that of water. In the presence of iron,
tissue susceptibility is changed proportional to the amount of iron
present. This alteration is detected, allowing non-invasive
measurement of LIC
Magnetizing coil
Dewar
Patient preparation: No special patient preparation is required.
Ultrasound is used to evaluate the depth and size of the liver. The
patient lies on their back with their torso surrounded by a 5-L water bag
to minimize contributions from other tissues
Step 1. The susceptometer applies a low-power (114 T and 7.7 Hz)
homogeneous magnetizing field in the hepatic region. Sensitive
detectors measure the interference of tissue iron vs the water reference
medium within the field
Step 2. LIC corresponds to the variation of magnetization detected and
is calculated using custom-made Matlab 6.5 software
Overall: The procedure is carried out by a qualified radiologist in a
hospital. It is fast (not longer than 5 min) and the patient is discharged
immediately after. Processing could be done on the spot and is faster
then LIC histopathological examination
22
G-EXJ-1030713
May 2012
Carneiro AA, et al. Reson Med. 2005;43:122-8.
Liquid
helium
H2O
SQUID
Pick to coil
Water bag
Patient
Mattress
Bed
Piston
SQUID: pros and cons
Pros
Cons
● Non-invasive1
● Wide linear range1
● Good correlation with LIC by biopsy2
● Requires expensive, specialized
equipment and expertise1
● Not widely available1
● Each machine should be
individually calibrated1
● SQUID can underestimate LIC3
Hepatic iron (biopsy)
(mol Fe/g wet wt)
250
R = 0.99
p < 0.001
200
150
100
50
0
0
50
100
150
200
250
Hepatic iron (magnetic) (mol Fe/g wet wt)
SQUID is a non-invasive method that has been calibrated,
validated, and used in clinical studies, but the complexity, cost
and technical demands limit its use
23
G-EXJ-1030713
May 2012
1TIF.
Guidelines for the Clinical Management of Thalassaemia. 2nd rev. ed. Cyprus: TIF; 2008. Available from:
www.thalassaemia.org.cy/pdf/Guidelines_2nd_revised_edition_EN.pdf. Accessed December 2010. 2Sheth S. Pediatr
Radiol. 2003;33:373-7. 3Piga A, et al. Blood. 2005;106:[abstract 2689].
Liver MRI
G-EXJ-1030713
May 2012
MRI
Principle of the technique: A strong magnetic field is used to
organize the protons in the tissue in 1 direction. Then radiofrequency is
used to “knock” them off. The time for them to re-align with the
magnetic field and the energy they release during the process depend
on the interactions of the proton with other ions, notably iron ions.
These events could be measured at various TEs and then analysed to
reveal the iron content in the tissue
Patient preparation: All infusion and medication pumps should be
removed. The scan does not require contrast agent, and so no
peripheral vein access is needed
Main
magnet
coils
x,y,z
gradient
coils
Step 1. Image acquisition: Images are taken at various TEs
Step 2. Post-processing: As TE increases, the image’s SI decreases.
The relationship between TE and SI in a selected part of the image (i.e.
ROI) is analysed with specialized software or manually. Data are
reported as relaxation times (T2 or T2*), depending on the acquisition
method
Overall: The procedure is carried out by a qualified radiologist in a
hospital. Acquisition is fast (approx. 5 min), and the patient is
discharged immediately after. Processing may require specialized
software and is done afterwards
25
G-EXJ-1030713
May 2012
ROI = region of interest; SI = signal intensity; TE = echo time.
Brittenham GM, Badman DG. Blood. 2003;101:15-9.
Ridgway JP. J Cardiovasc Magn Reson. 2010;12:71.
Patient table
Integral
radiofrequency
transmitter
(body) coil
Main
magnet
coils
MRI is increasingly being used
as a non-invasive method to measure LIC
Pros
Cons
●
●
●
●
● Indirect measurement of LIC2
● Requires MRI with dedicated imaging method2
● Sensitivity depends on type of scanner, degree
of iron overload, presence of fibrosis, and
inflammation7
Non-invasive1,2
Assesses iron content throughout the liver2
Increasingly and widely available worldwide2
Pathological status of liver and heart can be
assessed in parallel2
● Validated relationship with biopsy LIC3‒6
1Chavhan
26
G-EXJ-1030713
May 2012
GB, et al. Radiographics. 2009;29:1433-49. 2TIF. Guidelines for the Clinical Management of Thalassaemia.
2nd rev. ed. Cyprus: TIF; 2008. Available from: www.thalassaemia.org.cy/pdf/Guidelines_2nd_ revised_edition_EN.pdf.
Accessed December 2010. 3Christoforidis A, et al. Eur J Haematol. 2009;82:388-92.
4St Pierre TG, et al. Blood. 2005;105:855-61. 5Wood JC, et al. Blood. 2005;106:1460-5. 6Hankins JS, et al. Blood.
2009;113:4853-5. 7Sirlin CB, Reeder SB. Magn Reson Imaging Clin N Am. 2010;18:359-81.
MRI scanners
●
Manufacturers
– Siemens Healthcare (Erlangen, Germany; www.siemensmedical.com)
– GE Healthcare (Milwaukee, WI, USA; www.gemedicalsystems.com)
– Philips Healthcare (Best, the Netherlands; www.medical.philips.com)
●
Magnetic field strength
●
–
most imaging is done on 1.5 T machines
–
3 T machines give
• better signal:noise ratio1
• worse susceptibility artefacts1
• The upper detection limit is halved, therefore it is too low for many patients1
• lower T2 and T2* values than 1.5 T machines2
Liver package (including standard sequences and analysis of the data)
is included in the software provided together with the MRI machine
–
specialized LIC analysis software can be bought separately
27
G-EXJ-1030713
May 2012
1Wood
JC, Ghugre N. Hemoglobin. 2008;32:85-96. 2Storey P, et al. J Magn Reson Imaging. 2007;25:540-7.
Overview of MRI techniques used to
measure LIC
DATA
ACQUISITION
Signal Intensity
Ratio (SIR) method
(Gandon/Ernst)
DATA ANALYSIS
A combination of
gradient and spin
echos
Free website
Gradient echo (same
technique as cardiac
iron measurement)
(1 min)
Manually (free xls
sheet) or with
dedicated software
(e.g CMR tool 3,000
GBP per year)
Liver MRI
Technique
R2*(T2*)
Relaxometry
method
R2(T2)
(Ferriscan®)
28
G-EXJ-1030713
May 2012
Spin echo
(15min)
Done centrally by
Resonance Health
(300 USD per scan)
MAJOR PROS
AND CONS
+ Fast acquisition
Simple data analysis
− Limited sensitivity
Reproducibility
+ Fast acquisition
Correlates well with LIC
− Susceptible to artefacts
Training needs
+ Gold Standard
Little training need
− Longer data acquisition time
Cost of analysis
MRI measurement of LIC: techniques
● There are 2 broad groups of techniques
– SIR methods (Gandon et al. methods)
– relaxometry methods (FerriScan® and T2* (R2*) methods)
SIR
method
Relaxometry
method
29
G-EXJ-1030713
May 2012
Pros
Cons
● Fast data acquisition
● Relatively simple algorithms and
data analysis
● Limited range of sensitivity (upper limit
is 21 mg Fe/g dry wt [380 mol/L])
● Assumptions on reference tissue
● Can be used in scanners with
different magnetic strengths
(0.5, 1.0, 1.5 T)
● Greater range of sensitivity
● Does not rely on reference tissue
assumptions
● T2* (or R2*) is very quick
(requires a single breath-hold)
● Not reliable in cirrhosis
● Smaller reproducibility
● Has only been calibrated at 1.5 T
● Takes longer to acquire data, when
done as T2 (or R2)
Argyropoulou MI, Astrakas L. Pediatr Radiol. 2007;37:1191-200. Gandon Y, et al. Lancet. 2004;363:357-62. St
Pierre TG, et al. Ann N Y Acad Sci. 2005;1054:379-85. Wood JC. Curr Opin Hematol. 2007;14:183-90. Wood JC,
et al. Blood. 2005;106:1460-5.
SIR methods
1. Patient
preparation
2. Image
acquisition
(5 min)
●
3. Data analysis
(depends on
experience)
(approx. 5-20 min)
Most common protocol includes
–
4-gradient echo sequences with different TEs
–
1 spin-echo sequence
400
MRI LIC
(µmol Fe/g dry wt)
Study group
Validation group
300
200
100
0
0
30
G-EXJ-1030713
May 2012
100
200
Biopsy LIC (µmol Fe/g dry wt)
Gandon Y, et al. Lancet. 2004;363:357-62.
300
400
SIR methods (cont.)
1. Patient
preparation
(5 min)
2. Image
acquisition
(approx. 5-20 min)
3. Data analysis
(relatively fast)
● The ROI is selected in the liver
and the reference tissue
(muscle or fat), in each image
● The SI of the liver region is
divided by that of the
reference tissue
● A calculation algorithm to
assist has been developed
for 0.5, 1.0, and 1.5 T
MRI machines1
31
G-EXJ-1030713
May 2012
1Gandon
Y. Available from: http://www.radio.univ-rennes1.fr/Sources/EN/HemoResult.html. Accessed December 2010.
Relaxometry methods: T2, T2*, T2′, R2, and R2*
● If a spin-echo sequence is used, the relaxation time is T2
● If a gradient-echo sequence is used, it is T2*
● These are related by the equation1
1/T2* = 1/T2 + 1/T2′
● T2′ is the magnetic field inhomogeneity of the tissue
● To attain a positive linear relationship with HIC
– T2* can be transformed into reciprocal
R2*: R2* [Hz] = 1,000/T2* [ms]
– T2 can be transformed into reciprocal
R2: R2 [Hz] = 1,000/T2 [ms]
32
G-EXJ-1030713
May 2012
1Anderson
LJ, et al. Eur Heart J. 2001;22:2171-9. 2Wood JC, Ghugre N. Hemoglobin. 2008;32:85-96.
Relaxometry methods: R2 and R2*
● Several pulse sequences are included in the MRI software package
R2 (for FerriScan®)
spin echo sequence
T2* (and R2*) gradient
echo sequence
FOV (mm)
300 x 225
350 x 300
Matrix (lines)
256 x 176
128 x 80
1.17 x 1.28 x 5.0
2.73 x 3.75 x 10.0
TR (ms)
2500
200
TE (ms)
6, 9, 12, 15, 18
Minimum possible (ideally < 2.0 ms)
NEX (n)
1
1
Flip angle (°)
90
20
BW (Hz/px)
300
1,950
–
8
On
On
Parameters
Resolution (mm)
Segments (n)
FatSat
33
G-EXJ-1030713
May 2012
Wood JC, Ghugre N. Hemoglobin. 2008;32:85-96.
Correlation between R2-estimated LIC
and LIC by biopsy
R2-LIC calibration curve
by Wood et al. 20051
R2-LIC calibration curve
by St Pierre et al. 20052
300
300
250
250
200
Mean R2 (Hz)
R2 (Hz)
350
200
150
100
LIC by biopsy, R = 0.98
Linear fit using biopsy data
Controls, LIC by norms alone
50
150
-thalassaemia/Hb E
-thalassaemia
Hepatitis
Hereditary haemochromatosis
100
50
0
0
0
10
20
30
40
Biopsy LIC (mg Fe/g dry wt)
34
G-EXJ-1030713
May 2012
1Wood
50
60
0
10
20
30
Biopsy LIC (mg Fe/g dry wt)
JC, et al. Blood. 2005;106:1460-5. 2St Pierre TG, et al. Blood. 2005;105:855-61.
40
Correlation between R2*-estimated LIC and LIC
by biopsy
R2*-LIC calibration curve
by Wood et al.1
R2*-LIC calibration curve
by Hankins et al.2
Patients
Controls
Fit
2,000
1,800
1,600
30
25
LIC (mg Fe/g dry wt)
R2* (Hz)
1,400
1,200
1,000
800
600
400
20
15
10
5
R = 0.97
200
0
Correlation coefficient = 0.98
p < 0.001
0
0
10
20
30
40
50
60
0
200
Biopsy LIC (mg Fe/g dry wt)
35
G-EXJ-1030713
May 2012
1Wood
JC, et al. Blood. 2005;106:1460-5. 2Hankins JS, et al. Blood. 2009;113:4853-5.
400
600
R2*MRI (Hz)
800
1000
LIC estimated with R2 and R2* MRI correlate
well with each other
Estimated HIC (mg/dry) by R2-SP
50
40
30
20
Patient data
Linear fit, R=0.94
10
0
0
10
20
30
Estimated HIC (mg/dry) by R2*
36
G-EXJ-1030713
May 2012
Wood JC, et al. Blood. 2005;106:1460-5.
40
50
Gradient relaxometry (T2*, R2*) can
conveniently measure cardiac and liver iron
Liver MRI
HIC (mg Fe/g of dry weight liver)
Cardiac MRI
[Fe] (mg/g dry wt)
14
12
10
8
6
4
R2 = 0.82540
2
0
0
100
200
300
400
30
Hankins, et al.
25
20
Wood, et al.
15
10
Anderson, et al.
5
0
0
Cardiac R2* (Hz)
200
400
Liver R2* (Hz)
Cardiac and liver iron can be assessed together conveniently
by gradient echo during a single MRI measurement.
37
G-EXJ-1030713
May 2012
HIC = hepatic iron concentration
Carpenter JP, et al. J Cardiovasc Magn Reson. 2009;11 Suppl 1:P224.
Hankins et al Blood. 2009;113:4853-4855.
600
800
1000
Relaxometry methods: pros and cons
Pros
Cons
R2*
● Correlate well to biopsy LIC1–4
● Greater sensitivity to iron deposits5
● Faster (images can be obtained in a single
breath-hold) and easier6
● Can perform cardiac and liver iron assessment
at the same time
● More susceptible to artefacts
● Requires expert training of a
technician/ radiologist for data
acquisition and data analysis
● Multiple breath-holds required which
increases MRI time
R2
(Ferriscan®)
● Correlate well to biopsy LIC1–4
● Less affected by susceptibility artefacts6
● Highly sensitive and specific over a large range
of LIC, including patients with severe
haemosiderosis7
● The gold standard method in clinical trials
● Requires no training for data analysis (done
centralized by Resonance Health)
38
G-EXJ-1030713
May 2012
1Christoforidis
● Cost of analysis (300 USD per scan)
A, et al. Eur J Haematol. 2009;82:388-92. 2St Pierre TG, et al. Blood. 2005;105:855-61. 3Wood JC,
et al. Blood. 2005;106:1460-5. 4Hankins JS, et al. Blood. 2009;113:4853-5. 5Anderson LJ, et al. Eur Heart J.
2001;22:2171-9. 6Wood JC, Ghugre N. Hemoglobin. 2008;32:85-96. 7Papakonstantinou, O, et al. J Magn Reson
Imaging. 2009;29:853-9.
Relaxometry methods: R2 and R2* (cont.)
1. Patient
preparation
(5 min)
2. Image
acquisition
(approx. 5-20 min)
3. Data analysis
(depends on
experience)
● Correct position is important so that the LIC across the whole liver
can be measured
● Images are taken at various TEs
Red line indicates correct
position of the slice
39
G-EXJ-1030713
May 2012
Liver R2* MRI
Liver with normal iron levels
TE=1.3ms
TE=3.6ms
TE=7.1ms
T2* = 15.7 ms or R2* = 63.7 Hz or LIC = 1.3mg/g
Liver with severe iron overload
TE=1.3ms
TE=3.6ms
TE=7.1ms
T2* = 1.1 ms or R2* = 909 Hz or LIC = 25.0 mg/g
40
G-EXJ-1030713
May 2012
Images courtesy of Dr J. de Lara Fernandes.
FAQ: artefacts
How frequent are artefacts in liver MRI?
In contrast to cardiac MRI, the risk for motion artefacts (e.g.
due to breathing) or susceptibility artefacts is much lower
when performing liver MRI. As in cardiac MRI, if artefacts are
present and too severe, scans may have to be repeated
How can I avoid artefacts when assessing LIC by MRI?
When assessing LIC, one thing that is really important is to
use fat saturation (usually automatically included in all the
sequences). This is especially important if a patient has
steatosis (e.g. adults with haemochromatosis)
41
G-EXJ-1030713
May 2012
Questions and answers were prepared under the review of Dr J. de Lara Fernandes, University of Campinas, Brazil.
Relaxometry methods: R2 and R2* (cont.)
1. Patient
preparation
(5 min)
2. Image
acquisition
(approx. 5-20 min)
3. Data analysis
(depends on
experience)
● Determine ROI
– entire liver boundary, excluding obvious
hilar vessels1
● Slice thickness
– varies, generally 5–15 mm1–4
● Number of slices
Red outline shows position of ROI
– anything from about 1 to 20 slices
can be studied1–4
42
G-EXJ-1030713
May 2012
1Wood
JC, et al. Blood. 2005;106:1460-5. 2St Pierre TG, et al. Blood. 2005;105:855-61.
O, et al. J Magn Reson Imaging. 2009;29:853-9. 4Hankins JS, et al. Blood. 2009;113:4853-5.
3Papakonstantinou
Relaxometry methods: R2 and R2* (cont.)
1. Patient
preparation
As TE increases, SI should decrease
●
When plotted on a graph
–
as iron load increases, the curve gets steeper
–
T2 or T2* can be calculated from the curve
–
R2 and R2* can also be calculated
Calculations are done
–
manually, or
–
by specific licensed software
(e.g. CMRtools®), or
–
images could be directly sent to a validated
centre performing FerriScan® for analysis
43
G-EXJ-1030713
May 2012
(depends on
experience)
(approx. 5-20 min)
●
●
3. Data analysis
100
Typical non-iron-loaded tissue
80
SI
(5 min)
2. Image
acquisition
60
40
20
0
0
5
10
TE (ms)
15
20
Analysis of the data
● The data can be analysed manually or using
post-processing software
44
G-EXJ-1030713
May 2012
Manually
Post-processing software
•Excel spreadsheet
•ThalassaemiaTools (CMRtools)
•cmr42
•FerriScan
•MRmap
•MATLAB
Analysis of the data (cont.)
Method
Pros
Cons
Excel spreadsheet
Low cost
Time-consuming
Tedious
ThalassaemiaTools
(CMRtools)1
Fast (1 min)2
Easy to use
FDA approved
GBP 3,000 per year
cmr42(3)
Easy to use
FDA approved3
Can generate T2*/R2* and T2/R2 maps with
same software
Allows different forms of analysis
Generates pixel-wise fitting with colour maps
40,000 USD first year costs
12,000 USD per year after
45
G-EXJ-1030713
May 2012
FDA = Food and Drug Administration.
1www.cmrtools.com/cmrweb/ThalassaemiaToolsIntroduction.htm. Accessed Dec 2010.
2Pennell DJ. JACC Cardiovasc Imaging. 2008;1:579-81.
3www.circlecvi.com. Accessed Dec 2010.
Analysis of the data (cont.)
Method
Pros
Cons
FerriScan1
Centralized analysis of locally acquired data (206
active sites across 25 countries)
Easy set-up on most MRI machines
EU approved
Validated on GE, Philips, and Siemens scanners
USD 300 per scan
Patients data are sent to reference
centre
MRmap2
Uses IDL runtime, which is a commercial software
(less expensive than cmr42/CMRtools)
Can quantify T1 and T2 map with the same
software
Purely a research tool
Not intended for diagnostic or clinical
use
MATLAB3
Low cost
Available only locally
Physicists or engineers need to write
a MATLAB program for display and
T2* measurement
46
G-EXJ-1030713
May 2012
1www.resonancehealth.com/resonance/ferriscan.
Accessed Dec 2010.
Accessed Dec 2010.
3Wood JC, Noetzli L. Ann N Y Acad Sci. 2010;1202:173-9.
2www.cmr-berlin.org/forschung/mrmapengl/index.html.
FAQ: mistakes in manual analysis
of liver MRI data
What is truncation?
After the selection of the ROI, the signal decay can
be fitted using different models. In the truncation
model, the late points in the curve (the plateau) are
subjectively discarded to obtain a curve with an R2
> 0.995. A new single exponential curve is made by
fitting the remaining signals.
What is the most frequent mistake made
when interpreting the data from an MRI scan?
47
G-EXJ-1030713
May 2012
Interpreting a liver MRI is more challenging than for a
cardiac MRI, especially in patients with severe liver iron
overload. Correcting the data using truncation analysis is
very important (done automatically by some software).
The example (see following slide) clearly shows what
happens, if the truncation is not done correctly
Questions and answers were prepared under the review of Dr J. de Lara Fernandes, University of Campinas, Brazil.
FAQ: mistakes in manual analysis
of liver MRI data (cont.)
Analysis without
truncation of the data
Non-truncated analysis with results with a poor R2
(< 0.995). The apparent LIC of 4.65 suggests mild LICs.
Observe the flat plateau of the data points after a
TE of 3.62 ms
48
G-EXJ-1030713
May 2012
Analysis with
truncation of the data
The same patient, but analysing the data with only the 3 first
data points results in a better (although not perfect) R 2.
The LIC results in severe iron overload, reflecting the real
concentrations of iron
FAQ: how to start measuring liver iron loading?
How to start measuring liver iron loading in a
hospital? What steps need to be taken?
To start assessing liver iron loading by MRI, these steps can be followed
1. Check MRI machine requirements
• 0.5–1.5 T (1.5 T is highly recommended for T2* and T2
calculations; 0.5 T only for SIR)
• calibrated
• includes a liver package
2. Optional: buy software for analysing the data (otherwise, Excel
spreadsheet can be used)
3. Optional: training of personnel for acquiring MRI images
4. Optional: training of personnel on how to analyse the data
49
G-EXJ-1030713
May 2012
Questions and answers were prepared under the review of Dr J. de Lara Fernandes, University of Campinas, Brazil.
LIC: interpretation of results
● LIC threshold values for classification of iron overload
Iron levels
LIC (mg Fe per
g dry weight)
LIC (µmol Fe
per g dry wt)
R2 (s−1)†
R2* (s−1)
T2* (ms)
Normal
<2
< 35.6
< 50
< 88
> 11.4
Mild
overload
≥ 2−7
≥ 35.6 − 125.0
≥ 50 – 100
≥ 88 – 263
> 3.8 – 11.4
Moderate
overload
≥ 7−15
≥ 125 − 269
≥ 100 – 155
≥ 263 – 555
> 1.8 – 3.8
Severe
overload
≥ 15
≥ 269
≥ 155
≥ 555
≤ 1.8
50
G-EXJ-1030713
May 2012
†Values
estimated based on R2 LIC calibration curve; R2, R2* and T2* values valid for MRI machines
with 1.5T only.
St Pierre TG, et al. Blood 2005;105:855–861; Wood JC, et al. Blood 2005;106:1460–1465.
Implementation of liver and cardiac MRI
1.5T MRI Scanner
US$1.000.000
Yes
½ day training
Liver
Analysis
Experienced radiologist
No
1 day training
Post-processing analysis
Cardiac acquisition package
US$50.000
Yes
US$40.000 or US$4.000/y
or in-house or outsource
1-2 day training
Routine cardiac MR exams
No
51
G-EXJ-1030713
May 2012
4 day training
Slide presented at Global Iron Summit 2011 - With the permission of Juliano de Lara Fernandes
Heart
Analysis
Summary
G-EXJ-1030713
May 2012
Summary
● Iron overload is a serious problem among patients who require blood
transfusions to treat anaemia and associated conditions
● Analysing liver iron overload is important
– to predict risk of hepatic and extra-hepatic complications
● The extent of iron accumulation in the liver is a key prognostic indicator
for morbidity and mortality
● MRI has the added advantage that iron levels throughout the liver can
be analysed, rather than just the biopsied section (iron levels throughout
the liver can vary)
– R2 is the most commonly used technique in clinical practice,
although R2* is a comparable alternative across most ranges
of iron overload and is faster
53
G-EXJ-1030713
May 2012
GLOSSARY
OF TERMS
G-EXJ-1030713
May 2012
GLOSSARY
● AML = acute myeloid leukemia
● APFR = Atrialp peak filling rate
● BA = basilar artery
● ß-TM = Beta Thalassemia Major
● ß-TI = Beta Thalassemia Intermedia
● BM = bone marrow
● BTM = bone marrow transplantation
● BW = bandwidth
● CFU = colony-forming unit
● CMML = chronic myelomonocytic leukemia
● CT2 = cardiac T2*.
● DAPI = 4',6-diamidino-2-phenylindole
55
G-EXJ-1030713
May 2012
GLOSSARY
● DFS = = disease-free survival.
● DysE = dyserythropoiesis
● ECG = electrocardiography
● EDV = end-diastolic velocity
● EF = ejection fraction
● EPFR = early peak filling rate
● FatSat = fat saturation
● FAQ = frequently asked questions
● FDA = Food and Drug Administration
● FISH = fluorescence in situ hybridization.
● FOV = field of view
● GBP = Currency, pound sterling (£)
56
G-EXJ-1030713
May 2012
GLOSSARY
● Hb = hemoglobin
● HbE = hemoglobin E
● HbF = fetal hemoglobin
● HbS = sickle cell hemoglobin.
● HbSS = sickle cell anemia.
● HIC = hepatic iron concentration
● HU = hydroxyurea
● ICA = internal carotid artery.
● ICT = iron chelation therapy
● IDL = interface description language
● IPSS = International Prognostic Scoring System
● iso = isochromosome
57
G-EXJ-1030713
May 2012
GLOSSARY
● LIC = liver iron concentration
● LVEF = left-ventricular ejection fraction
● MCA = middle cerebral artery
● MDS = Myelodysplastic syndromes
● MDS-U = myelodysplastic syndrome, unclassified
● MRA = magnetic resonance angiography
● MRI = magnetic resonance imaging
● MV = mean velocity.
● N = neutropenia
● NEX = number of excitations
● NIH = National Institute of Health
● OS = overall survival
58
G-EXJ-1030713
May 2012
GLOSSARY
● pB = peripheral blood
● PI = pulsatility index
● PSV = peak systolic Velocity
● RA =refractory anemia
● RAEB = refractory anemia with excess blasts
● RAEB -T = refractory anemia with excess blasts in transformation
● RARS = refractory anemia with ringed sideroblasts
● RBC = red blood cells
● RF = radio-frequency
● RCMD = refractory cytopenia with multilineage dysplasia
● RCMD-RS = refractory cytopenia with multilineage dysplasia with
ringed sideroblasts
● RCUD = refractory cytopenia with unilineage dysplasia
59
G-EXJ-1030713
May 2012
GLOSSARY
● RN = refractory neutropenia
● ROI = region of interest
● RT = refractory thrombocytopenia
● SCD = sickle cell disease
● SD = standard deviation
● SI = signal intensity
● SIR = signal intensity ratio
● SF = serum ferritin
● SNP-a = single-nucleotide polymorphism
● SQUID = superconducting quantum interface device.
● STOP = = Stroke Prevention Trial in Sickle Cell Anemia
● STOP II = Optimizing Primary Stroke Prevention in Sickle Cell Anemia
60
G-EXJ-1030713
May 2012
GLOSSARY
● T = thrombocytopenia
● TAMMV = time-averaged mean of the maximum velocity.
● TCCS = transcranial colour-coded sonography
● TCD = transcranial doppler ultrasonography
● TCDI = duplex (imaging TCD)
● TE = echo time
● TR = repetition time
● WHO = World Health Organization
● WPSS = WHO classification-based Prognostic Scoring System
61
G-EXJ-1030713
May 2012