International Journal of Cardiac Imaging 14: I 17- 122, 1998.
`© 1998 Kluwer Academic Publishers. Printed in the Netherlands.
`
`117
`
`Myocardial iron deposition in {3-Thalassemia studied by magnetic
`resonance imaging
`
`Sophie I. Mavrogeni, Thomas Maris 1, Athanasios Gouliamos 1, Lambros Vlahos 1 &
`Dimitrios T. Kremastinos
`Onassis Cardiac Surgery Center; 1 Areteion Hospital, Athens, Greece
`
`Received I 0 February 1997; accepted 5 December 1997
`
`Key words: ,B-thalassemia, heart, iron overload, MRI
`
`Abstract
`
`Myocardial iron deposition is a common finding in ,B-thalassemia. The iron content of the myocardium was assessed
`using the T2 relaxation time of the heart. The T2 relaxation time of the liver and skeletal muscle was also assessed
`in order to study the relation of iron deposition between heart, liver and skeletal muscle. ECG gated spin echo
`images were obtained from thirty-eight consecutive adult thalassemic patients examined in an outpatient clinic,
`aged (x±SD) 25±6 years, usin~ a 0.5 T system. Patients were divided into groups A and B, according to their
`average serum ferritin levels ofllie...preceding five years(> or< 2000 ng/ml). Results were compared with nine
`controls, aged 24±7 years. Heart T2 relaxation time in the control group (x±SD)(48.3±5.5 msec) was higher
`compared to group A (28.4±6.7 msec, p <0.00 1) but not to group B (43.4±7.4 msec). The T2 relaxation time
`of the heart correlated positively with the T2 relaxation time of the liver (r=0.68, p < 0.001) and negatively with
`fenitin levels (r=-0.67, p<0.001). There was no correlation with the T2 relaxation time of skeletal muscle. This
`study indicates that regularly transfused !'.3-thalassemia patients may present with a broad variation of heart iron
`deposition which, however, is related to serum ferritin levels.
`
`Introduction
`
`The thalassemias are a diverse group of congenital
`disorders in whieh there is a defect in the synthesis
`of one or more of the subunits of hemoglobin. In ,B(cid:173)
`thalassemia, the !'.3-chains of hemoglobin have a normal
`structure but are produced in reduced or, occasional(cid:173)
`ly, undetectable amounts. The gene frequency for ,B(cid:173)
`thalassemia approaches 0.1 in Southern Meditenanean
`areas and both sexes are equally affected. Diagnosis is
`based on detection of an abnormal hemoglobin pat(cid:173)
`tern by electrophoresis. Patients present all signs and
`symptoms of severe anemia, including findings related
`to intramedullary and peripheral hemolysis, and iron
`overload. Adequate blood transfusions and continuous
`chelation are the milestones of treatment. However,
`the combination of chronic hypoxia and myocardial
`siderosis leads to cardiac arrythmias, congestive heart
`failure and, finally, death. Currently, bone marrow
`transplantation is perfonned in eligible cases, and gene
`
`therapy offers a promise for the future. Iron deposition
`in the heart and other organs is the causative factor of
`the main complications in ,B-Thalassemia Major [1-4,
`27]. Excessive iron is primarily retained in the retic(cid:173)
`uloendothelial system and, when the capacity of this
`system is exceeded, secondary deposition in parenchy(cid:173)
`mal organs will follow [5].
`The estimation of iron stores in each individual
`organ and the total iron load in !'.3-thalassemic patients
`would be helpful in order to evaluate the efficacy
`of chelation therapy. However, this is difficult to be
`assessed since hemosiderin and ferritin, the iron stor(cid:173)
`age proteins, are mostly intracellular. On the oth(cid:173)
`er hand, serum ferritin is highly correlated with the
`amount of iron deposited in the tissues and is consid(cid:173)
`ered to be a very sensitive index [6]. Ferritin levels,
`however, may be affected by factors such as fever or
`inflammation [7, 8]. Presently, the most accurate way
`to estimate total body iron deposition is the direct mea(cid:173)
`surement of iron content in liver biopsy specimens [9],
`
`•
`
`EXHIBIT
`I t{X
`tOSS-
`1tJfhf1&- it~
`
`1 of 6
`
`Taro Pharmaceuticals, Ltd.
`Exhibit 1055
`
`

`

`·.
`
`118
`
`Table I. Study group stratification (values as x±SD or x(range)).
`
`Table 2. Individual patient Ferritin levels (ng/ml).
`
`Group A
`
`Group B
`
`Group C
`
`Group
`
`No
`
`Fer. aver
`
`Fer. min
`
`Fer. max
`
`17
`9
`21
`Subjects (n)
`Fcrritin (ng/ml) 4357 (2500-9000) 1550 (I 060-2000)* -
`Blood units
`Transfused (n) 1025 ± 274
`Chelation (yrs) 13.3 ± 4.4
`
`965 ± 429
`12.7 ± 5.0
`
`+ p< 0.01.
`Group A: high fe rritin, Group 8: low ferritin, Group C: controls.
`
`an invasive procedure which can not be applied for rou-
`tine follow-up and does not give any infonnation about
`iron deposition in the heart. The ability of stored intra-
`cellular tissue iron to enhance magnetic susceptibility
`is the basis by which it can be detected by magnetic
`resonance imaging (MRI). Recent studies in experi-
`mental animals have shown that the T2 relaxation time
`has a linear correlation with the total iron contents for
`all organs, including the heart [10].
`The aim of this study was to assess heart iron con-
`tent using the T2 relaxation time of the heart, and to
`compare it to the T2 relaxation time ofliver and skelc-
`tal muscle in a group of regularly transfused patients
`with ,8-thalassemia major, and also to correlate these
`parameters with the average serum ferritin levels of th e
`last five years.
`
`Patients and methods
`
`Patient population (Tables 1, 2)
`
`Thirty-eight consecutive patients with ,B-thalassemia
`major, 16 males and 22 females, aged (x±SD) 25±6
`years were studied. Thirty-one of them presented a
`normal heart function and seven were diagnosed with
`heart failure. All patients were transfused monthly to
`maintain haemog lobin levels at I 0-13 g/dL. Transfu-
`sion therapy had started before the age of S years and
`iron chelation therapy with deferoxamine (IM, IV or
`SC administration) had started before the age of 15
`years. A scrum ferritin level below 2000 ng/mL was
`considered the target value for a patient successfully
`treated with deferoxamine [28]. The average serum
`ferritin level of each patient was derived from 30 val-
`ues obtained bi-monthly over the preceding 5 years.
`Thirty of the patients were splenectomized. Patients
`were stratified into two groups, according to their
`average ferritin values (> or < 2000 ng/ml): Group
`
`Group A
`(high Fer.)
`[No: 1-21]
`
`Group B
`(low Fer.)
`[No: 22-38]
`
`•I
`2
`3
`•4
`5
`6
`7
`8
`9
`10
`I I
`•12
`13
`14
`· 15
`16
`17
`18
`19
`·20
`21
`
`22
`23
`24
`25
`26
`- 27
`28
`29
`30
`31
`32
`33
`34
`35
`36
`37
`·38
`
`9000
`7000
`7000
`6500
`5500
`5000
`4500
`4500
`4000
`4000
`4000
`3600
`3500
`3500
`3000
`3000
`3000
`2800
`2800
`2800
`2500
`
`2000
`2000
`2000
`2000
`1800
`1700
`1500
`1500
`1500
`1500
`1500
`1400
`1200
`1200
`1200
`11 00
`1060
`
`5000
`3000
`6000
`5000
`4000
`4500
`4000
`2000
`3300
`2000
`3600
`2600
`1250
`2500
`3200
`800
`3000
`680
`1800
`2300
`2000
`
`1500
`1100
`750
`1800
`480
`1100
`1100
`940
`1062
`270
`1000
`650
`680
`600
`600
`1000
`950
`
`9500
`!0000
`7500
`10000
`7000
`10000
`8000
`5000
`5600
`12000
`6700
`7500
`6000
`5500
`7000
`3300
`8500
`7000
`7500
`6000
`4000
`
`6000
`5000
`5000
`5000
`2200
`2100
`2700
`3400
`4000
`5000
`4000
`2500
`3500
`2200
`2500
`2500
`8700
`
`aver: average, n1in: minimum, max: maxi1nu1n.
`• patients with symptomatic heart failure.
`
`A (n=2 l), with ferri tin levels of (mean (range)) 4357
`(2500-9000) ngliill and group B (n= 17), w ith ferritin
`levels of 1550 ( 1060- 2000) ng/ml. T he total number
`of transfusions and years of chelation therapy were
`similar in all groups. The patients were compared with
`9 normal controls (Group C), aged 24±7 years. An
`informed consent was obtained from all subjects and
`
`2 of 6
`
`Taro Pharmaceuticals, Ltd.
`Exhibit 1055
`
`

`

`the study was approved by the local Ethics Committee
`of the hospital.
`
`MR techniques
`
`Imaging technique
`All MR studies were performed using a 0.5 T supercon(cid:173)
`ducting imaging system (MR-Max-Plus, GE/CGR).
`Body quadrature coil was used for b oth excitation and
`signal detection. The single oblique orientation of the
`imaging slices (head-left to feet-right) was determined
`from a scout coronal T 1 localiser image in order to
`depict a short axis view of the heart. These slices were
`chosen since the relative images that they represented
`depicted the heart, liver and skeletal muscle (latissimus
`dorsi) in one picture. The slice with the best left ven(cid:173)
`tricular delineation was selected. The MR study was
`ECG-gated with TR=heart rate, slice thickness lOmm
`and TE= 12-120 msec in ten symmetrically repeatable
`echoes. These parameters were chosen on the basis
`of phantom studies, where a mean accuracy of 5.5%
`and a mean reproducibility of 4.5% was demonstrat(cid:173)
`ed for T2 values ranging from 10 up to 80 ms. For
`both sequences, the field of v iew was 42 cm and the
`image reconstruction matrix was 160 x 224. The longer
`anatomical axis was chosen as the frequency encoding
`axis, and the shorter as the phase encoding axis. Res(cid:173)
`p irat01y phase encoding and presaturation pulses were
`used to compensate for respiratory motion and blood
`flow artifacts respect ively. Rectilinear pixels were used
`in order to improve the signal to noise ratio.
`
`Quantitative image analysis
`T2 relaxation time measurements for the left ven(cid:173)
`tricle of the heart (T2LV), the liver (T2Ll) and the
`skeletal muscle (T2MU) were perfo1med from the
`T2 calcul ated imaging maps. These imaging maps
`were calculated using 10 successive T2-weighted
`images, obtained using the afore-mentioned single(cid:173)
`slice multi-echo imaging technique. The T2-weighted
`MRI images were transferred from the MRl machine
`to a PC-workstation using appropriately selected hard(cid:173)
`ware (calibrated HI-RES video frame grabber). Images
`were then archived as convent ional PC-based imaging
`files (256 grey level, T IFF format, LZW compression).
`Assuming single exponential behaviour of all tis(cid:173)
`sues presented on conventional spin-echo images,
`a three point fit (in-house created fitted algorithm)
`was used. T2 calculated imaging maps were finally
`obtained on a pixel-by-pixel basis. The method applied
`
`11 9
`
`Table 3. T2 relaxation time in the study groups (values as x±SD).
`
`Groups
`subjects
`
`Group A
`(n=21)
`
`Group B
`(n=17)
`
`Group C
`(n=9)
`
`T2 heart (msec)
`T2 liver (msec)
`T2 muscle (msec)
`
`28.4±6.7
`22.3±6.7
`28.4±2.7
`
`43.4±7.4a
`27.1 ±6.6b
`29.2±2.4
`
`48.3±5.5a
`34.9±5.oc
`30.0±1 .9
`
`a p< 0.00 1, compared to group A
`b p<0.05, compared to group A
`c p<O.O I, compared to group A and B
`Group A: high ferritin, Group B: low ferriti n, Group C: controls.
`
`for the needs of this study is the imaging representa(cid:173)
`tion of the quantitative T2 analysis (T2-QMRl) tech(cid:173)
`nique, applied partially by certain of the authors else(cid:173)
`where [11]. T2 values were calculated from the imag(cid:173)
`ing maps, using five appropriately selected regions of
`interest (ROls), with each one situated on the chosen
`organ (heart, liver, muscle) for every patient. Observers
`were unaware of the serum ferritin measurements. The
`intra-observer and inter-observer coefficient of varia(cid:173)
`tion for T2 measurements from the imaging maps was
`7% and 13% respectively.
`
`Statistical analysis
`
`The resu lts are expressed as x±SD (or x (range)) and
`compared by means of an unpaired two-tailed Studen(cid:173)
`t's t-test. The chi-square test was used for comparisons
`of percentages between groups. Correlations between
`various parameters were sought with Pearson's correla(cid:173)
`tion coefficient. Statistical signifi cance was considered
`for p < 0.05.
`
`Results
`
`The T2 relaxation time of the left ventricle (T2LY)
`in the high fe1Titin group (group A) was 28.4±6.7
`msec, in the low ferritin group (group B) it was
`43.4±7.4 msec and in the control group (group C)
`it was 48.3±3.5 msec. T2LV in group C was high(cid:173)
`er compared to group A (p<0.00 1) but not compared
`to group 8. There also was a significant difference
`between groups A and B (p<0.00 1) (Table 3 and Fig(cid:173)
`ures l , 2).
`There was a significant difference in the T2 relax(cid:173)
`ation time of the liver (T2Ll) between groups A and
`B (22.3±6.7 vs. 27. 1 ±6.6 msec, p <0.05). T2 relax(cid:173)
`ation time in both populations was lower compared
`
`3 of 6
`
`Taro Pharmaceuticals, Ltd.
`Exhibit 1055
`
`

`

`120
`
`Figure I. Short axis view of the heart of a thalassemic patient with
`mild iron overload. Iron deposition can be visualized as scanty black
`patches over the LV surface (left ventricle between arrowheads).
`
`Figure 2. Short axis view of the heart of a thalassemic patient with
`severe iron overload and impaired LY function (dilated cardiomy(cid:173)
`opathy). The enlargement of the LV is accompanied by wide-spread
`black areas of iron deposition (left ventricle between arrowheads).
`
`. .
`
`..
`..
`..
`
`"
`
`10
`
`lOOO
`
`2000
`
`3000
`
`eooo
`5000
`4000
`F•n'ltlni (AQ/otdl
`
`1000
`
`8000
`
`9000
`
`10000
`
`Figure 3. Inverse relationship between the T2 relaxation time of
`the heart and the average ferritin levels of the thalassemic patients
`(r=-0.67, p< 0.001 ).
`
`to controls in group C (34.3±5 msec, p< 0.01). The
`T2 relaxation time of the skeletal muscle (T2MU) was
`similar in the three groups.
`Five patients in group A (23.8%) exhibited heart
`failure, compared with only 2 in group B (11.8%,
`p=NS). There was no difference between patients
`with or without heart failure for T2LY (30.0± 8.5 vs.
`36.2± 10.4msec, p:NS), T2Ll (21.4 ±5.9 vs. 25. l ±6.5
`msec, p:NS), T2MU (31 .5±6.0 vs. 35.2±5.9 msec,
`p:NS) and ferritin levels (x (range)) (3951 ( 1060-9000)
`vs. 2909 (1100-7000) ng/ml, p:NS).
`A positive correlation was found between heart and
`liver T2 relaxation times of the thalassemic population
`in our study (r=0.68, p< 0.00 I) .
`The average ferritin levels of the preceding five
`years were inversely correlated with both the T2 relax(cid:173)
`ation time of the heart (r=-0.67, p< 0.001) (Figure 3)
`and the liver (r=-0.58, p< 0.001). No correlation was
`fow1d between the T2 relaxation time of the skeletal
`muscle and ferritin levels.
`
`Discussion
`
`The presence of iron affects tissue Tl and T2 relax(cid:173)
`ation times. The effect on TI relaxation time, however,
`is less significant and most studies use the T2 relax(cid:173)
`ation time to asS(:ss iron deposition [ 12]. This effect
`is proportional to the tissue iron content and depends
`[ 13]. We employed
`on the applied magnetic field (B 0 )
`this principle to evaluate a group of adult thalasscmic
`patients.
`A 0.5 T imaging system enabled the implementa(cid:173)
`tion of a ten echo pulse sequence, in which an ear-
`
`4 of 6
`
`Taro Pharmaceuticals, Ltd.
`Exhibit 1055
`
`

`

`..
`
`ly TE=l2ms and several repeated TEs were applied.
`Errors were eliminated by measuring five regions of
`interest for each site. The signal to noise ratio was
`increased by a shorter TE, as suggested by other authors
`[ 12].
`In our patient population, the heart and liver T2
`relaxation times were sign ificantly reduced, w ith high(cid:173)
`er iron concentrations resulting in lower T2 values.
`There was a significant difference in the heart T2 relax(cid:173)
`ation time between the high and low ferritin groups,
`but not between the controls and the low ferritin group.
`This indicated that the myocardium is usually unaffect(cid:173)
`ed in the early stages of iron deposition. Heart failure,
`a fatal complication of ,6-thalassemia, was present in
`both the high and low ferritin groups, possibly due to
`other factors than iron involvement in the pathogenesis
`of heart failure [ 14].
`The liver T2 relaxation time was severely affected
`in both the high and low fetTitin groups, compared to
`controls. This finding is suggestive of early and severe
`liver iron overload, even in cases with low total iron
`deposition.
`Serum ferritin is currently the most accurate index
`of body iron content. Variations correspond to changes
`in the reticuloendothelial system storage iron but not to
`parenchymal iron content [ 15]. Liver disease, inflam(cid:173)
`mation, infection and assay problems may influence
`measurements [7, 8, 16]. The average ferritin of the
`preceding five years was used in order to overcome
`these limitations. These levels inversely correlated with
`both heart and liver T2 relaxation times. The positive
`correlation between heart and liver T2 relaxation times
`suggests that both organs have a similar iron deposition
`pattern, although the liver ~eems to be more affected.
`The absence of changes in 'tlfo skeletal muscle T2 relax(cid:173)
`ation time indicates that this muscle is unaffected by
`iron overload.
`There is a controversy about the precise relationship
`between the myocardial iron content and the degree of
`heart dysfunction. Zaino et al. [17], using nuclear res(cid:173)
`onance scattering (NRS), found symptomatic cardiac
`disease in the patients with the highest cardiac levels
`of iron. Correlations between the heart T2 relaxation
`time and cardiac biopsy are absent since right ventricu(cid:173)
`lar biopsy is subject to serious sampling errors and iron
`deposition is patchy and not uniform [ 18]. A good cor(cid:173)
`relation was found between iron content and T2 relax(cid:173)
`ation time only when whole rat hearts were sampled
`[IO]. Recently, studies of thalassemic patients using
`0.5 T indicated that cardiac complications are related
`to a low heart T2 relaxation time [ 19]. Additionally,
`
`121
`
`the heart/muscle signal intensity ratio was sensitive to
`iron in patients treated by multiple blood transfusions
`[20] and impaired heart function was evident in the
`more heavily transfused patients, although there was
`no correlation with serum ferritin levels. MRI is use(cid:173)
`ful in addressing this issue. In our study, heart failure
`in the low ferritin group suggests that iron deposition
`may not be the only factor for the development of heart
`failure [14].
`Hepatic iron load, as studied by MRI, has been con(cid:173)
`sidered as an alternative to liver biopsy to assess the
`total iron load [2 1]. Quantitative image analysis (T2-
`QMRI) is a reliable technique, well correlated with
`liver biopsy [ 11 ]. In vitro measurements of the T2
`relaxation time of liver samples from iron-overloaded
`rats [22] and spleen samples from thalasscmic patients
`[ 12], demonstrated a linear correlation between relax(cid:173)
`ation rate ( l/T2) and iron content. The reticuloendothe(cid:173)
`lial system preferentially accumulates iron from the
`breakdown of transfused erythrocytes, before deposi(cid:173)
`tion occurs in parenchymal organs, e.g. heart. Our data
`are consistent w ith Buja et al. [23] who indicated that
`cardiac iron deposition is accompanied by heavy iron
`deposits in the liver.
`Some authors consider the application of the T2
`relaxation time in assessing heart iron deposition to
`be imprecise since it may be unmeasureable in sever(cid:173)
`ly iron-overloaded patients, due to low signal inten(cid:173)
`sity equal to background noise [20]. The use of a
`0.5 T machine (less prominent magnetic susceptibil(cid:173)
`ity phenomenon) [ 13], the use of shorter echo times
`(TE) [ 12] and the application of MR spectroscopy [21]
`appear to be more appropriate for the study of severe
`iron overload. Other non-invasive techniques such as
`dual energy computed tomography [24, 25], supercon(cid:173)
`ducted quantum interference device (SQUID) [26] and
`nuclear resonance scattering (NRS) [ I 7] can overcome
`this limitation.
`
`In conclusion, this study shows that T2-QMR1 is a non(cid:173)
`invasive tissue characterization method enabling the
`simultaneous examination ofheai1 function, and heart
`and liver iron content. It is easily repeated, facilitating
`thalassemic patient follow-up and chelation therapy
`efficacy evaluation.
`...
`
`Acknowledgment
`
`We are grateful to Ms. Sonja Phillips, M.Sc., B.Sc., for
`carefully proofing the manuscript.
`
`5 of 6
`
`Taro Pharmaceuticals, Ltd.
`Exhibit 1055
`
`

`

`122
`
`References
`
`5.
`
`I . Kattamis C, Touliatos N, Haidas S, Matsaniotis N. Growth
`of children with thalasscmia: effect of different transfusion
`regimens. Arch Dis Childhood 1970; 45: 502- 7.
`2. Propper RD, Cooper B, Rufo RR, et al. Continuous subcuta(cid:173)
`neous administration of desferoxamine in patients with iron
`overload. N Engl J Med 1977; 297: 418- 23.
`3. Aldouri MA, Wonke B, Hoffbrand AV, et al. Iron state and
`hepatic disease in patients with thalassemia major treated with
`Iong-tem1 subcutaneous desferrioxamine. J Clin Pathol 1987;
`40: 1353- 9.
`4. Engle MA. Cardiac involvement in Cooley's anemia. Ann NY
`Acad Sci 1969; 119: 694-702.
`Jacobs A. The pathology of iron overload. In: Jacobs A, Wor(cid:173)
`wood M (eds). Iron in biochemistry and medicine, II. New
`York: Academic Press, 1980: 439- 52.
`6. Jacobs A, Miller F, Worwood M, Beamisch R, Wardrop C.
`Ferritin in the serum of normal subjects wilh iron deficiency
`and iron overload. Br Med J 1972: 206-11.
`7. Crosby WH. Serum ferri tin fails to indicate hemochromatosis:
`nothing gold can stay. N Engl J Med 1976; 294: 333- 34.
`8. Green R, Watson LR, Saab GA, Crosby WH. 'Nomrnl ' serum
`ferritin: a caution. Blood I 977; 50: 545-47 .
`.
`9. Ridson RA, Barry M, Flynn DN. Transfusional iron overload:
`the relationship between tissue iron concentration and hepatic
`fibrosis in lhalassemia. J Pathol 1975; I 16: 83-95.
`10. Liu P, Henkelman M, Joshi J, Waien S, Butany J, Olivieri
`Q uantification of myocardial tissue iron contents using NMR
`relaxation. Validation in a novel murine thalassemia model.
`JACC 1992; 19(3): I 87 A (abstract 935-76).
`11. Papakonstantinou 0, Maris T, Kostaridou V, Gouliamos A,
`Koutoulas G, Kalovidouris A, et al. Assessment of liver iron
`overload by T2-quantitative magnetic resonance imaging. Cor(cid:173)
`relation of T2-QMRI measurements with serum ferritin con(cid:173)
`centration and histologic grading of siderosis. Magn Res I mag-
`ing 1995; 13: 967- 77.
`,
`12. Gomori JM, Grossman Rl , Drott HR. MR relaxation times and
`iron content of thalassemic spleens! an in vitro study. AJR
`I
`1988; 150: 567-69.
`13. Vymazal J, Brooks RA, Zak 0, McRill C, Shen C. Di Chiro
`G. Tl and T2 of ferritin at different field slrengths: Effect on
`MRI. Magn Res Med 1992; 2~ 368- 74.
`14. Kremastinos DT, Tiniakos G. Theodorakis GN. Katritsis DG,
`Toutouzas PK. Myocarditis in ,B-thalassemia major. A cause
`for heart fail ure. Circulation 1995; 91: 66-71.
`15. Gomori JM, Horev G, Tamary H, et al. Hepatic iron overload:
`Quantitative MR imaging. Radiology 1991 ; 179: 367- 69.
`16. Worwood M, Cragg SJ, Jacobs A, McLaren C, Ricketts C.
`Economidou J. Binding of serum ferritin to concanavallin A:
`Patients with homozygous beta-thalassemia and transfusional
`iron overload. Br J Haematol 1980; 46: 409- 16.
`
`20.
`
`17. Zaino EC, Wielopolski L. In vivo measurement of hepatic and
`cardiac iron in Cooleys anemia by nuclear resonance scattering.
`Ann NY Acad Sci 1990; 612: 573- 76.
`18. Olson LJ, Edwards WD, McCall JT, Ilstrup DM, Gersh BJ.
`Cardiac iron deposition in idiopathic hemochromatosis: histo(cid:173)
`logic and analytic assessment of 14 hearts from autopsy. JACC
`1987; I 0: 1239-43.
`19. Liu P, Olivieri N, Sullivan H, Henkelman M. Magnetic reso(cid:173)
`nance imaging of cardiac iron overload in thalassemia: Detec(cid:173)
`tion of iron content and association with cardiac complications.
`JACC 1993; 21 (2): 409A (abstract 948- 119).
`Johnston DL, Rice L, Vick W, Hedrick TD, Rockey R. Assess(cid:173)
`ment of tissue iron overload by nuclear magnetic resonance
`imaging. Am J Med 1989; 87: 40-47.
`21. Dixon RM, Styles P, Al-Refaie F, et al. Assessment of hepatic
`iron overload in thalasscmic patients by magnetic resonance
`spectroscopy. Hcpatology 1994; 19: 904-10.
`22. Stark DD, Moseley ME, Bacon BR, et al. Magnetic resonance
`imaging and spectroscopy of hepatic iron overload. Radiology
`1985; 154: 137-42.
`23. Buja LM, Roberts WC. Iron in the heart: etiology and clinical
`signifi cance Am J Med 1971 ; 51: 209- 2 1.
`24. Houang MT, Skalicka A, Arozena X, Huehns ER, Shaw DG.
`Correlation between computed tomographic values and liver
`iron content in thalassemia major with iron overload. Lancet
`1979; I: 1322- 33.
`25. Olivieri NF, Grisaru D, Dancman A, Martin DJ, Rose V, Freed(cid:173)
`man MH. Computed tomography scanning of the liver to deter(cid:173)
`mine efficacy of iron chelation therapy in thalassemia major. J
`Pediatrics 1989; 114: 427- 30.
`26. Brinenham G, Griffith PM, Nienhuis AW, McLaren CE, Young
`NS, Tucker EE. et al. Efficacy of deferoxamine in preventing
`complications of iron overload in patients with thalassemia
`major. N Engl J Med 1994; 33 1: 567-73.
`lsselbacher, Braunwald, Martin, Fauci and Kaspar, editors.
`Harrison's Principles of Internal Medicine. New York: Mc
`Graw Hill, 1994: 1741-43.
`28. Olivieri NF, Nathan DG, MacMillan JH, Wayne AS, Liu PP,
`McGee A, et al. Survival in medically treated patients with
`homozygous ,B-thalassemia. N Engl J Med 1994; 331: 574-
`78.
`
`27.
`
`Address for correspondence: S. Mavrogeni, 50 Esperou Street,
`175-61 P.Faliro, Athens, Greece. Tel/Fax: + 30- 1-98.82.797, E-mail:
`som@aias.net.gr
`
`6 of 6
`
`Taro Pharmaceuticals, Ltd.
`Exhibit 1055
`
`