In Vivo Comparison of Myocardial T1 With T2 and T2* in Thalassaemia Major

Transcription

1 CME JOURNAL OF MAGNETIC RESONANCE IMAGING 38: (2013) Original Research In Vivo Comparison of Myocardial T1 With T2 and T2* in Thalassaemia Major Yanqiu Feng, PhD, 1,2 Taigang He, PhD, 2,3 * John-Paul Carpenter, MD, 2,3 Andrew Jabbour, MD, 2,3 Mohammed Harith Alam, MB, 2,3 Peter D. Gatehouse, PhD, 2,3 Andreas Greiser, PhD, 4 Daniel Messroghli, MD, 5 David N. Firmin, PhD, 2,3 and Dudley J. Pennell, MD, FRCP 2,3 Purpose: To compare myocardial T1 against T2 and T2* in patients with thalassemia major (TM) for myocardial iron characterization. Materials and Methods: A total of 106 TM patients ( years; 58 males) were studied on a 1.5 Tesla scanner using dedicated T1, T2*, and T2 relaxometry sequences. A single mid-ventricular short axis slice was acquired within a breath-hold. Results: In patients with myocardial iron overload (T2* < 20 ms; n ¼ 52), there were linear correlations between T2 and T2* (r ¼ 0.82; P ¼ 0.0), and between T1 and T2* (r ¼ 0.83; P ¼ 0.0). In patients with no myocardial iron (n ¼ 54), T2* values were scattered with no significant correlation against T2 or T1. For all patients (n ¼ 106) there was a strong linear correlation (r ¼ 0.93; P ¼ 0.0) between myocardial T1 and T2. Conclusion: In patients with iron overload, myocardial T2 and T1 are correlated with T2*. In patients with low or normal myocardial iron concentration, other factors become dominant in affecting T2* values as shown by scattered T2* data. Myocardial T1 correlates linearly with T2 measurements in all patients suggesting that these two relaxation parameters avoid extrinsic magnetic field 1 School of Biomedical Engineering, Southern Medical University, Guangzhou, China. 2 Cardiovascular Magnetic Resonance Unit, Royal Brompton Hospital, London, United Kingdom. 3 National Heart and Lung Institute, Imperial College, London, United Kingdom. 4 Healthcare Sector, Siemens AG, Erlangen, Germany. 5 Department of Congenital Heart Disease and Pediatric Cardiology, Deutsches Herzzentrum Berlin, Berlin, Germany. Contract grant sponsor: National Basic Research Program of China (973 Program); Contract grant number: 2010CB732502; Contract grant sponsor: National Natural Science Funds of China; Contract grant number: ; Contract grant sponsor: British Heart Foundation (BHF) Intermediate Basic Science Fellowship; Contract grant number: FS/08/26225; Contract grant sponsor: the UK NIHR Cardiovascular Biomedical Research Unit of Royal Brompton & Harefield NHS Foundation Trust and Imperial College, London. *Address reprint requests to: T.H., Cardiovascular Magnetic Resonance Unit, Royal Brompton Hospital, Sydney Street, London SW3 6NP, UK. t.he@imperial.ac.uk Received July 12, 2012; Accepted November 30, DOI /jmri View this article online at wileyonlinelibrary.com. inhomogeneity effects and may potentially provide improved myocardial tissue characterization. Key Words: MRI relaxometry; T2*; T2; T1; tissue characterization; myocardium J. Magn. Reson. Imaging 2013;38: VC 2013 Wiley Periodicals, Inc. BETA-THALASSEMIA MAJOR (TM) is an inherited hemoglobin disorder which results in a regular blood transfusion requirement that leads to tissue iron overload. Myocardial iron quantification is essential in preventing cardiomyopathy, managing the iron chelation therapy, and monitoring the response to treatment during follow-up (1 3). The clinical technique used to assess myocardial iron measures myocardial T2* (4 6), and this has been calibrated against ex vivo myocardial levels (7), predicts heart failure, and is robust between centers across the world (8,9). More recently, a robust T2 relaxometry MR sequence was developed and may also be used for evaluation of myocardial iron overload (10,11). Both T2* and T2 of hydrogen nuclei change inversely with the amount of hemosiderin iron, an effect termed susceptibilityinduced relaxation (12). Tissue iron is dominant in determining myocardial T2* and T2 relaxation, and there is a linear correlation between these measures (13). There is however currently little data on in vivo human myocardial iron assessment using the longitudinal relaxation time T1. The technical challenge of myocardial T1 measurement, due in particular to motion and heart rate variability, has limited exploration of its clinical application. Early studies have suggested that T2 is more sensitive than T1 to the presence of liver iron, both in vivo (14) and in vitro (15). More recently, Wood et al calibrated T2*, T2, and T1 against cardiac iron using a gerbil iron overload model (16), and showed that T1 changes with tissue iron, but with smaller effect relative to T2 and T2* changes. Recently developments with the modified Look-Locker Inversion recovery (MOLLI) sequence for myocardial T1 quantification now show good in vivo accuracy and reproducibility (17,18), and it is now practical to VC 2013 Wiley Periodicals, Inc. 588

2 MRI T2*, T2 and T1 in the Human Heart 589 Table 1 Major Parameters in the MRI Protocol Sequence T2* T2 T1 MOLLI Readout GRE TSE bssfp Magnetic preparation DIR DIR Non-Sel IR Bandwidth (Hz/pixel) RF mode / Gradient Fast Fast Fast mode Flip angle FoV (mm) FoV phase (%), typical TR/TE (ms) 20/ / /1.5 Trigger ECG ECG ECG Trigger pulses Data acquisition window (ms) Number of images Slice thickness (mm) Base matrix Resolution (mm x mm) 3.1 x x x 1.6 Resulting scan time (s) GRE ¼ gradient echo; TSE ¼ turbo spin echo; bssfp ¼ balanced steady-state free precession; DIR ¼ double inversion recovery; IR ¼ inversion recovery; RF ¼ radiofrequency; FoV ¼ field of view; TR ¼ repetition time; TE ¼ echo time. investigate T1 relaxation time changes to myocardial iron deposition in TM patients. In this study, therefore, we compared in vivo myocardial T2*, T2, and T1 measurements in a substantial population of thalassemia patients to establish the relations between T1, T2, and T2* in patients with and without cardiac iron loading. MATERIALS AND METHODS Patient Population Regularly transfused patients with TM were scanned (n ¼ 106; 58 males; age years). All patients had received iron chelation therapy since the mid-tolate 1970s, or from early childhood in patients born after this time. The study protocol was approved by the local ethics committee. All patients gave written informed consent ms) were acquired. Both sequences used double inversion recovery pulses to suppress the blood signal and data was acquired every other cardiac cycle, which made the effective repetition time (TR) two cardiac cycles. 2. T1 MOLLI, as previously described (17), consisted of three inversion recovery (IR) prepared ECGsynchronized Look-Locker trains performed consecutively within a breathhold. Each of the three trains starts with an inversion pulse that uses a specific inversion time (100, 180, and 260 ms), after which multiple single-shot images are acquired in consecutive heartbeats. The three trains result in a set of 11 source images with different effective inversion times. A map of T1 in the imaging slice can then be generated (17). All images were acquired with a trigger delay timed to end-diastole. Typical parameters for the sequences that were used are summarized in Table 1. Analysis For all MR relaxometry analysis, a homogeneous full thickness region of interest (ROI), was chosen in the mid-ventricular septum (6,10,13). Care was taken to exclude epicardial structures and blood pool from the contours (Fig. 1). For T2* and T2 measurements, the mean signal intensity of ROI was measured in the series of images with increasing TE, and the data were plotted against the TEs to form an exponential decay curve. To derive T2*, the data were fitted with an equation of: SI ¼ P 0 e TE=T2 where P 0 represents a constant of magnetization, TE represents the echo time and SI represents the image signal intensity. T2 measurements were made in a similar way. All T2* and T2 analysis were carried out on a personal computer using Thalassemia-Tools, a plug-in of CMRtools (Cardiovascular Imaging Solutions, London, UK). The truncation algorithm was implemented ½1Š MR Protocol This study was carried out on a 1.5 Tesla (T) MR scanner (Siemens Magnetom Sonata, Erlangen, Germany) equipped with high performance gradients having a maximum strength of 40 mt/m and maximum slew rate of 200 T/m/s on each axis independently. A four-element cardiac phased array coil was used. All patients were scanned using the previously described black-blood T2* sequence (6), the black-blood T2 sequence (10), and the T1 MOLLI sequence (17), each within a breathhold. A single mid-ventricular short axis slice was imaged at the same position for T2*, T2, and T1, respectively, and the following optimized protocols were used: 1. For T2*, eight TE images ranging from 1.5 ms to 16.9 ms (increment 2.18 ms), and for T2, 12 TE images ranging from 4.8 ms to ms (increment Figure 1. A typical short axis mid-ventricular image obtained from a thalassemia patient. The ROI for MR relaxometry measurement was restricted to the septum. [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.]

3 590 Feng et al. left ventricular septum in the T1 maps to produce T1 measurements by use of in-house software developed with Matlab (Version , Mathworks). Statistics Histograms were drawn to demonstrate distributions of T2*, T2, and T1 measurements in the patient population. Summary data were expressed with 95%confidence intervals with range indicated. Correlation analysis was performed using Pearson s test. A P value of <0.05 was considered statistically significant. Figure 2. Typical short-axis mid-ventricular T2* (top), T2 (middle), and T1 images (bottom) obtained from two TM patients, one with iron overload (left, T2* ¼ 11.6 ms) and one normal (right, T2* ¼ 30.9 ms). Due to the technical limits, the resolutions of T2*, T2, and T1 were different, with the resolution of the T2 images being least good. The timing of data acquisition also varies. for both T2* and T2 measurement (19). T1 maps from MOLLI image sets were automatically calculated off-line on the personal computer by using an opensource software tool (20); ROIs were then drawn in the RESULTS Typical T2*, T2, T1 Images Representative images acquired from two TM patients of the same midventricular slice for T2*, T2, and T1 sequences are shown in Figure 2. When considering all the patient images, three issues were noted: (A) the T2 images were of lower resolution and therefore appeared a little blurred; (B) Motion artifacts were more pronounced in the T1 images because the data acquisition window was relatively longer; (C) The contrast-to-noise-ratio of the T1 image was low because of the inversion recovery. However, image quality was adequate for all T2*, T2, and T1 measurements. Distribution of Myocardial T2*, T2, and T1 Measurements The distribution of myocardial relaxometry measurements of all patients are shown in Figure 3. T2* Figure 3. The histogram charts for all patients show the distribution of myocardial T2* (left), T2 (middle), and T1 (right) measurements. Cumulative percentage, indicating the cumulative frequency distribution, is shown as an additional line to the chart with scale given on the right-hand side of each subplot. 48.1%of the patients had cardiac iron overload (T2* < 20 ms). [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.]

4 MRI T2*, T2 and T1 in the Human Heart 591 Figure 4. Correlation curve between T2 and T2* drawn from 106 TM patients. The vertical broken line (red) represents the previously established cut-off myocardial T2* value to distinguish significant myocardial iron loading. Data of patients with T2* < 20 ms were linearly correlated (solid green line, r ¼ 0.82). This direct line was extended to cover the whole data range (dashed green line). The whole data were fitted by a quadratic curve (red). ranged from 4.2 to 41.7 ms (mean, ms) corresponding to myocardial tissue iron loading of 8 mg/g to 0.5 mg/g dry weight (7). Approximately half (n ¼ 52; 48.1%) of the patients had myocardial iron loading (T2* < 20 ms). T2 ranged from 25.9 to 65.5 ms (mean, ms). Both T2* and T2 ranges were in accordance with previous reports (13). T1 ranged from 474 to 1033 ms (mean, ms). In patients with abnormal T2* (<20 ms), the T1 value ranged from 474 to 804 ms (n ¼ 52; mean, ms), which is clearly shortened compared with that of normal volunteers (17). Relationship Between T2 and T2* Figure 4 shows the relation between T2 and T2* from all patients in this study. There is a linear relationship (straight line, green, n ¼ 52, r ¼ 0.82, P ¼ 0.0) between T2 and T2* for patients with myocardial iron overload. In the patients with normal myocardial iron, by contrast, no such strong linear relation is present. The whole data was also fitted with a quadratic curve (red, r ¼ 0.94). Note that within the abnormal T2* range, this nonlinear curve agrees with the linear line (solid green). Figure 5. Correlation between T1 and T2* drawn from 106 TM patients. The vertical broken line (red) represents the previously established cut-off myocardial T2* value to distinguish significant myocardial iron loading. Data from patients with T2* < 20 ms were linearly correlated (green, r ¼ 0.83, P ¼ 0.0). This direct line was extended to cover the whole data range (dashed green line). The whole data were fitted by a quadratic curve (red). Relationship Between T1 and T2 Figure 6 shows the relation between T1 and T2 from all patients in this study. There is a linear relationship (n ¼ 106; r ¼ 0.93; P ¼ 0.0) between T1 and T2 for patients with or without myocardial iron overload. DISCUSSION To our knowledge, this is the first report relating T1 to T2 and T2* in the in vivo human heart of TM patients. In a previous work, it has been shown that T2* is important for identifying patients at risk and monitoring treatment in myocardial siderosis (3,4), and that T2 can also be used to assess myocardial iron (10,11), Relationship Between T1 and T2* Figure 5 shows the relation between T1 and T2* from all patients in this study. There is a linear relationship (n ¼ 52; r ¼ 0.83; P ¼ 0.0) between T1 and T2* for patients with myocardial iron overload, whilst no such correlation is present in patients with normal myocardial iron. The whole data was likely fitted with a quadratic curve (red, r ¼ 0.93; P ¼ 0.0). Note that within the abnormal T2* range, this nonlinear curve agrees with the linear line (solid green). Figure 6. Correlation between T1 and T2 drawn from 106 TM patients. All data were fitted by linear regression (green, r ¼ 0.93, P ¼ 0.0).

5 592 Feng et al. although there is little clinical validation of its use. In the current study, it has been further demonstrated that there is a linear correlation between T1 and T2 in the in vivo human heart, indicating that T1 can also be used to assess myocardial iron. In consistent with our previous finding (13), T2 correlates linearly with T2* in the presence of heart iron in the current study. T1 correlates linearly with T2* in the iron overloaded heart is further evidence that iron dominates the relaxation process in this scenario. The exact manner in which the iron interferes with the myocardium is complex. Iron deposition dominates the relaxation process in the myocardium which directly affects both transverse and longitudinal relaxation. Intracellular iron particles can be considered as microscopic magnets which interfere with the local magnetic field and desynchronize neighboring protons resulting in enhanced T2* and T2 relaxation. Paramagnetic interactions also give rise to the transfer of energy from the spin system to the lattice, hence affecting longitudinal relaxation. Although early studies suggest that T1 is less sensitive to iron compared with T2 or T2*, the current study demonstrates T1 relaxation responds to the iron similarly to T2. In TM patients with abnormal T2*, the maximum T1 value was approximately 800 ms in this study. To note, however, there is an off-resonance effect produced by the iron; in this scenario, not all of the magnetization is inverted hence the measured T1 could be shortened as T1 MOLLI uses bssfp (17); in addition, the ironshortened T2 might also contributes to the T1 measurement as bssfp images in general exhibit T2/T1 contrast (21). In the presence of iron, therefore, the underlying mechanism of T1 shortening can be complex and further studies are warranted to confirm a preliminary finding and to determine the cut-off T1 value for detecting myocardial tissue iron overload in thalassemia. Two possible uses for T1 relaxometry, combined with T2 or T2* relaxometry, might be: (A) for attempting to identify whether differences in chemical speciation of iron have differential transverse and longitudinal relaxation effects, and whether these have clinical utility; and (B) in diseases with mild iron overload where T2* should be used with caution as other factors affecting field homogeneity are important, which may confound the contribution of T2* change caused by the disease itself. In this scenario, T2 or T1 measurement might be a better option as they are less affected by local field inhomogeneity. An example of measurement of a disease associated with low levels of iron and change with therapy is in Friedreich s ataxia (22); and (C) due to technical limits, very short TE images cannot be obtained using conventional techniques so that very severe iron overload may not be accurately by T2*. T1 may offer help in this regard, but improved sequence and optimized protocols are needed. Despite considerable enthusiasm, the use of quantitative cardiovascular MR relaxometry has not fully lived up to expectation because of the wide scatter of these parameters for in vivo applications. This is in particular a challenge to myocardial tissue characterization due to the complexity of the structure, flow, and issues with respiratory and cardiac motion. We closely observed the patient for breathing during the scans. If there was breathing or motion artifacts, we simply repeated the breath-hold scan, and gave additional training to the patient on breathhold to improve the image quality. These simple measures helped improve the mis-registration. In the postprocessing, assiduous care was taken to exclude blood pool in drawing the ROIs. In the current study, each of the sequences had been previously validated with protocols optimized (6,11,18). Nevertheless, we were not able to make the imaging parameters identical due to inherent limitations of each sequence. One such issue is the resolution differences between the sequences used in this study and comparison of these relaxometry measurements at the same resolution is of interest. It is possible to improve the current T2* and T1 techniques to enable a resolution of 1.6 mm by 1.6 mm by use of parallel and partial Fourier techniques (23). To achieve this goal for T2, however, substantial technical improvement may be needed. For T1 MOLLI, measures should be taken to address motion artifacts for a more reliable quantification, and the sampling range can be further optimized to accommodate the shortened T1 range in TM patients. There are potential limitations to this study. The long acquisition window and the heart rate variability could potentially affect the measurement. Consequently, the correlation coefficient r between T1, T2, and T2* is possibly affected, but this is likely a minor effect and should not affect the overall conclusions in our study. Comparison was based on measurements in the interventricular septum, as it has been shown that myocardial T2* can only be measured reliably here due to noted susceptibility artifacts found in other regions. By restricting the measurements to the septum, the intrinsic relationships between T2*, T2, and T1, which are not confounded by other factors affecting T2* could be properly explored. Mid-ventricular septum T2* is known to be highly representative of global iron concentration (7,24), and there is little significant segmental variation of iron deposition across the heart (7); therefore, this limitation should not affect the conclusions. Although the exploration of the relations between T2*, T2, and T1 outside the mid-ventricular septum is of both scientific and clinical interest, we believe this is beyond the scope of the current approach and a further comprehensive relaxometry study of the whole heart is warranted. Another limitation of this study is that we did not have software to do all the analysis in the same framework so had to use MRMap, CMRtools and MATLAB for the analysis and comparison. Effort should be made to develop more reliable software to do all analysis in the same framework. In conclusion, this study shows that iron overload is a dominant factor in determining both transverse and longitudinal relaxation of the human heart, suggesting that MRI T2*, T2, and T1 can all provide noninvasive means for myocardial tissue iron quantification in transfusion-dependent patients. For patients with other diseases with low iron burden, such as disorders affecting iron handling with only mild

6 MRI T2*, T2 and T1 in the Human Heart 593 increases in tissue iron, T2 and T1 relaxometry may offer improved tissue characterization by avoiding extrinsic field inhomogeneity issues which may confound T2* relaxometry. It is also possible that further investigation into relations between T2*, T2, and T1 relaxometry may offer novel and complementary information on iron loading in the heart, perhaps through differential effects of chemical species of iron. ACKNOWLEDGMENTS Drs. Feng and He contribute equally to this study. Dr. Carpenter has received honoraria from Novartis, Apotex, and Swedish Orphan. Professor Pennell is a consultant to Novartis, Apotex, and Siemens, and a director of Cardiovascular Imaging Solutions. Dr. He has received honoraria from Novartis and Apotex, and is supported by the British Heart Foundation. The other authors declare no conflicts. Dr. Feng was funded by the National Basic Research Program of China and National Natural Science Funds of China. Dr. He was supported by Wellcome Trust Value In People (VIP) award and now holds a British Heart Foundation (BHF) Intermediate Basic Science Fellowship. REFERENCES 1. Borgna-Pignatti C, Rugolotto S, De Stefano P, et al. Survival and disease complications in thalassemia major. Ann NY Acad Sci 1998;850: Modell B, Khan M, Darlison M. Survival in beta thalassaemia major in the UK: data from the UK Thalassaemia Register. Lancet 2000;355: Modell B, Khan M, Darlison M, Westwood MA, Ingram D, Pennell DJ. Improved survival of thalassaemia major in the UK and relation to T2* cardiovascular magnetic resonance. J Cardiovasc Magn Reson 2008;10: Anderson LJ, Holden S, Davies B, et al. Cardiovascular T2* (T2 star) magnetic resonance for the early diagnosis of myocardial iron overload. Eur Heart J 2001;22: Westwood M, Anderson LJ, Firmin DN, et al. A single breath-hold multiecho T2* cardiovascular magnetic resonance technique for diagnosis of myocardial iron overload. J Magn Reson Imaging 2003;18: He T, Gatehouse PD, Kirk P, et al. Black-blood T2* technique for myocardial iron measurement in thalassemia. J Magn Reson Imaging 2007;25: Carpenter JP, He T, Kirk P, et al. On T2* magnetic resonance and cardiac iron. Circulation 2011;123: Tanner MA, He T, Westwood MA, Firmin DN, Pennell DJ. Thalassemia International Federation Heart T2* Investigators. Multi-center validation of the transferability of the magnetic resonance T2* technique for the quantification of tissue iron. Haematologica 2006;91: Kirk P, He T, Anderson LJ, et al. International reproducibility of single breathhold T2* MR for cardiac and liver iron assessment among five thalassemia centers. J Magn Reson Imaging 2010;32: He T, Gatehouse PD, Anderson LJ, et al. Development of a novel optimized breathhold technique for myocardial T2 measurement in thalassemia. J Magn Reson Imaging 2006;24: He T, Kirk P, Firmin DN, Lam WM, et al. Multi-center transferability of a breath-hold T2 technique for myocardial iron assessment. J Cardiovasc Magn Reson 2008;10: Gillis P, Roch A, Brooks RA. Corrected equations for susceptibility induced T2-shortening. J Magn Reson 1999;137: He T, Smith GC, Gatehouse PD, Mohiaddin RH, Firmin DN, Pennell DJ. On using T2 to assess extrinsic magnetic field inhomogeneity effects on T2* measurements in myocardial siderosis in thalassemia. Magn Reson Med 2009;61: Engelhardt R, Langkowsky JH, Fischer R, et al. Liver iron quantification: studies in aqueous iron solutions, iron overloaded rats, and patients with hereditary hemochromatosis. Magn Reson Imaging 1994;12: Stark DD, Moseley ME, Bacon BR, et al. Magnetic resonance imaging and spectroscopy of hepatic iron overload. Radiology 1985;154: Wood JC, Otto-Duessel M, Aguilar M, et al. Cardiac iron determines cardiac T2*, T2, and T1 in the gerbil model of iron cardiomyopathy. Circulation 2005;112: Messroghli DR, Radjenovic A, Kozerke S, Higgins DM, Sivananthan MU, Ridgway JP. Modified Look-Locker inversion recovery (MOLLI) for high-resolution T1 mapping of the heart. Magn Reson Med 2004;52: Messroghli DR, Plein S, Higgins DM, et al. Human myocardium: single-breath-hold MR T1 mapping with high spatial resolutionreproducibility study. Radiology 2006;238: He T, Zhang J, Carpenter JP, et al. Automated truncation method for myocardial T2* measurement in thalassemia. J Magn Reson Imaging [Epub ahead of print]. 20. Messroghli DR, Rudolph A, Abdel-Aty H, et al. An open-source software tool for the generation of relaxation time maps in magnetic resonance imaging. BMC Med Imaging 2010;10: Scheffler K. and S. Lehnhardt, Principles and applications of balanced SSFP techniques. Eur Radiol 2003;13: Abbruzzese G, Cossu G, Balocco M, Marchese R, Murgia D, Melis M, Galanello R, Barella S, Matta G, Ruffinengo U, Bonuccelli U, Forni GL. A pilot trial of deferiprone for neurodegeneration with brain iron accumulation. Haematologica 2011; 96: Messroghli DR, Greiser A, Frohlich M, Dietz R, Schulz-Menger. Optimization and validation of a fully-integrated pulse sequence for modified look-locker inversion-recovery (MOLLI) T1 mapping of the heart. J Magn Reson Imaging 2007;26: Pepe A, Positano V, Santarelli MF, et al. Multislice multiecho T2* cardiovascular magnetic resonance for detection of the heterogeneous distribution of myocardial iron overload. J Magn Reson Imaging 2006;23: