Applications Guide. Quantitative Cardiac Parameter Mapping (T1 T2 T2*)

Transcription

1 Applications Guide Quantitative Cardiac Parameter Mapping (T1 T2 T2*) Work-in-Progress # 780 (VD13A SP4) N4_VD13A_CV_TXMAP_Z001KZZF_WIP780 MAGNETOM Aera MAGNETOM Skyra MAGNETOM Avanto + Dot MAGNETOM Verio + Dot syngo MR Numaris 4 VD13A SP4 July _732C&780-ASD-S _732C&780-ASD-S01-01 Page 1/37

2 Important Note This document provides a description of techniques developed by Siemens. Siemens has tested the software provided with this work-in-progress package in combination with the proposed clinical application. However, each user should be aware of the fact that incorrect use of this software may produce unknown results. The sequences contained in this software package do not exceed the FDA safety performance parameter guidelines for MRI exams. Specifically, there is no change to patient risk as compared to routine operation of the MAGNETOM with regard to: static magnetic field; the time rate of change of the gradient magnetic fields; the rate at which RF power is deposited into the body (SAR); or the acoustic noise created by the MAGNETOM. The software has been tested internally but not yet in a clinical environment. For routine applications, its functionality may not be complete, and use of this software will remain investigational. In general, the clinical user will, in his/her sole responsibility, decide on the use of this application package or on subsequent therapeutic or diagnostic techniques and shall apply such techniques in his/her sole responsibility. Siemens will not take responsibility for the correct application of, or consequences arising from use of, this applications package. The software in this package may change in the future, or may not be available in future software versions. Siemens has the right to remove this software at any point. In case of any questions that are related to the use of this package please contact the Siemens representative in charge of this package (contact information is provided in the next page). Siemens representatives in charge of the package: _780-ASD-S02-01 Page 2/37

3 North America:... Bruce Spottiswoode, Ph.D. Siemens Healthcare MR Research and Development Chicago, IL Rest of the world:... Andreas Greiser, Ph.D. Siemens AG Healthcare Sector MR PI CARD Erlangen, Germany Additional developers of this works-in-progress package: Christopher Glielmi, Ph.D. Siemens Healthcare MR Research and Development New York, NY Shivraman Giri, Ph.D. Siemens Healthcare MR Research and Development Chicago, IL Randall Kroeker, Ph.D. Siemens Healthcare MR Research and Development Winnipeg, Canada _780-ASD-S02-01 Page 3/37

4 Table of Contents Improvements since previous versions of the WIP... 5 Introduction... 6 T1 mapping using MOLLI... 7 Sequence description... 7 User interface... 9 Example images Scanning tips T2 mapping using T2-prepared TrueFISP Sequence description User interface Example images Scanning tips T2* mapping using multi-echo GRE Sequence description Use interface Example images Scanning tips Installation Protocols User feedback References Acknowledgements Appendix 1: Hints for the use of the Cardiac Shim _780-ASD-S02-01 Page 4/37

5 Improvements since previous versions of the WIP The first version of the TX mapping WIP for VD11D, WIP732B, was ported directly from the original VB17A WIP448. The only practical difference between WIP448 and WIP732B relates to the TI times in T1 mapping. In WIP448 the TI time was defined as the duration between the beginning of the adiabatic inversion pulse and the center of k-space, whereas in WIP732B the TI time is more correctly defined from the end of the adiabatic inversion pulse to the center of k-space. More specifically, the reported TI times in WIP732B are s shorter than those in WIP448, and in WIP732C and WIP780 the reported TI times are 2560 s shorter than in WIP448. This version of the WIP, WIP780 (VD13A), incorporates a number of improvements compared to WIP 732B, including: T1 mapping: New adiabatic inversion pulse for improved inversion efficiency. The hypersec adiabatic inversion pulse in the previous WIP versions achieved an inversion factor of about A new shorter tan/tanh design with optimized parameters is now incorporated, resulting in an inversion factor of about Correction factor to accommodate for imperfect inversion. For the new tan/tanh adiabatic inversion pulse, the T1 estimate is divided by This correction is possible because tan/tanh pulse results in a reduced dependence on both T1 and T2. Phase sensitive inversion recovery (PSIR) fitting technique for more reliable T1 estimation and shorter processing times. New outputs: o T1* maps, which haven t experienced the Look-Locker correction, and can thus be used for more accurate measurement of T1 in blood, where fresh spins are flowing into the imaging plane _780-ASD-S02-01 Page 5/37

6 o Goodness of fit maps introduced as an optional quality assurance metric. o Optional synthetic PSIR and MagIR image series derived from the pixel-wise T1 maps. Protocol improvements offering a shorter scan time and minimal T1 measurement dependence on heart rate. T2 mapping: Improved adiabatic T2 preparation with only two refocusing pulses for shortening the achievable T2 prep time and for reduced RF power requirements. Dummy gradients played out during recovery heartbeats. Protocol improvements offering a centric gradient echo readout strategy for more accurate T2 measurements with minimal T1 dependence. T2* mapping: Improved slice profile to minimize errors due to dephasing. Introduction In recent years, quantitative MRI has become increasingly important for assessment of a range of cardiac diseases. Such techniques can reduce the subjectivity encountered in traditional non-quantitative techniques, and can detect global pathological changes within the heart. Quantification of T1 relaxation is of great importance for the characterization of myocardial tissue to assess both ischemic and non-ischemic cardiomyopathies [1]. Prior to the administration of contrast, an elevated value of myocardial T1 is associated with edema which may be related to the inflammatory response to myocardial injury. Following the administration of a T1-shortening contrast agent, _780-ASD-S02-01 Page 6/37

7 a shortened T1 corresponding to increased contrast agent concentration is associated with fibrotic scar or diffuse fibrosis which has a greater extracellular volume than normal [2,7,8]. Quantitative T2 mapping is utilized to assess pathologic conditions, such as acute ischemia, myocarditis and heart transplant rejection, which alter the myocardial water content and consequently prolong the T2 relaxation times [9-14]. T2 mapping has been shown to accurately and reliably detect regions of edematous myocardial tissue without the limitations of qualitative T2-weighted imaging [15]. Similarly, myocardial T2* measurement is a valuable tool for non-invasive assessment of iron overload, and is clinically employed for planning and monitoring iron-chelating treatments for transfused thalassemia major patients [17-20]. This package consolidates T1, T2 and T2* estimation techniques to offer a comprehensive cardiac parameter mapping with advanced reconstruction techniques such as motion correction and inline map generation. T1 mapping using MOLLI Sequence description In this package, T1 Mapping is performed using ECG triggered Modified Look- Locker Inversion Recovery (MOLLI) [3,4]. This technique allows the acquisition of single shot TrueFISP images acquired at different inversion times after a single inversion pulse, all gated to the same cardiac phase, thereby enabling a pixelbased T1 quantification in the myocardium. After the inversion, the magnetization _780-ASD-S02-01 Page 7/37

8 following the T1 relaxation curve is repetitively sampled over several heartbeats until the longitudinal magnetization has fully recovered. By combining several (typically 2-3) inversions with slightly shifted TI times within one protocol, the relaxation curve is sampled in an interleaved manner, resulting in a sufficient number of points for accurate T1 quantification acquired within a single breathhold. Figure 1 shows the original MOLLI implementation, comprising three inversion pulses with T1 sampling performed over 3, 3 then 5 heartbeats, and T1 recovery periods of 3 heartbeats. Figure 1: MOLLI pulse sequence scheme. There are three Look-Locker (LL) experiments, each prepared by a separate 180 degree inversion pulse (inv). The inversion time (TI) of the first LL experiment is defined as TIminimum. TI of the second and third LL experiment are determined by TIminimum + TIincrement and TIminimum + 2TIincrement, respectively. After the inversion pulses, images are read out in a non segmented fashion with a constant flip angle ( ). There are a defined number of pausing heart cycles in between LL experiments in order to allow for undisturbed signal recovery. Sourced from Ref. 2. For a group of N frames as inversion recovery images with different inversion times TI n, the signal for each pixel in position (x,y) is given by: _780-ASD-S02-01 Page 8/37

9 S(x,y,TI n ) = A(x,y) - B(x,y) * exp(-ti n /T1(x,y)) (1) T1(x,y) = T1*(x,y) * (B(x,y)/A(x,y) - 1) (2) Where A, B, and T1* are estimated by a three parameter fit on the measured data [5]. To generate the inline T1 map, the acquired inversion recovery images are first registered using a novel motion correction algorithm [5,6] which is based on estimating synthetic images presenting contrast changes similar to the acquired images by solving a variational energy minimization problem. Thereafter, the T1 estimate is computed on a per-pixel basis by performing a non-linear curve fitting using the three parameter signal model (Eq. 1,2). Note that this version of the WIP also outputs a T1* map, which can be used to provide a more realistic estimate of T1 in blood since the inflow of fresh spins obviates the need for the Look-Locker correction. Also, an improved phase sensitive fitting algorithm has been implemented for more reliable T1 estimation and shorter processing time. User interface Figure 2 shows the Sequence Special card for T1 mapping, which provides flexible control of the sequence timing and acquisition parameters. Additional processing steps and optional images can also be set using a number of checkboxes _780-ASD-S02-01 Page 9/37

10 Figure 2: User interface parameters for T1 mapping. Parameter map type This selection controls the type of parameter map (T1/T2/T2*). For MOLLI T1 mapping, this should be set to T1 Map. It can also be set to None, in which case the sequence behaves just like the normal product CV sequence. No. of inversions This parameter controls the number of inversions, which is equivalent to the number of passes of the T1 relaxation curve to sample the signal for various effective TI times. MOLLI TI start This parameter defines the initial TI time TI 1 as realized in the first heartbeat. According to the MOLLI acquisition scheme, the next heartbeat has an effective TI of TI 1 +RR 1. MOLLI TI increment _780-ASD-S02-01 Page 10/37

11 For the normal case where the number of inversions is greater than 1, this parameter defines the time shift relative to the preceding TI (TI increment in Figure 1). These shifts in TI ensure a more uniform sampling coverage along the T1 relaxation curve. Acq HB (array) This parameter array controls the number of heartbeats that are used for data acquisition after each inversion. Recov HB (array) This parameter array controls the number of heartbeats that are used for recovery of the magnetization before the next inversion is played out. During the recovery heartbeats, the sequence plays gradient noises so that the subject doesn t assume that the scan has finished and inadvertently starts breathing. Motion correction This checkbox activates registration of individual inversion images. If selected, non linear motion correction will be performed before curve-fitting and map generation. Both the original and registered images will be output as separate series. If the deformation exceeds a pre-defined threshold for a specific TI image, then the original (underformed) image for that specific TI will be used for the T1 curve fitting. Goodness of fit map Enabling this checkbox outputs a goodness of fit image defined as the sum of the square of the fitting residuals. This can be used for quality assurance. Note that this image will be inherently scaled by the magnitude of the MR image, so areas such as the lungs will appear dark. Synth MagIR and PSIR _780-ASD-S02-01 Page 11/37

12 This checkbox uses the known T1 for each pixel to create a series of synthetic images describing the magnitude inversion recovery (MagIR) and phase sensitive inversion recovery (PSIR). Each series comprises 40 images with TI increments of 25 ms, starting at 200 ms. Systolic imaging This checkbox should be enabled if imaging is performed during systole. It prevents possible SAR warnings in the case of a short trigger delay by including the time during diastole in the SAR calculations _780-ASD-S02-01 Page 12/37

13 Example images Figure 3 shows the improvement offered by the non rigid motion correction. Figures 4 and 5 provide example images from a normal volunteer, and Figures 6 and 7 show the utility of T1 maps in cases with pathology. Figure 3: Example of MOLLI motion correction. a c: Original images showing noticeable motion. d f: Results by directly applying nonrigid registration causing incorrect deformation. g i: Motion correction based on synthetic image estimation. Image from Ref _780-ASD-S02-01 Page 13/37

14 Figure 4: Example images output by MOLLI T1-mapping in a healthy subject. Top rows show the original images acquired at different TI times. The bottom rows show the motion corrected images using a novel non-rigid registration algorithm described in Ref. [5, 6]. Thereafter, a three parameter fit on the pixel-wise signal values is utilized to produce T1 estimates. (a) (b) (c) _780-ASD-S02-01 Page 14/37

15 (d) Figure 5: Additional images output by the T1 mapping sequence (a) T1 map, (b) T1* map (see Eq. 2), (c) goodness of fit map, and (d) select synthetic MagIR images. (a) (b) (c) Figure 6: Pre-contrast T1-maps (top row), post-contrast T1-maps (2nd row), late gadolinium enhancement (3rd row) for patients with: (a) chronic MI, (b) acute myocarditis, and (c) HCM. Adapted from Ref _780-ASD-S02-01 Page 15/37

16 (a) (b) Figure 7: (a) Conventional (FLASH segmented) LGE for a patient with extensive MI, showing a large apical thrombus. (b) The MOLLI T1 map acquired post contrast shows that the thrombus has long T1 as expected. Courtesy: Dr. Peter Kellman, NHLBI, Bethesda, MD, USA. Scanning tips The MOLLI T1 mapping is performed in a breathhold fashion. Please ensure that the trigger delay and acquisition window are modified so that the scan window is positioned during diastole. Also, we recommend reviewing the source images in addition to the parameter map to ensure that they are co-registered and free of artifacts. A number of protocols are described towards the end of this document, including the original MOLLI implementation as well as shorter sequence versions tailored for pre-contrast, post-contrast, slow and fast heart rates. For T1 mapping at 3T the use of the Cardiac Shim option is recommended. For further details, please check appendix 1. T2 mapping using T2-prepared TrueFISP _780-ASD-S02-01 Page 16/37

17 Sequence description This package provides a T2 Mapping technique [15] that uses a T2-prepared (T2p) TrueFISP sequence to produce single-shot T2-weighted images, each with different T2 preparation times. An optional fast non-rigid registration algorithm is utilized to compensate for in-plane motion between these images. Finally, a pixel-wise myocardial T2-map is generated using unsupervised curve-fitting based on the following two parameter equation. S(x,y) = M 0 (x,y) * exp( TE T2P /T2(x,y)) (3) The schematic in Figure 8 shows the sequence operation, which comprises varying T2-preparations followed by single-shot TrueFISP readout, and separated by several heartbeats for full T1 regrowth. Figure 8. Schematic of data acquisition and reconstruction for T2 Mapping. T2p- TrueFisp images are acquired at intervals of at least 3 RR intervals to allow for sufficient _780-ASD-S02-01 Page 17/37

18 magnetization recovery in between acquisitions. For each image, the acquisition window is in the same diastolic phase. An optional motion correction is applied to correct for misregistration between the different images. Finally, a pixel-wise T2 fit is done assuming mono-exponential signal decay User interface Figure 9 shows the Sequence Special card for T2 mapping, which provides flexible control of the number of T2 preps, the T2 prep duration (TE), and the recovery period. Figure 9: User Interface parameters for T2 Mapping Parameter map type This selection controls the type of parameter map (T1/T2/T2*). For T2 mapping, this should be set to T2 Map. It can also be set to None, in which case the sequence behaves just like the normal product CV sequence _780-ASD-S02-01 Page 18/37

19 No. Of T2 preps Represents the number of T2-prepared images used for T2 fitting. The minimum value is 3 and the maximum value is 12. T2 prep duration (array) Represents the T2 prep duration. The array ranges from 1 to the number selected in No. of T2 Preps. Motion correction This checkbox activates registration of individual T2 prep images. If selected, motion correction will be performed before curve-fitting and map generation. Recov HB This parameter controls the number of heartbeats that are used for recovery of the magnetization before the next T2p and readout. During the recovery heartbeats, the sequence plays gradient noises so that the subject doesn t assume that the scan has finished and inadvertently start breathing. Systolic imaging This checkbox should be enabled if imaging is performed during systole. It prevents possible SAR warnings in the case of a short trigger delay by including the time during diastole in the SAR calculations. Example images _780-ASD-S02-01 Page 19/37

20 Figure 10: Advantage of Motion Correction. Top row shows the original uncorrected T2- weighted images and the corresponding T2 map. Slight variations in cardiac cycle and/or inconsistent breathhold can produce erroneous map due to the pixel-wise nature of the fit. On the other hand, with non-rigid motion correction, such minor inconsistencies can be eliminated and more accurate T2 map can be generated. Figure 11: Comparison of T2 map images with T2-weighted STIR and late gadolinium enhancement (LGE) images. The patient > 90% occlusion of RCA, confirmed by x-ray angiogram. STIR images show no hyper-intensity in inferior regions and high stagnant endocardial blood signal. The T2 map nicely delineates increased T2 values in inferior regions and at endocardial borders. LGE image show the extent of the infarct _780-ASD-S02-01 Page 20/37

21 T1 map T2 map PSIR LGE Figure 12: Patient with a myocardial infarct, showing an elevation in both T1 and T2 maps, and a correlation with late gadolinium enhancement image. Courtesy: Drs. Arai and Kellman, NHLBI, Bethesda, MD, USA Scanning tips The T2 mapping is performed in a breathhold fashion. Ideally, the trigger delay and acquisition window should be modified so that the scan window is positioned during diastole, but acquisition is also possible during systole if a thicker myocardium is desired. In the case of systole imaging, be sure to select Systolic imaging on the Special card. A number of imaging protocols are described towards the end of this document. The sequence acquires multiple images in one breath-hold, each at a different T2 preparation time. The possible values for T2p are: 0 (i.e. no T2p) and ms; It is recommended that the maximum T2p value be in the vicinity of the expected T2 of the tissue being imaged. For instance, T2 of healthy human myocardium is reported to be around 55 ms. Accordingly, the maximum T2p value should be about 55 ms. To minimize breath-hold time in cardiac imaging, we recommend the acquisition of 3 T2P images, i.e. in the WIP special card, set the No of T2 Preps to 3. These 3 images are acquired in single-shot mode, with a gap of 3-4 RR intervals _780-ASD-S02-01 Page 21/37

22 in between to allow for sufficient T1 recovery. If the patient has a sufficiently low heart rate (RR interval > 1000) -, then a gap of 2-3 RR intervals may be used (this is done by setting Recov HB in the T2 mapping Special card). The default protocol is set up accordingly to acquire three T2p at 0ms, 25ms, and 55 ms; these times were chosen to optimize linear least squares fitting for myocardial T2 Mapping. If the sequence is to be used to generate a T2 map of a different tissue, then a different set of T2p times should be used. A detailed list of T1 and T2 values of different tissues can be found in reference [16]. Note that both conventional and adiabatic T2 preparation can be selected for T2 mapping. A centric FLASH readout is more accurate in measuring T2 values, whereas linear TrueFisp slightly overestimates T2. On the other hand, TrueFisp results in a higher SNR than FLASH. The reason for keeping a linear readout for TrueFisp is to avoid artifacts due to off resonance. The error in T2 estimate using a linear TrueFisp is due to T1 effects between the T2p and the center of the readout. T2 mapping is not recommended post contrast, but its worth noting that if the linear TrueFisp version is run post contrast, the shorter T1 may result in large errors. As such, we recommend using the centric FLASH version of the T2 mapping sequence. Figure 13 shows examples of T2 mapping acquired using centric FLASH and linear TrueFisp readouts. T2Prep TE = 0 ms T2Prep TE = 25 ms T2Prep TE = 55 ms _780-ASD-S02-01 Page 22/37

23 T2Prep TE = 0 ms T2Prep TE = 25 ms T2Prep TE = 55 ms Figure 13. T2 Maps using FLASH (top) and TrueFisp (bottom) readout schemes. For T2 mapping at 3T the use of the Cardiac Shim option is recommended. For further details, please check appendix _780-ASD-S02-01 Page 23/37

24 T2* mapping using multi-echo GRE Sequence description This WIP uses ECG-triggered GRE sequence to acquire multiple signal echoes during the T2* decay. Figure 14: Schematic of data acquisition for T2* mapping. Data is acquired over several heartbeats and at the same diastolic phase of the cardiac cycle. As all echoes are acquired in a short time after an RF pulse, all multi-echo images are intrinsically registered. Therefore, no motion correction is required during map generation. The signal at each echo-time (TE) is given by, S(t n ) = ρ 0 exp(-te n /T * 2 ) (4) Where, TE n = n th * echo-time, ρ 0 = initial signal intensity, and T 2 = effective * T2 for the voxel. Equation (4) assumes a unique T 2 * value per voxel. In voxels where multiple T 2 * components are present, this estimation will provide an effective T 2 * estimate. To generate an inline T2* map, an integrated image reconstruction performs pixel-wise T2* estimation using a robust fitting technique [20,21], in which the signal at each TE is iteratively weighted to reflect its fidelity to monoexponential decay curve. Signal points farther from the ideal relaxation curve are weighted lower, _780-ASD-S02-01 Page 24/37

25 reducing their influence on the fit. The weights of outlier points can be completely zero, eliminating their negative impact on the fit. Use interface A dedicated UI element is available in the Sequence Special card to generate T2* maps. Figure 15: User Interface parameters for T2* Mapping Parameter map type This selection controls the type of parameter map (T1/T2/T2*). For T2* mapping, this should be set to T2* Map. This will convert the protocol to multi-echo GRE protocol, required for T2* mapping. It can also be set to None, in which case the sequence behaves just like the normal product CV sequence. Systolic imaging _780-ASD-S02-01 Page 25/37

26 This checkbox should be enabled if imaging is performed during systole. It prevents possible SAR warnings in the case of a short trigger delay by including the time during diastole in the SAR calculations. Example images Following figures provide some examples and clinical cases using the inline myocardial T2* mapping. Figure 16: DB-prep GRE images [A-1, B-1] and corresponding T2*-map [A-2, B-2] produced using inline analysis in two healthy volunteers. The contours in these images mark septal regions from which the average T2* value was estimated. The average T2* value within septal regions were 29.8 ± 4.0 ms and 27.2 ± 3.3 ms for these two subjects, which are significantly above T2* < 20 ms range indicating cardiac iron overload _780-ASD-S02-01 Page 26/37

27 Figure 17: T 2 * estimates in 3 patients with suspected iron overload. [A-1, B-1, C-1] A DBprep GRE image showing the region used for T2* estimate within CMRTools. The estimated T2* is listed directly below each image. [A-2, B-2, C-2] Corresponding T2* maps obtained with inline analysis. Average of pixel-wise T2* estimate was obtained from indicated septal region. In all 3 cases, the average value obtained from T2*-map closely matches the one calculated using CMRTools _780-ASD-S02-01 Page 27/37

28 Scanning tips The T2*-map measurement should be performed in a breathhold fashion. The readout is typically positioned at mid-diastolic period so as to minimize cardiac motion. Moreover, it is important that the images are acquired in the same cardiac phase. Therefore, if there is high variation in the length of the cardiac cycle, then imaging during systole may improve phase consistency. For T2* estimation in the heart, dark-blood preparation is utilized to minimize the effects of moving blood. Such dark-blood protocols produce more homogeneous appearance and more accurate T2* estimates [19]. Installation Prerequisite scanner software baseline The required baseline for this software is: VD13A (WIP 780): N4_VD13A_LATEST_ _P16 (SP4) Installation of sequence / ICE files - Copy installation folder CD_contents_780 (VD13A) into a temporary folder on the scanner. - Double-click the InstallFiles_TXMapping.bat file to copy the sequence/ice files to appropriate directories. Please check the output to ensure that all files have been copied across properly. If this is a reinstallation then the ICE dll s may not install properly. If this happens, please reboot the scanner and try again _780-ASD-S02-01 Page 28/37

29 Protocols This package delivers protocols targeting heart T1, T2, T2* mapping. The protocols for heart are ECG-triggered protocols. These protocols should be used with breathholding and the trigger delay should be adjusted so that data is acquired during the diastolic period of cardiac motion. Import of protocols With reference to Figure 18: Open the EXAM EXPLORER and select the USER tree, press the right mouse button and select the dialog Exam Explorer Import. In the next dialog window, select Import and the appropriate drive where the protocol file (with extension.edx ) is located. This can be found in the folder protocols in the installation folder. Import the protocols, which will appear as a new element at the bottom of the USER tree. The inserted protocols are shown in Figure 19, and discussed below _780-ASD-S02-01 Page 29/37

30 Figure 18: Importing protocols T1 mapping: T2 mapping: T2* mapping: _780-ASD-S02-01 Page 30/37

31 Figure 19. Recommended protocols for T1, T2, and T2* mapping. T1 mapping protocols pre-con_molli_5(3)3_192mat (~8 second breath hold) Use pre-contrast with fast heart rates (> 90 bpm). The 5(3)3 nomenclature refers to 5 acquisition heartbeats, followed by 3 recovery heartbeats, then a further 3 acquisition heartbeats. This is faster than the original MOLLI 3(3)3(3)5 configuration, and the resulting T1 estimate is less dependent on heart rate. The 192 matrix is to minimize the readout time to reduce blurring during fast heart rates. post-con_molli_4(1)3(1)2_192mat (~8 second breath hold) Use post-contrast with fast heart rates (> 90 bpm). The shorter T1 values post contrast mean that less recovery time is necessary between acquisition periods. The 192 matrix is to minimize the readout time to reduce blurring during fast heart rates. pre-con_molli_5(3)3_256mat (~10 second breath hold) Use pre-contrast with slower heart rates (< 90 bpm). This is similar to precon_molli_5(3)3_192mat, except a 256 matrix can be acquired given the longer quiescent period associated with a slower heart rate. post-con_molli_4(1)3(1)2_256mat (~10 second breath hold) Use post-contrast with slower heart rates (< 90 bpm). This is similar to precon_molli_5(3)3_192mat, except a 256 matrix can be acquired given the longer quiescent period associated with a slower heart rate _780-ASD-S02-01 Page 31/37

32 MOLLI_5(3)3(3)5_256mat (~15 second breath hold) This is the original MOLLI protocols described in Ref. 3, and is included for consistency. Note that this protocol will produce slightly different (larger) T1 estimates than the counterpart in WIP 732B because of the new adiabatic inversion pulse and the adiabatic correction factor. T2 mapping protocols T2_3pt_3recovHB_gre (~11 second breath hold) This involves 3 T2 preparations (0, 25 and 55 ms) separated by recovery periods of 3 heart beats. The gradient echo readout is recommended for more accurate T2 estimates with less dependence on T1. The gradient echo readout is also more robust to artifacts, but offers a lower SNR than the TrueFISP counterpart. T2_3pt_3recovHB_trufi (~11 second breath hold) This uses the identical acquisition strategy to the previous sequence, but with a TrueFISP readout. This protocol is included for consistency with previous version of this WIP, and should be used if a higher SNR is required. T2* mapping protocols T2star_8echo_db_gre_1recovHB (~16 second breath hold) A segmented 8 echo monopolar gradient echo readout is used with a dark blood preparation pulse to sample the T2* decay. Similar to T2star_8echo_db_gre except a recovery heartbeat is inserted between each segment for more _780-ASD-S02-01 Page 32/37

33 consistent blood/artifact suppression. This protocol also includes ipat2, fat suppression and a larger flip angle. Recommended for 3T systems. T2star_8echo_db_gre (~16 second breath hold) This protocol remains unchanged from the previous WIP version. Recommended for 1.5T systems. User feedback The authors would be grateful for feedback from collaborators using this WIP. In particular, any comments on the following items would be of high interest: Clinical applications General performance and effectiveness of the package Specific protocol improvement Suggestions for addition and/or removal of features In addition, DICOM images or clinical case studies using this WIP package are very welcome. References 1. Schulz-Menger J, Friedrich MG. Magnetic Resonance Imaging in Patients with Cardiomyopathies: When and Why. Herz 2000;25(4): Ugander M, Bagi PS, Oki AJ, Chen B, Hsu L-Y, Aletras AH, Shah S, Greiser A, Kellman P, Arai AE. Quantitative T1-maps delineate myocardium at risk as accurately as T2-maps - experimental validation with microspheres. Journal of Cardiovascular Magnetic Resonance 2011;13(Supplement 1): _780-ASD-S02-01 Page 33/37

34 3. Messroghli DR, Radjenovic A, Kozerke S, Higgins DM, Sivananthan MU, Ridgway JP. Modified Look-Locker inversion recovery (MOLLI) for highresolution T1 mapping of the heart. Magn Reson Med 2004; 52: Messroghli DR, Greiser A, Fröhlich M, Dietz R, Schulz-Menger J. Optimization and Validation of a Fully-Integrated Pulse Sequence for Modified Look-Locker Inversion-Recovery (MOLLI) T1 Mapping of the Heart. J Magn Reson Imag 2007; 26: Xue H, Shah S, Greiser A, Guetter C, Chefd hotel C, Zuehlsdorff S, Guerhing J, Kellman P. Improved motion correction using image registration based on variational synthetic image estimation: application to inline T1 mapping of myocardium, Journal of Cardiovascular Magnetic Resonance 2011, 13(Suppl 1):P21 6. Xue H, Shah S, Greiser A, Guetter C, Littmann A, Jolly MP, Arai AE, Zuehlsdorff S, Guehring J, Kellman P. Motion correction for myocardial T1 mapping using image registration with synthetic image estimation. Magn Reson Med Jun;67(6): Kellman P, Wilson JR, Xue H, Ugander M, Arai AE. Extracellular volume fraction mapping in the myocardium, part 1: evaluation of an automated method. J Cardiovasc Magn Reson Sep 10;14:63 8. Kellman P, Wilson JR, Xue H, Bandettini WP, Shanbhag SM, Druey KM, Ugander M, Arai AE. Extracellular volume fraction mapping in the myocardium, part 2: initial clinical experience J Cardiovasc Magn Reson Sep 11;14: Abdel-Aty H, Simonetti OP, Friedrich MG. T2-Weighted Cardiovascular Magnetic Resonance Imaging. J Magn Reson Imag 2007; 26: _780-ASD-S02-01 Page 34/37

35 10. Abdel-Aty H, Cocker M, Meek C, Tyberg JV, Friedrich MG: Edema as a very early marker for acute myocardial ischemia: a cardiovascular magnetic resonance study. J Am Coll Cardiol 2009, 53: Abdel-Aty H, Boye P, Zagrosek A, Wassmuth R, Kumar A, Messroghli D, Bock P, Dietz R, Friedrich MG, Schulz-Menger J: Diagnostic performance of cardiovascular magnetic resonance in patients with suspected acute myocarditis: comparison of different approaches. J Am Coll Cardiol 2005, 45: Butler CR, Thompson R, Haykowsky M, Toma M, Paterson I: Cardiovascular magnetic resonance in the diagnosis of acute heart transplant rejection: a review. J Cardiovasc Magn Reson 2009, 11: Arai AE: Using magnetic resonance imaging to characterize recent myocardial injury: utility in acute coronary syndrome and other clinical scenarios. Circulation 2008, 118: Kellman P, Aletras AH, Mancini C, McVeigh ER, Arai AE: T2-prepared SSFP improves diagnostic confidence in edema imaging in acute myocardial infarction compared to turbo spin echo. Magn Reson Med 2007, 57: Giri S, Chung YC, Merchant A, et al. T2 quantification for improved detection of myocardial edema. J Cardiovasc Magn Reson 2009;11: Bottomley PA, Foster TH, Argersinger RE, Pfeifer LM: A review of normal tissue hydrogen NMR relaxation times and relaxation mechanisms from MHz: dependence on tissue type, NMR frequency, temperature, species, excision, and age. Med Phys 1984, 11: Pennell DJ, T2* Magnetic Resonance and Myocardial Iron in Thalassemia, Ann. N.Y. Acad. Sci. 1054: , He T, Gatehouse PD, Kirk P, Mohiaddin RH, Pennell DJ, Firmin DN, Myocardial T2* measurement in iron-overloaded thalassemia: an in vivo _780-ASD-S02-01 Page 35/37

36 study to investigate optimal methods of quantification, Magn Reson Med 2008 Nov; 60(5): He T, Gatehouse PD, Kirk P, Tanner MA, Smith GC, Keegan J, Mohiaddin RH, Pennell DJ, Firmin DN, Black-blood T2* technique for myocardial iron measurement in thalassemia, J Magn Reson Imaging Jun;25(6): Shah S, Xue H, Greiser A, Weale P, He T, Firmin DN, Pennell DJ, Zuehlsdorff S, Guehring J, Inline Myocardial T2* Mapping with Iterative Robust Fitting, Proc. Of SCMR / EuroCMR Joint Scientific Sessions, Feb Rousseeuw, R. J. & A. M. Leroy Robust Regression and Outlier Detection. New York: Wiley. Acknowledgements The authors gratefully acknowledge the invaluable collaboration with Dr. Peter Kellman and Hui Xue (NIH, Bethesda, MD), Dr. Martin Ugander (Karolinska University Hospital, Lund, Sweden), Dr. Daniel Messroghli and Cardiovacular MRI group at Charité (Berlin-Buch, Germany), Dr. Eric Schelbert (UPMC, Pittsburgh, PA), Dr. Orlando Simonetti (Ohio State University, Columbus, OH), Dr. David Firmin, Dr. Dudley Pennell, Dr. Taigang He (Royal Brompton Hospital, London, UK). Moreover, authors are also thankful to many Siemens colleagues including Dr. Wolfgang Rehwald, Peter Weale, Dr. Xiaoming Bi, Saurabh Shah, Dr. Jens Guehring, Dr. Aurelien Stalder, Dr Marie-Pierre Jolly and Dr. Sven Zuehlsdorff _780-ASD-S02-01 Page 36/37

37 Appendix 1: Hints for the use of the Cardiac Shim For TrueFISP based mapping protocols, in particular at 3T the use of the Cardiac Shim can be beneficial to reduce off-resonance artefacts. With VD11D and VD13A the Cardiac Shim is provided as a product option, as part of the adjustment framework. In the protocols provided with this package, the Cardiac Shim can be activated as the shim option in the Adjustments subcard. While its use can help a lot to improve the results, some caution is indicated regarding the workflow if the procedure is used outside the Dot framework. The adjustment volume that shall be adopted for covering the heart volume is not automatically propagated to new successive protocols if they are dragged from the exam database to the exam queue. Therefore, it is recommended to check the adjstment volume before a mapping protocol is started. The display of the adjustment volume represented as a green box can be activated in the graphical slice positioning tool in the View dropdown menue. The best approach to propagate the adopted adjustment volume from a previous scan to a protocol open for editing is to right click on the desired source protocol in the scan queue and select Copy Parameter > Adjust volume or if desired Slices & Adjustvolumes. Figure 20: Propagation of the adjustment volume to the open protocol _780-ASD-S02-01 Page 37/37