How to Analyze Inversion Recovery Magnetic Resonance Images

Abstract

A method for analyzing MRI images is described, including acquiring at least three medical MR images of a subject, analyzing the at least three medical MR images to determine a water T1 map, applying field strength correction and iron correction to the determined water T1 map to generate a corrected water T1 map, generating one or more simulated MRI images based on the water T1 map, and fitting the one or more simulated MR images to determine a standard cT1 image of the subject.

Description

[Technical Field]

[0001] The present invention relates to a method for analyzing magnetic resonance images (MRI) to generate a composite MR signal relaxation curve using multiple images that may be acquired from a variety of different MRI scanners. When used, the terms "LiverMultiScan" (registered trademark), "Siemens" (registered trademark), "Matlab" (registered trademark), and "Mathworks" (registered trademark) are acknowledged as registered trademarks. [Background technology]

[0002] Magnetic resonance imaging (MRI) scanning techniques can be used to obtain images of the human body with contrast that depends on the nuclear magnetic resonance (NMR) relaxation properties of the atomic nuclei being imaged (usually hydrogen atoms in water or fat). These properties have long been known to depend on the environment that produces the spin properties of the atoms. For example, T1, T2, and T2 * The properties depend on the magnetic environment of the atoms and also on the motion of the molecules within this environment.

[0003] In non-viscous liquids (e.g., cerebrospinal fluid (CSF)), the uniform magnetic field environment and unhindered, rapid motion of water molecules results in long T1 and T2 for hydrogen nuclei of water. Protons bound to or interacting with proteins have hindered motion, resulting in significantly shorter T2 and T1. These relaxation properties have proven useful because the ensemble-averaged fluid environment of fat and water hydrogen nuclei often differs between healthy and diseased tissues.

[0004] By collecting a series of images showing the same anatomical structures and applying image contrast with different sensitivities to these relaxation properties, it is possible to create parametric maps of the relaxation properties. This presents several challenges:

[0005] First, while it is desirable for collected images to be perfectly registered, data acquisition in abdominal tissues is difficult due to the twin challenges of cardiac and respiratory motion, which must be managed in acquisition and processing by methods such as motion suppression, motion compensation, and redundant acquisition to discard erroneous data.

[0006] Second, extracting parametric maps from images with different contrasts is difficult due to the noise level in the data, the small number of data points, and the complexity of fitting functions due to imperfect acquisition and mixed sources of MRI signal intensity.

[0007] Applicant pioneered the use of T1-based image contrast based on T1 determined by the Modified Look-Locker Inversion Recovery (MOLLI) method, as described in GB 2498254 and Messroghli DR, Radjenovic A, Kozerke S, Higgins DM, Sivananthan MU, and Ridgway JP, "Modified Look-Locker inversion recovery (MOLLI) for high-resolution T1 mapping of the heart," Magn Reson Med 2004;52:141-146. Because iron in the body influences T1 and because iron concentrations in the liver vary widely, Applicant employed a pioneering approach to correct T1 for the iron concentration present in each liver. Applicant calls this metric cT1 (corrected T1), and the LMS (LiverMultiScan (RTM)) product determines cT1 maps in human livers normalized to standard levels of liver iron. T1 also depends on the magnetic field strength of the MRI scanner (typically 1.5T and 3T scanners). Measurements are standardized to values obtained with a 3T scanner. Finally, because MRI scanners vary slightly across manufacturers, all values presented here are standardized to measurements on a Siemens (RTM) 3T scanner. Being able to determine stable, robust, and standardized measurements for patients using any commercially available scanner is an important foundation for commercial delivery, because it allows for initial insights such as, "If your cT1 is greater than 850 milliseconds, you belong to a population that will benefit from a particular treatment." Without standardization, such simplification is impossible.Parametric mapping using MRI is an inherently complex approach, but in settings where it needs to be delivered in a simple manner, the development of standardized metrics is needed that could potentially provide parameter ranges that could lead to stratification decisions (e.g., cT1 > 825 ms indicates disease and therefore the patient should be administered medication). This makes it an attractive and scalable technology that can be used at any MRI center worldwide.

[0008] Figure 1 shows a normalized cT1 map derived using a 1.5T MOLLI acquisition, with the region of interest highlighted in the diagram.

[0009] Figures 2(a) to 2(c) show cT1, T2, and T3 images acquired using different image acquisition methods. * , and proton density fat fraction (PDFF) values. * and PDFF images are acquired by multi-echo spoiled gradient echo acquisition. cT1, MOLLI and T2 * Derived from data.

[0010] It should be noted that while cT1 is standardized, it is not a suitable (metrological) measure of T1. The MOLLI approach to T1 measurement does not rely solely on T1 (as it should), but also on T2, magnetization transfer, the level of fat in the tissue (PDFF, proton density fat fraction, reported as a percentage of the signal from fat compared to the signal from fat + water), and other effects. These deficiencies stem from the use of MOLLI acquisition, which requires data collection within a short breath-hold period and synchronization to the cardiac cycle (although this approach can be used asynchronously in some cases).

[0011] Although cT1 is imperfect, it has been used in many studies. This means that it has been validated in biopsies, prognosis, and multiple clinical trials, and therefore can be considered a standard, albeit undoubtedly imperfect. Therefore, cT1 is a metric of interest because it shows clear correlations, even though it is not scientifically pure from the perspective of MR physics. Confirming a threshold (e.g., the 825 milliseconds mentioned above) requires much research and is impossible without a stable method.

[0012] Previously, cT1 could only be determined from MOLLI acquisition, but this has limitations. For example, -If the MRI scanner does not support the MOLLI acquisition method - If the MRI scanner does not support the specific timing of the MOLLI sequence required for accurate cT1 measurements - If the user is interested in 3D coverage of the liver (MOLLI is a 2D sequence, and as a result, acquiring a 3D volume requires many breath-holds, making it impractical) If the user is interested in collecting information at very high spatial resolution (not supported by MOLLI acquisition) - When breath-holding is an issue and the MOLLI sequence cannot be used · When the MOLLI sequence is unreliable due to spatial variations in B1+ (RF excitation field) or B0 (static field homogeneity).

[0013] In such situations, users may need cT1 map information, which would not be available using conventional MOLLI-based approaches. Summary of the Invention

[0014] According to an embodiment of the present invention, there is provided a method for analyzing MR images, comprising: acquiring at least three medical MR images of a subject and a reference MR image of the subject acquired without an inversion pulse; analyzing the at least three medical MR images to determine a water T1 map; applying field strength correction and iron correction to the determined water T1 map to generate a corrected water T1 map; generating one or more simulated MRI images based on the water T1 map; and fitting the one or more simulated MR images to determine a standardized cT1 image of the subject.

[0015] In a preferred exemplary embodiment of the present invention, the standardized cT1 image is a corrected cT1 image obtained from an image acquired on a reference MRI scanner for a subject with normal iron levels.

[0016] More preferably, the simulated MR images are generated using data from a dictionary of synthetic MR signal relaxation curves.

[0017] In an exemplary embodiment of the present invention, the data in the dictionary of synthetic MR signal relaxation curves is selected to match at least one of the corrected water T1 map and PDFF map for the acquired medical MR image.

[0018] Preferably, each of the acquired medical MR images comprises a pair of images having the same inversion time. More preferably, each successive medical MR image has a longer inversion time than the previously acquired medical MR image. In an exemplary embodiment of the present invention, the minimum inversion time for the first medical image is between 0.010 and 1.0 seconds.

[0019] In a preferred exemplary embodiment of the present invention, said at least three medical MR images include eight MR images.

[0020] Preferably, said at least three medical MR images are acquired such that there are redundant MR images.

[0021] In an exemplary embodiment of the present invention, the at least three medical MR images are inversion recovery MR images.

[0022] Preferably, one or more of the at least three medical MR images are acquired without the use of an inversion pulse.

[0023] In a preferred exemplary embodiment of the invention, said multiple acquisition medical MR images are acquired using a single shot acquisition, more preferably said single shot acquisition comprising at least one of EPI or single shot fast echo.

[0024] In an exemplary embodiment of the present invention, the magnetic field strength correction and iron correction for generating a corrected water T1 map uses a forward Bloch simulation. More preferably, the forward Bloch simulation uses a PDFF value, a T2 * The inputs include one or more of a T2 value, a water T1 value, and a pulse sequence for an MRI scanner used to acquire the medical MR images. Preferably, the field strength correction is based on a modification of the nominal field strength used to acquire at least three medical inversion recovery MR images. More preferably, the iron correction is based on a T2 value. * At least one of the map and the B0 field strength is used to correct for variations in iron concentration from normal values.

[0025] In a preferred exemplary embodiment of the present invention, a composite image is generated by analysis of the plurality of medical MR images.

[0026] Furthermore, in an exemplary embodiment of the present invention, the water T1 map is determined for the composite image.

[0027] More preferably, the single inversion pulse is an adiabatic pulse.

[0028] In an exemplary embodiment of the invention, the method further comprises the step of acquiring an additional medical MR image immediately before or immediately after the acquisition of said plurality of medical MR images, preferably said additional medical MR image being a multi-echo spoiled gradient-echo acquired image.

[0029] In a preferred embodiment of the invention, the original MRI images are acquired at 0.3-3.0T. [Brief explanation of the drawings]

[0030] Details, aspects, and examples of the present invention are described, by way of example only, with reference to the drawings, in which like reference numbers are used to identify like or functionally similar elements, and in which elements are illustrated for simplicity and clarity and have not necessarily been drawn to scale. [Figure 1] 1 shows a cT1 map derived using the prior art method, 1.5T MOLLI. [Figure 2(a)] 1 shows a cT1 image slice acquired using the prior art MOLLI method. [Figure 2(b)] 1 shows a T2* image slice acquired using the prior art MOLLI method. [Figure 2(c)] 1 shows a PDFF image slice obtained using the prior art MOLLI method. [Figure 3(a)-(d)] 1 shows raw data of images acquired at different inversion times (T1) according to an exemplary embodiment of the present invention. [Figure 3(e)] 10 shows an uncorrected TI map including T1 values in milliseconds (ms) according to an exemplary embodiment of the present invention. [Figure 4] The composite image of FIG. 1 shows water T1 as a subject. [Figure 5] Figures 4 and 2b and 2c show cT1 maps derived from the images. [Figure 6] 3 is a flow chart illustrating various method steps in a preferred exemplary embodiment of the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0031] The present invention will now be described with reference to the accompanying drawings, which illustrate examples of methods and apparatus for generating a composite cT1 MR image from a plurality of acquired MR images. Preferably, the MR images are inversion recovery images. However, it will be understood that the present invention is not limited to the specific examples described herein and illustrated in the accompanying drawings. Figure 6 is a flow chart illustrating the steps of the method in a preferred exemplary embodiment of the present invention.

[0032] In this first exemplary embodiment of the present invention, a GE 1.5T MRI scanner and a GE 3T scanner were used to acquire the 2D water T1 maps. However, other MRI scanners with different nominal magnetic field strengths can be used. For example, MRI scanners operating at 0.3-3T. The method is as follows: A subject is placed in the MRI scanner and imaged, preferably using a phased array abdominal coil, although other imaging coil arrays can also be used.

[0033] In an exemplary embodiment of the present invention, a single-shot fast spin-echo (IR SS-FSE) acquisition is preferably used to acquire medical images as acquired images, such as an MRI scan of the subject shown in step 602. Other acquisition methods, such as echo-planar imaging (EPI), may alternatively be used. In a preferred exemplary embodiment of the method, a fast 2D spin-echo medical MR image is acquired in a single shot after a preparatory inversion pulse. In a preferred exemplary embodiment of the present invention, the acquisition is fat-suppressed, as shown in step 604. If fat suppression is used, T1 fitting is performed in step 606 to model inversion recovery from the measured water signal. The method then proceeds directly to step 610, where a water T1 map is generated. If fat suppression was not performed during the MR image acquisition, a sequence model is used to model signals from fat and water components for the acquired image in step 608, and the sequence model is used to generate the water T1 map in step 610.

[0034] For fast 2D spin-echo acquisitions, "half-Fourier" imaging was used to reduce the acquisition time of MR image slices. In this case, just over half (typically 63%) of the k-space data is acquired directly, and the remaining k-space lines are estimated based on phase-conjugate symmetry (depending on the MRI scanner). However, while other imaging techniques can be used, the advantage of "half-Fourier" over single-spin fast spin-echo is that this technique results in a very high signal-to-noise ratio and limits unwanted T2 weighting. Alternatively, full k-space imaging can be used.

[0035] Preferably, the method requires the acquisition and analysis of at least three medical images (MR images) of the subject. In a preferred exemplary embodiment of the invention, the method requires the acquisition and analysis of eight MRI scans (acquired images) of the subject. Acquiring multiple MRI scans provides redundant MR images for analysis. In a preferred exemplary embodiment of the invention, the at least three medical MR images are inversion recovery MR images.

[0036] In one embodiment of the invention, five MR image slices (8 mm thick, 15 mm apart) were acquired during a single breath-hold with a repetition time (TR) ranging from 500 ms to 20 s. Preferably, the TR is 2000 ms and the effective echo time is 35-40 ms. In an exemplary embodiment of the invention, the echo time is minimized, preferably in the range of 1-100 ms, to maximize the signal-to-noise ratio. Using the above parameters, data acquired at a resolution of 1.72 x 1.72 x 8 mm was reconstructed at a resolution of 0.86 x 0.86 x 8 mm. In a preferred exemplary embodiment of the invention, multiple acquired MR images are acquired using a single-shot acquisition. Preferably, the single-shot acquisition includes at least one of echo-planar imaging (EPI) or single-shot fast spin-echo acquisition. In an exemplary embodiment of the invention, one or more of the at least three medical MR images are acquired without the use of an inversion pulse.

[0037] The above single breath-hold medical MR scan acquisition was repeated at various inversion times (TI), as shown in Figures 3(a) through 3(d). Figure 3(a) shows a medical MR image acquired at TI = 50 ms and 1.5 T, Figure 3(b) shows a medical MR image acquired at TI = 800 ms and 1.5 T, Figure 3(c) shows a medical MR image acquired at TI = 1200 ms, and Figure 3(d) is a medical MR image acquired without a preparatory inversion pulse.

[0038] As shown in Figure 3(a), in an embodiment of the present invention, the shortest TI was 50 ms at 1.5 T and 75 ms at 3 T (not shown). Other inversion times and field strengths may also be used in alternative exemplary embodiments of the present invention. In an embodiment of the present invention, the minimum TI for the first medical MR image ranges from 0.01 to 1.0 s. At both exemplary field strengths (1.5 T and 3.0 T), the first medical MR image was followed by at least two subsequent medical MR images with TIs of 800 ms (Figure 3(b)) and 1200 ms (Figure 3(c)) at both field strengths, followed by a reference medical MR scan (Figure 3(d)) acquired without a preparatory inversion pulse (effectively, TI = infinity).

[0039] In one embodiment of the present invention, two sets of four medical MR images with different TI values are acquired (requiring a total of eight breath-holds) to obtain a total of eight medical images. Preferably, each acquired medical MR image includes a pair of images with the same inversion time. Acquiring multiple MR images in this manner creates data redundancy and increases the robustness of the acquired MR images against motion-related artifacts in post-processing. As long as at least three medical MR images are acquired for subsequent analysis, various methods can be used, such as combining multiple repetitions of the same acquisition, multiple sets of long and short inversion times (TIs), and data sets with or without non-inversion acquisitions. In one example of the present invention, each successively acquired medical MR image has a longer inversion time than the previous medical MR image. This allows for fine-tuning of the characteristics of the final map depending on whether the MR image acquisition prioritizes acquisition time, spatial resolution, or noise reduction.

[0040] Typically, the TI is in the range of 50 milliseconds to 2000 milliseconds, with additional reference medical MR image scans having significantly longer TIs, each of the four medical MR images having a different TI value. Preferably, the TI increases with each subsequent medical MR scan. The TI value is designed to optimize sensitivity to T1 changes of a particular tissue of interest. Thus, in a preferred exemplary embodiment of the present invention, each medical MR image has its own unique inversion time, and each successive medical MR image in a sequence of MR images acquired in a single acquisition session has a longer inversion time than the previously acquired medical MR image. Preferably, the method may further include acquiring an additional reference medical MR image immediately before or after the acquisition of the multiple medical MR images. Preferably, the additional medical MR images are multi-echo spoiled gradient-echo acquired images.

[0041] To compare the method of the present invention with the prior LMS (LiverMultiScan (RTM)) method, standardized MR images required for LMS (LiverMultiScan (RTM)) acquisition were also collected (i.e., MOLLI-T1, LMS-MOST (for iron) [multi-echo spoiled gradient-echo acquisition], LMS-IDEAL (for fat in this case) [multi-echo spoiled gradient-echo acquisition], etc., see, e.g., https: / / doi.orq / 10.1002 / jmri.20831 Step 630 in FIG. 6 shows the LMS-MOST acquisition for iron, and step 632 shows the T2 * In step 638, the IDEAL acquisition used for fat is shown, and a PDFF map is obtained in step 640. In step 634, the T2 * The map and PDFF map are fitted to the model (step 634). * The image is the result of steps 630 and 632 in Figure 6, and the PDFF image shown in Figure 2(c) is the result of steps 638 and 640 in Figure 6. Step 636 is the determination of liver concentration, and step 642 is the determination of liver fat content from the PDFF.

[0042] In a preferred exemplary embodiment of the present invention, the LMS-MOST acquisition in step 630 is acquired using a 5-(1)-1-(1)-1 acquisition scheme with a 35 degree excitation pulse and a balanced bSSFP (balanced steady-state free-flow) readout, where the acquisition is ECG-gated, the slice thickness is 6 mm, the acquisition time is approximately 10 seconds (less than 10 heartbeats), and a shMOLLI acquisition (but not shMOLLI processing) (see https: / / patents.qooqle.com / patent / US20120078Q84A1 / en, Figure 2B and / or https: / / jcmr-online.biomedcentral.com / articles / 10.1186 / 1532-429X-12-69).

[0043] The LMS-MOST acquisition in step 630 acquires thin slice (3 mm) spoiled gradient echo MR images at multiple echo times, specifically for the T2 * It is designed to measure the B0 artifact by selectively combining the same images multiple times (preferably seven times at 1.5T) to minimize variability, and by using thin slices to reduce the effect of inter-slice dephasing, it is less susceptible to respiratory and B0 artifacts. Therefore, the acquisition time for an LMS-MOST acquisition is approximately 10 seconds.

[0044] The LMS-IDEAL acquisition in step 638 also uses multi-echo spoiled gradient echo acquisition with low excitation flip angles to minimize differential T1 weighting between fat and water species and generate accurate PDFF maps (after processing) in step 640.

[0045] Preferably, acquired medical MR images are adjusted for signal intensity using the presence or absence of an inversion pulse and adjusting the inversion time relative to the inversion pulse. Preferably, there are multiple different times between the inversion pulse and acquisition ("inversion time"), with the initial time delay being less than 0.5 seconds. Preferably, acquisition should use a single-shot acquisition (EPI or single-shot fast spin echo) to minimize motion-induced image artifacts. Preferably, images using the same acquisition module but without the inversion pulse should also be prepared. Preferably, there is multiple redundancy in the acquisition, with more images collected than there are parameters to be fitted. Preferably, multiple slices are taken during each breath-hold, and all of these images are displayed with the same image contrast.

[0046] In an exemplary embodiment of the invention, fat suppression should be used during acquisition (see: https: / / mriquestions.eom / uploads / 3 / 4 / 5 / 7 / 34572113 / haase frahm chess.pdf) to remove fat signal from the image. This is shown in step 604 of FIG. 6. In a second exemplary embodiment of the invention, water excitation should be used during acquisition (using excitation pulses designed to excite water but not fat; these are typically binomial RF pulses) (see: https: / / mriquestions.eom / uploads / 3 / 4 / 5 / 7 / 34572113 / water excitation radiol 2e2243011227.pdf) to remove fat signal. In other embodiments, other techniques of fat suppression (e.g., DIXON (see: https: / / pubmed.ncbi.nlm.nih.gOv / 6089263 / ) and / or fat suppression) may be used to reduce fat signal from the data.

[0047] In an exemplary embodiment of the present invention, medical MR image data are collected during eight separate breath-holds. Preferably, four different MR image contrasts are acquired using fat-suppressed single-shot fast spin-echo acquisition, and the acquisition of these four contrasts is repeated. Preferably, each acquired medical MR image consists of a pair of images having the same inversion time. In a preferred embodiment of the present invention, each successively acquired medical MR image has a longer inversion time than the previous medical MR image. More preferably, the minimum inversion time for the first medical MR image is between 0.010 and 1.0 seconds. The image contrasts in the four images span a range of "inversion times," and in one of the four images, no inversion pulse is applied. Preferably, the type of RF pulse used for the inversion pulse is an "adiabatic pulse," ensuring consistent inversion of magnetization that is robust to small variations in B1+ and B0. In a preferred exemplary embodiment of the present invention, the single inversion pulse is an adiabatic pulse (see: https: / / mriquestions.eom / uploads / 3 / 4 / 5 / 7 / 34572113 / tannus-adiabaticpulses.pdf). An example of an inversion pulse for this purpose is a hyperbolic secant pulse of 10 millisecond duration.

[0048] In an exemplary embodiment of the present invention, medical image data for a plurality of medical images is acquired with echo planar imaging (EPI) rather than single shot fast spin echo.

[0049] The processing of these medical MR images may include preprocessing to assess the impact of breathing on the data. The processing of these MR images may also include a registration step to align the images to reduce the effects of patient motion and compensate for it. This typically involves optimizing a cost function using a realistic spatial transformation of the image data. One approach to this goal is to apply a registration algorithm to each pair to estimate the level of displacement locally in a portion of the field of view or globally across the entire field of view. Another approach is to generate a local similarity or dissimilarity map for each image pair. This approach can reduce errors in the T1 map caused by motion occurring within the plane of image acquisition. Motion estimates can be used as part of a data subset selection strategy to determine which data to include and which to exclude in pixel-by-pixel model fitting. Motion estimates can also be used to define a check against the entire set of acquired data to determine whether a minimally viable set that successfully fits is available or whether the entire dataset should be rejected.

[0050] Processing these MR images may also include excluding certain images from the final analysis to reduce the effects of patient motion. Calibration can be performed to set a tolerance threshold for the level of motion acceptable for subsequent pixel-by-pixel model fitting to water T1. Image rejection is applied when significant out-of-plane motion occurs and different tissues are sampled from one image to the next. This is distinct from the effects of in-plane motion. To detect and reject out-of-plane motion, a similarity metric is used when comparing two or more images. Images that are determined to be outliers relative to the other images in terms of the similarity metric are rejected, leaving a set of images that are always from the same slice. Another approach utilizes a separately acquired 3D dataset to assess slice motion and registration. This 3D acquisition provides a reference for the data that can be compared and registered. Typically, zero, one, or two images are rejected. Rejecting a large number of images can lead to poor fitting conditions and a poor fit of the data.

[0051] These processed MR images are used in the next fitting step. For some datasets, no processing may be necessary, i.e., if there is no patient motion or if the patient motion is not significant.

[0052] Uncorrected (i.e., non-iron-corrected) water T1 maps are generated from multiple medical MR images (preferably at least three medical MR images) using a pixel-by-pixel fitting approach in single or multiple slices using the following equation: S(TI,x,y,z) = A(x,y,z) + B(x,y,z).exp(-TI / water T1(x,y,z)) + epsilon(TI,x,y,z) where S(TI,x,y,z) is the signal intensity of the image at spatial coordinates x,y,z, and TI is the inversion time in the image to be fitted. where A and B are fitting parameters that vary as a function of image position. water T1 (x,y,z) is the uncorrected water T1 as a function of position in the data. Epsilon is a noise term that is minimized during fitting, typically using a least squares error criterion. The fitting is performed on the magnitude / absolute value of S(TI,x,y,z) or on the complex value of S(TI,x,y,z). In this embodiment, the fitting is performed in the magnitude domain.

[0053] For medical MR images where no inversion pulse is applied, the TI is set to a very large time so that the exponential term is equal to zero. For each spatial location (pixel coordinate), data fitting is performed by modeling the signal for each TI. Fitting can also be performed in complex space, where A and B include phase angle terms that vary with image location. Fitting can also be performed in the magnitude domain by magnitude transforming both the data and the fitting function.

[0054] In an alternative embodiment of the present invention, NOLLI_water T1 (water T1 determined by the NOLLI method) was determined by running a forward Bloch simulation and selecting the water T1 that best explains the data.

[0055] Preferably, the method also includes acquiring an additional medical MR image immediately before or after the acquisition of the first and second medical MR images, and preferably the additional medical MR image is a multi-echo spoiled gradient-echo acquired MR image.

[0056] In an exemplary embodiment of the present invention, MR images are acquired at 0.3-3.0T.

[0057] In an exemplary embodiment of the present invention, for the 8 breath-hold NOLLI_water T1 dataset, the data was loaded into Matlab (RTM) from Mathworks (RTM), Natick, Massachusetts, and fitted to the above equation in the magnitude domain by minimizing the least squares error at each image location. In other exemplary embodiments of the present invention, alternative software may be used for analysis of medical image data.

[0058] FIG. 3 illustrates the output from a NOLLI acquisition in a first exemplary embodiment of the present invention, followed by the initial processing of multiple medical MR images to generate a composite image of the subject. Preferably, a water T1 map is determined for the composite image. The uncorrected NOLLI T1 map was determined using Equation (1). Bloch equation simulation can be used to fit simulated data to different T1 values based on the actual pulse sequence used. While this approach was not used in this case, an alternative exemplary embodiment of the present invention could use a T1 optimization approach in a Bloch simulation of the pulse sequence to select simulated data, e.g., a T1 that most closely matches the data acquired by the scanner, to generate a simulated MR medical image.

[0059] PDFF maps can be calculated from a single MR image slice, a single voxel (spectroscopy), multiple MR image slices, or a complete 3D volume. In practice, in many situations (homogeneous liver disease), the variation in hepatic PDFF across the entire liver can be quite small, and a single value of PDFF can likely be used for these signal simulations, although PDFF maps may also be useful. Image registration of PDFF maps to other maps may be necessary when performing signal simulations.

[0060] Figure 4 shows a liver water T1 map (a parametric map of water T1) obtained for the original subject using fat-suppressed NOLLI data acquired on a 1.5T GE scanner. Fat suppression was used to minimize the effect of fat on the calculated water T1 values. A liver mask was generated using an exemplary embodiment of the present invention based on a machine learning algorithm that replicates the performance of a manually drawn mask. This figure illustrates the expected homogeneity of the organ of interest in the subject. As shown, the map shows no fat artifacts because fat was suppressed during acquisition of these images. The data used to create this image was acquired on a Siemens (RTM) 1.5T scanner. Regions containing blood flow are subject to blood flow artifacts in this method. Masking was used to remove background noise. Various masking approaches are available, none of which are essential to the present invention. In this case, the mask was generated based on a machine learning algorithm that replicates the performance of a manually drawn mask around the liver. While a manual approach could have been used, a machine learning approach was used for efficiency reasons.

[0061] For the LMS acquisition, data were fitted using the LMS Discover tool (in Matlab), which has similar general performance characteristics to the LMS Medical Device and provides flexibility for exploratory work through rapid prototyping.

[0062] Iron and PDFF (fat) measurements were obtained from Matlab (RTM) processing and NOLLI-water T1 maps to determine the cT1 that would have been measured if the sample had normal iron content and been scanned on a Siemens (RTM) 3T scanner using the MOLLI method. In this study, the T2 at 3T was used for ease of calculation. * is standardized to 23.1 ms. Alternative standards include other T2 * You may use the value.

[0063] This was done by first correcting the measured NOLLI-water T1 for the effect of iron, and then calculating the water T1 that the tissue would have had if it were at 3 Tesla (instead of the 1.5T at which the data were acquired). Iron correction is * The map and B0 field strength are used to determine iron concentration, using a known equation: (R1(3T) = R1o(3T) + HIC × 0.029 g / mg.s; where R1 = 1 / T1, and HIC is the liver iron concentration) to determine the effect on T1 due to deviations in iron concentration from normal. In an exemplary embodiment of the invention, field strength correction and iron correction are used to generate a corrected water T1 map in step 614. In an exemplary embodiment of the invention, a forward Bloch simulation is used. Preferably, the field strength correction is based on modifying the nominal field strength used to acquire at least three medical inversion recovery MR images. More preferably, the iron correction is used to generate a corrected water T1 map. * At least one of the map and the B0 field strength is used to correct for variations in iron concentration from normal values.

[0064] FIG. 6 illustrates the determination of a forward Bloch simulation at 650. A Bloch simulation is performed at step 652, providing input to a model signal at step 654. Based on this, a dictionary is constructed at step 656. This step is repeated across the parameter space (including water, T1, T2, and PDFF). The output of the forward Bloch simulation is provided to a dictionary of synthetic MR signal relaxation curve data at step 660. The dictionary is then used for dictionary matching at step 616, where raw data from the dictionary is selected that matches the measured corrected water T1 and PDFF from step 614. Preferably, simulated MR images are generated using data from the dictionary of synthetic MR signal relaxation curves. In a preferred exemplary embodiment of the present invention, the data in the dictionary of synthetic MR signal relaxation curves is selected to match at least one of the corrected water T1 and PDFF of the acquired medical MR images.

[0065] Preferably, the forward Bloch simulation has inputs including one or more of a PDFF value, a T2 estimate, a water T1 value, and details of the pulse sequence of the scanner used to acquire the medical MR images.

[0066] The field correction performed in step 612 is based on an empirically determined mapping of the effect of magnetic field strength on T1, assessed from a group of subjects scanned at each magnetic field strength. The liver iron concentration determination output from step 636 is provided in step 612 and used for iron correction of the water T1 map previously generated in step 610. In exemplary embodiments of the invention, magnetic field correction can be performed before or after iron correction. In some exemplary embodiments of the invention, scanner-dependent correction can also be performed.

[0067] Once this field strength and iron-corrected water T1 are calculated, this value and PDFF are used in the Bloch equation for the MOLLI sequence to obtain the signal expected if this subject's iron levels were normal (in this case, a T2 of 23.1 ms at 3T). * (expressed as ) is determined. This normalization does not vary depending on the subject's weight, age, or gender. If the PDFF was not included in the simulation, this would be the standardized value for a person with normal iron levels, a heart rate of 60 bpm, and no liver fat. Bloch equation simulation of the MOLLI sequence acquires iron- and field-strength-corrected water T1, PDFF, and the correct pulse sequence (implemented in a reference MRI scanner, such as a Siemens (RTM) 3T scanner). Starting with a fully relaxed magnetization vector (Mz = 1, Mx = 0, My = 0) at time = 0, the magnetization vector is evaluated at short time increments (typically 50 microseconds) to assess how it changes due to the effects of RF pulses, nonresonant effects, spin-destruction gradients, and spin relaxation (T1 and T2). The simulated magnetization vector is recorded at the time the signal is sampled in the image sequence. Fat and water signals are simulated separately and then combined according to their PDFF ratios.

[0068] Preferably, the forward Bloch simulation uses known characteristics of the pulse sequence of a 3T reference MRI scanner, which is run sequentially in a simplified manner for each pixel, resulting in a series of simulated signals at each pixel. The simulation uses a synthetic heart rate of 60 beats per minute, and for pixels with a PDFF > 30%, the user can run the simulation with a fixed PDFF of 30%. These values are selected as standard parameters, but other values for the standardization parameters can also be used. Fat is simulated using a standard six-peak fat model known to represent liver fat. Additionally, the T2 relaxation of water is fixed to the T2 of a liver with normal iron levels. The fat and water signals are combined using known concentrations from the PDFF(x, y, z) map (this can be location-dependent, but alternatively, a global measurement can be used). The MOLLI sequence involves acquiring seven or more images at different inversion times (TI), resulting in a signal sequence called S(TI, x, y, z).

[0069] Finally, these simulated MOLLI signals obtained in step 618 are fed into a standard LMS cT1 fitting pipeline (at normal iron levels, since iron has already been corrected) to determine cT1. This is output in step 620 as the super-normalized cT1. That is, the resulting S(T1,x,y,z) matrix is fitted to each pixel to generate a map of cT1(x,y,z). This cT1 should be equivalent to the super-normalized MOLLI derived cT1, which is known to be normalized for field strength and MRI vendor. This final fitting step performs a least-squares fit for each pixel of the following function: S(TI) = (A - B exp(-TI / Ti*) Then, T1 is determined as follows. T1 = T1* ((B / A) -1)

[0070] Fitting MOLLI data typically involves creating a dictionary of varying A, B, and T1*, constructing a fitting function for each value in the dictionary, and identifying the one that best represents the data (perhaps by least-squares estimation, or by selecting the combination with the maximum value of the dot product between the normalized data and the normalized dictionary and comparing the two curves). Alternatively, A, B, and T1* can be fitted using an iterative search approach (such as least-squares or the Levenberg-Marquardt method). In practice, this fitting is not particularly difficult to perform using standard approaches. Knowing A, B, and T1*, T1 can be determined. In this case, S(TI) is generated so that the resulting T1 is normalized to cTI. This process results in a standardized cT1 measurement.

[0071] Some systematic differences (on the order of 5%) are expected between this modeling approach and the direct MOLLI acquisition approach. These offsets are due to several sources, most obviously magnetization transfer effects, as well as subtle biases in water T1 mapping. Users apply a fixed offset to cT1 for each acquisition pipeline based on the results of comparing known systems with human data acquired using this approach. Each acquisition pipeline (described above) requires specific calibration, with the goal that these methods return the same value for a subject with a cT1 of 825 ms by modeling the effects of fat, iron, etc. This allows for the use of a single offset per scanner, ensuring that cT1 from different scanners is standardized across the range of fat, iron, water T1, etc. This offset may be applied to cT1, or it may be applied to another parameter (e.g., water T1) to achieve the same effect.

[0072] Figure 5 shows a standard cT1 MR image of the liver acquired at 1.5T using the acquisition method of the present invention, mapped to cT1 at 3T using the novel approach described above. In this example, raw MR data was acquired on a 1.5T GE MRI scanner using a fat-suppressed sequence as shown in Figure 6. Preferably, at least three medical MR images of the subject are acquired and analyzed to determine a water T1 map. Field strength correction and iron correction are applied to the water T1 to generate a corrected water T1 map. Based on the water T1 map, one or more simulated MR images are generated and then fitted to determine the standard cT1 MR image described above. As described above, the simulated MR images are provided to the cT1 fitting pipeline to determine the corrected cT1 image. The cT1 map shown shows mapped cT1 values corresponding to a standard MOLLI sequence on a 3T Siemens (RTM) Prisma scanner. The cT1 map was derived from data indicating the ROI location on the subject. The data were from a Siemens (RTM) 1.5T scanner. 3 × ROI = 629, 624, 663 ms. These are shown as highlighted circles in the images. The pooled median for the images was cT1 = 640 + / - 51 ms. The pooled median is typically used, but other metrics such as mean, median, or pooled mean can also be used. However, the pooled median is preferred as it is more robust.

[0073] The mapping algorithm is only applicable to regions within the liver; regions outside the liver are not accurately mapped.

[0074] In the exemplary embodiments of the present invention described above, image processing can be performed in different dimensions. For example, the methods of the present invention can be applied at the level of a single large voxel (as in spectroscopy), or over a single region of interest. They can also be applied pixel-by-pixel in a 2D image, or over an entire 3D volume. The most likely use cases are generating cT1 maps for a single 2D slice, multiple 2D slices, or an entire 3D volume.

[0075] This novel acquisition and processing pipeline can provide synthetic MR signal relaxation curves that exhibit spatial uniformity similar to that of the standard LMS MOLLI approach. The quantitative values determined with this novel acquisition and processing pipeline are consistent with those of the standard LMS MOLLI approach. Therefore, this novel acquisition and processing pipeline provides a mechanism for providing an alternative approach to LMS MOLLI cT1. A further advantage of this invention is that cT1 maps can be acquired from any MRI scanner, regardless of the scanner's magnet strength or the company that manufactured the scanner. Furthermore, cT1 maps acquired using this invention may have greater reproducibility, higher spatial resolution, or other characteristics. This means that synthetic cT1 maps acquired using this invention are superior to cT1 maps determined with prior art MOLLI acquisition techniques.

[0076] The invention has been described with reference to the accompanying drawings.

[0077] It will be understood, however, that the present invention is not limited to the particular examples described herein, as illustrated in the accompanying drawings. Moreover, because the illustrated embodiments of the present invention can be implemented, for the most part, using electronic components and circuits well known to those skilled in the art, details will not be described beyond what is deemed necessary for an understanding and appreciation of the basic concepts of the present invention, as illustrated above, and will not be described beyond what is deemed necessary so as not to obscure or confuse the teachings of the present invention.

[0078] The present invention can be embodied in a computer program for running on a computer system, and comprises at least code portions for carrying out the steps of a method according to the invention, or for enabling a programmable device to carry out the functions of a device or system according to the invention, when the computer program is run on a programmable device such as a computer system.

[0079] A computer program is a list of instructions, such as for a particular application program and / or operating system. A computer program may include, for example, one or more of the following: subroutines, functions, procedures, object methods, object implementations, executable applications, applets, servlets, source code, object code, shared libraries / dynamic load libraries, and / or other sequences of instructions designed to run on a computer system. A computer program may be stored internally on a tangible, non-transitory computer-readable storage medium or transmitted to a computer system via a computer-readable transmission medium. All or part of a computer program may be provided on a computer-readable medium permanently, removably, or remotely coupled to an information processing system.

[0080] A computer process typically includes an executing (running) program or part of a program, current program values and state information, and the resources used by the operating system to manage process execution. An operating system (OS) is software that manages the sharing of a computer's resources and provides programmers with an interface for accessing those resources. An operating system processes system data and user input and responds by allocating and managing tasks and internal system resources as a service to system users and programs.

[0081] A computer system may include, for example, at least one processing unit, associated memory, and a number of input / output (I / O) devices. When executing a computer program, the computer system processes information in accordance with the computer program and generates resulting output information via the I / O devices.

[0082] In the foregoing specification, the invention has been described with reference to specific embodiments thereof. However, it will be apparent that various modifications and changes can be made therein without departing from the scope of the invention as set forth in the appended claims. Those skilled in the art will recognize that boundaries between logic blocks are merely exemplary, and that alternative embodiments may combine logic blocks or circuit elements or impose alternative functional decompositions on various logic blocks or circuit elements. Accordingly, it should be understood that the architectures described herein are merely exemplary, and that many other architectures that achieve the same functionality may in fact be implemented. Any arrangement of components that achieve the same functionality is effectively "associated" such that the desired functionality is achieved. Thus, any two components described herein that combine to achieve particular functionality can be considered to be "associated" with each other such that the desired functionality is achieved, regardless of the architecture or intermediate components. Similarly, any two components so associated can also be considered to be "operably connected" or "operably coupled" to each other to achieve the intended functionality.

[0083] Moreover, those skilled in the art will recognize that the boundaries between operations described above are merely illustrative. Operations may be combined into a single operation, a single operation may be distributed among additional operations, or operations may perform operations that at least partially overlap in time. Moreover, alternative embodiments include multiple instances of a particular operation, and the order of operations may be changed in various other embodiments. However, other modifications, variations, and alternatives are possible. Accordingly, the specification and drawings should be interpreted in an illustrative rather than a restrictive sense. Unless otherwise noted, terms such as "first" and "second" are used to arbitrarily distinguish between elements described by such terms. Thus, these terms are not necessarily intended to indicate a temporal or other priority of such elements. The mere fact that certain actions are recited in mutually different claims does not indicate that a combination of these actions cannot be advantageously used.

Claims

1. 1. A method for analyzing magnetic resonance (MR) images, comprising: acquiring at least three medical inversion recovery MR images of a subject and a reference MR image of the subject acquired without an inversion pulse; analyzing the at least three medical MR images to determine a water T1 map; applying magnetic field strength correction and iron correction to the determined water T1 map to generate a corrected water T1 map; generating a plurality of simulated magnetic resonance images (MRIs) based on the water T1 map and proton density fat fraction (PDFF) map for the acquired medical MR images; fitting the plurality of simulated MR images to determine a standardized corrected T1 (cT1) image for the subject; A method comprising:

2. 2. The method of claim 1, wherein the standardized cT1 image is a corrected cT1 image obtained from an image acquired with a reference MRI scanner for a subject with normal iron levels.

3. The method of claim 1 or 2, wherein the simulated MR image is generated using data from a dictionary of synthetic MR signal relaxation curves.

4. 4. The method of claim 3, wherein the data in the dictionary of synthetic MR signal relaxation curves is selected to match at least one of the corrected water T1 map and proton density fat fraction (PDFF) map for the acquired medical MR image.

5. 5. The method of claim 1, wherein each of the acquired medical MR images comprises a pair of images having the same inversion time.

6. The method of claim 1 , wherein each successively acquired medical MR image has a longer inversion time than the previously acquired medical MR image.

7. The method according to claim 5 or 6, wherein the minimum inversion time of the first medical MR image is between 0.010 and 1.0 seconds.

8. The method of claim 1 , wherein the at least three medical MR images include eight MR images.

9. The method of claim 1 , wherein the at least three medical MR images are acquired such that redundant MR images exist.

10. The method of claim 1 , wherein the multiple acquired medical MR images are acquired using a single-shot acquisition.

11. The method of claim 10 , wherein the single-shot acquisition comprises at least one of EPI or single-shot fast echo.

12. 12. The method of claim 1, wherein the magnetic field strength correction and iron correction to generate a corrected water T1 map uses forward Bloch simulation.

13. The method of claim 12 , wherein the magnetic field strength correction is based on a modification of a nominal magnetic field strength used to acquire the at least three medical inversion recovery MR images.

14. The iron correction is ＊ 14. The method of claim 12 or 13, wherein at least one of the map and the B0 field strength is used to correct for differences in iron concentration from normal levels.

15. The forward Bloch simulations were performed using PDFF values, T ＊ 15. The method of claim 12, having inputs including one or more of a T1 value, a water T1 value, and a pulse sequence for an MRI scanner used to acquire the medical MR images.

16. 16. The method of claim 1, wherein a composite image is generated by analysis of the plurality of medical MR images.

17. The method of claim 16 , wherein the water T1 map is determined for the composite image.

18. 18. The method of claim 1, wherein the single inversion pulse is an adiabatic pulse.

19. 19. The method of claim 1, comprising acquiring an additional medical MR image immediately before or after acquiring the plurality of medical MR images.

20. 20. The method of claim 19, wherein the additional medical MR image is a multi-echo spoiled gradient echo acquired image.

21. The method of any one of claims 1 to 20, wherein the original MRI images are acquired at 0.3 to 3.0 T.

Images