Method for analyzing, inverting and recovering magnetic resonance image

Abstract

A method of analyzing an MRI image is described. The method comprises the following steps: 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 intensity 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; the one or more simulated MR images are fitted to determine a standard cT1 image of the subject.

Description

Technical Field

[0001] The present invention relates to a method of analyzing Magnetic Resonance Imaging (MRI) images to generate a synthetic MR signal relaxation curve using multiple images that can be acquired from a series of different MRI scanners. The terms "LiverMultiScan", "Siemens", "Matlab" and "Mathworks" used herein are all registered trademarks. Background Art

[0002] Magnetic resonance imaging (MRI) scanning technology can be used to obtain images of the human body, and its contrast depends on the nuclear magnetic resonance (NMR) relaxation properties of the imaging nuclei (usually hydrogen atoms in water and fat). It has long been known that these properties depend on the environment of the atoms that produce these spin properties. The T1, T2 and T2* properties (for example) depend on the magnetic environment of the atoms and also on the movement of these molecules in this environment.

[0003] In non-viscous fluids such as cerebrospinal fluid (CSF), hydrogen nuclei of water have long T1 and T2 due to the uniform magnetic field environment and the rapid unhindered motion of water molecules. Protons bound to or interacting with proteins have hindered motion and may have much shorter T2 and T1. These relaxation properties have been found to be useful because the overall average fluid environment of hydrogen nuclei in fat and water is often different in diseased tissue than in healthy tissue.

[0004] By collecting a series of images showing the same anatomical structure but with different image contrasts for these relaxation properties, a parametric map of the relaxation properties can be created. There are several challenges:

[0005] First, data acquisition of abdominal tissue is challenging due to the dual challenges of cardiac and respiratory motion, as the images collected need to be absolutely aligned. Acquisition and processing must manage this motion by methods such as freezing the motion, correcting for the effects of the motion, or acquiring redundant data and discarding erroneous data.

[0006] Second, it can be difficult to extract parameter maps from images with varying contrast. This difficulty can be due to noise levels in the data, a small number of data points, complications with fitting functions due to imperfect acquisition, or confounding sources of MRI signal intensity.

[0007] Applicants have pioneered the use of T1-based imaging contrast, which is based on T1 determined by the MOLLI (Modified Look-Looker Inversion Recovery) method. This method is described in patent GB2498254 and in the article "Modified Look-Locker inversion recovery (MOLLI) for high-resolution T1 mapping of the heart" by Messroghli DR, Radjenovic A, Kozerke S, Higgins DM, Sivananthan MU and Ridgway JP. Magn Reson Med 2004; 52: 141-146. Since iron in the body affects T1, and the iron concentration in the liver varies greatly, applicants used a pioneering method that corrects T1 according to the iron concentration present in each liver. The applicant refers to this metric as cT1 (corrected T1), and the LMS (LiverMultiScan (RTM)) product determines a map of cT1 in the human liver, which is normalized to a standard level of liver iron. The T1 measurement also depends on the magnetic field strength of the MRI scanner (1.5T and 3T scanners are typically used). The measurement is standardized to the measurement results obtained on a 3T scanner. Finally, there are some subtle differences between MRI scanners from different manufacturers, so the values ​​discussed in this article are all based on measurements on a Siemens (RTM) 3T scanner. The ability to determine stable and reliable standardized measurements for patients using any commercially available scanner is an important cornerstone of commercial products because it potentially enables statements such as "If your cT1 is above 850ms, you belong to a group that benefits from a specific treatment", and this simplification is impossible without standardization. Parametric mapping using MRI is a complex approach in itself, but it needs to be used in settings where it needs to be provided in a simple way. The development of a standardized metric that has the potential to provide parameter ranges leading to stratified decisions (e.g., cT1>825ms indicates disease, so the patient should receive medical treatment) can be used in any MRI center in the world and is an attractive scalable technology.

[0008] Figure 1 A standard cT1 map obtained using the 1.5T MOLLI acquisition method is shown, with the region of interest highlighted.

[0009] Figure 2(a) to Figure 2(c) Images representing cT1, T2*, and PDFF (Proton Density Fat Fraction) obtained using different image acquisition methods are shown. T2* and PDFF images were acquired using Multi-echo Spoiled Gradient Echo Acquisition. cT1 was derived from MOLLI and T2* data.

[0010] It should be noted that although cT1 is standardized, it is not a good (metrical) measure of T1. The MOLLI method of measuring T1 depends not only on T1 (as it should), but also on T2, magnetization transfer, fat levels in the tissue (PDFF proton density fat fraction, expressed as a percentage of fat signal to fat+water signal), and other influencing factors. These deficiencies arise from the use of MOLLI acquisition, which requires data to be collected during a short breath-hold time and gated to the cardiac cycle (this method can also be used in a non-gated manner in some cases).

[0011] Although cT1 is not perfect, it has been used in many studies. This means that it has been validated for biopsies, prognosis, and in multiple clinical trials, so while it is undoubtedly not perfect, it does represent a standard of some kind. Therefore, cT1 is a very interesting metric because it has clear relevance, even if it is not scientifically pure from an MR physics perspective. Many studies are needed to determine a threshold (such as the 825ms defined above), so this is impossible without a stable method.

[0012] Conventionally, cT1 can be determined from MOLLI acquisitions alone, but this has limitations, such as:

[0013] - If the MRI scanner does not have the MOLLI acquisition method;

[0014] - If the MRI scanner cannot support the specific timing required by the MOLLI sequence to achieve accurate cT1 measurement;

[0015] - If the user is interested in 3D coverage of the liver (MOLLI is a 2D sequence, so collecting a 3D volume would require multiple breath holds, which is impractical);

[0016] - If the user is interested in collecting very high spatial resolution information (not supported by MOLLI acquisition);

[0017] - If the MOLLI sequence cannot be used due to breath-holding problems;

[0018] - If the MOLLI sequence is unreliable due to spatial variations of B1+ (RF excitation field) or B0 (homogeneity of static magnetic field).

[0019] In these cases, users need information from cT1 maps, which cannot be achieved using conventional MOLLI-based methods. Summary of the invention

[0020] According to an example of the present invention, a method for analyzing MR images is provided, comprising: acquiring at least three medical MR images of a subject, and a reference MR image of the subject acquired without using 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.

[0021] In a preferred exemplary embodiment of the present invention, the standardized cT1 image is a cT1 image corrected to an image acquired on a reference MRI scanner of a subject with normal iron levels.

[0022] Further preferably, the simulated MR image is generated using data in a dictionary of synthetic MR signal relaxation curves.

[0023] In an exemplary embodiment of the present invention, data in a dictionary of synthetic MR signal relaxation curves are selected to match at least one of a corrected water T1 map and a PDFF map for an acquired medical MR image.

[0024] Preferably, each acquired medical MR image comprises an image pair having the same inversion time. Further preferably, the inversion time of each consecutively acquired medical MR image is longer than the inversion time of the previously acquired medical MR image. In an exemplary embodiment of the present invention, the minimum value of the inversion time of the first medical image is between 0.010-1.0 seconds.

[0025] In a preferred exemplary embodiment of the present invention, the at least three medical MR images include eight MR images.

[0026] Preferably, the at least three medical MR images are acquired such that there will be redundant MR images.

[0027] In an exemplary embodiment of the present invention, the at least three medical MR images are inversion recovery MR images.

[0028] Preferably, one or more of the at least three medical MR images are acquired without using inversion pulses.

[0029] In a preferred exemplary embodiment of the present invention, the multiple acquired medical MR images are acquired using single shot acquisition. Further preferably, the single shot acquisition includes at least one of the following: EPI or single shot fast echo.

[0030] In an exemplary embodiment of the present invention, the field strength correction and iron correction used to generate the corrected water T1 map use forward Bloch simulation. Further preferably, the input of the forward Bloch simulation includes one or more of the PDFF value, the T2* value, the water T1 value and the pulse sequence of the MRI scanner used to acquire the medical MR image. Preferably, the field strength correction is based on the modification of the nominal field strength used to acquire at least three medical inversion recovery MR images. Further preferably, the iron correction uses at least one of the T2* map and the B0 field strength to correct the difference in iron concentration from normal levels.

[0031] In a preferred exemplary embodiment of the present invention, the analysis result of the plurality of medical MR images is a composite image.

[0032] In a further exemplary embodiment of the present invention, the water T1 map is determined for the composite image.

[0033] Further preferably, the single inversion pulse is an adiabatic pulse.

[0034] In an exemplary embodiment of the present invention, the method further comprises the step of acquiring another medical MR image immediately before or after acquiring the plurality of medical MR images. Preferably, the another medical MR image is a Multi-Echo Spoiled Gradient Echo Acquisition image.

[0035] In a preferred exemplary embodiment of the present invention, the original MRI image is acquired at 0.3-3.0T. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] Further details, aspects and embodiments of the present invention will be described by way of example only with reference to the accompanying drawings. In the drawings, the same figure numbers are used to identify the same or functionally similar elements. The elements in the drawings are drawn for simplicity and clarity and are not necessarily drawn to scale.

[0037] Figure 1 The cT1 map obtained using the 1.5T MOLLI prior art method is shown.

[0038] FIG2( a ) shows a cT1 image slice acquired using the prior art MOLLI method;

[0039] FIG2( b ) shows a T2* image slice acquired using the prior art MOLLI method; and

[0040] FIG2( c ) shows a PDFF image slice acquired using the prior art MOLLI method.

[0041] Figure 3 (a)-(d) show raw data of images acquired at different inversion times (T1) according to an exemplary embodiment of the present invention; and

[0042] Figure 3 (e) shows an uncorrected TI map according to an exemplary embodiment of the present invention, where the T1 value is in milliseconds (ms).

[0043] Figure 4 Shows Figure 1 Water T1 (Water T1) of the subject in the composite image.

[0044] Figure 5 Shows from Figure 4 and Figure 2b as well as Figure 2c cT1 map derived from the images in .

[0045] Figure 6 is a flow chart showing various method steps in a preferred exemplary embodiment of the present invention. DETAILED DESCRIPTION

[0046] The present invention will now be described with reference to the accompanying drawings, which illustrate an example of a method 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 should be understood that the present invention is not limited to the specific examples described herein and shown in the accompanying drawings. Figure 6 is a flow chart showing the method steps in a preferred exemplary embodiment of the present invention.

[0047] In a first exemplary embodiment of the present invention, 2D water T1 maps were acquired using a GE 1.5T MRI scanner and a GE 3T scanner, respectively, although other MRI scanners with different nominal magnetic field strengths may also be used. For example, an MRI scanner may operate between 0.3-3T. The method for acquiring is as follows. A subject is placed in an MRI scanner and imaging is preferably performed using a phased array abdominal coil, although other imaging coil arrays may also be used.

[0048] In an exemplary embodiment of the present invention, a single-shot fast spin echo (IR SS-FSE) acquisition is used to acquire a medical image as an acquisition image, preferably an MRI scan of a subject as shown in step 602. Alternatively, other acquisition methods, such as echo planarimaging (EPI), may 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, as shown in step 604, the acquisition is fat suppressed. If fat suppression is used, a T1 fitting step is performed at 606 to simulate inversion recovery based on the measured water signal. The method then directly proceeds to step 610 for generating a water T1 map. If fat suppression is not used during the MR image acquisition process, for the acquisition image, step 608 uses a sequence model to simulate the signals of fat and water components, and uses the sequence model to generate a water T1 map at 610.

[0049] For fast 2D spin echo acquisitions, "half Fourier" imaging is used to reduce MR image slice acquisition time, in which case slightly more than half (typically 63%) of the k-space data is acquired directly, and the remaining k-space lines are estimated based on phase conjugate symmetry (by the MRI scanner). However, other imaging methods can also be used, and the advantage of "half Fourier" combined with single-shot spin-fast spin echo is that this method can achieve a very high signal-to-noise ratio and limited unnecessary T2 weighting. Full k-space imaging can also be used.

[0050] Preferably, the method requires acquiring and analyzing at least three medical images (MR images) of the subject. In a preferred exemplary embodiment of the present invention, the method requires acquiring and analyzing eight MRI scan images (acquisition images) of the subject. Multiple MRI scan images are acquired so that there are redundant MR images in the analysis. In an exemplary embodiment of the present invention, at least three medical MR images are inversion recovery MR images.

[0051] In one example of the present invention, five MR image slices (each 8 mm thick and 15 mm apart) are acquired during a single breath hold, using a repetition time (TR) in the range of 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 present invention, the echo time is minimized and preferably in the range of 1-100 ms to maximize the signal-to-noise ratio. The MR image acquisition with the above parameters obtains data with a resolution of 1.72×1.72×8 mm and is reconstructed at a resolution of 0.86×0.86×8 mm. In a preferred exemplary embodiment of the present invention, multiple acquired MR images are acquired using a single acquisition. Preferably, the single acquisition includes at least one of echo planar imaging EPI or a single fast spin echo acquisition. In an exemplary embodiment of the present invention, one or more of the at least three medical MR images are acquired without the use of an inversion pulse.

[0052] The above-mentioned single breath-hold medical MR scan acquisition is repeated at a series of inversion times (TI), such as Figure 3 (a) to Figure 3 (d) as shown. Figure 3 (a) shows a medical MR image acquired at 1.5T, TI = 50ms; Figure 3 (b) shows a medical MR image acquired at 1.5T, TI = 800ms; Figure 3 (c) shows a medical MR image acquired at TI = 1200 ms; Figure 3 (d) is a medical MR image acquired without preparing an inversion pulse.

[0053] In the example of the present invention, as shown in the figure, the shortest TI at 1.5T is 50 milliseconds (eg Figure 3 (a)), and 75 milliseconds at 3T (not shown). In other exemplary embodiments of the present invention, other inversion times and field strengths may also be used. In one example of the present invention, the minimum TI of the first medical MR image is in the range of 0.01 to 1.0 seconds. At two example field strengths (1.5T and 3.0T), the first medical MR image is followed (at two field strengths) by at least two subsequent medical MR images, whose TIs are 800 ms ( Figure 3 (b)) and 1200ms( Figure 3 (c)) followed by a reference medical MR scan ( Figure 3 (d)), the scan was acquired without preparing an inversion pulse (in effect TI = close to infinity).

[0054] In one example of the present invention, a set of four medical MR image acquisitions, with different TI values, are acquired twice (requiring a total of 8 breath holds), resulting in a total of eight medical images. Preferably, each acquired medical MR image contains an image pair with the same inversion time. Acquiring such multiple MR images provides redundancy for the data and increases the robustness of the acquired MR images to motion-related artifacts in post-processing. Different schemes can be used, including more or fewer repetitions of the same acquisition, more or fewer different inversion times (TI), and selecting a data set that includes or does not include non-inversion acquisitions, as long as at least three MR medical images are acquired for subsequent analysis. In one example of the present invention, each consecutively acquired medical MR image has a longer inversion time than the previously acquired medical MR image. These allow the characteristics of the final map to be fine-tuned depending on whether the MR image acquisition prioritizes acquisition time, spatial resolution, or noise reduction.

[0055] Typically, the TI ranges from 50 ms to 2000 ms, and there is an additional reference medical MR image scan with a very long TI, with each of the four medical MR images having a different TI value. Preferably, the TI of each subsequent medical MR scan increases. The value of the TI is intended to optimize the sensitivity to T1 changes in a particular tissue of interest. Therefore, in a preferred exemplary embodiment of the present invention, each medical MR image has a unique inversion time, and each consecutive medical MR image in a sequence of MR images acquired in one acquisition session has a longer inversion time than a previously acquired medical MR image. Preferably, the method may further include the step of acquiring a further reference medical MR image immediately before or after acquiring the plurality of medical MR images. Preferably, the further medical MR image is a multi-echo spoiled gradient echo acquisition image.

[0056] In order to compare the method in the present invention with the previous LMS (LiverMultiScan (RTM)) method, standardized MRI images required for LMS (LiverMultiScan (RTM)) acquisition (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]) were also collected, see, for example, https: / / doi.org / 10.1002 / jmri.20831. Figure 6 Step 630 in FIG. 2 shows an LMS most acquisition for iron, which results in a T2* map in step 632. The IDEAL acquisition for fat is shown in step 638, and a PDFF map is obtained in step 640. Step 634 is fitting the T2* map and PDFF map to the model. The T2* image in FIG. 2(b) is Figure 6As a result of steps 630 and 632 in FIG. 2( c ), the PDFF image shown in FIG. 2( c ) is Figure 6 The results of step 638 and step 640 in step 636 are to determine the liver concentration and step 642 is to determine the liver fat content from the PDFF.

[0057] In a preferred exemplary embodiment of the present invention, the LMS most acquisition in step 630 uses a 5-(1)-1-(1)-1 acquisition scheme, employs a 35 degree excitation pulse and a balanced bSSFP (balanced steady state free precession) readout, the acquisition is cardiac gated, the slice thickness is 6 mm, the acquisition time is approximately 10 seconds (less than 10 heart beats), and the acquisition is performed using shMOLLI (although not shMOLLI processed) [ See also https: / / patents.google.com / patent / US20120078084A1 / en Figure 2B and / or https: / / jcmr-online.biomedcentral.com / articles / 10.1186 / 1532-429X-12-69].

[0058] The LMS-MOST acquisition in step 630 acquires thin-section (3 mm) spoiled gradient echo MR images with multiple echo times. It is specifically designed to measure T2* in the liver and is robust to respiratory and B0 artifacts by using multiple repetitions of the same image (preferably 7 at 1.5 T) that are selectively combined to minimize variance, and by using thin sections that help reduce through-slice dephasing effects. The imaging time of the LMS-MOST acquisition is approximately 10 seconds.

[0059] The LMS-IDEAL acquisition in step 638 also uses multi-echo spoiled gradient echo MR image acquisition with a low excitation flip angle to minimize the T1-weighted difference between fat and water material, thereby obtaining an accurate PDFF map (after processing) in step 640.

[0060] Preferably, the acquired medical MR images will use the presence or absence of an inversion pulse and adjustment of the inversion time relative to the inversion pulse to adjust the signal intensity. Preferably, there are multiple different times between the inversion pulse and the acquisition ("inversion times"), with the first time delay being less than 0.5 seconds. Preferably, the acquisition should use a single acquisition (whether EPI or single fast spin echo) to minimize image artifacts due to motion. Preferably, there should also be an image that uses the same acquisition module but does not include the inversion pulse. Preferably, there is multiple redundancy in the acquisition so that more images are collected than the parameters are fitted. Preferably, multiple slices are imaged in each breath hold, and each image has the same image contrast.

[0061] In an exemplary embodiment of the present invention, fat suppression (see https: / / mriquestions.com / uploads / 3 / 4 / 5 / 7 / 34572113 / haase_frahm_chess.pdf) should be used in the acquisition to eliminate the fat signal in the image. Figure 6 604. In a second exemplary embodiment of the invention, water excitation (using excitation pulses intended to excite water rather than fat, these are typically binomial RF pulses) [see https: / / mriquestions.com / uploads / 3 / 4 / 5 / 7 / 34572113 / water_excitation_radiol_2e2243011227.pdf] should be used to eliminate fat signals in the acquisition. In other embodiments, other fat suppression techniques (e.g., DIXON [see https: / / pubmed.ncbi.nlm.nih.gov / 6089263 / ]) may also be used to reduce fat signals in the data.

[0062] In an exemplary embodiment of the invention, data for the medical MR images are collected during 8 separate breath holds. Preferably, four different MR image contrasts are collected using a single fast spin echo acquisition with fat suppression, and these 4 contrasts are acquired repeatedly. Preferably, each acquired medical MR image in the claim comprises an image pair with the same inversion time. In a preferred example of the invention, the inversion time of each consecutively acquired medical MR image is longer than that of the previously acquired medical MR image. It is further preferred that the minimum value of the inversion time of the first medical MR image is between 0.010-1.0 seconds. The image contrasts in the 4 contrasts span a range of "inversion times" and no inversion pulse is applied in one of the 4 images. Preferably, the RF pulse type used for the inversion pulse is an "adiabatic pulse", which ensures a consistent inversion of the magnetization intensity, which is robust to small changes in B1+ and BO. In a preferred exemplary embodiment of the invention, the single inversion pulse is an adiabatic pulse [see https: / / mriquestions.com / 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 milliseconds duration.

[0063] In an exemplary embodiment of the present invention, echo planar imaging (EPI) is used to collect medical image data for multiple medical images rather than a single-shot turbo spin echo.

[0064] Processing of these medical MR images may involve preprocessing to assess the effects of breathing on the data. Processing of these MR images may involve a registration step to reduce the effects of patient motion and align the images to compensate for patient motion. Typically, this will optimize a cost function by applying an actual spatial transformation to the image data. One possible approach is to apply a registration algorithm to each pair of images and estimate the level of displacement locally in a portion of the field of view or across the entire field of view. Another approach is to generate local similarity or dissimilarity maps for the image pairs. This approach can reduce errors in the T1 map due to motion occurring within the image acquisition plane. Motion estimation can be used to determine which data should be included in each pixel model fit and which data should be excluded as part of a data subset selection strategy. Motion estimation can also be used to examine the entire acquired data set to determine if there is a minimum viable set of data available for a successful fit or if the data set should be rejected as a whole.

[0065] Processing of these MR images may also involve a step of rejecting 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 that is acceptable for any subsequent per-pixel model fit relative to the water T1. Image rejection can be applied when large through-plane motion occurs, so the tissue sampled is not the same in each image, which is different from the within-plane motion effect. To detect and reject through-plane motion, similarity measures are used when comparing two or more images. In the context of these similarity measures, images that are found to be outliers compared to the other images will be rejected, leaving a set of images that always belong to the same slice. Another approach can utilize a separately acquired 3D data set to evaluate slice motion and alignment, this 3D acquisition will provide a reference to which the data can be compared and registered. Typically, 0, 1, or 2 images will be rejected, and rejecting a large number of images will result in a poor fit to the data as the fit may become under-conditioned.

[0066] These processed MR images will be used in the next fitting step. There may be some datasets that do not require any processing, i.e. when there is no patient motion or the patient motion is not significant.

[0067] An uncorrected (ie, uncorrected for iron) water T1 map will be generated from multiple medical MR images (preferably at least three medical MR images) using a single or multiple slice pixel-by-pixel fitting method using the following formula:

[0068] 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)

[0069] Where S(TI,x,y,z) is the signal intensity of the image at the spatial coordinates x,y,z, and TI is the inversion time in the image being fitted.

[0070] Where: A and B are fitting parameters that vary as a function of image position.

[0071] waterT1(x,y,z) is the uncorrected water T1 as a function of position in the data.

[0072] Epsilon is the noise term that is minimized in the fit, usually using the least squares error criterion.

[0073] The fitting can be performed on the magnitude / absolute value of S(TI,x,y,z) or the complex value of S(TI,x,y,z). In this embodiment, we perform the fitting in the magnitude domain.

[0074] In the case of medical MR images without an applied inversion pulse, the TI is set to a very large time such that the exponential term is equal to zero. For each spatial location (pixel coordinate), the data is fitted by modeling the signal for each TI. The fit can be done in complex space, in which case A and B will include a phase angle term that varies with image position. The fit can also be done in the amplitude domain, with both the data and the fit function being amplitude transformed.

[0075] In an alternative example of the present invention, NOLLI_water T1 (water T1 determined by the NOLLI method) is determined by simulating a forward Bloch simulation and selecting the water T1 that best explains the data.

[0076] Preferably, the method of the present invention further comprises the step of acquiring a further medical MR image immediately before or after acquiring the first and second medical MR images. Preferably, the further medical MR image is a multi-echo spoiled gradient echo acquisition MR image.

[0077] In an exemplary embodiment of the present invention, the MR images are acquired at 0.3-3.0T.

[0078] In an exemplary embodiment of the present invention, for the NOLLI_water T1 dataset of 8 breath holds, the data was read into Matlab (RTM) (Mathworks (RTM)), Natick, MA), and the above equation was fitted in the amplitude domain by minimizing the least squares error at each image position. Other exemplary embodiments of the present invention may use other software to analyze medical image data.

[0079] Figure 3 The NOLLI acquisition output of a first exemplary embodiment of the present invention is shown, as well as preliminary processing of multiple medical MR images to generate a composite image of a subject. Preferably, a water T1 map is determined for the composite image. The uncorrected NOLLI-T1 map is determined using equation (1). The data can be fitted using a Bloch equation simulation that simulates the magnetization vector for different T1 values ​​based on the actual pulse sequence used. In this example, this method is not used, but another exemplary embodiment of the present invention may involve a method for optimizing the T1 in the Bloch simulation of the pulse sequence and selecting the T1 that produces simulated data, such as a simulated MR medical image, which is most similar to the data acquired by the scanner.

[0080] The PDFF map can be calculated from a single MR image slice, a single voxel (via spectroscopy), multiple MR image slices, or the complete 3D volume. In practice, in many cases (homogeneous liver disease), the variation in PDFF in the liver may be relatively small, so it may be possible to use a single PDFF value in these signal simulations, but a PDFF map may also be used. When performing signal simulations, it may be necessary to image-register the PDFF map with other maps.

[0081] Figure 4 An original subject's liver water T1 map (parametric map of water T1) acquired using a 1.5T GE scanner is shown, using fat suppressed NOLLI data. Fat suppression is used to minimize the effect of fat on the calculated water T1 values. A liver mask was generated based on a machine learning algorithm that replicates the performance of a manually drawn mask, using an exemplary embodiment of the present invention. The figure shows the expected uniformity in the subject's organ of interest. As shown, there is no fat effect in this figure because fat was suppressed during the acquisition of these images. The data used to generate this image was acquired on a Siemens (RTM) 1.5T scanner. In this method, areas containing flowing blood are affected by flow artifacts. Masking is used to remove background noise. A variety of masking methods can be used, none of which are critical to the present invention. In this example, a mask was generated based on a machine learning algorithm that replicates the performance of a manually drawn liver mask, although a manual method could have been used, a machine learning method was used for efficiency reasons.

[0082] For the LMS acquisition, the data were fit using the LMSDiscover tool in Matlab, which has the same general performance characteristics as the LMS Medical devices but has the additional flexibility to be used for rapid prototyping exploratory work.

[0083] Iron and PDFF (fat) measurements were obtained from Matlab (RTM) processing and NOLLI-water T1 maps, cT1 was determined as if the sample had normal levels of iron and was scanned using the MOLLI method on a Siemens (RTM) 3T scanner. In the present invention, T2* was normalized to 23.1 ms at 3T for ease of calculation. Other T2* values ​​may also be used as alternative standards.

[0084] The measured NOLLI-water T1 is first corrected for the effects of iron, and then the water T1 is calculated if the tissue is at 3 Tesla (instead of the 1.5T at which the data was acquired). The iron concentration is determined by using the T2* map and the BO field strength, and then the iron correction is performed using a known equation. The equation (R1(3T)=R10(3T)+HIC×0.029g / mg.s; where R1=1 / T1 and HIC is the liver iron concentration) determines the effect of T1 due to the difference in iron concentration from normal levels. In an exemplary embodiment of the present invention, field strength correction and iron correction are used to generate a corrected water T1 map, as shown in step 614. In an exemplary embodiment of the present invention, this will use forward Bloch simulation. 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. It is further preferred that the iron correction uses at least one of the T2* map and the B0 field strength to correct for the difference in iron concentration from normal levels.

[0085] The forward Bloch simulation is determined in Figure 6 650 of . A Bloch simulation is performed at step 652 and provides input for the simulated signal in step 654. Starting from this step, a dictionary is constructed at step 656. These steps will loop in the parameter space (including water, T1, T2 and PDFF) and be repeated. The output of the forward Bloch simulation is provided to a dictionary of synthetic MR signal relaxation curve data at step 660. Then, in step 616, the dictionary is used for dictionary matching, where raw data that matches the measured corrected water T1 and PDFF from step 614 is selected from the dictionary. 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, 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 image.

[0086] Preferably, the forward Bloch simulation has inputs including one or more of PDFF values, T2 estimates, water T1 values, and scanner pulse sequence details used to acquire the medical MR images.

[0087] The field strength correction performed in step 612 is based on an empirically determined map of the effect of field strength on T1 and is evaluated on a group of subjects scanned at each field strength. The liver iron concentration output determined in step 636 is provided in step 612 for iron correction of the water T1 map previously generated in step 610. In this exemplary embodiment of the invention, field strength correction may be performed before or after iron correction. In some exemplary embodiments of the invention, scanner dependent correction may also be performed.

[0088] Once the field- and iron-corrected water T1 is calculated, this value is used in conjunction with the PDFF in the Bloch equation for the MOLLI sequence to determine what the expected signal would be if the subject had normal iron levels (in this case represented by a T2* at 3T of 23.1 ms). This normalization does not change due to the subject's weight, age, or sex. If the PDFF is not included in the simulation, this will be a normalized value for a person with normal iron levels, a heart rate of 60 bpm, and no liver fat. The Bloch equation simulation of the MOLLI sequence uses the iron- and field-corrected water T1, the PDFF, and an exact pulse sequence implemented on a reference MRI scanner, such as a Siemens (RTM) 3T scanner, starting at time = 0, when the magnetization vector is fully relaxed (Mz = 1, Mx = 0, My = 0), and evaluating the effects of the RF pulse, off-resonance effects, spoiler gradients, and spin relaxation (T1 and T2) on the magnetization vector for short time increments (typically 50 microseconds). At the time point in the imaging sequence where the signal is sampled, the simulated magnetization vector is recorded. Fat and water signals were simulated separately and combined proportionally according to the PDFF.

[0089] Preferably, the forward Bloch simulation uses the known characteristics of the 3T reference MRI scanner pulse sequence and is performed pixel by pixel in a simplified manner, resulting in a series of simulated signals at each pixel. The simulation can be performed using a synthetic heart rate of 60 beats per minute, and for pixels with PDFF>30%, the user can fix the PDFF to 30%. These values ​​are chosen as standard parameters, although other values ​​can also be used as normalization parameters. Fat will be simulated using a standard 6-peak fat model known to represent liver fat. In addition, the T2 relaxation of water will be fixed to the T2 of a liver with normal iron levels. The fat and water signals will be merged using known concentrations in the PDFF(x,y,z) map (this can be position-dependent or a global measurement can be used). The MOLLI sequence collects 7 or more images at different inversion times (TI), which will derive a signal array S(TI,x,y,z).

[0090] Finally, in step 618, these simulated MOLLI signals obtained are input into the standard LMS cT1 fitting pipeline (iron corrected, so normal iron levels) to determine cT1 and output as super-normalized cT1 in step 620. That is, the resulting S(TI,x,y,z) matrix will be fitted at each pixel to produce a map of cT1(x,y,z). This cT1 should be equivalent to the cT1 obtained using the super-normalized MOLLI method, which is known to be normalized for field strength and MRI manufacturer. This final fitting step performs a pixel-by-pixel least squares fit of the function:

[0091] S(TI)=(AB exp(-TI / T1*)

[0092] And determine T1 as:

[0093] T1=T1*((B / A)-1)

[0094] The way MOLLI data are usually fit is by building a dictionary of different A, B, and T1* and building a fitting function for each parameter in the dictionary, and then seeing which one best represents the data (perhaps by least squares estimation, or by comparing the two curves by choosing the combination that maximizes the dot product of the normalized data and the normalized dictionary). Alternatively, it is also possible to fit A, B, and T1* using iterative search methods (i.e., minimizing least squares and Levenberg-Marquart). In practice, it is not difficult to perform such a fit using standard methods. Once A, B, and T1* are determined, T1 can be determined. In this case, S(TI) is generated in a way that ensures that the resulting T1 is normalized to cT1. This process produces a standardized cT1 measurement.

[0095] It is expected that there will be some small (~5%) systematic differences between this modeling approach and the direct MOLLI acquisition method. These biases come from multiple sources, most notably magnetization transfer effects, but also subtle biases in water T1 mapping. The user will apply a fixed bias to cT1 for each acquisition process based on comparisons to human data acquired using known systems and this method. Each acquisition process (as described above) requires specific calibration, with the goal that these methods give the same value for a subject with a cT1 of 825ms by modeling the effects of fat, iron, etc. A single offset can be used for each scanner, and cT1 from different scanners can be reliably normalized across fat, iron, water T1, etc. This bias can be applied to cT1, but also to other parameters to achieve the same effect (e.g., water T1).

[0096] Figure 5 A standard cT1 MR image of the liver acquired at 1.5T using the acquisition method of the present invention is shown and mapped to cT1 at 3T using the novel method described above. In this example of the present invention, a standard cT1 MR image of the liver acquired at 1.5T using the acquisition method of the present invention is shown. Figure 6The fat suppression sequence shown acquires raw MR data. Preferably, at least three medical MR images of the subject are acquired and analyzed to determine the water T1 map. Field strength correction and iron correction are applied to the water T1 to generate a corrected water T1 map. One or more simulated MR images are generated based on the water T1 map and fitted to determine the above-mentioned standard cT1 MR image. As described above, the simulated MR images are provided to the cT1 fitting process to determine the corrected cT1 image. The cT1 map shown in the figure shows the mapped cT1 values, corresponding to the standard MOLLI sequence on the 3T Siemens (RTM) Prisma scanner. The cT1 map is derived from data showing the location of the subject's ROI. The data is from a Siemens (RTM) 1.5T scanner. 3×ROI=629ms, 624ms, 663ms. These are represented by highlighted circles in the image. The summary median of cT1 in the image is 640+ / -51ms. The pooled median is often used, but other metrics such as mean, median, pooled mean, etc. can also be used, however the pooled median is more commonly used because it is more robust.

[0097] The mapping algorithm only works for the area inside the liver; areas outside the liver cannot be mapped correctly.

[0098] In the exemplary embodiments described herein, image processing can be performed in different dimensions. For example, the method of the present invention can be applied to a single large voxel (such as in spectral analysis) or a single region of interest, can be performed pixel by pixel on a two-dimensional image, or can be performed pixel by pixel on an entire three-dimensional volume. Preferably, the most likely use case is to generate a cT1 map in a single 2D slice, multiple 2D slices, or a 3D volume.

[0099] This novel acquisition and processing flow can provide a synthetic MR signal relaxation curve with similar spatial homogeneity as the standard LMS MOLLI method. The quantitative values ​​determined by this novel acquisition and processing flow produce values ​​consistent with the standard LMS MOLLI method. Therefore, this novel acquisition and processing flow provides a mechanism to replace the LMS MOLLI cT1 method. Another advantage of the present invention is that cT1 maps can be obtained from any MRI scanner, regardless of the magnetic field strength or manufacturer of the scanner. In addition, the cT1 maps obtained using the present invention may be more repeatable, or have a higher spatial resolution, or have some other characteristics, so that the synthetic cT1 obtained by the present invention is superior to the cT1 determined by the prior art MOLLI acquisition technology.

[0100] The present invention has been described in conjunction with the accompanying drawings. However, it should be understood that the present invention is not limited to the specific examples described herein and the contents shown in the accompanying drawings. In addition, since the embodiments shown in the present invention can be implemented to a large extent using electronic components and circuits known to those skilled in the art, in order to understand and appreciate the basic concepts of the present invention and in order not to confuse or deviate from the teachings of the present invention, the details will not be explained to a greater extent than is deemed necessary as shown above.

[0101] The present invention can be implemented in a computer program running on a computer system, which computer program comprises at least a code portion for executing the steps of the method according to the present invention or enabling the programmable device to execute the functions of the apparatus or system according to the present invention when running on a programmable device such as a computer system.

[0102] A computer program is a series of instructions, such as a specific application and / or operating system. A computer program may include one or more of the following: subroutines, functions, procedures, object methods, object implementations, executable applications, applets, servlets, source code, object code, shared libraries / dynamically loaded libraries, and / or other sequences of instructions designed for execution on a computer system. A computer program may be stored internally on a tangible and non-transitory computer-readable storage medium, or transmitted to a computer system via a computer-readable transmission medium. All or some of the computer programs may be provided on a computer-readable medium that is permanently, removably, or remotely coupled to an information processing system.

[0103] A computer process typically consists of 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 the execution of the process. An operating system (OS) is software that manages the sharing of computer resources and provides programmers with an interface to access 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 users and system programs.

[0104] A computer system may include at least one processing unit, associated memory, and a plurality of input / output (I / O) devices. When executing a computer program, the computer system processes information according to the computer program and generates resultant output information through the I / O devices.

[0105] In the above description, the present invention has been described in conjunction with specific embodiments of the present invention. However, it is apparent that various modifications and changes may be made to the present invention without departing from the scope of the present invention as described in the appended claims. Those skilled in the art will recognize that the boundaries between logic blocks are merely illustrative, and other embodiments may merge logic blocks or circuit elements, or perform alternative decompositions on the functions of various logic blocks or circuit elements. Therefore, it should be understood that the architecture described herein is merely exemplary, and that many other architectures capable of achieving the same functions may actually be implemented. Any arrangement of components capable of achieving the same functions is actually "associated" in order to achieve the desired functions. Therefore, any two components combined herein to achieve a particular function may be considered "associated" in order to achieve the desired functions, regardless of the architecture or intermediate components. Similarly, any two components so associated may also be considered "operably connected" or "operably coupled" in order to achieve the desired functions.

[0106] In addition, those skilled in the art will recognize that the boundaries between the operations described above are merely illustrative. Multiple operations can be combined into one operation, one operation can be distributed in more operations, and operations can be performed at least partially overlapping in time. In addition, other embodiments may include multiple instances of a specific operation, and the order of the operations can be changed in various other embodiments. However, other modifications, changes and substitutions may also be made. Therefore, the specification and the drawings should be regarded as having illustrative rather than restrictive meanings. Unless otherwise specified, terms such as "first" and "second" are arbitrarily distinguishing the elements described therein. Therefore, these terms are not necessarily intended to represent the time or other priority of these elements. The fact that certain measures are cited in mutually different claims does not mean that the combination of these measures cannot be used advantageously.

Claims

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

2. The method according to claim 1, wherein: The standardized cT1 image is a cT1 image corrected to an image acquired on a reference MRI scanner of a subject with normal iron levels.

3. The method according to claim 1 or claim 2, wherein: The simulated MR images are generated using data in a dictionary of synthetic MR signal relaxation curves.

4. The method according to claim 3, wherein: Data in the synthetic MR signal relaxation curve dictionary is selected to match at least one of a corrected water T1 map and a proton density fat fraction PDFF map of the acquired medical MR image.

5. A method according to any preceding claim, wherein: Each of the acquired medical MR images includes an image pair having the same inversion time.

6. The method according to claim 1, wherein: The inversion time of each successively acquired medical MR image is longer than the inversion time of the previously acquired medical MR image.

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

8. A method according to any preceding claim, wherein: The at least three medical MR images include eight MR images.

9. The method according to any one of claims 1 to 7, wherein: The at least three medical MR images are acquired such that there will be redundant MR images.

10. A method according to any preceding claim, wherein: Multiple acquired medical MR images are acquired using a single acquisition.

11. The method according to claim 10, wherein: The single acquisition includes at least one of the following: EPI or single fast echo.

12. A method according to any preceding claim, wherein: The field strength correction and iron correction used to generate the corrected water T1 map were performed using forward Bloch simulations.

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

14. The method according to claim 12 or claim 13, wherein: The iron correction uses at least one of a T2* map and a B0 field strength to correct for differences in iron concentration from normal levels.

15. The method according to any one of claims 12 to 14, wherein: Inputs to the forward Bloch simulation include one or more of a PDFF value, a T2* value, a water T1 value, and a pulse sequence of an MRI scanner used to acquire the medical MR image.

16. A method according to any preceding claim, wherein: The analysis result of the plurality of medical MR images is a composite image.

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

18. A method according to any preceding claim, wherein: The single inversion pulse is an adiabatic pulse.

19. The method according to any of the preceding claims, further comprising the step of acquiring a further medical MR image immediately before or after acquiring the plurality of medical MR images.

20. The method according to claim 19, wherein: The other medical MR image is a multi-echo spoiled gradient echo acquisition image.

21. A method according to any preceding claim, wherein: The original MRI images were acquired at 0.3-3.0T.

Images