Quantitative imaging of myocardium

Abstract

The present invention provides a method for non-invasive quantitative imaging of the heart, the method comprising acquiring an initial T1 map and a T2 map of the heart, and correcting the initial T1 map using the T2 map to obtain a corrected T1 map, where correcting the initial T1 map comprises subtracting weighted values ​​of the T2 map from values ​​in the initial T1 map and adding a constant. The present invention also provides further methods and devices for non-invasive quantitative imaging of the heart.

Description

Technical Field

[0001] The present invention relates to a method for non-invasive quantitative imaging of the myocardium and a method for visualizing T1 maps and T2 maps. The present invention also relates to a device for non-invasive imaging of the myocardium and a method and device for annotating medical data of the myocardium. Further, the method relates to a method and device for predicting the risk of cardiac events in patients suffering from myocardial inflammation.

[0002] The present invention also relates to a computer-readable storage medium storing program code, the program code comprising instructions for performing such a method.

Background Art

[0003] Heart failure is becoming increasingly widespread due to an aging population and the unexplained treatment of its predisposing conditions, resulting in a significant morbidity, mortality, and social burden. In current clinical practice, diagnostic and treatment steps are induced by the onset of symptoms of heart failure. Available treatment options for heart failure are applied to alleviate symptoms and prevent disease progression. In many patients, the time point of clinical diagnosis coincides with a fairly advanced stage of structural heart disease and significant structural heart damage. Underlying tissue characteristics are not used to guide individual therapies or to monitor the response to treatment and its success. The pathophysiological processes leading to end-stage myocardial damage involve adaptation and repair, namely myocardial inflammation and fibrosis, respectively, following a recurrent cycle of injury. These processes typically precede the onset of phenotypic and clinically overt heart disease over months to decades. They also represent a uniform response to various toxic triggers, regardless of the underlying disease (e.g., non-ischemic (cardiotoxic, autoimmune, genetic) or ischemic due to coronary vessel occlusion).

[0004] A deeper understanding of the pathophysiological processes and molecular mechanisms underlying the development of heart disease, also known as myocardial remodeling, which clinically determines its progression to symptomatic heart failure, is needed. In short, these involve a multitude of inflammatory cell and interstitial changes, including cardiomyocyte necrosis and apoptosis, inflammatory cell infiltration, myofibroblast activation, and extracellular matrix remodeling. Lymphocyte and mononuclear cell infiltration of the myocardium, increased expression of pro-inflammatory chemokines and cytokines, and the contribution of autoimmune inflammation accelerate myocardial remodeling. The inflammatory processes of myocardial remodeling are increasingly associated with non-infarcted myocardium in ischemic heart disease. These molecular histopathological insights have been derived primarily through studies in disease animal models, which have been partially validated in human myocardial biopsies and autopsies.

[0005] To date, there are no readily applicable diagnostic tests that enable direct detection and real-time monitoring of these processes in daily clinical practice. Recognition of cardiac impairment relies on abnormal myocardial structure (increased dimensions of the LV ventricle, increased wall thickness), impaired pumping function (reduced systolic function or ejection fraction), or decreased filling velocity as detected by echocardiography. Often, symptoms exist in the absence of structural impairment and with only preserved or mildly impaired function (heart failure with preserved or mildly impaired function). On the other hand, significant improvements in function may be measured in patients with severe impairment, while smaller changes are often undetectable anyway. The proportion of patients observed with improved heart failure function, indirectly involving elements of reversible myocardial remodeling, is small. If detected early and more directly, this can inform specific treatments, monitor the effectiveness of therapy, and individualize medical procedures. Myocardial biopsy is an invasive approach to histological examination of the myocardium, as it is performed using invasive cardiac catheterization, which relies on a skilled surgeon, and therefore should only be used in selected cases. Procedural complications are high, especially in critically ill patients. Furthermore, interpretation relies on a skilled and very high-quality histopathological reference laboratory. The yield of myocardial immunohistopathological examination is excellent for recognizing acute inflammatory processes and some infiltrative diseases, while processes related to myocardial remodeling cannot be easily distinguished from those involved in chronic, low-grade inflammation. Practical problems and limited diagnostic rates limit the clinical value of invasive myocardial histopathology, making widespread use and continuous application in patients with heart failure impossible. Magnetic resonance imaging (MRI) provides a non-invasive approach to assessing tissue characteristics and generating images using a strong magnetic field and, by its manipulation, magnetic field gradients and radio waves. Within the bounds of sufficient safe MRI practice, there are no known pathological consequences of single or repeated exposures. Myocardial magnetic relaxation mapping techniques use a magnetic field-prepared prepulse to measure the rate of magnetic field recovery to its original position after displacement. Two main mapping approaches, also referred to as T1 and T2 mapping, concern the longitudinal and transverse vector components of magnetic field relaxation.The recovery rate is directly related to the underlying tissue composition and its extent, which enables quantifiable tissue characterization of the myocardium. 1. Diagnosis of abnormal myocardium: Distinguishing abnormal tissue from normal tissue based on different range values ​​(normal reference range vs. abnormal tissue (above the mean of the normal range, ≥ 2 standard deviations (SD) distance)). 2. Differentiation between different types of conditions based on different underlying pathological ITEs, with different rates of magnetic relaxation (different values ​​for water, lipids, iron, fibrosis, and scarring), enabling non-invasive differentiation of different tissues. 3. Prognosis: Based on recognition of disease activity, disease stage, and association with cardiovascular outcomes.

[0006] Therefore, there is a need for tools to assist physicians in diagnosing and / or treating patients. [Overview of the project] [Means for solving the problem]

[0007] The object of the present invention is to provide a method for non-invasive quantitative imaging of the heart and a method for visualizing quantitative imaging of the heart, which overcomes one or more of the aforementioned problems of the prior art.

[0008] A first aspect of the present invention is to provide a method for non-invasive quantitative imaging of the heart, the method being To obtain the initial T1 and T2 maps of the heart, To obtain a corrected T1 map, the process includes correcting the initial T1 map using the T2 map, and correcting the initial T1 map involves subtracting the weighted values ​​of the T2 map from the values ​​in the initial T1 map and adding a constant.

[0009] In this specification, a T1 map may refer to a T1 value, for example, an image with a 2D image or 3D image volume, or it may refer to a set of one or more T1 values. Preferably, one or more T1 values ​​are stored together with their corresponding spatial coordinates. The same applies to the T2 map.

[0010] A T1 map can refer to the result of obtaining a T1 mapping. The same applies to T2 maps.

[0011] Obtaining the initial T1 and T2 maps of the heart may refer to obtaining the initial T1 mapping and T2 mapping of the heart.

[0012] Correcting the initial T1 map may refer to correcting the initial T1 mapping values, while using the T2 map may refer to using the T2 mapping values.

[0013] Acquiring a T1 map as described herein may mean, for example, obtaining a T1 map using an MR imaging device, or receiving a T1 map from another device, for example, from a network device where the T1 map is stored. In the same manner, a T2 map may be received from another device, for example, from a network device.

[0014] The first aspect of the method allows for obtaining a corrected T1 map, which provides improved diagnostic values ​​compared to the initial T1 map. As outlined above, obtaining a corrected T1 map may refer to obtaining one or more corrected T1 values.

[0015] In the first implementation of the method according to the first aspect, the weight of the weighted value is 10 to 15, preferably 11 to 14, particularly 12 to 13, and the constant is 350 to 500, preferably 400 to 450.

[0016] Further implementation of the method according to the first aspect further, This method uses a 2D color map to map values ​​from a corrected T1 map to values ​​from corresponding positions in a T2 map, wherein in the 2D color map, the lowest value in the first direction corresponds to the first color, the highest value in the first direction corresponds to the second color, the lowest value in the second direction corresponds to the third color, the highest value in the second direction corresponds to the fourth color, and the first to fourth colors are different colors. Outputting the output color on the screen, Includes.

[0017] In this specification, a 2D color map may refer to a mapping that maps two input values ​​to an output color.

[0018] In this implementation, please understand that the values ​​between the lowest and highest values ​​in the first direction and / or between the lowest and highest values ​​in the second direction are the colors assigned by the corresponding interpolated color values.

[0019] The experiment demonstrates that this type of color visualization enables more efficient and accurate diagnosis.

[0020] In a further implementation of the method according to the first aspect, the first color is black, the second color is yellow, the third color is blue, and the fourth color is red.

[0021] The experiment demonstrates that this color combination is particularly useful for physicians to quickly grasp the information content of color-coded images.

[0022] Further implementation of the method according to the first aspect involves initial T1 and T2 maps comprising cardiac regions, and the method further includes predicting the risk of cardiac events in patients with myocarditis, and the prediction is, This involves obtaining T1 values ​​based on the cardiac region in a corrected T1 map and T2 values ​​based on the patient's cardiac region in a T2 map, where the region is, in particular, a short-axis slice of the mid-ventricular region. Calculating a risk based on a weighted sum of T1 value, T2 value, and one or more additional factors; including.

[0023] When calculating the risk of a cardiac event based on the weighted sum of the T1 values, the T2 values and one or more additional factors give a more accurate prediction compared to calculating the risk based only on the unweighted T1 and T2 values, as shown by experiments.

[0024] In embodiments, the method does not directly predict the risk of a cardiac event, but provides data that enables a physician to predict the risk of a cardiac event. In those embodiments, the calculated risk is simply a score that enables a physician to predict the risk of a cardiac event (which may include other information not available for presentation to the physician).

[0025] In some embodiments, predicting the risk of a cardiac event for a patient with myocarditis includes myocarditis in the sense of chronic asymptomatic inflammation. In other words, in these embodiments, myocarditis is not limited to severe acute diseases (such as a football player who died on the pitch or ICU admission, etc.) or some mild ones that heal without problems. Instead, this may also include "chronic inflammatory cardiac remodeling".

[0026] In a further implementation of the method according to the first aspect, the method further comprises obtaining a radiological image of the heart, particularly a magnetic resonance image; determining the longitudinal axis of the heart based on the radiological image; determining at least one cardiological coordinate for at least one selected location within the image, the cardiological coordinate comprising a longitudinal coordinate indicating the projection of the location on the longitudinal axis; a circumferential coordinate indicating the circumferential position of the location around the longitudinal axis; a radial coordinate indicating the radial distance of the location from the longitudinal axis; and; This includes annotating at least one selected location using cardiac coordinates.

[0027] It is understood herein that acquiring radiographic images may refer to obtaining radiographic images using medical imaging devices. In this specification, radiographic images may specifically refer to MRI images.

[0028] This implementation allows physicians to annotate radiographic images using at least one cardiac coordinate system. This system is designed so that corresponding locations within the heart of different patients are assigned identical or similar coordinate values. This has the advantage that annotations from different physicians and different patients can be directly compared.

[0029] Further implementation of the method by the first aspect involves annotating at least one selected location. Output cardiac coordinates on the screen. Storing cardiac coordinates in a data storage device, and / or Visualizing one or more lines overlaid on a radiographic image, wherein one or more lines correspond to constant longitudinal, circumferential, and / or radial coordinates of cardiac coordinates. Includes.

[0030] This implementation has the advantage of making cardiac coordinates available to physicians.

[0031] A second aspect of the present invention provides a device for non-invasive quantitative imaging of the heart, the device configured to perform one of the methods or implementations of the first aspect. The device of the second aspect may include a computer and / or monitor. The device may be configured to connect to an image acquisition device, in particular a magnetic resonance imaging device. The connection may be implemented, for example, through a computer network.

[0032] A third aspect of the present invention provides a method for visualizing T1 maps and T2 maps, the method being: This method uses a 2D color map to map values ​​from the T1 map to values ​​from corresponding positions in the T2 map, wherein in the 2D color map, the lowest value in the first direction corresponds to the first color, the highest value in the first direction corresponds to the second color, the lowest value in the second direction corresponds to the third color, the highest value in the second direction corresponds to the fourth color, and the first to fourth colors are different colors. Outputting the output color on the screen, Includes.

[0033] It should be understood that interpolation is performed regarding the mapping of output colors between the lowest and highest values ​​in the first direction. In other words, the four colors can correspond to the extreme values ​​in the color map, and the positions between them are interpolated as appropriate.

[0034] The experiment demonstrates that this visualization enables efficient and accurate diagnosis of information, particularly from T1 and T2 maps.

[0035] In the first implementation of the third aspect method, the T1 map is a corrected T1 map, which is corrected based on the T2 map, specifically using the first aspect method or one of the implementations of the first aspect. Experiments have shown that visualization is particularly useful and accurate when it is based on a corrected T1 map.

[0036] Preferably, the first to fourth colors are selected as colors that can be easily distinguished by a physician.

[0037] In a further implementation of the method according to the third aspect, the first color is black, the second color is yellow, the third color is blue, and the fourth color is red.

[0038] A fourth aspect of the present invention provides a device for visualizing T1 maps and T2 maps, the device being configured to perform one of the methods or implementations of the third aspect.

[0039] A fifth aspect of the present invention provides a method for annotating cardiac medical data, the method being Acquiring cardiac radiographic images, particularly magnetic resonance imaging, Based on radiographic images, determine the vertical axis of the heart, To determine at least one cardiac coordinate for at least one selected location in the image, wherein the cardiac coordinate is: The vertical axis shows the projection of the location, using vertical coordinates, The circumferential coordinate system indicates the circumferential position of locations around the vertical axis, Radial coordinates, which show the radial distance from the vertical axis, To be equipped with, Annotate at least one selected location using cardiac coordinates, Includes.

[0040] This implementation allows physicians to annotate radiographic images using at least one cardiac coordinate system. This system is designed so that corresponding locations within the heart of different patients are assigned identical or similar coordinate values. This has the advantage that annotations from different physicians and different patients can be directly compared.

[0041] In the first implementation of the fifth aspect of the method, annotating at least one selected location is: Output cardiac coordinates on the screen. Storing cardiac coordinates in a data storage device, and / or, Visualizing one or more lines overlaid on a radiographic image, wherein one or more lines correspond to constant longitudinal, circumferential, and / or radial coordinates of cardiac coordinates. Includes.

[0042] This on-screen visualization of one or more lines corresponding to constant coordinate values ​​has the advantage of allowing physicians to quickly determine the position within the image corresponding to the same vertical, circumferential, and / or radial coordinates.

[0043] Other methods may also exist for how cardiac coordinates are stored, processed, or visualized for physicians. For example, cardiac coordinates may be transmitted via a computer network to further devices within the computer network so that these devices can perform processing based on the cardiac coordinates. In other words, the determined cardiac coordinates can be used not only for direct use by physicians but also for automated processing of annotations.

[0044] In a preferred embodiment, a patient image can be warped to a reference image, which corresponds to standardized coordinates. In other words, pixel or voxel values ​​from the patient image can be mapped such that equidistant points in the warped image correspond to equidistant points in the cardiovascular coordinate system. In the second implementation of the fifth aspect of the method, this method further states that

[0045] The vertical position is rounded to one of a predetermined number of possible vertical positions, preferably the number of possible vertical positions being 50 to 200, preferably 100. The circumferential coordinates are rounded to one of a predetermined number of possible circumferential positions, preferably the number of possible circumferential positions is 360, and / or

[0046] The radial coordinates are rounded to one of a predetermined number of possible radial positions, preferably the number of possible radial positions being 50 to 200, preferably 100. Includes.

[0047] The experiment shows that the surrounding environment leads to a shorter processing time and therefore easier memory of cardiovascular logical coordinates.

[0048] In a third implementation of the fifth aspect of the method, the method further includes rounding the longitudinal position, preferably the initial rounding being used for visualization, in particular for overlaying radiographic images and indicating rough rounded boundaries.

[0049] In the fourth implementation of the fifth aspect of the method, the cardiac coordinates further include temporal coordinates, which are determined based on the temporal distance to a predefined point in the cardiac cycle.

[0050] In a further implementation of the fifth aspect of the method, the predefined point is defined relative to the cardiac P-spike, preferably the beginning of the P-spike. Determining the temporal coordinates relative to the cardiac P-spike has the advantage that the temporal coordinates are relative to a position that can be reproducibly and accurately determined in the cardiac cycle.

[0051] According to the sixth aspect, embodiments of the present invention provide a device for non-invasive quantitative imaging of the heart, the device configured to perform one of the methods or implementations of the fifth aspect.

[0052] A seventh aspect of the present invention provides a method for predicting the risk of cardiac events in patients suffering from myocarditis, the method being Obtaining T1 and T2 values ​​for a region of the patient's heart, particularly a short-axis slice of the mid-ventricular region, The risk is calculated based on the weighted sum of the T1 and T2 values ​​and one or more additional coefficients. Includes.

[0053] In the first implementation of the seventh aspect of the method, the T1 value of a region is the average of the T1 map values ​​within the region, and / or the T2 value of a region is the average of the T2 map values ​​within the region.

[0054] In the second implementation of the seventh aspect of the method, the additional coefficients include one or more of the following: age, sex, hematocrit value, hs-CRP, hsTropT, LV-EF, RV-EF, and myocardial LGE.

[0055] In the third implementation of the method according to the seventh aspect, the coefficients are determined using multivariate Cox regression.

[0056] A further aspect of the present invention refers to a computer-readable storage medium for storing program code, the program code comprising instructions that, when executed by a processor, perform one of the methods of the first, third, fourth, fifth, sixth, and seventh aspects or implementations of the first, third, fourth, fifth, sixth, and seventh aspects. The present invention provides, for example, the following: (Item 1) A method for non-invasive quantitative imaging of the heart, wherein the method is To obtain the initial T1 and T2 maps of the aforementioned heart, To obtain a corrected T1 map, the initial T1 map is corrected using the T2 map. Includes, Correcting the initial T1 map is a method that includes subtracting the weighted value of the T2 map from the value in the initial T1 map and adding a constant. (Item 2) The method according to item 1, wherein the weight of the weighted value is 10 to 15, preferably 11 to 14, particularly 12 to 13, and the constant is 350 to 500, preferably 400 to 450. (Item 3) The method involves mapping values ​​from the correction T1 map and values ​​from corresponding positions in the T2 map to the output color using a 2D color map, wherein in the 2D color map, the lowest value in the first direction corresponds to the first color, the highest value in the first direction corresponds to the second color, the lowest value in the second direction corresponds to the third color, the highest value in the second direction corresponds to the fourth color, and the first to fourth colors are different colors. Outputting the aforementioned output color on the screen The method described in item 1 or 2, further including the method described in item 1 or 2. (Item 4) Acquiring radiographic images of the aforementioned heart, particularly magnetic resonance images, Based on the aforementioned radiographic image, the vertical axis of the heart is determined, Determining at least one cardiac coordinate with respect to at least one selected location in the aforementioned image, wherein the cardiac coordinate is: A vertical coordinate system showing the projection of the location onto the aforementioned vertical axis, A circumferential coordinate indicating the circumferential position of the location around the vertical axis, A radial coordinate indicating the radial distance from the vertical axis to the location, To be equipped with, Annotating the at least one selected location using the cardiovascular coordinates, wherein annotating the at least one selected location is preferably, Output the cardiac coordinates on the screen. Storing the cardiovascular coordinates in a data storage device, and / or, Visualizing one or more lines on the screen that are overlaid across the aforementioned radiographic image, wherein the one or more lines correspond to constant longitudinal coordinates, constant circumferential coordinates, and / or constant radial coordinates of the cardiac coordinates. including and The method described in one of the above items, further including: (Item 5) A method for visualizing T1 maps and T2 maps, wherein the method is The method involves mapping values ​​from the T1 map to corresponding values ​​from the T2 map to output colors using a 2D color map, wherein in the 2D color map, the lowest value in the first direction corresponds to the first color, the highest value in the first direction corresponds to the second color, the lowest value in the second direction corresponds to the third color, the highest value in the second direction corresponds to the fourth color, and the first to fourth colors are different colors. The output color is output on the screen, and preferably the T1 map is a corrected T1 map that is corrected based on the T2 map using the method described in one of items 1-4. Methods that include... (Item 6) A method for annotating cardiac medical data, wherein the method is Acquiring radiographic images of the aforementioned heart, particularly magnetic resonance images, Based on the aforementioned radiographic image, the vertical axis of the heart is determined, Determining at least one cardiac coordinate with respect to at least one selected location in the aforementioned image, wherein the cardiac coordinate is: A vertical coordinate system showing the projection of the location onto the aforementioned vertical axis, A circumferential coordinate indicating the circumferential position of the location around the vertical axis, A radial coordinate indicating the radial distance from the vertical axis to the location, To be equipped with, Annotating the at least one selected location using the aforementioned cardiac coordinates. Methods that include... (Item 7) Annotating at least one of the selected locations is, Output the cardiac coordinates on the screen. Storing the cardiovascular coordinates in a data storage device, and / or, Visualizing on the screen one or more lines overlaid across the aforementioned radiographic image, wherein the one or more lines correspond to a fixed longitudinal coordinate, a fixed circumferential coordinate, and / or a fixed radial coordinate of the cardiac coordinates. It includes, preferably, further, The vertical position is rounded to one of a predetermined number of possible vertical positions, wherein the predetermined number of possible vertical positions is 50 to 200, preferably 100. The circumferential coordinates are rounded to one of a predetermined number of possible circumferential positions, preferably the number of possible circumferential positions is 360, and / or The radial coordinates are rounded to one of a predetermined number of possible radial positions, preferably the number of possible radial positions being 50 to 200, preferably 100. The method described in item 6, including the method described in item 6. (Item 8) The method according to one of items 6 and 7, further comprising rounding the longitudinal position, preferably the initial rounding being used for visualization, in particular for overlaying the radiographic image and indicating the boundary of the rough rounding. (Item 9) The cardiac coordinates further comprise temporal coordinates, the temporal coordinates being determined based on the temporal distance to a predefined point in the cardiac cycle, preferably the predefined point being defined relative to a P-spike of the heart, preferably the beginning of the P-spike, according to one of items 6-8. (Item 10) A method for predicting the risk of cardiac events in patients suffering from myocarditis, wherein the method is: To obtain T1 and T2 values ​​of the cardiac region of the patient, in particular, a short-axis slice of the mid-ventricular region, The risk is calculated based on the weighted sum of the T1 and T2 values ​​and one or more additional coefficients. Methods that include... (Item 11) The method according to item 10, wherein the T1 value of the region is the average of the T1 map values ​​within the region, and / or the T2 value of the region is the average of the T2 map values ​​within the region. (Item 12) The method according to item 10 or 11, wherein the additional coefficients include one or more of age, sex, hematocrit, hs-CRP, hsTropT, LV-EF, RV-EF, and myocardial LGE, and / or the coefficients are determined using multivariate Cox regression. (Item 13) A device for non-invasive quantitative imaging of the heart, wherein the device is configured to perform the method described in one of items 1-12. (Item 14) A computer-readable storage medium for storing program code, wherein the program code comprises instructions, and when executed by a processor, the instructions cause the processor to perform the method described in one of items 1-12. (Item 15) The initial T1 map and the T2 map comprise the region of the heart, and the method further includes predicting the risk of cardiac events in a patient suffering from myocarditis, and the prediction is The process involves obtaining T1 values ​​based on the cardiac region in the corrected T1 map, and obtaining T2 values ​​based on the patient's cardiac region in the T2 map, wherein the region is a single ventricular mid-short-axis slice. The risk is calculated based on the weighted sum of the T1 value, the T2 value, and one or more additional coefficients. The method described in one of items 1-4, including the method described in item 1-4. [Brief explanation of the drawing]

[0057] To more clearly illustrate the technical features of embodiments of the present invention, supplementary drawings provided to illustrate the embodiments are briefly introduced below. The supplementary drawings in the following description are merely some embodiments of the present invention, and modifications to these embodiments are possible without departing from the scope of the invention as defined in the claims.

[0058] [Figure 1] Figure 1 shows a plot of T1 vs. T2 values ​​for MRI images of patients with various cardiac diseases but no suspected inflammation. There is no systematic effect of T2 on T1.

[0059] [Figure 2] Figure 2 shows a plot of native T1 vs. native T2 values ​​for patients with suspected myocardial inflammation. A significant effect of T2 on T1 is observed.

[0060] [Figure 3] Figure 3 shows a corrected plot of native T1 and native T2 values ​​for the same patient as in Figure 2, according to one embodiment of the present invention. The effect that T2 has on T1 is no longer present.

[0061] [Figure 4] Figure 4 shows the predicted risk of heart failure hospitalization for cardiovascular death in patients with suspected myocarditis, based on a myocarditis risk score using unadjusted T1 values. For each score value, the number of patients is shown by combined gray and orange boxes, the number of observed events is shown by orange boxes, and the predicted probability of an event is shown by a black line (with confidence intervals as dark gray and dotted red lines).

[0062] [Figure 5] Figure 5 shows the improved risk score using the corrected T1 value (T1c), as shown in Figure 4.

[0063] [Figure 6] Figure 6 shows the correlation between native T1 before and after therapy in patients with sarcoid diseases, i.e., diseases involving severe myocardial inflammation. A significant change in native T1 is observed with therapies that reduce inflammation / myocardial water content.

[0064] [Figure 7] Figure 7 shows the same patient using corrected T1c values. No change in T1c is observed here, demonstrating that sensitivity to inflammation / myocardial water content is eliminated, due to the therapy.

[0065] [Figure 8] Figure 8 shows a standardized color overlay for depicting T1 (upper panel) representing fibrosis and T2 (lower panel) representing inflammation.

[0066] [Figure 9] Figure 9 shows the same patient with combined color overlays for fibrosis and inflammation, enabling immediate visual classification of the underlying disease.

[0067] [Figure 10]Figure 10 shows an illustrative T1 map (top left panel) and a local zoom view (orange box, top right panel) demonstrating the gray values ​​of the original voxels.

[0068] [Figure 11] Figure 11 shows a simplified concept of the coordinate system along the x and y axes.

[0069] [Figure 12] Figure 12 illustrates the concept of standardized color schemes.

[0070] [Figure 13] Figure 13 shows the basis for the concept of coordinate systems in the z-axis and trans-wall directions.

[0071] [Figure 14A] Figure 14A shows the myocarditis score relating to the probability of death or significant shock in patients with suspected myocarditis.

[0072] [Figure 14B] Figure 14B shows the myocarditis score relating to the probability of hospitalization for heart failure leading to death from heart failure in patients with suspected myocarditis.

[0073] [Figure 15A] Figures 15A, 15B, and 15C show the Kaplan-Meier curves for native T2 and hs-TropT (terniles) and LGE (binary) for MACE. [Figure 15B] Figures 15A, 15B, and 15C show the Kaplan-Meier curves for native T2 and hs-TropT (terniles) and LGE (binary) for MACE. [Figure 15C] Figures 15A, 15B, and 15C show the Kaplan-Meier curves for native T2 and hs-TropT (terniles) and LGE (binary) for MACE.

[0074] [Figure 16]Figure 16 shows the correlation between the most frequent echocardiographic measurement (E / e') and invasively measured left ventricular stiffness (TAU).

[0075] [Figure 17] Figure 17 shows the correlation between the novel CMR score and invasively measured left ventricular stiffness (TAU).

[0076] [Figure 18] Figure 18 shows a box plot of age-sex-risk factor-corresponding control, i.e., between asymptomatic post-COVID patients and patients with “long-term COVID”. [Modes for carrying out the invention]

[0077] Detailed description of the embodiment Several presented embodiments relate to the ability of native myocardial T1 and T2 mapping (without contrast agent administration within at least one week) and post-contrast T1 mapping (within 20 minutes after administration of a gadolinium-based contrast agent, by measuring either post-contrast T1 or extracellular volume fraction (ECV)) in detecting myocardial pathological remodeling processes that lead to the development of heart failure symptoms. Numerous studies have correlated myocardial mapping measurements of longitudinal relaxation in the myocardium with histologically determined fibrotic infiltration (amyloidosis), histologically desiccated edema, myocardial iron content, and validated findings in model diseases (inflammatory, hereditary, invasive, valvular, and ischemic heart disease).

[0078] T1 and T2 mapping indices are elevated in the presence of cardiomyopathy compared to controls. The T1 mapping index has been shown to be associated with prognosis in patients with non-ischemic cardiomyopathy, ischemic heart disease, and light chain amyloidosis. Notably, these data were obtained in model diseases to provide proof of concept regarding mapping measures and to demonstrate the difference between health and overt disease. However, previous studies have not published the use of these tools in non-selective populations as screening tools for detecting abnormal cardiomyopathy as a pre-stage of cardiomyopathy and heart failure.

[0079] Experiments have shown that by establishing sequence-specific normal ranges (mean ± 2 standard deviations (SD)) in subjects with unknown or unestablished cardiac diseases, including cardiac markers, without any medication or normal blood tests, it is possible to separate pathological remodeling from healthy myocardium and stratify patients in terms of prognosis. In largely unpublished data, it has been further demonstrated that T1 and T2 mapping indices are directly associated with myocardial injury (by high-sensitivity troponin T), increased wall stress (by NT-proBNP), inflammatory processes, and the degree of prognosis within disease subcategories. Furthermore, it has been demonstrated that combined readouts of T1 and T2 values ​​enable decoding of the underlying tissue matrix and enable differentiation of disease stages and state subcategories. Finally, our data reveal that inducing disease based on T1 and T2 mapping readouts, compared to symptoms alone, provides a more sensitive measure of myocardial recovery. Personalized insights based on underlying histopathology and prognostic risk provide a beneficial approach to treating predisposing conditions for heart failure.

[0080] Therefore, this concept aims to provide means and method systems and methods for the application of non-invasive myocardial histopathophysiology using cardiovascular magnetic resonance imaging for the diagnosis, prognosis, and therapeutic guidance of predisposing conditions leading to heart failure. A further objective is to provide means and methods for risk stratification of patients with predisposing conditions leading to heart failure. It also aims to support means and method systems for providing personalized insights based on underlying histopathology and prognostic risk, which will guide beneficial approaches to the treatment of predisposing conditions related to heart failure.

[0081] The objective can be achieved by using non-invasive imaging means and systems, employing cardiac T1 and T2 mapping with MRI, for the diagnosis, prognosis, and therapy of abnormal myocardial remodeling and heart failure.

[0082] The methods presented include the following: 1. In common, a. Use of imaging devices that support cardiac magnetic resonance imaging approved for clinical use. b. The use of T1 and T2 mapping sequences, defined by the ability to provide several images of the myocardium, using inversion or saturated prepulses, or a combination or modification thereof, with proper magnetization preparation and using any range of flip angles and any number of image / prepulse acquisition schemes. c. Use of any approach to generate an exponentially fitting curve of signal intensity based on a series of images, d. Readout values ​​for T1 and T2 mappings, including the time during 63% to 37% relaxation, respectively. e. Determining the T1 value of the myocardium from one or more images. This can be obtained, for example, using myocardial T1 and T2 mapping, which can be achieved using several technically different approaches. 2. Obtain myocardial delayed gadolinium enhancement, post-contrast T1 mapping, or equivalent imaging for localized replacement fibrosis. 3. Interpretation of values, which are sequences and conditions specific to the purpose of screening, diagnosis, prognosis, and induction of therapy. This objective is addressed by providing a screening tool for the presence of abnormal myocardium as a pre-stage of myocardial remodeling leading to heart failure. Detection of abnormal tissue characteristics by cardiovascular magnetic resonance imaging represents clinically abnormal outcomes. A person is classified as a "patient" when the person has an abnormal native T1. This concept is very new as it provides a new definition of abnormality.

[0083] In embodiments of the present invention, the term “heart failure” is used not only for delayed manifest cases but also for early signs / symptoms of heart failure. This includes “heart failure with preserved ejection fraction” and below-normal cardiac function. In particular, the term “heart failure” may include “heart failure with preserved ejection fraction (HFpEF)” and / or “heart failure with mildly reduced ejection fraction (HFmrEF).”

[0084] In embodiments, “heart failure with preserved ejection fraction” may also include “diastolic dysfunction.” An early sign of HfpEF is diastolic dysfunction, measured by echocardiography as E / e'. CMR may be particularly suitable for diagnosing HFpEF. The literature has so far described extracellular volume fraction (ECV), but not native T1 or T2. Our experiments show that a score developed from a combination of blood markers (BNP) and CMR parameters correlates much more strongly with diastolic function, which is determined more invasively than echocardiography.

[0085] This objective is further addressed by providing a sequence-specific normal reference range (defined as the mean of the normal range, SD of the normal range, obtained from a minimum of 200 healthy subjects, with no known medical history, no cardiac or systemic disease, normal cardiac function and structure on the CMR, no LGE, routine blood tests, and no cardiac serology or regular medication). The characteristics of the normal distribution are applied to separate normal from abnormal myocardium based on a distance of 2SD (or Z-score).

[0086] This objective is further addressed by providing a risk stratification scheme that is proportionally associated with disease severity and prognosis deterioration. Due to the linearity of the relationship between cardiomyopathy severity and quantitative histometric measurements, the standard deviation (SD) of the normal range (or z-score, quartiles, ternary quartiles, or similar statistical approaches) can be applied to estimate predictive associations with outcomes and derive risk scores.

[0087] This objective is further addressed by providing a diagnostic algorithm that is specific to the sequence and cardiac state. This is based, firstly, on the observation that different cardiac states exhibit different diagnostic patterns of T1 and T2 mapping values ​​based on prominent underlying ITE determinants. Secondly, different sequences have different ranges and limitations of analytical detection. Representative tissue signatures and their interpretations are presented in the table below. [Table 1]

[0088] This objective is further addressed by providing a risk stratification algorithm that is specific to the sequence and cardiac state. Default risk stratification, based on ischemic or non-ischemic categorization, shall be used when the specific underlying disease cannot be determined by clinical means. Thirdly, different sequences exhibit different prognostic associations in different cardiac states. This is translated into a risk stratification scheme based on cutoffs that are specific to the sequence and cardiac state.

[0089] This objective is further addressed by providing a therapeutic guidance scheme that is specific to the sequence and cardiac state. These are, firstly, guided by widely recognized ITE determinants of the diagnostic pattern of T1 and T2 mapping values. Secondly, they are guided by further disease severity, worsening of the estimated prognosis, and evidence of reversibility. Thirdly, they are adapted by repeating measurements that determine the effectiveness and continued need for a given therapy. (T2 correction of T1 map for cardiac imaging)

[0090] As can be seen from the above, the T1 map of the myocardium is an excellent parameter for detecting cardiac abnormalities and predicting prognosis. However, to the extent that they are more sensitive and reproducible, they are affected by the T2 effect. This invention makes it possible to correct the T1 map for the T2 effect and obtain three distinct values ​​(T2-influenced T1, T2, and corrected T1).

[0091] Based on large-scale T1 and T2 data and data acquired before and after dialysis, the inventors can model T2-corrected T1 values ​​using specific imaging sequences adopted and validated in their laboratory. This novel parameter, namely T2-corrected T1, is added to medical equipment for myocardial mapping by retaining the advantages of a highly sensitive and robust T1 map, resulting in T2-influenced T1 values, while enabling the calculation of true T1 and true T2.

[0092] T2 correction is a formula used during post-processing developed by the inventors from large databases.

[0093] The T2-corrected T1 value was similar to that before and after dialysis (it alters the amount of water in the myocardium, and therefore modifies T2 and T2-influenced T1, but not T2-corrected T1).

[0094] To date, T1 mapping, T2 sensitivity of T1 mapping, and T2 mapping have been described. T1 and T2 are understood to be highly diagnostic and prognostic. Many researchers have attempted to "optimize" T1 mapping to reduce its sensitivity to T2, however, this results in less robust data and lower diagnostic accuracy.

[0095] The primary advantage is that T2-corrected T1 is added to medical equipment for myocardial mapping by retaining the benefits of a highly sensitive and robust T1 map, which yields T2-influenced T1 values, while allowing for the calculation of true T1 and true T2.

[0096] Further advantages of T2 correction T1 (also known as T1c) include the following: This maintains excellent robustness and reproducibility, as seen in the deformation of T2-sensitive MOLLI (most notably the FFM-MOLLI3(2)3(2) 550° flip angle). This allows for the separation of fibrosis and water components in T1 measurement. This should provide a prognostic value in addition to uncorrected T1, especially in inflammatory diseases where T2-sensitive T1 is influenced by the T2 effect (water). This should remain unchanged when the inflammation is treated.

[0097] The inventors used four different groups of patients. A: A patient with various heart conditions but no suspected inflammation. B: Patients with suspected myocarditis C: Patients with sarcoid disease before or after therapy D: Patients with chronic kidney disease before or after dialysis.

[0098] In group A, the inventors assessed the minimal T2 effect and the normal correlation between T1 and T2. As can be seen from Figure 1, there is no systematic effect that T2 has on T1.

[0099] In group B, the inventors determined the correlation between T1 and T2 and developed a correction factor to achieve a similarity relationship between T1c and T2, as observed between T1 and T2 in patients without inflammation. As visible in Figure 2, there is a clear dependency that native T1 imposes on native T2 in patients with inflammation.

[0100] The experiment found that the correction was applied using a preferred constant, and the preferred correction formula was of the following form. T1c = T1 - a × T2 + b This indicates that it can be obtained by using [this method].

[0101] Preferably, the constant a is in the range of 10 to 15, particularly in the range of 11 to 14, for example, in the range of 12 to 13.

[0102] Preferably, the constant b is in the range of 300 to 500, particularly in the range of 375 to 450, for example, in the range of 410 to 440.

[0103] In a preferred embodiment, the following formula is used. T1c = T1 - (T2 - 34) × 12, 5

[0104] Applying one of the above formulas leads to a corrected T1 map, called T1c, which has values ​​that are essentially free from T2 influence.

[0105] Native T1 T2 correction works with all T2-sensitive MOLLI deformations.

[0106] The validation was specifically performed on a unique T1 mapping sequence (FFM-MOLLI3(2)3(2)550°) developed and used in the inventors' laboratory, which yields the best diagnostic discrimination, superior reproducibility, and strongest prognostic ability compared to other sequences. Different formulas need to be used for other sequence combinations.

[0107] Native T1, native T2, and T1c, combined with other clinical or imaging measurements, can be used to predict outcomes in various patient groups by constructing quantitative scores. Illustrative applications include myocarditis (data as shown above), non-ischemic and ischemic cardiomyopathy, risk assessment in athletes, risk assessment after COVID-19 infection, and risk assessment in young adults (cardiomyopathy screening).

[0108] As shown in Figure 3, after correction, the effect of T2 on T1c is eliminated.

[0109] To validate the corrected T1 map, T1c was then added to the model to predict outcomes in patients with suspected myocarditis, and it was found to improve predictions compared to the current model.

[0110] In particular, Figure 4 shows the results of a novel model for predicting risk in myocarditis without T1c, and Figure 5 shows the results for the improved model. Note the higher percentage of events in the high-risk group compared to the model without T1c.

[0111] To further verify the accuracy of the correction, the inventors first compared native T1 before and after therapy in patients with sarcoid diseases, i.e., diseases involving severe myocardial inflammation. The inventors then calculated T1c and performed similarity comparisons.

[0112] Specifically, Figure 5 illustrates how T1 significantly changed with therapy (due to the change in T2). In contrast, Figure 6 illustrates how T1c did not change with therapy (a slope of 1.00 and a linear correlation with R2 of 0.9237).

[0113] The experiment confirmed that the T2 correction for native T1 works with all T2-sensitive MOLLI deformities.

[0114] The validation was specifically performed on a unique T1 mapping sequence (FFM-MOLLI3(2)3(2)550°) developed and used in the inventors' laboratory, which yields the best diagnostic discrimination, superior reproducibility, and strongest prognostic ability compared to other sequences. Different formulas need to be used for other sequence combinations.

[0115] Native T1, native T2, and T1c, combined with other clinical or imaging measurements, can be used to predict outcomes in various patient groups by constructing quantitative scores. Illustrative applications include myocarditis (data as shown above), non-ischemic and ischemic cardiomyopathy, risk assessment in athletes, risk assessment after COVID-19 infection, and risk assessment in young adults (cardiomyopathy screening).

[0116] Further embodiments of the present invention relate to visualization using a novel multidimensional color map.

[0117] Conventional visualization techniques for mapping values ​​are performed on several images, each displaying different color schemes for different data such as T1, T2, or extracellular volume fraction (ECV). The color schemes vary between different centers and are not scaled by absolute values ​​or mean ± standard deviation or other statistically significant values. As a result, the image colors are meaningless and can only be interpreted by the attached scale. Furthermore, due to the nature of all currently used color schemes, the information is limited to a single value; i.e., combinations of two data pairs are not considered possible. A novel color scheme has been developed to display T1 and T2 data in a single image, enabling standardized display. Each color represents a combination of two values ​​(provided as mean ± standard deviation).

[0118] The systematic use of color and the integration of two measurements into a single scheme intuitively highlights abnormalities (e.g., red, yellow), characterizes diseases (red = fibrosis, yellow = edema), adds quantitative information (e.g., >2 or >standard deviation), and combines these to enable a rapid intuitive understanding of diffuse fibrosis and focal edema.

[0119] Figure 8 illustrates a novel visualization scheme. In particular, Figure 8 shows a standardized color overlay for depicting T1 (upper panel) representing fibrosis and T2 (lower panel) representing inflammation. The use of standardized colors allows for immediate visual classification into normal, mild, and severe abnormalities.

[0120] Figure 9 shows the same patient with combined color overlays for fibrosis and inflammation, enabling immediate visual classification of the underlying disease.

[0121] Figure 10 shows an exemplary T1 map (top left panel) and a local magnified view (orange box, top right panel) demonstrating the gray values ​​of the original voxels. For demonstration purposes, this is further magnified (bottom left panel). For demonstration purposes, each voxel contains a T1 value, and some gray values ​​with similar grays are labeled using arbitrary units. This information is usually lost when processing images.

[0122] Over time periods within the cardiac cycle, these voxels will move to different locations within the image (standard coordinate space) due to cardiac motion, representing different relative volumes of the heart due to thickening / thinning. Consequently, overlay / partitioning deformations are required to "find" the "identical" voxels again on different images, and data is lost during documentation because standards are unavailable and the position relative to the heart is not documented.

[0123] Figure 11 shows a simplified concept of the coordinate system on the x and y axes. The image represents a typical example of a myocardial scar (white, highlighted in orange using a standardized color overlay in the upper right panel, within the black myocardium in the upper left panel). Using a standard compartmentalization system with six compartments within this representative imaging plane results in a loss of information regarding the exact extent and location of the scar. By extending to 360-degree circumferential radial rods, the circumferential extent can be well documented (only a subset of the radial rods, shown in the right center panel). The lower panel shows a distinction to two layers from the inside to the outside of the heart, as used in clinical standards. A distinction to 100 layers (not shown) would allow for better distinction to the transmural extent of the abnormality. Additional extensions of the coordinate system would be in the longitudinal direction of the heart (100 slices, see Figure 13) and the temporal position of the cardiac cycle (100 temporal positions, see Figure 13).

[0124] This information can be kept completely confidential and stored in a database without human intervention, and can be compared between different patients or at different points in time.

[0125] Figure 12 illustrates a concept for a standardized color scheme to enable rapid and intuitive presentation of anomalies in two combined imaging markers. One imaging marker (e.g., T1) is depicted with increasing red for higher values ​​and increasing blue for lower values, where gray represents the normal range, while the second imaging marker (e.g., T2) is depicted with increasing yellow for higher values ​​and increasing black for lower values, where gray represents the normal range. Each combination of values ​​(e.g., T1 and T2) can be assigned a standardized color for the severity of the anomaly (e.g., above or below normal, >2 or >5 standard deviations). Thus, the resulting colors are no longer random but represent clearly defined anomalies. The scheme can be made persistent if the original numerical values ​​are required rather than the classification and standard deviation of the anomalies (upper right panel).

[0126] In the conventional 17-compartment model, all information about the precise transmural location within the myocardium (layer) is lost. Regarding circumferential location within the myocardium, the precise location and extent cannot be explained, and it is unclear when a compartment is positive; instead of 350 locations, there are 6 locations. Similarly, for longitudinal explanations, there are 3 levels (instead of 100) with identical results. This means that using the conventional model, it is possible to preserve only those compartments that are "unaffected," "partially affected," "strongly affected," and "fully affected." This is no longer sufficient for modern machine learning algorithms.

[0127] Figure 13 shows the basis for the concept of a coordinate system in the transmural direction, with respect to the z-axis. The z-axis (a longitudinal axis defined by the line from the apex to the midpoint of the mitral valve) is divided into 100 elements, starting from the apex. Thus, every longitudinal position of the heart is precisely defined. The circumferential position is then bounded as 360-degree equiangular radial rods, starting from the insertion point of the right ventricle (red dot). The radial rods are demonstrated in Figure 11. The transmural position is described as 100 layers (orange arrows) from the inside to the outside of the myocardium. The temporal position is provided as the 100th of the cardiac cycle, starting from the beginning of the p wave. In other embodiments, the peak of the R wave can be used as a reference starting point.

[0128] (Improved model for predicting the risk of myocarditis - Myocarditis Risk Score) Further embodiments provide a risk assessment tool for prognosis in patients with suspected myocarditis. Risk assessment in patients with suspected myocarditis can be based on symptoms, blood markers, and magnetic resonance imaging. Using outcome data from patients with suspected myocarditis, the inventors can construct a prognosis model. This model includes a large amount of data (blood parameters, imaging, symptoms, etc.) but functions almost the same as a small number of definitive parameters. Based on the reduction of these parameters (blood tests, magnetic resonance imaging), the risk of heart failure or death is predicted. This is a first score for assessing the risk in patients with clinically suspected myocarditis. Previous recommendations do not rely on prognosis data and require invasive biopsy.

[0129] To date, no scores for clinical use exist. Guidelines have recommended biopsy for diagnosis. While there are recommendations for diagnosis using MRI, these are not based on outcome data and do not enable risk assessment.

[0130] Myocarditis is an inflammatory disease of the myocardium associated with adverse cardiovascular outcomes, including non-ischemic dilated cardiomyopathy (DCM), heart failure (HF), and sudden cardiac death (SCD). A modern approach to clinical recognition requires a two-step process, involving establishing the pre-test likelihood of myocarditis, followed by formal confirmation using invasive endocardial biopsy (EMB). In practice, the invasive step is frequently omitted for several reasons, depending on EMB expertise and the availability of a skilled pathologist referral center. The need for the invasive step is often questioned and debated, as the majority of patients present with mild symptoms, no structural abnormalities, or only non-serious structural abnormalities. Clinical trials of EMB-inducing therapy have been neutral in terms of therapeutic benefit. Social guidelines are divided regarding the use of EMB; the European Society of Cardiology (ESC) considers EMB a definitive diagnostic tool in all patients with a high pre-test likelihood of myocarditis, while the American College of Cardiology (ACC) predicts its use in selected patients with HF and rapidly progressing clinical exacerbations of unclear etiology.

[0131] To date, prognostic markers in myocarditis include positive EMB-based immunohistochemical criteria (IHC), NYHA functional classification, beta-blocker therapy, male gender, and the presence of delayed gadolinium enhancement (LGE) using cardiovascular magnetic resonance imaging (CMR). However, none of these measurements support clinical decision-making. Anti-remodeling therapy is initiated based on HF symptoms and reduced left ventricular ejection fraction (LVEF). The prognostic relationships of emerging quantitative non-invasive tissue markers by T1 and T2 mapping have not been systematically assessed. The goals of this study were to examine these candidate variables as predictors of outcome and to develop and validate models for individualizing risk prediction in patients with myocarditis.

[0132] (method) The prognostic model was derived from prospective longitudinal multicenter cohort studies (NCT03749343, NCT02407197). First, predictive associations of candidate variables were examined to generate the required number of associated predictors for enrollment in the prognostic model. The pre-specified criterion was a 15% significance level for univariate prognostic association with the endpoint. Next, the prognostic model was developed using a Cox proportional risk model to provide individualized risk estimates for the probability of adverse events over a mean of two years. The model was validated internally using 10 cross-validations and externally by training the model for two centers and validating it for a third center. Finally, a clinical risk score to predict the likelihood of events at two years was developed based on the development of the model. All procedures were carried out in accordance with the Declaration of Helsinki (2013). The study protocol was reviewed and approved by the Ethics Committee, and written informed consent was obtained from all participants.

[0133] (Research population and participating centers) The study cohort consisted of sequentially evaluated patients pre-recruited from three European university hospitals (University Hospital Frankfurt, Kerckhoff Clinic Bad Nauheim (both in Germany), and Guy's and St Thomas's Hospital (London, United Kingdom)) between October 2011 and December 2019. Some of the patients enrolled in this dataset are included in previous publications. Only adult patients (≥18 years old) without a prior major adverse cardiovascular event (MACE) or prior HF hospitalization were studied. The patient inclusion criteria were predefined criteria for clinically suspected myocarditis consistent with the current ESC opinion, including ≥1 clinical symptom and ≥1 diagnostic criterion (or ≥2 diagnostic criteria from different categories for asymptomatic patients). Exclusion criteria, in short, included contraindications to contrast-enhanced CMR, including pre-existing cardiac conditions, heart transplantation, pregnancy, or inability to provide informed consent. This status was determined independently of the research team.

[0134] (Patient assessment and data collection) Clinical demographics, medication, blood tests, EMB results, CMR measurements, and follow-up data were recorded at baseline and during periodic clinical follow-up using the REDCap electronic data capture tool. Blood tests were included if obtained within 7 days of clinical diagnosis.

[0135] EMB was routinely employed in the diagnostic pathway by several physicians (Frankfurt and Bad Nauheim, Germany). Detailed procedural information is included in the supplementary materials. EMB analysis was performed by the surgeon's choice of authorized EMB referral center (details in the supplementary materials). Predefined EMB-based variables included the Dallas criteria (DC), immunohistochemical criteria (IHC), and the presence of viral genomes.

[0136] CMR was performed on three Tesla clinical scanning systems (Frankfurt, Bad Nauheim: Skyra, Siemens Healthineers, Erlangen, Germany; London: Philips Achieva, Eindhoven, The Netherlands) using standardized protocols and unified sequence parameters in the imaging sites involved. Details of the imaging parameters are included in the supplementary materials. Myocardial T1 and T2 mapping were obtained in a single mid-ventricular short-axis slice using a validated variation of the modified Look-Locker imaging sequence (GoetheCVI®-MOLLI), while for T2 mapping, validated sequences for measuring myocardial edema, T2-FLASH, or T2-GraSE were used on Siemens or Philips scanning systems, respectively. Post-contrast-enhanced T1 mapping was not routinely performed. Analysis and validation data regarding the mapping sequences are summarized elsewhere. LGE imaging was performed within approximately 10 minutes after administration of gadobutrol (Gadovist®, Bayer AG, Leverkusen, Germany) at a dose of 0.1 mmol / kg body weight. Interpretation and post-processing of mapping data were performed by core laboratory staff on an anonymized dataset, free from underlying clinical information, following standardized operating procedures. CMR-based variables are included in Table 1S.

[0137] (Clinical Outcomes) Patient follow-up was conducted either between routine clinical assessments or by telephone every 12 months. The cause of death was determined by reviewing patient records, death certificates, autopsy reports, or interviews with witnesses, and event adjudication was performed by a skilled cardiologist independently of the research team. The primary outcome endpoint was MACE, consisting of cardiovascular mortality, sudden cardiac death (SCD), and appropriate intravascular defibrillator (ICD) discharge. The secondary endpoint consisted of a composite of deaths attributable to HF or HF hospitalization (HF endpoint). The first single event per patient from the date of inclusion was included in the analysis. Event definitions are provided in the supplementary materials.

[0138] (Variable selection) The candidate variables tested were listed in Table 1S (Supplementary Materials) and included cardiac risk factors, blood markers (hematocrit, highly sensitive (hs) C-reactive protein (hs-CRP), and hs-troponin (hs-TropT), estimated glomerular filtration rate), EMB (IHC, DC, presence of virus), and CMR parameters (LVEF, RVEF, LV mass index, native T1 and T2). Univariate Cox regression analysis was performed for each candidate variable to test the assumption of linearity with outcomes in the examination dataset. The proportional risk assumption required by the Cox model was investigated using Schoenfeld residuals. The pre-specified criterion for enrollment in the prognostic model was a 15% significance level of univariate prognostic association with the endpoint.

[0139] (Model and risk score development) The final risk model was developed using the entire dataset (multivariate Cox regression, stepwise, retrospective Wald). The multivariate Cox regression model was fitted with predictors. Variables with abnormal distributions were logarithmically transformed prior to enrollment in the regression model. The model was developed without centers as predictors to allow the model to be used in other cohorts of myocarditis. Clinical risk scores based on the model allow for the estimation of individual probabilities of adverse events over two years. External validation against a third center by deuteronomy and training it on two centers also allows for the determination of healthcare system effects, i.e., the reported performance index includes calibration slope, intercept, and C-index. Sensitivity analyses regarding center / healthcare system effects, representing external validation, were performed as described above.

[0140] (Sample size) To ensure that the model's regression coefficients were estimated with adequate accuracy, a minimum of 10 equivalent events were required per variable in the final model. The 64 MACE and 129 HF equivalent endpoint events observed across the entire cohort allowed for the estimation of up to 6 and 12 regression coefficients, respectively, and sensitivity analysis was performed.

[0141] (Comparison with the standard of practice) HF guideline-oriented anti-remodeling therapy in patients with reduced LVEF was adopted as a standard for comparison in modern practice. Time-decomposed clinical risk scores based on LVEF ≤ 45% for MACE and HF events were developed, respectively.

[0142] (statistical analysis) Statistical analysis was performed using SPSS software (SPSS Inc. (Chicago, IL, USA), version 25.0) and RStudio version 1.2.5001 (RStudio Inc., packages "rms", "survival", "survminer", "caret", "nricens", "mice"). Variables were expressed as mean + standard deviation (SD), median (interquartile range, IQR), or total (percentage), as appropriate. Time to event analysis was performed using univariate and multivariate Cox proportional risk models from the date of consent to the date of reaching the study endpoint or the date of the most recent assessment. Kaplan-Meier graphs were used to present the time to event relationships for categorical values. Missing data were handled using standard imputation methods (predictive mean match) and checked for any biases suffered. Patients with >25% missing predictors were excluded from the final model. All tests were two-sided, and a p-value of <0.05 was considered statistically significant.

[0143] result (Baseline characteristics) The final study cohort consisted of 722 patients meeting the criteria for clinically suspected myocarditis. Patients were similar in terms of clinical symptoms, cardiovascular risk factors, and medication. Patients with events had significantly higher hsCRP, hs-TropT, native T1 and T2, and mildly reduced LVEF. The presence of LGE was more common in patients with events (38% vs. 64%, p<0.01). Hs-TropT significantly exceeded the upper reference limit (URL) in 144 patients (20%), and 87 of these also had evidence of LGE.

[0144] (Endpoint events during follow-up) During an average follow-up cycle of 19 (15–23) months, 64 patients (9%) experienced MACE, which included 46 HF deaths, 2 acute coronary syndromes, 13 appropriate ICD discharges, 2 SCDs, and 1 stroke. HF endpoints consisted of 30 HF deaths and 99 HF hospitalizations. A total of 12 (26%) MACE events and 29 (23%) HF events occurred within 6 months from baseline. During the observation cycle, 116 (16%) patients were treated with ICDs, while 2 received dual-ventricular pacemakers.

[0145] (Missing data) 116 patients (16%) had at least one missing predictor. No patients needed to be excluded from the analysis due to missing data. Missing data were most common with respect to NYHA classification (n=109, 15%) and blood tests (hs-CRP n=79, 10%, hs-TropT n=84, 12%).

[0146] (Model development) Examination analysis was performed on consecutive patients using EMB, which is routinely employed in the diagnostic pathway (results are detailed in the supplementary materials). Six predictors, including hs-CRP, hs-TropT, native T1 and T2, LGE, and LVEF, met the pre-specified criteria for enrollment in the prognostic model with MACE. For the HF endpoint, additional predictors also included hematocrit, age, sex, and RVEF. Cardiovascular risk factors, medication, EMB-based parameters, and other measures of cardiac structure did not meet the enrollment criteria.

[0147] The estimated risk ratios (HR, 95% confidence interval, CI) for the above predictors regarding MACE across the entire dataset are shown in Table 1. The predictors satisfied the proportional risk assumption. The final multivariate Cox regression model (backward, Wald) included native T2, hs-TropT, LGE (chi-squared, 89.8, p<0.001) for MACE, and HF endpoint (chi-squared: 157.3, p<0.001). The 2-year event risk for individual patients was given by the following equation: 2-year stochastic event = 1 - 0.998exp(PI), PI for MACE = 0.673 × LGE + 0.245 × native T2(ms) + 0.906 × hs-TropT(lg10),

[0148] The PI for the HF endpoint can be calculated from PI = 0.427 × LGE + 0.251 × native T2(ms) + 0.823 × hs - TropT(lg10).

[0149] (Model validation) Figure 14A shows the myocarditis score for the probability of death or significant shock in patients with suspected myocarditis, and Figure 14B shows the myocarditis score for the probability of hospitalization for heart failure leading to death from heart failure in patients with suspected myocarditis. In particular, Figures 14A and 14B illustrate the good agreement between observed and predicted risks for the MACE or HF endpoint at 2 years. Cross-validation of the prognostic models revealed a calibration slope of 1.07 (95% CI, 0.78, 1.36), an intercept of -0.52, and a C-index of 0.82 (95% CI, 0.65, 0.99) for MACE. The separate index for the HF endpoint included a calibration slope of 1.04 (95% CI, 0.74, 1.16), an intercept of 0.01, and a C-index of 0.79 (95% CI, 0.59, 0.86).

[0150] (Center / Sensitivity Analysis for Healthcare Effectiveness - External Verification) The HR estimates from model adjustment for center effects were similar to those of the model developed without the centers. The effect of individual centers on the model, achieved by training the model with data from two centers with validation for the remaining centers, showed good agreement for both endpoints (see table below). [Table 2]

[0151] (Myocarditis risk score) The inventors calculated the 2-year risk for each endpoint using predictor-based scores from model-based predicted probabilities (MACE: a score of 1.5 if native T2(ms): ≥38 and a score of 2.0 ≥40; a score of 1 if hs-TropT(lg10pg / l); ≥0.84; and an LGE score of 0.5 if present; HF endpoint: a score of 1.0 if native T2(ms): ≥38 and a score of 2.0 ≥40; a score of 1 if hs-TropT(lg10pg / l); ≥0.84; and an LGE score of 0.5 if present). Three risk groups (low, moderate, and high risk) per endpoint were created using the mean percentage of event occurrences per group, i.e., MACE (0-<1 (<1%), 1-<2 (5%), ≥2 (22%)); HF endpoint (0-<1 (3%), 1-2 (20%), ≥2 (41%)) (see Figures 14A and 14B). Stratifying patients into three risk groups by myocarditis risk score revealed that the majority (56 patients, 87%) of all MACE and HF endpoints (102 patients, 79%) occurred in patients classified as high-risk by score. Figures 15A, 15B, and 15C show Kaplan-Meier curves for native T2 and hs-TropT (terniles) and LGE (binary) for MACE.

[0152] (Comparison with the standard of practice) A model based on standards of practice for initiating anti-remodeling therapy in HF patients (LVEF ≤ 45%) yielded C-index values ​​of 0.59 (0.52–0.66) for MACE and 0.56 (0.51–0.61) for HF endpoints. Stratifying patients based on LVEF (>46 vs ≤ 45%) revealed nearly equal proportions of MACE (32 (8%) vs 32 (11%) and HF events (67 (52%) vs 63 (49%) (0.27)) within both groups (p=0.14). Net reclassification calculations based on myocarditis risk scores were performed in patients with LVEF > 46% (n=442), of which 141 patients overlapped with the high-risk group by score (totaling 29 patients) and had MACE. This identifies a significant currently untreated subgroup that could potentially benefit from treatment to reduce event rates.

[0153] (Discussion) This is the first study to provide a prognostic model for personalized risk prediction in patients with a clinical diagnosis of myocarditis. Our findings demonstrate that myocardial edema, injury, and scarring, using T2 mapping, hs-TropT, and non-ischemic LGE, respectively, provide the most accurate estimates of outcomes in patients with high pre-test likelihood of myocarditis. The prognostic model was derived from a large and diverse population of patients with clinically suspected myocarditis from several European centers and healthcare systems. Our findings reveal the relatively weak predictive power of histological parameters, making the current concept of EMB as a central diagnostic requirement in myocarditis difficult. The myocarditis risk score provides an easy-to-use clinical tool based on readily available clinical parameters, based on the prognostic model derived in this paper. Comparison with modern management practices reveals the potential to identify additional patients with respect to currently untreated therapies. Further research is needed to test whether the myocarditis risk score can improve outcomes by using either available therapies or novel approaches for other purposes.

[0154] Our results are based on large prospective longitudinal tertiary hospital datasets from two European countries. In German centers, clinicians routinely employed EMB-based diagnostic pathways in many cases, while omitting invasive steps was a central approach to clinical management elsewhere. This is explained in part by the more severe myocarditis population encountered in German tertiary centers, in contrast to practices in the UK where university hospitals play a primary role as the point of presentation for specialized medical treatment. Nevertheless, the prognostic model was shown to be independent of the center or healthcare system, reaffirming the robust performance and effectiveness of the broad clinical trial patient criteria. In this paper, we avoided pre-selecting patients based on any prior diagnostic criteria, including the Lake-Louise criteria, and conducted examination analyses to independently determine predictors to be considered for the prognostic model. T2 mapping was identified as the strongest independent predictor in the examination dataset, and findings were subsequently replicated in training and validation datasets based on the entire cohort and recruiting centers. Examination and analysis indicate that, despite considerable advancements in histological diagnostic criteria over time, prognostic models cannot be refined by the addition of EMB information. Expanding the diagnostic criteria to also include immunohistochemical readouts aimed to address heterogeneous clinical manifestations, given the recognition that DC alone frequently yielded negative results in patients with evident clinical manifestations. Current ESC guidelines define IHC in terms of the required cell number and type of inflammatory infiltration. However, the guidelines also allow for methodological openness through “unspecified immunohistochemical criteria” and permit the search for better measurements. High costs, expertise in the necessary procedures, and an overall lack of specialized cardiac pathologist reference laboratories present further practical limitations, which remain unexamined. In contrast, magnetic resonance imaging techniques are widely available (averaging 25 and 38 units per million people in Germany and the United States, respectively).Most units are deployed in non-cardiac scanning, but this is rapidly changing due to standardized, shorter imaging protocols that can be completed in under 30 minutes.

[0155] The myocarditis risk score relies on a small number of selected myocardial measurements, two of which, namely hsTropT and LGE, are already well-established in the standard clinical assessment of suspected myocarditis. For the first time, T2 mapping is firmly placed within the realm of clinical utility. The overall robustness, availability, and simplicity of the three measurements mean that the score can be used routinely, enabling a uniform approach to the diagnosis and risk assessment of myocarditis. Non-ischemic myocarditis-like LGE strongly predicts both outcome endpoints, reaffirming observations from numerous previous studies in myocarditis. The presence of LGE is a consequence of infarct-like myocarditis resulting from focal myocardial necrosis and can be associated with significant troponin leakage in acute disease. Since significant troponin release and LGE co-localize only slightly, troponin cannot be explained by ongoing focal necrosis alone; this is rather associated with diffuse myocardial processes that are not detectable by LGE. Therefore, LGE has a prognostic role distinct from diffuse destructive processes, and its significance in myocarditis is evidenced by the much stronger prognostic association with hs-TropT and T2 mapping to both outcome endpoints.

[0156] Mapping techniques measure the magnetic properties of the myocardium, which are altered in the presence of diffuse disease, and thus provide objective, absolute measures of severity, disease progression, or response to treatment. Myocardial T2 mapping is a specific measure of myocardial edema. Extensive reports on elevated native T2 in disease models of inflammatory cardiomyopathy are complemented by histological verification of inflammation and myocardial injury. The strong prognostic association of T2 mapping in this study supports the idea that active inflammatory processes are central to myocardial injury in myocarditis and contribute to cardiac dysfunction and poor outcomes. Our findings are consistent with emerging evidence regarding the overall role of chronic inflammation in cardiovascular disease outcomes.

[0157] Contrary to the common belief that most cases of myocarditis resolve without sequelae, our findings confirm a significant morbidity and poor prognosis. T2 mapping and troponin were significantly associated with a higher prognosis compared to diffuse fibrosis and remodeling due to native T1 or LVEF, respectively. These findings are relevant because markers of remodeling and dysfunction are fundamental to cardiovascular assessment and clinical management, even in myocarditis. The myocarditis risk score categorizes the majority of patients with the event into a high-risk group (>80%), while LVEF identifies only 50% of these, suggesting the potential for reclassification of those at the highest risk, particularly with preserved LVEF that remains untreated. In patients receiving therapy, the myocarditis risk score may induce personalized treatment enhancement. Early targeting of myocardial inflammation may also benefit from reduced scarring burden and mitigation of HF development, which are currently limited due to unresolved means of early and accurate recognition. In summary, the myocarditis risk score may offer an opportunity for a significant step change from the current standard of clinical management.

[0158] An early sign of HFpEF is diastolic dysfunction measured by echocardiography as E / e'. Our experimental results show that a score developed from a combination of CMR parameters and a blood marker (BNP) correlates much more strongly with invasively determined diastolic function and echocardiography.

[0159] Color scheme The inventors propose the "Goethe color scheme," a color scheme for the standardized representation of one- and two-dimensional biomarkers. Currently, many different color schemes exist for displaying medical data. These are, firstly, used randomly, and secondly, do not allow for two-dimensional representation.

[0160] The proposed color model enables 1) a standardized, categorized representation with intuitive reading of values ​​from color, and 2) a combined representation of two markers simultaneously through a two-dimensional color model.

[0161] 1.) By creating a two-dimensional color space (instead of a regular circle), both the center ("normal" / "normal") and the increase / decrease of one or two parameters can be clearly represented.

[0162] 2.) This color space can be represented continuously and clearly defined (e.g., normal = less than 2 standard deviations below or above the mean; increased / decreased = 2 to 5 standard deviations above / below the mean; strongly increased / decreased = >5 standard deviations above / below the mean). This allows for the uniform presentation of two biomarkers using a single color. Thus, the color coding of pixels can simultaneously represent two pieces of information, for example, (fibrosis and inflammation) or (function and blood circulation).

[0163] To date, many one-dimensional color schemes with linear, continuous scales have existed. These can be arbitrarily divided into specific areas. In the medical field, echocardiography comes closest to the proposed concept, where blood velocity, direction, or potential turbulence is represented by different colors. For example, red is the color of blood flowing toward the transducer, while blood flowing away from the transducer is shown in blue. The various subtle differences in these colors represent the velocity of blood flow, with lighter colors indicating higher velocity. In addition, a green component is mixed in when turbulence occurs.

[0164] The following advantages are achieved here: 1.) Possibility of immediate assessment based on color. 2.) Standardized comparisons of different patients or follow-ups can be presented intuitively. 3.) Instead of integrating two data points from different datasets within the observer's brain into a wide area with a task of precise spatial allocation, each spatial data point can be assigned clear information about two parameters. 4.) The schemas used as standards in color theory are not sufficiently intuitive for medical applications (for example, abnormal and normal are not sufficiently separated, and the intermediate is defined as "white," which is usually "majority" in medical settings).

[0165] (Coordinate system for cardiac data) The proposed model enables standardized, machine-readable documentation of cardiac measurements with high resolution. Thus, different parameters from different imaging techniques can be documented in a precisely defined matrix. This allows for comparison with normal values, other patients, and follow-up observations. To date, a 17-compartment model has been widely used in clinical practice. At higher resolutions, this model divides the heart into 34 subcompartments, including the outer and inner layers. Within this model, voxels are often used, but they lack a clear location within the heart and therefore cannot be documented in a machine-readable manner. Regarding biomedical modeling, concrete models with finite elements and higher resolutions exist, but these can only be used by specialists and generally do not allow for clear, communicable locations.

[0166] Model: Instead of the currently common 17 compartments of the left ventricle, the left ventricle is divided into a four-dimensional coordinate system (in the longitudinal direction, circumferentially around the longitudinal axis, radially perpendicular to the myocardium, and along the time axis). This uses 100 longitudinal positions, 360 circumferential positions, 100 radial positions, and 1,000 temporal positions. Further subdivision or averaging of positions with lower resolution is also possible without issue. As a version that is easier to handle in terms of imaging, a division into longitudinal positions, 36 circumferential positions, radial positions, and 100 temporal positions is used.

[0167] The coordinate system can be used in two versions. 1.) Linear, i.e., all motion components are assumed to be linear (i.e., longitudinal shortening is assumed to be similar shortening in all cardiac planes). 2.) Realistic, i.e., existing knowledge / models of the moving component are integrated into the description. The advantages of the model are machine readability, clear description of each position in the left ventricle using three coordinates, the first possibility to clearly combine different imaging techniques and multi-parameter information (with different spatial resolutions), and therefore a vast number of possibilities.

[0168] Figure 16 shows the correlation between the most frequent echocardiographic measurement (E / e') and invasively measured left ventricular stiffness (TAU), which is considered a reference criterion. The correlation r = 0.45.

[0169] Figure 17 shows the correlation between the novel CMR score and invasively measured left ventricular stiffness (TAU) as a reference criterion. The correlation is r = 0.86.

[0170] Figure 18 shows a box plot of age-sex-risk factor-corresponding control (1-control), i.e., between asymptomatic post-COVID patients (2-cases) and patients with "long-term COVID" (3-long). ns = not significant, * = p < 0.05, ** = p < 0.01, *** = p < 0.001, **** = p < 0.0001. Note the strong p values ​​for echocardiographic native T1 and native T2 relative to echocardiographic E / e' (E_e).

[0171] The method presented allows for the diagnosis of early diastolic abnormalities in otherwise healthy individuals who are physically weakened after COVID-19 infection. Our data clearly demonstrate cardiac changes in patients after COVID-19. These changes are most pronounced in those patients who have new or ongoing symptoms more than four weeks after infection, also referred to as "long-term COVID." Again, echocardiography captures changes in the symptomatic group versus the asymptomatic group, but it does not detect differences between controls and COVID patients (see also Figure 18).

[0172] The foregoing description is merely an implementation of the present invention, and the scope of the invention is not limited thereto. Any modifications or substitutions can be readily made by those skilled in the art. Accordingly, the scope of protection of the present invention should be limited to the scope of protection of the appended claims.

Claims

1. A method for non-invasive quantitative imaging of the heart, wherein the method is Obtaining initial T1 and T2 maps from the aforementioned cardiac magnetic resonance images, In order to obtain a corrected T1 map, the initial T1 map is corrected using the T2 map. Includes, Correcting the initial T1 map is a method that includes subtracting the weighted value of the T2 map from the value in the initial T1 map and adding a constant.

2. The method according to claim 1, wherein the weight of the weighted value is 10 to 15, and the constant is 350 to 500.

3. The aforementioned method, The method involves using a 2D color map to map values ​​from the corrected T1 map to values ​​from corresponding positions in the T2 map to output colors, wherein in the 2D color map, the lowest value in a first direction corresponds to a first color, the highest value in the first direction corresponds to a second color, the lowest value in the second direction corresponds to a third color, the highest value in the second direction corresponds to a fourth color, and the first to fourth colors are different colors. Outputting the aforementioned output color onto the screen The method according to claim 1 or claim 2, further comprising:

4. The aforementioned method, To acquire a magnetic resonance image (MR image) of the aforementioned heart, Based on the aforementioned MR image, the vertical axis of the heart is determined, Determining at least one cardiac coordinate for at least one selected location in the MR image, wherein the cardiac coordinate comprises a vertical coordinate showing the projection of the location onto the vertical axis, a circumferential coordinate showing the circumferential position of the location around the vertical axis, and a radial coordinate showing the radial distance of the location from the vertical axis. Annotating the at least one selected location using the aforementioned cardiac coordinates. The method according to any one of claims 1 to 3, further comprising:

5. Annotating at least one of the selected locations is Outputting the aforementioned cardiac coordinates on the screen, The cardiac coordinates are stored in a data storage device, and / or Visualizing one or more lines overlaid across the MR image on the screen, wherein the one or more lines correspond to fixed longitudinal coordinates, fixed circumferential coordinates, and / or fixed radial coordinates of the cardiac coordinates. The method according to claim 4, including the method described in claim 4.

6. A method for annotating cardiac medical data, wherein the method is To acquire a magnetic resonance image (MR image) of the aforementioned heart, Based on the aforementioned MR image, the vertical axis of the heart is determined, Determining at least one cardiac coordinate for at least one selected location in the MR image, wherein the cardiac coordinate comprises a vertical coordinate showing the projection of the location onto the vertical axis, a circumferential coordinate showing the circumferential position of the location around the vertical axis, and a radial coordinate showing the radial distance of the location from the vertical axis. Annotating the at least one selected location using the aforementioned cardiac coordinates. Methods that include...

7. Annotating at least one of the selected locations is Outputting the aforementioned cardiac coordinates on the screen, The cardiac coordinates are stored in a data storage device, and / or Visualizing one or more lines overlaid across the MR image on the screen, wherein the one or more lines correspond to fixed longitudinal coordinates, fixed circumferential coordinates, and / or fixed radial coordinates of the cardiac coordinates. The method according to claim 6, including the method described in claim 6.

8. Adding the aforementioned annotation means To round the vertical position to one of a predetermined number of possible vertical positions, Rounding the circumferential coordinates to one of a predetermined number of possible circumferential positions, and / or, Rounding the radial coordinates to one of a predetermined number of possible radial positions. The method according to claim 7, further comprising:

9. The method according to claim 8, further comprising rounding the vertical position, wherein the initial rounding is used for visualization, i.e., overlaying the MR image and indicating the rough rounding boundary.

10. The method according to any one of claims 6 to 9, wherein the cardiac coordinate further comprises a temporal coordinate, the temporal coordinate being determined based on the temporal distance to a predefined point in the cardiac cycle.

11. The method according to claim 10, wherein the predefined points are defined with respect to the P-spike of the heart.

12. A method for operating a device to predict the risk of cardiac events for a patient suffering from myocarditis, based on multiple magnetic resonance images (MR images) previously acquired from the patient, wherein the method is: The device determines the T1 and T2 values ​​of a region of the patient's heart based on a plurality of previously acquired MR images, wherein the region of the heart is a single mid-ventricular short-axis slice. The device calculates the risk based on the weighted sum of the T1 value, the T2 value, and one or more additional coefficients. Methods that include...

13. The method according to claim 12, wherein the T1 value of the region is the average of the T1 map values ​​within the region, and / or the T2 value of the region is the average of the T2 map values ​​within the region.

14. The method according to claim 12 or 13, wherein the additional coefficient includes one or more of age, sex, hematocrit value, hs-CRP, hsTropT, LV-EF, RV-EF, and myocardial LGE, and / or the additional coefficient is determined using multivariate Cox regression.

15. A device for non-invasive quantitative imaging of the heart, wherein the device is configured to perform the method according to any one of claims 1 to 14.

16. A computer-readable storage medium for storing program code, wherein the program code comprises instructions, and the instructions, when executed by a processor, perform the method according to any one of claims 1 to 14.

17. The initial T1 map and the T2 map comprise the region of the heart, and the method further includes predicting the risk of cardiac events in a patient suffering from myocarditis. The above prediction is, Based on the region of the heart in the corrected T1 map, the T1 value is obtained, and based on the region of the patient's heart in the T2 map, The risk is calculated based on the weighted sum of the T1 value, the T2 value, and one or more additional coefficients. The method according to any one of claims 1 to 4, including the method described in any one of claims 1 to 4.

Images