Systems and methods for myocardial imaging using delayed-phase dynamic contrast enhancement MRI

Abstract

A method for performing magnetic resonance (MR) imaging includes receiving imaging data associated with a region of interest of a subject prior to an injection of a contrast agent into the region of interest, after the injection of the contrast agent into the region of interest, or both; obtaining a plurality of initial contrast-sensitive images of the region of interest of the subject, each of the plurality of initial contrast-sensitive images being indicative of a state of the region of interest at a plurality of times within an initial period of time; and generating, based on the plurality of initial contrast-sensitive images, a subsequent contrast-sensitive image that is indicative of the state of the region of interest at a subsequent time after the initial period of time ends.

Description

SYSTEMS AND METHODS FOR MYOCARDIAL IMAGING USING DELAYED-PHASE DYNAMIC CONTRAST ENHANCEMENT MRICROSS-REFERENCE TO RELATED APPLICATIONS

[0001] This application claims the benefit of U.S. Provisional Patent Application No. 63 / 520,017, filed on August 16, 2023, which is incorporated herein by reference in its entirety.STATEMENT REGARDING FEDERALLY SPONSORED RESEARCH

[0002] This invention was made with government support under Grant No. HL148788 awarded by the National Institutes of Health. The government has certain rights in the invention.TECHNICAL FIELD

[0003] The present disclosure relates generally to systems and methods for performing myocardial imaging, and more particularly, to systems and methods for performing myocardial imaging using delayed-phase dynamic contrast enhancement magnetic resonance imaging.BACKGROUND

[0004] Ischemic heart disease is the top cause of mortality in the developed world. As such, robust and effective imaging is needed to assess myocardial viability in patients. However, common techniques for myocardial imaging often require long exam times, produce weak image contrast between blood and myocardial lesions, and are non-quantitative and highly dependent on the imaging protocol used. Thus, improved systems and methods for performing myocardial imaging are needed.SUMMARY

[0005] According to some implementations of the present disclosure, a method for performing magnetic resonance (MR) imaging includes receiving imaging data associated with a region of interest of a subj ect prior to an inj ection of a contrast agent into the region of interest, after the injection of the contrast agent into the region of interest, or both; obtaining a plurality of initial contrast-sensitive images of the region of interest of the subject, each of the plurality of initial contrast-sensitive images being indicative of a state of the region of interest at a plurality of times within an initial period of time; and generating, based on the plurality ofinitial contrast-sensitive images, a subsequent contrast-sensitive image that is indicative of the state of the region of interest at a subsequent time after the initial period of time ends.

[0006] The above summary is not intended to represent each implementation or every aspect of the present disclosure. Additional features and benefits of the present disclosure are apparent from the detailed description and figures set forth below.BRIEF DESCRIPTION OF THE DRAWINGS

[0007] The disclosure, and its advantages and drawings, will be better understood from the following description of representative embodiments together with reference to the accompanying drawings. These drawings depict only representative embodiments and are therefore not to be considered as limitations on the scope of the various embodiments or claims.

[0008] FIG. 1 shows different imaging contrasts used to perform multiparametric myocardial MR imaging, according to aspects of the present disclosure.

[0009] FIG. 2 is a block diagram of a system for obtaining dDCE images, according to aspects of the present disclosure.

[0010] FIG. 3 is a flowchart setting forth steps of a method, according to aspects of the present disclosure.

[0011] FIG. 4 is a flowchart setting forth steps of another method, according to aspects of the present disclosure.

[0012] FIG. 5 shows an example image processing pipeline, according to aspects of the present disclosure.

[0013] FIG. 6A is a graph illustrating an example concentration curve, according to aspects of the present disclosure.

[0014] FIG. 6B is a panel of images showing example dDCE parameter maps, according to aspects of the present disclosure.

[0015] FIG. 6C is a panel of images showing example synthetic images, according to aspects of the present disclosure.

[0016] FIG. 7 shows an example image acquisition sequence for obtaining dDCE images, according to aspects of the present disclosure.

[0017] FIG. 8 is a panel of example conventional multiparametric maps, according to aspects of the present disclosure.

[0018] FIG. 9 shows dDCE images corresponding to the multiparametric maps of FIG. 8, according to aspects of the present disclosure.

[0019] FIG. 10A is a graph showing example growth truth concentration measurement,according to aspects of the present disclosure.

[0020] FIG. 1 OB is a graph showing an example model applied to measurements in FIG. 10A, according to aspects of the present disclosure.

[0021] FIG. 10C is a graph showing another example model applied to measurements in FIG. 10A, according to aspects of the present disclosure.

[0022] FIG. 11A is a graph showing example concentration curves measured 30 minutes post-contrast injection and predictive dDCE concentration curves, according to aspects of the present disclosure.

[0023] FIG. 1 IB is a graph showing yet other example concentration curves measured 5 minutes post-contrast injection along with predictive dDCE concentration curves, according to aspects of the present disclosure.

[0024] FIG. 12 shows is a panel comparing measured DCE parameters for 30 minutes postcontrast injection and predictive dDCE-derived DCE parameters for 30 minutes post-contrast injection obtained from data acquired within 5 minutes post-contrast injection, according to aspects of the present disclosure.

[0025] FIG. 13 A is a graph showing one example dDCE parameter derived in different myocardial territories, according to aspects of the present disclosure.

[0026] FIG. 13B is a graph showing another example dDCE parameter derived in different myocardial territories, according to aspects of the present disclosure.

[0027] FIG. 13C is a graph showing yet another example dDCE parameter derived in different myocardial territories, according to aspects of the present disclosure.

[0028] FIG. 13D is a graph showing yet another example dDCE parameter derived in different myocardial territories, according to aspects of the present disclosure.

[0029] FIG. 14 is a graphical illustration showing example EGE and LGE images of the same animal under MI obtained using conventional techniques and obtained using a dDCE approach, according to aspects of the present disclosure.

[0030] FIG. 15A shows a Bland- Altman analysis and a linear regression comparing infarct area measurements made using conventional techniques and a dDCE approach, according to aspects of the present disclosure.

[0031] FIG. 15B shows a Bland- Altman analysis and a linear regression comparing transmurality measurements made using conventional techniques and a dDCE approach, according to aspects of the present disclosure.

[0032] FIG. 15C shows a Bland- Altman analysis and a linear regression comparing persistent microvascular obstructions measurements made using conventional techniques anda dDCE approach, according to aspects of the present disclosure.

[0033] FIG. 16A is a graphical illustration showing example dDCE maps, in accordance with aspects of the present disclosure.

[0034] FIG. 16B are graphs showing example dDCE parameter measurements in different myocardial territories, according to aspects of the present disclosure.

[0035] FIG. 17A is a graphical illustration comparing standard and example dDCE images for acute MI, according to aspects of the present disclosure.

[0036] FIG. 17B is a graphical illustration comparing standard and example dDCE images for chronic MI, according to aspects of the present disclosure.

[0037] FIG. 17C is a graphical illustration comparing standard and example dDCE images for persistent microvascular obstruction, according to aspects of the present disclosure.DETAILED DESCRIPTION

[0038] Ischemic heart disease is the top cause of mortality in the developed world. Major advances in acute care have now reversed the trend of immediate death from myocardial infarction (MI) in clinical practice. However, the long-term morbidity, particularly from ischemic heart failure (IHF), in these patients is greater than ever. The therapeutic options available to these patients (medical, surgical, and / or device-based) require knowledge of the presence of prior (chronic) infarction(s) and its characteristics, such as compromised contrast permeability in infarcted myocardium and reduced blood volume in microvascular obstruction.

[0039] Multiparametric myocardial magnetic resonance imaging (MRI) has been used to characterize the myocardium with different MRI contrasts. FIG. 1 shows example images with various imaging contrasts (e.g., Tl, T2, T2*, early gadolinium enhancement (EGE), late gadolinium enhancement (LGE), and extracellular volume (ECV)) of myocardial tissue suffering from acute myocardial infarction (MI) and chronic MI. However, multiparametric imaging can involve prolonged image acquisition time, which can significantly extend scanner sections for cardiac magnetic resonance (CMR) exams and diminish incentives for routine CMR prescription for regular healthcare providers. In most CMR practice today, less imaging than recommended image protocol is being performed to reduce patient strain.

[0040] Among multiparametric imaging, LGE is the most prescribed CMR protocol because it provides the most clinically relevant information for patient management. For evaluating myocardial viability, LGE imaging is today's gold standard recommended by the American Heart Association (AHA) guidelines. Yet despite its capabilities and wide adoption, conventional LGE imaging has several limitations.

[0041] One limitation is long scan time. For instance, conventional LGE imaging of myocardium tissue typically involves a wait time of about 10 minutes to about 15 minutes to allow for contrast agent washout in the myocardium and establish signal enhancement in infarcted regions. Wait times prolong scan time, and reduce applicability of conventional LGE imaging for time-sensitive clinical cases, such as acute MI. Conventional LGE imaging represents a snapshot of contrast washout dynamic, and varies based on imaging protocol. For instance, signal intensity is highly dependent on acquisition timing, contrast agent dose, and imaging protocols. The qualitative nature of conventional LGE imaging impairs comparability with longitudinal studies, and makes it challenging to evaluate diffused diseases that do not show focal lesions. Yet another limitation is that conventional LGE imaging can produce weak lesion / blood contrast. Because T1 values of blood and LGE-enhanced lesions are similar, it can be difficult to delineate LGE lesions from LV blood pool. Hence, detectability of subendocardium injuries using conventional LGE imaging may be compromised, and lead to misdiagnosis in some small but important lesions. Therefore, there is a need for improved approaches to myocardial imaging.

[0042] Turning to FIG. 2, a block diagram of an example system 200, in accordance with aspects of the present disclosure, is illustrated. As shown, in some embodiments, the system 200 may include an imaging apparatus 202 and a processing device 212.

[0043] The imaging apparatus 202 can include an MRI system, which may include a primary magnet 204, gradient coils 206, a radio-frequency (RF) transmission system 208, and an RF receiving system 210. The primary magnet 204 can be a permanent magnet, an electromagnet (such as a coil), or any other suitable magnet. The primary magnet 204 may be used to create an external magnet field that is applied to a sample or subject during imaging, and the gradient coils 206 create one or more secondary magnet field (i.e., gradient field(s)) that distorts the external magnetic field and cause the resonant frequency of the protons in the sample or subject to vary by position. The gradient coils 206 can thus be used to spatially encode the positions of protons throughout the sample or subject, e.g., can be used to select which plane intersecting the sample will be used for imaging.

[0044] The RF transmission system 208 may be used to apply one or more RF pulse sequence to image a region of interest of the subject (e.g., the subject’s heart or portion thereof, such as a myocardium). The RF transmission system 208 generally includes a frequency generator (such as an RF synthesizer), a power amplifier, and a transmitting coil. Generated RF excitations waveforms are applied by the RF system to perform the RF pulse sequence. Responsive to the RF pulse sequence, signals are emitted by tissue in the region of interest andreceived by the receiving system 210. The RF receiving system 210 can use a receiving coil to capture signals emitted from the region of interest, and a pre-amplifier to boost received signals and ensure that signals are suitable for processing. In some implementations, the RF receiving system 210 can include a signal preparation or processing component that can prepare and / or process received signals and provide signal data usable by the processing device 212 and / or another processing device (e.g., a user input device 220). As such, the RF receiving system 210 may be configured to filter, demodulate, and digitize captured signals.

[0045] As illustrated in FIG. 2, the processing device 212 can be communicatively coupled to the imaging apparatus 202 (e.g., via one or more wired and / or wireless communication link). In some embodiments, the processing device 212 may include a processor 214, a processorexecutable or processor-accessible memory 216, a display 318, and a user input device 220. In some implementations, the processing device 212 may be used to manage operations of the imaging apparatus 202 and can thus be configured to cause the imaging apparatus 202 to perform MR imaging as described herein.

[0046] While FIG. 2 illustrate a certain configuration of the system 200, it may be appreciated that the system 200 can also include one or more additional processing devices such that various tasks corresponding to MR imaging can be performed by different processing devices. The system 200 can also include one or more printers, one or more network interfaces, one or more other types of hardware, and so forth. Also, in some embodiments, the system 200, or various components therein, can include one or more housings.

[0047] In some implementations, a system to implement steps of methods described herein can include a control system, which may include one or more processors. In some implementations, a computer program product comprises instructions, which when executed by a computer (e.g., a control system, one or more processing devices and / or processors, etc.), carries out steps of methods described herein. The computer program product may be a non- transitory computer readable medium. In some implementations, a system to implement steps of methods described herein includes a memory device and a control system. The memory device has stored thereon machine-readable instructions. The control system includes one or more processors that are configured to execute the machine-readable instructions to carry out the steps of methods described herein.

[0048] DCE imaging involves administering contrast agent to a subject and tracking flow of the contrast agent over time. As described in more detail below, in accordance with aspects of the present disclosure, one or more synthetic image may be generated. In particular, one or more synthetic LGE image (that can mimic conventional LGE imaging) may be generated.Unlike conventional LGE imaging, synthetic imaging or delayed DCE (dDCE) imaging, in accordance with aspects of the present disclosure, may be produced with comparable diagnostic value but without appreciable wait time.

[0049] Turning now to FIG. 3, a flowchart setting forth steps of a method 300, in accordance with aspects of the present disclosure, is illustrated. In some implementations, steps of the method 300 may be carried using any system, such as system 200 described with reference to FIG. 2.

[0050] The method may begin at step 310 with receiving imaging data associated with a region of interest or sub-region therein. By way of example, a region of interest may include a subject’s heart, chest cavity, and so forth, while a sub-region, for instance, of the heart may include a myocardium, endocardium, epicardium, or portions or tissues therein. Imaging data received at step 310 is associated with imaging performed on the region of interest of the subject prior to the injection of a contrast agent, during the injection of the contrast agent, after the injection of the contrast agent, or any combination thereof. The contrast agent can be a gadolinium-based contrast agent, and / or any other suitable contrast agent. In some implementations, the imaging data may be DCE imaging data acquired using one or more RF pulse sequence applied to image the region of interest of the subject before, during, and / or after the contrast agent is injected. Acquired imaging data can be used to track a flow of the contrast agent, and measure various parameters (e.g., associated with the flow of the contrast agent).

[0051] At step 320, a plurality of initial contrast sensitive images of the region of interest are obtained. Each of these initial contrast-sensitive images is indicative of the state of the region of interest within an initial period of time (e.g., the contrast-sensitive images show the region of interest at one or more point in time within the initial period of time). In some implementations, the initial period of time begins prior to the administration of the contrast agent and ends after the administration of the contrast agent. In this manner, the initial contrastsensitive images can include at least one contrast-sensitive image that is indicative of the state of the region of interest before the injection of the contrast agent, and at least one contrastsensitive image that is indicative of the state of the region of interest after the injection of the contrast agent. In some implementations, the initial period of time begins prior to the injection of the contrast agent, at the time when the injection of the contrast agent occurs, or after the injection of the contrast agent. In some implementations, the initial period of time ends after the injection of the contrast agent. For example, the initial period of time can end about 1 minute after the injection of the contrast agent, about 2 minutes after the injection of the contrast agent, about 3 minutes after the injection of the contrast agent, about 4 minutes afterthe injection of the contrast agent, about 5 minutes after the injection of the contrast agent, about 10 minutes after the injection of the contrast agent, between about 1 minute and 5 minutes after the injection of the contrast agent, etc.

[0052] In some implementations, the plurality of initial contrast-sensitive images includes one or more maps of the concentration of the contrast agent within the region of interest, such as T1 maps, T2 maps, T2* maps, and so forth. The initial contrast-sensitive images could also include EGE images, ECV images, and / or other DCE images. In some implementations, the plurality of initial contrast-sensitive images can include at least one pre-contrast image and at least one post-contrast image. In implementations that include at least one pre-contrast image, the at least one pre-contrast image can include a pre-contrast T1 map, a pre-contrast T2 map, a pre-contrast T2* map, a pre-contrast EGE image, and so forth. In implementations that include at least one post-contrast image, the at least one post-contrast image can include a post-contrast T1 map, a post-contrast T2 map, a post-contrast T2* map, a post-contrast EGE image, etc. In some implementations, the plurality of initial contrast-sensitive images include multiple postcontrast images that are obtained at different points in time within the initial period of time after the injection of the contrast agent (e.g., a first post-contrast T1 map acquired one minute after the contrast agent injection, a second post-contrast T1 map acquired two minutes after the contrast agent injection (and one minute after acquisition of the first post-contrast T1 map), and so forth).

[0053] At step 330, a subsequent contrast-sensitive image is obtained that is indicative of the state of the region of interest at a subsequent time that occurs after the initial period of time ends. In some implementations, the subsequent time occurs about 10 minutes after the initial period of time ends. In some implementations, the subsequent time occurs about 15 minutes after the injection of the contrast agent. In implementations where the initial period of time ends about 5 minutes after the injection of the contrast agent, the subsequent time is both 10 about minutes after the initial period of time ends, and about 15 minutes after the injection of the contrast agent.

[0054] In general, the subsequent contrast-sensitive image is indicative of (e.g., shows) the state of the region of interest at the subsequent time, but can be generated prior to the subsequent time actually occurring. For example, if the first period of time ends about 5 minutes after the injection of the contrast agent, the subsequent contrast-sensitive image can be obtained generally immediately, without having to wait until the subsequent time, which may not be for another 10 minutes. In this manner, the contrast-sensitive image showing thestate of the region of interest at the subsequent time can be obtained faster, which can improve the imaging process and make it easier for patients to undergo.

[0055] In some implementations, the subsequent contrast-sensitive image (indicative of the state of the region of interest at the subsequent time) is based only on the plurality of initial contrast-sensitive images (indicative of the state of the region of interest within the initial period of time). In other implementations, the subsequent contrast-sensitive image may additionally be based on other data. However, the subsequent contrast-sensitive image is generally not based on any data that cannot be collected until the subsequent time.

[0056] In some implementations, the subsequent contrast-sensitive image includes an LGE image. In some implementations, the subsequent contrast-sensitive image includes a T1 map, a T2 map, a T2* map, an ECV image, and / or any other DCE image. In some implementations, the plurality of initial contrast-sensitive images includes multiple contrast agent concentration maps based on the type of contrast agent being used, and the subsequent contrast-sensitive image includes an LGE image. In some implementations, multiple subsequent contrastsensitive images are obtained. Each of these subsequent contrast-sensitive images can be obtained at a subsequent time that occurs after the initial period of time ends.

[0057] In some implementations, step 330 includes inputting the plurality of initial contrast sensitive images into a trained model and receiving the subsequent contrast-sensitive image (or multiple subsequent contrast-sensitive images) from the trained model. The trained model can be trained to generate a subsequent contrast-sensitive image indicative of the state of the region of interest at the subsequent time based on the plurality of initial contrast-sensitive images obtained within the initial period of time. Thus, the training data used to train the trained model can include multiple sets of (i) a plurality of initial contrast-sensitive images (such as T1 maps and / or other contrast agent concentration maps), and (ii) one or more corresponding subsequent contrast-sensitive images that were obtained using conventional techniques (e.g., by waiting until the subsequent time to apply a pulse sequence and obtained the resulting imaging data). In some implementations, the trained model includes a deep learning model, and / or other models.

[0058] In some implementations, one or more DCE parameters can be generated from the imaging data and / or from one or more of the plurality of initial contrast-sensitive images. The DCE parameters can include ktrans, which is the volume transfer constant for the contrast agent between blood plasma in the region of interest and the extravascular extracellular space (EES) within the region of interest; kep, which is the time contrast for the reflux of the contrast agent from the EES within the region of interest back into the portion of the subject’s vascular systemwithin the region of interest; ve, which is the volume of the EES within the region of interest as a percentage of the overall volume of the region of interest; vp, which is the volume of the blood plasma within the region of interest as a percentage of the overall volume of the region of interest; F, which is the flow of the contrast agent; PS, which is the vascular permeability within the region of interest; and / or other parameters.

[0059] In some implementations, generating the one or more DCE parameters can include inputting imaging data and / or contrast-sensitive image(s) into a model. In some implementations, the model may be same model that generates the subsequent contrastsensitive image or a different model. In general, the one or more DCE parameters can indicate the state of the region of interest at (i) at least one of the plurality of times within the initial period of time, (ii) the subsequent time after the initial period of time ends, (iii) a point in time that is after an end of the first period of time and before the subsequent time, or (iv) any combination of (i)-(iii).

[0060] Turning now to FIG. 4, a flowchart setting forth steps of another method 400, in accordance with aspects of the present disclosure, is illustrated. Steps of the method 400 may be carried out using any combination of suitable devices or systems, such as the system 200 described with reference to FIG. 2. In some embodiments, steps of the method 400 may be implemented as instructions stored in non-transitory computer-readable media, as a program, firmware or software, and executed by a general-purpose, programmed or programmable computer, processor or other computing device. In other embodiments, steps of the method 400 may be hardwired in an application-specific computer, processor, dedicated system, or module. Although the method 400 is illustrated and described as a sequence of steps, it is contemplated that the steps may be performed in any order or combination, need not include all illustrated steps, and may include additional steps.

[0061] The method may begin at step 410 with receiving imaging data associated with a region of interest in a subject. In some implementations, imaging data may be acquired using an imaging apparatus, such as an MRI system, by performing one or more RF pulse sequence, as described. Imaging data may be acquired at multiple time points, for instance, before, during, and after administration of a contrast agent to the subject. For instance, as described, imaging data may be acquired during an initial period of time, as well as at a subsequent time that occurs after the initial period of time ends.

[0062] Acquired imaging data may be then processed to generate one or more images or maps of the region of interest. As indicated by step 420, in some implementations, a plurality of contrast-sensitive images and / or maps may be generated using the imaging data acquired.For instance, one or more pre- and / or post-cost T1 map, T2 map, T2 map, and so forth, may be generated at step 420. For instance, in some implementations, an image registration algorithm may be performed. More particularly, one or more post-contrast contrast-sensitive image or map may be registered to one or more pre-contrast contrast-sensitive image, for example, using a non-rigid image registration algorithm. In some implementations, a series of time-resolved contrast-sensitive images and / or maps may be generated using the imaging data received at step 410. Various post-processing may also be performed.

[0063] Contrast-sensitive images and / or maps may then be analyzed to generate a plurality of dDCE parameters. In particular, a plurality of contrast concentration curves may be generated from contrast-sensitive images and / or maps, as indicated by step 430. In some implementations, each of the plurality of contrast concentration curves reflects a time evolution of contrast in a region of interest, or a sub-region therein, and may be generated by a pixelwise processing of time-resolved contrast-sensitive images and / or maps depicting the region of interest, or sub-region.

[0064] A plurality of dDCE parameters can then be derived by applying a model to the plurality of concentration curves, as indicated by step 440. More particularly, the concentration curves may be fit using a model that quantifies contrast agent dynamics. In some non-limiting examples, dDCE parameters may include ktrans, kep, ve, vp, F, PS, and so forth, as described. In some non-limiting examples, the model may include a tracer kinetic (TK) model, an Extended Tofts-Kety (ETK) model, a diffusion-perfusion (DP) model, a modified two-compartment exchange model (2CXM), a compartmental tissue uptake model (CTUM), a PM model, a TH model, a deep learning model, and so forth.

[0065] Derived dDCE parameters may then be utilized to generate one or more synthetic image, as indicated by step 450. In some implementations, one or more synthetic LGE image, one or more synthetic image, or a combination thereof, corresponding to a post-contrast point in time may be generated using derived dDCE parameters. By way of example, the postcontrast point in time may be approximately 1 minute after the injection of the contrast agent, approximately 2 minutes after the injection of the contrast agent, approximately 3 minutes after the injection of the contrast agent, approximately 4 minutes after the injection of the contrast agent, approximately 5 minutes after the injection of the contrast agent, approximately 10 minutes after the injection of the contrast agent, approximately 15 minutes after the injection of the contrast agent, or more. In some implementations, the post-contrast point in time may be less than an extended washout time. As appreciated from description herein, generatedsynthetic image(s) may be advantageously used for diagnosis without delays typically associated with conventional imaging.

[0066] Referring to FIG. 5, an example image processing pipeline, in accordance with aspects of the present disclosure. In this example, multiple Modified Look-Locker Imaging (MOLLI) T1 maps were acquired before and after Gd contrast injection to derive time-resolved concentration in the heart. A model-based image registration was applied for motion correction. Pixel -wise Gd concentration curves of a defined period post-contrast injection were introduced into a modified 2CXM model to derive corresponding dDCE parameters at 5 and 30 minutes after contrast administration. An LGEdDCE image was synthesized at 5 minutes after contrast administration using fitted dDCE.

[0067] To illustrate, FIGs. 6A-6C show example concentration curves, dDCE parameter maps, and concentration maps, respectively. In particular, the dynamic concentration curves of FIG. 6A show that a 2CXM model can align well with measured values, and can successfully differentiate, for instance, myocardial infarction (MI), remote, and blood pool. FIG. 6B shows PS, Fp, vp, and vemaps from dDCEsomin images, illustrating that corresponding changes related to MI region and obstruction (MVO). Elevated veand PS were demonstrated in the hyperintense late gadolinium enhancement (LGE) image (arrowheads). Fpand vpwere decreased in the hypointense cores on the early gadolinium enhancement (EGE) image (arrows). FIG. 6C shows dDCEso min synthetic images show that MI and remote regions may be consistent with Gd concentration images.

[0068] Example Study

[0069] T en animal subj ects (dogs) with MI were imaged 5 days and 8 weeks after surgically induced myocardial infarction to investigate the DCE parameter changes under acute and chronic Mis. To trace the contrast agent dynamic, myocardial T1 maps were acquired before and after gadolinium contrast injection. Post-contrast T1 maps were scanned consecutively with a 1-2 minute interval. Images were acquired up to 30 minutes post contrast injection. The pulse sequence, contrast injection, image acquisition, and map acquisition protocol are depicted in FIG. 7. In this example study, two DCE models (the extended Tofts-Kety (ETK) model and the two-compartment exchange (2CMX) model) were used to fit the delayed phase myocardial contrast dynamics. Pixel wise DCE parameters were derived in the myocardium. The DCE parameters were measured in different myocardium territories (Remote, MVO, MI, inflammation (Edema), and Hemorrhagic core (HMI)) to measure the underlying physiological changes.

[0070] To explore washout dynamic of contrast agent in the heart, post-contrast T1 maps were acquired continually starting at 1 -minute post-contrast injection for 30 minutes. In brief, post-contrast T1 maps were registered to the native T1 map using a novel image registration incorporating a physiological model to mitigate motion between scans. Contrast agent concentration in the heart was calculated using the R1 (1 / T1) differences between the pre and post-contrast myocardium using the following equation: Ct(t) = (R^ t) — R native) / RelaxitivityGd, where Ct(t) is the temporal gadolinium contrast agent concentration in the myocardium, is the post-contrast 1 / Ti at time t, Rlnative is the l / (native Ti), and RelaxivityGdis the Ti relaxivity of the contrast agent.

[0071] An Extended Toft and Kermode (ETK) model was adopted. The contrast agent concentration between capillary plasma Cp(t), arterial plasma Ca(t), and extracellular and extravascular spaces (EES, Ce(t)) is formulated by the coupled system of differential equations and specified by convolving the AIF with the tissue impulse response function=Rtrans \ Cp(t) ~ From the tissue response, DCE parameters (ktrans, ve, and vp) can besolved using the convolution form of the differential equation. To probe the potential of myocardial viability assessment from the dDCE model with a shortened washout time, dDCE parameters were fitted using the whole dataset (1-30 mins, dDCEsomin) and a subset of the postcontrast TI images (1, 3, and 5 minutes, dDCEsmin). In addition, the dDCEsmin parameters were used to synthesize EGEHDCE images at 2 minutes and LGEdDCE images at 15 minutes, adopting a populational model of the Arterial Input Function (AIF).

[0072] The capability of eliminating LGE wait time was examined with DCE models derived from a subset of the post-contrast TI maps that were obtained between about 1 minute and 4 minutes after injection of the contrast agent. The DCE parameters were used to derive the contrast concentration at 15 minutes after injection of the contrast agent. Standard early Gadolinium enhancement images were acquired 1 minute injection of the contrast agent, and LGE images were acquired 15 minutes injection of the contrast agent to validate the prediction from the developed technique.

[0073] To explore the washout dynamic of the contrast agent in the heart, post-contrast TI maps were acquired continually starting at 1-minute post-contrast injection for 30 minutes. In brief, post-contrast TI maps were registered to the native TI map using a novel image registration incorporating the physiological model to mitigate motion between scans. Contrast agent concentration in the heart was calculated using the R1 (1 / TI) differences between the pre and post-contrast myocardium using the following equation: Ct(t) = ( / ?i(t) — Rinattve) / RelaxitivityGd, where Ct(t) is the temporal gadolinium contrast agent concentration in the myocardium, is the post-contrast 1 / Ti at time t, Rlnative is the 1 / native Ti, and RelaxivityGdis the Ti relaxivity of the contrast agent.

[0074] FIG. 8 shows representative multiparametric maps (TI, T2, T2*, EGE, LGE, and ECV (extracellular volume)) from one of the subjects with acute MI. FIG. 9 shows dDCE images (ve, vp, ktrans, and kep) that correspond to the multiparametric maps of FIG. 8. Significant signal differences are seen in the affected territories. Notably, the reduced ktrans in the microvascular obstruction (MVO) territory and the elevated extracellular and extravascular space (Ve) in the infarct zone aligned with the underlying pathological condition and previously reported conditions.

[0075] To further show that the proposed model can accurately portray the gadolinium contrast washout dynamic in various microcirculation conditions, the fitted curve of the proposed model is measured in an acute MI animal and compared to the ground truth (monitoring 30 minutes post-contrast injection), as illustrated in FIGs. 10A-10C. The dDCE models showed high-quality fits for all territories and aligned nicely with the grown truth (in- vivo acquisition). Notably, the signal progress in different myocardial territories progressed in different regimes due to a wide spread of microvasculature environment, which shows that the proposed model can cover a wide pathological range of microcirculation and nicely portray the temporal dynamic of contrast agents under diseases.

[0076] The ability of dDCE models to reduce wait time for the LGE images is illustrated in FIGs. 11 A-14. For instance, FIGs. 11 A and 1 IB show the predictive dDCE signal curves at different times post-contrast injection, derived from healthy and infarcted myocardium. In particular, curves on FIG. 11A are standard curves using data that was acquired across 30 minutes post-contrast injection. By contrast, curves in FIG. 11B are dDCE curves predicted from data acquired within 5 minutes post-contrast injection. As appreciated from FIGs. 11A and 1 IB, dDCE curves predicted using 5 minutes of data acquisition show high similarity to the acquired 30 minutes curves. FIG. 12 shows the dDCE parameters derived from data acquired within 5 minutes post-contrast injection and parameters derived from data acquired across 30 minutes post-contrast injection. Comparable dDCE images with similar trends in delineating the diseased territories are shown in both dDCE maps.

[0077] In another example, FIGs. 13A-13D shows the quantitative comparisons of the dDCE parameters derived in different myocardial territories. Comparable values were extracted from 5-minute and 30-minute dDCE models, which show sufficient information for5 minutes of data acquisition to extract late enhancement images 15 minutes post contrast injection. In yet another example, FIG. 14 shows EGE and LGE images derived using a dDCE model, according to the present disclosure, in comparison with standard EGE and LGE images from a clinical protocol. As appreciated from FIG. 14, comparable lesion appearances are shown in the dDCE-derived images with data acquired in less than 5 minutes.

[0078] The results show that dDCE CMR can delineate myocardial territories from different pathological states in a single scan. It provides quantitative measures for lesions in acute and chronic infarctions and essential information for myocardial tissue characterization. Furthermore, EGE and LGE images derived from dDCE CMR with data acquired in less than 5 minutes can accurately detect MI comparable to standard LGE.

[0079] Data Analysis

[0080] In the dDCE maps, the myocardium was segmented manually with endo and epimyocardial contours using the native T1 map. The dDCE parameters were measured and compared in different myocardium territories (Remote, Microvascular obstruction (MVO), Myocardial infarction (MI)) to probe the underlying physiological changes. The region of interest (ROI) were defined using the standard LGE (LGEstandard) images. To test the model fidelity with truncated imaging time, dDCE parameters reconstructed with 5 minutes of images (dDCEsmin) were compared to dDCE parameters derived from the whole dataset (dDCEsomin).

[0081] For investigating the diagnostic property of the proposed method without the conventional LGE wait time, synthetic LGE images derived from the dDCEsmin parameters (LGEdDcr) were analyzed and compared to the LGEstandard. The images were randomized and independently read by two blinded reviewers, and areas of disagreement were resolved by consensus. Detection and quantification of MI and MVO were compared between the LGEdocr and LGEstandard images. Semi -quantitative standards were adopted for MI delineation (MI=mean+5SD of the remote territory, MVO=hypointense regions in MI). Slice-wise MI size, MI location, MI transmurally (%), Microvasculature obstruction (MVO) size, and MVO location were compared.

[0082] Statistical Analysis

[0083] Shapiro-Wilk test and quantile-quantile plots were used to test the normality of the data. The whole-slice infarct area, the mean transmurality, and the percentage of mean MVO area were compared between LGE images and dDCE images using either paired Student t-test or Wilcoxon signed-rank test based on the normality of the data. Bland- Altman analysis was performed to evaluate the agreement between the two techniques. Simple linear regression was done to estimate the correlation between the two techniques concerning infarct area andtransmurality measurements. The slope and intercept of the best-fit line were tested to be equal to 1 and 0, respectively. Using LGE images as the gold standard, the sensitivity and specificity of the dDCE images to detect the infarct area at the segmental level and the percentage of mean MVO area were measured respectively. The receiver operating characteristic curve (ROC curve) was performed to measure the area under curve (AUC) for the infarct area. The threshold of segmental infarct area and mean MVO area in LGE images was set at 1%.

[0084] Results

[0085] As appreciated from results herein, dynamic T1 maps acquired within 5 mins postcontrast injection (dDCEsmins) can depict a whole contrast washout dynamic. The ability of dDCE models to reduce wait time for the LGE images is illustrated in FIGs. 11A-11B, for example. The dDCE maps derived from data acquired between 1-5 minutes and 1-30 minutes post-contrast injection are presented and compared to the corresponding standard LGE images in the same animal undergoing acute and chronic Mis. Comparable dDCE maps with matched lesion locations and extend are presented in all maps. Notably, both dDCE maps can delineate the diseased territories reliably compared to the LGE. Although a higher noise level is presented, dDCEsmins parameters showed comparable measurements and good spatial correspondence to the dDCEsomins.

[0086] dDCEsmins can derive late Gd enhancement images (LGEdDcr) that provide a comparable diagnostic capability to the standard LGE images without the extended wait time. For instance, FIG. 14 depicts a set of representative mid-ventricular gadolinium-enhanced images of the same animal under acute and chronic MI. The images obtained using a standard phase-sensitive inversion recovery (PSIR) sequence and dDCE modeling exhibited a remarkable similarity in their enhancement patterns, demonstrating the excellent correspondence between the standard gadolinium enhancement images and the proposed method. Notably, the presence of MVO in acute MI was resolved during the chronic stage, which was accurately captured by both methods. The dDCE synthesized images achieve highly accurate EGE and LGE images without requiring extended contrast washout time in the standard imaging protocol.

[0087] To quantify the effectiveness of MI delineation using LGEdDCE, a comprehensive quantitative analysis was performed, which is shown in FIGS. 15A-15C. The results demonstrated a high degree of correspondence between LGEdDCE and the standard LGE images in both MI size and transmurality measurements, as evidenced by linear regression and Bland- Altman analysis (MI size: R2= 0.95; slope= 0.93, p < 0.01; bias= -1.74±6.60%; FIG. 15A. Transmurality R2 =0.97; slope, 0.93, p < 0.01; bias= 1.86±2.73%; p = 0.05; FIG. 15B).Moreover, the receiver operating characteristic (ROC) analysis demonstrated that LGEdDcr achieves high accuracy in MI detection, with a sensitivity of 94.4%, specificity of 96.7%, and an area under the curve (AUC) of 0.96. The ability of LGEdDCE to detect MVO was also evaluated. Among the 10 acute MI animals, 6 exhibited persistent MVO within the MI. LGEdDCE showed high accuracy in detecting persistent MVO compared to the standard LGE images (both sensitivity and specificity =100%, and AUC=97%), with high correspondence in MVO size (R2, 0.97; slope, 0.98, p < 0.01; bias, 0.73±2.11%; p = 0.373; FIG. 15C).

[0088] Discussion

[0089] Viability imaging plays a crucial role in the diagnosis and management of heart diseases. It guides treatment decisions particularly for diseases associated with scar and fiber infiltration. LGE MRI is today’s gold standard for viability imaging. It provides high-resolution images to identify myocardium with impaired contrast washout capability. In this study, quantitative cardiac MRI was used to explore the temporal dynamics of the contrast washout in the myocardium under myocardial infarction. Quantitative dDCE maps to characterize the microcirculation in different territories of the infarcted hearts were successfully derived. The parameters provide reproducible and physiologically meaningful measures of myocardial microcirculation. In addition to the quantitative maps, the ability of the dDCE model to perform viability imaging without the 10-15mins washout time was tested. It was demonstrated that images acquired within 5 minutes post-contrast injection contain sufficient information to portray the whole contrast washout dynamic. The results show that dDCE CMR can provide precise LGE and EGE images without requiring the extended washout time from conventional protocols and is a highly accurate and efficient method for MI and MVO evaluation.

[0090] Existing quantitative DCE CMR only covers a partial view of the myocardial contrast dynamic. FPP MRI is the standard way of DCE MRI adoption in the heart. Pharmacodynamic models have been adopted to measure quantitative flow in the heart during the first passage of the contrast agent. Because FFP is commonly acquired during a single breath-hold, it only covers the wash-in period of the CA dynamic. It provides high sensitivity to the speed of the contrast agent getting into the tissue but has a limited resolution to the contrast clearance ability of the myocardium. The dDCE CMR explores the washout period. It allows a larger contrast dynamic with slower changing contrast concentration.

[0091] The proposed method can address key limitations of the standard viability CMR techniques. Late gadolinium enhancement (LGE) is considered the gold standard for evaluating tissue viability in the heart. However, key limitations have made it unsuitable to certain cases and impeded its wider clinical spread. For instance, standard LGE images measure arbitrarysignal intensity that depends on imaging timing, contrast agent dose, patient hemodynamics, and imaging protocols. This limits its ability to compare between scan sections and to evaluate diffused diseases with no focal lesions. Quantitative approaches, such as extracellular volume (ECV) images, were developed to investigate myocardial fibrosis and extracellular volume quantitatively. However, both LGE and ECV require an extended wait time of 10-15 mins to establish stable signal enhancement in the diseased territories. This results in a prolonged procedure time for fibrosis CMR exams and limited its wider clinical adoption. Although new CMR protocols have been proposed to use the contrast washout period to collect structural images and reduce total scan time, shortened washout time can lead to an impaired contrast difference and result in sub-optimal imaging contrasts. Furthermore, it limits the ability of novel fast CMR sequences to further shorten the CMR protocols. Finally, due to the similar T1 values of blood and lesions at the late enhancement phase, suboptimal contrast in LGE often makes it challenging to delineate small but important subendocardial lesions from the blood pool. Similar limitations are shown in the ECV maps because of its indifference between plasma volume and EES.

[0092] The proposed dDCE CMR on the other hand, is specific to EES. The significant contrast dynamic difference between tissue and blood, provides high sensitivity to subendocardial lesions and is suitable for quantitative evaluation for changing microvascular diseases.

[0093] The proposed dDCE models provide quantitative physiological parameters and can enable novel insights to a wide spectrum of heart diseases. The example demonstrated that the necessary information of Late Gd enhancement can be acquired within 5 minutes post contrast injection. This setup the theoretical foundation for rapid viability acquisitions without the extended wait time. The integrity of the dDCE parameters depend on the dynamic range of the signal progression. The data showed that 5 minutes is sufficient to extract the underlying pathological information in MI for the whole washout process.

[0094] In the current work, an ETK model was adopted to describe contrast progression and reduce the wait time for LGE images. Different models such as two compartment exchange model and other spatial distribution models can be used to extract additional characters of the myocardial microvasculature (e.g., vascular integrity, myocardial blood flow, etc.). In addition, deep learning (DL) models have been adopted for modeling dynamic contrast enhancement in different organism, and their applications in the proposed model for the heart during contrast washout were examined. A U-net model was adopted to synthesize LGE images from a subsetof the data. The DL model can enable rapid LGE without mathematical assumptions and synthesize LGE images.

[0095] Due to the adoption of the standard T1 MOLLI sequence in the current study, single-slice T1 map acquisition with a moderate sampling rate (Isample / minute) was acquired. Adopting advanced imaging techniques such as machine learning, compressed sensing, and Low-Rank tensor techniques, whole heart coverage can be acquired with improved temporal resolution. This can potentially boost the SNR in the dDCE maps and enable quantitative myocardial viability assessment within 5 minutes.

[0096] Because the wash-in period was not monitored in the proposed method, the sensitivity of wash-in parameters (e.g., ktrans) can be lower compared to the standard FPP. Although the results showed reasonable estimation of ktrans, its accuracy need to be further investigated. With fast imaging techniques and free-breathing acquisition, the proposed method can be Combined with the first pass perfusion to reveal the whole progression of contrast dynamic and provide complimentary information for myocardial tissue characterization. InterScan motion can cause error in dDCE map derivations. Hence, a physical based image registration algorithm was adopted to improve the robustness against motion. Further studies are required to validate the dDCE parameters with histological measurements.

[0097] Conclusion

[0098] dDCE CMR provides a quantitative way to represent the myocardial contrast dynamic and provides physiologically meaningful quantitative measurements for tissue characterization. It shows the potential of reducing the 15-minute washout wait time for standard LGE images by more than three folds and resolving the current bottleneck for the prolonged acquisition time of clinical CMR protocols. With advanced imaging acquisitions (e.g., continuous acquisition methods and MR multitasking, compressed sensing, deep learning reconstructions), dDCE CMR can potentially make CMR-based viability assessment a fast, streamlined protocol like their CT counterparts and push CMR for broader adoption in the routine clinical practice.

[0099] Turning to FIGss. 16A-16B, another example illustrating the present approach is provided. Referring specifically to FIG. 16 A, dDCEsomin and dDCEsmin maps for various parameters, derived from Gd contrast-enhanced dynamic are presented, and compared to a conventional LGE image. As visually appreciated, maps between 5 mins and 30 mins and conventional LGE image are consistent. Further, as shown in FIG. 16B, dDCEsmin parameters (PS and ve) measured in the remote and MI demonstrated no statistical difference compared dDCEsOmin.

[0100] In yet another example of the present disclosure, depicted in FIGs. 17A-17C, conventional LGE imaging is compared to dDCE maps generated for acute MI (FIG. 17A), chronic MI (FIG. 17B), and persistent microvascular obstruction (FIG. 17C). Again, no significant difference between conventional LGE imaging and maps derived according to aspects of the present disclosure.

[0101] As shown herein, the present approach provides a quantitative way for rapid and comprehensive myocardial tissue characterization and viability imaging. Among various advantages and improvements, the present approach provides the potential of eliminating prolonged 15 minutes washout time for standard LGE images and appreciably shortening study time of clinical CMR. Combining with advanced imaging acquisitions (e.g., continuous acquisition methods and MR multitasking, compress sensing, deep learning reconstructions), dDCE CMR, according to the present disclosure, can make CMR tissue characterization a fast, streamlined protocol like their CT counterparts and encourage CMR for broader adoption in routine clinical practice.

[0102] According to some embodiments, a method for performing magnetic resonance (MR) imaging is provided. The method includes receiving imaging data associated with a region of interest of a subject prior to an injection of a contrast agent into the region of interest, after the injection of the contrast agent into the region of interest, or both. The method also includes obtaining a plurality of initial contrast-sensitive images of the region of interest of the subject, each of the plurality of initial contrast-sensitive images being indicative of a state of the region of interest at a plurality of times within an initial period of time. The method further includes generating, based on the plurality of initial contrast-sensitive images, a subsequent contrast-sensitive image that is indicative of the state of the region of interest at a subsequent time after the initial period of time ends. In some aspects, the generating includes inputting the plurality of initial contrast sensitive images into a trained model and receiving the subsequent contrast-sensitive image from the trained model. In some aspects, the trained model includes a deep learning model. In some aspects, the initial period of time begins prior to the injection of the contrast agent and ends after the injection of the contrast agent. In some aspects, the initial period of time begins (i) prior to the injection of the contrast agent, (ii) at a time when the injection of the contrast agent occurs, or (iii) after the injection of the contrast agent. In some aspects, the initial period of time ends after the injection of the contrast agent. In some aspects, the initial period of time ends within about five minutes after the injection of the contrast agent. In some aspects, the subsequent time is about ten minutes after the initial period of time ends, more than about fifteen minutes after the injection of the contrast agent, or both.In some aspects, the plurality of initial contrast-sensitive images includes one or more maps of a concentration of the contrast agent within the region of interest. In some aspects, the plurality of initial contrast-sensitive images includes one or more T1 maps, one or more T2 maps, one or more T2* maps, or any combination thereof. In some aspects, the plurality of initial contrastsensitive images includes at least one pre-contrast contrast-sensitive image obtained prior to the injection of the contrast agent, and at least one post-contrast contrast-sensitive image obtained after the injection of the contrast agent. In some aspects, the plurality of initial contrast-sensitive images includes at least one pre-contrast T1 map obtained prior to the injection of the contrast agent, and at least one post-contrast T1 map obtained after the injection of the contrast agent. In some aspects, the subsequent contrast-sensitive image includes a late gadolinium enhancement (LGE) image. In some aspects, the subsequent contrast-sensitive image is generated based only on the plurality of initial contrast-sensitive images. In some aspects, the method includes generating, based on the plurality of initial contrast-sensitive images, one or more dynamic contrast enhanced (DCE) parameters associated with the region of interest. In some aspects, the one or more DCE parameters include (i) a volume transfer constant ktrans for the contrast agent between blood plasma in the region of interest and an extravascular extracellular space (EES) in the region of interest, (ii) a time constant (kep) for the contrast agent reflux from the EES back into a portion of a vascular system of the subject in the region of interest, (iii) a volume (ve) of the EES as a percentage of a volume of the region of interest, (iv) a volume (vp) of the blood plasma as a percentage of the volume of the region of interest, or (v) any combination of (i)-(iv). In some aspects, the one or more DCE parameters are indicative of the state of the region of interest at (i) at least one of the plurality of times within the initial period of time, (ii) the subsequent time after the initial period of time ends, (iii) a point in time that is after an end of the first period of time and before the subsequent time, or (iv) any combination of (i)-(iii). In some aspects, the contrast agent is a gadolinium- based contrast agent. In some aspects, the region of interest includes a chest cavity of the subject, a heart of the subject, a myocardium of the heart of the subject, or any combination thereof.

[0103] According to some embodiments, a system comprising a control system is provided, configured to implement a method for performing magnetic resonance (MR) imaging is provided, as described. The method includes receiving imaging data associated with a region of interest of a subject prior to an injection of a contrast agent into the region of interest, after the injection of the contrast agent into the region of interest, or both. The method also includes obtaining a plurality of initial contrast-sensitive images of the region of interest of the subject,each of the plurality of initial contrast-sensitive images being indicative of a state of the region of interest at a plurality of times within an initial period of time. The method further includes generating, based on the plurality of initial contrast-sensitive images, a subsequent contrastsensitive image that is indicative of the state of the region of interest at a subsequent time after the initial period of time ends. In some aspects, the generating includes inputting the plurality of initial contrast sensitive images into a trained model and receiving the subsequent contrastsensitive image from the trained model. In some aspects, the trained model includes a deep learning model. In some aspects, the initial period of time begins prior to the injection of the contrast agent and ends after the injection of the contrast agent. In some aspects, the initial period of time begins (i) prior to the injection of the contrast agent, (ii) at a time when the injection of the contrast agent occurs, or (iii) after the injection of the contrast agent. In some aspects, the initial period of time ends after the injection of the contrast agent. In some aspects, the initial period of time ends within about five minutes after the injection of the contrast agent. In some aspects, the subsequent time is about ten minutes after the initial period of time ends, more than about fifteen minutes after the injection of the contrast agent, or both. In some aspects, the plurality of initial contrast-sensitive images includes one or more maps of a concentration of the contrast agent within the region of interest. In some aspects, the plurality of initial contrast-sensitive images includes one or more T1 maps, one or more T2 maps, one or more T2* maps, or any combination thereof. In some aspects, the plurality of initial contrastsensitive images includes at least one pre-contrast contrast-sensitive image obtained prior to the injection of the contrast agent, and at least one post-contrast contrast-sensitive image obtained after the injection of the contrast agent. In some aspects, the plurality of initial contrast-sensitive images includes at least one pre-contrast T1 map obtained prior to the injection of the contrast agent, and at least one post-contrast T1 map obtained after the injection of the contrast agent. In some aspects, the subsequent contrast-sensitive image includes a late gadolinium enhancement (LGE) image. In some aspects, the subsequent contrast-sensitive image is generated based only on the plurality of initial contrast-sensitive images. In some aspects, the method includes generating, based on the plurality of initial contrast-sensitive images, one or more dynamic contrast enhanced (DCE) parameters associated with the region of interest. In some aspects, the one or more DCE parameters include (i) a volume transfer constant ktrans for the contrast agent between blood plasma in the region of interest and an extravascular extracellular space (EES) in the region of interest, (ii) a time constant (kep) for the contrast agent reflux from the EES back into a portion of a vascular system of the subject in the region of interest, (iii) a volume (ve) of the EES as a percentage of a volume of the regionof interest, (iv) a volume (vp) of the blood plasma as a percentage of the volume of the region of interest, or (v) any combination of (i)-(iv). In some aspects, the one or more DCE parameters are indicative of the state of the region of interest at (i) at least one of the plurality of times within the initial period of time, (ii) the subsequent time after the initial period of time ends, (iii) a point in time that is after an end of the first period of time and before the subsequent time, or (iv) any combination of (i)-(iii). In some aspects, the contrast agent is a gadolinium- based contrast agent. In some aspects, the region of interest includes a chest cavity of the subject, a heart of the subject, a myocardium of the heart of the subject, or any combination thereof.

[0104] In some embodiments, a computer program product comprising instructions which, when executed by a computer, cause the computer to carry out a method, as described. The method includes receiving imaging data associated with a region of interest of a subject prior to an injection of a contrast agent into the region of interest, after the injection of the contrast agent into the region of interest, or both. The method also includes obtaining a plurality of initial contrast-sensitive images of the region of interest of the subject, each of the plurality of initial contrast-sensitive images being indicative of a state of the region of interest at a plurality of times within an initial period of time. The method further includes generating, based on the plurality of initial contrast-sensitive images, a subsequent contrast-sensitive image that is indicative of the state of the region of interest at a subsequent time after the initial period of time ends. In some aspects, the generating includes inputting the plurality of initial contrast sensitive images into a trained model and receiving the subsequent contrast-sensitive image from the trained model. In some aspects, the trained model includes a deep learning model. In some aspects, the initial period of time begins prior to the injection of the contrast agent and ends after the injection of the contrast agent. In some aspects, the initial period of time begins (i) prior to the injection of the contrast agent, (ii) at a time when the injection of the contrast agent occurs, or (iii) after the injection of the contrast agent. In some aspects, the initial period of time ends after the injection of the contrast agent. In some aspects, the initial period of time ends within about five minutes after the injection of the contrast agent. In some aspects, the subsequent time is about ten minutes after the initial period of time ends, more than about fifteen minutes after the injection of the contrast agent, or both. In some aspects, the plurality of initial contrast-sensitive images includes one or more maps of a concentration of the contrast agent within the region of interest. In some aspects, the plurality of initial contrast-sensitive images includes one or more T1 maps, one or more T2 maps, one or more T2* maps, or any combination thereof. In some aspects, the plurality of initial contrast-sensitive images includesat least one pre-contrast contrast-sensitive image obtained prior to the injection of the contrast agent, and at least one post-contrast contrast-sensitive image obtained after the injection of the contrast agent. In some aspects, the plurality of initial contrast-sensitive images includes at least one pre-contrast T1 map obtained prior to the injection of the contrast agent, and at least one post-contrast T1 map obtained after the injection of the contrast agent. In some aspects, the subsequent contrast-sensitive image includes a late gadolinium enhancement (LGE) image. In some aspects, the subsequent contrast-sensitive image is generated based only on the plurality of initial contrast-sensitive images. In some aspects, the method includes generating, based on the plurality of initial contrast-sensitive images, one or more dynamic contrast enhanced (DCE) parameters associated with the region of interest. In some aspects, the one or more DCE parameters include (i) a volume transfer constant ktrans for the contrast agent between blood plasma in the region of interest and an extravascular extracellular space (EES) in the region of interest, (ii) a time constant (kep) for the contrast agent reflux from the EES back into a portion of a vascular system of the subject in the region of interest, (iii) a volume (ve) of the EES as a percentage of a volume of the region of interest, (iv) a volume (vp) of the blood plasma as a percentage of the volume of the region of interest, or (v) any combination of (i)-(iv). In some aspects, the one or more DCE parameters are indicative of the state of the region of interest at (i) at least one of the plurality of times within the initial period of time, (ii) the subsequent time after the initial period of time ends, (iii) a point in time that is after an end of the first period of time and before the subsequent time, or (iv) any combination of (i)-(iii). In some aspects, the contrast agent is a gadolinium-based contrast agent. In some aspects, the region of interest includes a chest cavity of the subject, a heart of the subject, a myocardium of the heart of the subject, or any combination thereof. In some aspects, the computer program product is a non- transitory computer readable medium.

[0105] In some embodiments, a system is provided. The system includes a memory device having stored thereon machine-readable instructions and a control system including one or more processors configured to execute the machine-readable instructions to receive imaging data associated with a region of interest of a subject prior to an injection of a contrast agent into the region of interest, after the injection of the contrast agent into the region of interest, or both, obtain a plurality of initial contrast-sensitive images of the region of interest of the subject, each of the plurality of initial contrast-sensitive images being indicative of a state of the region of interest at a plurality of times within an initial period of time, and generate, based on the plurality of initial contrast-sensitive images, a subsequent contrast-sensitive image that is indicative of the state of the region of interest at a subsequent time after the initial period oftime ends. In some aspects, the one or more processors of the control system are further configured to execute the machine-readable instructions to perform a method, as described. The method includes receiving imaging data associated with a region of interest of a subject prior to an injection of a contrast agent into the region of interest, after the injection of the contrast agent into the region of interest, or both. The method also includes obtaining a plurality of initial contrast-sensitive images of the region of interest of the subject, each of the plurality of initial contrast-sensitive images being indicative of a state of the region of interest at a plurality of times within an initial period of time. The method further includes generating, based on the plurality of initial contrast-sensitive images, a subsequent contrast-sensitive image that is indicative of the state of the region of interest at a subsequent time after the initial period of time ends. In some aspects, the generating includes inputting the plurality of initial contrast sensitive images into a trained model and receiving the subsequent contrast-sensitive image from the trained model. In some aspects, the trained model includes a deep learning model. In some aspects, the initial period of time begins prior to the injection of the contrast agent and ends after the injection of the contrast agent. In some aspects, the initial period of time begins (i) prior to the injection of the contrast agent, (ii) at a time when the injection of the contrast agent occurs, or (iii) after the injection of the contrast agent. In some aspects, the initial period of time ends after the injection of the contrast agent. In some aspects, the initial period of time ends within about five minutes after the injection of the contrast agent. In some aspects, the subsequent time is about ten minutes after the initial period of time ends, more than about fifteen minutes after the injection of the contrast agent, or both. In some aspects, the plurality of initial contrast-sensitive images includes one or more maps of a concentration of the contrast agent within the region of interest. In some aspects, the plurality of initial contrast-sensitive images includes one or more T1 maps, one or more T2 maps, one or more T2* maps, or any combination thereof. In some aspects, the plurality of initial contrast-sensitive images includes at least one pre-contrast contrast-sensitive image obtained prior to the injection of the contrast agent, and at least one post-contrast contrast-sensitive image obtained after the injection of the contrast agent. In some aspects, the plurality of initial contrast-sensitive images includes at least one pre-contrast T1 map obtained prior to the injection of the contrast agent, and at least one post-contrast T1 map obtained after the injection of the contrast agent. In some aspects, the subsequent contrast-sensitive image includes a late gadolinium enhancement (LGE) image. In some aspects, the subsequent contrast-sensitive image is generated based only on the plurality of initial contrast-sensitive images. In some aspects, the method includes generating, based on the plurality of initial contrast-sensitive images, one or more dynamic contrast enhanced (DCE)parameters associated with the region of interest. In some aspects, the one or more DCE parameters include (i) a volume transfer constant ktrans for the contrast agent between blood plasma in the region of interest and an extravascular extracellular space (EES) in the region of interest, (ii) a time constant (kep) for the contrast agent reflux from the EES back into a portion of a vascular system of the subject in the region of interest, (iii) a volume (ve) of the EES as a percentage of a volume of the region of interest, (iv) a volume (vp) of the blood plasma as a percentage of the volume of the region of interest, or (v) any combination of (i)-(iv). In some aspects, the one or more DCE parameters are indicative of the state of the region of interest at (i) at least one of the plurality of times within the initial period of time, (ii) the subsequent time after the initial period of time ends, (iii) a point in time that is after an end of the first period of time and before the subsequent time, or (iv) any combination of (i)-(iii). In some aspects, the contrast agent is a gadolinium-based contrast agent. In some aspects, the region of interest includes a chest cavity of the subject, a heart of the subject, a myocardium of the heart of the subject, or any combination thereof.

[0001] One or more elements or aspects or steps, or any portion(s) thereof, from one or more of any of claims below can be combined with one or more elements or aspects or steps, or any portion(s) thereof, from one or more of any of the other claims or combinations thereof, to form one or more additional implementations and / or claims of the present disclosure.

[0106] One or more elements or aspects or steps, or any portion(s) thereof, from one or more of any of the claims can be combined with one or more elements or aspects or steps, or any portion(s) thereof, from one or more of any of the other claims or combinations thereof, to form one or more additional implementations and / or claims of the present disclosure.

[0107] While the present disclosure has been described with reference to one or more particular embodiments or implementations, those skilled in the art will recognize that many changes may be made thereto without departing from the spirit and scope of the present disclosure. Each of these implementations and obvious variations thereof is contemplated as falling within the spirit and scope of the present disclosure. It is also contemplated that additional implementations according to aspects of the present disclosure may combine any number of features from any of the implementations described herein.

Claims

CLAIMSWHAT IS CLAIMED IS:

1. A method for performing magnetic resonance (MR) imaging, the method comprising: receiving imaging data associated with a region of interest of a subject prior to an injection of a contrast agent into the region of interest, after the injection of the contrast agent into the region of interest, or both; obtaining a plurality of initial contrast-sensitive images of the region of interest of the subject, each of the plurality of initial contrast-sensitive images being indicative of a state of the region of interest at a plurality of times within an initial period of time; and generating, based on the plurality of initial contrast-sensitive images, a subsequent contrast-sensitive image that is indicative of the state of the region of interest at a subsequent time after the initial period of time ends.

2. The method of claim 1, wherein the generating includes: inputting the plurality of initial contrast sensitive images into a trained model; and receiving the subsequent contrast-sensitive image from the trained model.

3. The method of claim 2, wherein the trained model includes a deep learning model.

4. The method of claim 1, wherein the initial period of time begins prior to the injection of the contrast agent and ends after the injection of the contrast agent.

5. The method of claim 1 , wherein the initial period of time begins (i) prior to the inj ection of the contrast agent, (ii) at a time when the injection of the contrast agent occurs, or (iii) after the injection of the contrast agent.

6. The method of claim 1, wherein the initial period of time ends after the injection of the contrast agent.

7. The method of claim 1 , wherein the initial period of time ends within about five minutes after the injection of the contrast agent.

8. The method of claim 1, wherein the subsequent time is about ten minutes after the initial period of time ends, more than about fifteen minutes after the injection of the contrast agent, or both.

9. The method of claim 1 , wherein the plurality of initial contrast-sensitive images include one or more maps of a concentration of the contrast agent within the region of interest.

10. The method of claim 1, wherein the plurality of initial contrast-sensitive images includes one or more T1 maps, one or more T2 maps, one or more T2* maps, or any combination thereof.

11. The method of claim 1, wherein the plurality of initial contrast-sensitive images includes at least one pre-contrast contrast-sensitive image obtained prior to the injection of the contrast agent, and at least one post-contrast contrast-sensitive image obtained after the injection of the contrast agent.

12. The method of claim 1, wherein the plurality of initial contrast-sensitive images includes at least one pre-contrast T1 map obtained prior to the injection of the contrast agent, and at least one post-contrast T1 map obtained after the injection of the contrast agent.

13. The method of claim 1, wherein the subsequent contrast-sensitive image includes a late gadolinium enhancement (LGE) image.

14. The method of claim 1, wherein the subsequent contrast-sensitive image is generated based only on the plurality of initial contrast-sensitive images.

15. The method of claim 1, further comprising generating, based on the plurality of initial contrast-sensitive images, one or more dynamic contrast enhanced (DCE) parameters associated with the region of interest.

16. The method of claim 15, wherein the one or more DCE parameters include (i) a volume transfer constant ktrans for the contrast agent between blood plasma in the region of interest and an extravascular extracellular space (EES) in the region of interest, (ii) a time constant (keP) for the contrast agent reflux from the EES back into a portion of a vascular system of the subjectin the region of interest, (iii) a volume (ve) of the EES as a percentage of a volume of the region of interest, (iv) a volume (vp) of the blood plasma as a percentage of the volume of the region of interest, or (v) any combination of (i)-(iv).

17. The method of claim 15, wherein the one or more DCE parameters are indicative of the state of the region of interest at (i) at least one of the plurality of times within the initial period of time, (ii) the subsequent time after the initial period of time ends, (iii) a point in time that is after an end of the first period of time and before the subsequent time, or (iv) any combination of (i)-(iii).

18. The method of claim 1, wherein the contrast agent is a gadolinium-based contrast agent.

19. The method of claim 1, wherein the region of interest includes a chest cavity of the subject, a heart of the subject, a myocardium of the heart of the subject, or any combination thereof.

20. A system comprising a control system configured to: receive imaging data associated with a region of interest of a subject prior to an injection of a contrast agent into the region of interest, after the injection of the contrast agent into the region of interest, or both; obtain a plurality of initial contrast-sensitive images of the region of interest of the subject, each of the plurality of initial contrast-sensitive images being indicative of a state of the region of interest at a plurality of times within an initial period of time; and generate, based on the plurality of initial contrast-sensitive images, a subsequent contrast-sensitive image that is indicative of the state of the region of interest at a subsequent time after the initial period of time ends.

21. A computer program product comprising instructions which, when executed by a computer, cause the computer to: receive imaging data associated with a region of interest of a subject prior to an injection of a contrast agent into the region of interest, after the injection of the contrast agent into the region of interest, or both;obtain a plurality of initial contrast-sensitive images of the region of interest of the subject, each of the plurality of initial contrast-sensitive images being indicative of a state of the region of interest at a plurality of times within an initial period of time; and generate, based on the plurality of initial contrast-sensitive images, a subsequent contrast-sensitive image that is indicative of the state of the region of interest at a subsequent time after the initial period of time ends.

22. The computer program product of claim 21, wherein the computer program product is a non-transitory computer readable medium.

23. A system comprising: a memory device having stored thereon machine-readable instructions; and a control system including one or more processors configured to execute the machine- readable instructions to: receive imaging data associated with a region of interest of a subject prior to an injection of a contrast agent into the region of interest, after the injection of the contrast agent into the region of interest, or both; obtain a plurality of initial contrast-sensitive images of the region of interest of the subject, each of the plurality of initial contrast-sensitive images being indicative of a state of the region of interest at a plurality of times within an initial period of time; and generate, based on the plurality of initial contrast-sensitive images, a subsequent contrast-sensitive image that is indicative of the state of the region of interest at a subsequent time after the initial period of time ends.

24. The system of claim 23, wherein the one or more processors of the control system are further configured to execute the machine-readable instructions to: receive imaging data associated with a region of interest of a subject prior to an injection of a contrast agent into the region of interest, after the injection of the contrast agent into the region of interest, or both; obtain a plurality of initial contrast-sensitive images of the region of interest of the subject, each of the plurality of initial contrast-sensitive images being indicativeof a state of the region of interest at a plurality of times within an initial period of time; and generate, based on the plurality of initial contrast-sensitive images, a subsequent contrast-sensitive image that is indicative of the state of the region of interest at a subsequent time after the initial period of time ends.

Images