System and methods for standardized measurements and personalized parameter maps of the thoracic aorta from anatomic imaging using machine learning

Abstract

The present disclosure provides a system and method for analyzing an aorta of a subject using medical images. A machine learning algorithm is applied to provide aorta analysis based on the medical images. The aorta analysis includes standardized measurements of the aorta, which are mapped to a common aorta model to create personalized (e.g., subject specific) aorta heatmaps and / or to derive an aorta risk score for the subject.

Description

SYSTEM AND METHODS FOR STANDARDIZED MEASUREMENTS AND PERSONALIZED PARAMETER MAPS OF THE THORACIC AORTA FROM ANATOMIC IMAGING USING MACHINE LEARNINGCROSS-REFERENCE TO RELATED APPLICATIONS

[0001] This application claims the benefit of U.S. Provisional Patent Application Serial No. 63 / 596,134, filed on November 3, 2023, and entitled “SYSTEM AND METHODS FOR STANDARDIZED MEASUREMENTS AND PERSONALIZED PARAMETER MAPS OF THE THORACIC AORTA FROM ANATOMIC IMAGING USING MACHINE LEARNING / ’ which is herein incorporated by reference in its entirety.BACKGROUND

[0002] Aorta dimensions and growth are the cunent standard of care for riskstratification of patients with thoracic aorta disease. Patients ty pically require life-long imaging surveillance, usually performed every 6-12 months. Current clinical risk assessment is based on measurements of aortic dimensions and growth rates performed on aorta images acquired with either computed tomography (CT) angiography (CTA) or magnetic resonance (MR) angiography (MRA). These measurements are typically made manually, which is time consuming. Moreover, differences in measurement tools and the individual observer performing the measurements can lead to clinically relevant inconsistencies. In addition, many patients receive mixed modality follow-up imaging with CTA and other follow-ups with MRA, which can result in measurement modality based differences that may or may not be clinically meaningful. Thus, there is a need for consistent measurement tools. Moreover, the current paradigm for risk assessment in patient with thoracic aortic aneurysm (TAA) patients is based on primitive size thresholds (e.g., aortic diameter > 5.0 cm, aorta growth rates > 0.3 cm / year) taken at isolated anatomic locations, which has poor predictive value. Thus, new methods are needed to provide standardized and personalized aorta assessment.SUMMARY OF THE DISCLOSURE

[0003] It is an aspect of the present disclosure to provide a method for generating standardized aorta measurements and personalized (i.e., subject-specific or patient-specific) aorta heatmaps indicative of measurement parameters. Such aorta heatmaps can provide visualization and quantification of aortic regions, including at least one of absolutemeasurements or relative measurements, changes over time, and comparisons with an appropriate control population. The method includes accessing computed tomography (CT) or magnetic resonance imaging (MRI) data acquired from a subject. A machine learning model, trained to generate aorta analysis is applied to the CT or MRI data. Aorta analysis, which includes standardized aorta measurements and personalized (i.e., subject-specific or patientspecific) aorta heatmaps, is generated from the medical images in an automated way.

[0004] The present disclosure further provides a system for generating aorta analysis. The system includes a computer processor that is configured to access CT or MRI images from a subject and a machine learning model trained to generate aorta analysis. The computer processor is further configured to apply the CT or MRI data to the machine learning model to generate aorta analysis of the subject, where the aorta analysis includes standardized aorta measurements and personalized (i.e., subject-specific or patient-specific) aorta heatmaps.

[0005] The present disclosure also provides a method for generating an automated system for analyzing aortas in medical images. The method includes accessing training data using a computer system. The training data includes labeled CT or MRI images from a group of subjects. The method further includes using a computer system to train a machine learning model on the training data. The machine learning model is trained to generate aorta analysis, including standardized aorta measurements and personalized (i.e., subject-specific or patientspecific) aorta heatmaps. The machine learning model is stored by a computer system for future use.BRIEF DESCRIPTION OF THE DRAWINGS

[0006] FIG. l is a flowchart setting forth the steps of an example method for generating aorta analysis in a subject by applying magnetic resonance imaging (“MRI”) or computed tomography (“CT”) images to a machine learning model.

[0007] FIG. 2 is a flowchart setting forth the steps of an example method for training a machine learning model to generate aorta analysis based on MRI or CT data obtained from a subject.

[0008] FIG. 3 is a block diagram of an example aorta analysis system that can implement the methods of the present disclosure.

[0009] FIG. 4 is a block diagram of example components that can implement the system of FIG. 3.

[0010] FIG. 5 is a block diagram of an example MRI system that can implement the methods described in the present disclosure.

[0011] FIG. 6A illustrates an example CT system that can implement the methods described in the present disclosure.

[0012] FIG. 6B illustrates an example CT system that can implement the methods described in the present disclosure.

[0013] FIG. 7A illustrates an example clinical workflow for generating aorta analysis from MRA and / or CTA input data, including standardized aorta measurements and a personalized aorta heatmap, in accordance with the present disclosure.

[0014] FIG. 7B illustrates an example clinical workflow for generating standardized aorta measurements from MRA and / or CTA input data and a standardized report in accordance with the present disclosure.

[0015] FIG. 7C illustrates an example clinical workflow for generating a personalized aorta heatmap from MRA and / or CTA input data that provides individual 3D aortic dimension references in comparison to an appropriate control population, in accordance with the present disclosure.DETAILED DESCRIPTION

[0016] Described here are methods for generating standardized aorta measurements and personalized (i.e., subject-specific or patient-specific) aorta heatmaps from computed tomography (CT) and / or magnetic resonance imaging (MRI) data of the chest. Standardized aorta measurements can be mapped onto a common aortic model using a suitably trained machine learning model. Such methods can mitigate potential sources of error, such as differences between scan type (e.g., CT, MRI), acquisition methods, measurement techniques, and others. The proposed methods provide several advantages to current techniques. Mapping measurements onto a common aortic model can ensure consistency and facilitate direct comparisons between imaging modalities across multiple surveillance imaging scans. The automated process significantly reduces the time required for manual measurements, enabling faster and more efficient analysis of aortic parameters as well as improved reproducibility7. In addition, the mapped aorta 3D imaging data can be referenced or compared to an appropriate control population (e.g., matched for age. sex, race, ethnicity, risk profile, etc.) to generate a personalized aorta heatmap. The aorta heatmap can be used to visualize significant differences in regional aortic dimensions compared to the control population and quantify regional and / orglobal changes (e.g., percentage of the aorta with significant aortic dilatation) in the individual patient. Leveraging advanced machine learning techniques, the methods described in the present disclosure provide accurate and precise measurements, reducing human errors and inter-observer variability. Combining standardized aortic measurements and aorta heatmap data with additional demographic and clinical information obtained from the patient’s electronic medical record (EMR) can also provide improved risk assessment for aortic complications. The proposed method is able to extract multiple relevant metrics, allowing for comprehensive assessment of the thoracic aorta that can aid in the diagnosis, monitoring, and treatment planning of cardiovascular diseases.

[0017] As will be described, the present disclosure uses machine learning (e.g., deep learning or other machine learning techniques, models, or frameworks) to analyze MRI or CT medical images (e.g., MRI or CT) of the thoracic aorta. The system is able to perform 1) aorta 3D segmentation and identification of standardized anatomic locations, 2) extraction of aorta morphologic measurements (e.g., diameter, circumference, lumen area, volume, length, curvature) at standardized anatomic locations, 3) mapping of measurements onto a common aortic model to provide consistent comparison between imaging modalities (e.g., CT and MR) and longitudinal assessment (e.g., calculate aorta growth rates, measure changes in aorta morphologic and / or hemodynamic measurements across serial acquisitions), 4) visualization of an aorta model, which may be labeled with measurements and allow for user interaction, 5) creating a standardized report of aorta measurements, 6) calculation of personalized aorta heatmaps by referencing the mapped aorta data to an appropriate control population (e.g., matched for age, sex, race ethnicity’, risk profile, etc.), 7) visualization of significant differences in regional aortic dimensions compared to the control population and quantification of regional and global changes (e.g., % of the aorta with significant aortic dilatation). 8) combination of standardized aortic measurements and aorta heatmap data with additional demographic and clinical information obtained from the patient’s electronic medical record (EMR) to provide an individualized risk score for aortic complications.

[0018] Referring now to FIG. 1. a flowchart is illustrated as setting forth the steps of an example method for generating standardized aortic measurements using a suitably trained neural network or other machine learning algorithm. As will be described, the neural network or other machine learning algorithm takes imaging data as input data and generates standardized aortic measurements and analysis as output data. As an example, the standardized aortic measurements can be indicative of aorta diameters and / or volumes of various anatomiclandmarks, locations, or regions along the aorta. The standardized aortic measurements may further include personalized aorta heatmaps, longitudinal changes of aorta size or shape, risk scores for aortic complications, and so on. In some implementations, the neural network may also be applied to provide standardized measurements and analysis of other vasculature, such as the abdominal aorta, inferior vena cava, superior vena cava, brachiocephalic veins, pulmonary veins, pulmonary arteries, right coronary artery, brachiocephalic artery, right subclavian artery, right common carotid artery, left common artery, left subclavian artery, other vasculature, or a combination thereof.

[0019] The method includes accessing imaging data with a computer system, as indicated at step 102. Such medical imaging data may include anatomical images acquired with CT or MRI, with or without contrast. As a non-limiting example, the imaging data may be anatomical MRI angiographic images of the chest, such as Ti-weighted 3D gradient echo images acquired after the administration of a contrast agent (e.g., gadolinium). Accessing the imaging data may include retrieving such data from a memory or other suitable data storage device or medium such as a picture archiving system (PACS) and vendor neural archive (VNA). Additionally or alternatively, accessing the imaging data may include acquiring such data with a MRI or CT system and transferring or otherwise communicating the data to the computer system, which may be a part of the aorta analysis system.

[0020] In some embodiments, in addition to imaging data, clinical data may also be accessed by the computer system. For example, the clinical data may include patient demographics, genetic risk data, clinical history data, or other clinically relevant data about the subject’s current state, clinical history, genetic profile, or family history.

[0021] A trained neural network (or other suitable machine learning algorithm) is then accessed with the computer system, as indicated at step 104. In general, the neural network is trained, or has been trained, on training data in order to provide analysis of the aorta. Aorta analysis may include recognizing and measuring the relevant anatomical structures, deriving personalized (i.e., subject-specific or patient-specific) aorta heatmaps, and characterizing changes in aortic dimensions over time (e.g.. growth rate) of the aorta. The neural network may be trained to segment the aorta and to delineate anatomic landmarks based on a common aortic model.

[0022] The neural network may be trained to provide automated 3D segmentation, generating a 3D segmentation that identifies and delineates the boundaries of thoracic aorta. The neural network may also identify anatomic landmarks or features within the segmentedaorta to provide standardized reference points along the length of the aorta, as shown in FIGS. 7A, 7B and 7C, for example. As a non-limiting example, the anatomic landmarks may include anatomic locations, such as the right and left coronary arteries; the sinotubular junction; the proximal, mid, and distal ascending aorta; the brachiocephalic artery; the left common carotid artery; the left subclavian artery; the proximal, mid and distal descending aorta; and so on. The anatomic landmarks may also include regions and / or volumes defined by anatomic locations (e.g., a volume between two anatomic locations, a cross-sectional area at an anatomic location), aorta zones, or cross-sectional areas along the aorta. The anatomic landmarks may also include standard regions of the aorta and related vessels based on defined relationships between anatomic features. Regions may include the ascending aorta, aortic arch, descending aorta, thoracic aorta, abdominal aorta, suprarenal abdominal aorta, infrarenal aorta, right coronary artery, brachiocephalic artery, right subclavian artery, right common carotid artery, left common carotid artery, left subclavian artery, and so on. For example, the ascending aorta region may be defined from the sinotubular junction to the brachiocephalic trunk (TB), as shown in FIG. 7B. As another non-limiting example, the location of the proximal arch may be defined using the anatomic landmark of the brachiocephalic trunk (labeled TB in FIG. 7B. In some implementations, the neural network may be trained to segment and analyze the superior vena cava, inferior vena cava, pulmonary7artery, and / or pulmonary vein.

[0023] Once the aortic region is accurately segmented and landmarks have been identified, the neural network can extract measurements of interest within each anatomic landmark. These measurements may include but are not limited to aortic diameter, length, lumen area, circumference, volume, shape, curvature, width to height aspect ratio, and other relevant metrics. The metrics may be measured for the full aorta, each individual anatomic landmark, or a combination of anatomic landmark. Additionally or alternatively, the measurements can include measurements of hemodynamic parameters, such as blood pressure (e.g., systolic blood pressure, diastolic blood pressure, mean arterial blood pressure), systolic pressure variation, pulse pressure variation, right arterial pressure, right ventricular pressure, pulmonary artery7pressure, mean pulmonary artery pressure, pulmonary vascular resistance, systemic vascular resistance, central venous pressure, stroke volume, stroke volume variation, stroke volume index, cardiac output, cardiac index, and the like.

[0024] The neural network can be trained to align and map the extracted measurements onto a common aortic model. For example, the common model may represent the idealized anatomy of the thoracic aorta as shown in FIGS. 7 A and 7B. incorporating anatomic landmarksfor consistent measurement comparisons. By mapping measurements onto this model, the software tool enables clinicians and researchers to compare and evaluate aortic measurements obtained from different imaging modalities and multiple surveillance imaging measurements conducted over time in a standardized manner. Measurements may be compared between successively obtained imaging data to quantify changes in aortic measurements such as growth rates at anatomic landmarks, changes in aortic volumes, and so on.

[0025] A separate neural network can be trained to map the segmented 3D aorta geometry to an appropriate control population (matched for age, sex, race ethnicity, risk profile, etc.) to generate a personalized (i.e., subject-specific or patient-specific) aorta heatmap as shown in FIGS. 7A and 7C. The aorta heatmap may be used to quantify significant differences in aortic dimensions compared to the control population and / or quantify regional and global changes, e.g., % of the aorta with significant aortic dilatation, aortic surface area with 2 standard deviations (STDEV) outside the aortic dimensions of the normal control population, and so on.

[0026] Accessing the trained neural network may include accessing network and / or training parameters (e.g., weights, biases, number of layers, training epochs, or a combination of these) that have been optimized or otherwise estimated by training the neural network on training data. In some instances, retrieving the neural network can also include retrieving, constructing, or otherwise accessing the particular neural network architecture to be implemented. For instance, data pertaining to the layers in the neural network architecture (e.g.. number of layers, type of layers, ordering of layers, connections between layers, hyperparameters for layers) may be retrieved, selected, constructed, or otherwise accessed.

[0027] A neural network generally includes an input layer, one or more hidden layers (or nodes), and an output layer. Typically, the input layer includes as many nodes as inputs provided to the artificial neural network. The number (and the type) of inputs provided to the artificial neural network may vary based on the particular task for the artificial neural network.

[0028] The input layer connects to one or more hidden layers. The number of hidden layers varies and may depend on the particular task for the artificial neural network. Additionally, each hidden layer may have a different number of nodes and may be connected to the next layer differently. For example, each node of the input layer may be connected to each node of the first hidden layer. The connection between each node of the input layer and each node of the first hidden layer may be assigned a weight parameter. Additionally, each node of the neural network may also be assigned a bias value. In some configurations, eachnode of the first hidden layer may not be connected to each node of the second hidden layer. That is, there may be some nodes of the first hidden layer that are not connected to all of the nodes of the second hidden layer. The connections between the nodes of the first hidden layers and the second hidden layers are each assigned different weight parameters. Each node of the hidden layer is generally associated with an activation function. The activation function defines how the hidden layer is to process the input received from the input layer or from a previous input or hidden layer. These activation functions may van- and be based on the type of task associated with the artificial neural network and also on the specific type of hidden layer implemented.

[0029] Each hidden layer may perform a different function. For example, some hidden layers can be convolutional hidden layers which can. in some instances, reduce the dimensionality of the inputs. Other hidden layers can perform statistical functions such as max pooling, which may reduce a group of inputs to the maximum value; an averaging layer; batch normalization; and other such functions. In some of the hidden layers each node is connected to each node of the next hidden layer, which may be referred to then as dense layers. Some neural networks including more than, for example, three hidden layers may be considered deep neural networks.

[0030] The last hidden layer in the artificial neural network is connected to the output layer. Similar to the input layer, the output layer typically has the same number of nodes as the possible outputs. In an example in which the artificial neural network analyzes medical images of the aorta, the output layer may include, for example, a number of different nodes, where each different node corresponds to a different analysis. In some implementations, a first node may segment the aorta to provide a 3D model of the full aorta, a second node may identify landmarks or anatomic locations of the aorta, a third node may segment the 3D aorta model to provide a 3D model of aortic sub-volumes, a fourth node may provide subject-specific measurements of various features or landmarks of the aorta, a fifth node may determine a mapping of subject-specific measurements onto a common aorta model, a sixth node may generate a personalized aortic heatmap or parameter maps, which may be indicative of subjectspecific measurements or parameters, and a seventh node may determine a risk score based on the subject-specific measurements, personalized heatmap, and / or additional demographic and clinical patient factors.

[0031] The medical images are then used as input data to the one or more trained neural networks, generating aorta analysis data as an output, as indicated at step 106. For example.the aorta analysis data may include 3D segmentation of the aorta, or a 3D segmentation of various anatomic landmarks of the aorta. The output aorta analysis data may also include measurements of various features of the aorta. Such features may include the diameter, lumen area, length, circumference, volume, shape, curvature, width to height aspect ratio, or other relevant metrics of the aorta. Similarly, the features may include the length, circumference, volume, shape, curvature, width to height aspect ratio, or other relevant metrics of each of the anatomic landmarks of the aorta or comparison between them. The features may further include a calculation of longitudinal changes, such as aorta growth rates. These metrics can be compared to population norms to provide clinical insight.

[0032] The aorta analysis data generated in block 106 may also include a visualization of such the model outputs, such as 3D segmentation, standardized measurements mapped onto a common aortic model, and measurements through time or across patients. For example, the 3D segmentation or measurements can be mapped onto a common aortic model to provide consistent measurement comparison across populations regardless of the imaging modality used to acquire the medical images. In other implementations, the 3D segmentation or measurements can be mapped onto a common aortic model to combine multiple MRI or CT scans obtained during longitudinal surveillance imaging to provide comparisons over time. These measurements can be visualized with respect to the common aortic model. In some implementations, the visualization includes personalized heatmaps. For example, the heatmaps can indicate longitudinal measurement changes or measurement comparisons between an individual subject with reference data or population norms. As one non-limiting example, a heatmap is provided in FIG. 7C in w hich the heatmap indicates regions of the aorta that have abnormal diameter. In this implementation, the heatmap identifies regions with a diameter of ± 2<JSDin which / / represents the mean and <JSD represents the standard deviation for the corresponding region within a healthy control population. As noted above, in addition to anatomic measurements, the aorta analysis data may also include a visualization of hemodynamic parameter measurements.

[0033] Other data may also be input to the one or more trained neural netw orks that generate aorta analysis data as an output, as indicated at step 106. For example, this input data may include patient specific data, such as medical history data, genetic data, demographic data, or other clinically relevant data. Such clinical data may supplement the medical images to generate an aorta analysis.

[0034] In some embodiments, the aorta analysis data may include risk assessment data, such as a risk score. The risk score can provide physicians or other clinicians with a recommendation to consider additional monitoring for subjects whose medical images indicate the likelihood of the subject suffering from a particular medical condition, such as progressive aortic dilatation or risks for adverse outcomes (e.g., rupture, death, cardiovascular events).

[0035] As another example, the aorta analysis data may indicate the probability for a particular classification (i.e., the probability that the medical images include patterns, features, or characteristics indicative of detecting, differentiating, and / or determining the severity of one or more medical conditions).

[0036] Additionally or alternatively, the aorta analysis data may classify the medical images as indicating a particular medical condition or risk for complications and adverse outcomes. In these instances, the aorta analysis data can differentiate between different medical conditions. In still other embodiments, the aorta analysis data may indicate a severity of a medical condition and risk for subsequent complications. For example, the aorta analysis data may include a severity score that quantifies the severity of a medical condition and risk for subsequent complications.

[0037] The aorta analysis data generated by inputting the medical images to the trained neural network(s) can then be displayed to a user, stored for later use or further processing, or both, as indicated at step 108.

[0038] For example, the aorta measurements can be mapped onto the common aorta model. This model can be labeled and color coded to highlight regions that fall outside accepted measurement and grow th rate norms. Such visualization allows for ease of data interpretation and identification of aortic regions with abnormal anatomic measurements and provides a standardized report of aortic measurements. In addition, the comparison with normative data of a representative control group (matched for age, sex, race ethnicity, risk profile, etc.) can be used to generate a personalized aorta heatmap. The aorta heatmap provides visualization of differences in regional aortic dimensions compared to a control population and quantification of regional and global changes. The standardized measurement data and personalized heatmap findings data can be supplemented by patient specific clinical, demographic, genetic risks extracted from the electronic medical record to derive a multi-variate risk score for aortic complications. The model results can be visualized on a graphical user interface (GUI) that enables users to interact with the model for detailed visualization and additional measurement of any specific regions of interests. The GUI can also hnk and spatially register findings in themodel to the actual CT or MR imaging used to build the model. The end-user can save and send the model and specific annotations to a hospital PACS system or local storage.

[0039] Referring now to FIG. 2, a flowchart is illustrated as setting forth the steps of an example method for training one or more neural networks (or other suitable machine learning algorithms) on training data, such that the one or more neural networks are trained to receive medical images as input data in order to generate aorta analysis as output data. The aorta analysis data may be indicative of aorta segmentation and / or aorta feature measurements, which may be mapped onto a common aorta model. The aorta analysis data may also include personalized aorta heatmaps and / or indication of risk of cardiovascular disease or another clinical condition.

[0040] In general, the neural network(s) can implement any number of different neural network architectures. For instance, the neural network(s) could implement a convolutional neural network, a residual neural network, general adversarial networks (GANs), cycle GANs, diffusion models, or the like. Alternatively, the neural network(s) could be replaced with other suitable machine learning or artificial intelligence algorithms, such as those based on supervised learning, unsupervised learning, deep learning, ensemble learning, dimensionality reduction, and so on.

[0041] The method includes accessing training data with a computer system, as indicated at step 202. Accessing the training data may include retrieving such data from a memory or other suitable data storage device or medium. Alternatively, accessing the training data may include acquiring such data with an MRI system of CT system and transferring or otherwise communicating the data to the computer system.

[0042] In general, the training data can include segmented or non-segmented medical images. For example, the training data may include MRI images and CT images of the aorta with manual segmentation of the aorta boundary. The segmented or non-segmented training data may also be labeled to identify various anatomic landmarks of the aorta. The training data may also be labeled to identify anatomic landmarks, such as the right and left coronary' arteries, left and right subclavian arteries, and left and right common carotid arteries. Additionally, the training data may include other patient data, such clinical history, demographic data, genetic data, or other clinically relevant data. Such data may be accessed from the electronic medical record. As one non-limiting example, the labeled data may also include measurements of hemodynamic parameters, such as blood pressure (e g., systolic blood pressure, diastolic blood pressure, mean arterial blood pressure), systolic pressure variation, pulse pressure variation.right arterial pressure, right ventricular pressure, pulmonary artery pressure, mean pulmonary artery pressure, pulmonary vascular resistance, systemic vascular resistance, central venous pressure, stroke volume, stroke volume variation, stroke volume index, cardiac output, cardiac index, and the like. In some embodiments, the training data may include patient data that have been labeled (e.g.. labeled as containing patterns, features, or characteristics indicative of cardiovascular disease risk; and the like).

[0043] The method can include assembling training data from medical records using a computer system. This step may include assembling the patient data into an appropriate data structure on which the neural network or other machine learning algorithm can be trained. Assembling the training data may include accessing and processing medical images. For example, the medical images may include CT images or MRI images of the aorta or chest. Assembling the training data may include reconstructing images from raw imaging data and applying image processing, such as artifact correction, or other preprocessing steps (e.g., cropping, resampling, filtering etc.). The medical images may be manually segmented. Assembling the training data may also include accessing segmented medical images.

[0044] Assembling the training data may also include assembling clinical data, segmented images, and other relevant data. For instance, assembling the training data may include generating labeled data and including the labeled data in the training data. Labeled data may include clinical data, segmented medical images, or other relevant data that have been labeled as belonging to. or otherwise being associated with, one or more different classifications or categories. For instance, labeled data may include clinical data and / or segmented medical images that have been labeled as being associated with cardiovascular disease or a particular type of cardiovascular disease.

[0045] One or more neural networks (or other suitable machine learning algorithms) are trained on the training data, as indicated at step 204. In general, the neural network can be trained by optimizing network and / or training parameters (e.g., weights, biases, or a combination thereof) based on minimizing a loss function. As one non-limiting example, the loss function may be a mean squared error loss function. The training may be tuned by adjusting the training parameters (e.g., number of layers, training epochs, and so on).

[0046] Training a neural network may include initializing the neural network, such as by computing, estimating, or otherwise selecting initial network and / or training parameters (e.g., weights, biases, number of layers, training epochs, or a combination thereof). During training, an artificial neural network receives the inputs for a training example and generatesan output using the bias for each node, and the connections between each node and the corresponding weights. For instance, training data can be input to the initialized neural network, generating output as aorta analysis. The artificial neural network then compares the generated output with the actual output of the training example in order to evaluate the quality of the aorta analysis. For instance, the aorta analysis can be passed to a loss function to compute an error. The current neural network can then be updated based on the calculated error (e.g., using backpropagation methods based on the calculated error). For instance, the current neural network can be updated by updating the network and / or parameters (e.g., weights, biases, number of layers, training epochs, or a combination thereof) in order to minimize the loss according to the loss function. The training continues until a training condition is met. The training condition may correspond to. for example, a predetermined number of training examples being used, a minimum accuracy threshold being reached during training and validation, a predetermined number of validation iterations being completed, and the like. When the training condition has been met (e.g., by determining whether an error threshold or other stopping criterion has been satisfied), the current neural network and its associated network parameters represent the trained neural network. Different types of training processes can be used to adjust the bias values and the weights of the node connections based on the training examples. The training processes may include, for example, gradient descent, Newton's method, conjugate gradient, quasi-Newton, Levenberg-Marquardt, among others.

[0047] The artificial neural network can be constructed or otherwise trained based on training data using one or more different learning techniques, such as supervised learning, unsupervised learning, reinforcement learning, ensemble learning, active learning, transfer learning, or other suitable learning techniques for neural networks. As an example, supervised learning involves presenting a computer system with example inputs and their actual outputs (e g., categorizations). In these instances, the artificial neural network is configured to leam a general rule or model that maps the inputs to the outputs based on the provided example inputoutput pairs.

[0048] The one or more trained neural networks are then stored for later use. as indicated at step 206. Storing the neural network(s) may include storing network and / or training parameters (e.g., weights, biases, number of layers, training epochs, or a combination of these), which have been computed or otherwise estimated by training the neural network(s) on the training data. Storing the trained neural network(s) may also include storing the particular neural network architecture to be implemented. For instance, data pertaining to the layers inthe neural network architecture (e.g.. number of layers, type of layers, ordering of layers, connections between layers, hyperparameters for layers) may be stored.

[0049] Referring now to FIG. 3, an example of a system 300 for generating aorta analysis in accordance with some embodiments of the systems and methods described in the present disclosure is shown. As shown in FIG. 3, a computing device 350 can receive one or more types of data (e.g., medical images, clinical data, clinical history, family history, demographic data) from data source 302. In some embodiments, computing device 350 can execute at least a portion of an aorta analysis system 304 to generate aorta analysis from data received from the data source 302.

[0050] Additionally or alternatively, in some embodiments, the computing device 350 can communicate information about data received from the data source 302 to a server 352 over a communication network 354, which can execute at least a portion of the aorta analysis system 304. In such embodiments, the server 352 can return information to the computing device 350 (and / or any other suitable computing device) indicative of an output of the aorta analysis system 304.

[0051] In some embodiments, computing device 350 and / or server 352 can be any suitable computing device or combination of devices, such as a desktop computer, a laptop computer, a smartphone, a tablet computer, a wearable computer, a server computer, a virtual machine being executed by a physical computing device, a cloud-based computing system, a cloud-based computing resource (e.g.. CPU or GPU cluster), and so on. The computing device 350 and / or server 352 can also reconstruct images from the data.

[0052] In some embodiments, data source 302 can be any suitable source of data (e.g., measurement data, images reconstructed from measurement data, processed image data), such as a medical imaging system (e.g., MRI system, CT system), another computing device (e.g., a server storing measurement data, images reconstructed from measurement data, processed image data, cloud-based data storage, etc.), and so on. In some embodiments, data source 302 can be local to computing device 350. For example, data source 302 can be incorporated with computing device 350 (e.g., computing device 350 can be configured as part of a device for measuring, recording, estimating, acquiring, or otherwise collecting or storing data). As another example, data source 302 can be connected to computing device 350 by a cable, a direct wireless link, and so on. Additionally or alternatively, in some embodiments, data source 302 can be located locally and / or remotely from computing device 350, and can communicate datato computing device 350 (and / or server 352) via a communication network (e.g., communication network 354).

[0053] In some embodiments, communication network 354 can be any suitable communication network or combination of communication networks. For example, communication network 354 can include a Wi-Fi network (which can include one or more wireless routers, one or more switches, etc.), a peer-to-peer network (e.g., a Bluetooth network), a cellular network (e.g., a 3G network, a 4G network, etc., complying with any suitable standard, such as CDMA, GSM, LTE, LTE Advanced, WiMAX, etc.), other types of wireless netw ork, a wired netw ork, and so on. In some embodiments, communication network 354 can be a local area network, a wide area network, a public network (e.g., the Internet), a private or semi-pnvate network (e.g., a corporate or university intranet), any other suitable type of netw ork, or any suitable combination of networks. Communications links shown in FIG. 3 can each be any suitable communications link or combination of communications links, such as wired links, fiber optic links, Wi-Fi links. Bluetooth links, cellular links, and so on.

[0054] Referring now to FIG. 4, an example of hardware 400 that can be used to implement data source 302, computing device 350, and server 352 in accordance with some embodiments of the systems and methods described in the present disclosure is shown.

[0055] As shown in FIG. 4, in some embodiments, computing device 350 can include a processor 402, a display 404, one or more inputs 406, one or more communication systems 408, and / or memory 410. In some embodiments, processor 402 can be any suitable hardware processor or combination of processors, such as a central processing unit (‘'CPU”), a graphics processing unit (“GPU’’), and so on. In some embodiments, display 404 can include any suitable display devices, such as a liquid crystal display (“LCD”) screen, a light-emitting diode (“LED”) display, an organic LED (“OLED”) display, an electrophoretic display (e.g., an “e- ink” display), a computer monitor, a touchscreen, a television, and so on. In some embodiments, inputs 406 can include any suitable input devices and / or sensors that can be used to receive user input, such as a keyboard, a mouse, a touchscreen, a microphone, and so on.

[0056] In some embodiments, communications systems 408 can include any suitable hardware, firmware, and / or software for communicating information over communication network 354 and / or any other suitable communication networks. For example, communications systems 408 can include one or more transceivers, one or more communication chips and / or chip sets, and so on. In a more particular example, communications systems 408 can includehardware, firmware, and / or software that can be used to establish a Wi-Fi connection, a Bluetooth connection, a cellular connection, an Ethernet connection, and so on.

[0057] In some embodiments, memory 410 can include any suitable storage device or devices that can be used to store instructions, values, data, or the like, that can be used, for example, by processor 402 to present content using display 404, to communicate with server 352 via communications system(s) 408, and so on. Memory 410 can include any suitable volatile memory, non-volatile memory, storage, or any suitable combination thereof. For example, memory 410 can include random-access memory (“RAM”), read-only memory (“ROM”), electrically programmable ROM (“EPROM”), electrically erasable ROM (“EEPROM"), other forms of volatile memory, other forms of non-volatile memory, one or more forms of semi-volatile memory, one or more flash drives, one or more hard disks, one or more solid state drives, one or more optical drives, and so on. In some embodiments, memory 410 can have encoded thereon, or otherwise stored therein, a computer program for controlling operation of computing device 350. In such embodiments, processor 402 can execute at least a portion of the computer program to present content (e.g., images, user interfaces, graphics, tables), receive content from server 352, transmit information to server 352, and so on. For example, the processor 402 and the memory 410 can be configured to perform the methods described herein (e.g., the method of FIGS. 1 & 2).

[0058] In some embodiments, server 352 can include a processor 412. a display 414, one or more inputs 416. one or more communications systems 418. and / or memory 420. In some embodiments, processor 412 can be any suitable hardware processor or combination of processors, such as a CPU, a GPU, and so on. In some embodiments, display 414 can include any suitable display devices, such as an LCD screen, LED display, OLED display, electrophoretic display, a computer monitor, a touchscreen, a television, and so on. In some embodiments, inputs 416 can include any suitable input devices and / or sensors that can be used to receive user input, such as a keyboard, a mouse, a touchscreen, a microphone, and so on.

[0059] In some embodiments, communications systems 418 can include any suitable hardware, firmware, and / or software for communicating information over communication network 354 and / or any other suitable communication networks. For example, communications systems 418 can include one or more transceivers, one or more communication chips and / or chip sets, and so on. In a more particular example, communications systems 418 can include hardware, firmware, and / or software that can be used to establish a Wi-Fi connection, a Bluetooth connection, a cellular connection, an Ethernet connection, and so on.

[0060] In some embodiments, memory' 420 can include any suitable storage device or devices that can be used to store instructions, values, data, or the like, that can be used, for example, by processor 412 to present content using display 414, to communicate with one or more computing devices 350, and so on. Memory 420 can include any suitable volatile memory', non-volatile memory', storage, or any suitable combination thereof. For example, memory 420 can include RAM. ROM, EPROM, EEPROM, other ty pes of volatile memory, other ty pes of non-volatile memory, one or more types of semi-volatile memory, one or more flash drives, one or more hard disks, one or more solid state drives, one or more optical drives, and so on. In some embodiments, memory 420 can have encoded thereon a server program for controlling operation of server 352. In such embodiments, processor 412 can execute at least a portion of the server program to transmit information and / or content (e.g., data, images, a user interface) to one or more computing devices 350, receive information and / or content from one or more computing devices 350, receive instructions from one or more devices (e.g., a personal computer, a laptop computer, a tablet computer, a smartphone), and so on.

[0061] In some embodiments, the server 352 is configured to perform the methods described in the present disclosure. For example, the processor 412 and memory 420 can be configured to perform the methods described herein (e.g., the method of FIGS. 1 & 2).

[0062] In some embodiments, data source 302 can include a processor 422, one or more data acquisition systems 424, one or more communications systems 426, and / or memory 428. In some embodiments, processor 422 can be any suitable hardware processor or combination of processors, such as a CPU, a GPU, and so on. In some embodiments, the one or more data acquisition systems 424 are generally configured to acquire data, images, or both, and can include MRI and CT imaging systems. Additionally or alternatively, in some embodiments, the one or more data acquisition systems 424 can include any suitable hardware, firmware, and / or software for coupling to and / or controlling operations of MRI and CT imaging systems. In some embodiments, one or more portions of the data acquisition system(s) 424 can be removable and / or replaceable.

[0063] Note that, although not shown, data source 302 can include any suitable inputs and / or outputs. For example, data source 302 can include input devices and / or sensors that can be used to receive user input, such as a keyboard, a mouse, a touchscreen, a microphone, a trackpad, a trackball, and so on. As another example, data source 302 can include any suitable display devices, such as an LCD screen, an LED display, an OLED display, an electrophoretic display, a computer monitor, a touchscreen, a television, etc., one or more speakers, and so on.

[0064] In some embodiments, communications systems 426 can include any suitable hardware, firmware, and / or software for communicating information to computing device 350 (and, in some embodiments, over communication network 354 and / or any other suitable communication networks). For example, communications systems 426 can include one or more transceivers, one or more communication chips and / or chip sets, and so on. In a more particular example, communications systems 426 can include hardware, firmware, and / or software that can be used to establish a wired connection using any suitable port and / or communication standard (e.g., VGA, DVI video, USB, RS-232, etc.), Wi-Fi connection, a Bluetooth connection, a cellular connection, an Ethernet connection, and so on.

[0065] In some embodiments, memory’ 428 can include any suitable storage device or devices that can be used to store instructions, values, data, or the like, that can be used, for example, by processor 422 to control the one or more data acquisition systems 424, and / or receive data from the one or more data acquisition systems 424; to generate images from data; present content (e.g., data, images, a user interface) using a display; communicate with one or more computing devices 350; and so on. Memory 428 can include any suitable volatile memory, non-volatile memory, storage, or any suitable combination thereof. For example, memory 428 can include RAM, ROM, EPROM, EEPROM, other ty pes of volatile memory, other ty pes of non-volatile memory7, one or more types of semi-volatile memory, one or more flash drives, one or more hard disks, one or more solid state drives, one or more optical drives, and so on. In some embodiments, memory 428 can have encoded thereon, or otherwise stored therein, a program for controlling operation of data source 302. In such embodiments, processor 422 can execute at least a portion of the program to generate images, transmit information and / or content (e.g., data, images, a user interface) to one or more computing devices 350. receive information and / or content from one or more computing devices 350, receive instructions from one or more devices (e.g., a personal computer, a laptop computer, a tablet computer, a smartphone, etc.), and so on.

[0066] In some embodiments, any suitable computer-readable media can be used for storing instructions for performing the functions and / or processes described herein. For example, in some embodiments, computer-readable media can be transitory or non-transitory. For example, non-transitory7computer-readable media can include media such as magnetic media (e.g., hard disks, floppy disks), optical media (e.g., compact discs, digital video discs, Blu-ray discs), semiconductor media (e.g., RAM, flash memory7, EPROM, EEPROM), any suitable media that is not fleeting or devoid of any semblance of permanence duringtransmission, and / or any suitable tangible media. As another example, transitory computer- readable media can include signals on networks, in wires, conductors, optical fibers, circuits, or any suitable media that is fleeting and devoid of any semblance of permanence during transmission, and / or any suitable intangible media.

[0067] Referring particularly now to FIG. 5, an example of an MRI system 500 that can implement the methods described herein is illustrated. The MRI system 500 includes an operator workstation 502 that may include a display 504, one or more input devices 506 (e.g., a keyboard, a mouse), and a processor 508. The processor 508 may include a commercially available programmable machine running a commercially available operating system. The operator workstation 502 provides an operator interface that facilitates entering scan parameters into the MRI system 500. The operator workstation 502 may be coupled to different servers, including, for example, a pulse sequence server 510, a data acquisition server 512, a data processing server 514, and a data store server 516. The operator workstation 502 and the servers 510, 512, 514, and 516 may be connected via a communication system 540, which may include wired or wireless network connections.

[0068] The MRI system 500 also includes a magnet assembly 524 that includes a polarizing magnet 526, which may be a low-field magnet. The MRI system 500 may optionally include a whole-body RF coil 528 and a gradient system 518 that controls a gradient coil assembly 522.

[0069] The pulse sequence server 510 functions in response to instructions provided by the operator workstation 502 to operate a gradient system 518 and a radiofrequency (“RF”) system 520. Gradient waveforms for performing a prescribed scan are produced and applied to the gradient system 518, which then excited gradient coils in an assembly 522 to produce the magnetic field gradients (e.g., Gx. Gy. and Gz) that can be used for spatially encoding magnetic resonance signals. The gradient coil assembly 522 forms part of a magnet assembly 524 that includes a polarizing magnet 526 and a whole-body RF coil 528.

[0070] RF waveforms are applied by the RF system 520 to the RF coil 528, or a separate local coil to perform the prescribed magnetic resonance pulse sequence. Responsive magnetic resonance signals detected by the RF coil 528, or a separate local coil, are received by the RF system 520. The responsive magnetic resonance signals may be amplified, demodulated, filtered, and digitized under direction of commands produced by the pulse sequence server 510. The RF system 520 includes an RF transmitter for producing a wide variety of RF pulses used in MRI pulse sequences. The RF transmitter is responsive to the prescribed scan and directionfrom the pulse sequence server 510 to produce RF pulses of the desired frequency, phase, and pulse amplitude waveform. The generated RF pulses may be applied to the whole-body RF coil 528 or to one or more local coils or coil arrays.

[0071] The RF system 520 also includes one or more RF receiver channels. An RF receiver channel includes an RF preamplifier that amplifies the magnetic resonance signal received by the coil 528 to which it is connected, and a detector that detects and digitizes the I and Q quadrature components of the received magnetic resonance signal. The magnitude of the received magnetic resonance signal may, therefore, be determined at a sampled point by the square root of the sum of the squares of the I and Q components:M = J( / 2+ Q2)

[0072] and the phase of the received magnetic resonance signal may also be determined according to the following relationship:

[0073] The pulse sequence server 510 may receive patient data from a physiological acquisition controller 530. By way of example, the physiological acquisition controller 530 may receive signals from a number of different sensors connected to the patient, including electrocardiograph ("ECG") signals from electrodes, or respiratory signals from a respiratory bellows or other respirator}' monitoring devices. These signals may be used by the pulse sequence server 510 to synchronize, or “gate,” the performance of the scan with the subject’s heartbeat or respiration.

[0074] The pulse sequence server 510 may also connect to a scan room interface circuit 532 that receives signals from various sensors associated with the condition of the patient and the magnet system. Through the scan room interface circuit 532, a patient positioning system 534 can receive commands to move the patient to desired positions during the scan.

[0075] The digitized magnetic resonance signal samples produced by the RF system 520 are received by the data acquisition server 512. The data acquisition server 512 operates in response to instructions downloaded from the operator workstation 502 to receive the realtime magnetic resonance data and provide buffer storage, so that data are not lost by data overrun. In some scans, the data acquisition server 512 passes the acquired magnetic resonance data to the data processor server 514. In scans that require information derived from acquired magnetic resonance data to control the further performance of the scan, the data acquisitionserv er 512 may be programmed to produce such information and convey it to the pulse sequence server 510. For example, during pre-scans, magnetic resonance data may be acquired and used to calibrate the pulse sequence performed by the pulse sequence server 510. As another example, navigator signals may be acquired and used to adjust the operating parameters of the RF system 520 or the gradient system 518, or to control the view order in which k-space is sampled. In still another example, the data acquisition server 512 may also process magnetic resonance signals used to detect the arrival of a contrast agent in a magnetic resonance angiography (“MRA”) scan. For example, the data acquisition server 512 may acquire magnetic resonance data and processes it in real-time to produce information that is used to control the scan.

[0076] The data processing server 514 receives magnetic resonance data from the data acquisition server 512 and processes the magnetic resonance data in accordance with instructions provided by the operator workstation 502. Such processing may include, for example, reconstructing two-dimensional or three-dimensional images by performing a Fourier transformation of raw k-space data, performing other image reconstruction algorithms (e.g., iterative or backproj ection reconstruction algorithms), applying filters to raw k-space data or to reconstructed images, generating functional magnetic resonance images, or calculating motion or flow images.

[0077] Images reconstructed by the data processing server 514 are conveyed back to the operator workstation 502 for storage. Real-time images may be stored in a data base memory cache, from which they may be output to operator display 502 or a display 536. Batch mode images or selected real time images may be stored in a host database on disc storage 538. When such images have been reconstructed and transferred to storage, the data processing server 514 may notify the data store server 516 on the operator workstation 502. The operator workstation 502 may be used by an operator to archive the images, produce films, or send the images via a network to other facilities.

[0078] The MRI system 500 may also include one or more networked workstations 542. For example, a networked workstation 542 may include a display 544, one or more input devices 546 (e.g., a keyboard, a mouse), and a processor 548. The networked workstation 542 may be located within the same facility as the operator workstation 502, or in a different facility, such as a different healthcare institution or clinic.

[0079] The networked workstation 542 may gain remote access to the data processing server 514 or data store server 516 via the communication system 540. Accordingly, multiplenetworked workstations 542 may have access to the data processing server 514 and the data store server 516. In this manner, magnetic resonance data, reconstructed images, or other data may be exchanged between the data processing server 514 or the data store server 516 and the networked workstations 542, such that the data or images may be remotely processed by a networked workstation 542.

[0080] Referring particularly now to FIGS. 6A and 6B. an example of an x-ray computed tomography (“CT”) imaging system 600 is illustrated. The CT system includes a gantry 602, to which at least one x-ray source 604 is coupled. The x-ray source 604 projects an x-ray beam 606, which may be a fan-beam or cone-beam of x-rays, towards a detector array 608 on the opposite side of the gantry 602. The detector array 608 includes a number of x-ray detector elements 610. Together, the x-ray detector elements 610 sense the projected x-rays 606 that pass through a subject 612, such as a medical patient or an object undergoing examination, that is positioned in the CT system 600. Each x-ray detector element 610 produces an electrical signal that may represent the intensity of an impinging x-ray beam and, hence, the attenuation of the beam as it passes through the subject 612. In some configurations, each x- ray detector 610 is capable of counting the number of x-ray photons that impinge upon the detector 610. During a scan to acquire x-ray projection data, the gantry 602 and the components mounted thereon rotate about a center of rotation 614 located within the CT system 600.

[0081] The CT system 600 also includes an operator workstation 616, which ty pically includes a display 618; one or more input devices 620. such as a keyboard and mouse; and a computer processor 622. The computer processor 622 may include a commercially available programmable machine running a commercially available operating system. The operator workstation 616 provides the operator interface that enables scanning control parameters (e.g., acquisition parameters) to be entered into the CT system 600. In general, the operator workstation 616 is in communication with a data store server 624 and an image reconstruction system 626. By way of example, the operator workstation 616, data store sever 624, and image reconstruction system 626 may be connected via a communication system 628, which may include any suitable network connection, whether wired, wireless, or a combination of both. As an example, the communication system 628 may include both proprietary or dedicated networks, as well as open networks, such as the internet.

[0082] The operator workstation 616 is also in communication with a control system 630 that controls operation of the CT system 600. The control system 630 generally includes an x-ray controller 632, a table controller 634, a gantry controller 636. and a data acquisitionsystem 638. The x-ray controller 632 provides power and timing signals to the x-ray source 604 and the gantry controller 636 controls the rotational speed and position of the gantry 602. The table controller 634 controls a table 640 to position the subject 612 in the gantry 602 of the CT system 600.

[0083] The DAS 638 samples data from the detector elements 610 and converts the data to digital signals for subsequent processing. For instance, digitized x-ray data is communicated from the DAS 638 to the data store server 624. The image reconstruction system 626 then retrieves the x-ray data from the data store server 624 and reconstructs an image therefrom. The image reconstruction system 626 may include a commercially available computer processor, or may be a highly parallel computer architecture, such as a system that includes multiple-core processors and massively parallel, high-density computing devices. Optionally, image reconstruction can also be performed on the processor 622 in the operator workstation 616. Reconstructed images can then be communicated back to the data store server 624 for storage or to the operator workstation 616 to be displayed to the operator or clinician.

[0084] The CT system 600 may also include one or more networked workstations 642. By way of example, a networked workstation 642 may include a display 644; one or more input devices 646, such as a keyboard and mouse; and a processor 648. The networked workstation 642 may be located within the same facility as the operator workstation 616, or in a different facility, such as a different healthcare institution or clinic.

[0085] The networked workstation 642, whether within the same facility or in a different facility as the operator workstation 616, may gain remote access to the data store server 624 and / or the image reconstruction system 626 via the communication system 628. Accordingly, multiple networked workstations 642 may have access to the data store server 624 and / or image reconstruction system 626. In this manner, x-ray data, reconstructed images, or other data may be exchanged between the data store server 624, the image reconstruction system 626, and the networked workstations 642, such that the data or images may be remotely processed by a networked workstation 642. This data may be exchanged in any suitable format, such as in accordance with the transmission control protocol ( 'TCP"), the internet protocol C’lP”), or other known or suitable protocols.

[0086] Referring to FIGS. 7A, 7B and 7C, example workflows are illustrated, which can be used to provide aorta analysis according to the present disclosure. The aorta analysis may be automated once the machine learning models are trained on manually labeled imaging data.

[0087] As shown in FIG. 7 A, CT or MRI images may be acquired from a patient at a hospital or imaging center. In the present example, the CT and MRI images were acquired with contrast. The images can be automatically or manually de-identified and uploaded onto a cloudbased server to perform analysis. The deep learning analysis may be performed with a suitable processor on the chest or aorta images CT or MRI images. The analysis includes mapping the aorta measurements onto a common aorta model and deriving a personalized aorta heatmap, regardless or whether CT or MRI images are used. The standardized aortic measurements are sent back to the hospital, imaging center, or other clinic to be used by the patient’s care team for improved diagnosis, risk-stratification, or treatment planning.

[0088] The workflow for standardized aortic measurements and mapping onto a common aorta model is shown in greater detail in FIG. 7B. Two deep learning networks are trained on manually labeled data. The first is trained on contrast-enhanced CT images from a group of subjects, while the second network is trained on contrast-enhanced MRI images from a group of subjects. The training data (CT or MRI images) are manually labeled to segment the aorta, providing aorta dimensions, volumes, and other relevant metrics. The training data also delineate anatomic landmarks defined by the common aorta model. In the present example, 8 anatomic locations are delineated in the training data. Each of these landmarks can be used to divide the aorta in sub-volumes, including proximal, mid, and distal ascending aorta; aortic arc; and proximal, mid. and distal descending aorta.

[0089] The workflow for deriving a personalized aorta heatmap is shown in greater detail in FIG. 7C. A deep learning network is trained on the common aorta model with standardized measurements that have been referenced or related to representative control groups (matched for age, sex, race ethnicity, risk profile, etc ). The aorta heatmap allows for visualization of differences in regional aortic dimensions compared to the control population and quantification regional and global changes, e g., by color coding aorta regions with dimension that are above or below 2 standard deviations (STDEV) relative to the distribution of aortic dimensions of the control population. The personalized heatmap can be used to quantify significant differences in aortic dimensions compared to the control population and quantify regional and global changes (e.g., percentage of the aorta with significant aortic dilatation).

[0090] After training models on CT data and MRI data, a patient may receive either CT or MRI imaging. The corresponding machine learning model is applied to provide standardized aorta measurements and identity’ anatomic landmarks and / or anatomic locations of the aorta.These measurements are mapped onto a common aorta model so that the process provides standardized aortic measurements. These measurements may include characterization of the volume, diameter, and shape of the identified anatomic landmarks. The measurements may include absolute values or comparisons to population norms or previous measurements. Further, a personalized aorta heatmap is derived based on the common aorta model and appropriate control population. Such standardized measurements and heatmaps can be compared longitudinally for a given subject to determine disease progression and aorta growth rates, which may be compared to population norms. The aorta measurements are used in conjunction with patient data from the electronic medical record (EMR) to provide a risk score for aortic complications such as progressive aortic dilatation or adverse outcomes (e g., rupture, death, cardiovascular events, or other cardiovascular conditions).

[0091] As used herein in the context of computer implementation, unless otherwise specified or limited, the terms “component,” “system,” “module,” “framework,” and the like are intended to encompass part or all of computer-related systems that include hardware, software, a combination of hardware and software, or software in execution. For example, a component may be, but is not limited to being, a processor device, a process being executed (or executable) by a processor device, an object, an executable, a thread of execution, a computer program, or a computer. By way of illustration, both an application running on a computer and the computer can be a component. One or more components (or system, module, and so on) may reside within a process or thread of execution, may be localized on one computer, may be distributed between two or more computers or other processor devices, or may be included within another component (or system, module, and so on).

[0092] In some implementations, devices or systems disclosed herein can be utilized or installed using methods embodying aspects of the disclosure. Correspondingly, description herein of particular features, capabilities, or intended purposes of a device or system is generally intended to inherently include disclosure of a method of using such features for the intended purposes, a method of implementing such capabilities, and a method of installing disclosed (or otherwise known) components to support these purposes or capabilities. Similarly, unless otherwise indicated or limited, discussion herein of any method of manufacturing or using a particular device or system, including installing the device or system, is intended to inherently include disclosure, as embodiments of the disclosure, of the utilized features and implemented capabilities of such device or system.

[0093] The present disclosure has described one or more preferred embodiments, and it should be appreciated that many equivalents, alternatives, variations, and modifications, aside from those expressly stated, are possible and within the scope of the invention.

Claims

CLAIMS1. A method for generating aorta analysis, the method comprising:(a) accessing imaging data using a computer system, wherein the imaging data comprises medical images acquired from a subject;(b) accessing a machine learning model with the computer system, wherein the machine learning model has been trained on training data to receive medical images as an input and generate aorta analysis data comprising standardized aorta measurements as an output;(c) generating aorta analysis data of the subject by applying the imaging data to the machine learning model using the computer system; and(d) outputting the aorta analysis data to a user.

2. The method of claim 1, wherein the aorta analysis data comprises a subjectspecific aorta heatmap.

3. The method of claim 1 or 2, wherein the medical images comprise at least one of computed tomography (CT) or magnetic resonance imaging (MR1) images.

4. The method of any one of claims 1-3, wherein the aorta analysis data comprise standardized aorta measurements mapped onto a common aorta model.

5. The method of claim 4, wherein the aorta analysis data comprise a comparison of the standardized aorta measurements to a control population.

6. The method of any one of claims 1-3, wherein the aorta analysis data comprise a 3D segmentation of an aorta of the subject.

7. The method of any one of claims 1-3, wherein the aorta analysis data comprise a delineation of one or more anatomic landmarks of the subject’s aorta.

8. The method of claim 7. wherein the one or more anatomic landmarks comprises at least one of anatomic locations, anatomic regions, sub-volumes, or aorta zones of the subject.

9. The method of claim 7. wherein the aorta analysis data comprise measurements of at least one of a diameter, a circumference, or a volume of each of the one or more anatomic landmarks.

10. A system for generating aorta analysis data, the system comprising: a memoiy to store medical imaging data and a machine learning model, wherein: the imaging data comprise medical images are acquired from a subject; the machine learning model has been trained on training data to generate aorta analysis that includes standardized aorta measurements; a computer processor in communication with the memory to: receive the imaging data from the memory; receive the machine learning model from the memory; apply the imaging data to the machine learning model, generating aorta analysis data of the subject as an output; and output the aorta analysis data to a user.

11. The system of claim 10, wherein the aorta analysis further includes a mapping of the standardized aorta measurements onto a common aorta model.

12. The system of claim 10 or 11, wherein the aorta analysis further includes a subject-specific aorta heatmap.

13. The system of any one of claims 10-12, wherein the aorta analysis further includes a risk score for aortic complications.

14. A method for training a machine learning model for analyzing aortas in medical images, the method comprising:(a) accessing training data using a computer system, wherein the training data comprises labeled medical images acquired from a group of subjects;(b) with the computer system, training a machine learning model on the training data to generate aorta analy sis data that include standardized aorta measurements; and(c) storing the machine learning model with a computer system.

15. The method of claim 14, wherein the labeled medical images comprise medical images that have been segmented to identify anatomic landmarks of an aorta for each subject in the group of subjects.

16. The method of claim 14 or 15, wherein the anatomic landmarks comprise at least one of anatomic locations, anatomic regions, sub-volumes, or aorta zones of each subject in the group of subjects.

17. The method of any one of claims 14-16, wherein the labeled medical images include labels comprising measurements of at least one of a diameter, a circumference, a volume, or a hemodynamic parameter of each of the anatomic landmarks.

18. The method of any one of claims 14-17, wherein the training data further comprise at least one of clinical history data, demographic data, or genetic risk data of each subject in the group of subjects.

19. The method of any one of claims 14-18, wherein the labeled medical images comprise at least one of computed tomography (CT) or magnetic resonance imaging (MRI) images.

Images