WO2025229488A1 - Prediction of left ventricular outflow tract obstruction - Google Patents

Abstract

Systems and methods are provided for determining a suitability of a patient for a medical procedure. A feature extractor generates a plurality of numerical parameters from a received image. The plurality of numerical parameters includes a dimension of a structure within a heart of the patient and a projected area of a left ventricular outflow track. A predictive model determines if the patient is a candidate for the medical procedure from the plurality of numerical parameters.

Description

PREDICTION OF LEFT VENTRICULAR OUTFLOW TRACT OBSTRUCTION CROSS-REFERENCE TO RELATED APPLICATIONS [0001] This application claims the benefit of U.S. Provisional Patent Application Serial No.63/639,791, filed April 29, 2024 and U.S. Provisional Patent Application Serial No.63/643,151, filed May 6, 2024, the entire contents of which are incorporated herein by reference. TECHNICAL FIELD [0002] This invention relates to clinical decision support systems, and more particularly, to prediction of left ventricular outflow tract obstruction. BACKGROUND [0003] Left ventricular outflow tract obstruction (LVOTO) is a condition in which blood flow out of the left ventricle is limited. This can occur naturally, due to aberrant anatomy or changes in valve function or as a result of medical intervention within the heart. Obstruction of the left ventricular outflow tract can result in hypertrophy of the left ventricle and associated pathologies. SUMMARY [0004] In accordance with one example, a system includes a processor and a non-transitory computer readable medium storing machine-readable instructions that are executable by the processor to determine if a patient is a candidate for a mitral valve repair. A feature extractor generates a plurality of numerical parameters from a received image. The plurality of numerical parameters includes a dimension of a structure within a heart of the patient and a projected area of a left ventricular outflow track. A predictive model determines if the patient is a candidate for the mitral valve repair from the plurality of numerical parameters. [0005] In accordance with another example, a method includes receiving a first image of a heart of a patient and extracting a plurality of numerical parameters from the first image of the heart of the patient. A second image of the heart of the patient is received and at least one numerical parameter is extracted from the second image of the heart of the patient. It is determined at a predictive model if the patient is a candidate for a medical procedure from a set of parameters including the plurality of numerical parameters extracted from the first image and the at least one numerical parameter extracted from the second image. [0006] In accordance with a further example, a system includes a processor and a non-transitory computer readable medium storing machine-readable instructions that are executable by the processor to determine if a patient is a candidate for a medical procedure. A feature extractor generates a plurality of numerical parameters from a first received image and generates at least one numerical parameter from a second received image to provide a set of numerical parameters. A predictive model determines if the patient is a candidate for the medical procedure from the set of numerical parameters. [0007] In accordance with a still further example, a method includes receiving a first image of a heart of a patient and determining a projected area of a left ventricular outflow track (LVOT) after a medical procedure from the first image. The patient is categorized as a candidate for a medical procedure if the projected area of the LVOT is greater than or equal to a first threshold value. The patient is categorized as unsuitable for the procedure if the projected area of the LVOT is less than or equal to a second threshold value. If the projected area of the LVOT for the patient is between the first threshold value and the second threshold value, a set of parameters are each extracted from either the first image or a second image, a score for the patient is determined from the set of parameters and the projected area of the LVOT, and the patient is categorized as a candidate for the medical procedure if the score meets a third threshold value. BRIEF DESCRIPTION OF THE DRAWINGS [0008] FIG.1 illustrates an example of system for determining if a patient is a candidate for a surgical procedure; [0009] FIG.2 illustrates another example of a system for determining if a patient is a candidate for a surgical procedure; [0010] FIG.3 illustrates one example of a method for determining if a patient is a candidate for a medical procedure; [0011] FIG.4 illustrates another example of a method for determining if a patient is a candidate for a medical procedure; [0012] FIG.5 illustrates a further example of a method for determining if a patient is a candidate for a medical procedure; and [0013] FIG.6 is a schematic block diagram illustrating an exemplary system of hardware components capable of implementing examples of the systems and methods disclosed herein. DEFINITIONS [0014] A “predictive model,” as used herein, is a mathematical or machine learning model that predicts a parameter associated with adverse patient outcomes. Examples of predictive models include artificial neural networks, convolutional neural networks, convolutional autoencoders, linear regression models, logistic regression models, Bayesian networks, such as naive-bayes, random forest models, boosting and bagging methods, decision trees, hidden Markov models, support vector machines, K-means clustering, and K-nearest neighbor classifiers. [0015] A “function” of a parameter is a mathematical expression that produces a unique output for each possible value of the parameter. As used herein, the identity function is explicitly included in the definition of the term “function.” DETAILED DESCRIPTION [0016] Systems and methods are provided herein for identifying patients who are suitable candidates for a medical procedure, specifically by evaluating a risk of left ventricular outflow tract obstruction (LVOTO) due to the procedure. Due to this risk, some patients are restricted from otherwise beneficial procedures due to concerns that the procedure may result in obstruction of the left ventricular outflow tract. By identifying patients who are suitable for the medical procedure with greater accuracy, treatment can be provided patients who might otherwise be unable to benefit from a procedure due to LVOTO risks. [0017] FIG.1 illustrates one example of a system 100 for employing a predictive model for determining if a patient is a candidate for a surgical procedure. In one example, the surgical procedure is a transcatheter mitral valve replacement. The system 100 includes a processor 102, a display 104, and a non-transitory computer readable medium 110 storing executable instructions, executed by the processor 102. It will be appreciated that the executable instructions can be spread across multiple non-transitory computer readable media that are operatively connected via an appropriate data connection, such that the executable instructions can be executed by multiple processors. [0018] The executable instructions stored on the non-transitory computer readable medium 110 include a feature extractor 114 that extracts a set of numerical parameters from a received image. The image can be received from an associated imager or retrieved from storage from a non-transitory computer readable medium via a local or network connection. In one implementation, the feature extractor 114 can include one or more automated segmentation algorithms that can be used to location various structures or other features within the image, and extract parameters representing a length, thickness, area, or volume of various features within the image. In another implementation, the feature extractor 114 can accept manual or semi-automated segmentations from a human expert and calculate the features from the human-assisted segmentation. In one example, the set of features extracted at the feature extractor can include two or more of an anterior leaflet length, an angle between the left ventricle (LV) and the left ventricle outflow tract (LVOT), an angle between the LV and the mitral valve (MV), a septal bulge thickness, an angle between the aortic valve (AV) and the MV, a parameter representing papillary muscle interaction, a parameter representing papillary muscle width, a chordae length, an anterior leaflet tip distance, a left ventricle height, a posterior wall thickness, a left ventricle end diastolic dimension, a relative wall thickness, a peak gradient of the mitral valve, an LV diameter, a distance from the MV to the septal bulge, a LV diameter at the septal bulge, a left ventricle end systolic diameter, an interventricular septum thickness, an LV volume, a projected LVOT area after a planned procedure (neoLVOT), and an LVOT volume. In some implementations, one or more of the values can be adjusted, for example, according to a size (e.g., height, weight, estimated surface area, or estimated volume) of the patient, to produce an indexed version of the parameter. [0019] The extracted set of parameters are provided to a predictive model 116 that generates a parameter representing the suitability of a patient for a medical procedure. In some implementations, the predictive model 116 can also be provided with biometric parameters of the patient, including one or more of a height, weight, estimated surface area, estimated volume, blood pressure, and ejection fraction. In one example, the predictive model 116 provides a categorical parameter for the patient that can assume a first value, representing a patient who is a candidate for the medical procedure, and a second value representing a patient who is not a good candidate for the medical procedure. In another example, the predictive model 116 provides a continuous or discrete parameter representing a risk of the surgical procedure to the patient or a suitability of the patient for the procedure. In a further example, the predictive model 116 can be a categorical parameter, with each category representing a range of values for a continuous or discrete parameter representing a risk of the surgical procedure to the patient or a suitability of the patient for the procedure. In a still further example, the predictive model 116 can be a categorical parameter with each category representing a procedure that is recommended for the patient. [0020] The predictive model 116 can be implemented as any of a plurality of machine learning algorithms or combinations of multiple machine learning algorithms, including artificial neural networks, rule-based classifiers, linear regression models, non-linear regression models, logistic regression models, Bayesian networks, boosted and bagging models, random forest models, hidden Markov models, and support vector machines. In one example, the predictive model 116 uses one or more supervised learning algorithms, each trained on a set of training samples, with a given training sample containing values for the set of parameters and a known outcome for the patient. In one example, the predictive model 116 can include a rule-based model that divides patients into candidate, non- candidate, and ambiguous classes for a given procedure, and a second model that divides patients from the ambiguous class into the candidate and non-candidate classes. The output of the predictive model 116 can be provided to a user at the display 104. [0021] FIG.2 illustrates another example of a system 200 for employing a predictive model for determining if a patient is a candidate for a surgical procedure. In one example, the surgical procedure is a transcatheter mitral valve replacement. The system 200 includes a processor 202, a display 204, and a non-transitory computer readable medium 210 storing executable instructions that are executed by the processor 202. It will be appreciated that the executable instructions can be spread across multiple non-transitory computer readable media that are operatively connected via an appropriate data connection, such that the executable instructions can be executed by multiple processors. [0022] The executable instructions stored on the non-transitory computer readable medium 210 include an imager interface 212 that receives an image of the heart of the patient from an associated imager. In one implementation, the imager interface 212 is configured to receive images from a computed tomography (CT) imager. It will be appreciated that the imager interface 212 can provide conditioning or formatting for the received image for further processing. In one implementation, the image can be provided as a three-dimensional volume of the heart, divided into multiple two-dimensional slices that can be analyzed individually. In another implementation, the imager interface 212 can be configured to receive images from a second imager, having a different imaging modality than the first imager. In one example, the imager interface 212 can be configured to receive and process both CT and ultrasound images. [0023] The executable instructions further include a network interface 214 with which the system 200 communicates with other systems (not shown) via a network connection, for example, an Internet connection and/or a connection to an internal network. In the illustrated example, the other systems can include an electronic health records (EHR) system that stores medical information for the patient, and the network interface 214 can include an application program interface (API) (not shown) for communicating with the EHR system. Data retrieved from the EHR can include, for example, a height, weight, estimated surface area, estimated volume, blood pressure, and ejection fraction of the patient. Where patient data is not available from the EHR, relevant information for the patient can be entered via an appropriate user interface 216. [0024] A feature extractor 218 extracts a set of numerical parameters from the image. In one implementation, the feature extractor 218 can include one or more automated segmentation algorithms that can be used to location various structures or other features within the image, and extract parameters representing a length, thickness, area, or volume of various features within the image. In another implementation, the feature extractor 218 can accept manual or semi-automated segmentations from a human expert via the user interface 216 and calculate the features from the human-assisted segmentation. In one example, the set of features extracted at the feature extractor can include two or more of an anterior leaflet length, an angle between the left ventricle (LV) and the left ventricle outflow tract (LVOT), an angle between the LV and the mitral valve (MV), an angle between the aortic valve (AV) and the MV, a parameter representing papillary muscle interaction, a parameter representing papillary muscle width, a chordae length, an anterior leaflet tip distance, a left ventricle height, a posterior wall thickness, a left ventricle end diastolic dimension, a relative wall thickness, a peak gradient of the mitral valve, an LV diameter, a distance from the MV to the septal bulge, a septal bulge thickness, a LV diameter at the septal bulge, a left ventricle end systolic diameter, an interventricular septum thickness, an LV volume, a projected LVOT area after a planned procedure (neoLVOT), a projected LVOT area after a planned procedure that is indexed for an estimated surface area of the patient, and an LVOT volume. [0025] The extracted set of parameters is provided to a predictive model 220 that generates a parameter representing the suitability of a patient for a medical procedure. In some implementations, the predictive model 220 can also be provided with biometric parameters of the patient, including one or more of a height, weight, estimated surface area, estimated volume, blood pressure, and ejection fraction. In the illustrated example, the predictive model 220 provides a categorical parameter for the patient that can assume a first value, representing a patient who is a candidate for the medical procedure, and a second value representing a patient who is not a good candidate for the medical procedure. [0026] The predictive model 220 can be implemented as any of a plurality of machine learning algorithms or combinations of multiple machine learning algorithms, including artificial neural networks, linear regression models, rule-based classifiers, non-linear regression models, logistic regression models, Bayesian networks, boosting and bagging models, random forest models, hidden Markov models, and support vector machines. In the illustrated example, the predictive model 220 uses one or more supervised learning algorithms, each trained on a set of training samples, with a given training sample containing values for the set of parameters and a known outcome for the patient. Where multiple classification or regression models are used, an arbitration element can be utilized to provide a coherent result from the plurality of models. The training process of a given classifier will vary with its implementation, but training generally involves a statistical aggregation of training data into one or more parameters associated with the output class. For rule-based models, such as decision trees, domain knowledge, for example, as provided by one or more human experts, can be used in place of or to supplement training data in selecting rules for classifying a patient using the extracted features. Any of a variety of techniques can be utilized for the classification algorithm, including support vector machines (SVMs), regression models, self-organized maps, fuzzy logic systems, data fusion processes, boosting and bagging methods, rule-based systems, or artificial neural networks. [0027] For example, a support vector machine (SVM) classifier can utilize a plurality of functions, referred to as hyperplanes, to conceptually divide boundaries in the N-dimensional feature space, where each of the N dimensions represents one associated feature of the feature vector. The boundaries define a range of feature values associated with each class. Accordingly, an output class and an associated confidence value can be determined for a given input feature vector according to its position in feature space relative to the boundaries. In one implementation, the SVM can be implemented via a kernel method using a linear or non-linear kernel. [0028] An artificial neural network includes a plurality of nodes having a plurality of interconnections, referred to as links. Input values are provided to a plurality of input nodes, which provide these input values to layers of one or more intermediate nodes, referred to as hidden nodes. A given intermediate node receives one or more output values from previous nodes, which are weighted according to a series of weights established during the training of the classifier. An intermediate node translates its received values into a single output according to a transfer function at the node. For example, the intermediate node can sum the received values and subject the sum to a binary step function, a sigmoid function, a hyperbolic tangent function, or a rectified linear unit function. A final layer of nodes provides the confidence values for the output classes of the artificial neural network, with each node having an associated value representing a confidence for one of the associated output classes of the classifier. [0029] A rule-based classifier applies a set of logical rules to the extracted features to select an output class. Generally, the rules are applied in order, with the logical result at each step influencing the analysis at later steps. The specific rules and their sequence can be determined from any or all of training data, analogical reasoning from previous cases, or existing domain knowledge. One example of a rule-based classifier is a decision tree algorithm, in which the values of features in a feature set are compared to corresponding threshold in a hierarchical tree structure to select a class for the feature vector. A random forest classifier is a modification of the decision tree algorithm using a bootstrap aggregating, or “bagging” approach. In this approach, multiple decision trees are trained on random samples of the training set, and an average (e.g., mean, median, or mode) result across the plurality of decision trees is returned. For a classification task, the result from each tree would be categorical, and thus a modal outcome can be used. [0030] In one implementation, the predictive model 220 includes a rule-based model that divides patients into candidate, non-candidate, and ambiguous classes for a given procedure, and a second model that divides patients from the ambiguous class into the candidate and non-candidate classes. The rule-based model can utilize, for example, a projected area of the LVOT after the procedure to perform the initial rule-based sorting of the patients, for example, by rejecting patients who have a projected area of the LVOT below a first threshold value and accepting patients who have a projected area of the LVOT above a second threshold value. Patients having projected LVOT areas between the two thresholds can be evaluated at the second model, where a score is calculated, representing the suitability of the patient for the procedure. In one example, the score for the patient is determined as a function of a thickness of a basal septum of the patient, a relative wall thickness of the patient, the projected area of the LVOT, and an estimated body surface area of the patient, with the patient rejected if the score fails to meet a third threshold value and accepted if the third threshold value is met. For example, the score can be calculated as a function of a value determined as a product of the relative wall thickness, a ratio of the projected area of the LVOT to the estimated body surface area, and a square of the thickness of the basal septum. [0031] In another implementation, the predictive model 220 includes a rule- based model that divides patients into non-candidate and ambiguous classes for a given procedure, and a second model that divides patients from the ambiguous class into candidate and non-candidate classes. The rule-based model can utilize, for example, a projected area of the LVOT after the procedure to perform the initial rule- based sorting of the patients, for example, by rejecting patients who have a projected area of the LVOT below a first threshold value. Patients having projected LVOT areas between above the first threshold value can be evaluated at the second model, where a score is calculated, representing the suitability of the patient for the procedure. In one example, the score for the patient is determined as a function of a thickness of a basal septum of the patient, a relative wall thickness of the patient, the projected area of the LVOT, and an estimated body surface area of the patient, with the patient rejected if the score fails to meet a second threshold value and accepted if the second threshold value is met. The output of the predictive model 220 can be provided to a user at the display 204. [0032] In view of the foregoing structural and functional features described above, methods in accordance with various aspects of the present invention will be better appreciated with reference to FIGS.3-5. While, for purposes of simplicity of explanation, the methods of FIGS.3-5 are shown and described as executing serially, it is to be understood and appreciated that the present invention is not limited by the illustrated order, as some aspects could, in accordance with the present invention, occur in different orders and/or concurrently with other aspects from that shown and described herein. Moreover, not all illustrated features may be required to implement a method in accordance with an aspect the present invention. [0033] FIG.3 illustrates one example of a method 300 for determining if a patient is a candidate for a medical procedure. At 302, a plurality of numerical parameters, including a dimension of a structure within a heart of the patient and a projected area of a left ventricular outflow track (LVOT) after the procedure, are extracted from at least one received image of the heart. In one example, the dimension of the structure within the heart can include one or more of a thickness of one or more walls of the left ventricle, a posterior wall thickness of the left ventricle, a left ventricle end diastolic dimension, and a thickness of a basal septum of the patient. The received image or images can be any appropriate modality for cardiac imaging, and in one example, features can be extracted from one or both of a computed tomography image and an ultrasound image. [0034] At 304, a predictive model determines if the patient is a candidate for the surgical procedure from the plurality of numerical parameters. The predictive model can include one or more machine learning models, each trained on training samples including all or a subset of the plurality of numerical parameters as well as a known outcome for the patient. It will be appreciated that a predictive model using multiple machine learning models can operate in parallel, with outputs from each machine learning model evaluated at an arbitrator to select a final output for the predictive model, or sequentially, with results from one machine learning algorithm provided as an input or filtering step for another algorithm. In this instance, additional acquisition of images and feature extraction may be performed after analysis of other extracted features at a given machine learning model to provide features for future models. For example, one model may only require features using a first imaging modality (e.g., computed tomography), and acquisition of images using other modalities may be reserved for patients for whom further analysis is necessary. [0035] FIG.4 illustrates another example of a method 400 for determining if a patient is a candidate for a medical procedure. In one example, the medical procedure is a mitral valve repair, such as a transcatheter mitral valve repair. At 402, a first image of a heart of a patient is received, for example, from a first associated imager or a local or remote storage medium. In one implementation, the first image of the heart is a computed tomography (CT) image. At 404, a plurality of numerical parameters are extracted from the first image of the heart of the patient. At 406. a second image of the heart of the patient is received, for example, from a first associated imager or a local or remote storage medium. In one example, the second image is an ultrasound image. At 408, at least one numerical parameter is extracted from the second image. A set of parameters extracted from the first and second images can include any of an anterior leaflet length, an angle between a left ventricle (LV) and a left ventricle outflow tract (LVOT), an angle between the LV and a mitral valve (MV), an angle between an aortic valve (AV) and the MV, a parameter representing papillary muscle interaction, a parameter representing papillary muscle width, a chordae length, an anterior leaflet tip distance, a septal bulge thickness, a left ventricle height, a posterior wall thickness of the left ventricle, a left ventricle end diastolic dimension, a relative wall thickness for the left ventricle, a peak gradient of the mitral valve, an LV diameter, a distance from the MV to a septal bulge, a LV diameter at the septal bulge, a left ventricle end systolic diameter, an interventricular septum thickness, an LV volume, a projected area of the LVOT, and an LVOT volume. [0036] At 410, it is determined, at a predictive model, if the patient is a candidate for a medical procedure from the set of parameters extracted from the first and second images. In one implementation, a score is calculated at the predictive model from the set of parameters and compared to one or more threshold values to determine if the patient is a candidate for a procedure. In one example, the score for the patient is determined as a function of a thickness of a basal septum of the patient, a relative wall thickness of the patient, the projected area of the LVOT, and an estimated body surface area of the patient. In other examples, the predictive model can have multiple sequential stages in which patients can be selected or rejected, with patients who cannot be classified with sufficient confidence passed to later classification stages. If a patient is determined to be a candidate for the procedure, the procedure can then be performed. [0037] FIG.5 illustrates a further example of a method 500 for determining if a patient is a candidate for a medical procedure. In one example, the medical procedure is a transcatheter mitral valve repair. At 502, a first image of a heart of a patient is received, and a projected area of a left ventricular outflow track (LVOT) after the medical procedure is determined from the first image at 504. In one example, the first image is a computed tomography image. At 506, it is determined if the projected area of the LVOT, neoLVOT, is greater than or equal to a first threshold value, T1. If so (Y), the patient is categorized as a candidate for the medical procedure at 508. If not (N), the method advances to 510, where it is determined if the projected area of the LVOT is less than or equal to a second threshold value, T2. If so (Y), the patient is categorized as unsuitable for the medical procedure at 512. [0038] If the projected area of the LVOT is greater than the second threshold value (N), the method advances to 514 where a set of parameters are extracted. Each of the set of parameters are extracted from either the first image or a second image. In one example, the second image is an ultrasound image. In one implementation, at least one parameter of the set of parameters is extracted from each of the first image and the second image. At 516, a score for the patient is generated from the set of parameters and the projected area of the LVOT. In one example, generating the score for the patient from the set of parameters and the projected area of the LVOT includes providing the set of parameters and the projected area of the LVOT to a predictive model. Additionally or alternatively, the score can be generated from a relative wall thickness associated with the patient, the projected area of the LVOT, an estimated body surface area of the patient, and a thickness of a basal septum of the patient. For example, the score can be generated as a function of a value determined as a product of the relative wall thickness, a ratio of the projected area of the LVOT to the estimated body surface area, and a square of the thickness of the basal septum. [0039] At 518, it is determined if the score meets a third threshold value. It will be appreciated that the score can be calculated such that either larger values or smaller values represent suitability for a procedure, and thus meeting the third threshold can include exceeding a threshold, falling below a threshold, or falling between multiple thresholds. If the score does not meet the threshold (N), the patient is categorized as unsuitable for the medical procedure at 512. If the score meets the threshold (Y), the patient is categorized as a candidate for the medical procedure at 508, and the procedure can be performed. [0040] FIG.6 is a schematic block diagram illustrating an exemplary system 600 of hardware components capable of implementing examples of the systems and methods disclosed herein. The system 600 can include various systems and subsystems. The system 600 can be a personal computer, a laptop computer, a workstation, a computer system, an appliance, an application-specific integrated circuit (ASIC), a server, a server BladeCenter, a server farm, etc. [0041] The system 600 can include a system bus 602, a processing unit 604, a system memory 606, memory devices 608 and 610, a communication interface 612 (e.g., a network interface), a communication link 614, a display 616 (e.g., a video screen), and an input device 618 (e.g., a keyboard, touch screen, and/or a mouse). The system bus 602 can be in communication with the processing unit 604 and the system memory 606. The additional memory devices 608 and 610, such as a hard disk drive, server, standalone database, or other non-volatile memory, can also be in communication with the system bus 602. The system bus 602 interconnects the processing unit 604, the memory devices 606-610, the communication interface 612, the display 616, and the input device 618. In some examples, the system bus 602 also interconnects an additional port (not shown), such as a universal serial bus (USB) port. [0042] The processing unit 604 can be a computing device and can include an application-specific integrated circuit (ASIC). The processing unit 604 executes a set of instructions to implement the operations of examples disclosed herein. The processing unit can include a processing core. [0043] The additional memory devices 606, 608, and 610 can store data, programs, instructions, database queries in text or compiled form, and any other information that may be needed to operate a computer. The memories 606, 608 and 610 can be implemented as computer-readable media (integrated or removable), such as a memory card, disk drive, compact disk (CD), or server accessible over a network. In certain examples, the memories 606, 608 and 610 can include text, images, video, and/or audio, portions of which can be available in formats comprehensible to human beings. [0044] Additionally or alternatively, the system 600 can access an external data source or query source through the communication interface 612, which can communicate with the system bus 602 and the communication link 614. [0045] In operation, the system 600 can be used to implement one or more parts of a system for determining a suitability of a patient for a surgical procedure in accordance with the present invention. Computer executable logic for implementing the diagnostic system resides on one or more of the system memory 606, and the memory devices 608 and 610 in accordance with certain examples. The processing unit 604 executes one or more computer executable instructions originating from the system memory 606 and the memory devices 608 and 610. The term "computer readable medium" as used herein refers to a medium that participates in providing instructions to the processing unit 604 for execution. This medium may be distributed across multiple discrete assemblies all operatively connected to a common processor or set of related processors. [0046] Implementation of the techniques, blocks, steps, and means described above can be done in various ways. For example, these techniques, blocks, steps, and means can be implemented in hardware, software, or a combination thereof. For a hardware implementation, the processing units can be implemented within one or more application specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field programmable gate arrays (FPGAs), processors, controllers, micro- controllers, microprocessors, other electronic units designed to perform the functions described above, and/or a combination thereof. [0047] Also, it is noted that the embodiments can be described as a process which is depicted as a flowchart, a flow diagram, a data flow diagram, a structure diagram, or a block diagram. Although a flowchart can describe the operations as a sequential process, many of the operations can be performed in parallel or concurrently. In addition, the order of the operations can be re-arranged. A process is terminated when its operations are completed but could have additional steps not included in the figure. A process can correspond to a method, a function, a procedure, a subroutine, a subprogram, etc. When a process corresponds to a function, its termination corresponds to a return of the function to the calling function or the main function. [0048] Furthermore, embodiments can be implemented by hardware, software, scripting languages, firmware, middleware, microcode, hardware description languages, and/or any combination thereof. When implemented in software, firmware, middleware, scripting language, and/or microcode, the program code or code segments to perform the necessary tasks can be stored in a machine readable medium such as a storage medium. A code segment or machine-executable instruction can represent a procedure, a function, a subprogram, a program, a routine, a subroutine, a module, a software package, a script, a class, or any combination of instructions, data structures, and/or program statements. A code segment can be coupled to another code segment or a hardware circuit by passing and/or receiving information, data, arguments, parameters, and/or memory contents. Information, arguments, parameters, data, etc. can be passed, forwarded, or transmitted via any suitable means including memory sharing, message passing, ticket passing, network transmission, etc. [0049] For a firmware and/or software implementation, the methodologies can be implemented with modules (e.g., procedures, functions, and so on) that perform the functions described herein. Any machine-readable medium tangibly embodying instructions can be used in implementing the methodologies described herein. For example, software codes can be stored in a memory. Memory can be implemented within the processor or external to the processor. As used herein the term "memory" refers to any type of long term, short term, volatile, nonvolatile, or other storage medium and is not to be limited to any particular type of memory or number of memories, or type of media upon which memory is stored. [0050] Moreover, as disclosed herein, the term "storage medium" can represent one or more memories for storing data, including read only memory (ROM), random access memory (RAM), magnetic RAM, core memory, magnetic disk storage mediums, optical storage mediums, flash memory devices and/or other machine readable mediums for storing information. The term "machine-readable medium" includes, but is not limited to portable or fixed storage devices, optical storage devices, wireless channels, and/or various other storage mediums capable of storing that contain or carry instruction(s) and/or data. [0051] In the preceding description, specific details have been set forth in order to provide a thorough understanding of example implementations of the invention described in the disclosure. However, it will be apparent that various implementations may be practiced without these specific details. For example, circuits, systems, networks, processes, and other components may be shown as components in block diagram form in order not to obscure the example implementations in unnecessary detail. In other instances, well-known circuits, processes, algorithms, structures, and techniques may be shown without unnecessary detail in order to avoid obscuring the examples. The description of the example implementations will provide those skilled in the art with an enabling description for implementing an example of the invention, but it should be understood that various changes may be made in the function and arrangement of elements without departing from the spirit and scope of the invention. Accordingly, the present invention is intended to embrace all such alterations, modifications, and variations that fall within the scope of the appended claims. [0052] The following examples are illustrative of the techniques described herein. [0053] Example 1. A system comprising: a processor; and a non-transitory computer readable medium storing machine-readable instructions that are executable by the processor to determine if a patient is a candidate for a mitral valve repair, the machine-readable instructions comprising: a feature extractor that generates a plurality of numerical parameters from a received image, the plurality of numerical parameters including a dimension of a structure within a heart of the patient and a projected area of a left ventricular outflow track (LVOT) after the procedure; and a predictive model that determines if the patient is a candidate for the mitral valve repair from the plurality of numerical parameters. [0054] Example 2. The system of Example 1, wherein the received image is a first image and the feature extractor receiving a second image and extracting at least one numerical parameter from the second image, the predictive model determining if the patient is a candidate for the mitral valve repair from the plurality of numerical parameters and the at least one numerical parameter extracted from the second image. [0055] Example 3. The system of Example 2, wherein a set of parameters comprising the plurality of numerical parameters and the at least one numerical parameter extracted from the second image includes at least two of an anterior leaflet length, an angle between a left ventricle (LV) and the LVOT, an angle between the LV and a mitral valve (MV), an angle between an aortic valve (AV) and the MV, a parameter representing papillary muscle interaction, a parameter representing papillary muscle width, a chordae length, an anterior leaflet tip distance, a septal bulge thickness, a left ventricle height, a posterior wall thickness, a left ventricle end diastolic dimension, a relative wall thickness, a peak gradient of the mitral valve, an LV diameter, a distance from the MV to a septal bulge, a LV diameter at the septal bulge, a left ventricle end systolic diameter, an interventricular septum thickness, an LV volume, and an LVOT volume. [0056] Example 4. The system of Example 3, wherein the predictive model comprises: a first classifier that classifies the patient into one of a first class, representing patients who are candidates for the mitral valve repair, a second class, representing patients who are not candidates for the mitral valve repair, and a third class from a projected area of the LVOT; and a second classifier that classifies patients who are classified into the third class into one of the first class and the second class from the set of parameters and the projected area of the LVOT. [0057] Example 5. The system of Example 4, wherein the second classifier classifies patients who are classified into the third class into one of the first class and the second class from a thickness of a basal septum of the patient, a relative wall thickness of the patient, the projected area of the LVOT, and an estimated body surface area of the patient. [0058] Example 6. The system of Example 1, the machine-readable instructions further comprising a network interface that receives at least one biometric parameter representing the patient from an external system, the predictive model determining if the patient is a candidate for the mitral valve repair from the plurality of numerical parameters and the at least one biometric parameter. [0059] Example 7. A method comprising: receiving a first image of a heart of a patient; extracting a plurality of numerical parameters from the first image of the heart of the patient; receiving a second image of the heart of the patient; extracting at least one numerical parameter from the second image of the heart of the patient; determining, at a predictive model, if the patient is a candidate for a medical procedure from a set of parameters comprising the plurality of numerical parameters extracted from the first image and the at least one numerical parameter extracted from the second image. [0060] Example 8. The method of Example 7, wherein the first image of the heart of the patient is a computed tomography image and the second image of the heart of the patient is an ultrasound image. [0061] Example 9. The method of Example 7, wherein the set of parameters comprises a projected area of a left ventricular outflow track (LVOT) and determining if the patient is a candidate for the medical procedure at the predictive model comprises: determining a projected area of the left ventricular outflow track (LVOT) from the first image; categorizing the patient as unsuitable for the medical procedure if the projected area of the LVOT is less than or equal to a first threshold value; determining a score for the patient from the set of parameters and the projected area of the LVOT if the projected area is greater than the first threshold value; and categorizing the patient as a candidate for the medical procedure if the score meets a second threshold value. [0062] Example 10. The method of Example 9, wherein the score for the patient is determined as a function of a thickness of a basal septum of the patient, a relative wall thickness of the patient, the projected area of the LVOT, and an estimated body surface area of the patient. [0063] Example 11. The method of Example 7, wherein the set of parameters includes at least two of an anterior leaflet length, an angle between a left ventricle (LV) and a left ventricle outflow tract (LVOT), an angle between the LV and a mitral valve (MV), an angle between an aortic valve (AV) and the MV, a parameter representing papillary muscle interaction, a parameter representing papillary muscle width, a chordae length, an anterior leaflet tip distance, a left ventricle height, a septal bulge thickness, a posterior wall thickness, a left ventricle end diastolic dimension, a relative wall thickness, a peak gradient of the mitral valve, an LV diameter, a distance from the MV to a septal bulge, a LV diameter at the septal bulge, a left ventricle end systolic diameter, an interventricular septum thickness, an LV volume, a projected area of the LVOT, and an LVOT volume. [0064] Example 12. The method of Example 7, wherein the procedure is a mitral valve repair, and the method further comprises performing the mitral valve repair if the patient is a candidate for the procedure. [0065] Example 13. A system comprising: a processor; and a non-transitory computer readable medium storing machine-readable instructions that are executable by the processor to determine if a patient is a candidate for a medical procedure, the machine-readable instructions comprising: a feature extractor that generates a plurality of numerical parameters from a first received image and generates at least one numerical parameter from a second received image to provide a set of numerical parameters; and a predictive model that determines if the patient is a candidate for the medical procedure from the set of numerical parameters. [0066] Example 14. The system of Example 13, wherein the plurality of numerical parameters includes a dimension of a structure within a heart of the patient and a projected area of a left ventricular outflow track (LVOT). [0067] Example 15. The system of Example 13, wherein the predictive model comprises: a first machine learning model that classifies the patient into one of a first class, representing patients who are candidates for the medical procedure, a second class, representing patients who are not candidates for the medical procedure, and a third class from at least one of the plurality of numerical parameters; and a second machine learning model that classifies patients who are classified into the third class into one of the first class and the second class from the set of numerical parameters. [0068] Example 16. The system of Example 13, wherein the predictive model determines if the patient is a candidate for a transcatheter mitral valve repair from the set of numerical parameters. [0069] Example 17. The system of Example 13, wherein the plurality of numerical parameters comprises a projected area of a left ventricular outflow track (LVOT) and a thickness of a basal septum of the patient and the at least one numerical parameter comprises a relative wall thickness associated with the patient, the predictive model determining if the patient is a candidate for the medical procedure from the relative wall thickness, the projected area of the LVOT, an estimated body surface area of the patient, and the thickness of a basal septum. [0070] Example 18. The system of Example 17, wherein the predictive model generates a score as a function of a value determined as a product of the relative wall thickness, a ration of the projected area of the LVOT to the estimated body surface area, and a square of the thickness of the basal septum and compares the score to a threshold value to determine if the patient is a candidate for the medical procedure. [0071] Example 19. A method comprising: receiving a first image of a heart of a patient; determining a projected area of a left ventricular outflow track (LVOT) after a medical procedure from the first image; categorizing the patient as a candidate for the medical procedure if the projected area of the LVOT is greater than or equal to a first threshold value; categorizing the patient as unsuitable for the medical procedure if the projected area of the LVOT is less than or equal to a second threshold value; performing the following steps if the projected area of the LVOT for the patient is between the first threshold value and the second threshold value: extracting a set of parameters, each of the set of parameters being extracted from one of the first image and a second image; generating a score for the patient from the set of parameters and the projected area of the LVOT; and categorizing the patient as a candidate for the medical procedure if the score meets a third threshold value. [0072] Example 20. The method of Example 19, wherein generating the score for the patient from the set of parameters and the projected area of the LVOT comprises providing the set of parameters and the projected area of the LVOT to a predictive model. [0073] Example 21. The method of Example 19, wherein extracting the set of parameters from one of the first image and the second image comprises extracting at least one of the set of parameters from the first image and extracting at least one of the set of parameters from the second image. [0074] Example 22. The method of Example 21, wherein the first image is a computed tomography image and the second image is an ultrasound image. [0075] Example 23. The method of Example 19, wherein generating the score for the patient from the set of parameters and the projected area of the LVOT comprises generating the score from a relative wall thickness associated with the patient, the projected area of the LVOT, an estimated body surface area of the patient, and a thickness of a basal septum of the patient. [0076] Example 24. The method of Example 23, wherein the generating the score for the patient from the set of parameters and the projected area of the LVOT comprises generating the score as a function of a value determined as a product of the relative wall thickness, a ration of the projected area of the LVOT to the estimated body surface area, and a square of the thickness of the basal septum. [0077] Example 25. The method of Example 19, wherein the medical procedure is a transcatheter mitral valve repair, and the method further comprises performing the transcatheter mitral valve repair if the patient is categorized as a candidate for the medical procedure.

Claims

What is claimed is: 1. A system comprising: a processor; and a non-transitory computer readable medium storing machine-readable instructions that are executable by the processor to determine if a patient is a candidate for a medical procedure, the machine-readable instructions comprising: a feature extractor that generates a plurality of numerical parameters from a received image, the plurality of numerical parameters including a dimension of a structure within a heart of the patient and a projected area of a left ventricular outflow track (LVOT) after the procedure; and a predictive model that determines if the patient is a candidate for the medical procedure from the plurality of numerical parameters.

2. The system of claim 1, wherein the received image is a first image and the feature extractor receiving a second image and extracting at least one numerical parameter from the second image, the predictive model determining if the patient is a candidate for the medical procedure from the plurality of numerical parameters and the at least one numerical parameter extracted from the second image.

3. The system of claim 2, wherein a set of parameters comprising the plurality of numerical parameters and the at least one numerical parameter extracted from the second image includes at least two of an anterior leaflet length, an angle between a left ventricle (LV) and the LVOT, an angle between the LV and a mitral valve (MV), an angle between an aortic valve (AV) and the MV, a parameter representing papillary muscle interaction, a parameter representing papillary muscle width, a chordae length, an anterior leaflet tip distance, a septal bulge thickness, a left ventricle height, a posterior wall thickness, a left ventricle end diastolic dimension, a relative wall thickness, a peak gradient of the mitral valve, an LV diameter, a distance from the MV to a septal bulge, a LV diameter at the septal bulge, a left ventricle end systolic diameter, an interventricular septum thickness, an LV volume, and an LVOT volume.

4. The system of claim 3, wherein the predictive model comprises: a first classifier that classifies the patient into one of a first class, representing patients who are candidates for the medical procedure, a second class, representing patients who are not candidates for the medical procedure, and a third class from a projected area of the LVOT; and a second classifier that classifies patients who are classified into the third class into one of the first class and the second class from the set of parameters and the projected area of the LVOT.

5. The system of claim 4, wherein the second classifier classifies patients who are classified into the third class into one of the first class and the second class from a thickness of a basal septum of the patient, a relative wall thickness of the patient, the projected area of the LVOT, and an estimated body surface area of the patient.

6. The system of any of claims 2-5, wherein the first image is a computed tomography image and the second image is an ultrasound image.

7. The system of any of claims 1-6, the machine-readable instructions further comprising a network interface that receives at least one biometric parameter representing the patient from an external system, the predictive model determining if the patient is a candidate for the medical procedure from the plurality of numerical parameters and the at least one biometric parameter.

8. The system of any of claims 1-7, wherein the plurality of numerical parameters comprises a projected area of a left ventricular outflow track (LVOT) and a thickness of a basal septum of the patient and the at least one numerical parameter comprises a relative wall thickness associated with the patient, the predictive model determining if the patient is a candidate for the medical procedure from the relative wall thickness, the projected area of the LVOT, an estimated body surface area of the patient, and the thickness of a basal septum.

9. The system of claim 8, wherein the predictive model generates a score as a function of a value determined as a product of the relative wall thickness, a ratio of the projected area of the LVOT to the estimated body surface area, and a square of the thickness of the basal septum and compares the score to a threshold value to determine if the patient is a candidate for the medical procedure.

10. The system of any of claims 1-9, wherein the medical procedure is a mitral valve repair 11. The system of claim 10, wherein the medical procedure is a transcatheter mitral valve repair. 12. A computer-implemented method comprising: receiving an image of a heart of a patient; extracting a plurality of numerical parameters from the first image of the heart of the patient, the plurality of numerical parameters including a dimension of a structure within a heart of the patient and a projected area of a left ventricular outflow track (LVOT) after the procedure; determining, at a predictive model, if the patient is a candidate for a medical procedure from the plurality of numerical parameters 13. The method of claim 12, wherein the image is a computer tomography image and the method further comprises: receiving an ultrasound image of the heart of the patient; and extracting at least one numerical parameter from the ultrasound image; wherein determining if the patient is a candidate for the medical procedure comprises determining if the patient is a candidate for the medical procedure comprises from the plurality of numerical parameters and the at least one numerical parameter. 14. The method of claim 12 or claim 13, wherein determining if the patient is a candidate for the medical procedure comprises: categorizing the patient as a candidate for the medical procedure if the projected area of the LVOT is greater than or equal to a first threshold value; categorizing the patient as unsuitable for the medical procedure if the projected area of the LVOT is less than or equal to a second threshold value; and performing the following steps if the projected area of the LVOT for the patient is between the first threshold value and the second threshold value: generating a score for the patient from at least the plurality of numerical parameters; and categorizing the patient as a candidate for the medical procedure if the score meets a third threshold value. 15. The method of claim 14, wherein the score for the patient is determined as a function of a thickness of a basal septum of the patient, a relative wall thickness of the patient, the projected area of the LVOT, and an estimated body surface area of the patient.