Identification of ferrous metal from single-energy CT image data using machine learning

Abstract

Ferrous and non-ferrous materials are identified and classified in a subject using a machine learning model that receives single-energy CT image data as an input. The machine learning model may be trained on single-energy CT image data. Additionally or alternatively, the machine learning model may be trained on dual-energy CT image data, geometrical data about metal objects, CT number data derived from the single-energy and / or dual-energy CT image data, and / or acquisition parameter data pertaining to acquisition parameters used when acquiring the single-energy and / or dual-energy CT image data. Single-energy CT image data acquired from a subject is input to the machine learning model, generating classified feature data as an output. The classified feature data indicate the presence of ferrous metal in the subject.

Description

IDENTIFICATION OF FERROUS METAL FROM SINGLE-ENERGY CT IMAGE DATA USING MACHINE LEARNINGCROSS-REFERENCE TO RELATED APPLICATIONS

[0001] This application claims the benefit of U.S. Provisional Patent Application SerialNo. 63 / 515,126, filed on July 23, 2023, and entitled “IDENTIFICATION OF FERROUS METAL FROM SINGLE-ENERGY CT IMAGE DATA USING MACHINE LEARNING." which is herein incorporated by reference in its entirety.BACKGROUND

[0002] Patients with implanted medical devices or shrapnel may encounter safety issues or limitations if they are scheduled for magnetic resonance imaging (MRI) studies since ferrous metals contained in medical implants or shrapnel may introduce risks such as material movement or device malfunction during an MRI exam. Normally, medical implants come with a vendor-provided document that contains safety guidelines to minimize these risks. However, there are patients whose documents are unavailable. Similarly, for patients with detained metal fragments, the components of fragments are unknown in most of the events. Risks of undergoing MRI examination for these patients are unpredictable and may potentially be severe. Therefore, it is clinically imperative to have a tool to identify if these undocumented medical implants or shrapnel contain ferromagnetic components before such patients undergo MRI exams.

[0003] Current methods for medical device clearance in routine clinical practice using radiography or computer tomography (CT) are effective for identifying the presence and location of device, but limited in their capability to identify if the device contains ferrous components. Conventional single-energy CT images have ambiguity in differentiating ferrous from non-ferrous materials based on CT number measures such as mean and standard deviation. In addition, CT number of metals varies with different protocols in the fashion of intra-scanner and inter-scanner variability, leading to poor accuracy in metal type differentiation.SUMMARY OF THE DISCLOSURE

[0004] It is an aspect of the present disclosure to provide a method for identifying ferrous metals from single-energy computed tomography (CT) image data. The methodincludes accessing single-energy CT image data with a computer system, where the singleenergy CT image data have been acquired from a subject. A machine learning model is also accessed with the computer system, where the machine learning model has been trained on training data to receive single-energy CT image data as an input and generate classified feature data indicating a presence of ferrous metal as an output. The single-energy CT image data are then applied to the machine learning model using the computer system, generating classified feature data as an output. The generated classified feature data indicate a presence of ferrous metal in the subject. The classified feature data may then be presented to a user by the computer system.

[0005] It is another aspect of the present disclosure to provide a method for identifying ferrous metals from single-energy CT image data by accessing single-energy CT image data with a computer system, where the single-energy CT image data have been acquired from a subject with a CT system and determining, with the computer system, a region-of-interest (ROI) in the single-energy CT image data that contains a metal object. CT number data are generated by determining CT number statistical metrics from the single-energy CT image data in the ROI. Acquisition parameter data are also received with the computer system, where the acquisition parameter data include data acquisition parameters used by the CT system when acquiring the single-energy CT image data. A machine learning model is accessed with the computer system, where the machine learning model has been trained on training data to indicate a presence of ferrous metal in an object based on CT number statistical metrics and data acquisition parameters. The CT number data and acquisition parameter data are input to the machine learning model using the computer system, generating classified feature data as an output. The classified feature data indicate a presence of ferrous metal in the metal object. The classified feature data may then be output to a user by the computer system. Other embodiments of this aspect include corresponding systems (e g., computer systems), programs, algorithms, and / or modules, each configured to perform the steps of the methods.BRIEF DESCRIPTION OF THE DRAWINGS

[0006] FIG. 1 is a flow chart setting forth the steps of an example method for identifying whether a subject contains ferrous metals by classifying single-energy CT images as containing ferrous metal, non-ferrous metal, or no metal, using a suitably trained machine learning model.

[0007] FIG. 2 is a flowchart setting forth the steps of an example method for training a machine learning model to differentiate between ferrous and non-ferrous metals based on single-energy CT images.

[0008] FIG. 3 shows confusion matrices demonstrating an evaluation of a machine learning model used to differentiate between ferrous and non-ferrous metals in single-energy CT images in an example study.

[0009] FIG. 4 illustrates example of single-energy CT images from subject having an implanted metal object. In each image, a region-of-interest (ROI) is shown as containing the identified metal objects.

[0010] FIG. 5 is a block diagram of an example system for classifying metallic objects in single-energy CT images as ferrous or non-ferrous.

[0011] FIG. 6 is a block diagram of example components that can implement the system of FIG. 5.

[0012] FIGS. 7 A and 7B illustrate an example CT system that can implement the methods described in the present disclosure.DETAILED DESCRIPTION

[0013] Described here are systems and methods for identifying and classifying ferrous and non-ferrous materials in a subject using a machine learning model that receives singleenergy CT image data as an input. Additionally or alternatively, the machine learning model may receive one or more metrics derived from the single-energy CT image data as an input. For example, the machine learning model may additionally or alternatively receive CT number data and / or geometrical data associated with metal objects in the subject as input data. The machine learning model may also additionally or alternatively receive as an input acquisition parameter data indicating the data acquisition parameters used to acquire the single-energy7CT image data.

[0014] As a non-limiting example, training data used to train such a machine learning model can be assembled by performing CT experiments on anthropomorphic phantoms with different metallic rods and analyzing retrospective CT image data in patients at different CT scanners (e.g., single-energy CT scanners, dual-energy CT scanners, or both). In this way, a training database of metals with geometrical metrics (size and distance to isocenter), CT number histogram (maximum, 90%, 80%, 70%, 60%, 50%, mean, standard deviation, skewness, kurtosis, entropy, energy, etc.) and protocol parameters (CT scanner station name,CT scanner model name, body part examined, scan option, kV / keV, effective mAs, bow tie filter, convolution kernel, image reconstruction algorithm, collimation width, slice thickness, etc.) may be established and used to train a machine learning model to differentiate ferrous and non-ferrous materials from single-energy CT image data input to the model. Additionally or alternatively, a machine learning model may incorporate a subset of the features described above (e.g., geometrical metrics, CT number histogram, protocol parameters) and trained on the training database.

[0015] Referring now to FIG. 1, a flowchart is illustrated as setting forth the steps of an example method for generating classified feature data using a suitably trained machine learning model (e g., a neural network or other machine learning algorithm). As will be described, the machine learning model takes single-energy CT image data as input data and generates classified feature data as output data. As an example, the classified feature data can indicate regions in the single-energy CT image data that contain ferrous metal, non-ferrous metal, or no metal at all.

[0016] The method includes accessing single-energy CT image data with a computer system, as indicated at step 102. Accessing the single-energy CT image data may include retrieving such data from a memory or other suitable data storage device or medium. Additionally or alternatively, accessing the single-energy CT image data may include acquiring such data with a CT system and transferring or otherwise communicating the data to the computer system, which may be a part of the CT system.

[0017] In some examples, additional relevant data may also be accessed by the computer system. For instance, CT number data, geometrical data, and / or acquisition parameter data may be accessed and used as additional inputs to the machine learning model. For instance, CT numbers associated with different acquisition parameters may be incorporated into the feature space of a machine learning-based classification learner. Geometrical data may include size (e.g., diameter) and distance to isocenter (e.g., locations) of regions segmented or otherwise identified in single-energy CT images as corresponding to probable metal objects. CT number data may include CT numbers computed from the single-energy’ CT image data, or may include data computed or otherwise obtained from a CT number histogram (e.g., maximum, 90%, 80%, 70%, 60%, 50%, mean, standard deviation, skewness, kurtosis, entropy, energy, etc.). Acquisition parameter data may include CT scanner station name, CT scanner model name, body part examined, scan option, kV / keV, effective mAs, bow tie filter, convolution kernel, image reconstruction algorithm, collimation width, slice thickness, etc.Additionally or alternatively, the additional relevant data may include dual-energy CT image data acquired from the same subject as the single-energy CT image data. In these instances, the dual-energy CT image data may include complementary information to the single-energy CT image data, which may improve the classification accuracy of the machine learning model(s).

[0018] A trained machine learning model (e.g., a neural network or other suitable machine learning algorithm) is then accessed with the computer system, as indicated at step 104. In general, the machine learning model is trained, or has been trained, on training data in order to classify metal containing regions of the single-energy CT image data as being associated with ferrous metal or non-ferrous metal.

[0019] Accessing the trained machine learning model may include accessing model parameters (e.g.. weights, biases, or both, for a neural network) that have been optimized or otherwise estimated by training the machine learning model on training data. In some instances, retrieving the machine learning model can also include retrieving, constructing, or otherwise accessing the particular model architecture to be implemented. For instance, data pertaining to the layers in a 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.

[0020] As one example, the machine learning model may be an ensemble classifier constructed for training and prediction. In some implementations, the ensemble classifier may employ a decision tree as the base classifier, bagging and adaptive boosting as the ensemble method, and Bayesian optimization as the optimizer.

[0021] As another example, the machine learning model may be a neural network. An artificial 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.

[0022] 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.

[0023] 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.

[0024] 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 classifies regions in single-energy CT image data as containing ferrous metal or non-ferrous metal, the output layer may include, for example, a number of different nodes, where each different node corresponds to a different classification. A first node may indicate pixels or regions containing ferrous metal, a second node may indicate pixels or regions containing non-ferrous metal, and a third node may indicate pixels or regions that do not contain metal.

[0025] The single-energy CT image data are then input to the trained machine learning model, generating output as classified feature data, as indicated at step 106. As described above, in some implementations additional data may also be input to the machine learning model, including CT number data, geometrical data, and / or acquisition parameter data. These additional data can be extracted, derived, or otherwise generated from the single-energy CT image data. For instance, a region-of-interest can be placed on a suspected implant to extract the geometrical and CT number statistical metrics from the ROI, which can then be combinedwith the data acquisition parameters used to acquire the single-energy CT image data to construct a feature vector that can be input to the machine learning model.

[0026] The classified feature data may indicate the probability for a particular classification (i.e., the probability that the single-energy CT image data include ferrous metal and / or non-ferrous metal). Additionally or alternatively, the classified feature data may classify the single-energy’ CT image data as having images with pixels or regions containing ferrous metal and / or non-ferrous metal. As one example, the classified feature data may include labeling the single-energy CT image data as containing ferrous metal or not containing ferrous metal. As another example, the classified feature data may include a classification map whose pixels indicate a classification (e.g., ferrous metal, non-ferrous metal, no metal) of corresponding pixels in a single-energy CT image.

[0027] The classified feature data generated by inputting the single energy CT image data to the trained machine learning model(s) can then be displayed to a user, stored for later use or further processing, or both, as indicated at step 108. As an example, the single-energy CT images can be displayed together with the classified feature data. For instance, the classified feature data may include a classification label for single-energy CT images, which can indicate those images identified as containing pixels or regions in the subject that contain ferrous metal. Additionally or alternatively, the classified feature data may be overlaid with the single-energy CT images to visualize the pixels and / or regions in the single-energy’ CT images that contain ferrous metal, non-ferrous metal, or no metal. In some implementations, the classified feature data may’ include a classification map whose pixels correspond to different classes, such as ferrous metal, non-ferrous metal, and no metal. Such a classification map may be overlaid on the corresponding single-energy CT image to visualize to the clinician those regions containing ferrous metal and / or non-ferrous metal.

[0028] Referring now to FIG. 2, a flowchart is illustrated as setting forth the steps of an example method for training one or more machine learning models (e.g., neural networks or other suitable machine learning algorithms) on training data, such that the one or more machine learning models are trained to receive single-energy CT images as input data in order to generate classified feature data as output data, where the classified feature data are indicative of pixels or regions in the single-energy CT images that are classified as containing ferrous metal, non-ferrous metal, or no metal.

[0029] In general, the machine learning model(s) may be neural network(s) that can implement any number of different neural network architectures. For instance, the neuralnetwork(s) could implement a convolutional neural network, a residual neural network, 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. For example, the machine learning model may additionally or alternatively include an ensemble classifier. The ensemble classifier may be based on one or more decision trees.

[0030] 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 a CT system and transferring or otherwise communicating the data to the computer system.

[0031] In general, the training data can include single-energy CT images and dualenergy CT images. Alternatively, the training data may include only single-energy CT images. Additionally, whether the training data includes single-energy CT images or both single-energy and dual-energy CT images, the training data may also include other data, such as CT number data, geometrical data, acquisition parameter data, and so on. In some embodiments, the training data may include single-energy' CT images and / or dual-energy CT images that have been labeled (e.g., labeled as containing patterns, features, or characteristics indicative of pixels or regions containing ferrous metal, non-ferrous metal, or no metal: and the like).

[0032] The method can include assembling training data from single-energy CT images and / or dual-energy' CT images using a computer system. For instance, the single-energy CT images and / or dual-energy CT images may have been acquired from one or more phantoms containing different types of metal common the medical implants (e.g., aluminum, copper, stainless steel, titanium). For example, four types of metallic rods (aluminum (Al), titanium (Ti), stainless steel (SS), copper (Cu)) with different diameters (10, 9, 8,7, 6, 5, 4, 3, 2 and 1 mm) may be collected as inserts into 16 / 32 cm diameter acrylic or anthropomorphic phantoms for single-energy' and dual-energy CT image data acquisition. The measured histogram measures of CT numbers (e.g., maximum, 90%, 80%, 70%, 60%, 50%. mean, standard deviation, skewness, kurtosis, entropy, energy, etc.) from metallic rods and associated acquisition parameters may also be used as features of a training dataset. For example, these histogram measures of CT numbers from the metallic rods associated w ith the rod diameter and location as well as acquisition parameters such as scanner model, body region imaged, scan option, pitch, kV. keV. FOV. bowtie filer, reconstruction kernel, metal rod diameter andlocation, collimation, and slice thickness, etc., may be selected as features for the machine learning model-based classifier.

[0033] Additionally or alternatively, the single-energy CT images and / or dual-energy CT images may have been acquired from subject having metallic implants of a known composition.

[0034] Assembling the training data may include assembling the single-energy CT images and / or dual-energy CT images into an appropriate data structure on which the machine learning model can be trained. Assembling the training data may include assembling singleenergy CT images and / or dual-energy CT images, segmented single-energy' CT images and / or dual-energy’ CT 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 single-energy CT images and / or dual-energy CT images, segmented single-energy CT images and / or dual-energy CT 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 single-energy CT images and / or dual-energy CT images and / or segmented single-energy CT images and / or dual-energy CT images that have pixels and / or regions that have been labeled as being associated with ferrous metal, non-ferrous metal, or no metal.

[0035] In some embodiments, the training data may include CT number data and acquisition parameter data extracted, derived, or otherwise generated from the single-energy CT images and / or dual-energy' CT images. Additionally, the training data may include geometrical data extracted, derived, or otherwise generated from the single-energy’ CT images and / or dual-energy CT images. In these instances, the single-energy' and / or dual-energy CT images may not be stored as part of the training data.

[0036] One or more machine learning models (e.g., neural networks or other suitable machine learning algorithms) are trained on the training data, as indicated at step 204. In general, the machine learning model can be trained by optimizing model parameters (e.g., weights, biases, or both for a neural network) based on minimizing a loss function. As one nonlimiting example, the loss function may be a mean squared error loss function.

[0037] Training a neural network may include initializing the neural network, such as by computing, estimating, or otherwise selecting initial network parameters (e.g., weights, biases, or both). During training, an artificial neural network receives the inputs for a training example and generates an output using the bias for each node, and the connections betweeneach node and the corresponding weights. For instance, training data can be input to the initialized neural network, generating output as classified feature data. 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 classified feature data. For instance, the classified feature data 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 parameters (e.g., weights, biases, or both) 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.

[0038] 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.

[0039] Additionally or alternatively, the machine learning model may be an ensemble classifier, such as an ensemble classifier based on one or more decision trees. As anon-limiting example, such a classifier can be trained on training data that includes CT number data and acquisition parameter data. Alternatively, the training data may also include geometrical data. In these instances, the ensemble classifier can be trained on training data by training multiple different machine learning models (e.g.. multiple different decision trees, other types of classifier models). Each of these models are then used to generate predictions based on thetraining data. The resulting predictions can be combined in the ensemble model to generate the final ensemble classifier that can be deployed in an inference stage to generate classified feature data by inputting new data to the ensemble classifier.

[0040] The one or more trained machine learning models are then stored for later use, as indicated at step 206. Storing the machine learning model(s) may include storing model parameters (e.g., weights, biases, or both for a neural network), which have been computed or otherwise estimated by training the machine learning model(s) on the training data. Storing the trained machine learning model(s) may also include storing the particular model architecture to be implemented. For instance, data pertaining to the layers in a neural network architecture (e.g., number of layers, type of layers, ordering of layers, connections between layers, hyperparameters for layers) may be stored.

[0041] In an example study, anthropomorphic phantoms with imbedded cast-off medical devices and ferrous / nonferrous rods were fabricated to mimic scenarios of patients with implants. Data acquisitions were performed with routine clinical protocols on singleenergy (SECT) and dual-energy CT (DECT) scanners. For each image, a region-of-interest (ROI) was placed on each metal object to extract its geometrical and CT number statistical metrics, which were combined with data acquisition parameters to construct a feature vector in a machine learning framework. An ensemble classifier based on decision trees was constructed for training and testing. A cross-validation method with a confusion matrix was utilized to evaluate classifier performance. Data from 29 clinical patient cases with medical implants were used for evaluation.

[0042] For the SECT images, the classifier had an accuracy of 72.6% (a single feature: mean CT number) versus 84.5% (multiple features). For the DECT images, the classifier had an accuracy of 76.5% (a single feature: mean CT number) versus 90.2% (multiple features). Out of 29 patient cases, the classifier identified 24 cases as containing nonferrous metals and 5 cases as containing ferrous metals. An example performance of the classifier is illustrated in FIG. 3. Representative images from the patients are illustrated in FIG. 4. The ROI containing an identified metal object is shown on each image.

[0043] Thus, the systems and methods described in the present application can effectively increase the accuracy of differentiation between ferrous and non-ferrous metals in medical implants with CT images by incorporating multiple features other than mean CT number into a machine learning model framework.

[0044] FIG. 5 shows an example of a system 500 for identifying ferrous metals in a subject in accordance with some embodiments of the systems and methods described in the present disclosure. As shown in FIG. 5, a computing device 550 can receive one or more types of data (e.g., single-energy' CT image data, acquisition parameter data, dual-energy' CT image data) from data source 502. In some embodiments, computing device 550 can execute at least a portion of a ferrous metal identification and classification system 504 to identify ferrous metals in a subject by classifying single-energy CT image data received from the data source 502.

[0045] Additionally or alternatively, in some embodiments, the computing device 550 can communicate information about data received from the data source 502 to a server 552 over a communication network 554. which can execute at least a portion of the ferrous metal identification and classification system 504. In such embodiments, the server 552 can return information to the computing device 550 (and / or any other suitable computing device) indicative of an output of the ferrous metal identification and classification system 504.

[0046] In some embodiments, computing device 550 and / or server 552 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, and so on. The computing device 550 and / or server 552 can also reconstruct images from the data.

[0047] In some embodiments, data source 502 can be any suitable source of data (e.g.. measurement data, images reconstructed from measurement data, processed image data), such as a single-energy CT system, another computing device (e.g., a server storing measurement data, images reconstructed from measurement data, processed image data), and so on. In some embodiments, data source 502 can be local to computing device 550. For example, data source 502 can be incorporated with computing device 550 (e.g., computing device 550 can be configured as part of a device for measuring, recording, estimating, acquiring, or otherwise collecting or storing data). As another example, data source 502 can be connected to computing device 550 by a cable, a direct wireless link, and so on. Additionally or alternatively, in some embodiments, data source 502 can be located locally and / or remotely from computing device 550, and can communicate data to computing device 550 (and / or server 552) via a communication network (e.g., communication network 554).

[0048] In some embodiments, communication network 554 can be any suitable communication network or combination of communication networks. For example.communication network 554 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 network, a wired network, and so on. In some embodiments, communication network 554 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 network, or any suitable combination of networks. Communications links show n in FIG. 5 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.

[0049] Referring now to FIG. 6, an example of hardware 600 that can be used to implement data source 502, computing device 550, and server 552 in accordance with some embodiments of the systems and methods described in the present disclosure is shown.

[0050] As shown in FIG. 6, in some embodiments, computing device 550 can include a processor 602, a display 604, one or more inputs 606, one or more communication systems 608, and / or memory 610. In some embodiments, processor 602 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 604 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 606 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.

[0051] In some embodiments, communications systems 608 can include any suitable hardware, firmware, and / or software for communicating information over communication network 554 and / or any other suitable communication networks. For example, communications systems 608 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 608 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.

[0052] In some embodiments, memory' 610 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 602 to present content using display 604. to communicate with server552 via communications system(s) 608, and so on. Memory 610 can include any suitable volatile memory, non-volatile memory, storage, or any suitable combination thereof. For example, memory 610 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 610 can have encoded thereon, or otherwise stored therein, a computer program for controlling operation of computing device 550. In such embodiments, processor 602 can execute at least a portion of the computer program to present content (e.g., images, user interfaces, graphics, tables), receive content from server 552, transmit information to server 552, and so on. For example, the processor 602 and the memory 610 can be configured to perform the methods described herein (e.g., the method of FIG. 1, the method of FIG. 2).

[0053] In some embodiments, server 552 can include a processor 612, a display 614, one or more inputs 616. one or more communications systems 618. and / or memory 620. In some embodiments, processor 612 can be any suitable hardware processor or combination of processors, such as a CPU, a GPU, and so on. In some embodiments, display 614 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 616 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.

[0054] In some embodiments, communications systems 618 can include any suitable hardware, firmware, and / or software for communicating information over communication network 554 and / or any other suitable communication networks. For example, communications systems 618 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 618 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.

[0055] In some embodiments, memory 620 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 612 to present content using display 614, to communicate with one or more computing devices 550, and so on. Memory 620 can include any suitable volatile memory, non-volatile memory, storage, or any suitable combination thereof. For example.memory 620 can include RAM, ROM, EPROM, EEPROM, other types of volatile memory', other types 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' 620 can have encoded thereon a server program for controlling operation of server 552. In such embodiments, processor 612 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 550, receive information and / or content from one or more computing devices 550, receive instructions from one or more devices (e.g., a personal computer, a laptop computer, a tablet computer, a smartphone), and so on.

[0056] In some embodiments, the server 552 is configured to perform the methods described in the present disclosure. For example, the processor 612 and memory 620 can be configured to perform the methods described herein (e.g., the method of FIG. 1, the method of FIG. 2).

[0057] In some embodiments, data source 502 can include a processor 622, one or more data acquisition systems 624, one or more communications systems 626. and / or memory 628. In some embodiments, processor 622 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 624 are generally configured to acquire data, images, or both, and can include a single-energy’ CT system and / or a dual-energy CT system. Additionally or alternatively, in some embodiments, the one or more data acquisition systems 624 can include any suitable hardware, firmware, and / or software for coupling to and / or controlling operations of a single-energy CT system and / or a dual-energy CT system. In some embodiments, one or more portions of the data acquisition system(s) 624 can be removable and / or replaceable.

[0058] Note that, although not shown, data source 502 can include any suitable inputs and / or outputs. For example, data source 502 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 502 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.

[0059] In some embodiments, communications systems 626 can include any suitable hardware, firmware, and / or software for communicating information to computing device 550 (and, in some embodiments, over communication network 554 and / or any other suitable communication networks). For example, communications systems 626 can include one or moretransceivers, one or more communication chips and / or chip sets, and so on. In a more particular example, communications systems 626 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.

[0060] In some embodiments, memory 628 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 622 to control the one or more data acquisition systems 624, and / or receive data from the one or more data acquisition systems 624; to generate images from data; present content (e.g.. data, images, a user interface) using a display; communicate with one or more computing devices 550; and so on. Memory 628 can include any suitable volatile memory, non-volatile memory, storage, or any suitable combination thereof. For example, memory' 628 can include RAM, ROM, EPROM, EEPROM, other types of volatile memory, other types 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 628 can have encoded thereon, or otherwise stored therein, a program for controlling operation of data source 502. In such embodiments, processor 622 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 550. receive information and / or content from one or more computing devices 550. receive instructions from one or more devices (e.g., a personal computer, a laptop computer, a tablet computer, a smartphone, etc.), and so on.

[0061] 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-transitory computer-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 memory, EPROM. EEPROM), any suitable media that is not fleeting or devoid of any semblance of permanence during transmission, 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.

[0062] 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).

[0063] 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.

[0064] Referring particularly now to FIGS. 7A and 7B, an example of an x-ray computed tomography (“CT”) imaging system 700 is illustrated. The CT system includes a gantry 702. to which at least one x-ray source 704 is coupled. The x-ray source 704 projects an x-ray beam 706, which may be a fan-beam or cone-beam of x-rays, towards a detector array 708 on the opposite side of the gantry 702. The detector array 708 includes a number of x-ray detector elements 710. Together, the x-ray detector elements 710 sense the projected x-rays 706 that pass through a subject 712, such as a medical patient or an object undergoing examination, that is positioned in the CT system 700. Each x-ray detector element 710 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 712. In some configurations, each x- ray detector 710 is capable of counting the number of x-ray photons that impinge upon thedetector 710. During a scan to acquire x-ray projection data, the gantry 702 and the components mounted thereon rotate about a center of rotation 714 located within the CT system 700.

[0065] The CT system 700 also includes an operator workstation 716, which typically includes a display 718; one or more input devices 720, such as a keyboard and mouse; and a computer processor 722. The computer processor 722 may include a commercially available programmable machine running a commercially available operating system. The operator workstation 716 provides the operator interface that enables scanning control parameters (e.g., acquisition parameters) to be entered into the CT system 700. In general, the operator workstation 716 is in communication with a data store server 724 and an image reconstruction system 726. By way of example, the operator workstation 716. data store server 724, and image reconstruction system 726 may be connected via a communication system 728, which may include any suitable network connection, whether wired, wireless, or a combination of both. As an example, the communication system 728 may include both proprietary or dedicated networks, as well as open networks, such as the internet.

[0066] The operator workstation 716 is also in communication with a control system 730 that controls operation of the CT system 700. The control system 730 generally includes an x-ray controller 732, a table controller 734, a gantry controller 736, and a data acquisition system 738. The x-ray controller 732 provides power and timing signals to the x-ray source 704 and the gantry controller 736 controls the rotational speed and position of the gantry 702. The table controller 734 controls a table 740 to position the subject 712 in the gantry 702 of the CT system 700.

[0067] The DAS 738 samples data from the detector elements 710 and converts the data to digital signals for subsequent processing. For instance, digitized x-ray data is communicated from the DAS 738 to the data store server 724. The image reconstruction system 726 then retrieves the x-ray data from the data store server 724 and reconstructs an image therefrom. The image reconstruction system 726 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 722 in the operator workstation 716. Reconstructed images can then be communicated back to the data store server 724 for storage or to the operator workstation 716 to be displayed to the operator or clinician.

[0068] The CT system 700 may also include one or more networked workstations 742. By way of example, a networked workstation 742 may include a display 744; one or more inputdevices 746, such as a keyboard and mouse; and a processor 748. The networked workstation 742 may be located within the same facility as the operator workstation 716, or in a different facility, such as a different healthcare institution or clinic.

[0069] The networked workstation 742, whether within the same facility’ or in a different facility as the operator workstation 716, may gain remote access to the data store server 724 and / or the image reconstruction system 726 via the communication system 728. Accordingly, multiple networked workstations 742 may have access to the data store server 724 and / or image reconstruction system 726. In this manner, x-ray data, reconstructed images, or other data may be exchanged between the data store server 724, the image reconstruction system 726, and the networked workstations 742, such that the data or images may be remotely processed by a networked workstation 742. This data may be exchanged in any suitable format, such as in accordance with the transmission control protocol (“TCP”), the internet protocol (“IP”), or other known or suitable protocols.

[0070] 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 identifying ferrous metals from single-energy computed tomography (CT) image data, the method comprising:(a) accessing single-energy CT image data with a computer system, wherein the single-energy' CT image data have been 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 singleenergy CT image data as an input and generate classified feature data indicating a presence of ferrous metal as an output;(c) applying the single-energy CT image data to the machine learning model using the computer system, generating classified feature data as an output, wherein the classified feature data indicate a presence of ferrous metal in the subject; and(d) presenting the classified feature data to a user by the computer system.

2. The method of claim 1. wherein the machine learning model comprises an ensemble classifier model.

3. The method of claim 2. wherein the ensemble classifier model comprises a decision tree-based model.

4. The method of claim 3, wherein the ensemble classifier model implements a bagging and adaptive boosting ensemble method.

5. The method of claim 4, wherein the ensemble classifier model is optimized using a Bayesian optimizer.

6. The method of claim 1. wherein the machine learning model comprises a neural network.

7. The method of claim 1, wherein the single-energy CT image data comprise single-energy CT images acquired from the subject.

8. The method of claim 7. wherein the single-energy CT image data further comprise CT numbers estimated from the single-energy CT images.

9. The method of claim 8, wherein the single-energy CT image data further comprise at least one metric computed from a CT number histogram generated from the CT numbers.

10. The method of claim 9, wherein the metric computed from the CT number histogram comprises at least one of maximum, percentile, mean, standard deviation, skewness, kurtosis, entropy, or energy.

11. The method of claim 1 , comprising accessing acquisition parameter data indicating acquisition parameters used when acquiring the single-energy CT image data from the subject and inputting the acquisition parameter data as an additional input to the machine learning model.

12. The method of claim 11, wherein the acquisition parameter data include at least one of CT scanner model of the CT system used to acquire the single-energy CT data, body region of the subject imaged, scan option, pitch. kV / keV, effective mAs, field-of-view. bowtie filer, reconstruction kernel, collimation width, or slice thickness.

13. The method of claim 1. comprising accessing geometrical data indicating at least one of a size or location of metal objects in the subject, and inputting the geometrical data as an additional input to the machine learning model.

14. The method of claim 1, wherein the training data comprise at least singleenergy CT image data acquired from a group of subjects.

15. The method of claim 14, wherein the training data further comprise dualenergy CT image data acquired from the group of subjects.

16. The method of claim 14, wherein the group of subjects comprises a plurality of imaging phantoms including one or more metallic objects.

17. A method for identity'ing ferrous metals from single-energy computed tomography (CT) image data, the method comprising:(a) accessing single-energy CT image data with a computer system, wherein the single-energy CT image data have been acquired from a subject with a CT system;(b) determining, with the computer system, a region-of-interest (ROI) in the single-energy CT image data that contains a metal object;(c) generating, with the computer system, CT number data by determining CT number statistical metrics from the single-energy CT image data in the ROI;(d) receiving acquisition parameter data with the computer system, wherein the acquisition parameter data comprise data acquisition parameters used by the CT system when acquiring the single-energy CT image data;(e) accessing a machine learning model with the computer system, wherein the machine learning model has been trained on training data to indicate a presence of ferrous metal in an object based on CT number statistical metrics and data acquisition parameters;(1) applying the CT number data and acquisition parameter data to the machine learning model using the computer system, generating classified feature data as an output, wherein the classified feature data indicate a presence of ferrous metal in the metal object; and(g) outputting the classified feature data to a user by the computer system.

18. The method of claim 17, further comprising constructing a feature vector with the computer system by combining the CT number data and acquisition parameter data in the feature vector, wherein applying the CT number data and data acquisition parameter data to the machine learning model comprising inputting the feature vector to the machine learning model.

19. The method of claim 17, wherein the machine learning model comprises an ensemble classifier.

20. The method of claim 19, wherein the ensemble classifier is based on a decision tree model.

21. The method of claim 17, wherein the CT number statistical metrics comprise at least one metric computed from a CT number histogram generated from CT numbers estimated from the single-energy CT image data in the ROI.

22. The method of claim 21, wherein the at least one metric computed from the CT number histogram comprises at least one of maximum, percentile, mean, standard deviation, skewness, kurtosis, entropy, or energy.

23. The method of claim 17, wherein the acquisition parameter data include at least one of CT scanner model of the CT system used to acquire the single-energy CT data, body region of the subject imaged, scan option, pitch. kV / keV, effective mAs, field-of-view. bowtie filer, reconstruction kernel, collimation width, or slice thickness.

24. The method of claim 17, further comprising generating, with the computer system, geometrical data associated with the metal object from the single-energy CT image data in the ROI, wherein the geometncal data are input to the machine learning model as an additional input.

25. The method of claim 24, wherein the geometrical data indicate at least one of a size or location of the metal object.

Images