Rotational scan spin delay with injector coupling

Abstract

An imaging system includes a processor configured to receive an X-ray image of a subject acquired using an X-ray imaging device; extract one or more features of the subject from the X-ray image; predict a spin initiation time delay for the X-ray imaging device based on the extracted one or more features of the subject; determine an injection time of administration of an intravascular contrast agent injection to the subject; and initiate acquisition of an angiography image of the subject using the X-ray imaging device at a start time determined based on the injection time and the spin initiation time delay.

Description

ROTATIONAL SCAN SPIN DELAY WITH INJECTOR COUPLINGFIELD

[0001] The following relates generally to the medical imaging arts, angiographic imaging arts, X-ray imaging arts, C-arm imaging arts, and related arts.BACKGROUND

[0002] Interventional procedures often require three-dimensional (3D) image acquisition at particular phases of the procedure. For example, in many endovascular procedures contrast is injected into the patient vasculature during 3D image acquisition in order to visualize the vasculature in 3D. The ability to visualize the vasculature in 3D allows physicians to measure the target to be treated and plan treatment delivery. For instance, when treating an intracranial aneurysm by a coiling procedure, a 3D image of the aneurysm allows physicians to measure the neck and dome size of the aneurysm, which in turn allows them to select the appropriate size coils to be inserted into the aneurysm. Higher quality 3D images allow for more accurate measurements and, therefore, can allow physicians to make better and more confident treatment decisions.

[0003] 3D image acquisition with a C-arm X-ray imaging system utilizes a C-arm spin, during which X-ray images are acquired from various different angles and are used to reconstruct the 3D image. When contrast is injected into the vasculature during this spin in order to acquire a 3D rotational angiography (3DRA) image, the contrast injection and the C-arm spin should be timed accurately such that the image acquisition starts when the contrast is in the region of interest (ROI) being imaged. If the spin starts too early, images will be acquired without contrast in the ROI and cannot be used in the 3DRA reconstruction. This exposes the patient to unnecessary radiation. If the spin starts too late, contrast may reach the ROI too early, significantly before images are acquired, resulting in fewer images acquired during the spin with contrast in the ROI which can produce lower quality 3DRA images. This can result in users needing to re-acquire 3DRA images, exposing the patient to additional radiation and contrast.

[0004] The following discloses certain improvements to overcome these problems and others.SUMMARY

[0005] In some embodiments disclosed herein, an imaging system includes a processor communicatively coupled to memory. The processor is configured to receive an X-ray image of a subject acquired using an X-ray imaging device; extract one or more features of the subject from the X-ray image; predict a spin initiation time delay for the X-ray imaging device based on the extracted one or more features of the subject; determine an injection time of administration of an intravascular contrast agent injection to the subject; and initiate acquisition of an angiography image of the subject using the X-ray imaging device at a start time determined based on the injection time and the spin initiation time delay.

[0006] In some embodiments disclosed herein, an imaging method includes receiving an X-ray image of a subject acquired using an X-ray imaging device; extracting one or more features of the subject from the X-ray image; predicting a spin initiation time delay for the X-ray imaging device based on the extracted one or more features of the subject; determining an injection time of administration of an intravascular contrast agent injection to the subject; and initiating acquisition of an angiography image of the subject using the X-ray imaging device at a start time determined based on the injection time and the spin initiation time delay.

[0007] One advantage, in some embodiments, resides in determining a spin delay of a C- arm imaging device on a per-patient basis to sync image acquisition with contrast flow through the ROI being imaged.

[0008] Another advantage, in some embodiments, resides in accurately evaluating a delay in C-arm spin during 3DRA acquisition, taking into account the specific patient’s anatomy, such that when the image acquisition starts, contrast is entering the ROI being imaged.

[0009] Another advantage, in some embodiments, resides in analyzing procedure information and patient information to determine a spin delay of a C-arm imaging device.

[0010] Another advantage, in some embodiments, resides in using an artificial neural network (ANN) to analyze procedure information and patient information to predict a spin delay of a C-arm imaging device.

[0011] A given embodiment may provide none, one, two, more, or all of the foregoing advantages, and / or may provide other advantages as will become apparent to one of ordinary skill in the art upon reading and understanding the present disclosure.BRIEF DESCRIPTION OF THE DRAWINGS

[0012] The disclosure may take form in various components and arrangements of components, and in various steps and arrangements of steps. The drawings are only for purposes of illustrating the preferred embodiments and are not to be construed as limiting the disclosure.

[0013] FIGURE 1 illustrates an imaging device in accordance with an example embodiment of the present disclosure.

[0014] FIGURE 2 illustrates an imaging method using the imaging device of FIGURE 1 in accordance with an example embodiment of the present disclosure.DETAILED DESCRIPTION

[0015] There is a desire to accurately evaluate the delay in imaging system spin (e.g., C- arm spin) for image acquisition (e.g., 3DRA acquisition) during a procedure, such that when the image acquisition starts, contrast is entering the ROI being imaged. Embodiments of the disclosed system address the challenges presented by the manual timing of imaging system spin delay in order to sync image acquisition with contrast flow through the ROI being imaged. The imaging system delay is dependent on several factors. These factors may include factors that can be determined from one or more features in images acquired during the procedure, such as patient vasculature (e.g., tortuosity, thickness, etc.) and the placement or positioning of a delivery catheter delivering the contrast agent in the vasculature. Other factors may also include patient health information (e.g., patient height, blood pressure, etc.) and procedural information including access site used, current procedure phase (e.g., is the 3D acquisition closer to or further away from the access site where contrast is being injected from), whether the injection is atrial or venous, and contrast injection parameters like injection rate.

[0016] These factors are difficult to evaluate mentally by the physician. To simplify decision making, systems may provide a limited number of options for spin delay to choose from or may offer a set of discrete values to choose from. For instance, the Philips Azurion image- guided therapy (IGT) system (available from Koninklijke Philips N.V., Eindhoven, the Netherlands) provides spin delay options at 0.5 second intervals to choose from (0, 0.5, 1, 1.5... seconds). These discrete values may result in suboptimal delay. For example, if the ideal delay is 0.75 seconds, then the physician must select either 0.5 seconds or 1.0 seconds, leading to an error of 0.25 seconds. Suboptimal delay may result in treatment decisions based on lower quality images and can require reacquisition, expositing patients to additional radiation and contrast. This alsoassumes the physician at least selects the closest discrete spin delay option. If the physician fails to do so, then the error can be even larger. In the above example, if the physician selected 1.5 seconds, then the error is 0.75 seconds. It is to be appreciated that contrast inflow / washout triggered by a contrast agent bolus can be on the order of seconds; hence such errors in the spin delay can significantly degrade the acquired 3D angiographic image.

[0017] In embodiments disclosed herein, the spin delay is automatically evaluated based on an X-ray image of the actual patient, optionally along with other patient- specific information, such as patient height or blood pressure. This provides a spin delay that is accurately tailored to the specific patient. By automatically evaluating the patient’s X-ray image and optionally other patient-specific factors, embodiments of the disclosed system can provide more accurate C-arm spin delay estimates. This helps not only generate consistently better quality images, but also improve both patient and clinician experience by reducing patient exposure to additional radiation and contrast from re-acquisitions and by reducing cognitive load on clinicians.

[0018] With reference to FIGURE 1, an illustrative medical or imaging system including a medical device 1 is shown. The medical system includes a contrast injector 10 for delivering a contrast agent to the patient via a suitable tube 11 or the like, and an X-ray imaging device 1 for acquiring an X-ray image (e.g., 3D X-ray image) such as, for example, a C-arm imaging device (as shown), although other imaging devices capable of acquiring images (e.g., 3D images) are also contemplated, such as a computed tomography (CT) imaging device, or other X-ray imaging device that utilizes X-ray imaging (hereinafter referred to as an “X-ray imaging device” or variants thereof). The X-ray imaging device 1 may be part of an image guided therapy (IGT) system. As shown in FIGURE 1 , the illustrative X-ray imaging device l is a C-arm imaging device having a C-arm 12. The X-ray imaging device 1 also includes or is in operative communication with a device controller 14 configured to control operation of the X-ray imaging device 1. The device controller 14 is also in communication with the contrast injector 10 to determine an injection time of administration of an intravascular contrast agent injection by the injector 10 to the subject. In some embodiments, the device controller 14 may control the contrast injector 10 to administer the contrast agent injection (in which case the controller 14 determines the injection time as the time at which the controller triggers the injection). In other embodiments, the contrast injector 10 may administer the contrast agent injection independently. In these embodiments, the controller 14determines the injection time using a signal sent from the contrast injector 10 to the controller 14 when the injector 10 triggers the injection.

[0019] The X-ray imaging device 1 includes an X-ray detector 16 that is configured to detect the X-ray radiation emitted by an X-ray source 19 (e.g., an X-ray tube) after the X-rays pass through an examination region 17. In the illustrative C-arm design of FIGURE 1 , the X-ray source 19 (e.g., comprising an X-ray tube) and the X-ray detector 16 are mounted on opposite ends of the C-arm 12. In other designs, the configuration of the C-arm may be different. In another contemplated embodiment employing a CT scanner, the X-ray tube and X-ray detector are mounted on a rotating gantry. During operation to acquire imaging data of a patient or other imaging subject (not shown), that imaging subject is disposed in the examination region 17. As shown in FIGURE 1, the X-ray detector 16 typically comprises a detector array, and the X-ray tube 19 emits a cone beam of X-rays that pass through the examination region 17 and thence impinge on the detector array 16.

[0020] The detector 16 is also in communication (e.g., electronic communication) with the device controller 14, such as a workstation computer, or more generally a computer. Images produced by the X-ray imaging device 1 via the X-ray radiation generated by the X-ray tube 19 are processed by the device controller 14 and / or another processing device 18 (e.g., electronic processing device), which may, for example, be implemented as a server computer 18 as shown. The illustrative X-ray imaging device employs a C-arm configuration in which the X-ray tube 19 and the detector 16 can be moved by movement of the C-arm 12 to different vantage points (called “views”) around the patient to, for example, provide clinically significant views of the brain, heart, or in an interventional procedure to provide a chosen view of a region of interest (ROI) that is the target of the interventional procedure. As a nonlimiting example, the illustrative C-arm-based imaging device 1 could be the Philips Azurion IGT system, and, as shown, further includes a patient support 6 and a large display 8 for presenting images for guiding an operator in performing a neuro-interventional or cardiology procedure or other image-guided therapy. In some embodiments, the controller 14 and the display 8 may be integrated as a single unit.

[0021] The processing device 18 is an optional component that may include a workstation, a server computer (as shown) or a plurality of server computers, e.g., interconnected to form a server cluster, cloud computing resource, various combinations thereof, or so forth, to perform more complex computational tasks. In the illustrative embodiment of FIGURE 1, the devicecontroller 14 could optionally be integrated into an on-board computer of the display system 8 of the imaging device 1, and / or an additional computing component 9 (such as a touch-screen display for a user to interact with the display system 8 and / or device controller 14 and / or server 18), for example. In a common configuration, the device controller 14 is provided for controlling the imaging device 1 to perform image acquisition. The device controller 14 and / or server 18 may include typical components, such as a hardware processor 20 (e.g., a microprocessor), memory, at least one user input device (e.g., a mouse, a keyboard, a trackball, and / or the like) 22, and a display device 24 (e.g., an LCD display, plasma display, cathode ray tube display, and / or so forth).

[0022] The processor 20 is operatively connected with one or more non-transitory storage media 26. The non-transitory storage media 26 may, by way of non-limiting illustrative example, include one or more of a magnetic disk, RAID, or other magnetic storage medium; a solid-state drive, flash drive, electronically erasable read-only memory (EEROM) or other computer memory; an optical disk or other optical storage; various combinations thereof; or so forth; and may be for example a network storage, an internal hard drive of the workstation 18, various combinations thereof, or so forth. It is to be understood that any reference to a non-transitory medium or media 26 herein is to be broadly construed as encompassing a single medium or multiple media of the same or different types. Likewise, the processor 20 may be embodied as a single processor or as two or more processors. The non-transitory storage media 26 stores instructions executable by the at least one processor 20. The instructions include instructions to control the X-ray imaging device 1 to acquire and generate an angiographic image for display on the display device 24 (and / or elsewhere, for example on the imaging system display 8 and / or auxiliary display 9).

[0023] With reference to FIGURE 2, and with continuing reference to FIGURE 1, an illustrative embodiment of an instance of an imaging method 100 is diagrammatically shown as a flowchart. To begin the method 100, a patient or subject is loaded into the examination zone 17. At an operation 102, the controller 14 is configured to control the X-ray imaging device 1 to acquire an initial X-ray image of the patient in the examination zone 17. The X-rays emitted by the X-ray tube 19 pass through the examination region 17 and thence are detected by the X-ray detector array 16. The device controller 14 and / or the server 18 is configured to receive the X-ray image of the patient in the examination zone 17. This initial image acquired at the operation 102 may include at least one two-dimensional (2D) X-ray image of the subject. In some embodiments, this initial image is not a 3D image of the subject, and hence can be acquired quickly, e.g., withoutmoving the C-arm 12. However, it is also contemplated for the initial image acquired in the operation 102 to be acquired with rotation of the C-arm 12 so that the acquired image is a 3D image. The initial image acquired in the operation 102 may be acquired without contrast agent; or, in other embodiments, the initial image acquired in the operation 102 may be an angiographic image acquired with contrast agent provided by the contrast injector 10 via the tube 11.

[0024] At an optional operation 104, one or more items of additional information are received by the device controller 14 and / or the server 18. In one example, the additional information may include subject information of the subject being imaged. The subject information may include a height of the subject (e.g., taller patients may have a larger distance between access site and ROI), blood pressure of the subject, age of the subject (i.e., older patients may have thinner and more tortuous vasculature), and so forth.

[0025] In another example, the additional information may include information about the procedure being performed and / or steps of the procedure performed so far. The procedural information may include an access site used (e.g., ROIs in the subject head may require longer delay when using femoral access site as compared to when using radial access site), a current procedure phase (e.g., when using femoral access, a spin delay for a ROI in the torso may be shorter than for a subsequent ROI in the subject’s head), and so forth. The ROIs may be identified manually (e.g., by the user via the GUI 28 or mouse 22), or automatically via algorithms that can identify context from scenes by analyzing images or videos using, for example, feature extraction or feature identification techniques such as object detection, segmentation, etc.

[0026] In another example, the additional information may include contrast injection information about the administration of the intravascular contrast agent injection to the subject before the injection time. The contrast injection information may include a type of contrast injection (e.g., atrial or venous), an injection rate or speed, an amount of contrast, and so forth.

[0027] In some embodiments, at least one of the subject information, the procedural information, and the contrast injection information can be acquired from one or more sensors (not shown) disposed in an imaging bay where the X-ray image is acquired or anywhere in the procedure room. Such sensors may include, for example, RGB or RGB-D cameras that are able to capture and detect relevant information.

[0028] As reflected in FIGURE 2, the operations 102 and 104 may be performed in either order (i.e., the X-ray image may be acquired before the additional information is received, or vice versa), or they may be performed simultaneously.

[0029] At an operation 106, one or more features in the acquired X-ray image of the subject are extracted from the acquired X-ray image. The extracted features may include one or more of a location of a ROI being imaged in the X-ray image (e.g., in the patient torso or head, where in the torso or head); a tortuosity of a blood vessel shown in the X-ray image (e.g., more tortuous vessels may require longer spin delay); a thickness of a blood vessel shown in the X-ray image (e.g., thicker vessels may require shorter spin delay). In another example, the extracted features may include a position of a catheter (e.g., a tip of a catheter) in the vasculature of the subject shown in the X-ray image relative to a target location in the vasculature of the subject. For example, contrast flow is different at the delivery catheter delivering the contrast agent compared to in the vasculature, where heart rate and blood pressure have a larger effect on contrast flow. In another example, the one or more features include other features of the vasculature of the subject or of another anatomy of the subject.

[0030] At an operation 108, a spin initiation time delay is determined based on the extracted feature(s). In some embodiments, the determination of the spin initiation time delay is also based on the additional received information from operation 104. In some embodiments, the spin initiation time delay is determined as a continuous value. In other embodiments, the X-ray device 1 is configured with a plurality of discrete, selectable spin initiation time delays, and the determination of the spin initiation time delay includes automatically selecting one selectable spin initiation time delay from the plurality of selectable spin initiation time delays. In some embodiments, the spin initiation time delay determination operation 108 may determine relationships or correlations between the extracted feature(s) and / or additional received information to estimate the contrast flow through the vasculature of the subject and determine the spin initiation time delay from the estimated contrast flow.

[0031] In some embodiments, the feature extraction 106 and / or the spin initiation time delay determination operations 108 can be performed by a machine-learning model 34, such as an artificial neural network (ANN), implemented, for example, in the device controller 14 or the server 18. Inputs are provided to the model 34 to generate the spin initiation time delay. The inputs to the model 34 may include the X-ray image acquired in operation 102, from which themodel extracts the features of operation 106. The inputs to the model 34 can also include, for example, the additional information of operation 104, such as subject information and / or procedural information and / or contrast injection information. The spin initiation time delay may be determined based on one or more of these inputs.

[0032] The model (e.g., ANN) 34 can be trained with previously-acquired X-ray images of a plurality of historical subjects to predict spin initiation time delay. To do so, subject information of the historical subjects and / or procedural information of the historical procedures associated with the previously-acquired X-ray images and / or contrast injection information associated with the previously-acquired X-ray images and / or the previously-acquired X-ray images are received by the device controller 14 or the server 18 as input. Features of the historical subjects may be extracted by the model from the previously-acquired X-ray images (e.g., location of a ROI, tortuosity of a blood vessel, thickness of a blood vessel, position of the delivery catheter, etc.). The previously-acquired X-ray images may have been acquired across different times after acquisition of contrast agent. A ground truth amount of delay in imaging system spin applied to acquire a 3DRA of a prerequisite quality is also received as input. The amount of delay may be determined from retrospective data or from simulated data generated by simulators like Mentice, etc. The model 34 is trained using these inputs and outputs the predicted spin initiation time delay of an imaging system spin (e.g., C-arm spin) to be applied in order to acquire a 3DRA of a prerequisite quality. The model 34 may determine and analyze relationships or correlations between features or parameters in these inputs to predict the spin initiation time delay. In the model, the predicted and ground truth delay may be compared using any loss function including LI loss, L2 loss, etc. Values of loss functions may be used to update parameters (e.g., weights and biases) of neural networks using backpropagation. The value of the loss function is typically minimized, and training is terminated when the value of the loss function satisfies a stopping criterion. Sometimes, training is terminated when the value of the loss function satisfies one or more of multiple criteria. Training is often performed using a Graphics Processing Unit (GPU) or a dedicated neural processor such as a Neural Processing Unit (NPU) or a Tensor Processing Unit (TPU). The ANN 34 may be of any type including a convolutional neural network (CNN) or a transformer network. The ANN 34 may be a classification network that is trained to output one of n classes of delays (e.g., 5 classes that match 5 available options). Alternatively, the ANN 34 may be a regression network that is trained to regress a value of a delay. This value may be applieddirectly to the X-ray imaging system or may be used to choose one of the available options (e.g., the closest available option). For instance, if the predicted delay is 1.4 seconds, the 1.5 second delay option may be set on the Philips Azurion system.

[0033] In some embodiments, an identified ROI may be used for training and inference. For instance, the model 34 may be trained on input X-ray images and corresponding bounding boxes identifying ROIs. This trained model 34 will require both an X-ray image and a bounding box at inference. Alternatively, the identified bounding box may be used for training only. For instance, an encoder-decoder architecture may be used to learn a reduced dimensional representation of the inputted image by weighting the reconstruction loss such that a higher weight is applied near the identified ROI and lower weight away from the ROI. The reduced dimensional representation may then be used to predict the delay in imaging system spin. The representations learned by any of these neural networks describe features associated with various types of ROIs that require a particular amount of delay (e.g., intracranial aneurysms typically require a longer delay since they are far away from the access site). Representations may also learn features associated with vessel thickness or vessel tortuosity that require a particular amount of delay.

[0034] In some embodiments, in order for numerical information like PHI, procedural information, contrast injection information, etc. to not be overshadowed by the much larger X-ray image data, such data may be concatenated to a reduced dimensional representation or feature vector (e.g., ID vector or linear layer a few layers before the output layer). The numerical value of some information (e.g., patient height, blood pressure, injection rate, etc.) may be concatenated directly. Information such as access site, procedure phase, etc. may be assigned a label and the label may be concatenated. For instance, in the case of access sites, femoral access site may be assigned a label of 0, radial access a label of 1 , and so on.

[0035] During inference or application time, the trained model 34 receives the acquired X- ray image 102 and additional information 104, and outputs the predicted imaging system spin initiation delay determined by the trained model 34 in operation 108 that should be applied in order to acquire a 3DRA of a prerequisite quality.

[0036] In some embodiments, a confidence metric of the determined spin initiation time delay can be generated. Confidence may be related to errors generated during neural network training allowing the neural network to learn to associate lower confidence with features that tend to generate higher errors and higher confidence with features that generate lower errors.Confidence may similarly be related to a quality of the image acquired in operation 102 or may be related to an amount of additional information received in operation 104. Confidence may also be computed using dropout layers in the neural network that randomly drop the outputs of a specified number of nodes in the neural network, generating slightly different outputs for the same input at different iterations. The variance from the multiple outputs can inform confidence. For instance, a smaller variance may indicate high confidence (consistent outputs), while a larger variance may indicate low confidence (inconsistent outputs).

[0037] In some embodiments, the spin initiation time delay is determined based on a distance through the vasculature between the tip of the delivery catheter and the target location, with this distance being estimated from the image acquired in operation 102. In these embodiments, the distal end of the delivery catheter used for contrast injection is visible in the acquired image. The position of delivery catheter with respect to the ROI can affect contrast flow into the ROI, since contrast flow is different in the delivery catheter compared to in the vasculature, where heart rate and blood pressure have a larger effect on contrast flow. The delivery catheter may be identified in the same way as the ROI (i.e., manually or automatically) and may be used for both training and inference or for training only as described in the case of the ROI.

[0038] In some embodiments, the determination of the spin initiation time delay includes simulating traversal or flow of contrast agent through the vasculature of the subject in response to the administration of the intravascular contrast agent injection to the subject and simulation of the extracted features of the vasculature of the subject. A simulated traversal time of the contrast agent through the vasculature is determined based on the simulation. The spin initiation time delay is determined from the simulated traversal time.

[0039] In this embodiment, a simulator system such as the Mentice simulator (available from Mentice AB, Gothenburg, Sweden) may used, paired with an X-ray imaging system such as the Philips Azurion. The pairing enables the simulator to generate a simulated 3DRA based on the simulated contrast injected and the X-ray imaging device settings for 3DRA acquisition. Here, a reinforcement learning (RL) based method may be used to allow the system to set C-arm spin delays given an X-ray image and any additional information about the patient, access site, etc. such that the ANN 34 is given higher rewards for generating 3DRAs that match the optimal 3DRA that can be generated.

[0040] In some embodiments, a digital subtraction angiography (DSA) sequence is input into the model 34 instead of a single angiographic X-ray image. The DSA sequence can provide better information on contrast flow near the ROI including better information on vessel tortuosity. Such input may require a recurrent neural network (RNN) or temporal convolutional network (TCN) or transformer network architecture to accommodate temporal data.

[0041] In another embodiment, contrast enhanced images acquired throughout the imaging procedure are input into the network. This allows the model 34 to have a better estimate of patient vasculature throughout the navigation path, not just near the ROI.

[0042] In another embodiment, a previously acquired 3D image is available (e.g., preprocedural CT angiography) and can be used in addition to the input X-ray image from the operation 102 to the model 34. The previously acquired 3D image can better inform features such as vessel thickness and vessel tortuosity.

[0043] In another embodiment, if multiple imaging system views of the ROI are available (e.g., data acquired from a biplane system), this data may be used by the model 34 to compute a more precise estimate of the delay since this data would essentially provide more information about the ROI and surrounding vasculature and, hence, increase the confidence of the model 34 in its predictions (e.g., reduce ambiguities from vessel foreshortening). The multiple views could be inputted as separate input channels into the model 34 or as separate inputs into a Siamese network architecture, where parallel convolutional layers process the multiple views separately in the early layers of the model 34 and merge the weights in later layers to provide a combined output.

[0044] At an operation 110, an injection time of administration of an intravascular contrast agent injection to the subject is determined. How this is done depends on the configuration of the imaging system, and relies on the device controller 14 being in communication with the contrast injector 10 to determine the injection time. In some configurations, the device controller 14 controls the contrast injector 10 and thus triggers the contrast injector 10 to administer the contrast agent injection. In such configurations, the controller 14 determines the injection time as the time at which the controller 14 triggers the contrast injector 10 to administer the injection. In other configurations, the contrast injector 10 operates independently of the device controller 14, but is operatively connected with the device controller 14 to send a signal to the device controller 14 indicating when the contrast injector 10 administers the contrast agent injection. In these embodiments, the controller 14 determines the injection time using the signal sent from thecontrast injector 10 to the controller 14 when the injector 10 triggers the injection. It should be noted that the contrast agent injection may be a discrete bolus injection (that is, a rapid injection of contrast agent over a brief period), or may be a start of a continuous contrast agent delivery over a longer time interval, or could be a combination of an initial bolus of contrast followed by a lower volumetric rate of contrast injection.

[0045] At an operation 112, acquisition of an angiography image 36 of the subject is initiated using the X-ray imaging device 1 at a start time determined based on the injection time (from the operation 110) and the spin initiation time delay (from the operation 108), for example, determined based on the injection time plus the spin initiation time delay. The angiography image 36 may be a three-dimensional (3D) image that is generated by a suitable image reconstruction (e.g., filtered backproj ection reconstruction, fast Fourier transform reconstruction, or so forth) applied to a series of X-ray images acquired during the spinning of the C-arm 12 (and hence, the corresponding movement of the X-ray source 19 and detector 16) over an angular range. This may be considered as a rotational angiography image. The spinning of the C-arm 12 may be initiated at the start time and continues for a chosen time duration and at a chosen spin speed. The angiography image 36 may be be displayed on the large display 8 or the display device 24 of the device controller 14.

[0046] In addition to the spin initiation time delay, one or more of an amount of the intravascular contrast agent administered to the subject, a speed at which the intravascular contrast agent is administered to the subject, or a framerate to acquire the angiography image may be displayed on the large display 8 or the display device 24 of the device controller 14.

[0047] In some embodiments, the model 34 may additionally output information about the contrast injection. For instance, the model 34 may output the delay in C-arm spin along with the amount of contrast and / or contrast speed (or rate of contrast injection) such that the optimal image may be acquired. This reduces the effort required for the user to calculate the appropriate amount of contrast and / or contrast speed for optimal image acquisition. This also allows for improved image acquisition on X-ray imaging device 1 with a limited set of options for C-arm spin delay. These outputs can be displayed, for example, on the large display 8, the touch-screen 9, and / or the display device 24 of the device controller 14.

[0048] In one illustrative implementation of the method 100, an interface to the C-arm imaging system 1 receives the estimated delay from operation 108 and applies it to the 3DRAacquisition settings of the imaging system so that the operations 110 and 112 of the X-ray device 1 waits the prescribed amount after contrast injection begins in order to initiate the C-arm spin just as the contrast is entering the ROI. The delay may also be displayed on the large display 8 or the display device 24 of the device controller 14 so that the user can make any changes before confirming the 3DRA acquisition settings.

[0049] In an embodiment, the X-ray imaging device 1 may acquire X-ray images (at the operation 102) from an initial C-arm position before the 3DRA acquisition spin is initiated. The device controller 14 and / or the server 18 then determines, from the acquired X-ray images, when the intravascular contrast agent arrives at a portion of the subject shown in the acquired X-ray images. The operation 112, including the rotational spin of the X-ray imaging device 1, is initiated when the intravascular contrast agent arrives at a portion of the subject shown in the acquired X- ray images.

[0050] In this embodiment, the X-ray imaging device 1 acquires 2D X-ray images at varying frame rates after contrast has been injected or when imaging is initiated. That is, while the system waits to begin its rotational scan or 3DRA acquisition spin, it may acquire low dose X- ray images at varying frequencies to determine when the contrast has arrived at the ROI in order to begin its spin. The varying framerates may be additional outputs of the model 34. That is, in addition to estimating the delay in C-arm spin, the model 34 may also output the varying framerates for 2D X-ray image acquisition before the start of the C-arm spin. For instance, if the predicted delay is 1.5 seconds, the X-ray imaging device 1 may also output recommended framerate for 2D X-ray acquisition to be low (e.g., 0 fps) in the first 0.5 seconds, medium in the next 0.5 seconds (10 fps), and high in the final 0.5 seconds (e.g., 20 fps) so that X-ray acquisition (and, consequently, radiation exposure) is low when contrast is far away from the ROI, but increases when contrast is closer to the ROI so that 3D acquisition (e.g., at 30 fps) can begin as soon as contrast is seen in the field of view.

[0051] In some embodiments, the device controller 14 and / or the server 18 is configured to monitor contrast agent flow to a target location in the vasculature of the subject using imaging data generated during the acquisition of the 3D angiography image (i.e., during the rotational scan or C-arm spin). A rotational scan generates a series of 2D X-ray images from varying C-arm positions that are used to reconstruct a 3D image. The device controller 14 and / or the server 18 then determines, from the acquired 2D X-ray images, whether contrast was present in the imagesat the start of the rotational scan and / or whether contrast arrived at the ROI during the rotational scan. A stop time for the acquisition of the angiography image 36 is extended if the contrast agent flow to the target location in the vasculature of the subject is delayed respective to the start time.

[0052] In this embodiment, when the rotational scan or C-arm spin is started, if no contrast is observed in the field of view, the X-ray device 1 extends its spin beyond the initial planned range. In this embodiment, the device controller 14 and / or the server 18 is additionally able to recognize whether contrast is in the field of view. Once contrast is observed in the field of view at angle a, the spin is performed for s+a degrees, where ,y is the initial planned spin range, if extending the spin does not exceed the C-arm’ s spin range and / or does not cause collision with other objects or personnel. Images from this part of the spin are used for the 3DRA reconstruction.

[0053] The disclosure has been described with reference to the preferred embodiments. Modifications and alterations may occur to others upon reading and understanding the preceding detailed description. Combinations of the disclosed embodiments, and other embodiments not specifically described herein, will be apparent to those of skill in the art upon reviewing the disclosure. It is intended that the exemplary embodiment be construed as including all such modifications and alterations insofar as they come within the scope of the appended claims or the equivalents thereof.

Claims

CLAIMS:

1. An imaging system, comprising: a processor (18) communicatively coupled to memory, the processor (18) configured to: receive an X-ray image of a subject acquired using an X-ray imaging device; extract one or more features of the subject from the X-ray image; predict a spin initiation time delay for the X-ray imaging device based on the extracted one or more features of the subject; determine an injection time of administration of an intravascular contrast agent injection to the subject; and initiate acquisition of an angiography image (36) of the subject using the X- ray imaging device at a start time determined based on the injection time and the spin initiation time delay.

2. The imaging system of claim 1 , wherein the spin initiation time delay is predicted using a model (34) trained with previously-acquired X-ray images of a plurality of historical subjects that were acquired at different times after injection of contrast agent.

3. The imaging system of claim 2, wherein the processor (18) is further configured to: receive subject information of the subject being imaged; input the subject information to the model (34); and predict, by the model, the spin initiation time delay based on the inputted subject information.

4. The imaging system of either one of claims 2 and 3, wherein the processor (18) is further configured to: receive procedural information of the imaging procedure acquiring the X-ray image;input the procedural information to the model (34); and predict, by the model, the spin initiation time delay based on the inputted procedural information.

5. The imaging system of any one of claims 2-4, wherein the processor (18) is further configured to: before the injection time, receive contrast injection information of the administration of the intravascular contrast agent injection to the subject; input the contrast injection information to the model (34); and predict, by the model, the spin initiation time delay further based on the inputted contrast injection information.

6. The imaging system of any one of claims 1-5, wherein the one or more features of the subject in the X-ray image includes at least one of: a thickness of a blood vessel shown in the X-ray image; a tortuosity of a blood vessel shown in the X-ray image; and a location of a region of interest (ROI) being imaged in the X-ray image.

7. The imaging system of any one of claims 1-6, wherein the one or more features of the subject in the X-ray image includes at least: a position of a tip of a catheter in vasculature of the subject shown in the X-ray image relative to a target location in the vasculature of the subject, wherein the spin initiation time delay is predicted based on a distance through the vasculature between the tip of the catheter and the target location.

8. The imaging system of any one of claims 1-7, wherein the one or more features of the subject include features of vasculature of the subject and, to predict the spin initiation time delay, the processor is further configured to: based on the features of the vasculature of the subject, simulate traversal of contrast agent through the vasculature of the subject in response to the administration of the intravascular contrast agent injection to the subject;determine a simulated traversal time of the contrast agent through the vasculature based on the simulation; and predict the spin initiation time delay based on the simulated traversal time.

9. The X-ray imaging system of any one of claims 1-8, wherein the processor (18) is further configured to: generate a confidence metric of the predicted spin initiation time delay.

10. The imaging system of any one of claims 1-9, wherein the processor (18) is further configured to: initiate acquisition of angiography images (36) of the subject using the X-ray imaging device (1) at the start time determined based on the injection time and the spin initiation time delay; determine, from the angiography images, when the intravascular contrast agent arrives at a portion of the subject shown in the angiography images; and initiate a rotational spin of the X-ray imaging device when the intravascular contrast agent arrives at a portion of the subject shown in the angiography images.

11. The X-ray imaging system of any one of claims 1-10, wherein the processor (18) is further configured to: monitor contrast agent flow to a target location in vasculature of the subject using imaging data generated during the acquisition of the angiography image; and extend a stop time for the acquisition of the angiography image if the contrast agent flow to the target location in the vasculature of the subject is delayed respective to the start time.

12. The imaging system of any one of claims 1-11, wherein the processor (18) is further configured to: determine one or more of an amount of the intravascular contrast agent administered to the subject, a speed at which the intravascular contrast agent is administered to the subject, and a framerate to acquire the angiography image.

13. The imaging system of any one of claims 1-12, further comprising: an X-ray imaging device (1) including an X-ray source (19) and an X-ray detector (16) arranged to detect X-rays output by the X-ray source after the X-rays pass through an examination region.

14. The imaging system of claim 13, wherein the X-ray imaging device (1) is configured with a plurality of selectable spin initiation time delays and, to predict the spin initiation time delay, the processor is further configured to select one selectable spin initiation time delay from the plurality of selectable spin initiation time delays based on the extracted one or more features of the subject.15 An imaging method (100), comprising: receiving an X-ray image of a subject acquired using an X-ray imaging device; extracting one or more features of the subject from the X-ray image; predicting a spin initiation time delay for the X-ray imaging device based on the extracted one or more features of the subject; determining an injection time of administration of an intravascular contrast agent injection to the subject; and initiating acquisition of an angiography image (36) of the subject using the X-ray imaging device at a start time determined based on the injection time and the spin initiation time delay.

Images