Automated analysis of image data for determination of fractional flow reserve

Abstract

A system and method for analyzing a tube automatically detects a pathology in a first image of the tube and attaches a virtual marker to the pathology in the first image. The system can detect the same pathology in a second image of the tube based on the virtual marker and can then provide an analysis of the pathology (e.g., determining an FFR value) based on the pathology detected in the first and second images.

Description

Technical Field

[0001] The present invention relates to automated conduit analysis from image data, including the automatic determination of functional measures such as fractional flow reserve, and to interfacing the analysis to a user. Background Art

[0002] Arterial disease involves circulation problems in which narrowed arteries reduce blood flow to the body's organs. For example, coronary artery disease (CAD) is the most common heart disease and involves reduced blood flow to the heart muscle due to a buildup of plaque in the heart's arteries.

[0003] Current clinical practice relies on visual assessment of diseased vessels. For example, angiography is an X-ray-based technique used to examine arteries, veins, and organs to diagnose and treat stenosis (narrowing, usually due to atherosclerosis) and other vascular problems. During angiography, a catheter is inserted through an access point into the artery or vein, and a contrast agent is injected through the catheter to make the blood vessels visible on the X-ray image.

[0004] While providing an anatomical overview of the diseased duct, visual assessment does not provide the functional significance of the obstruction, ie, the effect of the obstruction on blood flow through the duct.

[0005] Fractional flow reserve (FFR) is a technique used to measure the pressure difference across a stenosis to determine the likelihood that the stenosis is impeding oxygen delivery to the myocardium. FFR is defined as the pressure behind (distal to) the stenosis relative to the pressure in front of the stenosis, thus representing the maximum flow down the duct in the presence of the stenosis compared to the maximum flow assuming the stenosis is absent. Some techniques use a three-dimensional model or reconstruction of the duct from which a function, such as FFR, is measured. In some cases, 2D images of the duct obtained from different angles are used to construct a 3D model of the duct.

[0006] Some systems interface with health professionals to display 3D models and the results of calculations based on the 3D models.

[0007] However, reconstructing a complete 3D model from the images input to the system is typically a slow process that requires extensive use of readily available memory.

[0008] Furthermore, existing systems require user input to calculate functional measurements. For example, a health professional may be asked to mark locations on a canal image, and then calculations are performed on the marked locations. This required user interaction consumes user resources and means that results cannot be provided in real time.

[0009] Thus, existing systems are unable to provide immediate live results and have no or limited flexibility to improve results based on new or different images fed into the system live. Summary of the Invention

[0010] Embodiments of the present invention provide a fully automated solution for pipeline analysis based on image data. According to embodiments of the present invention, the system detects pathologies from images of the pipeline without requiring user input regarding the location of the pipeline or the pathology. The system can then track the pathology across a series of images of the pipeline, allowing the identification of the same pathology in different images, potentially captured from different angles.

[0011] In one embodiment, a system for analyzing a vessel (e.g., a coronary vessel) is provided. The system can automatically detect a pathology in a first image of the vessel and attach a virtual marker to the pathology, i.e., to the location of the pathology in the first image. The system can then detect the pathology in a second image of the vessel based on the virtual marker and then provide an analysis of the pathology (e.g., determining an FFR value) based on the pathology detected in the first and second images. The analysis can be displayed on a user interface device. For example, the FFR value and / or an indication of the pathology can be displayed on the user interface device.

[0012] The first image and the second image may each be captured at a different angle.

[0013] In one embodiment, a processor of the system uses computer vision techniques to detect the tube in the first image and the second image and detect the pathology in the first image and the second image. The processor can then create a virtual marker in the first image to indicate the location of the pathology relative to the tube and can determine that the pathology detected at the same location relative to the tube in the second image is the same pathology as in the first image.

[0014] As detailed herein, the ability to identify the same pathology in different images enables improved automatic solutions and facilitates user (eg, health professional) interaction with the system.

[0015] The virtual marker can indicate the location of the pathology within the portion of the conduit and / or relative to the structure of the conduit. In some embodiments, the processor indexes the pathology based on the location of the pathology relative to the conduit. The system processor can cause an indication of the pathology to be displayed based on the location of the pathology within the conduit upon a user request for the pathology.

[0016] According to an embodiment of the present invention, the processor may detect multiple pathologies in the first image, and the processor may create a distinctive virtual marker for each of the multiple pathologies.The processor may cause indications of the multiple pathologies to be displayed on a single display.

[0017] Each of the plurality of pathologies may be assigned a name based on the location of each pathology within the conduit, and the processor may cause an indication to be displayed including the names of the plurality of pathologies.

[0018] In one embodiment, a flexible system for automated pipeline analysis based on pipeline images is provided. Methods and systems according to embodiments of the present invention provide analysis results (such as pipeline attributes, diagnoses, functional measurements, etc.) based on pipeline images. Embodiments of the present invention enable computation of the accuracy of the results obtained from pipeline images and can provide user interaction to improve the accuracy of the analysis.

[0019] Embodiments of the present invention extract 3D-related features from a 2D image of a pipeline and output an indication of properties and / or functional measurements of the pipeline based on the extracted 3D-related features, without using processing time and storage space to construct or use a 3D model of the pipeline. Thus, embodiments of the present invention provide immediate, substantially real-time results and can improve the results substantially in real time based on new or different images provided in the field. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] The present invention will now be described in conjunction with certain embodiments and examples with reference to the following illustrative drawings so that the invention may be more fully understood. In the drawings:

[0021] Figure 1 Schematically illustrates a system for analyzing a pipeline according to an embodiment of the present invention;

[0022] Figure 2 Schematically illustrates a method for analyzing a duct based on tracking pathologies between images according to one embodiment of the present invention;

[0023] Figure 3 schematically illustrates a method for analyzing a duct based on tracking pathologies between images according to another embodiment of the present invention;

[0024] Figure 4 Schematically illustrates a method for analyzing a duct based on tracking pathologies between images according to yet another embodiment of the present invention;

[0025] Figure 5 Schematically illustrates a method for analyzing a pipeline and interfacing with a user according to an embodiment of the present invention;

[0026] Figure 6 Schematically illustrates a method for analyzing a pipeline using temporal and structural data according to an embodiment of the present invention;

[0027] Figure 7 Schematically illustrates a method for determining the significance of a condition and interfacing with a user according to an embodiment of the present invention; and

[0028] Figure 8A and Figure 8B The figure schematically shows a user interface according to an embodiment of the present invention. Specific embodiments

[0029] Embodiments of the present invention provide methods and systems for automatically analyzing a pipe from an image of the pipe or a portion of the pipe and displaying the analysis results.

[0030] According to embodiments of the present invention, the analysis may include information regarding properties of the conduit, such as information related to the conduit's geometry. The analysis may also include functional measurements that may be computed from one or more properties of the conduit. The analysis may also include diagnostic information, such as the presence of a pathology, identification of the pathology, location of the pathology, etc. Analysis results may be displayed to a user, including the functional measurements, conduit properties, and / or computed, diagnostic, or other information based on the image of the conduit.

[0031] "Conduit" may include a tube or channel that contains and transports or circulates body fluids. Thus, the term conduit may include blood veins or arteries, coronary vessels, lymphatic vessels, portions of the gastrointestinal tract, and the like.

[0032] Images of the conduit may be obtained using suitable imaging techniques, such as X-ray imaging, ultrasound imaging, magnetic resonance imaging (MRI), and other suitable imaging techniques.

[0033] "Conduit attributes" may include, for example, anatomical characteristics of the conduit and / or pathology within the conduit (e.g., the shape and / or size of a portion of the anatomy). For example, a pathology may include conduit narrowing (e.g., a stenosis or narrowing), a lesion within the conduit, etc. Thus, conduit attributes may include, for example, the shape and / or size of the conduit and / or portion of the conduit, the bend angle of the conduit, the diameter of the conduit (e.g., proximal and distal to the stenosis), the minimum lumen diameter (e.g., at the location of the stenosis), the length of the lesion, the entrance angle of the stenosis, the entrance length, the exit angle of the stenosis, the exit length, the percentage of the diameter blocked by the stenosis, the percentage of the area blocked by the stenosis, etc. An indication of the pathology or pathology and / or other diagnosis may be calculated based on these attributes.

[0034] "Functional measures" are measurements of the effect of a condition on flow through the duct. Functional measures may include measurements such as FFR, instantaneous flow reserve (iFR), coronary flow reserve (CFR), quantitative flow ratio (QFR), resting full cycle ratio (RFR), quantitative coronary analysis (QCA), and others.

[0035] In the following description, various aspects of the present invention will be described. For the purpose of explanation, specific configurations and details are set forth in order to provide a thorough understanding of the present invention. However, it will also be apparent to those skilled in the art that the present invention can be practiced without the specific details presented herein. In addition, well-known features may be omitted or simplified in order not to obscure the present invention.

[0036] Unless expressly stated otherwise, as will be apparent from the following discussion, it will be understood that terms such as "using," "analyzing," "processing," "calculating," "computing," "determining," "detecting," "identifying," and the like used throughout the specification discussion refer to the actions and / or processes of a computer or computing system or similar electronic computing device that manipulate and / or transform data represented as physical (such as electronic) quantities within the computing system's registers and / or memories into other data similarly represented as physical quantities within the computing system's memories, registers, or other such information storage, transmission, or display devices. Unless otherwise stated, these terms refer to the automatic actions of a processor, independent of and without the need for any action by a human operator.

[0037] exist Figure 1 In one embodiment, schematically illustrated in FIG, a system for analyzing a pipeline includes a processor 102 in communication with a user interface device 106. The processor 102 receives one or more images 103 of a pipeline 113, each of which may capture the pipeline 113 from a different angle. The processor 102 then performs analysis on the received image(s) and transmits analysis results and / or instructions or other communications based on the analysis results to a user via the user interface device 106. In some embodiments, user input may be received at the processor 102 via the user interface device 106.

[0038] Conduit 113 may include one or more conduits or portions of conduits, such as veins or arteries, a branching system of arteries (an arterial tree), or other portions and configurations of conduits.

[0039] Processor 102 may include, for example, one or more processors and may be a central processing unit (CPU), a graphics processing unit (GPU), a digital signal processor (DSP), a field programmable gate array (FPGA), a microprocessor, a controller, a chip, a microchip, an integrated circuit (IC), or any other suitable general-purpose or specialized processor or controller. Processor 102 may be locally embedded or remote, for example, in the cloud.

[0040] The processor 102 is typically in communication with a memory unit 112. In one embodiment, the memory unit 112 stores executable instructions that, when executed by the processor 102, facilitate the performance of operations of the processor 102, as described below. The memory unit 112 may also store image data for at least a portion of the image 103 (which may include data such as indicative of the intensity of reflected light and pixel values ​​for part or all of the image or video).

[0041] Memory unit 112 may include, for example, random access memory (RAM), dynamic RAM (DRAM), flash memory, volatile memory, nonvolatile memory, cache memory, buffer, short-term memory unit, long-term memory unit, or other suitable memory unit or storage unit.

[0042] The user interface device 106 may include a display, such as a monitor or screen, for displaying images, instructions and / or notifications (for example, via the graphics, images, text or other content displayed on the monitor) to the user. The user interface device 106 may also be designed to receive input from the user. For example, the user interface device 106 may include a mechanism for inputting data, such as a keyboard and / or mouse and / or touch screen, or may communicate with it so that the user can input data.

[0043] All or some components of the system may communicate wired or wirelessly and may include appropriate ports, such as USB connectors and / or network hubs.

[0044] In one embodiment, processor 102 may determine properties of the conduit from an image of the conduit, typically by applying computer vision techniques, such as by applying shape and / or color detection algorithms, object detection algorithms, and / or other suitable image analysis algorithms, to at least a portion of one or more of images 103. In some embodiments, a machine learning model may be used to detect portions of the conduit and determine properties of the conduit from image 103. In some embodiments, pathology and / or functional measurements of the conduit (e.g., at the location of the pathology) may be determined based on the determined conduit properties.

[0045] In some embodiments, pathology and / or functional measurements can be determined directly from one (or more) images of the conduit. For example, pathology and / or functional measurements can be determined based on a single 2D image of the conduit without having to determine properties of the conduit.

[0046] In some embodiments, properties and / or functional measurements of the conduit may be determined by using a combination of structural and temporal data obtained from images of the conduit, for example, as described below with reference to Figure 6 described.

[0047] Typically, each image 103 captures pipe 113 from a particular angle or viewpoint.

[0048] In one embodiment, determining properties of the conduit may include receiving a 2D image of the conduit and extracting 3D-related features from the image without constructing a 3D model of the conduit, e.g., without using voxels and / or point clouds or other 3D representations.

[0049] 3D-related features are image features that can be specific structures in the image (such as points, edges, or objects), or any other information in the image that can be used to determine the properties of the pipeline from the image. In some embodiments, 3D-related features are extracted from images obtained from different views. The features from these images can be used to teach a machine learning model to detect the properties of the pipeline from 2D images. For example, features extracted from 2D images obtained from different views can be combined using a neural network (e.g., a long short-term memory (LSTM) neural network) that can compute features for each imaging element, integrate the features, maintain a representation of the features in memory (hidden state), and update its output as more images are input. Such a neural network can be used to learn the implementation of the pipeline, which can then be used to determine the properties and / or functional measurements of the pipeline from 2D images without having to reconstruct a full 3D representation or using a 3D model of the pipeline or using voxels and / or point clouds or other 3D representations.

[0050] Extracting 3D-related features from 2D images and determining pipe properties and / or functional measurements from the 3D-related features without constructing a 3D representation and / or without using a 3D representation of the pipe provides a fast process with minimal accuracy cost.

[0051] In one embodiment, an indication of the characteristics and / or functional measurements of the conduit determined from one or more images of the conduit (eg, based on the extracted 3D-related features) may be displayed via the user interface device 106 .

[0052] AutoCathFFR by Medhub TM For example, a system for automatically calculating the FRR of a tube from an image of the tube. For example, Medhub's AutoCathIFR TM This is a system for automatically calculating iFR programs from pipeline images.

[0053] In one embodiment, the system described above includes a processor, such as processor 102, which is implemented in Figure 2 The method is schematically shown in FIG.

[0054] The system analyzes a sequence of images of the conduit, such as a video cine of angiographic images. Processor 102 selects a first image from the sequence of images using computer vision techniques (step 202) and detects a pathology, such as a stenosis or lesion, in the first image (step 204). Detection of the pathology and / or the location of the pathology is accomplished using computer vision techniques without requiring user input regarding the location of the pathology. The processor may automatically detect the pathology in a second image of the conduit (step 206), the second image being captured at a different angle than the first image, and the processor may then cause the first image and / or the second image of the conduit to be displayed on a user interface device, such as user interface device 106, with an indication of the pathology (step 208).

[0055] The condition indication displayed on the user interface device may include, for example, graphics such as letters, numbers, symbols, different colors and shapes, etc., which may be superimposed on the image.

[0056] Once a pathology is detected in a first image, the pathology may be tracked throughout a sequence of images (eg, a video) such that the same pathology may be detected in each image even if its shape or other visual characteristics change between images.

[0057] exist Figure 3 One method of tracking a pathology between images is schematically illustrated in FIG. As described above, a first image from a sequence of images is selected (step 302) and a pathology is detected in the first image (step 304). A processor attaches a virtual marker to the pathology (step 306).

[0058] In some embodiments, the virtual marker is location-based, for example, based on the location of the pathology within a portion of the conduit automatically detected by the processor 102. In some embodiments, the virtual marker includes the location of the pathology relative to the structure of the conduit. The structure of the conduit may include any visible indication of the anatomy of the conduit, such as the junctions of the conduit and / or a particular conduit that are typically present in a patient. The processor 102 may detect the conduit structure in the image using computer vision techniques and may then index the detected pathology based on the location of the pathology relative to the detected conduit structure.

[0059] For example, a segmentation algorithm can be used to determine which pixels in the image are part of a pathology, and the location of the pathology relative to the structure of the conduit can be recorded, for example, in a lookup table or other type of virtual index. For example, in a first image, a stenosis is detected at a particular location (e.g., in the distal left anterior descending artery (LAD)). The stenosis located at the same particular location (distal LAD) in the second image is determined to be the same stenosis detected in the first image. For example, if more than one stenosis is detected within the distal LAD, each stenosis is marked with its relative location relative to additional structures of the conduit (such as, relative to a junction of the conduit) so that the stenosis in the second image can be distinguished.

[0060] Thus, the processor 102 creates a virtual marker specific to each pathology and, in the event of multiple pathologies in a single image, distinguishes the multiple pathologies from one another.

[0061] The pathology may then be detected in the second image of the conduit based on the virtual marker (step 308). The processor 102 may then cause an indication of the pathology to be displayed based on the virtual marker (e.g., as described above). In some embodiments, the processor may assign a name to the pathology based on the location of the pathology within the conduit, and the indication of the pathology may include the name assigned to the pathology, as further described below.

[0062] In some cases, a tube or group of tubes may include more than one stenosis or other pathology, making it more difficult to detect the same pathology in different images. In some embodiments, the processor detects multiple pathologies in the first image and creates a distinct virtual marker for each of the multiple pathologies. The processor can then cause an indication of each pathology to be displayed based on the virtual marker. In some embodiments, the indications are displayed together on a single display.

[0063] Therefore, a processor according to an embodiment of the present invention can determine a functional measurement result of the pathology (e.g., an FFR value) based on the first image and the second image (e.g., based on the pathology detected in the first image and the second image), and can display an indication of the functional measurement result, for example on a user interface device.

[0064] In some embodiments, the processor may determine the accuracy level of the functional measurement result and may calculate the third image required to improve the accuracy level. The processor may then cause an indication of the third image to be displayed on the user interface to suggest to the user which image to add to improve the accuracy of the result.

[0065] The first image, the second image, and the third image are each typically captured at a different angle, and the indication displayed on the user interface device includes the angle of the third image.

[0066] In one embodiment, the best frame selected from the pipeline's sequence of images is used as the first image discussed above.

[0067] exist Figure 4 In the example schematically shown in FIG, a video of angiographic images is received (step 402) and the best image is detected from the video (step 404). A pathology is detected in the best image (step 406). The pathology can then be tracked in the sequence of images and thus detected in another frame (step 408), thereby enabling an indication of the pathology to be displayed in all images (step 410).

[0068] The best image is typically the one that shows the most detail. In the case of angiographic images, which involve injecting a contrast agent into a patient to make vessels (e.g., blood vessels) visible on an X-ray image, the best image may be the image of the vessel that shows the most / largest amount of contrast agent. Thus, the best image can be detected by applying an image analysis algorithm to a sequence of images.

[0069] In one embodiment, the image captured at the time corresponding to maximum diastole is the image showing the greatest amount of contrast agent. Thus, the optimal image can be detected based on comparison of the image capture time with a measurement of electrical activity, such as the patient's heartbeat (e.g., an ECG printout).

[0070] In one embodiment, the processor may calculate values ​​for the functional measure, such as the FFR value, for each pathology and may cause those value(s) to be displayed.

[0071] In some embodiments, the processor 102 calculates an accuracy level of the functional measurement value (eg, FFR value) based on the angle of capture of the first image and may cause an indication of the accuracy level to be displayed on the user interface device 106 .

[0072] exist Figure 5In one embodiment schematically shown in FIG, processor 102 receives an image of a conduit (e.g., image 103) (step 502) and provides analysis from the image (e.g., determining properties and / or functional measurements of the conduit) (step 504). For example, processor 102 may apply a computer vision algorithm (e.g., as described above) to the received image(s) 103 to determine one or more properties, such as the shape and / or size of the conduit and / or portion of the conduit, the bend angle of the conduit, the diameter of the conduit, the minimum lumen diameter, the pathological length, the entrance angle of the stenosis, the entrance length, the exit angle of the stenosis, the exit length, the percentage of the diameter obstructed by the stenosis, the percentage of the area obstructed by the stenosis, etc. Processor 102 may then determine the functional measurement based on the properties of the conduit. In other embodiments, processor 102 determines the functional measurement directly from image 103, for example, by running a regression algorithm using a machine learning model to predict the value of the functional measurement (e.g., FFR) from the image of the conduit.

[0073] In some embodiments, processor 102 computes a level of accuracy (also referred to as a "margin of error") of the analysis based on image(s) 103 (step 506) and may cause an indication of the level of accuracy to be displayed on user interface device 106 (step 508).

[0074] For example, the accuracy level can be calculated by obtaining functional measurements of the pipe using known methods (e.g., physical measurements) and comparing the obtained functional measurements with functional measurements obtained from images of the pipe according to embodiments of the present invention. Deviations from the measurements obtained using known methods can be used to determine the accuracy level of the determination based on embodiments of the present invention. This can be done for images obtained at all possible angles, thereby creating a mapping or regression analysis that connects images at different angles and / or image combinations to different accuracy levels. This analysis can be performed by performing empirical experiments or by using, for example, a predictive model to create a mapping function from the angle of the image to the accuracy level.

[0075] Thus, a processor according to an embodiment of the present invention may receive an image of a pipe that captures the pipe at an angle and may perform an analysis (eg, determine properties and / or functional measurements of the pipe) with a level of accuracy based on the angle.

[0076] Because any image acquired at any possible angle can be mapped to a level of accuracy, depending on the level of accuracy required, according to embodiments of the present invention, functional measurements and other analyses can be obtained based on a single 2D image.

[0077] Because the processor 102 can detect particular pathologies in different images of the conduit (e.g., images captured from different angles) and can determine a level of accuracy for each pathology based on the different images, the processor 102 can calculate which (if any) additional images (captured at which angles) are needed to adjust (e.g., improve) the accuracy of the analysis.

[0078] In one embodiment, the indication of the level of accuracy displayed on the user interface device in step 508 includes instructions or notification to the user (e.g., a health professional) regarding how many additional images to add, typically specifying the angle of each additional image, to improve the accuracy of the analysis results and lower the error margin.

[0079] In one embodiment, processor 102 may provide an indication of a single image angle that, when added to the image already supplied by the user, may provide the greatest improved level of accuracy.

[0080] exist Figure 6 In one embodiment schematically illustrated in FIG. , a sequence of images, such as a video 603 of angiographic images, is analyzed, for example, to determine properties of an imaged vessel and / or to compute functional measurements of the vessel. A processor obtains structural data 604 of the vessel from at least one image in video 603. The processor also obtains temporal data 605 of the vessel from the images in video 603. Structural data 604 and temporal data 605 are combined, and the combined information is analyzed, for example, by an encoder 610, to obtain relevant features from which properties of the vessel are determined and / or functional measurements of the vessel are computed.

[0081] In one embodiment, the processor determines a pathology from the image of the conduit and may cause an indication of the pathology to be displayed on the user interface device 606 .

[0082] In some embodiments, a functional measure of the conduit may be calculated based on properties of the conduit or based on relevant features obtained by encoder 610. An indication of the functional measure may then be displayed on user interface device 606.

[0083] The correlation features computed by encoder 610 may also be used to determine properties of the pipe, such as the shape or size of a portion of the pipe.

[0084] In all cases, an indication of the level of accuracy of the displayed analysis (pathology, functional measurements, properties of the conduit, etc.) may be calculated and displayed on the user interface device 606 .

[0085] As described above, angiographic images include contrast media injected into the patient to make vessels (e.g., blood vessels) visible on the X-ray image. Thus, in one embodiment, the image selected from the angiographic video (from which the structural data 604 is obtained) can be the best image, e.g., an image showing a blood vessel with the largest amount of contrast media.

[0086] The time data 605 can be obtained from a flow map that estimates the flow rate of blood (visible due to the contrast agent) at points within the conduit. Calculating flow and generating the flow map can be accomplished by applying a motion detection algorithm to the video 603 and / or using a neural network trained to estimate motion and output an optical flow map.

[0087] The structural data 604 may be obtained by using computer vision techniques, such as by applying a segmentation algorithm to at least one image from the video (e.g., the image showing the largest amount of contrast agent) to detect ducts and / or pathologies and / or geometry-related information or other information in the image.

[0088] In some embodiments, portions of the conduit may be detected and the location of the pathology in the conduit may be determined based on the associated features computed by encoder 610. The location of the pathology and / or other indications may then be displayed on user interface device 606.

[0089] In one embodiment, an example is Figure 7 As schematically shown in FIG, a processor (e.g., processor 102) receives an image of a conduit (step 702) and determines a pathology (e.g., a lesion or stenosis) from the image of the conduit (step 704). For example, the pathology can be determined from properties of the conduit and / or from relevant features extracted from the image of the conduit, e.g., as described above. The processor can then calculate a significance level for the pathology (step 706) and cause an indication of the pathology and / or a functional measure of the pathology to be displayed based on the significance level. For example, the significance level of the pathology can be determined based on parameters of the pathology, such as the size and / or shape of the pathology and / or the percentage of a diameter blocked by the pathology, the percentage of an area blocked by the pathology, etc.

[0090] In one embodiment, if the significance level is above a threshold (e.g., a predetermined threshold) (step 707), the condition and / or functional measurements associated with the condition are displayed to the user (708). However, if the significance level is below the predetermined threshold (step 707), the condition and / or functional measurements may not be displayed to the user (step 710). In some embodiments, the conditions may be scored based on their significance and may be displayed to the user along with their scores, e.g., each condition may be displayed in a table listing its significance, as described below.

[0091] In another embodiment, a significance level may be calculated by comparing multiple conditions to each other and / or to a predetermined standard.

[0092] Therefore, a system for analyzing a pipeline includes a processor in communication with a user interface device. The processor determines a pathology of the pipeline from an image of the pipeline, calculates a significance level for the pathology, and controls the device based on the calculated significance level. For example, the processor may control the user interface device to control its display based on the calculated significance level.

[0093] In some embodiments, a processor (such as processor 102) may classify a condition based on one or both of the location of the condition within the conduit and a functional measurement result (e.g., FFR value). The processor may accept a user request for a condition based on the location within the conduit and / or based on the FFR value and may display the condition according to the classification.

[0094] An example of a user interface according to an embodiment of the present invention is Figure 8A and Figure 8B It is shown schematically in FIG.

[0095] exist Figure 8A In one embodiment schematically shown in FIG, an image of a coronary vessel (eg, LAD 803 ) captured at a specific angle is displayed on a monitor 816 .

[0096] In one embodiment, the functional measurement value FFR 801 is displayed on a monitor 816 of a user interface device along with an indication 804 of the number of images used to calculate the functional measurement FFR 801. In one embodiment, a single image may be used from which the functional measurement is obtained. In some embodiments, even if multiple images are used to calculate the functional measurement, only one image of the LAD 803 including the best or most visible features is displayed on the monitor 816.

[0097] An indication of one or more pathologies (e.g., stenosis 807) can be displayed as a graphic superimposed on the displayed image. In some embodiments, the displayed image is a representation of the vessel. For example, the displayed representation can include a combination of images of the vessel, such as a combined image of several images (typically acquired at different angles) or an average of several images (acquired at different angles). The angle(s) 805 at which the displayed image(s) were acquired can be indicated on the display 816.

[0098] In some embodiments, as Figure 8B As shown, the image of the pipeline displayed on the monitor 816 is a graphical illustration 813 of the pipeline, rather than an actual image.

[0099] Graphics, which may include, for example, letters, numbers, symbols, different colors, and shapes, may be displayed superimposed on the representation of the tube. For example, references 811 may be made to different portions of the LAD and one or more pathologies. References 811 may be used to help the user locate the pathology, as shown in table 812. For example, the first row of table 812 is associated with tube number 7 in the middle LAD, both of which are shown as references 811 on the graphical illustration 813 of the tube.

[0100] References 811 may be assigned by the processor to different pipeline sections detected by the processor based on computer vision techniques.

[0101] In some embodiments, the monitor 816 includes a button 808 that enables a user to at least partially hide graphics superimposed on a representation (e.g., an image or graphical illustration) of a pipeline so that the user can view the pipeline unobstructed by the different graphics. For example, activating the button 808 can cause all or specified graphics to fade or become transparent.

[0102] The level of accuracy, the error limit 802, of the value of FFR 801 is also displayed on monitor 816. As described above, the error limit 802 can be known for each image obtained at a known angle. The level of accuracy can be similarly known or calculated for a combination of images obtained at different angles. Therefore, adding images obtained at different angles can change the level of accuracy of the currently displayed functional measurement result. For example, a user (e.g., a health professional) can add an image obtained at an angle different from angle 805 in order to change the error limit 802. In some embodiments, monitor 816 includes a window 815 for displaying to the user an indication of the angles of additional images that should be input into the system to improve the level of accuracy or minimize the error limit of the analysis result provided by the system.

[0103] In one embodiment, the processor 102 may classify the pathology based on one or both of the location of the pathology within the conduit and its FFR value. In one embodiment, the processor may accept a user request to display the pathology based on the location within the conduit and / or based on the FFR value.

[0104] Because embodiments of the present invention are capable of automatically detecting pathologies in images of a pipeline and are capable of marking the location of the pathologies relative to anatomical structures, pathologies can be retroactively identified and marked even in images of the pipeline captured before the pathology is identified. Therefore, according to embodiments of the present invention, the processor can detect the same pathology as in the first image in an image captured before the pathology in the first image is detected based on a virtual marker attached to the pathology in the first image. This enables the user to work offline as well as at the time of image capture. Offline work can include retroactively marking pathologies in images and classifying images based on desired parameters and displaying results based on the classification. In addition, offline work can include collecting analysis, as described below. For example, a user can request to view all stenosis detected in the middle LAD. The processor can then control the user interface device to display the stenosis according to the requested classification, for example, only displaying an image or representation of the middle LAD pipeline and its corresponding information (e.g., the first row in Table 812).

[0105] In another example, the user may request to view stenoses having FFR values ​​above a threshold, in which case the processor 102 may cause only stenoses having relevant FFR values ​​to be displayed or indicated on the user interface device.

[0106] Embodiments of the present invention may be used with images obtained using any suitable imaging method, for example, images obtained using quantitative angiography methods (such as quantitative superficial femoral angiography), ultrasound methods (such as intravascular ultrasound (IVUS)), tomography (such as optical coherence tomography (OCT)), etc.

[0107] Embodiments of the present invention provide systems and methods for obtaining functional measurements, such as FFR, whose accuracy can be improved in real time and can be customized to specific user requirements.

[0108] In some embodiments, medical data (such as life expectancy and lifespan) and / or other data (such as age, gender, medical history, etc.) can be input into the system and used with images of the pipes and the pathologies in the pipes to create big data. For example, embodiments of the present invention enable providing a user with an analysis involving functional measurements (such as FFR) collected from many subjects (e.g., angiograms of all patients examined at a particular facility or network of facilities). Such an analysis may include, for example, FFR for each gender and each age, as well as FFR for each anatomical region and / or each artery. A user interface device according to an embodiment of the present invention may provide a button for a user request for such an analysis and / or a window displaying a numerical and / or graphical representation of such an analysis.

[0109] For example, big data can be used to predict the risk level of disease conditions and the best possible treatment practices for each disease condition over the long term. TM It is a system for predicting risk conditions and the best treatment for the conditions based on big data analysis.

Claims

1. A system for analyzing a pipeline, the system comprising a processor in communication with a user interface device, the processor being configured to: (i) detecting a pathology in a first image of the conduit; (ii) creating a virtual marker specific to the pathology in the first image, wherein the virtual marker indicates a location of the pathology relative to a structure of the conduit; (iii) detecting the pathology in a second image of the conduit based on the virtual marker; (iv) determining an FFR value for the condition based on the condition detected in the first image and the second image; and (v) causing an indication of the FFR value to be displayed on the user interface device.

2. The system according to claim 1, wherein: The conduits include coronary vessels.

3. The system according to claim 1, wherein: The processor is configured to cause an indication of the condition to be displayed on the user interface device.

4. The system according to claim 1, wherein: The processor is configured to use computer vision techniques, detecting the pipe in the first image and the second image; detecting a pathology in the first image and the second image; creating the virtual marker in the first image to indicate a location of the pathology relative to a structure of the conduit; and A determination is made that a pathology detected in the second image at the same location relative to the structure of the vessel is the same pathology as in the first image.

5. The system according to claim 4, wherein: The virtual marker indicates the location of the pathology relative to the structure of the conduit.

6. The system according to claim 4, wherein: The processor is configured to index the pathology based on a location of the pathology relative to a structure of the conduit; and The processor is configured to control a display of the user interface device according to the index.

7. The system according to claim 1, wherein: The processor is configured to: causing a representation of the pipeline to be displayed on the user interface device; and An indication of the condition is caused to be displayed as a graphic superimposed on the representation.

8. The system according to claim 1, wherein: The processor is configured to: detecting a plurality of pathologies in the first image; and A distinctive virtual marker is created for each of the plurality of pathologies.

9. The system according to claim 8, wherein: The processor is configured to cause indications of the plurality of pathologies to be displayed on a single display.

10. The system according to claim 8, wherein: The processor is configured to assign a name to each of the plurality of pathologies based on a location of each pathology within the conduit.

11. The system according to claim 10, wherein: The processor is configured to cause an indication including the name of the pipe to be displayed.

12. The system according to claim 1, wherein: The first image and the second image are each captured at a different angle.

13. The system of claim 1, wherein: The processor is configured to calculate a significance level for the condition and cause the indication of FFR to be displayed based on the significance level.

14. The system according to claim 1, wherein: The processor is configured to retroactively execute steps (i) to (iv) in an offline state.

15. The system of claim 1, wherein: The user interface device is configured to receive a user request for an analysis related to FFR and to display the analysis.

Images