Simultaneous display of hemodynamic parameters and damaged brain tissue

Abstract

An image can be generated having an overlay indicating the amount of brain tissue damage based on the disruption of blood supply. The imaging data can be analyzed to identify a perfusion parameter for a region of the individual's brain. A perfusion parameter threshold for the presence of damaged brain tissue can be based on a time period elapsed since onset of a physiological condition that disrupts blood flow to one or more regions of the individual's brain. The imaging data can also be analyzed to identify a hypodensity measure for a region of the individual's brain. A likelihood that the hypodensity measure corresponds to brain tissue damage can also be determined based on a time period elapsed since onset of the physiological condition.

Description

[Technical field]

[0001] For the simultaneous display of hemodynamic parameters and damaged brain tissue. [Background technology]

[0002] Damage can occur in the human body when blood flow to tissue is obstructed. Reduction in blood flow to tissue can have several causes, such as vascular occlusion, vascular rupture, vascular narrowing, or vascular compression. The severity of damage to tissue can depend on the degree of obstruction of blood flow to the tissue and the amount of time blood flow to the tissue is obstructed. In situations where blood supply to tissue is inadequate for an extended period of time, the tissue can become infarcted. Infarcted tissue can result from tissue cells dying due to lack of blood supply. Summary of the Invention [Problem to be solved by the invention]

[0003] For the simultaneous display of hemodynamic parameters and damaged brain tissue. [Means for solving the problem]

[0004] In the drawings, which are not necessarily drawn to scale, like numbers may describe like components in various figures. To easily identify any particular element or discussion of an operation, the first digit or digits in a reference number refer to the figure number in which that element is first introduced. Some embodiments are provided by way of example and not by way of limitation. [Brief description of the drawings]

[0005] [Figure 1] 1 is a diagrammatic representation of an example architecture for aggregating data from different imaging modalities to generate an image that includes an overlay indicating potential damage to brain tissue, according to one or more example embodiments. [Diagram 2] 1 is a diagrammatic representation of an example of an architecture for determining perfusion parameters related to blood flow through an individual's brain, according to one or more example embodiments. [Diagram 3]1 is a diagrammatic representation of an example architecture for identifying low concentration measures for brain tissue, according to one or more example embodiments. [Figure 4] 1 is a diagrammatic representation of an example architecture for aggregating information generated from different imaging modalities to identify the amount of damage to brain tissue, according to one or more example embodiments. [Diagram 5] FIG. 1 is a diagrammatic representation of an example architecture for generating image data indicative of the amount of damage to brain tissue based on the timing of onset of biological conditions corresponding to one or more blood vessels in an individual's brain, according to one or more example embodiments. [Figure 6] 1 is a flowchart illustrating example operations of a process for identifying an aggregate image based on image data generated by different contrast modalities and generating an overlay of the aggregate image indicative of the amount of potential damage to brain tissue, according to one or more example embodiments. [Figure 7] 1 is a flowchart illustrating example operations of a process for identifying perfusion parameters and hypoconcentration measures related to brain tissue and identifying an amount of potential damage to at least a portion of the brain tissue, according to one or more example embodiments. [Figure 8] 1 is a flowchart illustrating example operations of a process for generating an output image indicative of damaged brain tissue based on the onset of a physiological condition corresponding to one or more blood vessels in an individual's brain, according to one or more example embodiments. [Figure 9] FIG. 1 illustrates an example user interface including several slices of perfusion-based CT imaging data with overlays showing regions of interest in an individual's brain identified according to different perfusion parameters and low-density analysis, in accordance with one or more example embodiments. [Figure 10] FIG. 1 is a block diagram illustrating machine components in the form of a computer system that can read and execute instructions from one or more machine-readable media to perform any one or more methodologies described herein, in accordance with one or more example embodiments. [Figure 11]FIG. 2 is a block diagram illustrating a representative software architecture that may be used in conjunction with one or more hardware architectures described herein, according to one or more example embodiments. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0006] Various imaging techniques can be performed to identify the extent of damage and / or risk to tissue due to reduced blood flow to the tissue. In one or more examples, a perfusion-based imaging technique can be performed to identify areas of the individual's brain that have experienced reduced blood supply. Another embodiment is to perform the assessment through cerebral angiography. For example, a computed tomography (CT) imaging technique can be used to capture images of the individual's brain. The images generated by the CT imaging technique can be analyzed to determine that damaged areas of the individual's brain are at high risk of irreversible damage due to lack of blood supply to those areas. In various examples, a contrast agent can be delivered to the individual and a CT image can be captured that shows the flow of the contrast agent through the blood vessels that supply blood to the individual's brain. If the image is taken while the contrast agent is still in the arteries, this is referred to as a CT angiogram (CTA). A CTA imaging technique can include capturing images of the blood vessels that carry blood to the brain and can provide an indication of narrowing or blockage of the blood vessels. Severe stenosis or blockage of blood vessels that carry blood to the brain can cause damage to tissues where blood flow is impaired or blocked. Thus, in at least some cases, the amount of stenosis or blockage of one or more blood vessels can raise suspicion of areas of the brain that may sustain or be at risk of sustaining tissue damage in the future. When images are taken dynamically during the time that contrast agent is passing through large arteries, capillary beds, and draining veins for the purpose of deriving hemodynamic parameters such as blood flow or blood arrival time in tissue, the technique is called CT perfusion (CTP). Blockage of the flow of contrast agent to an area of ​​the brain can indicate a lack of blood supply to the area of ​​the brain that can cause damage to the brain tissue in that area. In one or more examples, the contrast agent can include iodine or gadolinium disposed in a carrier solution. Perfusion-based magnetic resonance (MR) imaging techniques that implement contrast agents can also be used to identify areas of an individual's brain that have a blocked blood supply.

[0007] Additionally, non-contrast-based imaging techniques can be performed to identify regions of an individual's brain that may be damaged due to lack of blood supply. For illustrative purposes, diffusion-based MR imaging techniques can be used to identify portions of an individual's brain that have been impeded from blood supply and thus damaged. In at least some examples, damage to brain tissue may be irreversible. In one or more illustrative examples, diffusion-based MR imaging techniques can be performed to identify brain tissue that has been damaged due to lack of blood supply. Additionally, non-contrast CT imaging techniques can be used to identify regions of brain tissue that have been damaged due to insufficient blood supply to the region. In the hyperacute phase of an infarction, subtle intensity and morphological changes in non-contrast CT head images can reveal areas of brain tissue that can no longer be salvaged. Although non-contrast CT is typically a separate CT acquisition, information can also be derived from a dynamic CT perfusion scan phase taken before the contrast agent reaches the brain. The scan phase captured before the contrast agent enters the brain can be referred to as the baseline time point. The baseline time point can also be derived from a non-contrast CT image of the brain.

[0008] Identifying the amount of tissue damage to an individual's brain and the amount of brain that may be at risk if blood flow to the tissue cannot be restored can be used to determine the individual's treatment options. In various examples, reperfusion therapy can be performed to restore blood supply to the region of the brain that has been blocked. In various examples, the effectiveness of reperfusion therapy can be based on the degree of existing damage to the brain tissue following the situation that caused the lack of blood flow to the brain tissue. In one or more examples, various interventions to restore blood flow to the region of the brain based on the amount of existing tissue damage to the individual's brain may be ineffective or may cause further damage to the individual. Therefore, since the safety and effectiveness of interventions to restore blood flow to brain tissue may depend on the amount of damage to brain tissue, it is desirable to improve the accuracy of the techniques used to identify the amount of damage to brain tissue due to the obstruction of blood supply.

[0009] Existing imaging techniques may underestimate the amount of damage to brain tissue due to lack of blood supply. For example, a region of the brain where blood flow from a first blood source is blocked can be supplied with blood from a second blood source. However, in various situations, blood supplied to the region from the second blood source may reach the region after damage has already occurred due to the blockage of blood supply from the first blood source. Thus, in situations where perfusion imaging is used to identify tissue damage to a region of the brain, images captured using perfusion imaging techniques may show that blood supply to the region has been restored despite existing damage to the brain tissue of the region. Because images captured using perfusion imaging techniques may show reperfusion, identification of tissue damage based on images captured using reperfusion imaging techniques may be underestimated. In these situations, treatment options provided to an individual may not be effective on the actual amount of damage to brain tissue, but rather on the apparent amount of damage to brain tissue shown by images captured using perfusion-based imaging techniques. Thus, the treatment offered to an individual in these situations may be less effective than in situations where the treatment options offered to an individual correspond to the actual amount of brain tissue damage.

[0010] Perfusion imaging is used in acute stroke patients to infer that tissue is damaged if blood flow in the region is substantially reduced. For example, damaged brain tissue can be identified in situations where blood flow to a region in the first hemisphere of the brain is less than 30% of the value in the corresponding region in the second hemisphere of the brain. Empirically, such a large reduction in blood flow has been found to be a good, but not perfect, predictor of permanently damaged brain tissue. However, the absence of these reductions in cerebral blood flow is not always interpreted as a sign that the brain tissue is not damaged. For example, a thrombolytic agent may have been given to dissolve the clot or the clot may have migrated more distally in the blood vessel and, in so doing, released a segment that contains the outlet (mouth) of a branch vessel. That is, some regions can be at least partially reperfused. In such cases, perfusion-based imaging may not show a reduction in blood flow in these regions as a result of successful reperfusion of these regions. However, the underlying tissue in the region may already be irreversibly damaged. Since the total amount of irreversibly damaged tissue is a key factor in deciding whether to treat a patient or not, it is important knowledge to know as precisely as possible the amount of tissue that cannot be salvaged. Therefore, it is necessary to include additional information to account for irreversibly damaged tissue that is not detectable in perfusion images. This is where additional information from non-contrast CT can be useful. The combination of perfusion-based images and low density values ​​identified using non-contrast CT images gives a truer estimate of infarct size than either method alone.

[0011] The techniques, systems, processes, and methods described herein relate to generating images that provide a more accurate indication of the amount of brain tissue damage based on the disruption of blood supply than existing techniques. In one or more embodiments, the first imaging data can be generated by a first imaging technique, and the second imaging data can be generated by a second imaging technique. The first imaging technique can be a perfusion-based imaging technique. For illustration, the first imaging technique can include CT perfusion imaging. The second imaging technique can be a non-perfusion-based imaging technique. For example, the second imaging technique can include non-contrast CT imaging.

[0012] The voxel intensities of the first imaging data can be analyzed to identify several perfusion parameters indicative of blood supply to several regions of the individual's brain. The several perfusion parameters can be used to identify one or more first regions of the individual's damaged brain tissue. The one or more first regions can be displayed as an overlay placed on one or more images of the individual's brain generated using the first imaging technique. In one or more examples, the perfusion parameters can indicate that one or more first regions of the individual's brain have suffered irreversible damage. Furthermore, the perfusion parameters can indicate that at least one threshold probability of damage has occurred for the one or more first regions.

[0013] Further, the voxel intensities of the second imaging data can be analyzed to identify one or more second regions of the individual's brain tissue that have been damaged. In one or more examples, the second imaging data can be analyzed to generate an indicator of low density associated with a region of the individual's brain. Low density refers to a decrease in density of the brain tissue region. The decrease in density of the brain tissue region can be a result of a higher amount of water in the region than is found in healthy brain tissue. In various examples, the low density can be an indicator of brain tissue damage. In one or more examples, one or more regions of the brain having at least a threshold measure of low density corresponding to brain tissue damage can be shown as an overlay placed on one or more images of the individual's brain generated using the second imaging technique. In one or more examples, the indicator of low density can indicate that one or more second regions of the individual's brain have suffered irreversible damage. Furthermore, the indicator of low density can indicate that there is at least a threshold probability that damage has occurred with respect to the one or more second regions.

[0014] In one or more embodiments, the first region identified based on the first imaging data generated using the first imaging technique and the second region identified based on the second imaging data generated using the second imaging technique can have at least a partial amount of overlap. In various examples, the one or more second regions can include portions of brain tissue that are not included in the one or more first regions. In these circumstances, the combination of the one or more first regions and the one or more second regions can provide a more accurate indication of brain tissue damage for an individual than either the one or more first regions alone or the one or more second regions alone. In one or more illustrative examples, a user interface can be generated that includes an image derived from the first imaging data and includes both a first overlay displaying the one or more first regions and a second overlay displaying the one or more second regions. In this manner, a more accurate view of brain tissue damage for an individual can be provided to a medical professional viewing the user interface via a computing device. Thus, medical personnel can be directed to individual treatment options that may be safer and more effective than when looking at information provided by existing systems related to the amount of damage to an individual's brain tissue.

[0015] FIG. 1 is a diagrammatic representation of an architecture 100 that aggregates data from different imaging modalities and generates an image including an overlay indicating potential damage to brain tissue, according to one or more example embodiments. In one or more examples, the architecture 100 can be implemented to generate a user interface that indicates irreversibly damaged brain tissue. The architecture 100 can include an image processing system 102. The image processing system 102 can be implemented by one or more computing devices 104. The one or more computing devices 104 can include one or more server computing devices, one or more desktop computing devices, one or more laptop computing devices, one or more tablet computing devices, one or more mobile computing devices, or a combination thereof. In certain embodiments, at least a portion of the one or more computing devices 104 can be implemented in a distributed computing environment. For example, at least a portion of the one or more computing devices 104 can be implemented in a cloud computing architecture.

[0016] The image processing system 102 may include a perfusion imaging system 106. The perfusion imaging system 106 may acquire first imaging data 108 generated by a first imaging technique 110. The first imaging technique 110 may be a perfusion-based imaging technique. In one or more examples, the first imaging technique 110 may be a computed tomography (CT)-based imaging technique. In one or more illustrative examples, the first imaging technique 110 may be a CT perfusion-based imaging technique. In one or more additional illustrative examples, the first imaging technique 110 may be a CT angiography-based imaging technique.

[0017] The first imaging data 108 may correspond to one or more images of the brain 112 of the individual 114 captured by a CT imager. In one or more examples, the CT imager may capture several images of the brain 112 of the individual 114 over a period of time. In this manner, a series of images of the brain 112 may be captured consecutively over the period of time. Each image in the series of images may be referred to herein as a "slice." The slices may correspond to different regions of the brain 112. For example, the CT imager may start capturing slices at the shoulder or neck of the individual 114 and move up through the base of the skull, the brain 112, and to the top of the head of the individual 114. In one or more examples, the CT imager may capture multiple images of the same or similar regions of the brain 112 of the individual 114 over time. As part of the imaging process, an imaging agent may be delivered to the individual 114, such as via intravenous injection. The first contrast data 108 may include an image of the brain 112 of the individual 114 showing the brain 112 before the contrast agent is delivered, and an image of the brain 112 of the individual 114 showing the presence of the contrast agent in one or more regions of the brain 112.

[0018] The first contrast data 108 may indicate intensity values ​​of voxels of a captured image of the brain 112 of the individual 114. The intensity values ​​may be in Hounsfield units. In one or more examples, the intensity values ​​of voxels corresponding to regions of the brain 112 in which contrast agent is present may be greater than the intensity values ​​of voxels corresponding to regions of the brain 112 in which contrast agent is not present. In various examples, the intensity values ​​of voxels included in the first contrast data 108 may indicate an amount of contrast agent present in a region of the brain 112. To illustrate, as an amount of contrast agent present in a region of the brain 112 increases, the intensity values ​​of voxels corresponding to that region may increase. Additionally, as an amount of contrast agent present in a region of the brain 112 decreases, the intensity values ​​of voxels corresponding to that region may decrease.

[0019] In one or more examples, the first imaging data 108 may be formatted according to the Digital Imaging and Communications in Medicine (DICOM) standard. In addition to data corresponding to an image captured by the CT imaging device, the first imaging data 108 may include additional information about the image captured by the CT imaging device. For example, the first imaging data 108 may include timing data indicating the time at which each slice of the brain 112 of the individual 114 was captured and / or the time interval at which each slice of the brain 112 of the individual 114 was captured. Additionally, the first imaging data 108 may indicate characteristics of the slice. To illustrate, the first imaging data 108 may indicate at least one of a slice thickness, an inter-slice distance, a voxel dimension, or a voxel location. The first contrast data 108 may also indicate further information, such as at least one of information corresponding to the individual 114, information corresponding to the facility at which the first contrast data 108 was generated, or information corresponding to the CT imaging device that captured the image of the brain 112 of the individual 114 contained in the first contrast data 108.

[0020] The perfusion imaging system 106 may analyze the first imaging data 108 to generate one or more perfusion parameters 116. The one or more perfusion parameters 116 may be indicative of blood flow through one or more regions of the brain 112 of the individual 114. The one or more perfusion parameters 116 may include a measure of cerebral blood flow (CBF). In addition, the one or more perfusion parameters 116 may include a measure of cerebral blood volume (CBV). Furthermore, the one or more perfusion parameters 116 may include a measure of mean tracer transit time (MTT). The MTT may correspond to a mean transit time of the contrast agent through the region of the brain. The one or more perfusion parameters 116 may include a measure of a T max The tissue retention function may also include a function that indicates the probability that an amount of contrast agent that has entered a voxel will subsequently remain within that voxel. One or more perfusion parameters 116 may be determined for each voxel in at least a portion of the first contrast data 108.

[0021] In various examples, the perfusion image processing system 106 can analyze intensity values ​​of voxels of the first contrast image data 108 for one or more regions of the brain 112 to identify one or more perfusion parameters 116. For example, the perfusion image processing system 106 can register one or more images included in the first contrast image data 108 with a template image. The template image can include an anatomical template derived from images of the brains of many individuals. After being registered with the template image, the one or more images included in the first contrast image data 108 can be aligned with an atlas that indicates several regions of the human brain. In various examples, the atlas can indicate the location of blood vessels, parenchyma, etc. In this manner, several regions can be identified for one or more images included in the first contrast image data 108. In one or more examples, the one or more images of the first contrast image data 108 can be labeled according to several regions included in the atlas. The perfusion imaging system 106 may then analyze intensity values ​​of voxels corresponding to one or more of the regions over a period of time to determine one or more perfusion parameters 116. To illustrate, the perfusion imaging system 106 may analyze the intensity of voxels corresponding to blood vessels in the brain 112 to determine one or more perfusion parameters 116.

[0022] The perfusion imaging system 106 may also analyze one or more perfusion parameters 116 in conjunction with the first imaging data 108 to identify one or more regions of the brain 112 that are tissue damaged due to an obstruction in blood supply to the one or more regions. For example, the first imaging data 108 may include a first image 118 of the brain 112 of the individual 114. In one or more illustrative examples, the first image 118 may include a slice captured by a CT imager at a given time. The first image 118 may show blood vessels 120 in which contrast agent is present. The perfusion imaging system 106 may analyze the one or more perfusion parameters 116 and intensity values ​​of voxels corresponding to blood vessels 120 that supply blood to a section 122 of the brain 112 to identify the degree of obstruction in blood flow to the section 122 and the amount of time that blood flow to the section 122 is obstructed. Based on the degree of obstruction of blood flow to section 122, perfusion imaging system 106 can identify a first region of interest 124 in section 122 where the probability of tissue damage is at least a threshold probability. In one or more embodiments, perfusion imaging system 106 can generate first overlay contrast data 126 corresponding to a first overlay indicative of first region of interest 124. In one or more examples, the first overlay can be displayed in conjunction with first image 118.

[0023] The perfusion imaging system 106 can identify regions having damaged brain tissue based on the individual perfusion parameters 116. For example, the perfusion imaging system 106 can identify regions having a probability of containing damaged brain tissue that is at least a threshold probability based on one or more measures of cerebral flood flow of voxels contained in the region. In one or more additional examples, the perfusion imaging system 106 can include a T maxIn one or more further examples, the perfusion imaging system 106 can identify regions having damaged brain tissue based on one or more values ​​of cerebral blood volume. The perfusion imaging system 106 can identify regions having damaged brain tissue based on one or more values ​​of mean tracer transit time. In yet an additional example, the perfusion imaging system 106 can identify regions having damaged brain tissue based on one or more values ​​of cerebral blood flow, cerebral blood volume, T max Regions having damaged brain tissue can be identified based on one or more values ​​of at least one of the mean tracer transit time, or the mean tracer transit time.

[0024] In one or more illustrative examples, the perfusion imaging system 106 may identify the amount of damaged brain tissue and / or the predicted amount of damaged brain tissue based on the difference between the regions identified using two or more of the perfusion parameters 116. To illustrate, the perfusion imaging system 106 may identify a first region having a probability of containing damaged brain tissue that is at least a threshold probability using the value of cerebral blood flow, and a second region having a probability of containing damaged brain tissue that is at least a threshold probability using the value of cerebral blood flow. max and a second region having a probability of containing damaged brain tissue that is at least a threshold probability using a value of . In one or more examples, the first region can correspond to an approximation of the amount of damaged brain tissue at a first time, and the second region can correspond to an approximation of the amount of damaged brain tissue at a later second time, the volume of the second region being larger than the first region. In these situations, the volumetric difference between the first region and the second region can indicate that the volume of the first region can increase over time to the volume of the second region. In various examples, the intervention prescribed to treat the damaged brain tissue or minimize the amount of damaged brain tissue can be based on the volumetric difference between the first region and the second region. For example, a first treatment can be prescribed in situations where the volumetric difference between the first region and the second region is less than a threshold difference, and a second treatment can be prescribed in situations where the volumetric difference between the first region and the second region is equal to or greater than the threshold difference.

[0025] The perfusion image processing system 106 may also implement one or more machine learning techniques to identify regions of tissue in the brain 112 of the individual 114 that are damaged. In various examples, one or more machine learning techniques may be used to identify regions of tissue in the brain 112 of the individual 114 where the probability of damage is at least a threshold probability. In one or more examples, one or more convolutional neural networks may be implemented to identify regions of potentially damaged tissue in the brain 112 of the individual 114. For example, a U-Net architecture may be implemented to identify regions of damaged tissue in the brain 112 of the individual 114. Additionally, one or more classification convolutional neural networks may be implemented to identify regions of damaged tissue in the brain 112 of the individual 114. In one or more illustrative examples, the low-concentration analysis system 128 may acquire several CT perfusion images as training images. The training images may include a first several images of a brain of a first individual having one or more damaged regions and a second several images of a brain of a second individual not including damaged regions. In one or more circumstances, the first few images can be classified as having one or more damaged regions, and the second few images can be classified as not having damaged regions. In one or more further examples, a designated region of the first individual's brain contained in the first images can be classified as a damaged region. Values ​​for parameters of the one or more models generated in conjunction with the one or more machine learning techniques can be identified through a training process. After the training process is complete and the one or more models are validated using additional sets of images, the one or more models can be used to classify damaged regions of new images.

[0026] The image processing system 102 may also include a low-density analysis system 128. The low-density analysis system 128 may identify density values ​​of one or more regions of the brain 112 of the individual 114 based on the second imaging data 130. In one or more examples, the low-density analysis system 128 may identify one or more regions of the brain 112 of the individual 114 that include tissue having density values ​​below one or more threshold values. In various examples, brain tissue having density values ​​below one or more threshold values ​​may indicate that damage has occurred in the brain tissue.

[0027] The second imaging data 130 can be generated by a second imaging technique 132. The second imaging technique 132 can be a non-perfusion-based imaging technique. In various examples, the second imaging technique 132 can be a non-contrast-based imaging technique. In one or more illustrative examples, the second imaging technique 132 can perform one or more non-contrast CT imaging techniques. However, the second imaging technique 132 can also be replaced with one or more images from the first imaging technique 108, particularly images acquired before the contrast agent reaches the brain. In these situations, the first imaging technique 110 and the second imaging technique 132 can include the same imaging modality as the first imaging data 108 and the second imaging data 130 captured at different times. For purposes of illustration, the first contrast data 108 can be acquired using a contrast technique such as CT perfusion during a period of time when contrast agent is present in the brain 112 of the individual 114, and the second contrast data 130 can be acquired using the same contrast modality during a period of time when contrast agent is not present in the brain 112 of the individual 114.

[0028] The second imaging data 130 may correspond to one or more images of the brain 112 of the individual 114 captured by a CT imager. In one or more examples, the CT imager may capture several images of the brain 112 of the individual 114 over a period of time. In this manner, a series of images of the brain 112 may be captured sequentially over a period of time. The slices may correspond to different regions of the brain 112. For example, the CT imager may start capturing images of the shoulder or neck of the individual 114 and move up through the base of the skull, the brain 112, and to the top of the head of the individual 114. In one or more examples, the CT imager may capture multiple images of the same or similar regions of the brain 112 of the individual 114 over time.

[0029] The second contrast data 130 may represent intensity values ​​of voxels of a captured image of the brain 112 of the individual 114. The intensity values ​​may be represented in Hounsfield units. In one or more examples, the intensity values ​​of voxels corresponding to regions of the brain 112 having a relatively low density have relatively low intensity values ​​compared to regions of the brain 112 having a relatively high density. In these circumstances, the intensity values ​​of voxels included in the second contrast data 130 increase as the density of the brain tissue corresponding to the voxels increases.

[0030] In one or more examples, the second contrast data 130 may be formatted according to the Digital Imaging and Communications in Medicine (DICOM) standard. In addition to data corresponding to an image captured by the CT imaging device, the second contrast data 130 may include additional information about the image captured by the CT imaging device. For example, the second contrast data 130 may include timing data indicating the time at which each slice of the brain 112 of the individual 114 was captured and / or the time interval at which each slice of the brain 112 of the individual 114 was captured. Additionally, the second contrast data 130 may indicate characteristics of the slices. To illustrate, the second contrast data 130 may indicate at least one of slice thickness, inter-slice distance, voxel dimensions, or voxel locations. The second contrast data 130 may also indicate further information, such as at least one of information corresponding to the individual 114, information corresponding to the facility where the first contrast data 108 was generated, or information corresponding to the CT imaging device that captured the image of the brain 112 of the individual 114 included in the second contrast data 130.

[0031] In various examples, the low-density analysis system 128 can analyze the intensity values ​​of voxels in the second contrast data 130 for one or more regions of the brain 112 to identify Hounsfield density values ​​for the regions of the brain 112 and identify regions of the brain 112 that are damaged tissue based on the Hounsfield density values. In various examples, the low-density analysis system 128 can analyze the intensity values ​​of voxels in the second contrast data 130 to identify regions of the brain 112 of the individual 114 where the probability of damage is at least a threshold probability. In one or more examples, the low-density analysis system 128 can register one or more images in the second contrast data 130 with a template image. The template image can include an anatomical template derived from images of the brains of many individuals. After being registered with the template image, one or more images in the second contrast data 130 can be aligned with an atlas that represents several regions of the human brain. In this manner, several regions can be identified for one or more images included in the second contrast image data 130. In one or more examples, one or more images of the second contrast image data 130 can be labeled according to several regions included in the atlas. To illustrate, the low-intensity analysis system 128 can use the atlas to identify the ventricles of the brain 112, the cerebrospinal fluid within the brain 112, and the soft tissues of the brain 112, such as the parenchyma and additional blood vessels.

[0032] In one or more examples, the low-density analysis system 128 can analyze voxels in different hemispheres of the brain 112 to identify one or more regions of the brain 112 that include low-density tissue. In one or more illustrative examples, the second imaging data 130 can include a second image 134 of the brain 112 of the individual 114. In various examples, the second image 134 can include slices captured by a CT imager at a given time using one or more non-contrast CT imaging techniques. The low-density analysis system 128 can identify spatial correlations between voxels included in a first hemisphere 136 of the brain 112 and voxels included in a second hemisphere 138 of the brain 112. In one or more examples, the first hemisphere 136 and the second hemisphere 138 can be referred to herein as contralateral to one another. Additionally, a first voxel located in the first hemisphere 136 that spatially corresponds to a second voxel located in the second hemisphere 138 may be referred to herein as contralateral to one another.

[0033] The low-intensity analysis system 128 can then analyze the intensity values ​​of voxels located in the first hemisphere 136 with respect to the intensity values ​​of voxels located in the second hemisphere 138. In various examples, the low-intensity analysis system 128 can identify a difference between a contralateral voxel location in the first hemisphere 136 and a contralateral voxel location in the second hemisphere 138. In one or more examples, the low-intensity analysis system 128 can identify one or more regions of the brain 112 having first voxels that have intensity values ​​that differ from contralateral second voxels by at least a threshold difference. In the illustrative example of FIG. 1 , the low-intensity analysis system 128 includes voxels in a second region of interest 140 in the second hemisphere 138 that have intensity values ​​that differ from voxel locations in the first hemisphere 136 that are contralateral to the voxel locations in the second region of interest 140 by at least a threshold difference. In one or more examples, the second region of interest 140 may indicate damaged brain tissue due to an obstruction in the blood supply to the second region of interest 140. In one or more additional examples, the low-intensity analysis system 128 may identify the second region of interest 140 using one or more machine learning techniques instead of or in addition to a contralateral analysis of intensity values ​​of voxel locations in the first hemisphere 136 and the second hemisphere 138. In one or more embodiments, the low-intensity analysis system 128 may generate second overlay contrast data 142 corresponding to a second overlay showing the second region of interest 140. In one or more examples, the first overlay may be displayed in conjunction with the second image 134.

[0034] The low-density analysis system 128 may also implement one or more machine learning techniques to identify regions of low-density tissue in the brain 112 of the individual 114. In one or more examples, one or more convolutional neural networks may be implemented to identify regions of low-density tissue in the brain 112 of the individual 114. For example, a U-Net architecture may be implemented to identify regions of low-density tissue in the brain 112 of the individual 114. Additionally, one or more classification convolutional neural networks may be implemented to identify regions of low-density tissue in the brain 112 of the individual 114. In one or more illustrative examples, the low-density analysis system 128 may acquire several non-contrast CT images as training images. The training images may include a first number of images of a first individual's brain having one or more low-density regions and a second number of images of a second individual's brain not including low-density regions. In one or more situations, the first number of images may be classified as having one or more low-density regions and the second number of images may be classified as not including low-density regions. In one or more further examples, a designated region of the first individual's brain contained in the first image can be classified as a low density region. Values ​​for parameters of the one or more models generated in conjunction with the one or more machine learning techniques can be identified through a training process. After the training process is complete and the one or more models are validated using additional sets of images, the one or more models can be used to classify low density regions of new images.

[0035] The image processing system 106 may include an output image system 144. The output image system 144 may acquire the first overlay contrast imaging data 126 and the second overlay contrast imaging data 142 to generate one or more aggregate images 146. The one or more aggregate images 146 may include a first overlay corresponding to the first overlay contrast imaging data 126 and a second overlay corresponding to the second overlay contrast imaging data 142. In one or more examples, the output image system 144 may generate the one or more aggregate images 146 using at least one of the first contrast imaging data 108 or the second contrast imaging data 130 in conjunction with the first overlay contrast imaging data 126 and the second overlay contrast imaging data 142. For example, the output imaging system 144 may generate one or more aggregate images 146 to include an image of the first contrast imaging data 108 with a first overlay corresponding to the first overlay contrast imaging data 126 and a second overlay corresponding to the second overlay contrast imaging data 142. In one or more illustrative examples, the output imaging system 144 may generate one or more aggregate images 146 to include a first image 118 having a first overlay corresponding to the first region of interest 124 and a second overlay corresponding to the second region of interest 140.

[0036] In one or more instances, when viewed individually, the first region of interest 124 may indicate a volume of damaged brain tissue that is less than the actual volume of damaged brain tissue. In these circumstances, the first region of interest 124 may have a volume that is less than the volume of the second region of interest 140. Thus, by generating one or more summary images 146 that indicate the volumetric difference between the first region of interest 124 and the second region of interest 140, the output imaging system 144 may provide a user interface to medical personnel that includes the summary image 146 that indicates a more accurate estimate of the damage to the tissue of the brain 112. As a result, patient selection for reperfusion therapy may be improved and potentially ineffective procedures may be avoided, such as when the entire hypo-dense region is already infarcted.

[0037] Although not shown in the illustrative example of FIG. 1 , the image processing system 102 may also include a brain tissue damage and risk analysis system. The brain tissue damage and risk analysis system may identify volumetric differences of regions of interest identified using different perfusion parameters 116. For example, the brain tissue damage and risk analysis system may identify a first volume of a first region of interest, e.g., an infarct region, according to a first perfusion parameter at a first threshold, and identify a second volume of a second region of interest, e.g., the risk region plus the infarct region, according to a second perfusion parameter at a second threshold. To illustrate, the brain tissue damage and risk analysis system may identify a first volume of a first region of the brain 112 based on the relative cerebral blood flow in the first region being at least 30% less than the relative cerebral blood flow in a contralateral region of the brain 112. Additionally, the brain tissue damage and risk analysis system may identify a volumetric difference of a region of interest identified using different perfusion parameters 116. max A second volume of a second region of the brain 112 having a brain tissue injury probability of at least 100% can be identified. A volumetric difference between the first volume and the second volume, i.e., tissue at risk, can be calculated and displayed within the user interface. In one or more illustrative examples, the difference between the first volume and the second volume can be referred to herein as a mismatch volume. Additionally, a ratio between the first volume and the second volume can also be calculated and displayed within the user interface. The ratio between the first volume and the second volume can be referred to herein as a mismatch ratio. In various examples, the mismatch volume and mismatch ratio can be used by a medical professional to determine a treatment recommendation for an individual having a brain tissue injury probability of at least a threshold probability.

[0038] Further, although not shown in the illustrative example of FIG. 1 , the first imaging technique 110 can include a CT angiography system, and the first imaging data 108 can include a CT angiography image. In these circumstances, the image processing system 102 can include an additional image processing system to identify one or more regions of the brain 112 having damaged tissue based on the CT angiography images. In one or more examples, the additional image processing system can derive core regions of the brain 112 by identifying regions of the one or more CT angiography images in which slight signal enhancement is absent, and identifying stenosis of blood vessels of the brain 112 based on the one or more CT angiography images. Regions of the brain 112. The stenosis of blood vessels in one or more regions of the brain 112 can be analyzed to identify a probability of one or more regions having damaged tissue or to identify a measure of damage to one or more regions. In various examples, one or more machine learning techniques can be implemented to analyze cerebral blood vessel narrowing and identify regions of the brain 112 of an individual 114 that contain damaged tissue. For illustrative purposes, several CT angiography training images can be acquired. The training images can include a first number of images of a brain of a first individual having narrowing of blood vessels in a damaged region of the brain, and a second number of images of a brain of a second individual that does not contain narrowing of blood vessels caused by damaged tissue. The first number of images can be classified as having at least one region with damaged tissue, and the second number of images can be classified as having no damaged region. In one or more additional examples, a designated region of the brain of the first individual contained in the first image can be classified as a damaged region. Values ​​for parameters of the one or more models generated in conjunction with the one or more machine learning techniques can be identified through a training process. After the training process is completed and the one or more models are validated using additional sets of images, the one or more models can be used to classify brain regions contained in new images as containing or not containing damaged tissue.

[0039] FIG. 2 is a diagrammatic representation of an architecture 200 for determining perfusion parameters related to blood flow through an individual's brain, according to one or more example embodiments. The CT perfusion image data 202 may include a perfusion image processing system 106. The perfusion image processing system 106 may acquire CT perfusion image data 202. The CT perfusion image data 202 may be captured by a CT imaging device. During a perfusion-based CT imaging process, a contrast agent may be delivered to the individual, and during the perfusion-based imaging process, the presence and movement of the contrast agent through the individual's brain may be captured. The presence and movement of the contrast agent through the individual's brain may correspond to the presence and movement of blood through the individual's brain. Additionally, the CT perfusion image data 202 may include several images captured over a period of time. The several images included in the CT perfusion image data 202 may be referred to herein as slices. The CT perfusion image data 202 may be formatted and may include information that complies with the DICOM standard.

[0040] The perfusion image processing system 106 may perform several processes, such as generating one or more pre-contrast images in operation 204. To illustrate, the perfusion image processing system 106 may perform several operations in operation 206 to analyze the CT perfusion image data 202 for several rules in order for the CT perfusion image data 202 to be processed by the perfusion image processing system 106 in operation 206. For example, the perfusion image processing system 106 may analyze the CT perfusion image data 202 in operation 206 to determine whether the CT perfusion image data 202 includes a tag, such as a tag, associated with the timing of the capture of the image. The timing of the sampling of the images included in the CT perfusion image data 202 may be determined in operation 206 based on the timing tag included in the CT perfusion image data 202. In one or more examples, the perfusion image processing system 106 may determine the amount of CT perfusion image data 202 to be processed based on a maximum time threshold of the scan used to capture the CT perfusion image data 202, such as 1000 seconds. The perfusion image processing system 106 may also determine in operation 206 whether the CT perfusion image data 202 includes tags indicating image location, spacing, and orientation. Information included in the CT perfusion image data 202 may be used by the perfusion image processing system 106 to determine slice overlay. Operation 206 may also determine whether the CT perfusion image data 202 includes information that the perfusion image processing system 106 may use to calculate a region of interest that may indicate damaged brain tissue at risk of infarction.

[0041] In operation 208, the perfusion image processing system 106 may perform a motion correction process. In various examples, patient motion during image acquisition may degrade the quality of the perfusion image. The motion correction process in operation 208 may reduce the effects of patient motion during image acquisition. The motion correction process in operation 208 may include three-dimensional (3D) rigid coregistration for spatial misregistration. In one or more examples, a time point in the perfusion time series is registered with a reference volume. The reference volume may be a volume that is most similar to a number of other volumes in the CT perfusion image data 202. To illustrate, individual slices included in the CT perfusion image data 202 may be analyzed to identify a similarity metric relative to one another. The similarity metric of the individual slices may be optimized to identify a slice with a maximum value of the similarity metric with respect to the largest number of additional images. In one or more embodiments, at least a portion of the slices that were not determined to be reference images may be analyzed with respect to the reference image and may be identified as slices that correspond to the motion of the individual during the imaging process. In one or more illustrative examples, the similarity measure may be determined using a mean squared difference procedure.

[0042] For each image slice, rotation and / or translation parameters that optimize the similarity between the individual image slice and the reference image can be identified. The translation and / or rotation parameters can be used to reregister the image segments. In one or more illustrative examples, slices can be resampled with respect to the reference image to change at least one of the positions or sizes of the voxels of the slice to correspond to the positions and / or sizes of the voxels of the reference image. For illustrative purposes, slices capturing images during a time period in which the individual is moving can be resampled to a new position using the rotation and translation parameters in situations where the position difference from the reference image exceeds a specified amount, such as 10% of the voxels in any dimension, 25% of the voxels in any dimension, half of the voxels in any dimension, or 75% of the voxels in any dimension. The dimensions of the voxels can be identified by the perfusion image processing system 106 from information contained in the CT perfusion image data 202. In this manner, slices are resampled in situations where a cost function is optimized and the repositioning results in at least a threshold amount of improvement in the registration between the reference image and some additional slices.

[0043] In various examples, the motion correction process in act 206 can be used to identify corrupted slices that are less than a threshold amount of registration with the reference image. The corrupted slices can be labeled by the perfusion image processing system 106 and may not be utilized in the calculation of perfusion parameters by the perfusion image processing system 106. The end result of the motion correction process in act 208 can be to generate motion corrected image data that maximizes the number of slices included in the CT perfusion image data 202 that are registered such that the anatomical structures included in the slices of the CT perfusion image data 202 are in relatively the same or similar positions as the anatomical structures of the reference image. In one or more examples, non-anatomical structures, such as a head holder that holds the individual's head during the imaging process, can be removed.

[0044] In operation 210, a time correction process may be performed by the perfusion image processing system 106 based on slices of the CT perfusion image data 202 acquired at various time intervals. The time correction process performed in operation 210 may be performed on motion corrected data generated by the motion correction process performed in operation 208. The time correction operation may include resampling the CT perfusion image data 202 onto a common time axis having a regular interval of time. The regular interval of time may be about 0.1 seconds to about 2 seconds, about 0.1 seconds to 1 second, about 0.5 seconds to 2 seconds, about 1 second to 2 seconds, or about 0.5 seconds to about 1 second. In one or more illustrative examples, the regular interval of time may be 1 second. In one or more additional illustrative examples, the regular interval of time may be 0.5 seconds. In one or more further illustrative examples, the regular interval of time may be 2 seconds. In yet another illustrative example, the regular interval of time may be 0.25 seconds. The time correction process performed in operation 210 may include using portions of the CT perfusion image data 202 associated with the voxels of the slice to identify a piecewise linear curve at each spatial location represented by coordinates on the X-axis, Y-axis, and Z-axis based on timing data included in the CT perfusion image data 202. The piecewise linear curves at each spatial location may be resampled to a defined constant time interval to generate a time-resolved data set having a common time interval. The timing data of the corrupted slice identified from the motion correction process performed in operation 208 and removed from the motion-corrected data may be interpolated using linear interpolation based on slices captured at a time proximate to the corrupted slice.

[0045] The perfusion image processing system 106 may also perform a process to estimate the properties and timing of the contrast agent with respect to the CT perfusion image data 202 in operation 212. For example, in operation 212, the perfusion image processing system 106 may determine the arrival time of the contrast agent for voxels corresponding to brain tissue. The brain tissue may include blood vessels and blood within the blood vessels. In various examples, an average contrast agent passage curve may be generated by determining the average of the signal change across the voxels corresponding to the brain tissue at a given time point. The contrast agent arrival time of one or more voxels may be determined using information determined from the average contrast agent passage curve. The contrast agent arrival time may be used to determine a baseline time frame that includes the time from the start of the CT scan to the contrast agent arrival time. The baseline time frame may then be used by the perfusion image processing system 106 to determine perfusion parameters of the CT perfusion image data 202. For example, in act 214, a pre-contrast baseline image 216 of the CT perfusion image data 202 can be identified as an average of slices within a baseline time range before contrast arrives. The pre-contrast baseline image 216 can be provided for subsequent processing by the perfusion image processing system 106. For example, the pre-contrast baseline image 216 can be used in act 218 to register the image data to an anatomical template and identify regions of the individual's brain. The pre-contrast baseline image 216 can also be provided to the output image system 144 for use in generating an aggregate image with an overlay. In various examples, the pre-contrast baseline image 216 can be used to perform registration of a non-perfusion based image to a perfusion based contrast space.

[0046] The registration of the CT perfusion image data 202 to the anatomical template and the identification of the individual's brain regions based on the CT perfusion image data 202 may include, in operation 220, segmenting brain tissue and generating a brain mask. The brain mask may be determined based on a pre-contrast baseline image 216 calculated before the arrival of contrast. The brain mask may include a two-dimensional image of the individual's brain features, such as brain tissue including blood vessels and parenchyma. Additional features included in the pre-contrast baseline image 216, such as skull, dura, fat, skin, muscle, eyes, and bone, may be removed. The brain mask may be generated by performing morphological operations, such as opening and closing, on the voxels of the pre-contrast baseline image 216. The morphological operations may be followed by connected component analysis techniques to remove non-brain features from the pre-contrast baseline image 216 to generate the brain mask. One or more intensity value thresholds may be used to determine the voxels used to generate the brain mask. In one or more examples, at least one of a relative or absolute intensity value threshold may be used.

[0047] In an embodiment in which the CT perfusion image data 202 is generated using a CT imaging technique having an anatomical coverage along the Z-axis less than a threshold amount, such as 5 mm, 10 mm, 20 mm, 30 mm, 40 mm, 50 mm, 60 mm, or 75 mm, a segmentation process may be performed in operation 220 to generate a brain mask. The anatomical coverage of the CT imaging technique may correspond to at least one of a number of data channels, a number of detector rows, a pitch, a section thickness, a scan time, or a gantry rotation time. The segmentation process may determine a segmentation threshold comprising various intensity values. The range of intensity values ​​may correspond to brain tissue based on previously identified intensity values ​​of brain tissue. A voxel may be included in the brain mask based on a determination that the intensity value of the voxel falls within a range of values ​​corresponding to the segmentation threshold. Additionally, a voxel may be excluded from the brain mask based on a determination that the intensity value of the voxel falls outside a range of values ​​corresponding to the segmentation threshold.

[0048] In situations where the CT perfusion image data 202 is generated using a CT imaging technique having at least a threshold amount of anatomical coverage along the Z-axis, a brain mask can be determined in act 222 using an anatomical template that is co-registered with the CT perfusion image data 202 to identify intracranial tissue. The anatomical template can include a composite image of a human skull and brain generated from several CT images of the human brain. The CT perfusion image data 202 can be elastically co-registered with the anatomical template to generate modified image data. The modified image data can correspond to a parameter indicative of an amount of modification of the CT perfusion image data 202 to correspond to the anatomical template. In various examples, a deformation field can be generated based on the registration process. In one or more examples, the deformation field can be applied to an atlas that is indicative of several regions of an individual's brain. The atlas can be generated based on the locations of individual brain regions derived from several images of the brains of several individuals. The atlas can be transformed to correspond to the CT perfusion image data 202, and a brain mask for the CT perfusion image data 202 can be generated based on regions of the CT perfusion image data 202 that correspond to the modified atlas. In one or more embodiments, the brain mask determined using the procedure performed with respect to operation 220 in situations where the anatomical coverage along the Z axis is at least a threshold amount can be combined with the brain mask determined using the procedure performed in operation 222 in situations where the anatomical coverage along the Z axis is equal to or less than the threshold amount.

[0049] Act 224 may include generating intensity difference values ​​for voxels corresponding to the effect of a contrast effect present in a voxel of the CT perfusion image data 202. The CT perfusion image data 202 may be modified to indicate the contrast effect by removing information included in the CT perfusion image data 202 corresponding to a time period during which contrast was not present at one or more voxel locations. In various examples, the acts used to identify the intensity values ​​of voxels corresponding to the presence of contrast may be based on one or more contrast techniques generating the image data. For example, in a situation where computed tomography is used to generate the CT perfusion image data 202, the intensity values ​​of pixels corresponding to the presence of contrast may be identified based on values ​​of a contrast concentration curve indicating the intensity values ​​of voxels at several locations at a given time in Hounsfield units (HU). In one or more examples, the concentration curve values ​​may be based on the difference between the intensity values ​​in the CT perfusion image data at a given location at a specified time and a baseline intensity value. In one or more illustrative examples, the baseline value may be identified based on the average pre-contrast baseline slice identified in act 214.

[0050] The process of registering the image data to the anatomical template and identifying regions of the individual's brain in act 218 may also include determining an arterial input function (AIF) and a venous output function (VOF) in act 226. Act 226 may include determining the arterial input function by identifying clusters of voxels corresponding to large blood vessels in the brain. The arterial input function may indicate the concentration of contrast agent over time in the relatively large input arteries. In one or more examples, the arterial input function may be based on blood vessels in the middle cerebral artery or anterior cerebral artery of the brain. In additional examples, such as when the anatomical coverage is below the computed tomography contrast threshold, candidate clusters of voxels for the arterial input function may be determined in portions of the CT perfusion image data 202 corresponding to the anterior and posterior portions of the brain after bone and other non-brain tissue have been removed.

[0051] The venous output function can be identified by identifying clusters of voxels that correspond to large blood vessels in the brain. The venous input function can indicate the concentration of contrast agent over time in the larger output veins. In one or more examples, the venous output function can be based on blood vessels in the sagittal and straight sinuses of the brain venous areas. In additional examples, such as when anatomical coverage is below the computed tomography contrast threshold, cluster candidates can be determined in portions of the image data that correspond to the anterior and posterior portions of the brain after bone and other non-brain tissue has been removed.

[0052] In one or more examples, the process in act 226 of identifying the arterial input function and the venous output function can include identifying the contrast concentration curve amplitude, the time of the contrast concentration curve peak, and the width of the contrast concentration curve peak at each voxel representing brain tissue for the voxel candidate. In various examples, the highest peak of a given voxel can be used to identify the contrast concentration curve amplitude, the time of the contrast concentration curve peak, and the width of the contrast concentration curve peak for a given voxel. The peak information of each voxel can be analyzed for target peak shapes corresponding to a range of amplitude values, a range of time values, and a range of width values. A subset of the voxel candidates can be identified based on an analysis of the voxel peak information according to one or more criteria. The one or more criteria can include a negative amplitude, a peak width less than a threshold peak width, or a peak outside the remainder of at least a threshold number of peaks of voxels included in the CT perfusion image data 202. For the subset of voxels that meet one or more criteria, the perfusion imaging system 106 may determine the average values ​​of the peak amplitude, offset, and width. The perfusion imaging system 106 may also determine the standard deviation of the peak amplitude, offset, and width values ​​for the subset of voxel candidates. The value of the peak information for each voxel in the subset of voxel candidates may be normalized with respect to the average values ​​of the peak amplitude, offset, and width.

[0053] An arterial input function score may be determined for each voxel in the subset of voxel candidates. A score may be determined for each voxel candidate based on the amount that the amplitude value of the individual voxel is greater than the average amplitude value. A score may also be determined for each voxel candidate based on the amount that the offset value of the individual voxel is ahead of the average offset value. Additionally, a score may be determined for each voxel candidate based on the amount that the peak is narrower than the average peak value.

[0054] The perfusion imaging system 106 may identify at least a threshold number of voxels with the highest scores. The threshold number of voxels may be 10 to 50, 12 to 45, 15 to 30, or 18 to 25. The perfusion imaging system 106 may also identify voxels from this additional subset that do not have neighboring voxels included in the second group. The mean arterial input function may be determined using the time vs. contrast agent concentration curves of the remaining voxels.

[0055] A score for the venous output function of each voxel included in the subset of voxel candidates may also be determined. A score for each voxel candidate may be determined based on the amount by which the amplitude value of the individual voxel is greater than the average amplitude value. A score for a voxel candidate having a contrast arrival time that is at least 0.5 seconds, at least 1 second, at least 2 seconds, at least 3 seconds, or at least 4 seconds after the arrival time of the contrast agent for the arterial input function may be determined. In one or more examples, a score for a voxel candidate having a contrast arrival time that is about 2 seconds to about 15 seconds after the arrival time of the contrast agent for the arterial input function, about 3 seconds to about 12 seconds after the arrival time of the contrast agent for the arterial input function, about 4 seconds to about 10 seconds after the arrival time of the contrast agent for the arterial input function, or about 2 seconds to about 10 seconds after the arrival time of the contrast agent for the arterial input function may be determined. The perfusion imaging system 106 may identify at least a threshold number of voxels having the highest scores. The threshold number of voxels can be approximately 10 to 50, 12 to 45, 15 to 30, or 18 to 25. The perfusion imaging system 106 can also identify voxels from this additional subset that do not have neighboring voxels included in the second group. The mean arterial power function can be determined using the time vs. contrast concentration curves of the remaining voxels to determine the mean venous power function.

[0056] The perfusion image processing system 106 may also generate perfusion parameters in act 228. Generating the perfusion parameters may include one or more spatial filtering operations to reduce noise in at least a portion of the CT perfusion image data 202 in act 230. For example, the one or more spatial filtering operations may remove abrupt intensity changes. For illustrative purposes, a Gaussian smoothing kernel with a specified filter width that includes a minimum amount of spatial filtering to a maximum amount of spatial filtering applied. In one or more illustrative examples, the specified minimum amount of filtering may be 0.0 mm, 0.1 mm, 0.25 mm, 0.5 mm, 0.75 mm, or 1 mm. In one or more additional illustrative examples, the maximum amount of spatial filtering may be 7 mm, 8 mm, 9 mm, 10 mm, 11 mm, 12 mm, or 13 mm. In various examples, data corresponding to voxels used in the calculation of the arterial input function and the venous output function are excluded from the one or more spatial filtering operations. Furthermore, in one or more embodiments, voxels included in the CT perfusion image data 202 that do not correspond to brain tissue may be excluded from one or more spatial filtering operations.

[0057] Further, in operation 232, one or more deconvolution operations may be performed to identify a tissue retention function. The tissue retention function may describe the probability that an amount of contrast agent that entered a voxel at a given time remains in the voxel at a later time. The one or more deconvolution operations may be performed on the contrast agent concentration versus time curves of the voxels corresponding to brain tissue. In one or more examples, the one or more deconvolution operations may include performing one or more normalized Fourier transforms using one or more filters applied to the Fourier transform spectrum.

[0058] In operation 234, one or more perfusion parameters for individual voxels included in at least some individual slices of the CT perfusion image data 202 may be identified. The one or more perfusion parameters may be included in perfusion parameter data 236. The perfusion parameter data 236 may be used by the perfusion image processing system 106 to identify brain tissue having at least a threshold probability of being damaged and / or to identify brain tissue at risk of damage. In one or more circumstances, the damaged brain tissue may become infarcted. Furthermore, in various examples, the damaged brain tissue may become irreversibly damaged.

[0059] The one or more perfusion parameters can include relative cerebral blood volume (rCBV). The relative cerebral blood volume can be determined by identifying an area under a tissue residue function. Additionally, the one or more perfusion parameters can include relative cerebral blood flow (rCBF). The relative cerebral flood flow can be determined by identifying a peak in the tissue residue function. Additionally, the one or more perfusion parameters can include mean tracer transit time (MTT). The mean tracer transit time of an individual voxel can be determined using the central volume principle. The central volume principle indicates that MTT=CBV / CBF. In one or more additional examples, the one or more perfusion parameters can include T max It can include T max can be identified by identifying peaks in the tissue residue function.

[0060] In one or more illustrative examples, the cerebral blood volume, cerebral blood flow, and mean tracer transit time of individual voxels can be determined using techniques related to those described in "High resolution measurement of cerebral blood flow using intravascular tracer bolus passages" by L. Ostergaard, RM Weiskoff, DA Chesler, C. Gyldensted, and BR Rose, Part I: Mathematical approach and statistical analysis, Magn Reson Med., November 1996; 36(5):715-25. The values ​​of cerebral blood volume, cerebral blood flow, and mean tracer transit time can be relative values. In one or more further illustrative examples, the T of individual voxels can be determined based on the maximum value of the tissue retention function for each time slice. max It is possible to identify T max The value of can be an absolute value given in seconds.

[0061] 3 is a diagrammatic representation of an example of an architecture 300 for identifying hypodensity measures for brain tissue, according to one or more example embodiments. Hypodensity brain areas can include voxels having lower intensity values ​​than corresponding regions of the brain in the contralateral hemisphere. In one or more examples, lower intensity values ​​of the hypodensity regions can indicate increased water content in the tissue corresponding to the voxels of the hypodensity regions.

[0062] The low-density analysis of the CT image data can be identified using non-contrast CT image data 302. The non-contrast CT image data 302 can include several two-dimensional slices that, when combined, form a three-dimensional image of the subject's brain. In one or more examples, the non-contrast CT image data 302 can cover at least 60 mm, at least 70 mm, at least 80 mm, at least 90 mm, at least 100 mm, at least 110 mm, at least 120 mm, at least 130 mm, at least 140 mm, or at least 150 mm of the subject's brain along an axial direction. In one or more additional examples, the non-contrast CT image data 302 can include slices having a thickness of about 0.5 mm to about 5 mm, about 0.1 mm to about 6 mm, about 1 mm to about 4 mm, about 2 mm to about 5 mm, or about 2 mm to about 6 mm. In various examples, the non-contrast CT image data 302 can have a slice spacing having a value similar to the slice thickness.

[0063] The low-concentration analysis system 128 may perform one or more image data pre-processing operations in operation 304, which may be performed on the non-contrast CT image data 302. For example, in a situation where the gantry is tilted during a CT scan, data corresponding to each slice of the non-contrast CT image data 302 may be translated according to the gantry tilt to correct for an offset of the image origin generated based on the tilt of the gantry. Additionally, one or more operations may be performed to generate slices of the non-contrast CT image data 302 having a consistent thickness. To illustrate, in various examples, the non-contrast CT image data 302 may include slices having different thicknesses due to slices at the base of the skull being thicker than slices corresponding to brain tissue to minimize partial volume effects and streak artifacts. In one or more examples, the one or more pre-processing operations that generate slices with a consistent thickness may include a first group of slices having a first thickness and a second group of slices having a second thickness that is thicker than the first thickness. In one or more illustrative examples, the first thickness can be 1 mm, 1.25 mm, 1.5 mm, 1.75 mm, 2 mm, 2.25 mm, 2.5 mm, 2.75 mm, 3 mm, 3.25 mm, or 3.5 mm, and the second thickness can be 4 mm, 4.5 mm, 5 mm, 5.5 mm, 6 mm, 6.5 mm, or 7 mm. The slices included in the first group can be resampled in the axial direction to correspond to the slice spacing of the second group. Furthermore, in situations where the slice thickness is less than a threshold thickness, some slices can be merged. In one or more illustrative examples, the threshold thickness for merging can be 0.5 mm, 0.75 mm, 1 mm, 1.25 mm, 1.5 mm, 2 mm, or 2.5 mm. Merging slices with a thickness less than the threshold thickness can improve the signal-to-noise ratio of individual slices. The number of slices to be merged can be based on a target spacing metric between slices. In various examples, the target spacing metric can be 2 mm, 2.5 mm, 3 mm, 3.5 mm, 4 mm, 4.5 mm, or 5 mm.

[0064] In operation 306, a contralateral mirroring operation may be performed. The contralateral mirroring operation may include an operation to determine a portion of the non-contrast CT image data 302 to be used to analyze the non-contrast CT image data 302 for low density. In this manner, a subset of the non-contrast CT image data 302 is subjected to low density analysis to provide an accurate rate of low density of voxels of the non-contrast CT image data 302 and reduce processing time for performing the low density analysis. In one or more examples, the non-contrast CT image data 302 may be separated into soft tissue, ventricles, and cerebrospinal fluid by performing a coregistration process of the non-contrast CT image data 302 with the atlas utilized in the perfusion image processing system 106.

[0065] Additionally, a median intensity value of voxels included in the non-contrast CT image data 302 may be identified. The median intensity value may be a threshold intensity value used to identify voxels of the non-contrast CT image data 302 corresponding to cerebrospinal fluid. In various examples, voxels having intensity values ​​between 0 HU and 60 HU are utilized to identify the median intensity value. The median intensity value of the voxels of the CT image data is used to identify at least one of a location of cerebrospinal fluid, a location of intracerebral hemorrhage, or a volume effect associated with bone. In one or more illustrative examples, the threshold intensity value identifying cerebrospinal fluid may be between 6 HU and 15 HU, between 8 HU and 12 HU, or between 10 HU and 15 HU. In one or more additional illustrative examples, the threshold intensity value for determining intracerebral hemorrhage and volume effect may be between 25 HU and 50 HU, between 35 HU and 45 HU, or between 40 HU and 50 HU. In various examples, the cerebrospinal fluid, intracerebral hemorrhage, and / or volume effect thresholds can be modified with an offset based on one or more thresholds and different median intensity values. The threshold intensity value for determining cerebrospinal fluid can be used in conjunction with an active contour model to identify pixels in the slice that correspond to cerebrospinal fluid. In these circumstances, a cerebrospinal fluid mask can be generated. Portions of the non-contrast CT image data 302 that do not correspond to cerebrospinal fluid can be used in the low-density analysis.

[0066] Additionally, spatial correlations of soft tissue voxels (e.g., voxels not corresponding to ventricles and cerebrospinal fluid) in a first hemisphere of the individual's brain associated with soft tissue voxels in a second hemisphere of the individual's brain can be identified. In various examples, the non-contrast CT image data 302 can be mirrored to the sagittal plane and the mirrored data can be registered with the non-contrast CT image data 302, followed by one or more non-rigid registration processes. The ventricles and cerebrospinal fluid can be identified in both the mirrored data and the non-contrast CT image data 302. The mirrored data and the non-contrast CT image data 302 can then be sliced ​​pre-processed by median filtering and Gaussian filtering. In one or more illustrative examples, the one or more median filtering operations may be performed using a kernel radius of about 0.5 mm to about 2.5 mm, about 0.8 mm to about 1.6 mm, or about 1.2 mm to about 2 mm, and the one or more Gaussian filtering operations may be performed using a kernel radius of about 0.2 mm to about 2 mm, about 0.4 mm to about 1.2 mm, or about 0.5 mm to about 1.5 mm.

[0067] Identifying the hypo-density of brain tissue corresponding to the non-contrast CT image data 302 can be based on input data including the non-contrast CT image data 302 and one or more additional parameters. The one or more additional parameters can include a minimum intensity value difference between healthy tissue and hypo-dense tissue in opposing hemispheres of the brain. In operation 308, a hypo-density analysis can be performed. The hypo-density analysis can include identifying intensity values ​​of voxels corresponding to regions of the brain in a first hemisphere that can be analyzed with respect to intensity values ​​of voxels in a contralateral region in a second hemisphere of the brain.

[0068] In operation 310, several thresholding operations can be performed to identify a hypodense measure of a voxel undergoing hypodense analysis. In situations where the difference between the intensity values ​​of the contralateral hemisphere is greater than a threshold, the voxel can be determined to be associated with hypodense tissue. In one or more illustrative examples, the threshold difference in intensity values ​​for identifying hypodense regions can be at least 2 HU, at least 3 HU, at least 4 HU, at least 5 HU, at least 6 HU, at least 7 HU, at least 8 HU, at least 9 HU, or at least 10 HU. In one or more additional illustrative examples, the threshold difference in intensity values ​​for identifying hypodense regions can be 22 HU or less, 21 HU or less, 20 HU or less, 19 HU or less, 18 HU or less, 17 HU or less, 16 HU or less, 15 HU or less, 14 HU or less, 13 HU or less, or 12 HU or less. In one or more further illustrative examples, the threshold difference in intensity values ​​for identifying a low density region may include several intensity value ranges selected from the intensity values ​​listed above, such as 2 HU to 22 HU, 3 HU to 20 HU, 4 HU to 18 HU, or 5 HU to 12 HU.

[0069] The low density threshold can be used to generate one or more regions of interest indicative of damaged brain tissue and / or brain tissue having a probability of damage that is at least the threshold probability in operation 312. In one or more examples, the damaged brain tissue can be infarcted.

[0070] The low density region of interest 314 can be included in one or more Insight Segmentation and Registration Toolkit (ITK) images. The one or more ITK images can correspond to the initial non-contrast CT image data 302 modified based on the gantry tilt of the CT contrast device used to generate the initial CT image data. The one or more ITK images can also correspond to the initial non-contrast CT image data 302 modified based on variable slice thickness in the initial CT image data and resampled to merge modified slices having consistent thickness. The one or more ITK images can include one or more contrast maps with two or more labels. For example, the one or more ITK images can include a contrast map that labels subacute brain tissue as "1" and background and / or non-subacute brain tissue as "0". Additionally, the one or more ITK images can include a contrast map with labels indicating differences from one or more Hounsfield unit thresholds. In situations where one or more machine learning techniques are implemented, a threshold may not be used to determine infarction using the non-contrast CT image data.

[0071] The results of the low-density analysis can be used for mismatch analysis of the perfusion parameter data 236. The CT perfusion mismatch analysis can access the low-density analysis ITK file via a reference included in one or more JSON files. The non-contrast CT image data 302 can be registered in the coordinate space of the CT perfusion image data 202 together with the pre-contrast baseline image 216. The low-density analysis can be integrated with the mismatch output of the perfusion image processing system 106 to visualize the regions detected in the low-density analysis and identify slices having low-density regions. In a situation where the non-contrast CT image data 302 is not acquired from a separate scan and is based on the pre-contrast baseline image, a registration process between the non-contrast CT image data 302 and the pre-contrast baseline image 216 may not be performed. In a situation where the non-contrast CT image data 302 is not acquired from a separate scan and is based on the pre-contrast baseline image, the CT perfusion image data 202 can be corrected for patient motion to avoid misregistration. The time series of images captured to generate CT perfusion image data 202, which includes images captured before and after contrast agent enters the individual's brain, may be motion corrected.

[0072] The volume of the perfusion region of interest corresponding to brain tissue that is damaged and / or has a threshold probability of damage, e.g., 25% of normal CBF value, 30% of normal CBF value, 35% of normal CBF value, or 40% of normal CBF value, and the volume of the hypodense region in the NCCT (or pre-contrast baseline CT scan) can be determined by counting the voxels contained in the labeled region of the labeled image and multiplying the number of voxels by the volume of a single voxel. In situations where multiple regions overlap, indicating different degrees of hypodense, different amounts of tissue damage, and / or different probabilities of tissue damage being present, the larger region can include the volume of the smaller region within the larger region. For example, in a situation where a first region includes a second region, the volume of the first region can include the volume of the voxels corresponding to the first region and the second region.

[0073] A mismatch ratio and a volume can be identified that indicate the difference between the volumes of regional perfusion images for different perfusion parameters. The mismatch ratio and the mismatch volume can correspond to the amount of brain tissue that may be at risk of being damaged if blood flow to the brain tissue is subsequently impeded. In one or more illustrative examples, for CT perfusion, the mismatch ratio is

[0074]

number

[0075] and the mismatch volume is Mismatch volume = ROIvolume(Tmax>th1)-ROI(rCBF <th2) where th is the threshold for a given region.

[0076] In one or more examples, T max The th1 value of the relative cerebral blood flow (rCBF) can be at least 4 seconds, at least 6 seconds, at least 8 seconds, or at least 10 seconds. Additionally, the th2 value of the relative cerebral blood flow (rCBF) can be less than 20%, less than 30%, less than 34%, or less than 38%. In various examples, the th value of the relative cerebral blood volume (rCBV) can be less than 34%, less than 38%, or less than 42%. In one or more illustrative examples, the thresholds th1 and th2 can be arbitrarily selected at the discretion of a medical professional.

[0077] Although the illustrative example of Figure 3 relates to one or more example implementations of performing low-density analysis on non-contrast CT image data, different implementations can be used to identify low-density regions of interest. For example, machine learning techniques can be implemented to identify low-density regions of interest, as described with respect to Figure 1 and low-density analysis system 128.

[0078] 4 is a diagrammatic representation of an example architecture 400 for aggregating information generated from different contrast modalities to determine an estimate of the extent of irreversible damage to brain tissue, according to one or more example embodiments. The architecture 400 may include an output imaging system 144. In operation 402, the output imaging system 144 may acquire perfusion parameter data 236 and low density regions of interest 314 as well as the CT perfusion image data 202. In operation 404, the output imaging system 144 may remove portions of the CT perfusion image data 202 corresponding to the ventricles and cerebrospinal fluid. In operation 406, the output imaging system 144 may also register the low density regions of interest derived from the non-contrast CT image data 302 with the regions of interest identified by the perfusion image processing system 106 based on the CT perfusion image data 202. In various examples, the output image system 144 may generate one or more summary images 146 based on the regions of interest derived from the low-density analysis performed by the low-density analysis system 128 and the regions of interest derived from the analysis of the CT perfusion image data 202 by the perfusion image processing system 106. To illustrate, the one or more summary images 146 may include a first overlay indicative of the regions of interest derived from the CT perfusion image data 202 by the perfusion image processing system 106 and a second overlay indicative of the regions of interest derived from the low-density analysis performed by the low-density analysis system 128. The first and second overlays may be displayed on one or more slices of the CT perfusion image data 202.

[0079] 4 relates to one or more example implementations for generating an overlay showing regions of interest in the brain of a potentially injured individual, additional implementations may be implemented. Different portions of the CT perfusion image data 202 other than those corresponding to the ventricles and cerebrospinal fluid may be removed, i.e., the portions of the CT perfusion image data 202 corresponding to the ventricles and cerebrospinal fluid may be left. Additionally, perfusion regions of interest and low density regions of interest may be identified and combined prior to application to the output imaging system 144.

[0080] 5 is a diagrammatic representation of an example architecture for generating image data indicative of an amount of damage to brain tissue based on the timing of onset of a biological condition corresponding to one or more blood vessels in an individual's brain, according to one or more example embodiments. The image processing system 102 may acquire imaging data 502. The imaging data 502 may be acquired from one or more imaging data sources. The one or more imaging data sources may perform one or more imaging techniques to generate the imaging data 502.

[0081] In one or more examples, the imaging data 502 may include first imaging data generated by performing a first imaging technique. The first imaging technique may be a perfusion-based imaging technique. In one or more illustrative examples, the first imaging technique may be a computed tomography (CT)-based imaging technique. In one or more additional illustrative examples, the first imaging technique may be a CT perfusion-based imaging technique. In one or more additional illustrative examples, the first imaging technique may be a CT angiography-based imaging technique.

[0082] Additionally, the imaging data 502 may include second imaging data that may be generated by a second imaging technique. The second imaging technique may be a non-perfusion based imaging technique. In various examples, the second imaging technique may be a non-contrast agent based imaging technique. In one or more illustrative examples, the second imaging technique may implement one or more non-contrast CT imaging techniques.

[0083] In various examples, the imaging data 502 may correspond to one or more images of the individual's brain captured by a CT imaging device over a period of time such that a series of images of the brain may be captured successively over the period of time. Each image in the series of images may be referred to herein as a "slice." The imaging data may include an image of the individual's brain showing the brain before a contrast agent is delivered, and an additional image of the individual's brain showing the presence of the contrast agent in one or more regions of the brain.

[0084] The imaging data 502 may indicate intensity values ​​of voxels of a captured image of the individual's brain. The intensity values ​​may be in Hounsfield units. In one or more examples, the intensity values ​​of voxels corresponding to regions of the brain where contrast agent is present may be greater than the intensity values ​​of voxels corresponding to regions of the brain where contrast agent is not present. In various examples, the intensity values ​​of voxels included in the imaging data 502 may indicate an amount of contrast agent present in a region of the individual's brain. In one or more illustrative examples, the imaging data 502 may be formatted according to the Digital Imaging and Communications in Medicine (DICOM) standard. In various examples, the imaging data 502 may include at least one of the first imaging data 108 of FIG. 1 or the second imaging data 130 of FIG. 1. The image processing system 102 may also acquire onset timing data 504. The onset timing data 504 may indicate an amount of time that has elapsed since a biological condition became present in the individual's brain. In one or more examples, the onset timing data 504 can indicate a time period from minutes to hours. For example, the onset timing data 504 can be 1 minute to 48 hours, 10 minutes to 40 hours, 30 minutes to 36 hours, 1 hour to 30 hours, 2 hours to 24 hours, 30 minutes to 12 hours, 1 hour to 12 hours, 15 minutes to 6 hours, or 15 minutes to 4 hours. In one or more illustrative examples, the onset timing data 504 can indicate an amount of time that has elapsed since a biological condition was detected in the individual's brain. In one or more additional illustrative examples, the onset timing data 504 can indicate an amount of time that has elapsed since one or more symptoms of the biological condition were detected in the individual. In various examples, the biological condition can reduce blood flow to one or more regions of the individual's brain. In at least some examples, the biological condition can be associated with one or more blood vessels in the individual's brain. For illustrative purposes, the biological condition can include a stroke.

[0085] In one or more examples, the onset timing data 504 can be captured via one or more input devices of one or more computing devices. For example, the onset timing data 504 can be captured via one or more user interfaces displayed by the computing device. Additionally, the onset timing data 504 can be captured via one or more audio input devices, such as one or more microphones, of the computing device. In one or more illustrative examples, the onset timing data 504 can be captured by a computing device of a medical professional. In one or more further illustrative examples, the onset timing data 504 can be captured by a computing device of an individual in whom a physiological condition exists. In various examples, the onset timing data 504 can be determined based on an analysis of the imaging data 502.

[0086] The image processing system 102 may include a perfusion image processing system 106. The perfusion image processing system 106 may analyze the contrast data 502 to generate one or more perfusion parameters 116 for one or more voxels of the contrast data 502. For example, the perfusion image processing system 106 may analyze intensity values ​​of voxels of the contrast data 502 for one or more regions of the brain to identify one or more perfusion parameters 116. The one or more perfusion parameters 116 may be indicative of blood flow through one or more regions of the individual's brain. In one or more illustrative examples, the perfusion parameters 116 may be a measure of cerebral blood flow (CBF), a measure of cerebral blood volume (CBV), a measure of mean tracer transit time (MTT), or a T corresponding to the amount of contrast agent that entered and subsequently remains within the voxel. max It may include at least one of the following:

[0087] The perfusion imaging system 106 can analyze the one or more perfusion parameters 116 in conjunction with the imaging data 502 to identify one or more regions of the brain having damaged tissue due to an obstruction in blood supply to the one or more regions. For example, the perfusion imaging system 106 can analyze the one or more perfusion parameters 116 and intensity values ​​of voxels corresponding to blood vessels in the individual's brain to identify the degree of obstruction in blood flow to one or more regions of the individual's brain. In one or more examples, the one or more perfusion parameters 116 can be analyzed to identify an amount of time of obstruction in blood flow to a region of the individual's brain. In one or more illustrative examples, the perfusion imaging system 106 can analyze the imaging data 502 to identify at least a portion of the onset timing data 504. Based on the degree of obstruction in blood flow to one or more regions of the individual's brain, the perfusion imaging system 106 can identify one or more regions of interest having damaged tissue that have a probability of at least a threshold probability.

[0088] In one or more examples, the perfusion imaging system 106 can utilize the threshold data 506 to identify one or more regions of the individual's brain in which damaged tissue may be present. In one or more illustrative examples, the threshold data 506 can indicate one or more threshold levels of the perfusion parameter 116. The threshold level can indicate a value of the perfusion parameter 116 that corresponds to damaged tissue in the region of the individual's brain. In various examples, the threshold level can indicate a value of the perfusion parameter 116 that corresponds at least to a minimum probability that tissue in the region of the individual's brain has been damaged. In at least some examples, the threshold data 506 can indicate a threshold level of one or more parameters of cerebral blood flow, cerebral blood volume, MTT, or T max The method may include one or more thresholds for at least one of:

[0089] The threshold data 506 utilized by the perfusion imaging system 106 can be based on the onset imaging data 504. For example, as the time period from onset of the biological condition impeding blood flow to one or more regions of the individual's brain changes, the one or more thresholds used by the perfusion imaging system 106 can also change. In one or more examples, as the time period from onset of the biological condition increases, the one or more thresholds for one or more perfusion parameters can also increase. In various examples, the one or more first thresholds can be used by the perfusion imaging system 106 to determine whether damage has occurred in one or more regions of the individual's brain during a first time period from onset of the biological condition that impeded blood flow to one or more regions of the individual's brain to a current time. Additionally, the one or more second thresholds can be used by the perfusion imaging system 106 to determine whether damage has occurred in one or more regions of the individual's brain during a second time period from onset of the biological condition.

[0090] In one or more illustrative examples, the first time period from onset of a physiological condition impeding blood flow to one or more regions of the individual's brain can be 30 minutes or less, 1 hour or less, 2 hours or less, 4 hours or less, or 6 hours or less. In these circumstances, illustrative examples of threshold values ​​of the perfusion parameter 116 associated with the first time period can be a cerebral blood flow value that is less than 10% of the standard cerebral blood flow value, less than 15% of the standard cerebral blood flow value, less than 20% of the standard cerebral blood flow value, or less than 25% of the standard cerebral blood flow value, or less than 30% of the standard cerebral blood flow value. In one or more examples, the standard cerebral blood flow value can be identified using information from the individual from whom the imaging data 502 was acquired. The standard cerebral blood flow value can also be identified using information from several individuals. In various examples, the standard cerebral blood flow value can be an average cerebral blood flow value based on an analysis of cerebral blood flow values ​​of several individuals. In at least some examples, the standard cerebral blood flow value can be based on the age of the individual. In one or more additional examples, the standard cerebral blood flow value can be based on blood value values ​​of one or more blood vessels contained in one or more regions of the individual's brain. In one or more further examples, the standard cerebral blood flow value can be based on blood vessel diameters.

[0091] In one or more additional illustrative examples, the second time period from onset of the biological condition can be longer than the threshold value associated with the first time period. For illustrative purposes, the second time period from onset of the biological condition that impedes blood flow to one or more regions of the individual's brain can be at least 30 minutes, at least 1 hour, at least 2 hours, at least 4 hours, or at least 6 hours. In one or more examples, the second time period can also have an upper limit of 48 hours or less, 36 hours or less, 24 hours or less, 18 hours or less, 12 hours or less, 10 hours or less, 8 hours or less, 6 hours or less, or 4 hours or less, depending on the lower limit of the second time period. In various examples, the threshold value of the blood flow parameter associated with the second time period can be a value greater than the threshold value of the blood flow parameter associated with the first time period. For purposes of illustration, illustrative examples of threshold values ​​for the perfusion parameter 116 in the second time period can be cerebral blood flow values ​​that are less than 25% of the standard blood flow value, less than 30% of the standard blood flow value, less than 35% of the standard blood flow value, less than 40% of the standard blood flow value, less than 45% of the standard blood flow value, or less than 50% of the standard blood flow value. In various examples, the threshold value for the blood flow parameter associated with the second time period can be a value greater than the threshold value for the blood flow parameter associated with the first time period. In yet other examples, the threshold value for the blood flow parameter associated with the second time period can be a value less than the threshold value for the blood flow parameter associated with the first time period. If the blood flow parameter is cerebral blood flow, cerebral blood volume, MTT, or T max In a situation that includes at least one of: the first time period threshold can be less than the second time period threshold because the longer the time since onset of the biological condition that impedes blood flow to a region of the individual's brain, the higher the likelihood of reperfusion occurring in that region. Thus, blood flow to a region of the individual's brain can be relatively low in the early stages of the biological condition that caused the impediment of blood flow to the region, such as a stroke, while in later stages of the biological condition, blood flow may be increased from blood vessels other than those once or currently used to restrict blood flow.

[0092] The perfusion image processing system 106 can generate first system output data 508. The first system output data 508 can include one or more values ​​of the perfusion parameters 116. The first system output data 508 can also be indicative of a probability of damage to one or more regions of the individual's brain. In addition, the first system output data 508 can include overlay contrast data corresponding to an overlay indicative of regions of the individual's brain where tissue damage is present or has a probability of presence of at least a threshold probability. In various examples, the perfusion image processing system 106 can perform one or more computational models to identify at least one of the one or more values ​​of the perfusion parameters 116 or the first system output data 508. In one or more illustrative examples, the perfusion image processing system 106 can perform one or more machine learning techniques to generate at least one of the one or more values ​​of the perfusion parameters 116 or the first system output data 508.

[0093] The image processing system 102 may also include a low-density analysis system 128. The low-density analysis system 128 may analyze intensity values ​​of voxels in the contrast data 502 for one or more regions of the individual's brain to identify one or more measures of low density 510. The one or more measures of low density may be expressed as Hounsfield density values. In one or more examples, the one or more measures of low density 510 may be identified to identify one or more regions of the individual's brain in which tissue damage is present. In various examples, the low-density analysis system 128 may analyze intensity values ​​of voxels in the contrast data 502 to identify regions of the individual's brain in which the probability of damage is at least a threshold probability. In one or more additional examples, the low-density analysis system 128 may analyze voxels in different hemispheres of the individual's brain to identify one or more regions of the brain containing low-density tissue. In one or more illustrative examples, low-density analysis system 128 may also implement one or more machine learning techniques to identify regions of low-density tissue in the individual's brain.

[0094] The onset timing data 504 may be analyzed by the low-density analysis system 128 to identify a likelihood that tissue damage is present in a region of the individual's brain. In one or more examples, the likelihood that tissue damage is present in a region of the individual's brain may increase with increasing time from onset of a biological condition related to one or more blood vessels of the individual's brain. In one or more examples, at least one of the accuracy or precision of the results generated by the low-density analysis system 128 may increase with increasing time from onset of the biological condition. That is, low-density tissue in the individual's brain may be more reliably detected with increasing time from onset of the biological condition. Thus, the low-density analysis system 128 may identify one or more regions of the individual's brain in which tissue damage is present and a likelihood score corresponding to the likelihood of the presence of tissue damage. In one or more illustrative examples, the low-density analysis system 128 may generate second system output data 512 indicative of one or more regions of the individual's brain in which tissue damage is present. The hypodensity analysis system 128 may also generate a likelihood that a measure of hypodensity 510 generated using the imaging data 502 indicates tissue damage in one or more regions.

[0095] One or more computational models may be implemented to determine a likelihood for a measure of low concentration 510 indicating that tissue damage is present in the individual's brain. In various examples, the low concentration analysis system 128 may implement one or more machine learning algorithms to determine a likelihood for a measure of low concentration 510 indicating that tissue damage is present in the individual's brain. For example, one or more machine learning classification algorithms may be implemented to determine a likelihood for a measure of low concentration 510 indicating that tissue damage is present in the individual's brain. In one or more illustrative examples, one or more random forest algorithms may be implemented to determine a likelihood for a measure of low concentration 510 indicating that tissue damage is present in the individual's brain. In one or more further illustrative examples, the likelihood that tissue damage exists for a measure of low concentration 510 indicating that tissue damage is present in the individual's brain may be based on a branch number of the output of one or more random forest algorithms indicating tissue damage to a region of the individual's brain.

[0096] In one or more illustrative examples, the low-density analysis system 128 may analyze imaging data 502 of the individual's brain during a first time period from onset of the biological condition and identify a first measure of low density 510. The low-density analysis system 128 may identify one or more regions of the individual's brain in which tissue damage is present based on the first measure of low density 510. The low-density analysis system 128 may also identify a likelihood of the first measure of low density 510 indicating that tissue damage is present in one or more regions based on the onset timing data. In situations where the likelihood is less than a threshold likelihood, the low-density analysis system 128 may indicate in the second system output data 512 that tissue damage is not present in the one or more regions. Further, in situations where the likelihood is at least the threshold likelihood, the low-density analysis system 128 may indicate in the second system output data 512 that tissue damage is present in the one or more regions.

[0097] The low-density analysis system 128 may also analyze the individual's brain imaging data 502 during a second time period from onset of the biological condition to identify a second measure of low density. The second time period may follow the first time period. In various examples, the low-density analysis system 128 may identify an additional likelihood based on the onset timing data 504 that the second measure of low density 510 indicates that tissue damage is present in at least one or more regions. In situations where the additional likelihood is less than a threshold likelihood, the low-density analysis system 128 may indicate in the second system output data 512 that tissue damage is not present in the one or more regions. Furthermore, in situations where the additional likelihood is at least a threshold likelihood, the low-density analysis system 128 may indicate in the second system output data 512 that tissue damage is present in the one or more regions. In one or more illustrative examples, the additional likelihood identified based on the second measure of low density may be greater than the initial likelihood identified based on the first measure of low density.

[0098] The image processing system 102 can include an output image system 144. The output image system 144 can analyze the first system output data 508 and the second system output data 512 to identify one or more output images 514. In one or more examples, the output image system 144 can execute one or more computational models 516 to generate the one or more output images 514. In one or more examples, the output image system 144 can execute the one or more computational models 516 on the first system output data 508 to identify a first region of the individual's brain where tissue damage is present. Additionally, the output image system 144 can execute the one or more computational models 516 on the second system output data 512 to identify a second region of the individual's brain where tissue damage is present. In various examples, the second region can overlap at least a portion of the first region. Additionally, the second region can include a portion of the individual's brain not included in the first region.

[0099] In one or more illustrative examples, the output image system 144 can generate an aggregate image based on the first region and the second region. The aggregate image can include a first overlay corresponding to the first region and a second overlay corresponding to the second region. In one or more embodiments, the one or more computational models 516 can be implemented based on a likelihood included in the second system output data 512 associated with the second region. In one or more circumstances, the one or more computational models 516 can determine that the second region should not be included in the aggregate image based on the likelihood value. In one or more additional examples, the one or more computational models 516 can determine that a portion of the second region should not be included in the aggregate image based on the likelihood value. In one or more further examples, the second region or at least a portion of the second region may not be included in the aggregate image based on an instruction included in the second system output data 512 to exclude at least a portion of the second region from the aggregate image. The indication may be determined based on a likelihood value generated by the low concentration analysis system 128 associated with a measure of low concentration 510 corresponding to damaged tissue in the second region. The one or more output images may include at least one of at least a portion of the first region and the second region overlaid on an image included in the contrast data 502. For example, the one or more output images 514 may include one or more CT images of the individual's brain in the presence of contrast agent along with one or more overlays showing one or more regions of the individual's brain where damaged tissue is present. In one or more additional examples, the one or more output images 514 may include one or more non-contrast CT images of the individual's brain where damaged tissue is present. In one or more further examples, the one or more output images 514 may include at least a first CT image of the individual's brain in the presence of contrast agent and at least a second non-contrast CT image of the individual's brain along with one or more overlays showing regions of the individual's brain where damaged tissue is present or where the probability of damaged tissue being present is at least a threshold probability.In at least some examples, the CT images included in the one or more output images 514 may have one or more overlays that are different from the one or more overlays for the non-contrast CT images included in the one or more output images 514. In one or more situations, the one or more overlays of the perfusion-based CT images may correspond to one or more regions identified by the perfusion image processing system 106 as having tissue damage. The one or more overlays of the non-contrast CT images may correspond to one or more regions identified by the low-concentration analysis system 128 as having tissue damage.

[0100] 6-8 show a flowchart of a process for generating an image showing damaged brain tissue. The process may be embodied in computer readable instructions executed by one or more processors such that the operations of the process may be performed in part or in whole by the functional components of the image processing system 102. Thus, the process described below is, in some circumstances, by way of example with reference to FIGS. 6-8. However, in other embodiments, at least some of the operations of the process described with respect to FIGS. 6-8 may be deployed in a variety of other hardware configurations. Thus, the process described with respect to FIGS. 6-8 is not intended to be limited to the image processing system 102 and may be performed in whole or in part by one or more additional components. Although the depicted flowchart may depict operations as a sequential process, many of the operations may be performed in parallel or simultaneously. Additionally, the order of the operations may be reconfigured. A process ends when the operations are completed. A process may correspond to a method, a procedure, an algorithm, or the like. The method operations may be performed in whole or in part, may be performed in conjunction with some or all of the other method operations, and may be performed by any number of different systems, such as the systems described herein, or any part of a system, such as a processor included in any system.

[0101] FIG. 6 is a flow chart illustrating example operations of a process 600 for identifying an aggregate image based on image data generated by different imaging modalities and generating an overlay of the aggregate image indicative of a potential amount of damage to brain tissue, according to one or more example embodiments. In operation 602, the process 600 can include acquiring first imaging data of the individual's brain. In one or more examples, the first imaging data can be generated by a first imaging technique. The first imaging technique can relate to perfusion characteristics of fluids in the individual's brain. In various examples, the first imaging technique can generate an image of the individual's brain indicative of the presence of blood in the blood vessels of the individual's brain. In one or more illustrative examples, the first imaging technique can include a computed tomography perfusion imaging technique. In one or more additional examples, the first imaging technique can include a computed tomography angiography technique.

[0102] Additionally, the process 600 may acquire second imaging data at operation 604. In one or more examples, the second imaging data may be generated by a second imaging technique that is a non-perfusion based imaging technique. In one or more illustrative examples, the second imaging technique may include a non-contrast computed tomography imaging technique. In one or more additional illustrative examples, the second imaging technique may include a diffusion based magnetic resonance imaging technique. The second imaging data may include several second slices captured consecutively over a second time period. Each second slice may include a cross-section of a portion of the individual's brain. In various examples, the first imaging data and the second imaging data may be generated by different imaging devices. The first imaging data and the second imaging data may also be generated using the same imaging device. Furthermore, in one or more examples, the first imaging data and the second imaging data may be generated during different time periods. In at least some examples, the different time periods may not overlap.

[0103] In various examples, the first contrast imaging data and the second contrast imaging data can be captured using the same contrast modality at different times. For example, the first contrast imaging data can be captured using a contrast technique during a time period when contrast agent is present in the individual's brain. Additionally, the second contrast imaging data can be captured using a contrast technique during a time period when contrast agent is not present in the individual's brain. In one or more illustrative examples, the contrast technique can include a computed tomography perfusion imaging technique that captures images of the brain before contrast agent is delivered to the individual and captures images of the brain of the individual after contrast agent is delivered to the individual. In one or more illustrative examples, the first contrast imaging data can include a number of first slices captured consecutively over a first time period, and the second contrast imaging data can include a number of second slices captured consecutively over a second time period. Each of the first slices and the second slices can include a cross-section of a portion of the individual's brain.

[0104] The process 600 may also include, at operation 606, determining one or more perfusion parameters based on a first intensity value of a first voxel of the first contrast imaging data. The one or more perfusion parameters may be indicative of blood flow in the brain of the individual. The one or more perfusion parameters may be generated for each voxel included in the first contrast imaging data. In one or more examples, a number of voxels for which the one or more perfusion parameters are determined may correspond to brain tissue. In various examples, the first contrast imaging data may be analyzed to identify voxels included in the first contrast imaging data that correspond to brain tissue. In one or more illustrative examples, voxels corresponding to brain tissue in the first contrast imaging data may be identified by analyzing intensity values ​​of the voxels. Furthermore, voxels corresponding to brain tissue in the first contrast imaging data may be identified based on an atlas indicative of several regions of the brain, with at least a portion of the first contrast imaging data being aligned with the atlas.

[0105] The process 600 may also include, at operation 608, determining that damage to tissue has occurred in one or more first regions of the individual's brain based on the one or more perfusion parameters. In one or more examples, the one or more perfusion parameters of the one or more regions of interest may be analyzed relative to one or more thresholds. In various examples, the one or more thresholds may correspond to values ​​of the one or more perfusion parameters previously determined to correspond to damage to the brain tissue. In various examples, the one or more thresholds may correspond to a probability that damage is present in the brain tissue, such as at least a 50% probability that damage is present in the brain tissue, at least a 60% probability that damage is present in the brain tissue, at least a 70% probability that damage is present in the brain tissue, at least an 80% probability that damage is present in the brain tissue, at least a 90% probability that damage is present in the brain tissue, at least a 95% probability that damage is present in the brain tissue, or at least a 99% probability that damage is present in the brain tissue. In one or more illustrative examples, the one or more perfusion parameters may include cerebral blood flow. The measure of cerebral blood flow can be determined by analyzing intensity values ​​of voxels included in the first imaging data. The measure of cerebral blood flow of voxels corresponding to one or more regions of the individual's brain can be analyzed with respect to the one or more measures of cerebral blood flow. In situations where voxels of the region of the individual's brain are associated with measures of cerebral blood flow that correspond to one or more threshold values, a determination can be made that a probability that tissue damage is present in the region of the individual's brain is at least a threshold probability.

[0106] Additionally, the process 600 may include, at operation 610, identifying one or more second regions of the brain corresponding to one or more portions of brain tissue having a low density measure corresponding to a low density threshold measure. The low density measure may be identified based on intensity values ​​of voxels included in the one or more second regions. The low density threshold measure may correspond to a low density measure associated with a probability of brain tissue damage based on previously analyzed intensity values ​​of voxels included in images generated using a non-contrast computed tomography imaging technique.

[0107] Further, in operation 612, the process 600 may include generating an aggregate image including a first overlay showing the one or more first regions and a second overlay including the one or more second regions. In one or more examples, the aggregate image may correspond to an individual slice of a plurality of slices included in the first contrast data. In these circumstances, the aggregate image may include a slice of a plurality of slices captured by the perfusion-based computed tomography contrast technique. In various examples, the aggregate image may be included in a user interface. In one or more illustrative examples, user interface data corresponding to the user interface may be generated. The user interface may also display a plurality of slices included in the first contrast data. At least some of the individual slices may include each of the first overlays corresponding to one or more regions of brain tissue represented by the individual slices that have been damaged or have a damage probability of at least a threshold probability according to values ​​of the one or more perfusion parameters. Further, at least some of the individual slices may each include a second overlay corresponding to one or more regions of brain tissue represented by the individual slice that are damaged or have a damage probability that is at least a threshold probability according to a low density value in the one or more regions.

[0108] 7 is a flow chart illustrating example operations of a process 700 for identifying perfusion parameters and hypodensity measures for brain tissue to identify the extent of irreversible brain damage, according to one or more example embodiments. At operation 702, the process 700 may include obtaining first imaging data of the individual's brain generated by a first imaging technique corresponding to perfusion characteristics in the individual's brain. In one or more illustrative examples, the first imaging technique may include a computed tomography perfusion imaging technique.

[0109] Additionally, process 700 may include, at operation 704, acquiring second imaging data generated by a second imaging technique that is a non-perfusion based imaging technique. In one or more illustrative examples, the second imaging technique may include a non-contrast computed tomography imaging technique. The second imaging data may include a number of second slices captured consecutively over a second time period. Each second slice may include a cross-section of a portion of the individual's brain.

[0110] The process 700 may also include, at operation 706, determining that damage to tissue has occurred in association with one or more first regions of the individual's brain based on the first contrast data. In one or more examples, prior to identifying the one or more first regions, the first contrast data may be corrected for motion of the individual during acquisition of the first contrast data. For example, the first contrast data includes several first slices acquired consecutively over a period of time. In response to the motion of the individual during acquisition of the first contrast data, one or more of the first slices may be identified that are offset from the reference image by at least a threshold amount. In various examples, a registration process may be performed to align the one or more first slices with the reference image. Based on the registration process, one or more motion correction parameters may be determined. The one or more motion correction parameters may include at least one of one or more rotation parameters or one or more translation parameters that position voxels of the one or more first slices within a threshold distance of corresponding voxels of the reference image.

[0111] In one or more additional examples, the first imaging data can be subjected to a time correction process. The time correction process can include determining that a portion of a first slice is acquired at a time interval that is different from an additional time interval between acquisitions of additional portions of the first slice. Further, the time correction process can include changing the time interval and the additional time interval to a common time interval such that the first slices are arranged with a common time interval between successive first slices.

[0112] In various examples, identifying one or more first regions having tissue damage can include identifying a number of perfusion parameters. In one or more examples, the perfusion parameters can be identified based on a baseline intensity value of a voxel before the contrast agent arrives at the voxel. To identify the baseline intensity value, a portion of the first contrast imaging data captured before the contrast agent arrives at the voxel can be identified. For example, an intensity value change of at least a portion of the voxels included in the first contrast imaging data over a time period can be identified, and based on the intensity value change, a time when the contrast agent entered a portion of the individual's brain can be identified. That is, the intensity value of the voxel before the arrival of the contrast agent can be lower than the intensity value of the voxel in which the contrast agent is present. After identifying the arrival time of the contrast agent to the voxel included in the respective slice of the first contrast imaging data, an image of the individual's brain image without the contrast agent can be generated based on the portion of the first contrast imaging data captured before the arrival time.

[0113] Additionally, identifying one or more first regions of the brain having damaged brain tissue may include performing a registration process between at least a portion of the first contrast-enhanced data and an anatomical template of the human head, and generating a deformation field as a result of the registration process. The deformation field may indicate the extent to which voxels included in the first contrast-enhanced data should be deformed in order to align with the anatomical template. After registration with the anatomical structure, various regions of the individual's brain may be identified with respect to the first contrast-enhanced data. For example, the deformation field may be applied to an atlas showing several regions of the human brain to generate a modified atlas corresponding to the individual's brain. In this manner, portions of the first contrast-enhanced data corresponding to regions of the brain that are not used to identify one or more perfusion parameters may be removed. For illustration, each region included in the modified atlas may be used to identify an individual region of the individual's brain, and one or more portions of the first contrast-enhanced data that do not correspond to brain tissue may be removed. As a result, modified first contrast-enhanced data corresponding to brain tissue is generated without other parts of the individual's body that may be captured in the first contrast-enhanced data, such as the skull or eyes.

[0114] The one or more perfusion parameters can be determined based on intensity values ​​of voxels associated with the individual's brain tissue, and the intensity values ​​can be indicated by a contrast agent concentration curve for each voxel. In one or more examples, a contrast agent concentration curve can be determined for each voxel in the modified first contrast data based on the intensity values ​​over time of the each voxel relative to one or more reference intensity values. In various examples, an intensity value of a voxel in the modified first contrast data that differs from at least one of the one or more reference values ​​by more than a threshold difference indicates that a contrast agent is present in the voxel.

[0115] In various examples, the arterial input function and the venous output function can be used to identify one or more perfusion parameters. In one or more examples, one or more first candidate regions of an arterial input function of the modified first contrast data including at least a first blood vessel having a threshold diameter can be identified based on one or more first locations of a first blood vessel having a threshold diameter in the modified atlas. The one or more first candidate regions can include at least a portion of a middle cerebral artery or at least a portion of an anterior cerebral artery. The arterial input function can be generated based on a first contrast agent concentration curve of a first voxel included in a first candidate region of the one or more first candidate regions. Furthermore, one or more second candidate regions of a venous output function of the modified first contrast data including at least a second blood vessel having a threshold diameter can be identified based on one or more second locations of a second blood vessel having a threshold diameter in the modified atlas. The one or more second candidate regions include at least a portion of a sagittal sinus region or at least a portion of a straight sinus region. The venous output function may be determined based on a second contrast agent concentration curve of second voxels included in a second candidate region of the one or more second candidate regions.

[0116] Additionally, determining the value of the perfusion parameter associated with the first contrast data may be determined based on a tissue residue function of voxels in one or more regions of the individual's brain, the tissue residue function may be generated by performing one or more deconvolution operations on a contrast agent concentration curve of the voxels in the modified contrast data with respect to an arterial input function.

[0117] The one or more perfusion parameters can include a relative cerebral blood volume, which corresponds to the area under the tissue retention function of the individual voxel. The one or more perfusion parameters can also include a relative cerebral blood flow, which corresponds to a peak in the tissue retention function of the individual voxel. Additionally, the one or more perfusion parameters can include a mean tracer transit time, which corresponds to a ratio of cerebral blood volume to cerebral blood flow of the individual voxel. In one or more examples, the one or more perfusion parameters can include a T, which corresponds to a time of a peak in the tissue retention function of the individual voxel. max may include.

[0118] Process 700 may include generating first overlay contrast data corresponding to the first contrast data and showing the one or more first regions as a first overlay of the individual's brain in which the contrast agent is present, at operation 708. In one or more examples, the first overlay may be displayed on a slice of the perfusion-based computed tomography image.

[0119] At operation 710, the process 700 can include determining that the tissue occurs in association with one or more second regions of the individual's brain. A portion of the one or more second regions can overlap at least a portion of the one or more first regions. In various examples, a volume indicated by the second overlay can be larger than a volume indicated by the first overlay.

[0120] In one or more examples, damage to tissue associated with one or more second regions of the individual's brain can be identified based on a low density value associated with the one or more second regions. The low density value can be identified by analyzing intensity values ​​of voxels included in the one or more second regions. To illustrate, a correlation can be identified between a first voxel in a first hemisphere of the individual's brain and a second voxel in a second hemisphere of the individual's brain. In addition, a difference in a first intensity value of an individual first voxel relative to a second intensity value of the individual second voxel can be identified. The individual second voxels can be displayed contralaterally to the individual first voxels. The difference between the first intensity value and the second intensity value can indicate a low density measure of brain tissue corresponding to the individual first voxels and the individual second voxels.

[0121] In one or more embodiments, the low-density measure can be generated by analyzing the difference with respect to several threshold intensity differences to generate several groups of voxels, each group of voxels corresponding to one or more threshold intensity differences of the several threshold intensity difference values. The one or more threshold intensity difference values ​​corresponding to the respective groups of voxels can indicate a low-density measure of each portion of the brain tissue corresponding to the respective group of voxels. In various examples, a first group of voxels corresponding to at least a first threshold intensity difference value can be identified. The first group of voxels can correspond to a second region of the one or more second regions of the second overlay, and the first group of voxels can correspond to a first measure of low density of the first portion of the brain tissue. Furthermore, a second group of voxels corresponding to at least a second threshold intensity difference value can be identified. The second group of voxels can correspond to an additional second region of the one or more second regions of the second overlay. The second group of voxels can correspond to a second measure of low concentration in a second portion of the brain tissue, and the second measure of low concentration can be different from the first measure of low concentration. In one or more illustrative examples, different rates of low concentration corresponding to different portions of the brain tissue can be displayed to show different threshold intensity difference values. For example, the first group of voxels can be displayed in conjunction with a first color and the second group of voxels can be displayed in conjunction with a second color.

[0122] Process 700 may also include generating second overlay contrast data at operation 712. The second overlay contrast data may indicate one or more regions of the individual's brain as a second overlay on a second image of the individual's brain. The second image may be a slice included in the second contrast data. Additionally, the second image may indicate that contrast agent is absent from the individual's brain.

[0123] Further, in operation 714, the process 700 can include generating an aggregate image including the first image of the individual's brain, the first overlay, and the second overlay. In one or more illustrative examples, user interface data corresponding to a user interface and that can be provided to the computing device for display to a medical professional can be generated. The user interface can also display a plurality of slices included in the first contrast data. At least some of the individual slices can include each of the first overlays corresponding to one or more regions of brain tissue represented by the individual slices that are damaged or have a damage probability of at least a threshold probability according to values ​​of the one or more perfusion parameters. Further, at least some of the individual slices can include each of the second overlays corresponding to one or more regions of brain tissue represented by the individual slices that are damaged or have a damage probability of at least a threshold probability according to low density values ​​of the one or more regions.

[0124] FIG. 8 is a flow chart illustrating example operations of a process 800 for generating an output image indicative of damaged brain tissue based on the onset of a biological condition corresponding to one or more blood vessels of the individual's brain, according to one or more example embodiments. The process 800 may include, at operation 802, acquiring first imaging data of the individual's brain. The first imaging data may be generated by a first imaging technique. The first imaging data may correspond to a perfusion characteristic of a fluid in the individual's brain. At operation 804, the process 800 may also include acquiring second imaging data of the individual's brain. The second imaging data may be generated by a second imaging technique that is a non-perfusion based imaging modality. In one or more examples, the first imaging technique may implement one or more perfusion based computed tomography (CT) imaging techniques, and the second imaging technique implements one or more non-contrast CT imaging techniques. In one or more additional examples, the first imaging technique performs one or more computed tomography angiography (CTA) techniques and the second imaging technique performs one or more non-contrast CT imaging techniques. In one or more further examples, the first imaging data can be generated by an imaging technique, including a non-perfusion based imaging technique, and the second imaging data can be generated by an additional imaging technique, including a perfusion based imaging technique.

[0125] Additionally, at operation 806, process 800 can include determining an amount of time that has elapsed since the onset of a biological condition corresponding to one or more blood vessels of the individual's brain. In one or more examples, the biological condition is an obstruction of blood flow to at least one region of the individual's brain. In one or more illustrative examples, the biological condition can include a stroke.

[0126] The process 800 may include, at operation 808, determining one or more threshold levels for one or more blood flow parameters associated with a blood vessel of the individual's brain based on the first imaging data and the amount of time. The one or more threshold levels may include a first threshold level for the blood flow parameter corresponding to a first time period elapsed from onset of the biological condition. Additionally, the one or more threshold levels may include a second threshold level for the blood flow parameter corresponding to a second time period elapsed from onset of the biological condition. The second time period may be a time period longer than the first time period. In various examples, the first time period may be 15 minutes to 2 hours or 30 minutes to 4 hours. Additionally, the second time period may be 4 hours to 6 hours, 3 hours to 5 hours, or 5 hours to 8 hours. In one or more further examples, the second threshold level may correspond to a value of the blood flow parameter greater than the first threshold level. In one or more illustrative examples, the blood flow parameter may include cerebral blood flow.

[0127] Further, in operation 810, the process 800 may include determining that the blood flow parameter is at least a threshold level for the brain of the individual. One or more values ​​of the blood flow parameter may be determined by analyzing voxel information of the first imaging data. In one or more illustrative examples, the blood flow parameter may be analyzed for a first threshold level during a first time period. In these circumstances, damage to the region of the brain of the individual may be determined based on the blood flow parameter being at least the first threshold level. Additionally, the blood flow parameter may also be analyzed for a second threshold level during a second time period. In these circumstances, damage to the region of the brain of the individual may be determined based on the blood flow parameter being at least the second threshold level. In one or more examples, the blood flow parameter may be analyzed during both the first time period and the second time period. In one or more additional examples, in situations where damage to the region of the brain of the individual is detected during the first time period, subsequent analysis during the second time period may not be performed. In one or more further examples, in situations where damage to an area of ​​the individual's brain is not detected during a first time period, one or more subsequent analyses can be performed, such as during a second time period.

[0128] Further, the process 800 may include, at operation 812, identifying one or more hypodensity measures for the individual's brain based on the second imaging data. Intensity values ​​of voxels included in the imaging data may be analyzed to identify regions of the individual's brain where the damage probability is at least a threshold probability. In at least some examples, voxels in different hemispheres of the individual's brain may be analyzed to identify one or more regions of the brain that include hypodense tissue. In one or more examples, one or more machine learning techniques may be implemented to identify regions of hypodense tissue in the individual's brain. In one or more additional examples, one or more weights of one or more parameters of the model may be determined based on an amount of time to measure damage to tissue of the individual's brain. In one or more illustrative examples, intensity values ​​of some voxels of the additional imaging data may be analyzed to identify values ​​of parameters of the model, and the parameter weights may also be determined based on an amount of time. In various examples, the model may be run using the parameter weights to identify hypodense measures for regions of the individual's brain.

[0129] At operation 814, the process 800 can include determining that damage is present in one or more regions of the individual's brain based on the one or more hypoconcentration measures. In situations where the hypoconcentration measures are at least a threshold level of hypoconcentration, damage is likely present in the individual's brain.

[0130] Process 800 may also determine a likelihood that the one or more low density measures correspond to tissue damage in one or more regions based on the amount of time at operation 816. In one or more examples, one or more computational models, such as one or more machine learning classification models or one or more random forest models, may be performed to determine the likelihood of the low density measures indicating that tissue damage is present in the individual's brain.

[0131] At operation 818, process 800 may include generating an image indicating that at least one region of the individual's brain includes damaged tissue. In one or more examples, the image may be included in a user interface that displays at least one of the one or more values ​​of the blood flow parameter and the one or more hypodensity measures during a first time period. In at least some examples, the user interface may also display the one or more values ​​of the blood flow parameter and the one or more hypodensity measures during a second time period. Additionally, the user interface may include one or more overlays that indicate one or more regions of the individual's brain where the damage occurred. In one or more illustrative examples, the damage to the individual's brain may occur due to a stroke.

[0132] FIG. 9 is an illustration of an example of a user interface 900 including several slices of a perfusion-based CT image with overlays showing regions of interest in the individual's brain identified according to different perfusion parameters and low-density analysis, according to one or more example embodiments. For example, the user interface 900 can include several first slices 902. The several first slices 902 can be captured using a perfusion-based computed tomography imaging technique and can correspond to different sections of the individual's brain captured at different times. Some of the several first slices 902 can have a first overlay 904 that corresponds to a region of the individual's brain where the cerebral blood flow (CBF) value is below a threshold value (30%). The threshold value can correspond to an amount of difference between the cerebral blood flow value of the tissue included in the region associated with the first overlay 904 and healthy brain tissue.

[0133] Some of the slices 902 may also include a second overlay 906. The second overlay may correspond to regions of the brain having low density values ​​within a threshold Hounsfield unit difference range (≧5 HU and ≦12 HU). The threshold difference range of voxels included in the region associated with the second overlay 906 may relate to contralateral voxels in the brain hemisphere that are on the opposing hemisphere of the brain that includes the region associated with the first overlay 904 and the second overlay 906. In one or more illustrative examples, a combination of the volume associated with the first overlay 904 and the volume associated with the second overlay 906 may yield a total volume of damaged tissue in the individual's brain.

[0134] Although the illustrative example of FIG. 9 shows the first overlay 904 and the second overlay 906 as areas having different patterns, in additional examples, the first overlay 904 and the second overlay 906 may be displayed using different colors, such as a union of multiple colors or such that the overlapping areas may be semi-transparent and displayed as a blend of colors.

[0135] FIG. 10 is a block diagram illustrating components of a machine 1000 according to some example implementations that can read instructions from a machine-readable medium (e.g., a machine-readable storage medium) and perform any one or more of the methodologies discussed herein. In particular, FIG. 10 illustrates a diagrammatic representation of the machine 1000 in the form of an example computer system within which instructions 1002 (e.g., software, programs, applications, applets, apps, or other executable code) can be executed that cause the machine 1000 to perform any one or more of the methodologies discussed herein. Thus, the instructions 1002 can be used to implement modules or components described herein. The instructions 1002 transform a generic, unprogrammed machine 1000 into a specific machine 1000 that is programmed to perform the functions described and illustrated as described. In alternative implementations, the machine 1000 can operate as a standalone device or be coupled (e.g., networked) to other machines. In a networked deployment, the machine 1000 may operate in the capacity of a server or client machine in a server-client network environment, or as a peer machine in a peer-to-peer (or distributed) network environment. The machine 1000 may include, but is not limited to, a server computer, a client machine, a personal computer (PC), a tablet computer, a laptop computer, a netbook, a set-top box (STB), a personal digital assistant (PDA), an entertainment media system, a cellular phone, a smartphone, a mobile phone, a wearable device (e.g., a smart watch), a smart home device (e.g., a smart appliance), other smart devices, a web appliance, a network router, a network switch, a network bridge, or any machine capable of sequentially or otherwise executing instructions 1002 that specify operations to be performed by the machine 1000. Additionally, although only a single machine 1000 is shown, the term "machine" shall also be construed to include a collection of machines that individually or collectively execute instructions 1002 to perform any one or more of the methodologies discussed herein.

[0136] The machine 1000 may include a processor 1004, memory / storage 1006, and I / O components 1008, which may be configured to communicate with each other via a bus 1010 or the like. In this regard, a "processor" refers to any circuit or virtual circuit (a physical circuit emulated by logic executed in an actual processor 1004) that manipulates data values ​​in accordance with control signals (e.g., "commands," "op-codes," "machine codes," etc.) and generates corresponding output signals that are applied to operate the machine 1000. In an example embodiment, the processor 1004 (e.g., a central processing unit (CPU), a reduced instruction set computing (RISC) processor, a complex instruction set computing (CISC) processor, a graphics processing unit (GPU), a digital signal processor (DSP), an application specific integrated circuit (ASIC), a radio frequency integrated circuit (RFIC), another processor, or any suitable combination thereof) may include a processor 1012 and a processor 1014 that may execute instructions 1002, for example. The term "processor" is intended to encompass a multi-core processor 1004 that may include two or more independent processors (sometimes referred to as "cores") that may simultaneously execute instructions 1002. Although Figure 10 shows multiple processors 1004, machine 1000 may include a single processor 1012 having a single core, a single processor 1012 having multiple cores (e.g., a multi-core processor), multiple processors 1012, 1014 having a single core, multiple processors 1012, 1014 having multiple cores, or any combination thereof.

[0137] The memory / storage 1006 may include a memory and storage unit 1018, such as a main memory 1016 or other memory storage, both of which are accessible to the processor 1004 via a bus 1010 or the like. The storage unit 1018 and the main memory 1016 store instructions 1002 that embody any one or more of the methodologies or functions described herein. The instructions 1002 may reside, completely or partially, within the main memory 1016, within the storage unit 1018, within at least one of the processors 1004 (e.g., within a processor's cache memory), or any suitable combination thereof, during execution by the machine 1000. Thus, the main memory 1016, the storage unit 1018, and the memory of the processor 1004 are examples of machine-readable media. In this regard, a "machine-readable medium," also referred to herein as a "computer-readable storage medium," refers to a component, device, or other tangible medium capable of temporarily or permanently storing instructions 1002 and data, and may include, but is not limited to, random access memory (RAM), read-only memory (ROM), buffer memory, flash memory, optical media, magnetic media, cache memory, other types of storage (e.g., erasable programmable read-only memory (EEPROM)), and / or any suitable combination thereof. The term "machine-readable medium" may be interpreted as including a single medium or multiple media capable of storing instructions 1002 (e.g., a centralized or distributed database or associated caches and servers). The term "machine-readable medium" shall also be interpreted as including any medium or combination of media capable of storing instructions 1002 (e.g., code) executed by machine 1000 such that, when the instructions 1002 are executed by one or more processors 1004 of machine 1000, they cause machine 1000 to perform any one or more of the methodologies described herein. Thus, "machine-readable medium" refers to a single storage device or device as well as a "cloud-based" storage system or storage network including multiple storage devices or devices. The term "machine-readable medium" by itself excludes signals.

[0138] The I / O components 1008 may include a wide variety of components for receiving input, providing output, generating output, transmitting information, exchanging information, capturing measurements, and the like. The particular I / O components 1008 included in a particular machine 1000 will depend on the type of machine. For example, a portable machine such as a mobile phone will likely include a touch input device or other such input mechanism, while a headless server machine will likely not include such a touch input device. It will be appreciated that the I / O components 1008 may include many other components not shown in FIG. 10. For the sole purpose of simplification of the following discussion, the I / O components 1008 are grouped according to function, and this grouping is in no way limiting. In various example embodiments, the I / O components 1008 may include a user output component 1020 and a user input component 1022. User output components 1020 may include visual components (e.g., a display such as a plasma display panel (PDP), a light emitting diode (LED) display, a liquid crystal display (LCD), a projector, or a cathode ray tube (CRT)), acoustic components (e.g., speakers), haptic components (e.g., vibration motors, resistive mechanisms), other signal generators, etc. User input components 1022 may include alphanumeric input components (e.g., a keyboard, a touch screen configured to receive alphanumeric input, an optical keyboard, or other alphanumeric input component), point-based input components (e.g., a mouse, touchpad, trackball, joystick, motion sensor, or other pointing device), haptic input components (e.g., a physical button, a touch screen that provides the position or force of a touch or touch gesture, or other haptic input component), audio input components (e.g., a microphone), etc.

[0139] In further example embodiments, the I / O components 1008 may include a biometric component 1024, a motion component 1026, an environmental component 1026, or a position component 1030, among other components. For example, the biometric components 1024 may include components for detecting expressions (e.g., hand expressions, facial expressions, voice expressions, body gestures, or eye tracking), measuring biosignals (e.g., blood pressure, heart rate, body temperature, sweat, or brain waves), identifying people (e.g., voice identification, retina identification, face identification, fingerprint identification, or brain wave-based identification), etc. The motion components 1026 may include an acceleration sensor component (e.g., an accelerometer), a gravity sensor component, a rotation sensor component (e.g., a gyroscope), etc. The environmental components 1026 may include, for example, a lighting sensor component (e.g., a light meter), a temperature sensor component (e.g., one or more thermometers to detect ambient temperature), a humidity sensor component, a pressure sensor component (e.g., a barometer), an acoustic sensor component (e.g., one or more microphones to detect background noise), a proximity sensor component (e.g., an infrared sensor to detect nearby objects), a gas sensor (e.g., a gas detection sensor to detect concentrations of harmful gases or detect air pollution for safety purposes), or other components capable of providing indications, measurements, or signals corresponding to the surrounding physical environment. The location components 1030 may include a location sensor component (e.g., a Global Positioning System (GPS) receiver component), an altitude sensor component (e.g., an altimeter or barometer to detect air pressure from which altitude can be derived), a direction sensor component (e.g., a magnetometer), etc.

[0140] Communications may be implemented using a wide range of technologies. The I / O component 1008 may include a communication component 1032 operable to couple the machine 1000 to a network 1034 or a device 1036. For example, the communication component 1032 may include a network interface component or other device suitable for interfacing with the network 1034. In further examples, the communication component 1032 may include a wired communication component, a wireless communication component, a cellular communication component, a near field communication (NFC) component, a Bluetooth® component (e.g., Bluetooth® Low Energy), a Wi-Fi® component, and other communication components providing communication via other forms. The device 1036 may be another machine 1000 or any of a wide range of peripheral devices (e.g., a peripheral device coupled via USB).

[0141] Further, the communication component 1032 may detect the identifier or may include a component operable to detect the identifier. For example, the communication component 1032 may include a radio frequency identification (RFID) tag reader component, an NFC smart tag detection component, an optical reader component (e.g., an optical sensor for detecting one-dimensional barcodes such as Universal Product Code (UPC) barcodes, multi-dimensional barcodes such as Quick Response Codes (QR Codes), Aztec Code, Data Matrix, Dataglyph, MaxiCode, PDF417, Ultra Code, UCC RSS-2D barcodes, and other optical codes, etc.), or an acoustic detection component (e.g., a microphone for identifying tagged acoustic signals). In addition, various information may be derived via the communication component 1032, such as location via Internet Protocol (IP) geolocation, location via Wi-Fi signal triangulation, location via detection of NFC beacon signals that may indicate a particular location, etc.

[0142] In this regard, a "component" refers to a device, physical entity, or logic having boundaries defined by function or subroutine calls, branch points, APIs, or other techniques that provide partitioning or modularization of a particular processing or control function. A component may be coupled through interfaces with other components to perform a machine process. A component may be a packaged functional hardware unit designed to be used with other components and portions of a program that typically perform a particular function among the associated functions. A component may be comprised of either a software component (e.g., code embodied in a machine-readable medium) or a hardware component. A "hardware component" is a tangible unit capable of performing a particular operation and may be configured or arranged in a particular physical manner. In various example embodiments, one or more computer systems (e.g., a stand-alone computer system, a client computer system, or a server computer system), or one or more hardware components of a computer system (e.g., a processor or processors) may be configured by software (e.g., an application or application portions) as hardware components that operate to perform particular operations as described herein.

[0143] A hardware component may also be implemented mechanically, electronically, or any suitable combination thereof. For example, a hardware component may include dedicated circuitry or logic that is permanently configured to perform a particular operation. A hardware component may be a dedicated processor, such as a field programmable gate array (FPGA) or an ASIC. A hardware component may also include programmable logic or circuitry that is temporarily configured by software to perform a particular operation. For example, a hardware component may include software executed by a general purpose processor 1004 or other programmable processor. Once configured by such software, the hardware component becomes a specific machine (or a specific component of the machine 1000) uniquely adapted to perform the function for which it is configured, and is no longer a general purpose processor 1004. It will be appreciated that the decision to implement a mechanical hardware component with dedicated permanently configured circuitry or with temporarily configured circuitry (e.g., configured by software) may depend on cost and time considerations. Thus, the phrase "hardware component" (or "hardware-implemented component") should be understood to encompass tangible entities, whether the entities are physically constructed, permanently configured (e.g., hardwired), or temporarily configured (e.g., programmed) to operate in a particular manner or to perform particular operations described herein. We contemplate embodiments in which hardware components are temporarily configured (e.g., programmed), and each hardware component need not be configured or instantiated at all times. For example, if the hardware components include a general-purpose processor 1004 configured by software to be a special-purpose processor, the general-purpose processor 1004 may be configured as different special-purpose processors (e.g., including different hardware components) at different times. Thus, the software configures, for example, a particular processor 1012, 1014, or processor 1004 to configure a particular hardware component at one time and a different hardware component at another time.

[0144] Hardware components can provide information to and receive information from other hardware components. Thus, the described hardware components can be considered to be communicatively coupled. When multiple hardware components are present simultaneously, communication can be achieved through signal transmission between or among two or more of the hardware components (e.g., via appropriate circuits and buses). In embodiments in which multiple hardware components are configured or instantiated at different times, communication between such hardware components can be achieved, for example, through storage and retrieval of information in and from memory structures accessible to multiple hardware components. For example, one hardware component can perform an operation and store the output of that operation in a communicatively coupled memory device. Then, at a later time, an additional hardware component can access the memory device to retrieve and process the stored output.

[0145] A hardware component may also initiate communication with an input device or an output device and may operate on a resource (e.g., a collection of information). Various operations of the example methods described herein may be performed, at least in part, by one or more processors 1004 that are temporarily configured (e.g., by software) or permanently configured to perform the associated operations. Whether temporarily or permanently configured, such processors 1004 may constitute processor-implemented components that operate to perform one or more operations or functions described herein. As used herein, a "processor-implemented component" refers to a hardware component that is implemented using one or more processors 1004. Similarly, the methods described herein may be at least in part implemented in a processor, with a particular processor 1012, 1014 or processor 1004 being an example of hardware. For example, at least some of the operations of the methods may be performed by one or more processors 1004 or processor-implemented components. Additionally, one or more processors 1004 may also operate to support execution of associated operations in a "cloud computing" environment or as a "software as a service" (SaaS). For example, at least some of the operations may be performed by a cluster of computers (as an example of machine 1000 including processor 1004), and these operations may be accessible via a network 1034 (e.g., the Internet) and via one or more suitable interfaces (e.g., APIs). A particular implementation of the operations may not only reside within a single machine 1000, but may be distributed among processors deployed across several machines. In some example implementations, processor 1004 or processor-implemented components may be located in a single geographic location (e.g., a home environment, an office environment, or a server farm). In other example implementations, processor 1004 or processor-implemented components may be distributed across several geographic locations.

[0146] FIG. 11 is a block diagram illustrating a system 1100 including an example software architecture 1102 that may be used in conjunction with various hardware architectures described herein. It will be appreciated that FIG. 11 is a non-limiting example of a software architecture and that many other architectures may be implemented to facilitate the functionality described herein. The software architecture 1102 may be implemented in hardware such as the machine 1000 of FIG. 10, which includes, among other things, a processor 1004, memory / storage 1006, and input / output (I / O) components 1008. A representative hardware layer 1104 is shown and may represent, for example, the machine 1000 of FIG. 10. The representative hardware layer 1104 includes a processing unit 1106 having associated executable instructions 1108. The executable instructions 1108 represent executable instructions for the software architecture 1102, including implementations of methods, components, etc. described herein. The hardware layer 1104 also includes at least one of a memory or storage module memory / storage 1110, which also has executable instructions 1108. The hardware layer 1104 may include other hardware 1112 .

[0147] In the example architecture of FIG. 11, software architecture 1102 may be conceptualized as a stack of layers, with each layer providing a particular function. For example, software architecture 1102 may include layers such as operating system 1114, libraries 1116, framework / middleware 1118, application 1120, and presentation layer 1120. Operationally, application 1120 or other components within a layer may invoke API calls 1124 through the software stack and receive messages 1126 in response to API calls 1124. The layers shown are representative in nature, and not all software architectures have all layers. For example, some mobile or dedicated operating systems do not provide framework / middleware 1118, while others provide such a layer. Other software architectures may include additional or different layers.

[0148] The operating system 1114 may manage hardware resources and provide general services. The operating system 1114 may include, for example, a kernel 1128, services 1130, and drivers 1132. The kernel 1128 may act as an abstraction layer between the hardware and other software layers. For example, the kernel 1128 may be responsible for memory management, processor management (e.g., scheduling), component management, networking, security configuration, etc. The services 1130 may provide other common services to the other software layers. The drivers 1132 are responsible for controlling or interfacing with the underlying hardware. For example, the drivers 1132 may include a display driver, a camera driver, a Bluetooth® driver, a flash memory driver, a serial communication driver (e.g., a Universal Serial Bus (USB) driver), a Wi-Fi® driver, an audio driver, a power management driver, etc., depending on the hardware configuration.

[0149] The libraries 1116 provide a common foundation used by the applications 1120 and / or other components or layers. The libraries 1116 provide functions that allow other software components to perform tasks more easily than by directly interfacing with the underlying operating system 1114 functions (e.g., kernel 1128, services 1130, drivers 1132). The libraries 1116 may include system libraries 1134 (e.g., C standard libraries) that may provide functions such as memory allocation functions, string manipulation functions, mathematical functions, etc. Additionally, the libraries 1116 may include API libraries 1136 such as media libraries (e.g., libraries for supporting the presentation and manipulation of various media formats such as MPEG4, H.264, MP3, AAC, AMR, JPG, PNG, etc.), graphics libraries (e.g., OpenGL framework that may be used to render two- or three-dimensional in graphical content to a display), database libraries (e.g., SQLite that may provide various relational database functions), web libraries (e.g., WebKit that may provide web browsing functions), etc. The libraries 1116 may also include a wide variety of other libraries 1138 for providing many other APIs to the applications 1120 and other software components / modules.

[0150] Framework / middleware 1118 (sometimes referred to as middleware) provides a higher level common foundation that can be used by applications 1120 or other software components / modules. For example, framework / middleware 1118 may provide various graphical user interface functionality, high level resource management, high level location services, etc. Framework / middleware 1118 may provide a wide range of other APIs, some of which may be specific to a particular operating system 1114 or platform, that can be utilized by applications 1120 or other software components / modules.

[0151] Applications 1120 include built-in applications 1140 and third party applications 1142. Representative examples of built-in applications 1140 may include, but are not limited to, a contacts application, a browser application, a book reader application, a location application, a media application, a messaging application, or a gaming application. Third party applications 1142 may include applications developed using the ANDROID® or IOS™ Software Development Kit (SDK) by an entity other than a particular platform vendor, and may be mobile software running on a mobile operating system such as IOS™, ANDROID®, WINDOWS® Phone, or other mobile operating system. Third party applications 1142 may call API calls 1124 provided by the mobile operating system (e.g., operating system 1114) to facilitate the functionality described herein.

[0152] Applications 1120 may use built-in operating system facilities (e.g., kernel 1128, services 1130, drivers 1132), libraries 1116, and frameworks / middleware 1118 to create a UI for interacting with a user of the system. Alternatively or additionally, in some systems, interaction with the user may occur through a presentation layer, such as presentation layer 1120. In these systems, the application / component "logic" may be separate from the aspects of the application / component that interact with the user.

[0153] Changes and modifications can be made to the disclosed embodiments without departing from the scope of the disclosure. These and other changes or modifications are intended to be included within the scope of the disclosure, as expressed in the following claims.

[0154] A non-limiting, numbered list of aspects of the present subject matter is provided below. Aspect 1. A method comprising: acquiring, with one or more computing devices, each including a processor and a memory, first imaging data indicative of a first characteristic within the individual's brain based on a presence of an imaging agent within the individual's brain; acquiring, with at least one computing device of the one or more computing devices, second imaging data indicative of a second characteristic within the individual's brain in the absence of the imaging agent within the individual's brain; determining, with the at least one computing device of the one or more computing devices, that damage to tissue has occurred associated with one or more first regions of the individual's brain based on the first imaging data; and generating, with the at least one computing device of the one or more computing devices, a first overlay imaging data indicative of the one or more first regions as a first overlay of a first image of the individual's brain corresponding to the first imaging data. generating, by at least one computing device of the one or more computing devices, second overlay imaging data corresponding to the second contrast data, showing the one or more second regions as a second overlay of a first image of the individual's brain without the presence of contrast agent; and generating, by the at least one computing device of the one or more computing devices, an aggregate image including the first image, the first overlay, and the second overlay.

[0155] Aspect 2. The method of aspect 1, wherein the first imaging data is generated by a first imaging technique and the second imaging data is generated by a second imaging technique that is different from the first imaging technique.

[0156] Aspect 3. The method of aspect 2, wherein the first imaging technique performs one or more perfusion-based computed tomography (CT) imaging techniques and the second imaging technique performs one or more non-contrast CT imaging techniques.

[0157] Aspect 4. The method of aspect 2, wherein the first imaging technique performs one or more computed tomography angiography (CTA) imaging techniques and the second imaging technique performs one or more non-contrast CT imaging techniques.

[0158] Aspect 5. The method of any one of aspects 1-4, comprising acquiring, by at least one computing device of the one or more computing devices, training images captured during a time period during which contrast agent is present in the brain of an additional individual prior to the first and second images, the training images including a first group of images showing at least one damaged brain region and a second group of images showing an absence of the damaged brain region; and performing, by at least one computing device of the one or more computing devices, a training process using the training images, thereby identifying values ​​for parameters of one or more trained models corresponding to one or more convolutional neural networks, the one or more trained models being used to identify the one or more first regions of the brain of the damaged individual.

[0159] Aspect 6. The first imaging data includes several first images captured consecutively over a period of time, the method including: identifying, by at least one computing device of the one or more computing devices, one or more first images offset from a reference image by at least a threshold amount corresponding to movement of the individual during capture of the first imaging data; performing, by at least one computing device of the one or more computing devices, a registration process to thereby align the one or more first images with the reference image; and performing, by the at least one computing device of the one or more computing devices, a registration process to match voxels of the one or more first images with voxels of the reference image based on the registration process. The method of any one of aspects 1 to 5, comprising: identifying one or more motion correction parameters including at least one of one or more rotational parameters or one or more translational parameters that position a corresponding voxel within a threshold distance; determining, by at least one computing device of the one or more computing devices, that a portion of the several first images was captured at a time interval that is different from an additional time interval of capture of an additional portion of the several first images; and modifying, by at least one computing device of the one or more computing devices, the time interval and the additional time interval to be a common time interval such that the several first images are positioned in a common state with a time interval between successive first images.

[0160] Aspect 7. A method as described in any one of aspects 1 to 6, comprising: determining, by at least one computing device of the one or more computing devices, a change in intensity value of at least a portion of the voxels included in the first contrast data over a period of time; and determining, by at least one computing device of the one or more computing devices, an arrival time at which the contrast agent entered the portion of the individual's brain based on the change in intensity value.

[0161] Aspect 8. The method of aspect 7, comprising generating, by at least one computing device of the one or more computing devices, first contrast data based on a first image captured subsequent to the arrival time, and generating, by at least one computing device of the one or more computing devices, second contrast data based on a second image captured prior to the arrival time.

[0162] Aspect 9. The method of aspect 8, wherein the first contrast data and the second contrast data are generated using one or more perfusion-based computed tomography imaging techniques. Aspect 10. A method as described in aspect 8, comprising: performing, by at least one computing device of the one or more computing devices, one or more deconvolution operations on a contrast agent concentration curve of a voxel included in the first contrast data with respect to an arterial input function, thereby generating a tissue residue function; and determining, by at least one computing device of the one or more computing devices, a plurality of perfusion parameters based on the tissue residue function of an individual voxel included in each slice of the first contrast data, wherein the one or more first regions are identified based on the plurality of perfusion parameters.

[0163] Aspect 11. The multiple perfusion parameters include, for each voxel, a relative cerebral blood volume corresponding to the area under the tissue residue function of the individual voxel, a relative cerebral blood flow corresponding to the peak of the tissue residue function of the individual voxel, a mean tracer transit time corresponding to the ratio of the relative cerebral blood volume to the relative cerebral blood flow of the individual voxel, and a T corresponding to the time of the peak of the tissue residue function of the individual voxel. max The method of embodiment 10, comprising:

[0164] Aspect 12. A system, comprising: one or more hardware processors; and one or more non-transitory computer readable storage media comprising computer readable instructions that, when executed by the one or more hardware processors, cause the one or more hardware processors to perform operations, including acquiring first imaging data indicative of a first characteristic within the individual's brain based on a presence of an imaging agent within the individual's brain; acquiring second imaging data indicative of a second characteristic within the individual's brain in the absence of the imaging agent within the individual's brain; determining that damage to tissue has occurred associated with one or more first regions of the individual's brain based on the first imaging data; and determining an individual characteristic corresponding to the first imaging data. generating first overlay contrast data illustrating the one or more first regions as a first overlay on a first image of the brain of the individual; determining based on the second contrast data that tissue damage has occurred associated with one or more second regions of the individual's brain, where a portion of the one or more second regions overlaps with at least a portion of the one or more first regions; generating second overlay contrast data illustrating the one or more second regions as a second overlay on the first image of the individual's brain without the presence of contrast agent, the second overlay corresponding to the second contrast data; and generating an aggregate image including the first image of the individual's brain, the first overlay, and the second overlay.

[0165] Aspect 13. One or more non-transitory computer readable storage media include additional computer readable instructions that, when executed by one or more hardware processors, cause the one or more hardware processors to perform additional operations, including identifying a correlation between a first voxel in a first hemisphere of the individual's brain and a second voxel in a second hemisphere of the individual's brain, and identifying a difference in a first intensity value of each of the first voxels relative to a second intensity value of each of the second voxels, where each of the second voxels is greater than or equal to the first intensity value of each of the first voxels. The system of aspect 12, further comprising: an operation of identifying first and second intensity values ​​located on opposite sides of a cell, the difference between which indicates a low density measure for brain tissue corresponding to a respective first voxel and a respective second voxel; and an operation of analyzing the difference with respect to a number of threshold intensity difference values, thereby generating a number of voxel groups, each voxel group corresponding to one or more of the number of threshold intensity difference values, the one or more threshold intensity difference values ​​corresponding to the each voxel group indicating a low density measure for each portion of the brain tissue corresponding to the each voxel group.

[0166] Aspect 14. The system of Aspect 13, wherein the one or more non-transitory computer-readable storage media include additional computer-readable instructions which, when executed by the one or more hardware processors, cause the one or more hardware processors to perform additional operations, the additional operations including: identifying a first group of voxels corresponding to at least a first threshold intensity difference value, the first group of voxels corresponding to a second region of the one or more second regions of the second overlay, the first group of voxels corresponding to a first low-density measure of the first portion of the brain tissue; and identifying a second group of voxels corresponding to at least a second threshold intensity difference value, the second group of voxels corresponding to an additional second region of the one or more second regions of the second overlay, the second group of voxels corresponding to a second low-density measure of the second portion of the brain tissue, the second low-density measure different from the first low-density measure.

[0167] Aspect 15. The system of any one of Aspects 12-14, wherein the one or more non-transitory computer readable storage media includes additional computer readable instructions that, when executed by the one or more hardware processors, cause the one or more hardware processors to perform additional operations, including acquiring training images captured during a time period in which contrast agent is not present in a brain of a number of additional individuals prior to the first and second images, the training images including a first group of images showing at least one damaged brain region and a second group of images showing no damaged brain region, and performing, by at least one computing device of the one or more computing devices, a training process using the training images, thereby identifying values ​​for parameters of one or more trained models corresponding to the one or more convolutional neural networks, wherein a probability of irreversible damage to tissue occurring associated with the one or more second regions of the brain of the individual is identified using the one or more trained models.

[0168] Aspect 16. The system of any one of Aspects 12-15, wherein the one or more non-transitory computer-readable storage media includes additional computer-readable instructions which, when executed by the one or more hardware processors, cause the one or more hardware processors to perform additional operations, including identifying one or more perfusion parameters based on a first intensity value of a first voxel of the first contrast data, the one or more perfusion parameters being used to determine that damage to tissue has occurred in association with one or more first regions of the individual's brain.

[0169] Aspect 17. A method, comprising: acquiring, by one or more computing devices, each including a processor and a memory, first imaging data of a brain of an individual generated by a first imaging technique implementing one or more perfusion-based imaging techniques; acquiring, by at least one computing device of the one or more computing devices, second imaging data generated by a second imaging technique, where the second imaging technique is a non-perfusion-based imaging technique; determining, by the at least one computing device of the one or more computing devices, one or more perfusion parameters based on a first intensity value of a first voxel of the first imaging data; determining that tissue damage has occurred associated with one or more first regions of the individual's brain based on the one or more perfusion parameters; identifying, by at least one computing device of the one or more computing devices, one or more second regions of the individual's brain based on second intensity values ​​of second voxels of the second contrast data, where the one or more second regions correspond to one or more portions of brain tissue having a low density measure corresponding to a threshold measure of low density; and generating, by the at least one computing device of the one or more computing devices, an aggregate image including a first overlay indicative of the one or more first regions and a second overlay including the one or more second regions.

[0170] Aspect 18. The method of aspect 17, comprising: determining, by at least one computing device of the one or more computing devices, a cerebral blood flow of a region of the individual's brain based on the first imaging data; determining, by at least one computing device of the one or more computing devices, that the cerebral blood flow is below a threshold amount of cerebral blood flow; and determining, by at least one computing device of the one or more computing devices, that damage to tissue has occurred in association with the region based on the cerebral blood flow being below the threshold amount, wherein the region is included in the one or more first regions.

[0171] Aspect 19. A method as described in aspect 17 or 18, comprising: determining, by at least one computing device of the one or more computing devices, a measure of cerebral blood flow based on a first intensity value of a first voxel of the first contrast data; and determining, by at least one computing device of the one or more computing devices, one or more first regions based on a portion of the measure of cerebral blood flow of a portion of the first voxel that corresponds to a threshold measure of cerebral blood flow.

[0172] Aspect 20. A method as described in any one of aspects 17-20, wherein the first imaging data includes a plurality of slices captured over a period of time, each individual slice of the plurality of slices corresponding to a respective cross-section of the individual's brain, and the aggregate image corresponds to each individual slice of the plurality of slices, and the method includes generating, by at least one computing device of the one or more computing devices, user interface data corresponding to a user interface displaying each individual slice of the plurality of slices and displaying each of the first overlays and each of the second overlays of at least a portion of the individual slices of the plurality of slices.

[0173] Aspect 21. A method comprising: a computing system including one or more computing devices, each including a processor and a memory, acquiring imaging data indicative of a first characteristic within the brain of the individual based on the presence of an imaging agent in the brain of the individual; determining, by the computing system, an amount of time that has elapsed since onset of a biological condition corresponding to one or more blood vessels of the brain of the individual; determining, by the computing system, one or more threshold levels of one or more blood flow parameters associated with the blood vessels of the brain of the individual and based on the amount of time and the imaging data, wherein the one or more threshold levels are indicative of damage to the brain of the individual; determining, by the computing system, that one blood flow parameter of the one or more blood flow parameters is at least one of the one or more threshold levels for the blood vessels of the brain of the individual; and generating, by the computing system and based on the one blood flow parameter, an image indicative of at least one region of the brain of the individual including damaged tissue.

[0174] Aspect 22. The method of aspect 21, wherein the one or more thresholds include a first threshold level of the blood flow parameter corresponding to a first period of time elapsed since onset of the physiological condition, and the one or more thresholds include a second threshold level of the blood flow parameter corresponding to a second period of time elapsed since onset of the physiological condition.

[0175] Aspect 23. The method of aspect 22, comprising: determining, by the computing system, that the amount of time corresponds to a first amount of time; and determining, by the computing system, that damage has occurred to the region of the brain of the individual based on the blood flow parameter being at least the first threshold level.

[0176] Aspect 24. The method of aspect 22, wherein the second threshold level corresponds to a value of the blood flow parameter that is greater than the first threshold level. Aspect 25. The method of aspect 22, wherein the blood flow parameter comprises cerebral blood flow.

[0177] Aspect 26. A method according to any one of aspects 21 to 25, comprising determining the blood flow parameter by analyzing voxel information of the first contrast data by the computational system.

[0178] Aspect 27. A method as described in any one of aspects 21 to 26, comprising: acquiring, by the computing system, additional contrast data indicative of a second characteristic in the brain of the individual in the absence of the contrast agent in the brain of the individual; and determining, by the computing system and based on the amount of time, one or more weights of one or more parameters of a model for determining damage to tissue of the brain of the individual.

[0179] Aspect 28. The method of aspect 27, comprising: identifying values ​​of parameters of the model by analyzing intensity values ​​of several voxels of the additional contrast data by the calculation system; and determining weights of the parameters by the calculation system and based on the amount of time.

[0180] Aspect 29. The method of aspect 28, comprising causing the model run by the computing system using the weights of the parameters to identify a low concentration measure in the region of the brain of the individual, the computing system determining that the low concentration measure is at least a low concentration threshold level, and the computing system determining that tissue in the region of the brain of the individual is damaged.

[0181] Aspect 30. The method of any one of aspects 21 to 29, wherein the biological condition is an obstruction of blood flow to at least one region of the brain of the individual. Aspect 31. A system, comprising: one or more hardware processors; and one or more non-transitory computer readable storage media containing computer readable instructions that, when executed by the one or more hardware processors, cause the one or more hardware processors to perform operations, including: acquiring imaging data indicative of a first characteristic within the brain of the individual based on a presence of an imaging agent within the brain of the individual; identifying an amount of time that has elapsed since onset of a biological condition corresponding to one or more blood vessels of the brain of the individual; and determining a blood vessel density of the brain of the individual. determining one or more threshold levels for one or more blood flow parameters associated with a vessel based on the amount of time and the imaging data, the one or more threshold levels being indicative of damage to the individual's brain; determining that one of the one or more blood flow parameters is at least one of the one or more threshold levels for blood vessels of the individual's brain; and generating an image indicative of at least one region of the individual's brain including damaged tissue based on the blood vessels of the individual's brain.

[0182] Aspect 32. The system of aspect 31, wherein the one or more non-transitory computer-readable storage media include additional computer-readable instructions which, when executed by the one or more hardware processors, cause the one or more hardware processors to perform additional operations, including an operation of acquiring additional contrast data indicative of a second characteristic in the brain of the individual in the absence of the contrast agent in the brain of the individual, and an operation of determining one or more weights of one or more parameters of a model for determining damage to tissue of the brain of the individual based on the amount of time.

[0183] Aspect 33. The system of aspect 32, wherein the first imaging technique performs one or more perfusion-based computed tomography (CT) imaging techniques and the second imaging technique performs one or more non-contrast CT imaging techniques.

[0184] Aspect 34. The system of aspect 32, wherein the first imaging technique performs one or more computed tomography angiography (CTA) imaging techniques and the second imaging technique performs one or more non-contrast CT imaging techniques.

[0185] Aspect 35. The system of aspect 32, wherein the one or more non-transitory computer-readable storage media include additional computer-readable instructions that, when executed by the one or more hardware processors, cause the one or more hardware processors to perform additional operations, the additional operations including generating an image including a first overlay corresponding to the region of the brain of the individual and a second overlay corresponding to an additional region of the brain of the individual.

[0186] Aspect 36. A method comprising: acquiring, by a computing system including one or more processors and a memory, contrast data indicative of characteristics within the brain of the individual in the absence of a contrast agent in the brain of the individual; determining, by the computing system, an amount of time that has elapsed since onset of a biological condition corresponding to one or more blood vessels in the brain of the individual; determining, by the computing system and based on the amount of time, one or more weights of one or more parameters of a model for determining damage to tissue of the brain of the individual; identifying, by the computing system, a value of one of the one or more parameters of the model by analyzing intensity values ​​of several voxels of additional contrast data and determining, by the computing system and based on the amount of time, a weight of the one parameter; and generating, by the computing system and based on the blood vessels of the brain of the individual, an image indicative of at least one region of the brain of the individual including damaged tissue.

[0187] Aspect 37. The method of aspect 36, comprising: acquiring, by the computing system, additional contrast data indicative of additional characteristics within the individual's brain based on the presence of a contrast agent within the individual's brain; and determining, by the computing system and based on the amount of time, one or more threshold levels of one or more blood flow parameters associated with blood vessels in the individual's brain, the one or more threshold levels being indicative of damage to the individual's brain.

[0188] Aspect 38. The method of aspect 37, comprising determining, by the computational system and based on the amount of time, one or more threshold levels of one or more blood flow parameters associated with blood vessels of the brain of the individual, the one or more threshold levels being indicative of damage to the brain of the individual, and determining, by the computational system, that one blood flow parameter of the one or more blood flow parameters is at least one of the one or more threshold levels for blood vessels of the brain of the individual.

[0189] Aspect 39. The method of aspect 37, wherein the imaging data is generated by an imaging technique including a non-perfusion based imaging technique, and the additional imaging data is generated by an additional imaging technique including a perfusion based imaging technique.

[0190] Aspect 40. The method of aspect 38, comprising: generating, by the computing system, first user interface data corresponding to the one or more values ​​of the blood flow parameter and one or more low concentration measures during a first time period; and generating, by the computing system, second user interface data corresponding to the one or more values ​​of the blood flow parameter and one or more low concentration measures during a second time period.

Claims

1. A method comprising: a computing system including one or more computing devices, each including a processor and a memory, obtaining contrast data indicative of a first characteristic within the brain of an individual based on the presence of a contrast agent within the brain of the individual; the computing system identifying an amount of time elapsed since the onset of a physiological condition corresponding to one or more blood vessels of the brain of the individual; the computing system determining one or more threshold levels of one or more blood flow parameters related to the blood vessels of the brain of the individual and based on the amount of time and the contrast data, wherein the one or more threshold levels indicate damage to the brain of the individual; the computing system determining that one of the one or more blood flow parameters is at least one of the one or more threshold levels related to the blood vessels of the brain of the individual; the computing system generating an image indicative of at least one region of the brain of the individual including damaged tissue based on the computing system and based on the one blood flow parameter; A method comprising the above steps.

2. The one or more threshold levels include a first threshold level of a blood flow parameter corresponding to a first time period elapsed since the onset of the physiological condition, The method according to claim 1, wherein the one or more threshold levels include a second threshold level of the blood flow parameter corresponding to a second time period elapsed since the onset of the physiological condition.

3. the computing system determining that the amount of time corresponds to a first amount of time; the computing system determining that damage has occurred in the region of the brain of the individual based on the blood flow parameter being at least the first threshold level, the method according to claim 2.

4. The method according to claim 2, wherein the second threshold level corresponds to a value of the blood flow parameter that is greater than the first threshold level.

5. The method according to claim 2, wherein the blood flow parameter includes cerebral blood flow.

6. The method according to claim 1, wherein the computing system identifies the blood flow parameter by analyzing voxel information of the contrast data.

7. the computing system obtaining additional contrast data indicative of a second characteristic within the brain of the individual in the absence of the contrast agent within the brain of the individual; Determining, by the computing system and based on the amount of time, one or more weights of one or more parameters of a model for determining damage to the tissue of the individual's brain; The method of claim 1, comprising: **Claim 8** Identifying, by the computing system, values of parameters of the model by analyzing intensity values of some voxels of the additional contrast data; Determining, by the computing system and based on the amount of time, weights of the parameters; The method of claim 7, comprising: **Claim 9** Causing, by the computing system, the model executed using the weights of the parameters to identify a low concentration measurement of a region of the individual's brain; Determining, by the computing system, that the low concentration measurement is at least at a low concentration threshold level; Determining, by the computing system, that tissue of the region of the individual's brain is damaged; The method of claim 8, comprising: **Claim 10** The method of claim 1, wherein the physiological condition is an obstruction of blood flow to at least one region of the individual's brain. **Claim 11** A system comprising: One or more hardware processors; One or more non-transitory computer-readable storage media comprising computer-readable instructions; wherein when the computer-readable instructions are executed by the one or more hardware processors, the one or more hardware processors are caused to perform operations, the operations comprising: Obtaining contrast data indicative of a first characteristic within the individual's brain based on the presence of a contrast agent within the individual's brain; Identifying an amount of time elapsed since the onset of a physiological condition corresponding to one or more blood vessels of the individual's brain; Determining one or more threshold levels of one or more blood flow parameters related to the blood vessels of the individual's brain and based on the amount of time and the contrast data, wherein the one or more threshold levels indicate damage to the individual's brain; Determining that one of the one or more blood flow parameters is at least one of the one or more threshold levels related to the blood vessels of the individual's brain; Generating an image indicative of at least one region of the individual's brain including damaged tissue based on the blood vessels of the individual's brain; A system comprising: **Claim 12** The one or more non-transitory computer-readable storage media include additional computer-readable instructions that, when executed by the one or more hardware processors, cause the one or more hardware processors to perform additional operations, the additional operations including obtaining additional contrast data indicative of a second characteristic within the individual's brain in the absence of the contrast agent within the individual's brain; determining, based on the amount of time, one or more weights of one or more parameters of a model for determining damage to the tissue of the individual's brain; The system of claim 11, comprising. **Claim 13** The system of claim 12, wherein the contrast data is captured using a first contrast technique that performs one or more perfusion-based computed tomography (CT) contrast techniques, and the additional contrast data is captured using a second contrast technique that performs one or more non-contrast CT contrast techniques. **Claim 14** The system of claim 12, wherein the contrast data is captured using a first contrast technique that performs one or more computed tomography angiography (CTA) contrast techniques, and the additional contrast data is captured using a second contrast technique that performs one or more non-contrast CT contrast techniques. **Claim 15** The one or more non-transitory computer-readable storage media include additional computer-readable instructions that, when executed by the one or more hardware processors, cause the one or more hardware processors to perform additional operations, the additional operations including generating an image including a first overlay corresponding to the region of the individual's brain and a second overlay corresponding to an additional region of the individual's brain The system of claim 12, comprising.

Images