Method and analysis device for determining carotid artery compressibility by using ultrasound image

Abstract

This method for determining carotid artery compressibility by using an ultrasound image comprises steps in which an analysis device: receives an ultrasound image including a carotid artery region of a specific patient; classifies, by frame, a blood vessel region in the ultrasound image by using a segmentation model trained in advance; determines a final mask for a target frame in units of two consecutive frames with respect to a plurality of frames included in the ultrasound image; and calculates a carotid artery compressibility value for the target frame on the basis of the final mask.

Description

Method and analysis device for determining carotid artery compressibility using ultrasound imaging

[0001] The technology described below relates to a technique for predicting the prognosis of cardiac arrest patients. Specifically, the technology described below relates to a technique for predicting the prognosis of cardiac arrest patients using ultrasound imaging.

[0002] Point-of-care ultrasound (POCUS) is used for immediate patient diagnosis. It is also widely used in the management of cardiac arrest patients. POCUS-CAC (Carotid Artery Compressibility) is a technique for rapidly assessing carotid artery compressibility (CAC) under probe-induced compression. POCUS-CAC uses ultrasound to assess carotid artery compressibility (CAC) under probe-induced compression, reducing pulse assessment time to less than 2 seconds on average.

[0003] Recently, AI-based diagnostic and predictive techniques have been actively researched in the medical field. Quantifying carotid artery compressibility is essential for AI model learning and inference. However, during probe-induced compression, the carotid artery flattens, and its boundary is difficult to distinguish due to movement during compression. Consequently, building an AI model to assess carotid artery compressibility is challenging.

[0004] The technique described below aims to provide a method for predicting return of spontaneous circulation (ROSC) in cardiac arrest patients by quantifying carotid artery compressibility.

[0005] A method for determining carotid artery compressibility using an ultrasound image includes a step of an analysis device receiving an ultrasound image including a carotid artery region of a specific patient, a step of the analysis device segmenting a blood vessel region using a segmentation model learned in advance for each frame in the ultrasound image, a step of the analysis device determining a final mask for a target frame among two consecutive frames among a plurality of frames included in the ultrasound image, and a step of the analysis device calculating a carotid artery compressibility value for the target frame based on the final mask.

[0006] An analysis device for calculating carotid artery compressibility includes an interface device for receiving an ultrasound image including a carotid artery region of a specific patient, a storage device for storing a segmentation model for distinguishing a blood vessel region from the input image, and a calculation device for distinguishing a blood vessel region from the ultrasound image by frame using the segmentation model, determining a final mask for a target frame among two consecutive frames for a plurality of frames included in the ultrasound image, and calculating a carotid artery compressibility value for the target frame based on the final mask.

[0007] The technology described below improves the quality of emergency medical care by rapidly assessing a patient's condition using a vascular region detection algorithm for carotid artery compressibility.

[0008] Figure 1 is an example of a system for predicting ROSC in cardiac arrest patients.

[0009] Figure 2 is an example of a process for quantifying POCUS-CAC.

[0010] Figure 3 is an example of an ultrasound data set using the POCUS-CAC approach.

[0011] Figure 4 is an example of the results of blood vessel classification and CAC value quantification using the conventional model and the proposed technique.

[0012] Figure 5 is an example of classifying cardiac arrest and ROSC based on carotid artery CAC.

[0013] Figure 6 is an example of an analysis device that predicts a patient's ROSC.

[0014] The technology described below is susceptible to various modifications and embodiments. Specific embodiments are illustrated and described in detail in the drawings. However, this does not limit the technology described below to specific embodiments, and it should be understood that all modifications, equivalents, and alternatives fall within the spirit and scope of the technology described below.

[0015] Terms such as first, second, A, and B may be used to describe various components, but these components are not limited by these terms and are used solely to distinguish one component from another. For example, without departing from the scope of the technology described below, a first component may be referred to as a second component, and similarly, a second component may also be referred to as a first component. The term "and / or" includes a combination of a plurality of related described items or any of a plurality of related described items.

[0016] As used herein, the singular expressions should be understood to include the plural expressions unless the context clearly dictates otherwise, and the term "comprises" and the like should be understood to mean the presence of a described feature, number, step, operation, component, part, or combination thereof, but not to exclude the presence or addition of one or more other features, numbers, steps, operations, components, parts, or combinations thereof.

[0017] Before going into a detailed description of the drawings, it should be made clear that the division of components in this specification is merely a division based on the main function of each component. In other words, two or more components described below may be combined into a single component, or a single component may be further subdivided into two or more components with more detailed functions. In addition to its own main function, each component described below may additionally perform some or all of the functions of other components, and of course, some of the main functions of each component may be exclusively performed by other components.

[0018] Additionally, in performing a method or method of operation, each process constituting the method may occur in a different order than the stated order, unless the context clearly indicates a specific order. That is, each process may occur in the same order as the stated order, may be performed substantially simultaneously, or may be performed in the opposite order.

[0019] The technology described below relates to a technique for evaluating the condition of a patient in cardiac arrest using ultrasound imaging.

[0020] The technique described below is a technique that quantifies carotid artery compressibility (CAC) using carotid artery ultrasound and classifies ROSC based on the quantified carotid artery compressibility.

[0021] The technique described below segments the carotid artery region using a deep learning model that identifies regions of interest in an image. The model that identifies the region of interest is called a segmentation model. In the description below, the segmentation model can be any of a variety of models. For example, the segmentation model can be any of a variety of models, such as U-net or FCN (Fully Convolutional Networks).

[0022] In the following description, the analysis device uses a segmentation model to segment the carotid artery region from an ultrasound image and assesses carotid artery compressibility to classify or predict ROSC. The analysis device is a computing device that can analyze a patient's cardiac arrest status using image processing and a learned model. The analysis device can be implemented using various devices capable of data processing. For example, the analysis device can be implemented as a PC, a networked server, a smart device, or a chipset embedded with a dedicated program.

[0023] Fig. 1 is an example of a system (100) for predicting ROSC in a cardiac arrest patient. In Fig. 1, the analysis device is illustrated as a user terminal (130) and a server (140).

[0024] An ultrasound scanner (110) generates an ultrasound image of a specific area of ​​a patient. The ultrasound scanner (110) scans the patient's neck area to generate an image of an area including the carotid artery, which is the area of ​​interest.

[0025] The user terminal (130) receives the patient's ultrasound video. The user terminal (120) may be any of various types of devices, such as a smart device, a PC, or a wearable device. In some cases, the user terminal (130) may be an integrated device connected to the ultrasound scanner (110). The user terminal (130) may receive the patient's ultrasound video via a wired or wireless network.

[0026] The user terminal (130) segments the carotid artery region from an ultrasound video using a pre-trained segmentation model and calculates carotid artery compressibility. The user terminal (130) predicts the patient's return-to-severity score (ROSC) based on the calculated carotid artery compressibility. The user terminal (130) can classify the patient's condition as either cardiac arrest or ROSC based on the calculated carotid artery compressibility. User A can check the ROSC prediction result output from the user terminal (130).

[0027] The server (140) receives an ultrasound video of a patient. The server (140) may receive the ultrasound video through an ultrasound scanner (110) or a separate computer device.

[0028] The server (140) segments the carotid artery region from an ultrasound image using a pre-trained segmentation model and calculates carotid artery compressibility. Based on the calculated carotid artery compressibility, the server (140) predicts the patient's return-to-severity-score (ROSC). Based on the calculated carotid artery compressibility, the server (140) can classify the patient's condition as either cardiac arrest or ROSC. User A can access the server (140) via a user terminal and check the patient's ROSC prediction results.

[0029] Figure 2 is an example of a process (200) for quantifying POCUS-CAC.

[0030] The analysis device receives an ultrasound video containing a carotid artery region (210). The analysis device can perform analysis on a plurality of consecutive frames constituting the ultrasound video.

[0031] The analysis device can preprocess the input ultrasound image (220). The preprocessing can be performed on each frame. The analysis device can convert a 3D (dimensional) ultrasound image into a 2D image. Hereinafter, the analysis will be described based on the 2D image. Medical ultrasound images can include metadata such as text. Therefore, the analysis device can remove metadata from the ultrasound image. In addition, the analysis device can crop unnecessary black areas from the ultrasound image. Afterwards, the analysis device can perform histogram normalization on the ultrasound frame.

[0032] The segmentation model must be built using training data in advance. The researcher removed metadata and black areas in the frames during the process of preparing the training data set. The researcher used 11,958 ultrasound images obtained from 37 subjects collected from the affiliated institution as the data set. The ultrasound images included in the data set were stored as 8-bit bitmaps with pixel intensities in the range [0, 255]. Figure 3 shows an example of an ultrasound data set using the POCUS-CAC approach. Figure 3(A) is an ultrasound image (with compression) of a patient in cardiac arrest, Figure 3(B) is an ultrasound image (with compression) of a patient in ROSC, and Figure 3(C) is an ultrasound image (without compression) of a patient in ROSC. The blue arrows in Figure 3 indicate the probe-induced compression area. The areas shown in the image in Figure 3 are the CA (carotid artery), IJV (internal jugular vein), and SCM (sternocleidonastoid muscle).

[0033] The researchers used the TransUNet architecture (Jieneng Chen et al., TransUNet: Transformers Make Strong Encoders for Medical Image Segmentation, arXiv:2102.04306, 2021) as a segmentation model. TransUNet has shown high performance in segmenting specific organs in medical images. TransUNet includes a convolutional neural network (CNN) that generates a feature map of the input image, a transformer layer that divides the feature map into patches and then receives linearly projected embedding data, and a U-net that converts the features output from the transformer layer to 2D and then performs upsampling. The researchers built a segmentation model using training data to distinguish regions of interest (vascular regions). The vascular regions include the carotid artery (CA) region and the internal jugular vein (IJV) region.

[0034] Of course, segmentation models can also utilize other types of models.

[0035] The analysis device segments the blood vessel region in each frame using the learned segmentation model (230). Fig. 2 illustrates frame 1, frame t-1, and frame t among multiple frames as examples.

[0036] The analysis device then calculates the degree of carotid artery compressibility (CAC) in each frame (240).

[0037] The process of calculating the CAC value is illustrated on the right side of Fig. 2. The analysis device calculates the CAC value for one frame based on two consecutive frames. For convenience of explanation, the explanation is based on frames t-1 and t. The analysis device receives the mask of frame t-1 and the mask of frame t (241). Here, the mask corresponds to the blood vessel area distinguished from the ultrasound frame using the segmentation model.

[0038] The analysis device detects the mask outline for the vascular region using the mask in the frame (242). The analysis device can identify the mask edge based on changes in pixel intensity or color using an image processing technique. The vascular region may include the carotid artery (CA) and internal jugular vein (IJV). The mask outline can define the mask area.

[0039] The segmentation model should only distinguish between the carotid artery (CA) and the internal jugular vein (IJV), but noise in the ultrasound input may lead to incorrectly identifying the region. The segmentation model may also classify the IJV and muscle regions as the same group, extracting a larger mask containing the IJV and muscle regions. In Figure 2, frame t is an example of a mask segmentation that includes the IJV and muscle regions.

[0040] The analysis device optimizes the mask using the contours detected from two frames (243). The optimization process corresponds to the process of minimizing incorrect blood vessel masks. Two consecutive frames have a high probability of having masks of similar shapes. The analysis device can determine whether the mask of the current frame is appropriate based on the mask of the previous frame. The analysis device can compare the mask area of ​​frame t-1 with the mask area of ​​frame t, and if the overlapping area is less than a threshold value, the non-overlapping area can be deleted from the mask area of ​​frame t. For example, the analysis device can optimize the mask area of ​​the current frame if the overlapping area of ​​the masks of consecutive frames is less than 50% of the total mask area. Figure 2 shows an example in which the analysis device deletes non-overlapping areas from the initial mask area of ​​frame t. In other words, the analysis device calculates only the area in the mask area of ​​the current frame that overlaps with the mask of the previous frame as the final mask (optimal mask).

[0041] The analysis device calculates the CAC value based on the optimal mask (244). The analysis device can calculate the CAC value based on the contour area of ​​the optimal mask. The vascular area has an elliptical shape due to compression. The analysis device can calculate the CAC value based on the eccentricity of the ellipse encompassing the vascular area mask. In this case, the ellipse can be the smallest shape that includes the outer edge of the mask. The eccentricity (e) can be calculated as shown in Mathematical Formula 1 below.

[0042]

[0043] In mathematical equation 1, a is the length of the semi-major axis, and b is the length of the semi-minor axis. Eccentricity indicates the degree of compression of an ellipse. Eccentricity e has a value between 0 and 1, and a higher eccentricity indicates a flatter shape. In other words, the closer the eccentricity e is to 1, the more compressed it is.

[0044] The analysis device can determine the eccentricity of the ellipse encompassing the vascular area mask as the CAC value. In some cases, the analysis device can also process and calculate the CAC value based on the eccentricity value of the ellipse encompassing the vascular area mask.

[0045] The analysis device can derive the CAC value and the optimal mask for the current frame (frame t) through the CAC calculation process. The optimal mask for frame t is used as a mask for optimizing the mask for frame t+1. For example, the mask for frame t-1 used for optimizing frame t in FIG. 2 may be the result of optimization in the previous step.

[0046] The researchers compared the results of quantifying the blood vessel region and CAC value through the process described in Fig. 2 with the results of segmentation using a conventional segmentation model. The conventional models were constructed using U-Net, Mask R-CNN, and TransUNeT and the results were compared. In Fig. 4, GT refers to the ground truth. Table 1 shows the accuracy of the blood vessel region segmented by the segmentation model. Table 1 shows the mean ± standard deviation of the Dice Similarity Coefficient (DSC) and Intersection over Union (IOU) for the carotid artery (CA) and the internal jugular vein (IJV), respectively. Fig. 4 shows examples of the results of blood vessel segmentation and CAC value quantification using the conventional model and the proposed technique. In Fig. 4, the blue arrow indicates a situation where the blood vessel is compressed by the ultrasound probe. The proposed technique in Table 1 and Fig. 4 is a technique that calculates the CAC value through the process described in Fig. 2.

[0047] modelDSC (CA)DSC (IJV)IOU (CA)IOU (IJV)U-Net0.89 ± 0.130.54 ± 0.450.83 ± 0.160.50 ± 0.44Mask R-CNN0.90 ± 0.190.81 ± 0.370.87 ± 0.190.79 ± 0.37TransUNeT0.90 ± 0.120.63 ± 0.430.82 ± 0.150.59 ± 0.42Proposed technique0.90 ± 0.110.81 ± 0.340.84 ± 0.150.89 ± 0.18

[0048] Mask R-CNN showed a high DSC of 0.90, but there were instances where it failed to predict blood vessel regions when the probe was compressed, as shown by the red arrow in Figure 4. TransUNet also performed well overall, but there were instances where it incorrectly predicted muscle regions as blood vessels, as shown by the yellow arrow in Figure 4. In summary, the proposed technique achieved the highest performance in terms of blood vessel region identification and accuracy.

[0049] The analysis device can classify or predict a patient's ROSC in real time based on the calculated CAC value. The CAC value increases with ultrasound probe compression. The analysis device can classify ROSC if the compressibility value of the internal jugular vein (IJV) under the probe-induced compression becomes 1 and the CAC value does not increase above (is below) a certain threshold. The researcher experimentally set the threshold for classifying ROSC and cardiac arrest to 0.94. The analysis device classified the current patient status as ROSC if the compressibility value of the internal jugular vein (IJV) in consecutive frames was 1 and the CAC value was below the threshold (e.g., 0.94). The analysis device classified the current patient status as ROSC if the compressibility value of the internal jugular vein (IJV) was 1 and the CAC value was above the threshold (e.g., 0.94) in a certain section of consecutive frames. Figure 5 shows an example of classifying cardiac arrest and ROSC based on carotid artery CAC. Figure 5(A) is an example of a CAC value classified as ROSC, and Figure 5(B) is an example of a CAC value classified as cardiac arrest.

[0050] Table 1 below shows the performance of classifying ROSC based on the vascular region distinguished using the conventional model and the proposed technique.

[0051] ModelPrecision(Precision), Recall(Recall), Accuracy(ACC), F1-scoreU-Net0.950.950.950.95Mask R-CNN1.000.890.940.94TransUNet0.980.930.940.95Proposed technique0.980.950.960.97

[0052] As mentioned above, Mask R-CNN showed high DSC, but it sometimes failed to predict blood vessel regions during probe compression, resulting in relatively low ACC and low recall. TransUNet sometimes mispredicted muscles as blood vessels, resulting in a lower ACC of 0.94 compared to the proposed method. The proposed method achieved the highest accuracy in the ROSC classification task, with an accuracy of 0.96 and an F1-score of 0.97.

[0053] Figure 6 illustrates an example of an analysis device that predicts a patient's ROSC. The analysis device (300) corresponds to the aforementioned analysis device (130 or 140 of Figure 1). The analysis device (300) may be physically implemented in various forms. For example, the analysis device (300) may take the form of a computer device such as a PC, a network server, a data processing chipset, or the like.

[0054] The analysis device (300) may include a storage device (310), a memory (320), a computing device (330), an interface device (340), a communication device (350), and an output device (360).

[0055] The storage device (310) can store ultrasound images of a specific patient. In this case, the ultrasound images include the carotid artery area.

[0056] The storage device (310) can store a program or code for image processing.

[0057] The storage device (310) can store a learned segmentation model that distinguishes blood vessel regions.

[0058] The storage device (310) can store commands or program codes for calculating CAC values ​​based on the distinguished blood vessel area and classifying ROSC based on the CAC values.

[0059] The storage device (310) can store the results of vascular region classification, CAC value, and ROSC classification for a specific patient.

[0060] The memory (320) can store data and information generated during the process of the analysis device (300) predicting ROSC for a specific patient.

[0061] The interface device (340) is a device that receives certain commands and data from the outside.

[0062] The interface device (340) can receive ultrasonic video from a physically connected input device or an external storage device.

[0063] The interface device (340) can also transmit the CAC value and ROSC classification results to an external object.

[0064] Meanwhile, the interface device (340) is meant to include a device configuration that transmits data received from a communication device (350) into the analysis device (300).

[0065] A communication device (350) refers to a configuration that receives and transmits certain information through a wired or wireless network.

[0066] The communication device (350) can receive ultrasonic video from an external object.

[0067] The communication device (350) may also transmit the CAC value and ROSC classification results to an external object such as a user terminal or server.

[0068] The output device (360) is a device that outputs certain information. The output device (360) can output interfaces, ultrasound images, blood vessel areas, CAC values, ROSC classification results, etc. required for the data processing process.

[0069] The computational device (330) can preprocess real images in a uniform manner in response to the learning process of the 3D latent diffusion model. For example, the computational device (330) can truncate 3D images to a uniform size.

[0070] The computational unit (330) can uniformly preprocess an ultrasound video frame by frame. The computational unit (330) can perform metadata removal, black area removal, and histogram normalization in an ultrasound image.

[0071] The computational device (330) can segment a blood vessel region in each ultrasound frame using a segmentation model. The blood vessel region may include the carotid artery (CA) and the internal jugular vein (IJV).

[0072] The calculation device (330) can calculate the optimal mask and CAC value for a specific frame through the process described in Fig. 2. The calculation device (330) calculates the CAC value of the current frame based on the masks of two consecutive frames.

[0073] The computational device (330) detects contours in the mask of frame t-1 and the mask of frame t. The computational device (330) checks the overlapping area between the mask area of ​​frame t-1 and the mask area of ​​frame t. If the overlapping area between the mask area of ​​frame t and the mask area of ​​frame t-1 is less than a threshold value (e.g., 50%), the computational device (330) can remove the overlapping area from the mask area of ​​frame t (optimal mask generation).

[0074] The calculation device (330) can calculate the eccentricity by setting an ellipse encompassing the blood vessel area based on the optimal mask. The calculation device (330) can calculate the CAC value based on the eccentricity of the ellipse. The calculation device (330) uses the optimal mask of frame t as a mask for mask optimization of the next frame.

[0075] The computing device (330) can classify the current patient's ROSC status based on the CAC values ​​of consecutive frames. That is, the computing device (330) can classify the current patient's status as ROSC if the compressibility value of the internal jugular vein (IJV) in consecutive frames is 1 and the CAC value is below a threshold value. Conversely, the computing device (330) can classify the current patient's status as cardiac arrest if the compressibility value of the internal jugular vein (IJV) in consecutive frames is 1 and the CAC value is above.

[0076] The computing device (330) may be a device such as a processor, AP, or chip embedded with a program that processes data and performs certain operations.

[0077] Additionally, the ultrasound image analysis method, POCUS-CAC evaluation technique, or ROSC prediction technique described above may be implemented as a program (or application) including an executable algorithm that can be executed on a computer. The program may be stored and provided on a temporary or non-transitory computer-readable medium.

[0078] A non-transitory readable medium refers to a medium that stores data semi-permanently and can be read by a device, rather than a medium that stores data for a short period of time, such as a register, cache, or memory. Specifically, the various applications or programs described above may be stored and provided on a non-transitory readable medium, such as a CD, DVD, hard disk, Blu-ray disk, USB, memory card, ROM (read-only memory), PROM (programmable read only memory), EPROM (Erasable PROM, EPROM), EEPROM (Electrically EPROM), or flash memory.

[0079] Temporarily readable media refers to various types of RAM, such as Static RAM (SRAM), Dynamic RAM (DRAM), Synchronous DRAM (SDRAM), Double Data Rate SDRAM (DDR SDRAM), Enhanced SDRAM (ESDRAM), Synchronous DRAM (Synclink DRAM, SLDRAM), and Direct Rambus RAM (DRRAM).

[0080] The present embodiment and the drawings attached to the present specification only clearly illustrate a part of the technical idea included in the above-described technology, and it is obvious that all modified examples and specific embodiments that can be easily inferred by a person skilled in the art within the scope of the technical idea included in the specification and drawings of the above-described technology are included in the scope of the rights of the above-described technology.

Claims

1. A step in which an analysis device receives an ultrasound image including a carotid artery area of ​​a specific patient; A step in which the above analysis device segments a blood vessel region using a segmentation model learned in advance for each frame in the ultrasound image; A step in which the above analysis device determines a final mask for a target frame among two consecutive frames included in the ultrasound image; and A method for determining carotid artery compressibility using an ultrasound image, the method comprising the step of calculating a carotid artery compressibility value for the target frame based on the final mask by the above analysis device.

2. In paragraph 1, The step of determining the final mask is The above analysis device comprises a step of receiving a blood vessel area mask of frame t-1 and a blood vessel area mask of frame t from the plurality of frames; The above analysis device detects a blood vessel area mask of the frame t-1 and an outline of the blood vessel area mask of the frame t; The above analysis device comprises a step of identifying an overlapping area in the blood vessel area mask area of ​​the frame t-1 and the blood vessel area mask of the frame t based on the detected outline; A method for determining carotid artery compressibility using an ultrasound image, wherein the analysis device comprises a step of generating a final mask of the frame t by removing a non-overlapping area from the vascular region mask of the frame t if the overlapping area occupies less than a threshold value in the mask region of the frame t.

3. In paragraph 1, A method for determining carotid artery compressibility using an ultrasound image in which the above analysis device sets an ellipse encompassing the final mask and calculates the carotid artery compressibility value based on the eccentricity of the ellipse.

4. In paragraph 1, A method for determining carotid artery compressibility using ultrasound images, further comprising a step of classifying the cardiac arrest state of the specific patient based on the carotid artery compressibility values ​​for the plurality of frames.

5. In paragraph 4, A method for determining carotid artery compressibility using an ultrasound image in which the analysis device classifies the condition of the specific patient as ROSC (Return Of Spontaneous Circulation) when the compressibility value of the internal jugular vein is 1 for the plurality of frames and the carotid artery compressibility value is less than a threshold value.

6. An interface device for receiving an ultrasound image including the carotid artery area of ​​a specific patient; A storage device for storing a segmentation model that distinguishes a blood vessel region in an input image; and An analysis device for calculating carotid artery compressibility, comprising a calculation device that segments a blood vessel region frame by frame using the segmentation model in the ultrasound image, determines a final mask for a target frame among two consecutive frames for a plurality of frames included in the ultrasound image, and calculates a carotid artery compressibility value for the target frame based on the final mask.

7. In paragraph 6, The above-mentioned calculation device receives a blood vessel area mask of frame t-1 and a blood vessel area mask of frame t from the plurality of frames, checks an overlapping area in the blood vessel area mask of frame t-1 and the blood vessel area mask of frame t, and if the overlapping area is less than a threshold value in the blood vessel area mask of frame t, calculates the carotid artery compressibility by removing a non-overlapping area from the blood vessel area mask of frame t to generate a final mask of frame t.

8. In paragraph 6, The above calculation device is an analysis device that calculates the carotid artery compressibility by setting an ellipse encompassing the final mask and calculating the carotid artery compressibility value based on the eccentricity of the ellipse.

9. In paragraph 6, The above calculation device is an analysis device that calculates carotid artery compressibility to classify the cardiac arrest state of the specific patient based on the carotid artery compressibility values ​​for the plurality of frames.

10. In paragraph 9, The above calculation device is an analysis device that calculates carotid artery compressibility to classify the condition of the specific patient as ROSC (Return Of Spontaneous Circulation) when the compressibility value of the internal jugular vein is 1 for the plurality of frames and the carotid artery compressibility value is less than a threshold value.

Images