Systems and methods for diagnosing sarcopenia

Abstract

A system and method for assessing sarcopenia in a patient is provided, comprising receiving longitudinal ultrasound images acquired by longitudinal scanning of a muscle of interest, receiving transverse ultrasound images acquired by transverse scanning of the muscle of interest using ultrasound imaging, segmenting the longitudinal and transverse ultrasound images to identify a target muscle region, determining a global attenuation of the target muscle region, respectively, identifying regions of interest within the target muscle region in the longitudinal and transverse ultrasound images, respectively, determining a local attenuation coefficient of the region of interest, determining a global acoustic velocity within the target muscle region in the longitudinal and transverse scan ultrasound images, respectively, determining a local acoustic velocity within the region of interest within the target muscle region, respectively, and determining a level of sarcopenia in the patient based on the global attenuation and acoustic velocity within the target muscle region and the local attenuation coefficient and acoustic velocity within the region of interest.

Description

[Technical Field]

[0001] Sarcopenia is a chronic disease characterized by the progressive loss of skeletal muscle mass, composition, and function in the human body. In general, in sarcopenic patients, muscle fibers are replaced by intramuscular adipose tissue (fat content). The prevalence of sarcopenia varies between 10% and 27% worldwide, with severe sarcopenia occurring between 2% and 9% worldwide. Additional reported risk factors independently associated with sarcopenia include family conditions, lifestyle, physical inactivity, poor nutrition and dental health, and diseases (e.g., osteoporosis, metabolic diseases), although age may be the most important risk factor. In particular, the likelihood of developing sarcopenia correlates with many cardiometabolic risk factors, particularly diabetes, hypertension, dyslipidemia, motor neuron loss, underactive neuromuscular junctions, hormonal status, pro-inflammatory cytokines, reduced mitochondrial function, abnormal myokine production, and weight loss accompanied by decreased appetite. Screening and early diagnosis of sarcopenia are major medical and social challenges resulting from the generally aging world population. In particular, ultrasound imaging has emerged as an important tool for measuring muscle quantity and quality, and offers the unique advantage of being a non-invasive modality that can be used at the bedside to provide serial measurements. [Background technology]

[0002] Ultrasound has been used to assess muscle function, including texture analysis and speed of sound (SoS) ultrasound. For example, Figures 1A and 1B show transverse ultrasound images of the proximal third of the rectus femoris muscle of two female patients with similar ages and body mass indexes (BMIs). In particular, image 101 of Figure 1A shows the rectus femoris muscle of the first patient, who is 38 years old and has a BMI of 24.2 and is an avid runner, while image 102 of Figure 1B shows the rectus femoris muscle of the second patient, who is 42 years old and has a BMI of 25.1 and does not engage in any physical activity. In both patients, the first patient's cross-sectional area (CSA) index was 7.575 cm. 2 and the second patient's CSA was 7.351 cm 2Nevertheless, increased muscle belly echogenicity can be observed in the second patient due to fatty infiltration detectable in image 102. Summary of the Invention [Problem to be solved by the invention]

[0003] However, the conventional workflow for assessing sarcopenia using ultrasound imaging is based on simple parameters such as muscle thickness and echogenicity, as well as the CSA index, which poses several important methodological issues, leading to conflicting results from previous studies. For example, the conventional workflow for muscle ultrasound examination is complex and time-consuming, lacking advanced tools to help reduce workload. Furthermore, there is no clear standardization in clinical practice. [Means for solving the problem]

[0004] In an exemplary embodiment, 1. A method for diagnosing sarcopenia in a patient, the method comprising: receiving longitudinal ultrasound images acquired by longitudinal scanning of a muscle of interest in the patient using ultrasound imaging; receiving a transverse ultrasound image obtained by transverse scanning of a muscle of interest in the patient using ultrasound imaging; segmenting the longitudinal and transverse ultrasound images, respectively, to identify target muscle regions of the muscle of interest in the longitudinal and transverse ultrasound images; determining the overall attenuation in the target muscle region between boundaries in the longitudinal and transverse ultrasound images, respectively; identifying regions of interest in a target muscle region in the longitudinal ultrasound image and the transverse ultrasound image, respectively; determining local attenuation coefficients of regions of interest in the target muscle region, respectively; determining an overall acoustic velocity in the target muscle region between boundaries in the longitudinal and transverse ultrasound images, respectively; determining local acoustic velocities in regions of interest in the target muscle regions, respectively; determining a level of sarcopenia in the patient based on global attenuation and global acoustic velocity in the target muscle region and local attenuation coefficients and local acoustic velocities in the region of interest; A method is provided, comprising:

[0005] In another exemplary embodiment, 1. A system for diagnosing sarcopenia in a patient, the system comprising: A user interface; a processor in communication with the user interface; a non-transitory memory that, when executed by the processor, causes the processor to: receiving longitudinal ultrasound images acquired by longitudinal scanning of a muscle of interest in the patient using ultrasound imaging; receiving a transverse ultrasound image obtained by transverse scanning of a muscle of interest in the patient using ultrasound imaging; segmenting the longitudinal and transverse ultrasound images, respectively, to identify target muscle regions of the muscle of interest in the longitudinal and transverse ultrasound images; determining the overall attenuation in the target muscle region between boundaries in the longitudinal and transverse ultrasound images, respectively; identifying regions of interest in the target muscle region in the longitudinal ultrasound image and the transverse ultrasound image, respectively; determining local attenuation coefficients of regions of interest in the target muscle region, respectively; determining an overall acoustic velocity in the target muscle region between boundaries in the longitudinal and transverse ultrasound images, respectively; determining local acoustic velocities in regions of interest in the target muscle regions, respectively; determining a level of sarcopenia in the patient based on global attenuation and global acoustic velocity in the target muscle region and local attenuation coefficients and local acoustic velocities in the region of interest; a non-transitory memory for storing instructions for executing the The system further includes a display configured to display an indication of the level of sarcopenia.

[0006] In another exemplary embodiment, 1. A non-transitory computer-readable medium storing instructions for diagnosing sarcopenia in a patient, the instructions, when executed by a processor, causing the processor to: receiving longitudinal ultrasound images acquired by longitudinal scanning of a muscle of interest in the patient using ultrasound imaging; receiving a transverse ultrasound image obtained by transverse scanning of a muscle of interest in the patient using ultrasound imaging; segmenting the longitudinal and transverse ultrasound images, respectively, to identify target muscle regions of the muscle of interest in the longitudinal and transverse ultrasound images; determining the overall attenuation in a target muscle region between boundaries in the longitudinal and transverse ultrasound images, respectively; identifying regions of interest in a target muscle region in the longitudinal ultrasound image and the transverse ultrasound image, respectively; determining local attenuation coefficients of regions of interest in the target muscle region, respectively; determining an overall acoustic velocity in the target muscle region between boundaries in the longitudinal and transverse ultrasound images, respectively; determining local acoustic velocities in regions of interest in the target muscle regions, respectively; determining a level of sarcopenia in the patient based on global attenuation and global acoustic velocity in the target muscle region and local attenuation coefficients and local acoustic velocities in the region of interest; Displaying an indicator of the level of sarcopenia. A non-transitory computer-readable medium is provided that is configured to cause the execution of

[0007] The illustrative embodiments are best understood from the following detailed description when read in conjunction with the accompanying drawings. It is emphasized that the various features are not necessarily drawn to scale. In fact, dimensions may be arbitrarily increased or decreased for clarity of discussion. Where applicable and practical, like reference numerals refer to like elements. [Brief explanation of the drawings]

[0008] [Figure 1A] Shown is a lateral ultrasound image of the rectus femoris muscle of the first patient who exercises regularly and has low fatty infiltration of the muscle. [Figure 1B] A transverse ultrasound image of a second patient without exercise and with high fatty infiltration of the muscle is shown. [Figure 2] FIG. 1 is a simplified block diagram of a system for performing sarcopenia assessment using ultrasound images, according to a representative embodiment. [Figure 3A] 10 shows superior muscle segmentation results with a deep learning-based segmentation algorithm, according to a representative embodiment. [Figure 3B] 10 illustrates moderate to poor muscle segmentation results with a deep learning-based segmentation algorithm, according to a representative embodiment. [Figure 4] FIG. 1 is a flow diagram illustrating a method for assessing sarcopenia in a subject using ultrasound images, according to a representative embodiment. DETAILED DESCRIPTION OF THE INVENTION

[0009] In the following detailed description, for purposes of explanation and not limitation, exemplary embodiments disclosing specific details are set forth to provide a thorough understanding of embodiments in accordance with the present teachings. Descriptions of known systems, devices, materials, methods of operation, and methods of manufacture may be omitted so as to avoid obscuring the description of the exemplary embodiments. Nevertheless, systems, devices, materials, and methods within the purview of those skilled in the art are within the scope of the present teachings and may be used in accordance with the exemplary embodiments. It is to be understood that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting. Defined terms are in addition to the technical and scientific meaning of the defined terms as commonly understood and accepted in the art of the present teachings.

[0010] Terms such as "first," "second," and "third" may be used herein to describe various components or constituent elements, but it should be understood that these components or constituent elements should not be limited by these terms. These terms are used only to distinguish one component or constituent element from another. Thus, a first element or component described below could be referred to as a second element or component without departing from the teachings of the inventive concept.

[0011] The terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting. As used in this specification and the appended claims, the singular forms of "a," "an," and "the" are intended to include both the singular and the plural unless the context clearly dictates otherwise. Additionally, the terms "comprises," "comprises," and / or similar terms specify the presence of stated features, elements, and / or components, but do not exclude the presence or addition of one or more other features, elements, components, and / or groups thereof. As used herein, the term "and / or" includes any and all combinations of one or more of the associated listed items.

[0012] Unless otherwise specified, when a component or component is said to be "connected," "coupled," or "adjacent" to another component or component, it is understood that the component or component may be directly connected or coupled to the other component or component, or there may be intervening components or components. That is, these and similar terms encompass the case where one or more intermediate components or components may be used to connect the two components or components. However, when a component or component is said to be "directly connected" to another component or component, this only encompasses the case where the two components or components are connected to each other without any intermediate or intervening components or components.

[0013] Thus, the present disclosure is intended to derive one or more of the advantages, as specifically set forth below, through one or more of its various aspects, embodiments, and / or specific features or subcomponents. For purposes of explanation and not limitation, exemplary embodiments disclosing specific details are described to provide a thorough understanding of embodiments in accordance with the present teachings. However, other embodiments consistent with the present disclosure that depart from the specific details disclosed herein remain within the scope of the appended claims. Furthermore, descriptions of well-known devices and methods may be omitted so as not to obscure the description of the exemplary embodiments. Such methods and devices are within the scope of the present disclosure.

[0014] Generally, various embodiments described herein provide systems and methods for implementing an improved workflow for sarcopenia assessment using ultrasound images. The workflow includes automatically segmenting an ROI of the muscle under investigation from the ultrasound image, measuring global attenuation for the entire muscle portion from the ROI for both longitudinal and transverse scans, measuring local attenuation within a small ROI for the muscle portion, measuring acoustic velocity in both longitudinal and transverse scans, and assessing sarcopenia grading from a combination of measured ultrasound parameters.

[0015] This workflow reduces the time and effort required to perform a musculoskeletal (MSK) examination because data acquisition is the same as that required for a conventional ultrasound examination, where muscles appear clearly in longitudinal and transverse scans. The workflow also simplifies muscle quality assessment, especially for relatively inexperienced users (e.g., physicians and sonographers), thereby improving results and increasing user confidence. This workflow enables early diagnosis of patients at high risk for sarcopenia by using quantitative muscle ultrasound to reduce effort and examination time.

[0016] In patients with sarcopenia, muscle fibers are replaced by intramuscular adipose tissue (fat content), resulting in a decrease in acoustic velocity. This is due to the lower acoustic velocity in adipose tissue (1440 m / s) compared to muscle (1585 m / s). Correspondingly, the attenuation of ultrasound in ultrasound images also decreases with the proportion of intramuscular fat. Because muscle tissue is anisotropic, meaning that certain physical properties have different values ​​when measured in different directions, according to embodiments herein, ultrasound parameters from both longitudinal and transverse scans are integrated into a workflow for sarcopenia assessment.

[0017] FIG. 2 is a simplified block diagram of a system for performing sarcopenia assessment using ultrasound images, according to a representative embodiment.

[0018] 2 , system 100 includes a workstation 105 for implementing and / or managing the processes described herein for assessing sarcopenia in a subject 150 using ultrasound images from an ultrasound imaging device 140. Workstation 105 includes one or more processors represented by processor 120, one or more memories represented by memory 130, a user interface 122, and a display 124. Processor 120 communicates with ultrasound imaging device 140 via an imaging interface (not shown). Ultrasound imaging device 140 includes an ultrasound transducer probe 145 operable by a user to acquire musculoskeletal (MSK) ultrasound images of a portion of subject 150. Ultrasound transducer probe 145 can be operated manually by a user, automatically by a robot under the control of a robotic controller (not shown), or a combination of both.

[0019] The memory 130 stores instructions executable by the processor 120. When executed, the instructions cause the processor 120 to implement one or more processes for performing a sarcopenia assessment of the subject 150 using ultrasound images acquired by the ultrasound imaging device 140. The ultrasound images may be provided from the ultrasound imaging device 140 in real time or near real time during a scanning procedure, or may be retrieved from storage following the scanning procedure. For purposes of illustration, the memory 130 is shown to include software modules each including instructions executable by the processor 120 corresponding to relevant capabilities of the system 100.

[0020] Processor 120 represents one or more processing devices and may be implemented using any combination of hardware, software, firmware, hardwired logic circuitry, or combinations thereof, such as a general-purpose computer, a central processing unit (CPU), a digital signal processor (DSP), a graphical processing unit, a computer processor, a microprocessor, a state machine, a programmable logic device, a field-programmable gate array (FPGA), an application-specific integrated circuit (ASIC), or a combination thereof. Any processor or processing unit herein may include multiple processors, parallel processors, or both. Multiple processors may be included in or combined with a single device or multiple devices. As used herein, the term "processor" encompasses any electronic component capable of executing a program or machine-executable instructions. A processor may also refer to a collection of processors within a single computer system or distributed across multiple computer systems, such as in a cloud-based or other multi-site application. A program comprises software instructions executed by one or more processors, which may be within the same computing device or distributed across multiple computing devices.

[0021] Memory 130 may include main memory and / or static memory, which may communicate with each other and with processor 120 via one or more buses. Memory 130 may be implemented by any number, type, and combination of random access memory (RAM) and read-only memory (ROM), for example, and may store various types of information, such as software algorithms, artificial intelligence (AI) machine learning models, and computer programs, all of which are executable by processor 120. The various types of ROM and RAM may include any number, type, and combination of computer-readable storage media, such as disk drives, flash memory, electrically field-programmable gate array read-only memory (EPROM), electrically erasable field-programmable gate array read-only memory (EEPROM), registers, hard disks, removable disks, tape, compact disk read-only memory (CD-ROM), digital versatile disks (DVDs), floppy disks, Blu-ray disks, universal serial bus (USB) drives, or any other form of storage media. Memory 130 is a tangible storage medium that stores data and executable software instructions and is non-transitory while the software instructions are stored therein. As used herein, the term "non-transitory" should be interpreted as a characteristic of a state that persists over a period of time, rather than as a permanent characteristic of a state. The term "non-transitory" specifically negates fleeting characteristics, such as the characteristics of a carrier wave or signal or other formation that exists only temporarily at any place at any time. Memory 130 may store software instructions and / or computer-readable code that enable the performance of various functions. Memory 130 may be secure and / or encrypted, or non-secure and / or unencrypted.

[0022] System 100 may also include a database 112 for storing information that can be used by the various software modules in memory 130. For example, database 112 may include image data from previously acquired ultrasound images of subject 150 and / or other similarly positioned subjects. The stored image data may be used, for example, to train AI machine learning models, such as neural network models, as discussed below. Database 112 may be implemented, for example, by any number, type, and combination of RAM and ROM. The various types of ROM and RAM may include any number, type, and combination of computer-readable storage media, such as disk drives, flash memory, EPROM, EEPROM, registers, hard disks, removable disks, tape, CD-ROM, DVD, floppy disks, Blu-ray disks, USB drives, or any other form of storage media known in the art. Database 112 comprises a tangible storage medium for storing data and executable software instructions, and the data and software instructions are non-transitory while stored. Database 112 may be secure and / or encrypted, or may be unsecure and / or unencrypted. For purposes of illustration, database 112 is shown as a separate storage medium, but it will be understood that it may be combined with and / or included in memory 130 without departing from the scope of the present teachings.

[0023] Processor 120 may include or have access to an artificial intelligence (AI) engine, which may be implemented as software that provides artificial intelligence (e.g., neural network models) and applies machine learning as described herein. The AI ​​engine may reside in any of a variety of components in addition to or other than processor 120, such as, for example, memory 130, an external server, and / or the cloud. If the AI ​​engine is implemented in the cloud, such as in a data center, for example, the AI ​​engine may be connected to processor 120 via the Internet or other communications network using one or more wired and / or wireless connections.

[0024] The user interface 122 is configured to provide a user with information and data output by the processor 120, the memory 130, and / or the ultrasound imaging device 140, and / or receive information and data input by the user. That is, the user interface 122 allows a user to input data and control or manipulate aspects of the processes described herein, and also allows the processor 120 to indicate the effects of the user's input, which may include control or manipulation of the ultrasound transducer probe 145. All or a portion of the user interface 122 may be implemented by a graphical user interface (GUI), such as the GUI 128 viewable on the display 124, described below. The user interface 122 may include one or more interface devices, such as, for example, a mouse, keyboard, trackball, joystick, microphone, video camera, touchpad, touchscreen, voice or gesture recognition captured by the microphone or video camera, etc.

[0025] Display 124 may be, for example, a monitor such as a computer monitor, television, liquid crystal display (LCD), organic light-emitting diode (OLED), flat panel display, solid state display, or cathode ray tube (CRT) display, or an electronic whiteboard. Display 124 includes a screen 126 for viewing ultrasound images of subject 150 and includes a GUI 128 that conveys to the user the degree of sarcopenia (if any) shown in the images, along with various features described herein, and allows the user to interact with the displayed images and features. In one embodiment, ultrasound imaging device 140 may include a separate, dedicated display for acquiring ultrasound images, which is also represented by display 124.

[0026] Referring to the memory 130, various modules store data and sets of instructions executable by the processor 120 to perform sarcopenia assessment, as described above. The longitudinal scan image module 131 is configured to receive and process ultrasound images (referred to as "longitudinal ultrasound images") of a muscle of interest 155 in the subject 150 provided from a longitudinal scan by the ultrasound imaging device 140. The transverse scan image module 132 is configured to receive and process ultrasound images (referred to as "transverse ultrasound images") of the muscle of interest 155 from a transverse scan by the ultrasound imaging device 140. The longitudinal scan generally provides a length and depth perspective of the muscle of interest 155 from the ultrasound transducer probe 145, and the transverse scan generally provides a width and depth perspective of the muscle of interest 155 from the ultrasound transducer probe 145. The transverse scan may be provided at a perspective that is substantially perpendicular to the perspective of the longitudinal scan, as indicated by the arrow in FIG. 2 . However, it is understood that the vertical and horizontal scans can differ from each other by about 45 degrees to about 135 degrees without departing from the scope of the present teachings.

[0027] The longitudinal and transverse ultrasound images may be displayed on the display 124. The longitudinal and transverse ultrasound images may be received in real time or near real time from the ultrasound imaging device 140 during a simultaneous imaging session of the subject 150. The display of real time images, among other things, allows a user to visualize the anatomical structures of the subject 150 while manipulating the ultrasound transducer probe 145. Alternatively, or additionally, the longitudinal and transverse ultrasound images may be previously acquired images retrieved from a storage device (e.g., database 112), acquired during a previous imaging session.

[0028] The segmentation module 133 receives longitudinal ultrasound images from the longitudinal scan image module 131 and transverse ultrasound images from the transverse scan image module 132 of the same muscle of interest 155 and performs segmentation of the muscle of interest 155 in each of the received longitudinal and transverse ultrasound images. Segmentation may be performed, for example, using known conventional image segmentation methods or deep learning-based segmentation algorithms, as will be apparent to those skilled in the art. In one embodiment, the deep learning-based segmentation algorithm may be, for example, a convolutional neural network (CNN) such as U-Net. Alternatively, or in addition, at least some of the segmentation and / or editing for automated segmentation may be performed manually by a user. In various embodiments, segmentation of longitudinal ultrasound images may be performed using the same or different methods or algorithms as those used to perform segmentation of transverse ultrasound images. The segmentation provides a first target muscle region 151 of the muscle of interest 155 from the longitudinal ultrasound image and a second target muscle region 153 of the muscle of interest 155 from the transverse ultrasound image.

[0029] In one embodiment, the above-described deep learning-based segmentation algorithm can be implemented as an N-fold voting mechanism, where N is a positive integer, running N different trained predictive models. In one embodiment, N may be, for example, 10 or less, because an N greater than 10 can lead to undesirably long processing times. According to the N-fold voting mechanism, the training set and validation set are equally segmented into N folds. Then, to train N predictive models using these N different data selections, each fold is sequentially selected as a validation set, while the remaining N-1 folds provide the training set. All predictive models are used to predict muscle segmentation, and the final result is voted for, for example, by averaging the predicted segmentations to obtain the final result. Because the user has the opportunity to edit the boundary of the target muscle region, as discussed below, one predictive model (N = 1) is typically sufficient to run the deep learning-based segmentation algorithm. For example, a 10-fold mechanism (N = 10) generally slightly improves the performance of deep learning-based segmentation algorithms, but requires more time and computational power to run.

[0030] In one embodiment, the segmentation module 133 also determines the quality of the muscle segmentation of the longitudinal and transverse ultrasound images, and whether the quality is acceptable. For example, the segmentation module 133 can automatically determine the quality of the muscle segmentation by identifying which portions of the segmented ultrasound images have unusable region segmentation quality, for example, due to contours that are not smooth and / or have odd concave shapes. Alternatively, the segmentation module 133 can display the segmentation results on the display 124 to allow the user to judge the quality of the muscle segmentation. In this case, the segmentation module 133 receives a quality indication from the user via the user interface 122 after the user views the muscle segmentation.

[0031] If the quality of the muscle segmentation is unacceptable, the segmentation module 133 can perform fine-tuning of the muscle segmentation. This can be done by region re-segmentation. Alternatively, if the quality of the muscle segmentation is unacceptable, the user can manually edit the muscle segmentation using the user interface 122. For example, the user can perform fine-tuning by manually adjusting the boundaries of the target muscle region and / or small ROI in the longitudinal and / or transverse ultrasound images for the muscle of interest 155.

[0032] 3A illustrates excellent segmentation results from a deep learning-based segmentation algorithm according to a representative embodiment. Referring to FIG. 3A, a target muscle region 301 is shown with a ground truth boundary 302 and a segmented (predicted) boundary 303 following muscle segmentation. A comparison of the ground truth boundary 302 and the segmented boundary 303 shows excellent alignment, with a Dice similarity coefficient exceeding 0.98. Therefore, only minor fine-tuning is required by the user to adjust the segmented boundary 303 of the target muscle region 301 to substantially match the ground truth boundary.

[0033] In comparison, FIG. 3B illustrates relatively moderate to poor muscle segmentation results using a deep learning-based segmentation algorithm according to a representative embodiment. Referring to FIG. 3B, a target muscle region 311 is shown with a ground truth boundary 312 and a segmented (predicted) boundary 313 following muscle segmentation. A comparison of the ground truth boundary 312 and the segmented boundary 313 shows a fairly significant misalignment, with a Dice similarity coefficient of approximately 0.72. The misalignment may be due, for example, to the limited data size for the deep learning-based segmentation algorithm. Therefore, user editing is required to correct the mismatched boundary at the bottom of the target muscle region 311.

[0034] Referring again to FIG. 2 , the global attenuation module 134 is configured to automatically calculate the global attenuation (dB / cm) for each of the first and second target muscle regions 151 and 153 for the longitudinal and lateral scans of the muscle of interest 155. The global attenuation is determined within the boundaries of the first and second target muscle regions 151 and 153 using any known technique. For example, a first global attenuation in the first target muscle region 151 from the longitudinal scan may be determined between an upper boundary and a lower boundary, and a second global attenuation in the second target muscle region 153 from the lateral scan may be determined between a left boundary and a right boundary. The boundaries of the longitudinal and lateral scans are used to average the global attenuation for each of the entire first and second target muscle regions 151 and 153 within a single frame of the longitudinal and lateral ultrasound image. Generally, the value of the global attenuation decreases as the amount of fat relative to muscle increases. In other words, the denser the tissue (ie, the greater the proportion of muscle), the greater the overall attenuation of ultrasound through the tissue.

[0035] The local attenuation module 135 is configured to automatically calculate local attenuation coefficients (dB / cm / MHz) for small first and second regions of interest (ROIs) 152 and 154 within the first and second target muscle regions 151 and 153 for longitudinal and transverse scans, respectively, using any known technique. The local attenuation module 135 enables selection of a first ROI 152 within the first target muscle region 151 from a single image frame and a second ROI 154 within the second target muscle region 153 from a single image frame. Selection of the first and second ROIs 152 and 154 can be performed automatically by segmentation or manually by a user via the user interface 122. For example, each of the first and second ROIs 152 and 154 may be selected by a user via the user interface 122, or may be automatically set to a specific ROI in a central region of the target muscle regions 151 and 153, respectively. For example, central regions of target muscle regions 151 and 153 may be automatically identified and cropped from the respective muscle segmentations to provide first and second ROIs 152 and 154. Once the first and second ROIs 152 and 154 have been identified, a first local attenuation coefficient may be determined for the first ROI 152 from the longitudinal scan, and a second local attenuation coefficient may be determined for the second ROI 154 from the transverse scan. The local attenuation coefficients may be calculated for the first and second ROIs 152 and 154 directly or by averaging the corresponding attenuation maps.

[0036] For example, the method for determining the local attenuation coefficient for fatty liver quantification may be adapted to determine the local attenuation coefficient of a targeted muscle region. For example, a linear ultrasound probe may be used to image the targeted muscle region instead of a curved ultrasound probe used to image the liver. Generally, the average local attenuation coefficient of a normal liver is about 0.567 dB / cm / MHz, the average local attenuation coefficient of a mild fatty liver is about 0.659 dB / cm / MHz, and the average local attenuation coefficient of a severe fatty liver is about 0.789 dB / cm / MHz. The relative values ​​between the averages of the normal, mildly fatty, and severely fatty targeted muscle regions are similar.

[0037] The global acoustic velocity module 136 is configured to automatically calculate the global acoustic velocity (m / s) of ultrasound waves in each of the first and second target muscle regions 151 and 153 for longitudinal and transverse scans of the muscle of interest 155. Global acoustic velocity values ​​are determined within the boundaries of the first and second target muscle regions 151 and 153, respectively, by, for example, calculating the propagation velocity of sound waves through tissue using any known technique. For example, a first global acoustic velocity in the first target muscle region 151 from the longitudinal scan may be determined by averaging acoustic velocity values ​​through the upper and lower boundaries, and a second global acoustic velocity in the second target muscle region 153 from the transverse scan may be determined by averaging acoustic velocity values ​​through the left and right boundaries. Generally, the global acoustic velocity value decreases as the amount of fat relative to muscle increases. In other words, the acoustic velocity of ultrasound through fat tissue is approximately 1440 m / s, while the acoustic velocity of ultrasound through muscle tissue is approximately 1585 m / s, so the denser the tissue (i.e., the greater the ratio of muscle to fat), the faster the velocity of ultrasound through the tissue.

[0038] The local acoustic velocity module 137 is configured to automatically calculate, using any known technique, the local acoustic velocity (m / s) of ultrasound in each of the first and second ROIs 152 and 154 within the first and second target muscle regions 151 and 153 for longitudinal and transverse scans of the muscle of interest, respectively. For example, a first local acoustic velocity may be determined for the first ROI 152, and a second local acoustic velocity may be determined for the second ROI 154. Each of the first and second ROIs 152 and 154 may be selected by a user or may be set to a specific ROI in a central region of each of the first and second target muscle regions 151 and 153. The local acoustic velocity may be calculated, for example, for the first and second ROIs 152 and 154 directly or by averaging their respective attenuation maps.

[0039] The sarcopenia grading module 138 is configured to determine the level (grade) of sarcopenia (e.g., normal sarcopenia, mild sarcopenia, moderate sarcopenia, severe sarcopenia) in the muscle of interest 155, thereby grading the sarcopenia. For example, normal sarcopenia (i.e., little or no sarcopenia) may receive a grade of 0 or A, mild sarcopenia may receive a grade of 1 or B, moderate sarcopenia may receive a grade of 2 or C, and severe sarcopenia may receive a grade of 3 or D. Different levels are defined by corresponding predetermined thresholds. The level of sarcopenia may be calculated by applying a regression model to a combination of eight parameters output by the global attenuation module 134, the local attenuation module 135, the global acoustic velocity module 136, and the local acoustic velocity module 137, where four of the parameters are from the longitudinal ultrasound images and four of the parameters are from the transverse ultrasound images. The regression model determines weights for each of the eight parameters and combines the weighted parameters to output a level of sarcopenia based on a predefined threshold.

[0040] In various embodiments, the regression model may be a machine learning algorithm, such as a neural network model, such as an artificial neural network (ANN), a recurrent neural network (RNN), or a CNN. Thus, the regression model may be trained using previously acquired data providing the values ​​of the eight parameters and the corresponding sarcopenia levels of past patients. The training data may be stored, for example, in database 112. The training also compares the relative effects of the eight parameters on the resulting sarcopenia level to determine the extent to which the eight parameters influence the final level of sarcopenia. Based on these relative effects, the regression model determines and assigns respective weights to the eight parameters. Current sarcopenia results and corresponding parameters can be added to this training database to continue improving the accuracy of the regression model. That is, the training minimizes similarity loss to improve the accuracy of the regression model as the training dataset grows, as will be apparent to those skilled in the art.

[0041] The weights applied by the regression model can account for various different relationships between parameters depending on their corresponding impact on the overall determination of sarcopenia. For example, global parameters (i.e., global attenuation and global acoustic velocity) from each of the longitudinal and transverse ultrasound images may be weighted more heavily than local parameters (i.e., local attenuation coefficient and local acoustic velocity) from each of the longitudinal and transverse ultrasound images. Also, measurements of attenuation (i.e., global attenuation and local attenuation coefficient) in each of the longitudinal and transverse ultrasound images may be weighted slightly more heavily than measurements of acoustic velocity (i.e., global and local acoustic velocity) in each of the longitudinal and transverse ultrasound images.

[0042] In various embodiments, all or part of the process provided by the machine learning regression model may be implemented, for example, by the AI ​​engine described above. The neural network model may be trained using ground truth longitudinal and transverse ultrasound images of target muscle regions and ROIs in the muscles of interest of multiple patients. The longitudinal and transverse ultrasound images are associated with the corresponding determined global attenuation, local attenuation coefficient, global acoustic velocity, and local acoustic velocity, as well as the resulting sarcopenia grade. Thus, during the training phase, the neural network model learns the appropriate sarcopenia level from retrospective data t applied to the current parameters.

[0043] The resulting sarcopenia level indicator may be provided to a user, for example, via the display 124, to diagnose whether sarcopenia is present in the subject 150 and, if so, the level of sarcopenia (e.g., mild to severe) in the subject 150. In one embodiment, the sarcopenia level may be provided to a reporting module (not shown), which formats the information from the sarcopenia grading module 138 and causes the formatted information to be displayed on the display 124 via the GUI 128. In addition to the sarcopenia level, the displayed information may include all or some of the parameters used by the sarcopenia grading module 138 to determine the level of sarcopenia, including global attenuation, local attenuation coefficient, global acoustic velocity, and local acoustic velocity from the longitudinal and transverse ultrasound images. Also, in one embodiment, the longitudinal and transverse images themselves may be displayed along with the sarcopenia grade and any additional information. In light of the diagnosis, appropriate medical care may then be provided.

[0044] 4 is a flow diagram of a method for assessing sarcopenia in a subject using ultrasound images, according to a representative embodiment. The method may be implemented, for example, using instructions stored in memory 130 and executable by processor 120 in system 100. In one embodiment, the method may be performed by an online version of an ultrasound imaging device to improve MSK ultrasound workflow, for example, for assessing muscle function in subjects at high risk of developing severe sarcopenia.

[0045] 4, longitudinal ultrasound images are received by a processor (e.g., processor 120) in block S411, where the longitudinal ultrasound images are acquired by longitudinal scanning of a patient's muscle of interest using ultrasound imaging. The longitudinal scan is performed after pre-configuring the ultrasound imaging device for MSK imaging. The longitudinal ultrasound images may be received in real time or near real time from an ultrasound imaging device (e.g., ultrasound imaging device 140) during an ultrasound imaging session, or may be retrieved from a database (e.g., database 112) that stores images acquired during a previous ultrasound imaging session.

[0046] In block S412, a lateral ultrasound image is received by a processor, the lateral ultrasound image having been acquired by a lateral scan of a patient's muscle of interest using ultrasound imaging. The lateral ultrasound image may be received in real time or near real time by an ultrasound imaging device during an ultrasound imaging session, or may be retrieved from a database.

[0047] In block S413, the longitudinal and transverse ultrasound images are segmented to identify target muscle regions of the subject's muscles, i.e., the longitudinal ultrasound image is segmented to identify a first target muscle region and the transverse ultrasound image is segmented to identify a second target muscle region.

[0048] In block S414, the global attenuation in the target muscle regions is determined between the boundaries of the target muscle regions in the longitudinal and transverse ultrasound images, respectively: a first global attenuation in a first target muscle region is determined between upper and lower boundaries defining the first target muscle region, and a second global attenuation in a second target muscle region is determined between left and right boundaries defining the second target muscle region.

[0049] In block S415, regions of interest (ROIs) are identified in the target muscle regions in the longitudinal and transverse ultrasound images, respectively. That is, a first ROI is identified in the first target muscle region, and a second ROI is identified in the second target muscle region. Each of the first and second ROIs may be manually identified by a user or automatically identified by a processor. For example, the first ROI may be identified in a central region of the first target muscle region, and the second ROI may be identified in a central region of the second target muscle region, which may be automatically identified and cropped from the segmented target muscle regions.

[0050] In block S416, local attenuation coefficients of the ROIs are determined in each target muscle region, i.e., a first local attenuation coefficient is determined for the first ROI in the first target muscle region, and a second local attenuation coefficient is determined for the second ROI in the second target muscle region.

[0051] In block S417, overall acoustic velocities within the targeted muscle regions are determined between boundaries of the targeted muscle regions in the longitudinal and transverse ultrasound images, respectively: a first overall acoustic velocity in a first targeted muscle region is determined between upper and lower boundaries defining the first targeted muscle region, and a second overall acoustic velocity in a second targeted muscle region is determined between left and right boundaries defining the second targeted muscle region.

[0052] In block S418, local acoustic velocities within ROIs within the target muscle regions are determined within the longitudinal and transverse ultrasound images, respectively, i.e., a first local acoustic velocity within a first ROI is determined within the first target muscle region, and a second local acoustic velocity within a second ROI is determined within the second target muscle region.

[0053] In block S419, the level of sarcopenia in the subject is determined (graded) based on the global attenuation and global acoustic velocity in the first and second target muscle regions and the local attenuation coefficients and local acoustic velocities throughout the first and second ROIs. The level of sarcopenia may be determined by applying a regression model, as described above, to the global attenuation and global acoustic velocity in the first and second target muscle regions and the local attenuation coefficients and local acoustic velocities of each of the first and second ROIs.

[0054] In block S420, the level of sarcopenia is reported to the user via a display or GUI as a diagnosis of the patient. The user can then determine an appropriate course of medical treatment accordingly. The reporting may further include displaying representative image frames for each of the longitudinal scan and the transverse scan. Also, one or more of the first global attenuation, local attenuation coefficient, global acoustic velocity, and local acoustic velocity may be displayed with the image frame for the longitudinal scan, and one or more of the second global attenuation, local attenuation coefficient, global acoustic velocity, and local acoustic velocity may be displayed with the image frame for the transverse scan.

[0055] According to various embodiments of the present disclosure, the methods described herein may be implemented using a hardware computer system executing a software program stored on a non-transitory storage medium. Furthermore, in exemplary, non-limiting embodiments, implementations may include distributed processing, component / object distributed processing, and parallel processing. A virtual computer system process may implement one or more of the methods or functions described herein, and the processors described herein may be used to support a virtual processing environment.

[0056] While assessing sarcopenia using ultrasound images has been described with reference to exemplary embodiments, it should be understood that the words used are words of description and illustration, rather than words of limitation. Changes may be made within the purview of the appended claims, as presently stated and as amended, without departing from the scope and spirit of the embodiments. While assessing sarcopenia using ultrasound images has been described with reference to particular means, materials, and embodiments, it is not intended to be limited to the details disclosed, but rather to cover all functionally equivalent structures, methods, and uses, as fall within the scope of the appended claims.

[0057] The descriptions of the embodiments described herein are intended to provide a general understanding of the structure of various embodiments. The illustrations are not intended to serve as a complete description of all of the elements and features of the present disclosure described herein. Many other embodiments may be apparent to those skilled in the art upon reviewing the present disclosure. Other embodiments may be utilized and derived from the present disclosure, such that structural and logical substitutions and changes may be made without departing from the scope of the present disclosure. Additionally, the illustrations are merely representational and may not be drawn to scale. Certain proportions in the figures are exaggerated, while other proportions are minimized. Therefore, the present disclosure and the drawings should be considered illustrative and not limiting.

[0058] One or more embodiments of the present disclosure may be individually and / or collectively referred to herein by the term "invention" merely for convenience and without any intention to intentionally limit the scope of the present application to any particular invention or inventive concept. Furthermore, although specific likenesses have been illustrated and described herein, it should be understood that any subsequent configuration designed to achieve the same or similar purpose may be substituted for the specific likeness shown. The present disclosure is intended to cover any and all subsequent adaptations or modifications of the various embodiments. Combinations of the above embodiments, and other embodiments not specifically described herein, will become apparent to those skilled in the art upon reviewing the description.

[0059] This Abstract of the Disclosure is provided for purposes of compliance with 37 CFR §1.72(b) and is submitted with the understanding that it will not be used to interpret or limit the scope or meaning of the claims. Moreover, in the foregoing Detailed Description, various features may be grouped together or described in a single embodiment for the purpose of streamlining the disclosure. This disclosure should not be interpreted as reflecting an intention that the claimed embodiments require more features than are expressly recited in each claim. Rather, as the following claims reflect, inventive subject matter may be directed to less than all features of any of the disclosed embodiments. Accordingly, the following claims are hereby incorporated into the Detailed Description, with each claim standing on its own as defining separately claimed subject matter.

[0060] The foregoing description of the disclosed embodiments is provided to enable those skilled in the art to practice the concepts described in this disclosure. Accordingly, the subject matter disclosed above should be considered illustrative and not limiting, and the appended claims are intended to encompass all such modifications, enhancements, and other embodiments that fall within the true spirit and scope of the present disclosure. Accordingly, to the maximum extent permitted by law, the scope of the present disclosure should be determined by the broadest permissible interpretation of the following claims and their equivalents, and should not be limited or constrained by the foregoing detailed description.

Claims

1. 1. A method for diagnosing sarcopenia in a patient, the method comprising: receiving longitudinal ultrasound images acquired by longitudinal scanning of a muscle of interest in the patient using ultrasound imaging; receiving a transverse ultrasound image obtained by transverse scanning of a muscle of interest in the patient using ultrasound imaging; segmenting the longitudinal and transverse ultrasound images, respectively, to identify a target muscle region of a muscle of interest in the longitudinal and transverse ultrasound images; determining the overall attenuation in the target muscle region between boundaries in the longitudinal and transverse ultrasound images, respectively; identifying regions of interest in a target muscle region in the longitudinal ultrasound image and the transverse ultrasound image, respectively; determining local attenuation coefficients of regions of interest in the target muscle region, respectively; determining an overall acoustic velocity in the target muscle region between boundaries in the longitudinal and transverse ultrasound images, respectively; determining local acoustic velocities in regions of interest in the target muscle regions, respectively; determining a level of sarcopenia in the patient based on global attenuation and global acoustic velocity in the target muscle region and local attenuation coefficients and local acoustic velocities in the region of interest; A method comprising:

2. 10. The method of claim 1, wherein segmenting the longitudinal and transverse ultrasound images comprises applying a deep learning based segmentation algorithm to image data from the longitudinal and transverse images.

3. 3. The method of claim 2, wherein segmenting the longitudinal and transverse ultrasound images further comprises fine-tuning or editing at least one inconsistent boundary of at least one of the target muscle regions.

4. 2. The method of claim 1, wherein determining the level of sarcopenia in the patient comprises applying a regression model to global attenuation and global acoustic velocity in the target muscle region and local attenuation coefficients and local acoustic velocity in the region of interest.

5. The step of identifying a region of interest in the target muscle region comprises: receiving a selection of a local attenuation coefficient of the region of interest from a user at a user interface; or automatically selecting the region of interest in a central region of the target muscle region; 2. The method of claim 1, comprising:

6. The method of claim 5 , wherein determining the local attenuation coefficient of the region of interest comprises averaging attenuation maps from the region of interest, respectively.

7. The method of claim 1 , wherein determining the overall acoustic velocity through the target muscle region comprises averaging the acoustic velocities through each boundary of the target muscle region.

8. The method of claim 7 , wherein determining the local acoustic velocity through the region of interest comprises averaging acoustic velocity maps from each of the regions of interest.

9. displaying a representative image frame for each of the longitudinal scan and the lateral scan; displaying one or more of the global attenuation, the global attenuation coefficient, the global acoustic velocity, and a local acoustic velocity for the longitudinal ultrasound image together with a representative image frame for the longitudinal scan; and displaying one or more of the global attenuation, the local attenuation coefficient, the global acoustic velocity, and a local acoustic velocity for the transverse ultrasound image together with a representative image frame for the transverse scan. The method of claim 1 further comprising:

10. reporting a diagnosis result corresponding to the level of sarcopenia The method of claim 1 further comprising:

11. 1. A system for diagnosing sarcopenia in a patient, the system comprising: A user interface; a processor in communication with the user interface; a non-transitory memory that, when executed by the processor, causes the processor to: receiving longitudinal ultrasound images acquired by longitudinal scanning of a muscle of interest in the patient using ultrasound imaging; receiving a transverse ultrasound image obtained by transverse scanning of a muscle of interest in the patient using ultrasound imaging; segmenting the longitudinal and transverse ultrasound images, respectively, to identify a target muscle region of a muscle of interest in the longitudinal and transverse ultrasound images; determining the overall attenuation in the target muscle region between boundaries in the longitudinal and transverse ultrasound images, respectively; identifying regions of interest in the target muscle region in the longitudinal ultrasound image and the transverse ultrasound image, respectively; determining local attenuation coefficients of regions of interest in the target muscle region, respectively; determining an overall acoustic velocity in the target muscle region between boundaries in the longitudinal and transverse ultrasound images, respectively; determining local acoustic velocities in regions of interest in the target muscle regions, respectively; determining a level of sarcopenia in the patient based on global attenuation and global acoustic velocity in the target muscle region and local attenuation coefficients and local acoustic velocities in the region of interest; a non-transitory memory storing instructions for executing the a display configured to display an indicator of the level of sarcopenia; A system having:

12. 12. The system of claim 11, wherein the instructions cause the processor to perform the step of segmenting the longitudinal ultrasound image and the transverse ultrasound image by applying a deep learning based segmentation algorithm to image data from the longitudinal and transverse images.

13. 13. The system of claim 12, wherein segmenting the longitudinal and transverse ultrasound images further comprises fine-tuning or editing at least one inconsistent boundary of at least one of the target muscle regions.

14. 12. The system of claim 11, wherein the instructions cause the processor to perform a step of determining a level of sarcopenia in the patient by applying a regression model to global attenuation and global acoustic velocity in the target muscle region and local attenuation coefficients and local acoustic velocity in the region of interest.

15. The instructions direct the processor to: receiving a selection of a local attenuation coefficient of the region of interest from a user at the user interface; or automatically selecting the region of interest in a central region of the target muscle region; 12. The system of claim 11, wherein the step of identifying a region of interest in the target muscle region is performed by:

16. 16. The system of claim 15, wherein the instructions cause the processor to perform the step of determining a local attenuation coefficient for the region of interest by averaging attenuation maps from the region of interest individually.

17. 12. The system of claim 11, wherein the instructions cause the processor to perform the step of determining an overall acoustic velocity through the target muscle region by averaging acoustic velocities through each boundary of the target muscle region.

18. 20. The system of claim 17, wherein determining the local acoustic velocity through the region of interest comprises averaging acoustic velocity maps from each of the regions of interest.

19. The display further comprises: displaying a representative image frame for each of the longitudinal scan and the lateral scan; displaying one or more of the global attenuation, the global attenuation coefficient, the global acoustic velocity, and the local acoustic velocity for the longitudinal ultrasound image along with a representative image frame for the longitudinal scan; displaying one or more of the global attenuation, the local attenuation coefficient, the global acoustic velocity, and the local acoustic velocity for the transverse ultrasound image along with a representative image frame for the transverse scan; The system of claim 11 configured to:

20. 1. A non-transitory computer-readable medium storing instructions for diagnosing sarcopenia in a patient, the instructions, when executed by a processor, causing the processor to: receiving longitudinal ultrasound images acquired by longitudinal scanning of a muscle of interest in the patient using ultrasound imaging; receiving a transverse ultrasound image obtained by transverse scanning of a muscle of interest in the patient using ultrasound imaging; segmenting the longitudinal and transverse ultrasound images, respectively, to identify a target muscle region of a muscle of interest in the longitudinal and transverse ultrasound images; determining the overall attenuation in a target muscle region between boundaries in the longitudinal and transverse ultrasound images, respectively; identifying regions of interest in a target muscle region in the longitudinal ultrasound image and the transverse ultrasound image, respectively; determining local attenuation coefficients of regions of interest in the target muscle region, respectively; determining an overall acoustic velocity in the target muscle region between boundaries in the longitudinal and transverse ultrasound images, respectively; determining local acoustic velocities in regions of interest in the target muscle regions, respectively; determining a level of sarcopenia in the patient based on global attenuation and global acoustic velocity in the target muscle region and local attenuation coefficients and local acoustic velocities in the region of interest; Displaying an indicator of the level of sarcopenia.

1. A non-transitory computer-readable medium configured to cause execution of

Images