Method and system for automated analysis of cervical elastography

Abstract

The invention relates to a method and system for automated analysis of cervical elastography. Various methods (300) and systems for cervical elastography analysis are provided. In one embodiment, the method (300, 500, 400) comprises: acquiring an ultrasound scene (202) comprising a plurality of frames (226); generating a first lost image (206) for the plurality of frames (226); selecting a background (224) and a foreground (222) of the ultrasound scene (612) based on the first loss image (206); and generating a softness ratio (214) from the selected background (224) and foreground (222).

Description

Technical Field

[0001] Embodiments of the subject matter disclosed herein relate to an automated method of analyzing cervical elastography, such as for use in assessing the likelihood of preterm birth. Background Art

[0002] Clinical ultrasound is an imaging modality that uses ultrasound waves to detect the internal structures of a patient's body and produce corresponding images. An ultrasound probe, which includes multiple transducer elements, transmits ultrasound pulses, which are reflected or transmitted back, refracted, or absorbed by structures in the body. The ultrasound probe then receives the reflected echoes, which are processed into an image. For example, medical imaging devices such as ultrasound imaging devices can be used to obtain images of a patient's heart, uterus, liver, lungs, and various other anatomical parts.

[0003] One application of clinical ultrasound is elastography. Elastography attempts to assess the mechanical properties of tissues, specifically their stiffness and elasticity, by analyzing how tissues respond to external forces or physiological processes. One of the key parameters in elastography is strain, which quantifies the percentage of tissue deformation that occurs when static or oscillatory compression is applied. The strain metric measures tissue deformation under pressure, where softer tissues deform more easily, resulting in larger strain values. Conversely, stiffer tissues exhibit reduced deformation, resulting in lower strain values.

[0004] Cervical elastography is one method used to predict preterm birth. During cervical elastography, the sonographer selects the cervical tissue of interest and the reference tissue of the ultrasound image, known as the foreground and background, respectively, and calculates strain values ​​to obtain the softness ratio of the cervical tissue. The softness ratio is used to evaluate cervical changes that may indicate preterm birth. Shear wave imaging analysis is an alternative method for predicting preterm birth. However, the shear wave method utilizes high energy push pulses that can be sent to an area close to the fetal head. Summary of the invention

[0005] In one embodiment, a method includes acquiring an ultrasound scene including a plurality of frames; generating a first loss image for the plurality of frames; selecting a background and a foreground of the ultrasound scene based on the first loss image; and generating a softness ratio from the selected background and foreground.

[0006] It should be understood that the above brief description is provided to introduce in a simplified form selected concepts that are further described in the detailed description. It is not meant to identify key features or essential features of the claimed subject matter, the scope of which is uniquely defined by the claims that follow the detailed description. Furthermore, the claimed subject matter is not limited to implementations that solve any disadvantages noted above or in any part of this disclosure. BRIEF DESCRIPTION OF THE DRAWINGS

[0007] The present invention will be better understood by reading the following description of non-limiting embodiments with reference to the accompanying drawings, in which:

[0008] Figure 1 An ultrasound imaging system according to one or more embodiments of the present disclosure is shown.

[0009] Figure 2 An overview of a process for automated analysis of cervical elastography according to an exemplary embodiment is shown.

[0010] Figure 3 A flow chart illustrating a first method for automated analysis of cervical elastography according to an exemplary embodiment is shown.

[0011] Figure 4 A flow chart illustrating a second method for automated analysis of cervical elastography according to an exemplary embodiment is shown.

[0012] Figure 5 A flow chart illustrating a third method for automated analysis of cervical elastography according to an exemplary embodiment is shown.

[0013] Figure 6 A display device displaying a frame of a cervical ultrasound scene including user input is shown according to an exemplary embodiment.

[0014] Figure 7 A display device displaying a loss image generated for a cervical ultrasound scene including automatic foreground region of interest assignment is shown according to an exemplary embodiment.

[0015] Figure 8 A display device displaying a loss image generated for a cervical ultrasound scene including automatic background region of interest assignment is shown according to an exemplary embodiment.

[0016] Fig. 9 A display device displaying a loss image generated for a cervical ultrasound scene including automatic foreground selection is shown according to an exemplary embodiment.

[0017] Fig.10 A display device displaying a loss image generated for a cervical ultrasound scene including automatic background selection is shown according to an exemplary embodiment. DETAILED DESCRIPTION

[0018] The following description relates to various embodiments of systems and methods for automatic cervical elastic imaging. Predicting premature birth is an application of cervical elastic imaging. In clinical practice, an ultrasound technician performs cervical ultrasound and selects a foreground area and a background area (also referred to herein as foreground and background) of the cervix from an ultrasound image. The foreground area can represent cervical tissue of interest, and the background area can represent a baseline tissue that is compared with the cervical tissue of interest to estimate changes in cervical elasticity. Strain values ​​are obtained for the selected areas, wherein the strain value is a measurement of the deformation or displacement of a tissue structure in response to an applied mechanical force or stress. The strain value is used to obtain a softness ratio indicating the softness of the cervical tissue. The softness ratio can be used to evaluate cervical changes that may indicate premature birth.

[0019] Currently, selection of foreground and background regions of the cervix is ​​entirely manual, which presents challenges to the clinical practice of cervical elastography. As an example, reliance on user input may produce erroneous outputs due to the instability of cervical elastography and the variability of the technique in relation to the technician's expertise. Relatedly, manual selection of background and foreground regions may be difficult for inexperienced technicians. Furthermore, as currently practiced, cervical elastography analysis is performed on user-selected frames of ultrasound imaging data, which increases the complexity and potential for erroneous outputs for the technician. Another challenge is that manual selection of background and foreground regions is a time-consuming process for even highly skilled technicians.

[0020] To address the above challenges, a method for automatic cervical elastic imaging analysis is disclosed herein. In the disclosed method, an ultrasound technician (also referred to herein as a user) guides the cervical elastic imaging analysis by indicating the position of the cervix on a cervical ultrasound scene (also referred to herein as a scene) generated during cervical ultrasound imaging. As used herein, the cervical ultrasound scene includes a digital record of real-time images or frames captured during a cervical ultrasound examination. In some examples, the indication may include a user drawing a line or rectangle containing a cervical region on the cervical ultrasound scene, such as via a user input device coupled to an ultrasound imaging system. The method includes automatically analyzing a portion of the entire scene (e.g., all frames of the scene) adjacent to the user-guided cervical region using strain measurements to calculate a loss image, and suggesting a foreground and background to the user based on the loss image. In some examples, a first loss image may be generated by calculating the average and standard deviation of the strains measured on the image sequence, or in other words, by averaging the frames in the time direction. Additionally or alternatively, several loss images may be generated. In some examples, the user may confirm the selection, and in response to the user confirmation, the method includes automatically generating a softness ratio based on the automatically selected region by averaging the softness ratios of the foreground and background strain measurements over the entire scene.

[0021] By automatically selecting the foreground and background of the cervical ultrasound scene based on the loss image, a number of advantages are achieved. The loss image improves processing efficiency by focusing the analysis on tissue within the cervical region close to the user's guidance. In addition, by generating a softness ratio from the automatically selected foreground and background rather than the entire scene, using the loss image can be more efficient. The processing efficiency gain of the method supports increased accuracy because the selection of foreground and background areas and the measurements performed thereon are likely to be more accurate due to the analysis of all frames of the cervical ultrasound scene. The method reduces the reliance on the technician's experience, and further, as a benefit to experts and less experienced technicians, the method reduces the effort and time to perform cervical elastic imaging. As another advantage, the disclosed method is compatible with shear wave imaging as a method for automatically determining the background and foreground in shear wave analysis. However, the disclosed method can also replace shear wave analysis when a less invasive method is desired. Strain imaging does not utilize such high-energy push pulses to determine the elastic properties of cervical tissue, and is therefore more preferred than shear wave imaging for predicting preterm birth. In this manner, the disclosed method increases usability, increases efficiency, reduces the impact of noisy measurements, reduces reliance on expert-level knowledge, and is less invasive to the fetus.

[0022] Provided herein are methods for using an ultrasound imaging system such as Figure 1A method and system for automatically analyzing cervical elasticity imaging using system 100). Figure 2 A diagram illustrating the disclosed method for automatically analyzing cervical elasticity imaging using an ultrasound imaging system is shown in FIG. The ultrasound imaging system can perform one or more methods for cervical elasticity imaging, which are respectively illustrated by Figure 3 , Figure 4 and Figure 5 The flowcharts of the methods 300, 400 and 500 are shown. In the disclosed method, the operator guides the automatic cervical elastography analysis by indicating the cervical position on the ultrasound scene. The automatic cervical elastography analysis analyzes the entire scene and generates a cervical elastography analysis output. Figure 6 A display of an example ultrasound scene showing the position of the cervix including operator indications as part of automated cervical elastography analysis is shown in FIG. Figure 7 Figure 4 shows the foreground region of interest selection as part of automated cervical elastography analysis. Figure 6 display. Figure 8 Figure 3 shows the background region of interest selection included as part of automated cervical elastography analysis. Figure 6 display. Fig. 9 FIG. 4 shows a method for including foreground selection as part of automated cervical elastography analysis. Figure 6 display. Fig.10 Figure 2 shows the inclusion of background selection as part of automated cervical elastography analysis. Figure 6 In some examples, the systems and methods disclosed herein for automatic cervical elasticity imaging can be applied to elasticity imaging analysis of other tissues and organs, such as breast imaging.

[0023] Figure 1 A block diagram of a system 100 according to an embodiment is depicted. In the illustrated embodiment, the system 100 is an imaging system, and more specifically an ultrasound imaging system. As shown, the system 100 includes a plurality of components. These components may be coupled to each other to form a single structure, may be separate but located in a common room, or may be remotely located to each other. For example, one or more of the modules described herein may be operated in a data server that has different and remote locations for other components of the system 100 such as probes and user interfaces. Alternatively, in the case of an ultrasound system, the system 100 may be a single system that can be moved from one room (e.g., portable) to another room. For example, the system 100 may include wheels or be transported on a vehicle.

[0024] In the illustrated embodiment, the system 100 includes a transmit beamformer 101 and a transmitter 102, which drives an array of elements 104 (e.g., piezoelectric elements) within a diagnostic ultrasound probe 106 (or transducer), including piezoelectric ceramics, high dielectric ceramics, single crystals, etc., to transmit ultrasound signals (e.g., continuous signals or pulsed signals) into a body or volume (not shown) of a subject. The elements 104 and the probe 106 can have a variety of geometric shapes. The ultrasound signal is backscattered from a body structure (e.g., an inserted needle) to produce an echo that returns to the element 104. The echo is received by a receiver 108. The received echo is provided to a receive beamformer 110, which performs beamforming and outputs a radio frequency (RF) signal. The RF signal is then provided to an RF processor 112 that processes the RF signal. Alternatively, the RF processor 112 may include a complex demodulator (not shown) that demodulates the RF signal to form an I / Q data pair representing the echo signal. The RF or I / Q signal data may then be provided directly to the memory 114 for storage (e.g., temporary storage). The system 100 also includes a system controller 116, which may be part of a single processing unit (e.g., a processor) or distributed among multiple processing units. The system controller 116 is configured to control the operation of the system 100.

[0025] For example, the system controller 116 may include an image processing module that receives image data (e.g., an ultrasound signal in the form of RF signal data or an I / Q data pair) and processes the image data. For example, the image processing module may process the ultrasound signal to generate two-dimensional (2D) slices or frames of ultrasound information (e.g., ultrasound images) or ultrasound waveforms (e.g., continuous or pulsed wave Doppler spectra or waveforms) for display to the operator. The image processing module may be configured to perform one or more processing operations based on a plurality of optional ultrasound modalities on the acquired ultrasound information. By way of example only, ultrasound modalities may include color flow, acoustic radiation force imaging (ARFI), B mode, A mode, M mode, spectral Doppler, acoustic flow, tissue Doppler module, C scan, and elastic imaging. In addition, in some examples, one or more processing operations may include one or more image transformations, such as a Radon transform for identifying linear features in ultrasound images.

[0026] When the echo signal is received, the acquired ultrasound information can be processed in real time during the imaging session (or scanning session). Additionally or alternatively, the ultrasound information can be temporarily stored in the memory 114 during the imaging session and processed in a less than real-time manner in real-time or offline operation. An image memory 120 is included, which is used to store processed slices or waveforms of acquired ultrasound information that are not scheduled for immediate display. The image memory 120 may include any known data storage medium, such as a permanent storage medium, a removable storage medium, etc. In addition, the image memory 120 may be a non-transitory storage medium including instructions that, when executed, cause the processor 112 to perform automatic analysis of cervical elasticity imaging. For example, the instruction may cause the processor 112 to acquire an ultrasound scene including multiple frames and receive user input indicating the position of the cervix on the ultrasound scene. The instruction may cause the processor 112 to generate a first loss image for multiple frames and select the background and foreground of the ultrasound scene based on the first loss image. The instruction may cause the processor 112 to generate a softness ratio from the selected background and foreground. Refer to the following Figures 2 to 10 A method for automated analysis of cervical elastography is described in more detail.

[0027] In operation, the ultrasound system may acquire data (e.g., 2D data sets, spectral Doppler data sets, and / or volume data sets) by various techniques (e.g., three-dimensional (3D) scanning, real-time 3D imaging, volume scanning, 2D scanning using a probe with a positioning sensor, freehand scanning using voxel correlation techniques, scanning using a 2D or matrix array probe, etc.). An ultrasound spectrum (e.g., waveform) and / or image may be generated (at the controller 116) from the acquired data and displayed to an operator or user on the display device 118.

[0028] The system controller 116 is operably connected to a user interface 122 that enables an operator to control at least some of the operations of the system 100. The user interface 122 may include hardware, firmware, software, or a combination thereof that enables an individual (e.g., an operator) to directly or indirectly control the operation of the system 100 and its various components. As shown, the user interface 122 includes a display device 118 having a display area 117. In some embodiments, the user interface 122 may also include one or more user interface input devices 115, such as a physical keyboard, a mouse, and / or a touchpad. In one embodiment, a touchpad may be configured to the system controller 116 and the display area 117 so that when a user moves a finger / glove / stylus on the surface of the touchpad, a cursor on the ultrasound image or Doppler spectrum on the display device 118 moves in a corresponding manner.

[0029] In an exemplary embodiment, the display device 118 is a touch-sensitive display (e.g., a touch screen) that can detect the presence of an operator's touch on the display area 117 and can also identify the location of the touch in the display area 117. The touch can be applied by, for example, at least one of an individual's hand, a glove, a stylus, etc. In this way, the touch-sensitive display can also be characterized as a user input device configured to receive input from an operator (such as a request to adjust or update the orientation of the displayed image). The display device 118 also communicates information from the controller 116 to the operator by displaying the information to the operator. The display device 118 and / or the user interface 122 can also perform audio communication. The display device 118 is configured to present information to the operator during or after an imaging or data acquisition session. The presented information can include ultrasound images, graphic elements, measurement graphics of the displayed image, user-selectable elements, user settings, and other information (e.g., administrative information, personal information of the patient, etc.). In one example, the ultrasound image can include an ultrasound scene (or video) that includes multiple 2D frames.

[0030] In addition to the image processing module, the system controller 116 may also include one or more of a graphics module, an initialization module, a tracking module, and an analysis module. The image processing module, the graphics module, the initialization module, the tracking module, and / or the analysis module may coordinate with each other to present information to the operator during and / or after the imaging session. For example, the image processing module may be configured to display the acquired image on the display device 118, and the graphics module may be configured to display specified graphics with the displayed image, such as selectable icons related to the image (e.g., an image rotation icon) and measurement parameters (e.g., data). As an example, the analysis module may include automatic analysis of cervical elasticity imaging, referred to herein as Figures 2 to 10 Describe in more detail.

[0031] The screen of the display area 117 of the display device 118 is composed of a series of pixels that display data acquired using the probe 106. The acquired data includes one or more imaging parameters calculated for each pixel or a group of pixels of the display (e.g., a group of pixels assigned the same parameter value), wherein the one or more calculated image parameters include one or more of intensity, velocity (e.g., blood flow velocity), color flow velocity, texture, granularity, contractility, deformation, deformation rate value, and strain, strain value. This series of pixels then constitutes a display image and / or Doppler spectrum generated from the acquired ultrasound data.

[0032] In some examples, the ultrasound console 150 can house the RF processor 112, the memory 114, one or more user interface input devices 115, the system controller 116, the image memory 120, and the user interface 122 (including the display device 118 with the display area 117). The ultrasound console 150 can further house the transmit beamformer 101, the transmitter 102, the receiver 108, and the receive beamformer 110. In addition, in some examples, the cable 151 can communicatively couple the probe 106 to the console 150, so that the received ultrasound data can be transmitted from the probe 106 to the console 150 via the cable 151. Figure 1 In other embodiments not depicted, the probe 106 can be communicatively coupled to the console 150 via one or more wireless networks. Thus, data and signals can be sent between the probe 106 and the console 150 via the cable 151 or one or more wireless networks for generating and interpreting ultrasound images.

[0033] Figure 2 An overview of a process 200 for automated cervical elastography is schematically illustrated, the process referring to Figure 3 , Figure 4 and Figure 5 Described in more detail. Process 200 includes an automatic method for obtaining a foreground and background for cervical elastic imaging analysis, including obtaining strain measurements for cervical tissue and generating a flexibility index therefrom. The foreground can represent cervical tissue of interest, and the background can represent a baseline or reference tissue that is compared to the cervical tissue of interest to estimate changes in cervical elasticity. Process 200 analyzes an entire ultrasound scene, such as all frames of a scene, as opposed to analyzing only a single frame. Process 200 includes a manual subprocess and an automatic subprocess, as described below. Compared to existing methods, the disclosed method exemplified by process 200 automates steps of cervical elastic imaging analysis that would otherwise be performed manually by an ultrasound technician.

[0034] The process 200 includes acquiring a cervical ultrasound scene 202. The cervical ultrasound scene 202 is manually acquired by an ultrasound technician. For example, the ultrasound technician may operate a probe of an ultrasound imaging system, such as the probe 106 of the system 100, to acquire the cervical ultrasound scene 202. In one example, the technician may perform a patient examination according to an examination workflow that determines certain measurements of the cervix, including images of the cervix for structural assessment and measurements of cervical tissue in response to compression. The cervical ultrasound scene 202 may include a plurality of frames 226. The cervical ultrasound scene 202 may be displayed to the ultrasound technician via a display device, such as the display device 118 of the system 100.

[0035] The cervical location 216 may be manually marked by the sonographer in one or more frames 226 , as shown in the marked image 204 .

[0036] Next, process 200 includes automatic generation of a loss image 206 for a strain region of interest (ROI) 228 for the cervical ultrasound scene 202. In strain imaging, strain is measured in the strain ROI 228 and displayed as a color-coded representation. The color-coded representation of the measured strain can be overlaid on a brightness mode or "B-mode" ultrasound image, where in B-mode, the echoes generated by the ultrasound are represented as varying shades of gray. The loss image 206 can be calculated based on a loss function and the color-coded strain values ​​in the strain ROI 228. Areas outside the strain ROI 228 are not considered in process 200. In one example, the loss image 206 can be at least a first loss image. Refer to FIG. 200 for details. Figure 3 The loss image calculation is described in more detail.

[0037] The process 200 includes automatic determination of a foreground region of interest 218 within the loss image 206. The foreground region of interest 218 may be a region of the loss image 206 that is within a threshold proximity of the cervical location 216.

[0038] Similarly, process 200 includes automatic determination of a background region of interest 220 of loss image 206. Background region of interest 220 may be a second region of loss image 206 within a threshold proximity of cervical location 216, which may include a larger area surrounding cervical location 216 or an area at a greater distance from cervical location 216 than foreground region of interest 218. Alternatively, background region of interest may be selected as a region of a predetermined size representing a region having a minimum loss function.

[0039] The process 200 then includes automatically selecting a foreground 222 and a background 224 from the foreground region of interest 218 and the background region of interest 220, respectively. The foreground 222 and the background 224 are automatically selected based on the loss image. Figures 3 to 5 Describe the selection process in more detail.

[0040] From the foreground 222 and the background 224, the process 200 includes the automatic generation of the softness ratio 214. Figure 3 , Figure 4 and Figure 5 The generation of the softness ratio 214 is described.

[0041] Figure 3 is a flow chart illustrating an example high-level method 300 for automated analysis of cervical elastography for risk assessment of preterm birth according to an embodiment of the present disclosure. Figure 1The method 300 is described with reference to the system and components of FIG. 1 , but it should be understood that the method 300 may be implemented with other systems and components without departing from the scope of the present disclosure. The method 300 may be based on a system stored in, for example, Figure 1 In one non-limiting example, the method 300 may be executed according to the instructions in the non-volatile memory of the computing device of the image memory 120. Figure 2 process 200.

[0042] At 302, method 300 includes acquiring ultrasound data of a patient. An ultrasound probe (e.g., Figure 1 The ultrasound data may be acquired by an ultrasound probe 106 (e.g., an ultrasound probe 106). The ultrasound data may be processed to generate one or more displayable images that may be displayed on a display device (e.g., display device 118), such as a cervical ultrasound scene including multiple frames. The ultrasound data may be processed to generate 2D images and / or 3D renderings that may be displayed in real time as the images are acquired and / or may be displayed in a more persistent manner (e.g., a freeze indication on a given image) in response to user input. The ultrasound data may include B-mode ultrasound data, and may also include ultrasound elastography data, wherein tissue strain in a predetermined strain ROI is recorded and overlaid on the B-mode ultrasound data.

[0043] At 304, method 300 includes receiving user input indicating the location of the cervix on the selected displayed ultrasound image frame. The user input may be manually entered by the ultrasound technician. For example, the ultrasound technician may indicate the location of the cervix in the cervical ultrasound scene via an input device that communicates with the display and the ultrasound system (such as via the user interface input device 115 of the system 100). An operator of the ultrasound probe may perform a patient examination according to an examination workflow that dictates that certain measurements of the cervix be taken. The operator may indicate the location of the cervix via an input device of the display (such as an electronic stylus, a mouse, or a touch screen). For example, a user may draw a line along the selected displayed ultrasound image to indicate the location of the cervix (e.g., Figure 2 For another example, the user may draw a box or circle indicating the cervical position on the currently displayed ultrasound image, or indicate the cervical position in a different manner.

[0044] At 306, method 300 includes generating a loss image (eg, loss image 206) for the entire strain ROI. Figure 4 An example method for generating a loss image is described in more detail.

[0045] At 308, once the loss image is generated, the method 300 includes automatically selecting a background and a foreground of the cervical ultrasound scene, wherein the foreground can represent cervical tissue of interest and the background can represent surrounding cervical tissue (e.g., reference tissue) to compare the surrounding cervical tissue to the foreground region. The method may include selecting a first pixel with a maximum loss value within the foreground region of interest as the foreground. The method may include selecting a second pixel with a minimum loss value within the background region of interest as the background. Additionally or alternatively, the foreground and background may be selected by averaging the loss values ​​of a plurality of pixels within a circle having a center pixel p (e.g., see Figures 9 and 10 ). In such an example, the first pixel may be the central pixel p in the foreground region of interest, and the second pixel may be the central pixel p in the background region of interest. In this way, the strain measurement may be more stable and less affected by noise. Figure 5 The method for computing the loss image is described in more detail.

[0046] At 310, method 300 includes automatically determining a softness ratio for each frame of the scene. In one example, the softness ratio is calculated from the strain values ​​and can be a ratio of an average strain value for the foreground to an average strain value for the background. The ratio can be foreground strain / background strain or background strain / foreground strain. Determining the softness ratio can include taking a ratio of an average strain value for pixels in the foreground of all frames in a plurality of frames of the cervical ultrasound scene to an average strain value for pixels in the background of all frames of the cervical ultrasound scene. In some examples, an alternative method may include calculating the softness ratio based on a single image and excluding the averaged portion.

[0047] As an additional example, the softness ratio can be calculated based on shear wave imaging. For example, in shear wave imaging, elasticity values ​​or values ​​for the speed of sound are generated within the region of interest. If a shear wave image sequence is acquired for the region of interest, the method 300 can be similarly applied. In particular, instead of strain values, elasticity values ​​or speed of sound values ​​can be used to calculate the loss function.

[0048] At 312, method 300 includes automatically excluding outlier frames, such as noisy frames and low-quality frames, from downstream analysis. In one example, outlier frames may be characterized by high strain variance or low mean strain in the foreground region. Additionally or alternatively, since strain estimation reliability is related to sufficiently correlated ultrasound signals, in some examples, a correlation measure may be used to determine low-quality frames. For another example, outlier frames may be determined based on a softness ratio determined for each frame and an average softness ratio determined for all frames of the scene before excluding outliers. As an example, outlier frames may be determined based on a quality threshold. The quality threshold may be a positive non-zero threshold, such as a percentage deviation from an average softness ratio for all frames of the scene before excluding outliers.

[0049] At 314, method 300 includes determining an average softness ratio for all frames of the scene. In some examples, all frames may include a set of frames excluding outlier frames. Alternatively, the method may include calculating an average of foreground strain and background strain, and then calculating the softness ratio.

[0050] At 316, method 300 includes displaying the results of the cervical elastography analysis. The results may include an average softness ratio for all frames of the scene. In some examples, all frames may include a set of frames that exclude outlier frames. In some examples, a score or risk assessment for preterm birth may be output based on the softness ratio, or in other examples, a score or risk assessment for preterm birth may be output based on the softness ratio and additional medical factors such as patient history.

[0051] Figure 4 A flow chart illustrating an exemplary method 400 for generating a loss image as part of an automated analysis of cervical elastic imaging for risk assessment of preterm birth in accordance with an embodiment of the present disclosure is shown. In one example, a loss image can be determined based on a loss function that operates at the level of individual pixels across multiple frames that include a cervical ultrasound scene. For each pixel, the loss function produces a corresponding loss value. In one example, the loss image can include multiple loss values ​​determined for each individual pixel of each frame for the cervical ultrasound scene using the loss function. Reference Figure 1 The method 400 is described with reference to the system and components of FIG. 4 , but it should be understood that the method 400 may be implemented with other systems and components without departing from the scope of the present disclosure. The method 400 may be based on a storage device such as Figure 1 In one non-limiting example, the method 400 may be executed according to the instructions in the non-volatile memory of the computing device of the image memory 120. Figure 2 process 200.

[0052] At 402, method 400 includes extracting a region from a plurality of frames including an ultrasound scene. In one example, as above, the region may be a strain ROI (e.g., Figure 2 The strain ROI 228) can be manually set by the technician before the cervical elastic imaging examination. The extraction can be automatically performed by the processor of the ultrasound system using only user input including the cervical position and the ultrasound imaging data acquired by the ultrasound technician.

[0053] At 404, method 400 includes determining a loss value for each pixel of the region in multiple frames using the loss function. For example, method steps 406, 408, 410, and 412 described below may be performed for each pixel within the extracted region for each frame of the scene. By determining the loss value for the extracted region as opposed to the entire scene, the method may reduce processing time while achieving a high degree of stability.

[0054] At 406, method 400 includes determining a strain value for each pixel of the region for each frame. The method may include defining a set Sp that includes a plurality of strain values ​​for each pixel in a plurality of frames of the scene. In one example, the set may be calculated as follows:

[0055] S p ={s p (i), i=1,…,n} (1)

[0056] where s p (i) is the strain value at pixel p in frame i. In other words, for each pixel p, there is a corresponding strain value s for each frame i (e.g., from 1 to n) of the ultrasound scene. p (i).

[0057] At 408, method 400 includes calculating a strain value s associated with pixel p across multiple frames. p (i), and at 410, method 400 includes calculating the strain value s associated with pixel p across all frames. p (i) Variance.

[0058] At 412, method 400 includes determining a loss value for each pixel in the pixel set Sp based on the mean and variance of the strain values. The loss value is calculated for each pixel p as follows:

[0059] L p =Mean(S p )+Var(S p ). (2)

[0060] Mean(S p ) is the set S p The average value of the strain value, and Var(S p ) is the set S p In other words, to calculate the loss value L for pixel p, two components are considered: the average strain and the strain variance, and the loss value of a given pixel can be the sum of the average strain and the strain variance for pixel p of the set Sp.

[0061] In one example, in response to determining the loss value for each of the plurality of pixels of the cervical ultrasound scene, the method 400 may return. For example, the method 400 may return to Figure 3 The loss value can also be used to select foreground or background from pixels of a scene, as described below with reference to Figure 5 as described.

[0062] Method 500 illustrates an example of obtaining a loss function by calculating the mean and standard deviation of the estimated strain over a sequence of images. Another method may include calculating several loss images, for example, by using a narrow sliding window for calculating the mean and standard deviation in the time direction. Additionally or alternatively, the loss function may be obtained by adding weights to the mean and variance before summing (e.g., weighted sum). As another method, the loss function may be obtained by calculating the ratio of the mean to the variance (e.g., mean / variance).

[0063] Figure 5 A flow chart illustrating an exemplary method 500 for selecting background and foreground as part of automated analysis of cervical elastography for risk assessment of preterm birth according to an embodiment of the present disclosure is shown. Figure 1 The method 500 is described with reference to the system and components of FIG. 1 , but it should be understood that the method 500 may be implemented with other systems and components without departing from the scope of the present disclosure. The method 500 may be based on a storage device such as Figure 1 In one non-limiting example, the method 500 may be executed according to the instructions in the non-volatile memory of the computing device of the image memory 120. Figure 2 process 200.

[0064] At 502, method 500 includes receiving a loss image. In one example, as above, the loss image can be determined based on a loss function, as described above with reference to Figure 4 as described.

[0065] At 504, method 500 includes assigning a region of interest (ROI) based on user input indicating a cervical location. For example, a foreground ROI (e.g., a region of interest) may be assigned based on proximity to the cervical location (such as within a first threshold proximity to the cervical location). Figure 2218). Similarly, a background ROI (e.g., background region of interest 220) can be assigned based on proximity to the cervical location, such as within a second threshold proximity to the cervical location. In one example, the foreground region of interest can be closer to the cervical location, and the background region of interest can be a greater distance from the cervical location or a larger area surrounding the cervical location. In one example, the method 500 can determine the boundaries of the foreground ROI and the background ROI by analyzing the results of a manual cervical elastography analysis performed by an expert user. For example, data analysis using data acquired from an expert user can provide a baseline assessment of foreground and background thresholds to form a boundary.

[0066] At 506, method 500 includes selecting a foreground based on a maximum loss value within the foreground ROI. In one example, method 500 may select a first pixel region including a plurality of pixels (e.g., 10 pixels) within the foreground ROI where the loss value is the maximum loss value based on the loss image. In other words, the foreground is selected such that the average strain value of the foreground is maximized throughout the foreground ROI.

[0067] At 508, method 500 includes selecting a background based on a lowest loss value within the background ROI. In one example, method 500 may select a second pixel region including a plurality of pixels (e.g., 10 pixels) within the background ROI where the loss value is the lowest loss value based on the loss image. In other words, the background is selected such that the average strain value of the background is minimized across the background ROI.

[0068] At 510, method 500 includes displaying the background and foreground selections to a user.For example, the first pixel region and the second pixel region may be overlaid on one of the frames of the cervical ultrasound scene via a display of the ultrasound imaging system.

[0069] At 512, method 500 includes requesting the user to confirm the background and foreground selections. For example, the method may include generating a pop-up message that is displayed via a display of the ultrasound imaging system. The user may click a button to confirm or deny one or both of the selected background and foreground. As used herein, a button may refer to any type of user input that provides a mechanism for the user to select, confirm, or otherwise indicate a selection.

[0070] At 514, method 500 includes determining whether the user confirms the selection. If the user confirms the selection, method 500 may return. For example, method 500 may return to Figure 3 Method step 310 in.

[0071] At 516 , if the user does not confirm one or both of the selected background and foreground, method 500 may include prompting the user to manually input a background and / or foreground selection.

[0072] Figures 6 to 10 Example user interfaces 600, 700, 800, 900, 1000 are illustrated, which show a sequence of actions performed within a method for automatic analysis of cervical elasticity imaging for an exemplary ultrasound imaging system. The method for automatic analysis of cervical elasticity imaging can be used in conjunction with the above referenced Figures 2 to 5 The instructions for executing the methods described in the user interfaces 600, 700, 800, 900, 1000 may be executed by a controller (e.g., controller 116) based on instructions stored in a memory of the controller (e.g., image memory 120) and in combination with instructions from the above references. Figure 1 The automatic analysis may be performed based on feedback from components of the ultrasound imaging system described herein (e.g., the probe 106, the user interface input device 115, etc.). In an example, a user, such as an ultrasound technician, may interact with the user interfaces 600, 700, 800, 900, 1000 via one or more input devices, and the user interfaces 600, 700, 800, 900, 1000 may visually display user input and automatic output of the automatic analysis.

[0073] Figure 6 A user interface 600 is shown. The user interface 600 may be displayed on an ultrasound imaging system (eg, Figure 1 The user interface 600 may be displayed on a display 602 of the system 100 in FIG. 600. The user interface 600 may display a cervical ultrasound scene or scene 612 that may be obtained by an ultrasound technician operating a probe (e.g., the probe 106) during a cervical ultrasound. In addition, the user interface 600 may display frames 606 of a plurality of frames constituting the scene 612 one at a time. For example, the frames 606 may be displayed based on a user indication, such as in response to a user scanning the scene 612 and selecting the frames 606 via an input device. Additionally or alternatively, the frames 606 may be displayed automatically, such as based on an image processing algorithm.

[0074] The user can indicate the position of the cervix on the frame 606 via an input device such as an electronic stylus connected to the display. For example, the user can draw a line 608 indicating the position of the cervix on the display screen 604. As another example, the user can draw a boundary 610 (e.g., an organic shape, a rectangle, etc.) indicating the position of the cervix on the display screen 604. In one example, the cervical ultrasound scene 612 and the cervical position as indicated by the line 608 or the boundary 610 can be the only user input to the automatic analysis of cervical elastic imaging. Based on the cervical ultrasound scene and in response to the user indicating the cervical position, the automatic analysis of cervical elastic imaging generates a loss image for a sub-region of the scene 612, such as Figure 7 For example, a loss image may be generated for the portion of frame 606 within the strain ROI, which may be manually set by a technician prior to the examination.

[0075] Figure 7 The above reference shows Figure 6 The user interface 700 is shown on the display screen 604 of the display 602 described above. The user interface 700 displays an automatically generated loss image 714. As above, the loss image 714 can be generated from a plurality of loss values, one loss value per pixel, wherein the loss values ​​are determined based on a loss function. In addition, the user interface 700 displays a line 608 indicating a cervical position indicated by the user. The line 608 is shown herein for reference; however, in some examples, the line 608 can be excluded from the display after the loss image is generated. The automatic analysis of the cervical elastic imaging can assign a foreground region of interest displayed as a first shape 702. In one example, the foreground region of interest can be assigned based on proximity to the line 608 indicating the cervical position as above. For example, the first shape 702 can include an area of ​​the loss image 714 within a first threshold proximity 704 of the line 608. In one example, the first threshold proximity 704 can include a positive, non-zero number of pixels, a measurement of length, or other value.

[0076] Figure 8 The above reference shows Figure 6602 is depicted with a user interface 800 displayed on a display screen 604. As above, the user interface 800 displays the loss image 714. In addition, the user interface 800 displays the line 608 indicating the location of the cervix. The automatic analysis of the cervical elastic imaging can assign a background region of interest displayed as a second shape 802. In one example, the background region of interest can be assigned based on proximity to the line 608 indicating the location of the cervix as above. For example, the second shape 802 can include an area of ​​the loss image 714 within a second threshold proximity 804 of the line 608. In one example, the second threshold proximity 804 can include a positive, non-zero number of pixels, a measurement of length, or other value, which can be a larger area around the line 608 or a larger distance from the line. For another example, the second threshold proximity 804 can include an area of ​​a predetermined size representing an area with a minimum loss function.

[0077] Fig. 9 The above reference shows Figure 6 900 is shown on the display screen 604 of the depicted display 602. The user interface 900 displays a loss image 714 including a first shape 702 indicating a foreground region of interest assigned by the automatic analysis of cervical elastic imaging. The automatic analysis of cervical elastic imaging can select a foreground from within the foreground region of interest, shown as a first circle 902. The foreground can be determined based on the loss image 714 as above, where the foreground is selected so that the average strain value of the foreground is maximized throughout the foreground ROI. The first circle 902 includes a first pixel 904 within the first shape 702, the first pixel being shown as a small circle with a black dashed border, where the loss value is maximized based on the loss image 714. The first circle 902 can also include a plurality of adjacent pixels (e.g., 10 pixels) proximate to the first pixel.

[0078] Fig.10 The above reference shows Figure 6 1. The user interface 1000 is shown on the display screen 604 of the display 602. As above, the user interface 1000 displays a loss image 714 including a first shape 702 and a second shape 802, indicating a foreground region of interest and a background region of interest assigned by the automatic analysis of cervical elastic imaging, respectively. The user interface 1000 displays a first circle 902 indicating a selected foreground including a first pixel 904. The automatic analysis of cervical elastic imaging can assign a background displayed as a second circle 1002. The background can be determined based on the loss image 714 as above, wherein the background is selected so that the average strain value of the background is minimized in the entire background ROI. The second circle 1002 may include a second pixel 1004 within the second shape 802, the second pixel being shown as a black dot, where the loss value is lowest based on the loss image. The second circle 1002 may also include a plurality of adjacent pixels (e.g., 10 pixels) close to the second pixel.

[0079] In some examples, the user interface 1000 can display a pop-up window 1006 requesting user confirmation of the background and foreground selections displayed for the first circle 902 and the second circle 1002, respectively. In response to the user confirmation, the automatic analysis of the cervical elasticity imaging can include calculating a softness ratio based on the selected foreground and background. For example, for each frame, the softness ratio can include a ratio of the average strain value of the foreground (e.g., within the first circle 902) to the average strain value of the background (e.g., within the second circle 1002). The final softness ratio includes the average of the softness ratios for all frames. The user interface 1000 can display the softness ratio on a pop-up window 1008 of the automatic analysis of cervical elasticity.

[0080] In this way, the disclosed system and method increase usability, minimize noise effects, speed up processing, and reduce expert input dependency. The technical effect of the disclosed method and system for cervical elastic imaging is increased processing efficiency. By using loss images generated for a portion of all frames of the ultrasound scene, as opposed to analyzing every pixel of the entire scene, the tissue within the cervix is ​​automatically analyzed, focusing on a location close to the user's guidance. The processing efficiency gain of the method allows the determination of softness ratios from a large, stable data set, thereby increasing accuracy and reproducibility.

[0081] The present disclosure also provides support for a method, the method comprising: acquiring an ultrasound scene including a plurality of frames; generating a first loss image for the plurality of frames; selecting a background and a foreground of the ultrasound scene based on the first loss image; and generating a softness ratio from the selected background and foreground. In a first example of the method, generating the first loss image comprises: extracting a region from a plurality of frames including an ultrasound scene, and for each pixel of the ultrasound scene: defining a pixel set, determining an average strain and a strain variance for the pixel set, and determining a loss value, the loss value being the sum of the average strain and the strain variance for the pixel set. In a second example of the method, the method optionally comprises the first example, the pixel set comprising a plurality of strain values ​​for a single pixel at each of the plurality of frames. In a third example of the method, the method optionally comprises one or both of the first and second examples, the method further comprising: receiving a user input indicating a cervical position on the ultrasound scene, wherein the only user input is the ultrasound scene and the cervical position. In a fourth example of the method, the method optionally comprises one or more or each of the first to third examples, the method further comprising: allocating a foreground region of interest and a background region of interest based on proximity to the user input indicating a cervical position. In a fifth example of the method, the method optionally includes one or more or each of the first to fourth examples, the foreground region of interest is closer to the cervix than the background region of interest. In a sixth example of the method, the method optionally includes one or more or each of the first to fifth examples, the method further includes: selecting the foreground based on the maximum loss value in the foreground region of interest, and selecting the background based on the minimum loss value in the background region of interest. In a seventh example of the method, the method optionally includes one or more or each of the first to sixth examples, the ultrasound scene includes a cervical ultrasound scene. In an eighth example of the method, the method optionally includes one or more or each of the first to seventh examples, the ultrasound scene includes a shear wave image sequence. In a ninth example of the method, the method optionally includes one or more or each of the first to eighth examples, automatically performing generation of a first loss image, selection of a background and a foreground, and generation of a softness ratio.

[0082] The present disclosure also provides support for an ultrasound system, the ultrasound system comprising: an ultrasound probe; a display, the display being communicatively coupled to the ultrasound probe; a user input device, the user input device being communicatively coupled to the display; and a processor and a non-transitory memory communicatively coupled to the display, the non-transitory memory comprising instructions, which when executed cause the processor to: acquire an ultrasound scene comprising a plurality of frames; receive a user input indicating a cervical position on the ultrasound scene; generate a first loss image for the plurality of frames; select a background and a foreground of the ultrasound scene based on the first loss image; and generate a softness ratio from the selected background and foreground. In a first example of the system, the first loss image comprises a plurality of loss values, the plurality of loss values ​​being determined based on a sum of an average strain and a strain variance for each individual pixel of each frame for the ultrasound scene. In a second example of the system, the system optionally comprises the first example, the instructions causing the processor to automatically select a background and a foreground. In a third example of the system, the system optionally comprises one or both of the first example and the second example, the user input, namely the cervical position and the ultrasound scene, is the only user input. In a fourth example of the system, the system optionally includes one or more or each of the first to third examples, the instructions causing the processor to select a foreground based on a maximum loss value within a foreground region of interest, and to select a background based on a minimum loss value within a background region of interest.

[0083] The present disclosure also provides support for a method, the method comprising: acquiring a cervical ultrasound scene including a plurality of frames; receiving a user input indicating a cervical position on the cervical ultrasound scene; generating a first loss image for the cervical ultrasound scene; allocating a foreground region of interest and a background region of interest based on the cervical position; selecting a foreground within the foreground region of interest and a background within the background region of interest based on the first loss image; determining a softness ratio for each frame based on an average strain value in the foreground and an average strain value in the background; and determining a final softness ratio based on an average softness ratio of all frames. In a first example of the method, generating the first loss image comprises: extracting a region from a plurality of frames including the cervical ultrasound scene, and for each pixel of the cervical ultrasound scene: defining a pixel set, determining an average strain and a strain variance for the pixel set, and determining a loss value, the loss value being the sum of the average strain and the strain variance for the pixel set. In a second example of the method, the method optionally comprises the first example, the pixel set comprising a plurality of strain values ​​for a single pixel at each of the plurality of frames. In a third example of the method, the method optionally comprises one or both of the first and second examples, the cervical ultrasound scene and the cervical position being the only user input. In a fourth example of the method, the method optionally includes one or more or each of the first to third examples, the method further comprising: excluding outlier frames based on the softness ratio of each frame.

[0084] As used herein, the elements or steps listed in the singular and beginning with the word "one" or "a kind of" should be understood as not excluding a plurality of said elements or steps, unless such exclusion is explicitly stated. In addition, the reference to "one embodiment" of the present invention is not intended to be interpreted as excluding the existence of additional embodiments that also include the cited features. In addition, unless explicitly stated to the contrary, "comprising", "including" or "having" an embodiment of an element or multiple elements with a specific performance may include additional such elements without the performance. The terms "including" and "in..." are used as the concise language equivalents of the corresponding terms "including" and "wherein". In addition, the terms "first", "second" and "third" etc. are only used as marks, and are not intended to impose numerical requirements or specific positional order on their objects.

[0085] This written description uses examples to disclose the invention, including the best mode, and also to enable a person skilled in the relevant art to practice the invention, including making and using any devices or systems and performing any included methods. The patentable scope of the invention is defined by the claims, and may include other examples that occur to a person skilled in the art. Such other examples are intended to be within the scope of the claims if they have structural elements that do not differ from the literal language of the claims, or if they include equivalent structural elements with insignificant differences from the literal language of the claims.

Claims

1. A method (300, 500, 400), the method comprising: Acquiring an ultrasound scene (202) comprising a plurality of frames (226); generating a first loss image (206) for the plurality of frames (226); selecting a background (224) and a foreground (222) of the ultrasound scene (202) based on the first loss image (206); as well as A softness ratio (214) is generated from the selected background (224) and foreground (222).

2. The method (300, 500, 400) of claim 1, wherein generating the first loss image (206) comprises: extracting a region from the plurality of frames (226) including the ultrasound scene (202); and For each pixel of the ultrasound scene (202): Define a pixel set; determining a mean strain and a strain variance for the set of pixels; as well as A loss value is determined, the loss value being the sum of the mean strain and the strain variance for the set of pixels.

3. The method (300, 500, 400) of claim 2, wherein the set of pixels comprises a plurality of strain values ​​for a single pixel at each frame of the plurality of frames (226).

4. The method (300, 500, 400) of claim 1, further comprising receiving a user input indicating a cervical position (216) on the ultrasound scene (202), wherein the only user input is the ultrasound scene (202) and the cervical position (216).

5. The method (300, 500, 400) of claim 4, further comprising assigning a foreground region of interest (218) and a background region of interest (220) based on proximity to the user input indicating the cervical location (216).

6. The method (300, 500, 400) of claim 5, wherein the foreground region of interest (218) is closer to the cervical location (216) than the background region of interest (220).

7. The method (300, 500, 400) according to claim 5, further comprising selecting the foreground (222) based on a maximum loss value within the foreground region of interest (218), and selecting the background (224) based on a minimum loss value within the background region of interest (220).

8. The method (300, 500, 400) of claim 1, wherein the ultrasound scene (202) comprises a cervical ultrasound scene (202, 612).

9. The method (300, 500, 400) of claim 1, wherein the ultrasound scene (202) comprises a shear wave image sequence.

10. The method (300, 500, 400) of claim 1, wherein said generating the first loss image (206), said selecting the background (224) and the foreground (222), and said generating the softness ratio (214) are performed automatically.

11. An ultrasound system (100), comprising: Ultrasonic probe (106); a display (602) communicatively coupled to the ultrasound probe (106); a user input device communicatively coupled to the display (602); as well as A processor (112) and a non-transitory memory (114) communicatively coupled to the display (602), the non-transitory memory (114) comprising instructions that, when executed, cause the processor (112) to: Acquiring an ultrasound scene (612) comprising a plurality of frames (226); receiving user input indicating a location (216) of a cervix on the ultrasound scene (202); generating a first loss image (206) for the plurality of frames (226); selecting a background (224) and a foreground (222) of the ultrasound scene (612) based on the first loss image (206, 714); and A softness ratio (214) is generated from the selected background (224) and foreground (222).

12. The ultrasound system (100) of claim 11, wherein the first loss image (206) comprises a plurality of loss values ​​determined based on a sum of a mean strain and a strain variance for each individual pixel in each frame (606) for the ultrasound scene (612).

13. The ultrasound system (100) of claim 11, wherein the instructions cause the processor (112) to automatically select the background (224) and the foreground (222).

14. The ultrasound system (100) of claim 11, wherein the user inputs, namely the cervical position (216) and the ultrasound scene (202), are the only user inputs.

15. The ultrasound system (100) of claim 11, wherein the instructions cause the processor (112) to select the foreground (222) based on a maximum loss value within a foreground region of interest (218), and to select the background based on a minimum loss value within a background region of interest (220).

Images