Ultrasound data processing

Abstract

A method and apparatus for maintaining and updating a shared register storing one or more parameters related to features of an anatomical structure accessible by a plurality of different software applications adapted to derive quantized data from ultrasound images using the one or more parameters, in this way, the updating of the parameters can be shared by all the applications, it is avoided that the same parameters need to be recalculated multiple times, and it is avoided that the user needs to re-input the updated parameters into each individual application.

Description

Technical Field

[0001] The method relates to the field of ultrasonic data processing. Background Art

[0002] It is known to apply specialized image processing software modules to ultrasound imaging data that are configured to obtain specific outputs based on the image data. For example, there are modules for obtaining specific anatomical measurements from a set of ultrasound image data, such as dimensions associated with the heart, or standard dimensions associated with a fetus, or dynamic measurements associated with an organ such as the heart, such as cardiac output. There are applications that process data to synthesize new image views or modulate the imagery to better represent certain features, such as segmentation or artificial contrast enhancement.

[0003] US 2007 / 214235 A1 discloses a method and system for sharing intermediate medical image processing results between several different applications. Summary of the invention

[0004] In real-world clinical use, multiple such software applications can often be applied to a given data set to obtain all the desired information. For example, with respect to a single organ such as the heart, there may be multiple different applications that quantify different information associated with the heart or generate specific images associated with the heart.

[0005] The inventors have recognized that within this context, there is often duplication of computational steps between multiple applications, as multiple applications may utilize the same one or more parameters associated with the anatomical structure of interest. The inventors have recognized that there is therefore an opportunity to improve efficiency by providing a data sharing scheme by which the computation and subsequent updating of common parameters utilized by multiple software modules can be centralized to avoid duplicate processing. There is also an opportunity to improve the reliability of the outputs of the software modules, as in some cases, if algorithms or processing steps do not have access to certain anatomical information, they may fail or produce undesirable results.

[0006] The invention is defined by the claims.

[0007] According to an example in accordance with aspects of the present invention, a computer-implemented method is provided, comprising: receiving ultrasound imaging data of an anatomical structure of a patient obtained from an ultrasound acquisition system; communicating with an anatomical feature register, the anatomical feature register storing one or more parameters associated with anatomical features of the anatomical structure; a plurality of application modules, each application module processing the acquired ultrasound imaging data based on a current value of at least one of the one or more parameters associated with the anatomical feature currently stored in the anatomical feature register to generate output quantization or segmentation data; and receiving input data from a user interface.

[0008] The method further includes updating a value of at least one of the one or more parameters stored in the anatomical feature register based on input data received from the user interface and / or output quantization or segmentation data generated by any of the plurality of application modules.

[0009] Thus, the present invention provides a system (and associated methods and computer programs) for enabling multiple application modules to process ultrasound imaging data of a patient's structure in an efficient manner, so that any updates or corrections made to parameters associated with features in the context of a given application module can be easily used by other application modules. This avoids the need to recalculate parameters multiple times, and avoids the need for the user to re-enter updated parameters into each individual application multiple times. In the latter case, the user is able to perform one data input action for updating or setting parameter values, and this automatically updates or sets the parameters for all application modules. In addition, in some cases, if certain algorithms or processing steps cannot access certain anatomical information, they may fail or produce poor results. For example, a heart segmentation model such as the Philips HeartModel may require heart size as input in order to correctly segment very small or large hearts. Therefore, according to some embodiments, preventing algorithm failure or poor algorithm results is a further advantage.

[0010] The term "application module" means, for example, a software module or a hardware module encoded in software, in either case operable to perform processing of ultrasound image data to produce output quantification or segmentation data. It may include or take the form of one or more algorithms for performing the method steps attributed to it. It may include or take the form of a program or routine for performing one or more method steps attributed to it.

[0011] As a non-limiting example, the output quantitative data may include quantification of one or more anatomical measurements, such as a dimension or size parameter of an anatomical structure. For example, these may include a length parameter, an area parameter, or a volume parameter. The output quantitative data may include dynamic measurements, such as hemodynamic measurements in the case of cardiac imaging, such as cardiac output.

[0012] In some embodiments, the method further comprises generating initialization data for at least one of the one or more parameters of the anatomical feature, and storing the initialization data in the anatomical feature register as an initial value for the at least one parameter. In some embodiments, the initialization data is generated based on received ultrasound imaging data. In some embodiments, the initialization data may be generated based on retrieving the initialization data from a data repository. In some embodiments, the initialization data may be generated based on applying at least one of a plurality of application modules to received ultrasound imaging data, or based on applying at least one of a plurality of application modules to historical ultrasound imaging data of a patient retrieved from a data repository.

[0013] In some embodiments, the method further comprises receiving an examination start indicator signal, for example from an ultrasound acquisition system, indicating the start of an ultrasound examination session. In some embodiments, generating and storing initialization data is performed in response to receiving the examination start indicator signal. Thus, at the beginning of the patient examination, an initial value of at least one parameter is obtained. The initial value can be updated as the examination progresses.

[0014] In some embodiments, generating initialization data for at least one of the one or more parameters includes applying an artificial intelligence algorithm to the received ultrasound imaging data. In some embodiments, the artificial intelligence algorithm is a convolutional neural network.

[0015] In some embodiments, for at least a subset of the plurality of application modules, the output quantification or segmentation data may include a segmentation of at least one anatomical feature or structure. This may be in the form of a segmentation mask or mesh for the feature or structure. This may be stored in a register as a series of coordinate points corresponding to the vertices of the mesh or mask.

[0016] In some embodiments, at least a subset of the application modules are each adapted to generate a segmentation based on an initial shape template, and wherein the initial shape template is scaled according to a scaling parameter, and wherein the scaling parameter is at least one of the one or more parameters stored in the anatomical feature register or is calculated based on at least one of the one or more parameters stored in the anatomical feature register.

[0017] In some embodiments, the anatomical structure is a cardiac structure.

[0018] In some embodiments, the one or more parameters include one or more of: heart size, left ventricular size, left atrial size, right ventricular size, end-diastolic right atrial size, ejection fraction and / or left ventricular length.

[0019] In some embodiments, the received ultrasound image data is real-time image data. In some embodiments, the received ultrasound image data is an image data stream. In some embodiments, the plurality of application modules are each adapted to repeatedly or continuously operate on the received image data stream.

[0020] In some embodiments, the received ultrasound imaging data comprises 2D ultrasound imaging data for one or a series of time frames. In some embodiments, the received ultrasound imaging data comprises 3D ultrasound imaging data for one or a series of time frames.

[0021] In some embodiments, the anatomical feature register includes an update counter entry associated with at least one of the one or more parameters, and wherein the update counter entry is incremented concurrently with any update of the parameter value in the anatomical feature register based on the user input data. In some embodiments, in response to the current value of the update counter register exceeding a threshold, any update of the parameter in the anatomical feature register is prevented. In some embodiments, the threshold is zero. In this way, once a parameter has been corrected or updated once, other applications are prevented from further updating it. This avoids duplication of processing and also ensures consistency in the use of the same parameter.

[0022] In some embodiments, a reset indicator signal is generated at the end or start of any ultrasound examination session, and wherein the method includes resetting the update counter entry to zero in response to the reset indicator signal.

[0023] The present invention can also be implemented in software. Therefore, another aspect of the present invention is a computer program product comprising code modules, which are configured to perform a method according to any embodiment or example described in this document or according to any claim of this application when executed by a processor.

[0024] The present invention can also be implemented in hardware. Therefore, another aspect of the present invention is a processing unit, comprising: an input / output portion for use in connection with an ultrasound data acquisition device to receive ultrasound imaging data of an anatomical structure of a patient; an anatomical feature register, which is suitable for storing one or more parameters associated with anatomical features of the anatomical structure; a control module, which is operatively coupled to the anatomical feature register; and a plurality of application modules, which are operatively coupled to the control module, each application module being suitable for processing the ultrasound imaging data received at the input / output portion based on the current value of at least one parameter of the one or more parameters stored in the anatomical feature register to generate output quantization or segmentation data. The above-mentioned control module is suitable for: receiving user input data from a user interface; and updating the value of at least one parameter of the one or more parameters stored in the anatomical feature register based on the user input data provided by the user through the user interface and / or the output quantization or segmentation data generated by any application module of the plurality of application modules.

[0025] Any features or options described with respect to the method aspects of the present invention are equally applicable to the software and hardware aspects of the present invention.

[0026] In some embodiments, the processing unit further comprises a register initialization module adapted to generate initialization data for at least one of the one or more parameters based on the received ultrasound imaging data, the initialization data comprising an initial value for at least one of the one or more parameters.

[0027] In some embodiments, the control module is further adapted to trigger the initialization module to generate initialization data and store the data in the anatomical feature register as an initial value for at least one of the one or more parameters.

[0028] In some embodiments, the anatomical structure is a cardiac structure.In some embodiments, the one or more parameters include one or more of: heart size, left ventricular size, left atrial size, right ventricular size, end-diastolic right atrial size, ejection fraction, and / or left ventricular length.

[0029] Another aspect of the invention is a system comprising: a processing unit according to any embodiment described in this document, such as described above; and an ultrasound data acquisition device operatively coupled to an input / output portion of the processing unit. In some embodiments, the system further comprises a user interface for obtaining user input.

[0030] These and other aspects of the invention will be apparent from and elucidated with reference to the embodiment(s) described hereinafter. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] For a better understanding of the invention, and in order to more clearly show how the invention may be carried into effect, reference will now be made, by way of example only, to the accompanying drawings, in which:

[0032] Figure 1 The steps of an example method according to one or more embodiments of the present invention are outlined;

[0033] Figure 2 schematically outlines components of an example processing device and a process flow according to one or more embodiments of the present invention; and

[0034] Figure 3 Results of the convolutional neural network used to generate initialization data are shown. DETAILED DESCRIPTION

[0035] The present invention will be described with reference to the accompanying drawings.

[0036] It should be understood that the detailed description and specific examples, although indicating exemplary embodiments of the device, system and method, are intended only for illustrative purposes and are not intended to limit the scope of the invention. These and other features, aspects and advantages of the device, system and method of the present invention will be better understood from the following description, the appended claims and the accompanying drawings. It should be understood that the drawings are merely schematic and are not drawn to scale. It should also be understood that the same reference numerals are used in all drawings to indicate the same or similar parts.

[0037] The present invention provides a method and apparatus for maintaining and updating a shared register storing one or more parameters associated with features of an anatomical structure, the shared register being accessible by a plurality of different software applications adapted to derive quantitative data from an ultrasound image using the one or more parameters, such that updates to the parameters can be shared by all applications, avoiding the need to recalculate the same parameters multiple times, and avoiding the need for the user to re-enter the updated parameters into each individual application.

[0038] It is therefore a general object of embodiments of the present invention to provide improved connectivity between different applications. The aim is to ensure that information can flow more easily from one application to the next, particularly if corrections are made by the user.

[0039] Figure 1 The steps of an example method according to one or more embodiments are summarized in block diagram form.Before further explanation in the form of an example embodiment, the steps will be summarized.

[0040] The method includes receiving 12 ultrasound imaging data of an anatomical structure of a patient obtained from an ultrasound acquisition system. The method also includes communicating 14 with an anatomical feature register that stores one or more parameters associated with one or more anatomical features of the anatomical structure. The method also includes a plurality of application modules, each application module processing 16 the acquired ultrasound imaging data based on a current value of at least one of the one or more parameters associated with the anatomical feature currently stored in the anatomical feature register to produce output quantized or segmented data 22.

[0041] The method may further comprise receiving 18 input data from a user interface.

[0042] The method also includes dynamically updating 20 a value of at least one of the one or more parameters stored in the anatomical feature register. This may be based on input data received 18 from a user interface. Additionally or alternatively, it may be based on output quantization or segmentation data 22 produced by any of a plurality of application modules.

[0043] As mentioned above, the method can also be implemented in the form of hardware, for example in the form of a processing unit configured to perform the method according to any example or embodiment described in this document or according to any claim of this application.

[0044] To further aid understanding, Figure 2 A schematic representation of an example processing unit 32 configured to perform a method according to one or more embodiments of the invention is presented, and the processing flow is also schematically shown in more detail.

[0045] The processing unit 32 includes an input / output portion 34 , which is used to connect with the ultrasound data acquisition device 52 to receive ultrasound imaging data of the patient's anatomical structure.

[0046] The processing unit 32 further comprises an anatomical feature register 38 adapted to store one or more parameters associated with one or more anatomical features of the anatomical structure.

[0047] The processing unit 32 further comprises a control module 36 operatively coupled to the anatomical feature register 38. The control module 36 comprises, for example, one or more processors.

[0048] The processing unit 32 also includes a plurality of application modules 42a, 42b ... 42n operatively coupled to the control module. Each application module is adapted to process the ultrasound imaging data received at the input / output portion 34 based on the current value of at least one parameter of the one or more parameters stored in the anatomical feature register 38 to generate output quantization or segmentation data 22.

[0049] The control module 36 may also be adapted to receive user input data from the user interface 54 .

[0050] The control module 36 is adapted to dynamically update the value of at least one of the one or more parameters stored in the anatomical feature register 38. This may be based on user input data provided by a user via the user interface 54. Additionally or alternatively, this may be based on output quantization or segmentation data 22 produced by any of the plurality of application modules 42a, 42b ... 42n.

[0051] exist Figure 2 In the schematic illustration of , processing unit 32 is shown as a single integral component, and various components are located within the single integral component. However, this is not required. Processing 32 can be a distributed processing unit, wherein the various components are distributed in multiple locations. In addition, although the various components are shown as separate hardware modules, this is also not required. For example, different application modules 42 can be software modules that are all encoded and executed by a single hardware processing module.

[0052] For purposes of illustration, each application module 42 is shown as being communicatively coupled to the bus 48 to facilitate communication between these application modules and each of the input / output 34, the control module 36, and the anatomical feature registers 38. The link with the input / output allows the application modules in this example to directly access incoming ultrasound image data from the ultrasound data acquisition device 52. The link with the control module 36 allows the control module 36 to receive generated quantization data 22 from each application module, which in some cases can be used to update parameter entries in the anatomical feature registers 38. The link with the anatomical feature registers 38 allows each application module 42 to directly access the latest value of each parameter so that these values ​​can be used for quantification performed by each module.

[0053] One aspect of the invention is a separate processing unit 32. Another aspect of the invention is a system 30 including a processing unit 32 and an ultrasound data acquisition device 52 and / or a user interface device 54.

[0054] Regarding the ultrasound data acquisition device 52, this can be an ultrasound scanner. The ultrasound data acquisition device can include a transducer unit (e.g., an ultrasound probe or patch). It can also include a local dedicated user interface, which includes a display for displaying ultrasound images. The ultrasound data acquisition device 52 can include a processing component for processing the acquired echo data to generate ultrasound image data.

[0055] In operation, the processing and data flow may run, for example, as follows.

[0056] The ultrasound examination starts, and the user starts acquiring ultrasound image data using the ultrasound data acquisition device 52 .

[0057] The ultrasound data is received at the input / output 34 .

[0058] Various application modules 42 begin processing ultrasound data to derive quantification of anatomical structures. Some parameters stored in the anatomical feature registers 38 are quantities calculated as outputs of one or more application modules. These parameters can in turn be inputs used by other application modules. Thus, the output quantification or segmentation data of one or more application modules can be used to update the values ​​of one or more parameters in the anatomical feature registers. These parameters can be used by other application modules to perform their own quantification.

[0059] In addition, some of the one or more parameters in the feature register 38 can be transmitted to the user via the user interface 54 for confirmation or correction of the feature. For example, the graphical user interface may include an area that displays the current value of each parameter and optionally also displays the source of the current value (e.g., one of the application modules or the initial default value). In addition, in some examples, the segmentation generated by a specific application module for quantifying the size of the anatomical structure can be output to the user interface on a display. The user can input corrections or modifications to the segmentation using the user input element of the user interface. This can then be used by the control module 36 to update the relevant size parameters related to the anatomical structure in the anatomical feature register 38.

[0060] In a preferred embodiment, the processing unit may further include a register initialization module (not shown) adapted to generate initialization data for at least one of the one or more parameters stored in the anatomical feature register 38 based on the received ultrasound imaging data, and wherein the initialization data comprises an initial value for at least one of the one or more parameters. The control module 36 may further be adapted to trigger the initialization module to generate the initialization data and store the data in the anatomical feature register 38 as the initial value for at least one of the one or more parameters.

[0061] For example, the initialization data may be generated at the start of an ultrasound examination. For example, the method may further include receiving an examination start indicator signal indicating the start of an ultrasound examination session, for example from an ultrasound acquisition system, and wherein generating and storing the initialization data is performed in response to receiving the examination start indicator signal.

[0062] Now will describe the example implementation of the method according to a set of specific embodiments by illustrating the above summarized concepts of the present invention. It should be understood that not all features of this set of specific embodiments are necessary for the inventive concept and are described to help understand and provide examples to illustrate the inventive concept.

[0063] This particular set of embodiments relates to an ultrasound system comprising a plurality of application modules 42 for calculating anatomical and physiological quantities related to the heart.

[0064] Thus, in this set of embodiments, the anatomical structure is a cardiac structure. As an example, one or more parameters associated with the anatomical feature stored in the register may include one or more of the following: heart size; left ventricle (LV) size; left atrial size; right ventricle (RV) size; right atrial size; LV length; RV length; regional wall thickness; LV diameter. In some embodiments, the one or more parameters may also include one or more segmentation masks (or representations thereof) for one or more structures (such as for the LV or RV). The segmentation mask can be stored as a contour consisting of multiple x / y points. These points can be processed and stored in the register as anatomical parameters (scalars) that can be placed in a vector, for example, there can be 20 pairs of x / y image coordinates describing the LV contour. References to the above-mentioned "sizes" of, for example, the heart or LV or RV can be, for example, area parameters (e.g., across 2D planes) or volume parameters (e.g., estimated from 2D cross-sections or derived from 3D data).

[0065] Furthermore, there may also be parameter values ​​recorded for each parameter at both end-systole and end-diastole.The parameters may also include dynamic parameters, such as: ejection fraction, cardiac output or strain parameters.

[0066] The complete set of parameters may be stored as a vector, called a heart size or heart shape vector, where each element of the vector corresponds to one of the parameters.

[0067] The heart size or heart shape vector is saved in a register and is available to all application modules. The heart size or heart shape vector can be updated throughout the examination.

[0068] For example, initial values ​​for the different parameters contained in the heart size / shape vector may be set, for example, based on previous patient scan data or using a separate algorithm (described in more detail later).

[0069] Thereafter, more accurate values ​​for at least a subset of the parameters may be obtained from the quantification performed by one or more of the application modules 42. For example, at least one of the application modules may be adapted to output a value for heart size or left ventricle size, and this may be used to update these parameter values. Other applications may use the heart size and / or shape vectors to calculate other quantifications, such as cardiac output. Certain size or dimension parameters may be used as a basis for calculating other size or dimension parameters.

[0070] Additionally, based on the current values ​​of the parameters in the registers, acquisition parameters of the ultrasound data acquisition device may be determined and transmitted to the device.

[0071] As an example, in one system, there are multiple application modules, each receiving a 3D ultrasound image as input and generating a segmentation map of the heart (in the form of a grid) as output. Each of these example application modules begins its process of generating a segmentation map with an initial heart shape. The heart shape can be scaled by additional inputs such as heart size. The scaling then helps each application module converge to a better segmentation or prevent failure. In addition to heart size, other scaling parameters can be used, such as LV size or any other parameter mentioned in the present disclosure.

[0072] Thus, as a principle of more general application of any embodiment of the invention, for at least a subset of a plurality of application modules, the output quantification or segmentation data may include a segmentation of at least one anatomical feature or structure, and wherein the application module is adapted to generate the segmentation based on an initial shape template, and wherein the initial shape template is scaled according to a scaling parameter, and wherein the scaling parameter is at least one of one or more parameters stored in an anatomical feature register or calculated based on at least one of one or more parameters stored in an anatomical feature register. The resulting segmentation grid or mask can be used for a number of different purposes. It can, for example, be used to calculate one or more quantifications, thereby generating quantitative data. For example, one or more quantifications may include measurements such as ejection fraction, heart size, LV size, etc. It can also be used to derive a 2D planar image from a 3D input image. Thus, in some examples, the final output from the application module may be quantitative information, and wherein the segmentation is derived as an intermediate step based on one or more parameters in the register as input.

[0073] Additionally or alternatively, in some examples, the plurality of application modules may include one or more modules configured to quantify strain information, such as global longitudinal strain. For example, at least one module may be configured to receive a 2D ultrasound image as input and generate strain information as output. Here, one or more parameters in the feature register may be used to generate an initial estimate of the quantized strain information, and then, for example, iteratively refine the quantized strain information. The output is a strain measurement.

[0074] In some examples, one or more application modules can be adapted to perform a quality control check using one or more parameters stored in the anatomical feature registers. For example, a quality control check can be applied to a candidate value for a particular anatomical quantification based on a reference to one or more parameter values ​​in the registers, and if the candidate value passes the quality control check, the candidate value is output as output quantification data from the application module.

[0075] For example, the application module may be configured to calculate the heart size or LV size based on the input ultrasound image. For this purpose, the parameter values ​​in the feature register may not be needed. For example, an initial estimate of the heart size may not be needed. However, after the application module calculates a candidate value for the heart size, it can be compared with an existing value of the heart size in the feature register to check whether it is within a predefined range of the existing values ​​in the feature register. If it is outside the predefined range, the quality control check fails and the candidate value can be refined or discarded, for example.

[0076] Additionally or alternatively, in some embodiments, two or more of the application modules may be configured to calculate the same parameter in the parameters stored in the feature register, and wherein the method includes performing a comparison of multiple generated values ​​of the parameter with each other to check consistency. If the consistency meets a predefined consistency threshold, such as the values ​​of the parameter are within a predefined range of each other, the register can be updated based on the generated parameter values, such as selecting one of the parameter values ​​based on predefined criteria or calculating an average. If the consistency does not meet the predefined consistency threshold, the method can reject the generated parameter value and can, for example, wait for the value to be recalculated by the two or more application modules.

[0077] As described above, in some embodiments, an initial value of at least one parameter can be derived, for example, using an initialization module. For example, once an image covering a first cardiac chamber is acquired, a first estimate of the size of the heart or the size of the left ventricle can be derived by applying a dedicated initialization algorithm. As an example, a deep learning network can be applied to the acquired image data to obtain a first value of the size of the heart or the size of the left ventricle. As a more general principle, it is proposed that initialization data for at least one of the one or more parameters can be derived by applying an artificial intelligence algorithm to the received ultrasound imaging data. In a favorable example, the artificial intelligence algorithm can be a convolutional neural network.

[0078] Also as described above, additionally or alternatively, initialization data for at least one of the one or more parameters may be derived based on accessing data records in a data repository associated with the patient being examined. For example, in some embodiments, historical image data associated with the patient can be accessed and a quantization algorithm applied to the historical image data to derive an initial value for at least one of the one or more parameters. As a simpler example, the data repository may record the most recent value of a previously calculated parameter, and this may be used as initialization data for the parameter. For example, in the case of a longitudinal scan, a previous heart size can be determined from records in a PACS system and used as an initial estimate.

[0079] The initial values ​​for the one or more parameters may be used for initial quantization by one or more of the plurality of application modules.

[0080] As already discussed, a function is proposed by means of which the user can update or correct one or more parameters in the anatomical feature registers. For example, each parameter can be assigned an initial value at the beginning of the examination (according to initialization data), wherein this can be a default value, or can be a value obtained from a data record associated with the patient, or can be a value calculated by a dedicated algorithm or application module.

[0081] The user interface device 54 can display a graphical user interface window on its display unit, and this can include an area indicating the current value of each parameter in a plurality of parameters associated with the anatomical features of the anatomical structure. For example, there may be an area indicating the current estimate of the heart size, LV size, or any other parameter mentioned previously. In addition, the graphical user interface can also indicate the source of the parameter value for each parameter, such as initialization data (e.g., a deep learning neural network), historical patient data, a specified application module in an application module, or user input data. The user interface can be configured to allow the user to input a modified value of at least a subset of these values. In other words, the user can overwrite information. This may be valuable, especially at the beginning of the examination, before the application module 42 has generated quantitative data. Various parameters are stored in registers in the background.

[0082] In prior art ultrasound systems, all application modules used to perform quantification operate independently of each other. This means that if more than one application module utilizes the same input parameter related to heart size, the parameter must be updated multiple times, once for each application. Embodiments of the present invention provide improved parameter sharing as well as other benefits.

[0083] These initial values ​​may remain in place in registers for use by each application module 42 to which they are associated unless and until updated.

[0084] Additionally or alternatively, the user can input updates or corrections to one or more parameters via the user interface 54, such as manual corrections to the segmentation.

[0085] According to some embodiments it is proposed that once a user has updated a given parameter, editing or updating of that parameter should be locked for the remainder of a given examination. For example, this may be achieved by including an update counter entry in the anatomical feature register 38 associated with at least one of the one or more parameters. It is proposed that concurrently with any update of the parameter value in the anatomical feature register based on user input data, the update counter entry is incremented, and wherein, in response to the current value of the update counter register exceeding a threshold, any update of the parameter in the anatomical feature register is blocked. The threshold may be set to zero so that once a user has updated or corrected a parameter once in a given examination, it remains locked for the remainder of the examination.

[0086] Resetting between examination sessions may be achieved by a reset indicator signal generated at the end or start of any ultrasound examination session, and wherein the method comprises resetting the update counter entry to zero in response to the reset indicator signal.

[0087] As described above, in some embodiments, initialization data is generated to provide initial values ​​for one or more parameters stored in the anatomical feature registers. In some embodiments, it is proposed to calculate the initialization data using an artificial intelligence algorithm applied to the input ultrasound image data.

[0088] The inventors developed a new convolutional neural network for this purpose.

[0089] The generated model is a deep learning model for estimating heart size. It can be designed as a convolutional neural network (CNN) that takes a 2D, 3D, 2D+t, or 3D+t ultrasound image dataset as input and generates a heart size estimate as output. The convolutional neural network can be configured in an end-to-end manner so that the heart size parameters are directly regressed into continuous parameters.

[0090] A convolutional neural network (CNN) according to the present specification was implemented, trained and tested by the inventors. The CNN was trained to predict end-diastolic (ED) and end-systolic (ES) left ventricular (LV) volumes based on an input 2D+t ultrasound image set. The 2D+t image data set includes a series of 2D image frames, each 2D image frame corresponding to a different time point.

[0091] The CNN uses 3D convolutions and includes seven convolutional layers, a global average pooling layer, and an output layer with two outputs. There are three resolution levels. Downsampling between stages is performed inside the convolution with a stride of 2. The model is trained using a training dataset of 25,000 ultrasound DICOM (2D+t) image datasets.

[0092] The resulting performance of the CNN model is Figure 3 This shows the correlation between the predicted values ​​of heart size (y-axis) and the ground truth (x-axis). Quantitatively, the CNN achieved a mean absolute error of 18.60 ml for ED volume and 11.44 ml for ES volume over a large volume range (see Figure 2 ). The Pearson correlation coefficient was 0.864 for ED volume and 0.899 for ES volume. As an estimate of rough heart size, this performance is adequate.

[0093] In the case of implementing the invention in hardware, there are different options on how to configure the hardware. For example, physically separate processing units may be provided, such as Figure 2As shown. In other embodiments, some or all of the functions attributed above to a single processing unit may be distributed among multiple processing devices. In some embodiments, the functions of the processing unit 32 may be integrated into the processing unit of the ultrasound data acquisition device 52. In some embodiments, the processing unit may be a cloud-based processing unit, and wherein communication with the ultrasound device 52 and the user interface 54 is facilitated via an Internet link.

[0094] As mentioned above, the present invention can be embodied in software form. Therefore, another aspect of the present invention is a computer program product, which includes a code module, which is configured to cause the processor to perform a method according to the examples or embodiments of the present invention described in this document or according to any claim of this patent application when it is run on a processor.

[0095] Some embodiments of the present invention described above adopt processing units. The processing unit may generally include a single processor or multiple processors. It may be located in a single containing device, structure or unit, or it may be distributed between multiple different devices, structures or units. Therefore, references to a processing unit adapted or configured to perform a specific step or task may correspond to the step or task performed by any one or more of a plurality of processing components individually or in combination. Those skilled in the art will understand how such distributed processing arrangements can be implemented. The processing unit may include a communication module or input / output portion for receiving data and outputting data to other components.

[0096] The one or more processors of the processing unit can be implemented in a variety of ways using software and / or hardware to perform the various functions required. The processors typically employ one or more microprocessors that can be programmed using software (e.g., microcode) to perform the required functions. The processor can be implemented as a combination of dedicated hardware for performing some functions and one or more programmed microprocessors and associated circuits for performing other functions.

[0097] Examples of circuits that may be employed in various embodiments of the present disclosure include, but are not limited to, conventional microprocessors, application specific integrated circuits (ASICs), and field programmable gate arrays (FPGAs).

[0098] In various embodiments, the processor may be associated with one or more storage media, such as volatile and non-volatile computer memory, such as RAM, PROM, EPROM, and EEPROM. The storage media may be encoded with one or more programs that, when executed on one or more processors and / or controllers, perform the desired functions. The various storage media may be fixed within the processor or controller, or may be removable so that one or more programs stored thereon can be loaded into the processor.

[0099] By studying the drawings, the disclosure and the appended claims, those skilled in the art can understand and implement variations of the disclosed embodiments. In the claims, the word "comprising" does not exclude other elements or steps, and the indefinite article "a" or "an" does not exclude a plurality.

[0100] A single processor or other unit may fulfill the functions of several items recited in the claims.

[0101] The mere fact that certain measures are recited in mutually different dependent claims does not indicate that a combination of these measures cannot be used to advantage.

[0102] The computer program may be stored / distributed on suitable media such as optical storage media or solid-state media supplied together with or as part of other hardware, but may also be distributed in other forms such as via the Internet or other wired or wireless telecommunications systems.

[0103] If the term "suitable for" is used in the claims or the specification, it should be noted that the term "suitable for" is intended to be equivalent to the term "configured to".

[0104] Any reference signs in the claims should not be construed as limiting the scope.

Claims

1. A computer-implemented method comprising: receiving (12) ultrasound imaging data of an anatomical structure of a patient obtained from an ultrasound acquisition system; communicating with an anatomical feature register (14) that stores one or more parameters associated with anatomical features of the anatomical structure; a plurality of application modules, each application module processing (16) the acquired ultrasound imaging data based on a current value of at least one of the one or more parameters associated with the anatomical feature currently stored in the anatomical feature register to produce output quantized or segmented data (22); as well as receiving (18) input data from a user interface; Wherein, the method further comprises: The value of at least one of the one or more parameters stored in the anatomical feature register is updated (20) based on the input data received from the user interface and / or the output quantization or segmentation data generated by any of the plurality of application modules.

2. The method according to claim 1, further comprising: generating initialization data for at least one of the one or more parameters for the anatomical feature based on the received ultrasound imaging data, and The initialization data is stored in the anatomical feature register as an initial value for the at least one of the one or more parameters.

3. The method according to claim 2, wherein: The method further includes receiving an examination start indicator signal, for example from the ultrasound acquisition system, indicating the start of an ultrasound examination session, and Wherein, generating and storing the initialization data are performed in response to receiving the check start indicator signal.

4. The method according to claim 2 or 3, wherein: Generating the initialization data for the at least one of the one or more parameters includes applying an artificial intelligence algorithm to the received ultrasound imaging data.

5. The method according to claim 4, wherein: The artificial intelligence algorithm is a convolutional neural network.

6. The method according to any one of claims 1 to 5, wherein: The anatomical structure is a cardiac structure, and wherein the one or more parameters include one or more of: heart size, left ventricle size, left atrial size, right ventricle size, end-diastolic right atrial size, ejection fraction and / or left ventricular length.

7. The method according to any one of claims 1 to 6, wherein: The received ultrasound imaging data includes: 2D ultrasound imaging data for one or a series of time frames, or 3D ultrasound imaging data for one or a series of time frames.

8. The method according to any one of claims 1 to 7, wherein: The anatomical feature register comprises an update counter entry associated with at least one of the one or more parameters, and wherein the update counter entry is incremented concurrently with any update of the parameter value in the anatomical feature register based on the user input data, and wherein any update of the parameter in the anatomical feature register is prevented in response to a current value of the update counter register exceeding a threshold value.

9. The method according to claim 8, wherein: The threshold is zero.

10. The method according to claim 8 or 9, wherein: A reset indicator signal is generated at the end or start of any ultrasound examination session, and wherein the method comprises resetting the update counter entry to zero in response to the reset indicator signal.

11. The method according to any one of claims 1 to 10, wherein: At least a subset of the application modules are each adapted to generate a segmentation based on an initial shape template, wherein the initial shape template is scaled according to a scaling parameter, and wherein the scaling parameter is at least one of the one or more parameters stored in the anatomical feature register, or is calculated based on at least one of the one or more parameters stored in the anatomical feature register.

12. The method according to any one of claims 1 to 11, wherein: The received ultrasound image data is an image data stream, and wherein the plurality of application modules are each adapted to repeatedly or continuously operate on the received image data stream.

13. A computer program product comprising code modules, the code modules being configured to perform the method according to any one of claims 1 to 12 when run by a processor.

14. A processing unit (32), comprising: An input / output unit (34) is used to connect to an ultrasonic data acquisition device to receive ultrasonic imaging data of the patient's anatomical structure; an anatomical feature register (38) adapted to store one or more parameters associated with anatomical features of the anatomical structure; a control module (36) operatively coupled to the anatomical feature register; a plurality of application modules operatively coupled to the control module, each application module being adapted to process ultrasound imaging data received at the input / output portion based on a current value of at least one of the one or more parameters stored in the anatomical feature register to produce output quantized or segmented data (22); and Wherein, the control module is suitable for: receiving user input data from a user interface (54); and A value of at least one of the one or more parameters stored in the anatomical feature register is updated based on the user input data provided by a user through the user interface and / or the output quantization or segmentation data generated by any application module of the plurality of application modules.

15. The processing unit according to claim 14, further comprising: a register initialization module adapted to generate initialization data for at least one of the one or more parameters based on the received ultrasound imaging data, the initialization data comprising an initial value for at least one of the one or more parameters, and The control module is further adapted to trigger the initialization module to generate the initialization data and store the data in the anatomical feature register as an initial value for at least one of the one or more parameters.

16. The processing unit according to claim 14 or 15, wherein: The anatomical structure is a cardiac structure, and wherein the one or more parameters include one or more of: heart size, left ventricle size, left atrial size, right ventricle size, end-diastolic right atrial size, ejection fraction and / or left ventricular length.

17. A system (30), comprising: A processing unit (32) according to any one of claims 14 to 16; as well as an ultrasound data acquisition device (52) operatively coupled to the input / output portion of the processing unit, and Optionally, a user interface (54) for obtaining user input is also included.

Images