Obtaining medical image at target plane

Abstract

A method for obtaining a medical image at a target plane. The method includes obtaining a three-dimensional (3D) volume containing anatomical structures, and identifying the anatomical structures in the 3D volume. The geometry of the anatomical structure in the 3D volume is analyzed, and the target plane is registered to the 3D volume based on the analysis of the geometry of the anatomical structure. A medical image at the registered target plane is extracted from the 3D volume.

Description

Technical Field

[0001] The present invention relates to a method for obtaining a medical image at a target plane. In particular, the method relates to extracting a medical image at a target plane from a 3D volume. Background Art

[0002] Longitudinal medical studies utilizing 2D images can be difficult to read due to differences in acquisition (e.g., angle, field of view (FOV), etc.). In longitudinal studies, it is important to maximize the comparability of examinations performed at different time points. If scanning is performed in 2D, the angle of the imaging device (e.g., ultrasound transducer) determines the anatomical plane of the resulting scan, and deviations in this angle can result in non-comparable examinations compared to a previous reference examination.

[0003] Users must compensate for any effects of different acquisitions when comparing examinations over time in longitudinal studies. Simple registration of images may not work because rigid registration cannot always align views, and elastic registration may destroy subtle differences in anatomy.

[0004] Therefore, there is a need to improve the comparability of examinations in longitudinal studies.

[0005] US2021 / 338203 discloses a system for guiding users to acquire ideal 2D images. It utilizes 3D+t volume scanning and segmentation. The result is a single 2D projection of a given view plane determined at the time of the scan.

[0006] US2020 / 13482 relates to a general method for multiplanar reforming (MPR). Summary of the Invention

[0007] The invention is defined by the claims.

[0008] According to an example of one aspect of the present invention, there is provided a computer-implemented method for obtaining a medical image at a target plane, the method comprising:

[0009] Obtain a three-dimensional 3D volume containing anatomical structures;

[0010] identifying the anatomical structure in the 3D volume;

[0011] analyzing a geometry of the anatomical structure in the 3D volume;

[0012] registering the target plane to the 3D volume based on the analysis of the geometry of the anatomical structure; and

[0013] The medical image at a registered target plane is extracted from the 3D volume.

[0014] This allows the user to obtain a 3D volume / dataset without having to precisely utilize the orientation of the imaging device when capturing the target view / plane. This is because the target plane can be extracted from the 3D volume regardless of its orientation. 2D planes can be rendered (with comparable image quality) by cutting a dedicated MPR from the 3D volume.

[0015] In order to accurately extract the target plane from the 3D volume, the anatomical structure in the 3D volume is identified and the geometry of the anatomical structure is analyzed. This analysis of the geometry allows the target plane to be accurately aligned / registered with the 3D volume.

[0016] The first aspect involves:

[0017] extracting a plurality of medical images near a registered target plane from the 3D volume,

[0018] determining a quality score of the extracted medical image; and

[0019] A medical image is selected from the extracted plurality of medical images based on the quality score.

[0020] In this way, medical images near the target plane are selected based on the quality assessment, such that the medical images are both high-quality images and capture desired anatomical features (because the desired anatomical features are near the target plane).

[0021] The second aspect involves:

[0022] A target plane is selected, wherein the target plane comprises a plane of a previous 2D scan relative to the time when the 3D volume was acquired, or a plane of a future 2D scan relative to the time when the 3D volume was acquired.

[0023] The 2D target plane can be determined at review time rather than at scan time, which enables the application of matching the view plane to match any desired plane at a later time. This can correspond to a previous or future 2D scan (relative to the time the 3D volume was imaged).

[0024] Thus, a previous 2D scan refers to a scan performed at a time earlier than the time the 3D volume was acquired. A future 2D scan refers to a scan performed at a time later than the time the 3D volume was acquired. However, images can be extracted after the later 2D scan, for example, to fill in gaps in the longitudinal analysis of the 2D images. Thus, a 3D volume is captured at a certain time and can be analyzed later to provide 2D images corresponding to the 2D scan acquired at the later time.

[0025] The first aspect and the second aspect can be used alone or in combination.

[0026] In both aspects, analyzing the geometry of the anatomical structure may include matching features / landmarks in the identified anatomical structure with expected features in the medical image at the target plane. The geometry may include the size of the anatomical features, the distance between anatomical features, etc.

[0027] Registering a first plane / image / volume to a second plane / image / volume typically involves determining the transformation required to transform the first plane / image / volume to the coordinate system of the second plane / image / volume.

[0028] Identifying the anatomical structure may include using a model-based segmentation algorithm trained on a model of the anatomical structure.

[0029] Model-based segmentation improves segmentation accuracy because the segmentation algorithm has prior information about the anatomical structure it is segmenting. Often, medical scans are performed on a targeted anatomical structure. For this reason, model-based segmentation is particularly suitable because various algorithms can be trained on the target anatomical structure (e.g., a heart model, a fetal model, etc.).

[0030] Registering the target plane to the 3D volume may include registering the identified anatomical structure to a model of the anatomical structure, wherein the target plane is pre-registered to the model of the anatomical structure.

[0031] Different models can be used for different purposes. For example, for cardiac imaging, different models can be used for different types of hearts (e.g., pediatric, adult, etc.) and / or for different times during the cardiac cycle. A general model can also be transformed for different purposes.

[0032] Extracting (according to the first aspect) the plurality of medical images in the vicinity of the registered target plane from the 3D volume may comprise extracting the medical images at the registered target plane.

[0033] Medical images extracted at the target plane may not be of suitable quality (or may not be of optimal quality). For this reason, it is proposed to extract medical images at other planes near the target plane. Therefore, medical images with higher quality scores at nearby planes may be more suitable, for example, for comparison with other images taken at different times.

[0034] This is particularly advantageous when the user wants to compare 2D slices from two or more 3D volumes, as the 2D slices can all be extracted at planes near the target plane, which provides appropriate quality for all 2D slices.

[0035] Each medical image may have a single quality score. Alternatively, each medical image may have multiple quality scores.

[0036] The quality score can be based on noise level, artifact count, field of view coverage, visibility of anatomical structures, etc. In a first example, the quality score of a medical image is a composite score of the above criteria. In a second example, each medical image has multiple individual quality scores for the above criteria.

[0037] The 3D volume can be labeled as a first 3D volume, and the method may further include obtaining a second 3D volume containing the anatomical structure, identifying the anatomical structure in the second 3D volume, analyzing a geometry of the anatomical structure in the second 3D volume, registering the target plane to the second 3D volume based on the analysis of the geometry of the anatomical structure in the second 3D volume, and extracting a second medical image at the registered target plane from the second 3D volume.

[0038] Extracting the second medical image may include extracting a plurality of second medical images near the registered target plane from the second 3D volume, determining a quality score of the second medical images, and selecting a second medical image from the plurality of second medical images based on the quality score.

[0039] Extracting the first medical image (from the first 3D volume) and extracting the second medical image can be based on a quality score of the extracted first medical image (of the first 3D volume) and a quality score of the second medical image. The planes of the extracted first medical image and the extracted second medical image can be the same (or within a maximum displacement threshold). This enables comparability between the two extracted medical images with appropriate quality, without being limited to a strict target plane (which may, for example, be noisy).

[0040] During the 3D acquisition of the first 3D volume and / or the second 3D volume, the acquisition may sometimes be gated to obtain the anatomy at the same point during the cycle (eg, for cardiac imaging).

[0041] The method may further include obtaining a two-dimensional (2D) medical scan of the anatomical structure, and determining an anatomical plane from the 2D medical scan, wherein the target plane is the determined anatomical plane.

[0042] This enables extraction of medical images from the 3D volume that match planes of separately acquired medical scans, enabling comparability of such medical images (eg, comparing anatomical structures at a target plane from different examinations).

[0043] The method may further include obtaining a 3D reference volume registered with the 2D medical scan, wherein determining the anatomical plane from the 2D medical scan includes identifying the anatomical plane relative to the 3D reference volume.

[0044] The method may further include obtaining a first 2D medical scan of the anatomical structure at a first time, obtaining a second 2D medical scan of the anatomical structure at a third time, obtaining the 3D medical volume at a second time between the first time and the third time, and determining a first anatomical plane from the first 2D medical scan and / or a second anatomical plane from the second 2D medical scan, wherein the target plane is one of the first anatomical plane or the second anatomical plane.

[0045] Thus, the target plane may be selected as a plane of a previously acquired 2D scan (at a previous first time) or a plane of a future 2D scan (at a future third time).

[0046] For example, a user may be able to select whether the extracted image is based on an image obtained previously (eg, at a first time) or based on an image obtained later (eg, at a third time).

[0047] Determining anatomical planes from a 2D medical scan may include using a regression network trained to identify anatomical planes from any 2D medical scan that includes anatomical structures.

[0048] In other words, plane regression techniques can be used to identify anatomical planes of 2D medical scans.

[0049] Extracting the medical image from the 3D medical volume may include using multi-planar reformatting (MPR) on the 3D medical volume.

[0050] The 3D medical volume may be a 3D ultrasound volume.Of course, other types of 3D medical data may be used.

[0051] The invention also provides a computer program carrier comprising computer program code, which, when executed on a processing system, causes the processing system to perform all the steps of the aforementioned method.

[0052] The computer program carrier may be, for example, a data storage system (eg hard drive, solid-state drive, etc.) or a transient carrier (eg a bit stream).

[0053] The present invention also provides a system for obtaining a medical image at a target plane, the system comprising a processor, wherein the processor is configured to:

[0054] Obtain a three-dimensional 3D volume containing anatomical structures;

[0055] identifying the anatomical structure in the 3D volume;

[0056] analyzing a geometry of the anatomical structure in the 3D volume;

[0057] registering the target plane to the 3D volume based on the analysis of the geometry of the anatomical structure; and

[0058] The medical image at a registered target plane is extracted from the 3D volume.

[0059] In one aspect, a plurality of medical images in the vicinity of a registered target plane are extracted from the 3D volume, and the processor is further configured to:

[0060] determining a quality score of the extracted medical image; and

[0061] A medical image is selected from the extracted plurality of medical images based on the quality score.

[0062] In another aspect, the processor is further configured to receive as input a target plane, wherein the target plane comprises: a plane of a previous 2D scan relative to a time when the 3D volume was acquired, or a plane of a future 2D scan relative to a time when the 3D volume was acquired;

[0063] The system may further comprise an imaging device for obtaining the 3D volume. For example, the imaging device may be a 3D ultrasound transducer.

[0064] The processor may be configured to identify the anatomical structure by using a model-based segmentation algorithm trained on a model of the anatomical structure.

[0065] The 3D volume may be labeled as a first 3D volume, and the processor may be further configured to obtain a second 3D volume containing the anatomical structure, identify the anatomical structure in the second 3D volume, analyze a geometry of the anatomical structure in the second 3D volume, register the target plane to the second 3D volume based on the analysis of the geometry of the anatomical structure in the second 3D volume, and extract a second medical image at the registered target plane from the second 3D volume.

[0066] The processor may be further configured to obtain a two-dimensional (2D) medical scan of the anatomical structure, and determine an anatomical plane from the 2D medical scan, wherein the target plane is the determined anatomical plane.

[0067] 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

[0068] For a better understanding of the invention, and in order to show more clearly how it may be put into practice, reference will now be made, by way of example only, to the accompanying drawings, in which:

[0069] Figure 1 The diagram shows a 2D scan of the anatomy at three separate times;

[0070] Figure 2 The diagram shows a 3D scan of the anatomy at three separate times;

[0071] Figure 3 The 3D acquisition process is shown;

[0072] Figure 4 A method for obtaining a target plane from a 2D slice is illustrated;

[0073] Figure 5 showing a time series of scans, the time series of scans comprising a first 2D scan, a second 3D scan, a third 3D scan, a fourth 2D scan, and a fifth 3D scan; and

[0074] Figure 6 A method for obtaining a medical image at a target plane using a three-dimensional 3D volume containing an anatomical structure is shown. DETAILED DESCRIPTION

[0075] The present invention will be described with reference to the accompanying drawings.

[0076] It should be understood that the detailed description and specific examples, while indicating exemplary embodiments of the apparatus, system, and method, are intended for illustrative purposes only and are not intended to limit the scope of the invention. These and other features, aspects, and advantages of the apparatus, 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 schematic only and are not drawn to scale. It should also be understood that throughout the drawings, the same reference numerals are used to indicate the same or similar parts.

[0077] The present invention provides a method for obtaining a medical image at a target plane. The method includes obtaining a three-dimensional (3D) volume containing an anatomical structure and identifying the anatomical structure in the 3D volume. The geometry of the anatomical structure in the 3D volume is analyzed, and based on the analysis of the geometry of the anatomical structure, a target plane is registered to the 3D volume. A medical image at the registered target plane is extracted from the 3D volume.

[0078] Figure 1 The diagram shows a 2D scan of the anatomical structure 102 at three separate times. Figure 1 , for illustration purposes only, the anatomical structure 102 is shown as a heart. It should be understood that other anatomical structures (e.g., lungs, liver, fetus, etc.) may be imaged instead.

[0079] In order for a user to compare 2D acquisitions (i.e., 2D scans or 2D medical images) at three separate times t1, t2, and t3, an imaging device (e.g., a transducer for ultrasound imaging) (not shown) must be aligned all three times to the target plane 104. This means that the angle of the imaging device must be carefully reproduced (e.g., at each separate examination) so that the imaging device's field of view 106 captures the target plane 104 each time.

[0080] If the imaging device is not correctly aligned, the 2D acquisition may be at a different plane and therefore will not be comparable with earlier / later acquisitions obtained at the target plane.

[0081] It is proposed to solve this problem by acquiring a 3D volume of the anatomical structure 102 and then cutting 2D anatomical planes from the 3D acquisition.

[0082] Ultrasound imaging is currently moving towards 3D volume acquisition. Scanning can be performed using a 3D matrix probe. 2D slices can then be extracted from the 3D volume using predefined slices or parametric slices. However, such slice techniques may not always provide 2D slices at the exact target plane.

[0083] Figure 2 The 3D scan of the anatomical structure 102 at three separate times is shown. In this case, the angle of the imaging device (not shown) is not a critical condition because, unlike in a 2D acquisition (e.g. Figure 1 The field of view 202 is much larger during a 3D acquisition than during a 3D acquisition (as shown). As such, the target plane 104 can be sliced ​​out / extracted from the 3D acquisition. In other words, slicing out a 2D anatomical plane from a 3D acquisition relaxes the dependency on the imaging device angle during acquisition, since planes that are not axis-aligned with the imaging device, such as the plane 104, can be sliced ​​out of the volume. Figure 2 shown.

[0084] Because of this, the user does not have to be as precise as they would be in positioning / angling the imaging device at different times (e.g., during different examinations in a longitudinal study). In some cases, the user may want to compare the current 2D slice at the target plane 104 with a previous 2D slice. This is only possible if the user has also previously acquired a 2D slice at the target plane 104. However, the user may have previously acquired a 3D volume of the anatomical structure 102. Therefore, the user can extract a 2D slice at the target plane from the previously acquired 3D volume to compare with the current 2D slice.

[0085] Figure 3 The 3D acquisition process is shown. Specifically, Figure 3The anatomical structure scanned at three different times t1, t2, and t3 during three different examinations is shown. During each examination, the angle of the imaging device (not shown) is not a critical condition because the 3D acquisition is able to obtain all necessary information about the anatomical structure 102 and store it in the 3D volume 302 due to the large field of view 202.

[0086] The 3D volume 302 can be stored in, for example, Cartesian or polar coordinates. The 3D volume can be sliced ​​(e.g., at a later date) to obtain a medical image 304 at any target plane. Thus, even if a longitudinal study requires only 2D acquisition (i.e., multiple examinations over time), storing the 3D volume 302 enables the user to extract / slice out the 3D volume at the target plane as needed to obtain comparable medical images 304.

[0087] In order to obtain comparable (ie, at the same target plane) images from non-registered volumes, it is important to be able to accurately align the target plane to each volume so that the 2D slices are accurately aligned to the same target plane.

[0088] The target plane can be registered to the volume(s) by first identifying the anatomical structures in the 3D volume. For example, a segmentation algorithm can be used to identify the anatomical structures in the 3D volume. Additionally, anatomical model-based segmentation can be used to more accurately identify the anatomical structures. For example, a segmentation algorithm can be trained from one or more models of anatomical structures (e.g., heart, fetus, etc.) to more accurately identify the anatomical structures in the 3D volume.

[0089] The segmented structure can be treated as an anatomical model. This means that the geometry of the features expected in the target plane can be mapped to the features in the segmented anatomical structure. A standard target plane can be carefully defined and verified using a combination of vertices of a certain anatomical structure (e.g., center of gravity, axis, distance, etc.).

[0090] For example, in the case of cardiac ultrasound imaging, a target plane can be defined by taking the vertices of all the "mitral annulus" and calculating the mitral valve center (which also forms the image center). With the first axis point toward the left ventricular apex (which is defined by another set of vertices) and the second axis point toward the aortic valve center (calculated as described above), a coordinate system can be derived to align the target plane to the 3D volume. Finally, the size of the medical image can be given by a scaling factor used to adapt the anatomical model (e.g., during segmentation) to the acquired 3D volume.

[0091] In other words, a standard target plane can be defined in the cardiac coordinate system of the anatomical model and transformed to the 3D volume in the same way as the anatomical structure (eg, during model-based segmentation).

[0092] More generally, the geometry of features in the 3D volume can be matched to the geometry of features expected in the image at the target plane in order to register the target plane to the 3D volume.

[0093] The target plane may be a plane at or near a standard plane that has been registered to an anatomical coordinate system (eg, the coordinate system of a model of a cardiac structure).Thus, the target plane can be registered to the segmented anatomical structure using the geometry of the segmented anatomical structure.

[0094] Additionally, arbitrary planes can be mapped from one volume to another by describing the plane with reference to a segmented mesh of the anatomical structure in the first volume, transforming the plane reference to the second volume via corresponding vertices in the mesh (which can undergo affine but non-rigid deformations), and finally deriving the corresponding plane in the second volume. In the case of arbitrary planes, it is preferred to use an anatomical model to establish vertex-to-vertex correspondences so that arbitrary planes can be established for all volumes.

[0095] In the case of longitudinal studies, with at least one additional examination performed subsequently, it is possible to run a combined optimization on all studies consisting of: segmenting each 3D volume (e.g., via an anatomical model); extracting a target anatomical plane (e.g., a standard target plane or an arbitrary target plane) in which to examine; and selecting slices to maximize comparability across the various studies, allowing for small deviations from the target plane.

[0096] When selecting slices to compare, optimization can take into account noise level, artifact count, FOV coverage, visibility of anatomical structures, etc. This allows a series of optimized corresponding 2D slices to be comparable for each exam / time point, resulting in higher diagnostic confidence.

[0097] To compensate for motion, residual errors in the segmentation, or noise in the extracted 2D slices, it may be preferable to search for comparable 2D slices near the target plane. Of course, since some anatomical structures may be ambiguous, a global search can be performed over the 3D volume. This may mean, for example, finding an arbitrary target plane that provides the best comparison between the 3D volumes. However, this approach may yield suboptimal results, as the arbitrary plane may not show the desired features of the anatomical structure.

[0098] In some cases, a 2D scan of the anatomical structure may already exist at the target plane (e.g., from a previous study / examination). In these cases, the 2D scan can be used to obtain the target plane, thereby extracting additional 2D slices from the 3D volume of the anatomical structure.

[0099] Figure 4The diagram illustrates a method for obtaining a target plane from a 2D slice. In a longitudinal study, if at least one examination is a 2D scan 402 performed with a different technique or probe, or if no 3D information is available for the examination, it is recommended to use the 2D scan 402 to obtain the target plane 408. Thus, the 3D volume can be sliced ​​using the target plane 408, enabling comparison of the 2D slices.

[0100] For each 2D image 402, a plane regression technique can be used to determine an anatomical plane. The input 2D image 402 is segmented to identify the boundaries of the anatomical structure 404. The plane parameters can then be derived by regressing a convolutional neural network that has been trained using, for example, artificial contours and plane parameter pairs. As such, the plane parameters can be used to define a target plane 408 relative to the anatomical model 410.

[0101] As previously described, the 3D volume can then be segmented using a cardiac model or deep learning segmentation methods, and corresponding slices can be generated using the determined planar parameters of the target plane 408, for example using multi-planar reformat (MPR).

[0102] Longitudinal series can now be presented for any plane found in a 2D scan.For multiple 2D scans, the user can select any 2D scan and thus reformat the 3D acquisition to optimally examine changes over time in orientation relative to the selected scan.

[0103] Similar techniques can be used in 2D workflows performed with 3D probes. Most scans during the exam can still be recorded in 2D. However, at certain times (e.g., during each frame at end-diastole), the system can record a single frame 3D image (possibly with lower resolution) in the background. These recorded 2D slices will now have a 3D reference that can be analyzed. In other words, the plane of the 2D slice will be aligned to the 3D model of the anatomical structure (i.e., acquired via a 3D reference). In this way, the 2D workflow that many doctors are accustomed to is retained, and the target plane can be obtained for later use (e.g., to obtain additional 2D slices at the target plane from a later 3D acquisition). This can be used to not sacrifice frame rate when scanning (e.g., for valves or larger field of view scans).

[0104] Using a 3D reference can provide a more accurate target plane because the target plane is already defined relative to the 3D reference of the anatomy, so there is no need to apply plane regression to the 2D slices. For some target planes (e.g., the short axis plane in cardiac ultrasound), plane regression may not always be accurate.

[0105] Additionally, using the above techniques, subsequent scans can be compared to each other and 2D slices can be obtained from subsequent 3D acquisitions if, for example, the acquisition planes are not aligned with the previous acquisition.

[0106] If a 2D image at the target plane is not acquired, it is not possible to reformat the already acquired 2D image to best match across the longitudinal scan. However, it is still possible to derive a 2D image at the target plane using a 3D reference frame.

[0107] If knowledge of the target plane is used during scanning, the 2D parameters can be adapted to capture the corresponding target plane during scanning. Knowledge of the target plane can also be used retrospectively to identify deviations from the target plane. Thus, possible flaws in the measurement can also be discovered.

[0108] The gaps in the longitudinal series of 2D scans can be filled using 2D slices sliced ​​out / extracted from the 3D volume of the anatomy.

[0109] Figure 5 A time series of scans is shown, including a first 2D scan 502, a second 3D scan 504, a third 3D scan 508, a fourth 2D scan 512, and a fifth 3D scan 514. The first 2D scan 502 and the fourth 2D scan 512 may be scans of the anatomical structure at the same target plane. To fill the gap between the first 2D scan 502 and the fourth 2D scan 512, 3D scans acquired between (and possibly before and / or after) the 2D scans can be sliced ​​at the target plane corresponding to the first 2D scan 502 and / or the fourth 2D scan 512.

[0110] Extracting 2D scans from a 3D volume can involve using multi-planar reformatting (MPR) on the 3D scans at a target plane. The target plane can be obtained by identifying a plane of the first 2D scan 502 and / or identifying a plane of the fourth 2D scan 512. The second 3D scan 504 can then be sliced ​​at the target plane to obtain MPR slices 506. Similarly, the third 3D scan 508 can be sliced ​​at the target plane to obtain MPR slices 510, and the fifth 3D scan 514 can be sliced ​​at the target plane to obtain MPR slices 516. Thus, the first 2D scan 502 and the fourth 2D scan 512 can be compared to the MPR slices 506, 510, and 516.

[0111] Typically, if a 2D scan is missing from the time series, the target plane can be detected or estimated from the other 2D scans and then sliced ​​from the 3D volume. This allows the sonographer to review comparable slices for exams where the 2D planes were not initially acquired or were forgotten to be acquired.

[0112] Thus, the target plane may be selected as the plane of a previously acquired 2D scan or the plane of a future 2D scan.

[0113] The operator / user can decide to align the MPR slices with any 2D scan by selecting an MPR slice (e.g., by clicking on it), and thus match the MPR slices to the selected scan. In the case of selecting multiple 2D slices, the matching MPRs for each of these 2D slices can be displayed simultaneously. This can provide further insight into whether the differences in the 2D scans are simply due to different locations (e.g., if such differences were also present when displaying the corresponding MPRs from a single 3D scan) or due to disease progression.

[0114] Figure 6 A method for obtaining a medical image at a target plane using a three-dimensional (3D) volume containing an anatomical structure is shown. In step 602, the anatomical structure is identified in the 3D volume. For example, a segmentation algorithm can be applied to the 3D volume. The segmentation algorithm can be a model-based segmentation algorithm trained on one or more models of the anatomical structure.

[0115] The target plane is then registered to the 3D volume in step 604. This can be accomplished by analyzing the geometry of the segmented anatomical structure. For example, the segmented anatomical structure can be registered to a model of the anatomical structure to which the target plane is already registered (e.g., a segmented anatomical structure obtained from a previous examination) using the relative positions of features in the anatomical structure. The target plane can then be transformed to the segmented anatomical structure. A medical image at the registered target plane can then be extracted from the 3D volume in step 606.

[0116] Instead of extracting a single image at the target plane, multiple images near the registered target plane can be extracted. Multiple images can be extracted at a set of parallel planes adjacent to the target plane, or non-parallel planes can be defined as intersecting the target plane.

[0117] A parallel plane can be derived by shifting the original plane along its normal vector. For example, this can be done by shifting the plane using discrete offsets in two directions (e.g., 1 mm, 2 mm, 3 mm in two directions).

[0118] Non-parallel planes can be derived in a variety of ways, for example, by determining a point of the plane (e.g., the center of the plane or the center of the anatomical structure of interest) as the rotation point. An axis of the plane (e.g., the x-axis or y-axis of the plane) can be defined as the rotation axis. The original plane can be rotated about the origin and the rotation axis in discrete steps (e.g., + / - 2 degrees, + / - 4 degrees, and + / - 6 degrees).

[0119] The process can be repeated with different centers and axes of rotation (eg, first rotating around the x-axis, then around the y-axis, or also around axes between x and y).

[0120] The quality scores of the extracted medical images are then derived.

[0121] A medical image may then be selected from the extracted plurality of medical images based on the quality score.

[0122] In this way, the final extracted 2D slices are not only based on matching (standard) views of the anatomy but also on optimized quality scores, resulting in better comparability across longitudinal studies.

[0123] Optimization can take into account one or more of the following:

[0124] The noise level in the image;

[0125] Artifact count;

[0126] Field of view coverage;

[0127] Visibility of anatomical structures.

[0128] The concepts described herein are generally applicable to ultrasound imaging. However, the methods are also applicable to other medical imaging modalities with 3D acquisition modes.

[0129] The concept described herein is particularly advantageous for obtaining standard views from a 3D volume. For example, in cardiac or fetal ultrasound imaging, standard views are particularly important. The target plane described herein can correspond to the standard view.

[0130] A skilled person will be able to easily develop a processor for executing any method described herein. Therefore, each step of the flowchart may represent a different action performed by the processor and may be performed by a corresponding module of the processor.

[0131] As discussed above, the system utilizes a processor to perform data processing. The processor can be implemented in a variety of ways using software and / or hardware to perform the various functions required. The processor typically employs 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.

[0132] 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).

[0133] In various embodiments, the processor may be associated with one or more storage media (e.g., 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 transportable so that one or more programs stored thereon can be loaded into the processor.

[0134] By studying the drawings, the disclosure, and the appended claims, those skilled in the art can understand and effect variations of the disclosed embodiments in practicing the claimed invention. In the claims, the word "comprising" does not exclude other elements or steps, and the word "a" or "an" does not exclude a plurality.

[0135] A single processor or other unit may fulfill the functions of several items recited in the claims.

[0136] 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.

[0137] The computer program may be stored / distributed on a suitable medium (e.g. optical storage media or solid-state media provided with or as part of other hardware), but may also be distributed in other forms (e.g. via the Internet or other wired or wireless telecommunications systems).

[0138] If the term "adapted to" is used in the claims or the specification, it should be noted that the term "adapted to" is intended to be equivalent to the term "configured to."

[0139] Any reference signs in the claims should not be construed as limiting the scope.

Claims

1. A computer-implemented method for obtaining a medical image at a target plane, the method comprising: Obtain a three-dimensional 3D volume containing anatomical structures; identifying (602) the anatomical structure in the 3D volume; analyzing a geometry of the anatomical structure in the 3D volume; registering (604) the target plane to the 3D volume based on the analysis of the geometry of the anatomical structure; extracting (606) a plurality of medical images near a registered target plane from the 3D volume, determining a quality score of the extracted medical image; as well as A medical image is selected from the extracted plurality of medical images based on the quality score.

2. The method according to claim 1, further comprising: A target plane is selected, wherein the target plane comprises a plane of a previous 2D scan relative to the time when the 3D volume was acquired, or a plane of a future 2D scan relative to the time when the 3D volume was acquired.

3. The method according to claim 1 or 2, wherein: Identifying the anatomical structure includes using a model-based segmentation algorithm trained on a model of the anatomical structure.

4. The method according to any one of claims 1 to 3, wherein: The 3D volume is a first 3D volume, and the method further comprises: obtaining a second 3D volume containing the anatomical structure; identifying the anatomical structure in the second 3D volume; analyzing a geometry of the anatomical structure in the second 3D volume; registering the target plane to the second 3D volume based on the analysis of the geometry of the anatomical structure in the second 3D volume; and A second medical image at the registered target plane is extracted from the second 3D volume.

5. The method according to any one of claims 1 to 4, further comprising: Obtaining a two-dimensional 2D medical scan of the anatomical structure (402); as well as An anatomical plane is determined from the 2D medical scan, wherein the target plane is the determined anatomical plane.

6. The method according to claim 5, further comprising: A 3D reference volume is obtained that is registered with the 2D medical scan, wherein determining the anatomical plane from the 2D medical scan comprises identifying the anatomical plane relative to the 3D reference volume.

7. The method according to any one of claims 1 to 6, further comprising: obtaining, at a first time, a first 2D medical scan of the anatomical structure; obtaining a second 2D medical scan of the anatomical structure at a third time; obtaining the 3D medical volume at a second time between the first time and the third time; as well as A first anatomical plane from the first 2D medical scan and / or a second anatomical plane from the second 2D medical scan is determined, wherein the target plane is one of the first anatomical plane or the second anatomical plane.

8. The method according to any one of claims 1 to 7, wherein: Extracting the medical image from the 3D medical volume includes using multi-planar reformatting (MPR) on the 3D medical volume.

9. The method according to any one of claims 1 to 8, wherein: The 3D medical volume is a 3D ultrasound volume.

10. A computer program carrier comprising computer program code which, when executed on a processing system, causes the processing system to perform all the steps of the method according to any one of claims 1 to 9.

11. A system for obtaining a medical image at a target plane, the system comprising a processor configured to: Obtain a three-dimensional 3D volume containing anatomical structures; identifying (602) the anatomical structure in the 3D volume; analyzing a geometry of the anatomical structure in the 3D volume; registering (604) the target plane to the 3D volume based on the analysis of the geometry of the anatomical structure; extracting (606) a plurality of medical images near a registered target plane from the 3D volume; determining a quality score of the extracted medical image; as well as A medical image is selected from the extracted plurality of medical images based on the quality score.

12. The system according to claim 11, wherein: The processor is configured to receive as input a target plane, wherein the target plane comprises a plane of a previous 2D scan relative to a time when the 3D volume was acquired, or a plane of a future 2D scan relative to a time when the 3D volume was acquired.

13. The system according to claim 11 or 12, wherein: The processor is configured to identify the anatomical structure by using a model-based segmentation algorithm trained on a model of the anatomical structure.

14. The system according to any one of claims 11 to 13, wherein: The 3D volume is a first 3D volume, and the processor is further configured to: obtaining a second 3D volume containing the anatomical structure; identifying the anatomical structure in the second 3D volume; analyzing a geometry of the anatomical structure in the second 3D volume; registering the target plane to the second 3D volume based on the analysis of the geometry of the anatomical structure in the second 3D volume; and A second medical image at the registered target plane is extracted from the second 3D volume.

15. The system according to any one of claims 11 to 14, wherein: The processor is further configured to: obtaining a two-dimensional (2D) medical scan of the anatomical structure (402); and An anatomical plane is determined from the 2D medical scan, wherein the target plane is the determined anatomical plane.

Images