The present invention relates to assisting diagnosis in three-dimensional ultrasound imaging. In particular, diagnostically significant information is extracted from ultrasound data representing a volume.
For diagnosis with ultrasound images, a set of interrelated images may be acquired. For example, the American Society of Echocardiography (ASE) specifies standard two-dimensional tomograms for fetal and adult echocardiograms. One standard set includes a long axis view, a short axis view, an apical 2 chamber (A2C) view and an apical 4 chamber (A4C) view. Other standardized sets for a same application or different applications may be used. The standard may be set by a national organization, local medical group, insurance company, hospital or by an individual doctor.
In two-dimensional imaging, a clinician positions a transducer at various locations to acquire images at the desired views. However, such positioning may be time-consuming and result in images of the same organ at greatly different times rather than a same time. Clinicians may not be familiar with one or more views.
Ultrasound energy may be used for a volumetric scan (e.g., three- or four-dimensional imaging). A volume is scanned at a substantially same time. The data representing the volume may be used to generate various images. For example, a three-dimensional representation of the volume is rendered using projection or surface rendering. User control or manual cropping tools may be used to alter the rendering. The data representing the volume may also be used to generate orthogonal multi-plane images. Two orthogonal two-dimensional planes are positioned within the volume. The data associated with each of the planes is then used to generate two two-dimensional images. Rendering software may allow for users to position and select an arbitrary plane through the volume for generating a two-dimensional image. Where the volume scan included scanning along a plurality of different planes and different positions within the volume, images associated with each of the component frames may be separately generated. A plane may be tilted or positioned in different locations relative to the volume.
Bi-plane imaging may be provided where two orthogonal planes corresponding to an azimuth and elevation planes are used to generate images during volume acquisition. The planes are positioned within the volume as a function of the transducer position.
In one system, the volume is scanned. After obtaining data representing the volume, the user input provides an indication of the region, organ, tissue or other structure being imaged. For example, the user indicates the heart is being imaged. A template is then used to match with the data, providing an orientation and position of the feature within the volume. Two-dimensional images for different planes through the recognized anatomy are then generated automatically.
By way of introduction, the preferred embodiments described below include methods for assisting three-dimensional ultrasound imaging. Standardized or preset views for a given application are used to assist in volumetric scanning and diagnosis. By displaying one or more images of a standard view during acquisition, the scan may be more appropriately guided to assure proper positioning of the volumetric scan. The location of a user identified view within the volume is used to determine the location of an additional view. The spatial interrelationship of the views within the standard or preset set of views allows generation of images for each of the views after the user identification of one of the views within the volume. Identification of landmarks associated with a particular view may be used for more efficient or accurate feature recognition, more likely providing images for the standard views.
In a first aspect, a method is provided for assisting three-dimensional ultrasound imaging. A first location of a first view within a volume is determined as a function of a second location of a user-identified view within the volume. The first location is different than and non-orthogonal to the second location. An image of the first view is generated.
In a second aspect, a method is provided for assisting three-dimensional ultrasound imaging. A volume is scanned with ultrasound energy. A set of images representing regions with different spatial locations within the volume are displayed during the volume scan. The set of images correspond to preset spatial relationships within the volume.
In a third aspect, a method is provided for assisting three-dimensional ultrasound imaging. A volume is scanned with ultrasound energy from an acoustic window. A first plane of a first standard view associated with the acoustic window is identified relative to the volume. A second plane of a second standard view associated with the acoustic window is automatically extracted as a function of the first plane. The second plane is different than and non-orthogonal to the first plane.
The present invention is defined by the following claims, and nothing in this section should be taken as a limitation on those claims. Further aspects and advantages of the invention are discussed below in conjunction with the preferred embodiments and may be later claimed independently or in combination.
The components and the figures are not necessarily to scale, emphasis instead being placed upon illustrating the principles of the invention. Moreover, in the figures, like reference numerals designate corresponding parts throughout the different views.
By having preset spatial relationships of planes for different views, volume acquisition may be assisted by displaying images corresponding to one or more of the views. The scanning is guided by the view, such as the user orientating a transducer until a recognizable view is provided by a two-dimensional image. Other views of a standard set are then automatically provided given the spatial relationship between the different views. Immediate feedback is provided to the user for confirming desired volumetric scanning. In addition to or alternative to assisting in acquisition, the spatial relationship may be used to identify the position of planes corresponding to standard views within a volume in non-real time. The user identified view is used to determine other views. Where a user may more accurately identify one view, other views are provided without requiring user recognition. Accordingly, more inexperienced clinicians may provide desired images based on recognizing only one or less than all of the views of a set. The location of the different views relative to each other can then be automatically extracted using user placed landmarks to determine the orientation of the heart or other organs, and templates to match and identify the views whose location can be manually refined by the user.
The transducer 12 is a multidimensional transducer array, one-dimensional transducer array, wobbler transducer or other transducer operable to scan mechanically and/or electronically in a volume. For example, a wobbler transducer array is operable to scan a plurality of planes spaced in different positions within a volume. As another example, a one-dimensional array is rotated by hand or a mechanism within a plane along the face of the transducer array or an axis spaced away from the transducer array for scanning a plurality of planes within a volume. As yet another example, a multidimensional transducer array electronically scans along scan lines positioned at different locations within a volume. The scan is of any formats, such as sector scan along a plurality of frames in two dimensions and a linear or sector scan along a third dimension. Linear or vector scans may alternatively be used in any of the various dimensions.
The beamformer system 14 is a transmit beamformer, a receive beamformer, a controller for a wobbler array, filters, position sensor, combinations thereof or other now known or later developed components for scanning in three-dimensions. The beamformer system 14 is operable to generate waveforms and receive electrical echo signals for scanning the volume. The beamformer system 14 controls the beam spacing with electronic and/or mechanical scanning. For example, a wobbler transducer displaces a one-dimensional array to cause different planes within the volume to be scanned electronically in two-dimensions.
The detector 16 is a B-mode detector, Doppler detector, video filter, temporal filter, spatial filter, processor, image processor, combinations thereof or other now known or later developed components for generating image information from the acquired ultrasound data output by the beamformer system 14. In one embodiment, the detector 16 includes a scan converter for scan converting two-dimensional scans within a volume associated with frames of data to two-dimensional image representations. In other embodiments, the data is provided for representing the volume without scan conversion.
The three-dimensional processor 18 is a general processor, a data signal processor, graphics card, graphics chip, personal computer, motherboard, memories, buffers, scan converters, filters, interpolators, field programmable gate array, application specific integrated circuit, analog circuits, digital circuits, combinations thereof or any other now known or later developed device for generating three-dimensional or two-dimensional representations from input data in any one or more of various formats. The three-dimensional processor 18 includes software or hardware for rendering a three-dimensional representation, such as through alpha blending, minimum intensity projection, maximum intensity projection, surface rendering, or other now known or later developed rendering technique. The three-dimensional processor 18 also has software for generating a two dimensional image corresponding to any plane through the volume. The software may allow for a three-dimensional rendering bounded by a plane through the volume or a three-dimensional rendering for a region around the plane. The three-dimensional processor 18 is operable to render an ultrasound image representing the volume from data acquired by the beamformer system 14.
The display 20 is a monitor, CRT, LCD, plasma screen, flat panel, projector or other now known or later developed display device. The display 20 is operable to generate images for a two-dimensional view or a rendered three-dimensional representation. For example, a two-dimensional image representing a three-dimensional volume through rendering is displayed.
The user input 22 is a keyboard, touch screen, mouse, trackball, touchpad, dials, knobs, sliders, buttons, combinations thereof or other now known or later developed user input devices. The user input 22 connects with the beamformer system 14 and the three-dimensional processor 18. Input form the user input 22 controls the acquisition of data and the generation of images. For example, the user manipulates buttons and a track ball or mouse for indicating a viewing direction, a type of rendering, a type of examination, a specific type of image (e.g., an A4C image of a heart), an acoustic window being used, a type of display format, landmarks on an image, combinations thereof or other now known or later developed two-dimensional imaging and/or three-dimensional rendering controls. In one embodiment, the user control 22 is used during real time imaging, such as streaming volumes (i.e., four dimensional imaging) are acquired. In other embodiments, the user control 22 is used for rendering from a previously acquired set of data now stored in a memory (i.e., non-real time imaging).
In act 30, a set of standard views and corresponding spatial relationships are established. The set of standard views includes two or more preset, different views. The views may correspond to one-dimensional, two-dimensional or three-dimensional imaging. Each different view corresponds to a different imaging location, such as two two-dimensional planes at different positions within a same volume.
The standard views are standards based on any individual or organization. For example, a medical organization associated with a particular application, group of applications, ultrasound imaging, imaging, or other organizations may establish different sets of views useful for diagnosis.
Other sets of standard views for a same or different applications may be used. For example, a plurality of non-orthogonal planes that are at slight angles, such as 10° or less, to each other through a same region of the heart or other organ are provided as the standard views as shown in
Different sets of standard views may be provided for different acoustic windows in a same application. For example, cardiac imaging of the heart may provide for three or four different acoustic windows. One acoustic window is positioned by the neck, another by the sternum and two between different ribs. Other acoustic windows may be used, such as associated with imaging from the esophagus using a transesophageal probe. Different acoustic windows may be provided for different applications, such as for imaging different organs or body structures.
The corresponding spatial relationships are provided through experimentation, definition as a standard or known structural relationships. While some variation may be provided between different patients in the size, shape and orientation of an image organ, standard views may allow for likely identification of appropriate locations associated with each of the standard views.
Other sets of views may include user established standards or preset views. The user inputs a spatial relationship for one or more views. For example, the user desires a view of the heart not typically obtained using another standard set of views. The user inputs a spatial relationship of the desired view to a known view, such as a user identifiable A4C view. An algorithm provides tools for the user to encode the relative positions of non-standard views with respect to at least one standard view (e.g., A4C) into the system. By inputting the spatial relationship, the set of views includes a user set standard view. Alternatively, the set of views includes only user established views. Other information may be input by the user. For example, the user creates templates and landmark descriptions for these user established views using a training or other image data set. These templates, landmark descriptions and/or the training image data may be used in automatically identifying the non-standard views relative to a specified standard view when new image data is acquired. After at least one non-standard view is thus described, it can be used as if it were a standard view, in describing other non-standard views. This enables the system to function properly when only user established views are used by the clinician.
In act 32, a location of one view associated with an acoustic window or application is identified. For example, a plane associated with a standard view is identified. In the example provided in
For real time acquisition and imaging, a view is identified in response to user input or automated processes. A volume is scanned with ultrasound energy from an acoustic window. The acquired data is then used to generate a three-dimensional or other image. For example, both a three-dimensional rendering as represented in
In act 34, a location of a view within a volume is determined as a function of the location of the user identified or other view within the volume. The locations of the different views are different and may or may not be orthogonal. Since the spatial relationship of the different views within a set of standard or preset views is known and stored in a memory, user identification of one view provides the locational information for other views relative to the user identified view. Any number of different views may be determined based on spatially locating a first view. By identifying the acoustic window and/or the desired set of views, any number of views within the set may be determined by identifying the location or position of one view within the set. Identification of the acoustic window indicates a set or a plurality of different sets. Identification of a set with or without corresponding acoustic window information allows for the determination of spatial relationships of a known view to other views.
In the example embodiment of
The different views are determined automatically in response to user identification of the user identified view. For example, a processor obtains the spatial relationship from memory and identifies data corresponding to the different views. In one embodiment, the location relative to the volume of the different views within a set of standard or preset views is determined automatically in act 36 by the positioning of the transducer during imaging. By displaying an image associated with one desired view and positioning the transducer until the image corresponds to desired tissue structure, the various views are automatically positioned as a function of position of the transducer (e.g., acoustic window being used) and the spatial interrelationships. By the user identifying the location of one view relative to the volume, the position of the other views is automatically determined. Referring to
Other parameters may be altered based on the determined positions of the different views. For example, the volume scan rate is increased once the position of the views is determined. The volume scan rate is increased by limited the location and/or depth of scan lines used to image the volume. By scanning where needed to acquire data for the desired views and desired images of the views, less time may needed to scan portions of the volume not being imaged. For example, using the standard views shown in
In another embodiment for automatically extracting the position of one plane or view as a function of a position of a different plane or view, landmarks are used in act 38. In real time or non-real time, the user identifies one of the views within a set. An image corresponding to the view is displayed, such as by the user slicing or arbitrarily positioning planes or volumes for rendering within the scan volume. One or more landmarks associated with the identified view or image are then provided as input. For example, user input identifying a plurality of landmarks within the image is received. The landmarks entered may depend on the view being used. For example in an A4C view, three or more points are identified associated with the lateral tricuspid, lateral mitrol annulus, the crux of the heart and the LV apex. Other landmarks may be used. Continuous landmarks associated with tracing an outline or identifying a border automatically or with user input may also be used. In alternative embodiments, a processor automatically identifies various landmarks using pattern matching or correlation with a template. Where automated landmarks are used, the user indicates that a given image in an associated view position is of a particular view. The processor then identifies landmarks within the view for determining the orientation and/or size of the anatomy.
The landmarks are used to determine an orientation or size of the organ or structure being imaged within the volume. By spatially positioning the orientation or size of the anatomy as a function of the selected view with the volume and the landmarks, a more refined determination of the location of other views may be used. For example, the spatial relationship between different views is a function of structure within the anatomy. Where the heart or other organ is at a different orientation, different spatial relationships may be provided. The landmarks allow for selection of an appropriate spatial relationship. In fetal echocardiography, the orientation of the fetal heart relative to the transducer may vary depending on fetus position. Landmarks are used to determine the orientation of the fetal heart relative to the transducer. The desired views may then be located given the orientation and spatial relationships.
Further refinement of the spatial relationships is provided by allowing adjustment of the spatial relationship of one view relative to another view. In act 44, the adjustment corresponds to manual or user input based adjustment. As an alternative, the spatial relationship is adjusted automatically or with a processor. Spatial relationship provided with a set of views provides an approximate positioning of one view relative to another view. A preset spatial relationship allows extraction of approximate positions of different planes or regions. A template based on the structure within an image for a different view is matched to the corresponding data. Sample images from an image database, a likely geometric shape or other templates may be matched to identify a translation and/or rotation associated with adjustment of the relative spatial locations for a given examination. By matching the template with data representing planes or other regions near the approximated position, a more optimum position may be identified. Any of various matching may be used, such as correlation or pattern recognition.
In act 40, one or more images of the different views are generated. Different viewing formats may be provided. For example, different images for two or more different views are displayed substantially simultaneously, such as adjacent to each other.
In one embodiment represented in
The generated images are in any now known or later developed format. For example, an M-mode, B-mode, Doppler mode, contrast agent mode, harmonic mode, flow mode or combinations thereof is used. One-, two- or three-dimensional imaging may be provided. For example, a two-dimensional plane is used as a boundary for rendering a three-dimensional representation. One or more of the views of a standard set of views may be represented with a three-dimensional volume rendering bounded by the location of the view. As another example, a plurality of adjacent planes or grouping of data around a location of a particular view is used for rendering a three-dimensional representation of a slice. As yet another example, a two-dimensional image is generated from data along a two-dimensional plane. In one embodiment, one or more views are displayed as two-dimensional views and at least another view is volume rendered with an identified plane acting as a front cut-plane or boundary for the rendering. A three-dimensional rendering of the entire volume may be displayed at a same time or sequentially with images generated for any of the standard or preset views. The different images displayed for different views or a three-dimensional rendering may use the same or different light sources and the same or different viewing directions for generation of the images. Displayed images may be overlapping, such as one image overlapping another in an opaque or semi-opaque manner. A pulse or continuous wave image, such as provided for spectral Doppler imaging, may be provided as one of the views or in addition to any of the other generated images.
In act 42, the spatial relationship of the user identified view to other views is displayed. For example, the display format of images shown in
In act 44, the spatial relationship between different views is adjusted as a function of user input. After or during the display of images corresponding to the different views, the user may indicate an adjustment, such as a tilting, rotating or translation along any dimension or axis of a position of a view relative to another view. The spatial relationship is adjusted for a given examination or adjusted and stored as part of the set of views for later examinations. Adjustment allows for optimizing views for different patient conditions, such as orientations or size differences between different patients. The adjustment is performed after data is acquired, or while data is acquired for real time imaging. The adjustment may be stored for a given set of data representing a volume for a later use and diagnosis. In one embodiment, the user selects one view and identifies the location of that view relative to the volume. The spatial relationship between the user identified view and other views are adjusted as desired in real time or non-real time.
While the invention has been described above by reference to various embodiments, it should be understood that many changes and modifications can be made without departing from the scope of the invention. It is therefore intended that the foregoing detailed description be regarded as illustrative rather than limiting, and that it be understood that it is the following claims, including all equivalents, that are intended to define the spirit and scope of this invention.