Medical Imaging

Abstract

The present invention provides a method for positioning and / or defining the boundaries of a subject's scan volume (volume to be scanned) based on processing camera data with a predictive model that predicts the location of bony structures within the body, and then determining the scan volume based on reference to the bony structure locations, which provide spatial reference or anchor points for defining the extent of the volume to be scanned.

Description

[Technical Field]

[0001] The present invention relates to a method for setting up a medical imaging scan, and in particular to a method for setting the spatial boundaries of a volume to be scanned. [Background technology]

[0002] Proper positioning of the patient's target anatomy relative to the medical imaging system (e.g., computed tomography, MRI) is particularly important to optimize the signal-to-noise ratio (SNR) and radiation dose distribution for ionizing modalities, ensuring optimal image quality for subsequent diagnostic analysis.

[0003] Positioning typically includes positioning in the horizontal plane, which defines the field of view of the scan, and vertical positioning, which ensures proper centering of the target anatomy.

[0004] In this context, "positioning" is used to refer to the placement of the volume scanned by the scanner, i.e., the volume scan window, or scan field of view, to encompass a particular target anatomical structure. In practice, this scan window is implemented by physically positioning the patient relative to the scanner, either statically or with a particular motion path, so that the correct volume of the patient is scanned. Because the target anatomical structure is internal to the subject, its position relative to the scanner's coordinate system cannot be directly detected. Traditionally, it has been necessary for radiologists to use their experience to make their best estimate of the placement and boundaries of the scan window.

[0005] With the current state of the art, the selection of the patient's horizontal and vertical positions can be automated using a sensor-based approach, in which the location of the target anatomical structure is calculated by processing sensor data, typically camera images. However, a highly accurate model is required to predict the location of the target anatomical structure based on the sensor data. This makes the model highly sensitive to any errors in the input data or processing. Additionally, models currently known in the art are very inflexible with regard to customizing scan volume placement. These models provide a fixed definition of the start and end points of the scan window for a given input sensor data set. However, allowing for greater customization would be valuable. Summary of the Invention [Problem to be solved by the invention]

[0006] Improvements in this area of ​​the technology are generally sought. Predictive models for body applications such as spine, head / neck, and chest, heart, abdomen, liver, and pelvis would be highly valuable as these are important and routine clinical imaging applications. [Means for solving the problem]

[0007] The invention is defined by the claims.

[0008] According to an example in accordance with one aspect of the present invention, there is provided a method for use in setting up an imaging scan to be performed using an imaging scanner, the imaging scanner having an examination area for receiving a subject to be scanned, the method comprising: acquiring spatial sensor data outside the subject's body received within the examination region; applying a predetermined algorithm to the sensor data to determine the position of one or more anatomical landmarks on the subject's body positioned within the examination region relative to a coordinate system associated with the examination region; applying a prediction procedure, the prediction procedure comprising application of a prediction algorithm configured to generate an output comprising a prediction of a position of a plurality of bony structures of the subject within the examination region relative to the coordinate system based on an input to the prediction algorithm comprising a detected position within the examination region of one or more anatomical landmarks; defining the location and / or boundary points of the scan volume to be scanned relative to said coordinate system based on the predicted positions of bone structures output from the prediction algorithm; generating an output indicating defined locations and / or boundary points of the scan volume to be scanned; It has.

[0009] Although the general concept of the present invention is to provide an automated method for determining the scanned volume (or scan window), an intermediate step is introduced in which the location of bony structures is predicted from sensor data, and then the start and end points and placement of the scanned volume are determined with reference to the location of the bony structures. The location of the bony structures provides a very useful set of reference points within the patient, relative to which the scanned volume can be determined. Bone structures are useful reference points because the various anatomical structures that are typically scanned can, in most cases, be defined in very precise locations relative to the bones. The location of skeletal bony structures relative to organs is very predictable, making skeletal features useful and reliable reference points.

[0010] The one or more anatomical landmarks detected using the sensor data are, for example, visible landmarks on the outside of the body, i.e., landmarks whose positions are observable on the surface of the body, such as joints, or externally visible organs or parts thereof, such as eyes or ears, etc. Thus, the positions of such anatomical landmarks can be defined positions on the surface of the body, or can be positions within the body estimated based on observations on the surface of the body.

[0011] In some embodiments, the plurality of bony structures includes a plurality of spinal vertebrae.

[0012] Using the spine as a reference for anatomical position has the advantage of providing a natural alignment of human anatomy to a relative size-invariant coordinate system, and correlating well with articular structures such as the shoulders and hips that are typically detected on camera images.

[0013] For example, according to one preferred set of embodiments, defining the location and / or boundary points of the scanning volume includes obtaining first indicators of positions and / or boundary points, the first indicators being defined relative to the positions of one or more of the bone structures, and converting the first indicators into second representations of the positions and / or boundary points, the second representation being defined relative to the coordinate system, the converting including using predicted positions of the bone structures output from the prediction algorithm.

[0014] The method may include explicitly defining a body coordinate system using the detected spine positions as reference points.

[0015] The initial representation (for the bone structure) can be obtained from a user interface. The first indices can additionally or alternatively be obtained from a reference database.

[0016] The aforementioned coordinate system may be the coordinate system of the imaging scanner, which means, for example, the coordinate system in which the surgical imaging component of the scanner is defined and / or in which the imaging field of view is defined, which means, for example, the coordinate system in which the scan area acquired by the scanner is defined.

[0017] In some embodiments, the method includes determining the position and / or spatial extent of one or more target anatomical structures within the body relative to a coordinate system based on the predicted positions of the bony structures.

[0018] In some embodiments, determining the positions and / or spatial extent of one or more target anatomical structures within the body includes obtaining first indications of the positions and / or spatial extent, the first indications being defined relative to the positions of one or more bony structures, and transforming the first indications into a second representation of the positions and / or spatial extent, the second representation being defined relative to said coordinate system, the transformation including using predicted positions of the bony structures output from a prediction algorithm. Defining said placement and / or boundary points of the scan volume may then (optionally) be performed based on the determined positions and / or spatial extent of the one or more target anatomical structures.

[0019] For example, it may include obtaining first indications of location and / or boundary points, the first indications being defined relative to the positions of one or more target anatomical structures, and transforming the first indications into a second representation of the location and / or boundary points, the second representation being defined relative to the coordinate system, the transformation including using the determined positions and / or spatial extents of the one or more target anatomical structures.

[0020] In some embodiments, the method includes determining a vertical center of a spatial extent of one or more target anatomical structures within the body relative to a coordinate system based on the predicted location of the bony structure, and aligning a vertical center of the scan volume with the vertical center of the target anatomical structures within the body.

[0021] For example, the coordinate system may include a vertical dimension, and determining the location and / or spatial extent of one or more target anatomical structures within the body includes determining a vertical center of the spatial extent of at least one target anatomical structure within the body relative to the vertical dimension of the coordinate system.

[0022] In some embodiments, the placement and / or boundary points of the scan volume are determined to align the vertical center of the scan volume with said vertical center of at least one target anatomical structure within the body.

[0023] In some embodiments, the imaging scanner is a CT imaging scanner having a rotating gantry that rotates in a plane of rotation. In some embodiments, the imaging scanner is an MRI scanner.

[0024] As a general principle, in some embodiments, an imaging scanner has an imaging isocenter that defines the center point of a scan volume to be scanned in at least one dimension, where the scanner performs a scan of a given scan volume by changing the relative positioning of the scanner's operable scan section with respect to the subject being scanned, either by moving the scan section or by moving a subject support carrying the subject. For example, for a CT scanner, the surgical scan section can be a rotating gantry carrying x-ray emitters and detectors. For an MRI scanner, it can be an arrangement of coils.

[0025] In some embodiments, the method includes controlling a patient support table of an imaging scanner to adjust a vertical positioning of the patient support table defined relative to the vertical dimension of the coordinate system such that a vertical center of a target anatomical structure within the body is aligned with an imaging isocenter of the scanner.

[0026] In some embodiments, the method includes determining a vertical center of spatial extent of one or more target anatomical structures within the body relative to the coordinate system. The vertical dimension of the coordinate system corresponds to the vertical direction of the scanner. For a CT scanner, this may be, for example, the radial dimension of the plane of rotation of the gantry.

[0027] With reference to the predictive algorithm, in some embodiments the predictive algorithm may be a statistical model, such as a regression model.

[0028] In some embodiments, the predictive algorithm may be a trained machine learning algorithm. A machine learning algorithm is any self-learning algorithm that processes input data to generate or predict output data. Suitable machine learning algorithms for use in the present invention will be apparent to those skilled in the art. Examples of suitable machine learning algorithms include decision tree algorithms and artificial neural networks. Other machine learning algorithms, such as logistic regression, support vector machines, and naive Bayes models, are suitable alternatives.

[0029] In some embodiments, the aforementioned spatial sensor data includes camera data. This may be video camera data. In some examples, it may include 3D camera data. However, this is not required, and other examples may include 2D camera data. The camera data may include infrared camera data or visible spectrum camera data. An alternative to camera data may include laser-sensing spatial sensor data, such as LIDAR data.

[0030] In some embodiments, the one or more anatomical landmarks include one or more skeletal joints.

[0031] In some embodiments, the prediction procedure involves the application of multiple prediction algorithms, one for each of multiple bony structures whose positions are to be predicted.

[0032] In some embodiments, spatial sensor data is data acquired using a set of one or more sensor elements, e.g., cameras, having predefined spatial positions relative to the coordinate system mentioned above. For example, the set of sensor elements may include a spatial array of sensor elements, e.g., a 2D array, arranged in a grid formation.

[0033] In some embodiments, the one or more target anatomical structures include one or more of a section of the spine, a head or neck, a chest, a heart or portion thereof, an abdomen, a liver, and a pelvis.

[0034] Another aspect of the present invention is a computer program product comprising code means configured to, when executed on a processor, cause the processor to perform a method according to any embodiment or example described herein or according to any claim of the present application.

[0035] The invention can also be implemented in hardware form. One aspect of the invention is a processing unit for use in setting up an imaging scan to be performed using an imaging scanner, the imaging scanner having an examination area for receiving an object to be scanned. It is proposed that the processing unit has an input / output and one or more processors, the one or more processors: receiving at an input / output spatial sensor data outside the body of a subject received within an examination region; applying a predetermined algorithm to the sensor data to determine the position of one or more anatomical landmarks on the subject's body positioned within the examination region relative to a coordinate system associated with the examination region; applying a prediction procedure, the prediction procedure including applying a prediction algorithm configured to generate an output including a prediction of a position of a plurality of bony structures of the subject within the examination region relative to the coordinate system based on inputs to the prediction algorithm including detected positions within the examination region of one or more anatomical landmarks; defining the location and / or boundary points of the scan volume to be scanned relative to said coordinate system based on the predicted positions of bone structures output from the prediction algorithm; generating an output in the input / output section indicating defined locations and / or boundary points of the scan volume to be scanned; adapted / configured / programmed to perform

[0036] As another aspect of the present invention, a system may be provided having a processing unit as described above or according to any embodiment or claim of the present disclosure, and an imaging scanner including an examination area for receiving a subject.

[0037] In some embodiments, the imaging scanner is a computed tomography (CT) scanner or an MRI scanner. In some embodiments, the imaging scanner is an X-ray imaging scanner.

[0038] These and other aspects of the invention will be apparent from and elucidated with reference to the embodiments described hereinafter.

[0039] For a better understanding of the present invention, and to show more clearly how the same may be carried into effect, reference will now be made, by way of example only, to the accompanying drawings in which: [Brief explanation of the drawings]

[0040] [Figure 1] 1 is a schematic diagram of an example CT scanner known in the art. [Figure 2] FIG. 2 is a block diagram of example method steps in accordance with one or more embodiments of the present invention. [Figure 3] FIG. 2 is a block diagram of an example processing unit and system components in accordance with one or more embodiments of the present invention. [Figure 4] 10A-B show a schematic representation of the acquisition of sensor data outside the subject and the identification of landmark locations. [Figure 5] 1 shows a real-world example of acquired sensor data and identified landmark locations. [Figure 6] 10 shows prediction of vertebral positions based on identified landmark positions. [Figure 7] We demonstrate the accuracy of the regression model used to predict the scan range by comparing the ground truth with the model output for different predictions. DETAILED DESCRIPTION OF THE INVENTION

[0041] The present invention will now be described with reference to the drawings.

[0042] It should be understood that the detailed description and specific examples, while indicating exemplary embodiments of the devices, systems, and methods, are intended for purposes of illustration only and are not intended to limit the scope of the invention. These and other features, aspects, and advantages of the devices, systems, and methods of the present invention will become better understood from the following description, appended claims, and 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 throughout the drawings to indicate the same or similar parts.

[0043] The present invention provides a method for positioning and / or defining the boundaries of a scan volume or window (scanned volume) of a subject based on processing sensor data, such as camera data, with a predictive model to predict the location of bony structures, such as vertebrae, within the body, and determining the scan volume based on reference to the location of the bony structures, which provide spatial reference or anchor points for defining the extent of the scanned volume.

[0044] In practice, a particular scanned volume is implemented by physically positioning the patient relative to the scanner, either statically or with a particular motion path, so that the correct volume of the patient is scanned. The scanner has a local or native 3D coordinate system, and the relative positions of the motion scanning part / section of the scanner and the patient support part of the scanner can be adjusted with respect to this 3D coordinate system by moving the scanning part, the patient support part, or both. Purely by way of example, in a CT scanner, the scanning part is an emitter-detector structure rotatably supported by a rotating gantry, which is movable linearly along an axial direction relative to the patient support table (usually by physically moving the table).

[0045] Therefore, if the correct scan volume to be scanned can be determined within the scanner's 3D coordinate system, the scan volume can be correctly scanned by standard scan control procedures that are already encoded in the scanner's control software, which utilize the native scanner coordinate system.

[0046] To achieve a specific scan region, current approaches for automatic patient positioning typically rely on a fixed definition of the target anatomical structures relative to the scanner's 3D coordinate system, based on clinical guidelines. This includes defining the isocenter of the localizer / survey scan (for MRI) or the start and end points of the localizer scan (for CT) based on external landmarks such as the chin, shoulders, and hips, or internal landmarks such as the lungs, liver, and pelvic bones. A statistical model, such as a regression or machine learning (e.g., deep learning) model, is then trained to calculate the specified positions based on the input sensor data. However, clinical facilities often have local preferences for defining the center or start and end positions of the scan for each anatomical structure. Technical operators may also need to adapt these preferences to specific patients if a clinician's referral requests a specific scan. For these reasons, automatic positioning using fixed, pre-trained anatomical landmarks is only applicable to very standard examinations and may not be suitable for all cases. This is particularly true for the spine, where the start and end points of corresponding scans can frequently vary depending on the clinical case.

[0047] Therefore, an improved approach is needed.Although reference is made to localizer scans, embodiments of the present invention are applicable to any type of medical imaging scan.

[0048] Embodiments of the present invention propose an improved approach in which the positions of bone structures, such as the spine, pelvic bone, skull, shoulder bone, or any other bony feature or structure, are predicted and these are used as reference points for defining the planned scan volume. The use of spinal vertebrae as predicted bone structures, for example, represents one particularly advantageous example.

[0049] Embodiments of the present invention are applicable to a wide range of medical imaging modalities, including, for example, computed tomography imaging (CT), magnetic resonance imaging (MRI), X-ray imaging, and PET imaging.

[0050] To aid in understanding the description of the inventive concepts that follows, a brief description of the basic structure of an example CT scanner will now be provided, although it is emphasized that the present invention is not limited to use with this particular structure of CT scanner, or indeed to CT imaging at all.

[0051] FIG. 1 shows an imaging system 100, such as a computed tomography (CT) scanner.

[0052] Imaging system 100 includes a generally stationary gantry 102 and a rotating gantry 104. Rotating gantry 104 is rotatably supported by stationary gantry 102 and rotates about a longitudinal or z-axis around an examination region.

[0053] A patient support 120, such as a couch, supports an object or subject, such as a human patient, within the examination region. The support 120 is configured to move the object or subject for loading, scanning, and / or unloading the object or subject. The patient's position can be adjusted in all three dimensions: horizontally (the XZ plane according to the example coordinate system shown in FIG. 1 ) and vertically (the Y axis according to the example coordinate system shown in FIG. 1 ).

[0054] A radiation source 108, such as an x-ray tube, is rotatably supported by the rotating gantry 104. The radiation source 108 rotates with the rotating gantry 104 and emits radiation across an examination region 106.

[0055] The radiation sensitive detector array 110 corresponds to an angular arc across the examination region 106 opposite the radiation source 108. The detector array 110 includes one or more rows of detectors extending along the z-axis direction to detect radiation traversing the examination region 106 and generate projection data indicative thereof.

[0056] A general-purpose computing system or computer serves as an operator console 112 and includes input devices 114, such as a mouse and / or keyboard, and output devices 116, such as a display monitor or filmer. The console 112 allows an operator to control the operation of the system 100.

[0057] A reconstructor 118 processes the projection data and reconstructs volumetric image data, which can be displayed through one or more display monitors of the output device 116.

[0058] The reconstruction unit 118 may employ filtered backprojection (FBP) reconstruction, (image domain and / or projection domain) reduced noise reconstruction algorithms (e.g., iterative reconstruction), and / or other algorithms. It should be understood that the reconstruction unit 118 can be implemented through a microprocessor that executes computer-readable instructions encoded or embedded in a computer-readable storage medium such as physical memory and other non-transitory medium. Additionally or alternatively, the microprocessor can execute computer-readable instructions transmitted by a carrier wave, signal, and other transitory (or non-transitory) medium.

[0059] 2 outlines in block diagram form the steps of an exemplary method according to one or more embodiments. The steps are summarized before being further described in the form of an exemplary embodiment.

[0060] The method 10 is for use in setting up an imaging scan to be performed using an imaging scanner having an examination area for receiving a subject to be scanned.

[0061] The method comprises a step 12 of acquiring spatial sensor data outside the body of the subject received in the examination region. For example, the spatial sensor data may be camera data, e.g. image data, e.g. 3D camera data.

[0062] The method further includes step 14 of applying a predetermined algorithm to the sensor data to determine the position of one or more anatomical landmarks on the subject's body positioned within the examination region relative to a coordinate system associated with the examination region.

[0063] The coordinate system can be the coordinate system of the imaging scanner, for example the coordinate system of the reference frame of the imaging scanner, or the coordinate system of the reference frame of the patient support 120.

[0064] The one or more anatomical landmarks may include, for example, one or more skeletal joints, such as the shoulder or hip.

[0065] The method further comprises applying a prediction procedure, wherein the prediction procedure comprises application of at least one prediction algorithm adapted to generate an output comprising a prediction 16 of the positions of a plurality of bony structures of the subject within the examination region relative to the coordinate system based on an input to the prediction algorithm comprising the detected positions of one or more anatomical landmarks within the examination region. Thus, the positions of the landmarks are mapped to the positions of the bony structures. The at least one prediction algorithm may comprise application of a statistical model, such as a regression algorithm, or may comprise a trained machine learning algorithm.

[0066] The method further comprises a step 18 of defining the location and / or boundary points of a scan volume to be scanned relative to the coordinate system based on the predicted positions of the bone structures output from the prediction algorithm. The scan volume to be scanned may alternatively be referred to as a scan region, a scan window, or a scan field of view.

[0067] The method may further comprise the step 20 of generating an output indicative of the defined locations and / or boundary points of the scan volume to be scanned.

[0068] As mentioned above, the method may also be embodied in hardware form, for example in the form of a processing unit configured to perform the method according to any example or embodiment described herein or according to any claim of the present application.

[0069] To further aid understanding, FIG. 3 shows a schematic representation of an example processing unit 32 configured to perform methods according to one or more embodiments of the present invention.

[0070] The processing unit 32 includes an input / output unit 34 and one or more processors 36. The processing unit 32 is configured to receive spatial sensor data of an area outside the subject's body received within the examination region at the input / output unit 34. For example, this may be received from a sensor device 52 communicatively coupled to the input / output unit. The processing unit 32 is further configured to apply a predetermined algorithm to the sensor data to determine the positions of one or more anatomical landmarks of the subject's body disposed within the examination region relative to a coordinate system associated with the examination region. The processing unit 32 is further configured to apply a prediction procedure, the prediction procedure including application of a prediction algorithm configured to generate an output including predictions of the positions of a plurality of bony structures of the subject within the examination region relative to the coordinate system based on input to the prediction algorithm having detected positions of one or more anatomical landmarks within the examination region. The processing unit 32 is further configured to define the location and / or boundary points of the scan volume to be scanned relative to the coordinate system based on the predicted positions of the bony structures output from the prediction algorithm. The processing unit 32 is further configured to generate an output at the input / output 34 indicative of the defined locations and / or boundary points of the scan volume to be scanned. By way of example, this output may be transferred to an imaging scan 58, which may be communicatively coupled to the input / output.

[0071] The processing arrangement may further include a memory 38 that may store computer-executable instructions (computer program code) that, when executed by one or more processors 36, cause the one or more processors to perform the methods outlined above.

[0072] Also, one embodiment of the present invention is a system having a processing arrangement 32 in combination with one or both of a sensor device 52 and an imaging scanner 58 .

[0073] By way of example, the imaging scanner may be a computed tomography scanner or an MRI scanner.

[0074] In some embodiments, the bony structures referred to above are spinal vertebrae. This represents a particularly advantageous example, as the spine provides a natural alignment of human anatomy to a relative size-invariant coordinate system. The spine can be particularly advantageous when imaging upper body anatomy.

[0075] Other bony structures may be useful for imaging anatomical structures in the lower body, upper body extremities, or head, including, for example, the pelvic bone, the skull (or a portion thereof), the shoulder bone, the arm bone, or any other bony structure.

[0076] The bony structures may include a combination of spinal vertebrae and other bony structures in some instances.

[0077] As a general principle, in some embodiments, an imaging scanner may have an imaging isocenter that defines the center point of a scan volume to be scanned in at least one dimension, where the scanner performs a scan of a given scan volume by changing the relative positioning of the scanner's operational scan section with respect to the object being scanned, either by moving the scan section or by moving the object support carrying the object. For example, in a CT scanner, the operational scan section may be a rotating gantry carrying the x-ray emitters and detectors. In an MRI scanner, it may be the coil arrangement.

[0078] In some embodiments, the method includes determining a vertical center of a spatial extent of one or more target anatomical structures within the body relative to a coordinate system based on the predicted location of the bony structure, and aligning a vertical center of the scan volume with the vertical center of the target anatomical structures within the body.

[0079] For example, the coordinate system may include a vertical dimension, and determining the location and / or spatial extent of one or more target anatomical structures within the body includes determining a vertical center of the spatial extent of at least one target anatomical structure within the body relative to the vertical dimension of the coordinate system.

[0080] In some embodiments, the placement and / or boundary points of the scan volume are determined to align the vertical center of the scan volume with said vertical center of at least one target anatomical structure within the body.

[0081] This can be done, for example, by controlling the patient support table of the imaging scanner to adjust the vertical positioning of the patient support table defined relative to the vertical dimension of the coordinate system so that the vertical center of the target anatomical structure within the body is aligned with the imaging isocenter of the scanner.

[0082] With regard to the one or more spatial sensor elements mentioned above, preferably these are implemented with one or more cameras.

[0083] In some embodiments, spatial sensor data is data acquired using a set of one or more sensor elements, e.g., cameras, having predefined spatial positions relative to the coordinate system mentioned above.

[0084] For example, the set of sensor elements may comprise a spatial array of sensor elements, e.g., a 2D array, arranged in a grid configuration. In some embodiments, the method may include a calibration procedure in which the position, preferably the orientation, of each of the set of sensor elements relative to a coordinate system is determined. This can be achieved, for example, by using each sensor element to capture a respective sensor data set, e.g., an image, and identifying within each data set the positions of the same set of reference points within the imaged examination area, each of the reference points having a known position relative to the coordinate system, and using the positions of the reference points detected within each data set to determine the position of each sensing element, e.g., each camera.

[0085] One preferred approach for implementing the method outlined in the summary above is to define the position and / or boundary points of the scan volume in two stages: first, obtaining first indicators of the positions and / or boundary points, where the first indicators are defined relative to the positions of one or more of the bony structures; and second, transforming the first indicators into second indicators of the positions and / or boundary points, where the second indicators are defined relative to the coordinate system of the scanner, and the transformation includes using predicted positions of the bony structures output from a prediction algorithm.

[0086] The first indicator may be an indicator obtained from an input by a user received at a user interface.

[0087] As an example, the user interface may have a display unit controlled to display an image or avatar of a human body or to display a medical image (2D or 3D, e.g., an MRI or CT image), the location of bony structures being indicated, and the boundaries of a desired scan volume can be interactively selected by a user with respect to the displayed image or avatar. For example, the user controls a pointer device to spatially indicate the location and / or boundaries of the desired scan volume relative to the displayed locations of the bony structures.

[0088] The first index may alternatively be pre-stored in a reference database and simply retrieved, reflecting a preferred placement of the scan volume relative to the bone structures, for example, set at a policy level by a medical institution or clinic, or pre-set by an individual clinician.

[0089] In addition to or instead of obtaining indications of the intended scan volume relative to the positions of the bony structures, the method may include the steps of obtaining first indications of the positions and / or spatial extent of one or more target anatomical structures, the first indications being defined relative to the positions of one or more of the bony structures, and converting the first indications into second indications of the positions and / or spatial extent, the second indication being defined relative to the coordinate system, the conversion including use of predicted positions of the bony structures output from the prediction algorithm.

[0090] The placement and / or boundary points of the scan volume can then be determined, for example, based on the determined positions and / or spatial extents of one or more target anatomical structures. Again, the first indicators can be obtained from a user or a data store.

[0091] In accordance with the above approach, an example implementation of a method according to one or more embodiments of the present invention will now be described by way of illustration of the above-summarized concepts of the present invention. It will be understood that not all features of this particular set of embodiments are essential to the inventive concepts, but are described to aid understanding and to provide an example for illustrating the inventive concepts.

[0092] As a general introduction, this set of embodiments proposes using a predictive algorithm to predict the horizontal and vertical positions of at least a subset of spinal vertebrae (e.g., C1-C6, T1-T12, L1-L5, S1-S5, coccyx) based on externally observable landmarks such as the head, shoulders, and hips extracted from sensor data. Thus, in this case, the previously mentioned bony structures are spinal vertebrae. For example, detecting landmarks such as externally observable joints or other anatomical features using camera images is an active research area in computer vision. For example, segmentation or classification algorithms are known to perform such tasks, and skilled practitioners are aware of means to implement this feature.

[0093] From the positions of the anatomical landmarks, one or more prediction algorithms can be applied to predict the 3D positions of the vertebrae.

[0094] By way of example, suitable predictive algorithms include statistical models such as regression models, or trained machine learning algorithms. A machine learning algorithm is any self-learning algorithm that processes input data to generate or predict output data. Suitable machine learning algorithms for use in the present invention will be apparent to those skilled in the art. Examples of suitable machine learning algorithms include decision tree algorithms and artificial neural networks. Other machine learning algorithms, such as logistic regression, support vector machines, and naive Bayes models, are suitable alternatives.

[0095] For example, in the training phase of a regression model or other machine learning model, ground truth can be provided in the form of a set of 3D images of all or part of the spine that are annotated to label the 3D positions of the vertebrae and the 3D positions of externally detectable anatomical landmarks, such that one or more regression models can be trained to predict the position of each of the vertebrae (as output) based on the positions of the anatomical landmarks (as input).

[0096] The clinical user of the system then has the possibility to define the target anatomical structure and / or the end and start points and / or isocenter of the volumetric scan window (volume to be scanned) based on any one or more vertebrae.

[0097] This is particularly useful for body applications such as the head / neck, spine, and chest, heart, abdomen, liver, and pelvis.

[0098] As a brief summary, according to a particular set of embodiments, the steps of this method may include:

[0099] Access and apply a regression model (or other statistical model) to predict the 3D positions of at least a subset of the subject's vertebrae based on input indicating the 3D positions of a plurality of externally observable skeletal joints or other landmarks. As previously mentioned, the model can be trained based on appropriately annotated volumetric MRI or CT image data.

[0100] The system's user interface can be used by a clinical user to predefine and store in a data store or to select in real time the scan region to be scanned, or potentially on the fly, the start and end points (and / or center) of the anatomical structure to be scanned for one or more vertebrae, such as T1-T12 for a thoracic spine scan. Note that the scan region to be scanned may optionally be the scan region of a localizer or survey scan, based on which the final selection of the scan region is determined. The scan region in any embodiment may be a 2D or 3D scan region.

[0101] During operation, during patient exam setup, sensor data, preferably camera images, of a patient positioned for the exam are acquired. An algorithm for landmark detection, such as one or more skeletal joints or other anatomical features, from the sensor data is applied to obtain the landmark coordinates relative to a coordinate system. The aforementioned vertebral regression model is applied to predict the patient-specific vertebral positions. The clinical operator selects the start and end points (and / or center) of a scan region (e.g., a localizer scan region) based on the desired vertebra.

[0102] For a spine scan, the optimal vertical center of the selected spine segment is calculated based on, for example, a vertebral regression model.

[0103] This vertebra-based approach gives the clinical operator complete freedom to define the appropriate position based on a well-defined anatomical reference point (the vertebrae), in contrast to the standard approach where only a predefined set of positions can be selected from the system.

[0104] Using the spine as a reference for anatomical position has the advantage of providing a natural alignment of human anatomy to a relative size-invariant coordinate system and correlating well with articular structures such as the shoulders and hips that are typically detected on camera images.

[0105] According to any embodiment of the present invention, part of the method is acquiring subject sensor data, identifying anatomical landmarks, and predicting vertebral positions. These steps of the method are described in further detail with reference to Figures 4-6. The features described can be applied to any embodiment of the present invention.

[0106] FIG. 4 shows a schematic diagram of steps for acquiring spatial sensor data using a sensor device 52. FIG. 4 (left) shows an elevation view. FIG. 4 (bottom) shows a plan view. The sensor device in the illustrated example has multiple cameras, e.g., 3D cameras. Alternatively, only a single camera can be used. The use of a camera is not required. Other examples may include laser-based detection, non-camera optical detection, the use of electromagnetic sensing, etc.

[0107] In some embodiments, the spatial sensor data includes camera data, e.g., 3D camera data. 3D camera image data refers to depth images. These can be acquired with either a depth camera / 3D sensor, e.g., a time-of-flight camera, or a pair of stereo sensors. Using this 3D data, the outer contour of the body can be represented. By also taking into account the data of the empty table, the thickness of a specified region can be calculated, which can be used, for example, to find the vertical center.

[0108] 4 shows a patient 56 positioned within the examination region 64 of the imaging device 100. For purposes of illustration, this is shown as a CT scanner, and a rotatable gantry 104 is shown. Alternatively, other modalities can be used.

[0109] From the sensor data, the positions of a number of anatomical landmarks 62 within a coordinate system 66 of an examination region 64 are determined. The landmarks may be joints or other visible anatomical structures, such as the shoulders, head, and hips. This is shown schematically in the plan view of Figure 4 (bottom). In this example, the head, left shoulder, right shoulder, and left and right hips are detected.

[0110] Figure 5 shows a real-world example of spatial sensor data in the form of an image of a subject, with the locations of detected landmarks indicated.

[0111] By applying an algorithm such as a regression model, the positions of at least a subset of the subject's vertebrae can be detected.

[0112] Figure 6 shows a schematic representation of the predicted positions of the vertebrae of a subject's spine based on the previously detected positions of anatomical landmarks, projected onto the patient's MRI dataset.

[0113] Part of this method is the prediction of vertebral positions using algorithms such as regression or machine learning models.

[0114] As described above, the model can be trained based on multiple annotated datasets containing sensor data in the same format and modality as the sensor device 52 used in the operation, where the sensor data is annotated with both anatomical landmark locations and vertebral locations, so that the model can be trained to predict vertebral locations based on the landmark locations.

[0115] As an illustration, Figure 7 shows the accuracy of a pair of example regression models trained in this way to predict vertebral positions C6 (Figure 7, top) and L3 (Figure 7, bottom). These also correspond to the start and end points (along the z-axis, using the coordinate system already introduced above) of the ideal cardiac scan region for imaging the heart. Each graph shows the correlation between the model output and the ground truth (gt). We can see that the correlation is very strong, indicating good prediction results.

[0116] For example, by training a regression model for each single vertebra, every clinician can individually select the size of the scan region, for example, based on clinic guidelines. These selected regions can then be used to first calculate the vertical center within this region and, based thereon, perform, for example, a survey scan of the target anatomical structure.

[0117] In some embodiments, a further regression / machine learning model can be applied that is configured to predict the boundaries and / or placement of a scan volume to be scanned to capture a particular anatomical structure of interest, either based on the predicted positions of the vertebrae (already predicted using the first one or more predictive models) or directly based on the acquired spatial sensor data. In the latter case, such a model can be trained based on training data that includes multiple spatial sensor data sets annotated with landmark locations and annotated with the ideal boundary and / or center location of the volume to be scanned to capture the particular anatomical structure.

[0118] Although the preferred embodiment uses a regression model to predict vertebral position, other algorithm types can be used instead, such as other statistical models or machine learning or artificial intelligence algorithms trained in a manner similar to that already described above for regression models.

[0119] The present invention can be applied to a wide range of different image scanning techniques, including, by way of non-limiting example, MRI, CT, or DXR (Digital X-ray Radiogrammetry).

[0120] Furthermore, while the above-described embodiments refer to predicting the positions of multiple vertebrae, it should be noted that the sample principle can be equally applied to predict the positions of other bony structures, such as the pelvic bone, the skull or portions thereof, the shoulder bone, etc. Acquired sensor data (e.g., camera data) can detect visually observable anatomical landmarks but cannot detect internal structures. Thus, the general concept is to use sensor-detected landmarks to predict the positions of internal bony structures, which can effectively be used as a set of anchor points, relative to which a scan region or scanner placement can be defined, e.g., by predicting the positions of internal anatomical features relative to the bony structures.

[0121] The above-described embodiments of the invention employ a processing unit. A processing unit may generally have a single processor or multiple processors. It may be located within a single containing device, structure, or unit, or may be distributed among multiple different devices, structures, or units. Thus, reference to a processing unit adapted or configured to perform a particular step or task may correspond to the step or task performed by any one or more of multiple processing components, either alone or in combination. Those skilled in the art will understand how such a distributed processing configuration can be implemented. The processing unit may include a communication module or input / output unit for receiving data and outputting data to additional components.

[0122] The one or more processors of a processing unit can be implemented in a number of ways using software and / or hardware to perform the various functions required. The processor typically uses one or more microprocessors that are programmed using software (e.g., microcode) to perform the required functions. The processor may also be implemented as a combination of dedicated hardware to perform some functions and one or more programmed microprocessors and associated circuitry to perform other functions.

[0123] Examples of circuitry that may be used 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).

[0124] In various implementations, 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 required functions. The various storage media may be mounted within the processor or controller, or may be transportable such that one or more programs stored on the storage media can be read by the processor.

[0125] Variations to the disclosed embodiments can be understood and effected by those skilled in the art in practicing the claimed invention, from a study of the drawings, the disclosure, and the appended claims. 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.

[0126] A single processor or other unit may fulfill the functions of several items recited in the claims.

[0127] 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.

[0128] The computer program may be stored / distributed on a suitable medium such as an optical storage medium or a solid-state medium 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.

[0129] It is noted that when the term "suitable" is used in the claims or specification, the term "suitable" is intended to be equivalent to the term "configured with."

[0130] Any reference signs in the claims should not be construed as limiting the scope.

Claims

1. 1. A method for use in setting up an imaging scan to be performed using an imaging scanner, the imaging scanner having an examination area for receiving a subject to be scanned, the method comprising: acquiring spatial sensor data outside the subject's body received within the examination region; applying a predetermined algorithm to the sensor data to determine the position of one or more anatomical landmarks of the subject's body located within the examination region relative to a coordinate system associated with the examination region; applying a prediction procedure, the prediction procedure including application of a prediction algorithm configured to generate an output including a prediction of positions of a plurality of bony structures of the subject within the examination region relative to the coordinate system based on input to the prediction algorithm including the detected positions of the one or more anatomical landmarks within the examination region; defining a location and / or boundary points of a scan volume to be scanned relative to the coordinate system based on the predicted positions of the bone structures output from the prediction algorithm; generating an output indicative of defined locations and / or boundary points of the scan volume to be scanned; A method comprising:

2. The method of claim 1 , wherein the plurality of bony structures comprises a plurality of spinal vertebrae.

3. The step of defining the location and / or boundary points of the scan volume comprises: obtaining first indicators of the location and / or boundary points, the first indicators being defined relative to the location of one or more of the bone structures; - transforming the first indicators into second indicators of the positions and / or boundary points, the second indicators being defined relative to the coordinate system, the transformation comprising using the predicted positions of the bone structures output from the prediction algorithm; 3. The method of claim 1 or 2, comprising:

4. the first indication is obtained from a user interface; or the first indicator is obtained from a normative database; The method of claim 3.

5. 5. The method of claim 1, further comprising determining the positions and / or spatial extent of one or more target anatomical structures within the body relative to the coordinate system based on the predicted positions of the bony structures.

6. determining the location and / or spatial extent of the one or more target anatomical structures within the body, obtaining a first indication of the position and / or spatial extent of the one or more target anatomical structures, the first indication being defined relative to a position of one or more of the bony structures; transforming the first indicators into second indicators of the position and / or spatial extent of the one or more target anatomical structures, the second indicators being defined relative to the coordinate system, the transformation comprising use of the predicted positions of the bony structures output from the prediction algorithm; 6. The method of claim 5, comprising:

7. 7. The method of claim 5 or 6, wherein the step of defining the location and / or boundary points of the scan volume is performed based on the determined positions and / or spatial extents of the one or more target anatomical structures.

8. 8. The method of claim 5, wherein the coordinate system includes a vertical dimension, and wherein determining the position and / or spatial extent of the one or more target anatomical structures within the body comprises determining a vertical center of the spatial extent of at least one target anatomical structure within the body relative to the vertical dimension of the coordinate system, and wherein the location and / or boundary point of the scan volume is determined to align a vertical center of the scan volume with the vertical center of the at least one target anatomical structure within the body.

9. The method of claim 1 , wherein the one or more anatomical landmarks include one or more joints.

10. 10. The method according to claim 1, wherein the prediction procedure comprises the application of a plurality of prediction algorithms, one for each of the plurality of bony structures whose position is to be predicted.

11. 11. The method of claim 1, wherein the spatial sensor data is data acquired using a set of one or more sensor elements, e.g., cameras, having predefined spatial positions relative to the coordinate system.

12. 12. The method of any one of claims 1 to 11, wherein the one or more target anatomical structures comprise one or more of a portion of the spine, a head or neck, a chest, a heart or portion thereof, an abdomen, a liver, a kidney, a stomach, a prostate, a bladder, and a pelvis.

13. A computer program comprising code means adapted to, when executed on a processor, cause the processor to carry out the method according to any one of claims 1 to 12.

14. 1. A processing unit for use in setting up an imaging scan to be performed using an imaging scanner, the imaging scanner having an examination area for receiving a subject to be scanned, the processing unit comprising: Input / output section, receiving at the input / output spatial sensor data outside the subject's body received within the examination region; applying a predetermined algorithm to the sensor data to determine the positions of one or more anatomical landmarks of the subject's body located within the examination region relative to a coordinate system associated with the examination region; applying a prediction procedure, the prediction procedure including application of a prediction algorithm configured to generate an output including a prediction of positions of a plurality of bony structures of the subject within the examination region relative to the coordinate system based on input to the prediction algorithm including the detected positions of one or more anatomical landmarks within the examination region; defining a location and / or boundary points of a scan volume to be scanned relative to the coordinate system based on the predicted positions of the bone structures output from the prediction algorithm; generating, at the input / output unit, an output indicative of the defined location and / or boundary points of the scan volume to be scanned; one or more processors configured to a processing unit having:

15. A processing unit according to claim 14; an imaging scanner having an examination area for receiving a subject; A system having:

Images