WO2017162664A1

WO2017162664A1 - Apparatus for detecting deformation of a body part

Info

Publication number: WO2017162664A1
Application number: PCT/EP2017/056686
Authority: WO
Inventors: Dominik Benjamin KUTRA; Thomas Buelow
Original assignee: Koninklijke Philips N.V.
Priority date: 2016-03-21
Filing date: 2017-03-21
Publication date: 2017-09-28

Abstract

The present invention relates to an apparatus for detecting deformation of a body part. It is described to provide (210) a magnetic resonance image of a patient's body part, the magnetic resonance image of the body part comprising structural information of the body part, wherein the structural information comprises large scale structural information. A model of the body part is provided (220), the model of the body part comprising structural information of the body part, wherein the structural information comprises large scale structural information. At least one transformation is applied (230) to the model of the body part and/or to the magnetic resonance image of the body part, wherein the at least one transformation comprises transformation on the basis of at least some of the large scale structural information. A fit is determined (240) between the model of the body part and the magnetic resonance image of the body part comprising application of the at least one transformation. A deviation is determined (250) between the model of the body part and the magnetic resonance image of the body part on the basis of the structural information in the magnetic resonance image and the structural information in the model of the body part. Information relating to the deviation is provided (260).

Description

APPARATUS FOR DETECTING DEFORMATION OF A BODY PART

FIELD OF THE INVENTION

The present invention relates to an apparatus for detecting deformation of a body part, to a medical system for detecting deformation of a body part, and to a method for detecting deformation of a body part, as well as to a computer program element and a computer readable medium.

BACKGROUND OF THE INVENTION

Magnetic resonance (MR) breast imaging involves the usage of breast coils to ensure high signal and, thus, high image quality. The breasts might, furthermore, be supported by cushions to prevent the breasts from moving during the (lengthy) acquisition. As a result the breast is not hanging freely inside the breast coil. This has multiple disadvantages including reduced reproducibility between longitudinal studies and undefined loading of the breast for biomechanical simulations. For instance, positioning of the breasts, even using the same model of breast coil, varies between studies. In addition, using unsuitable coils can alter shape of the breast significantly. The same applies to MR imaging of other body parts.

US201 1/0142308A1 describes an information processing apparatus and an information processing method including a deformation shape model generation unit configured to generate, from information about a first shape and a first position of a feature region in a target object under a first deformation condition, a deformation of the first shape with the position of the feature region as a reference as a model, and a deformation estimation unit configured to, based on information about a second shape and a second position corresponding to the feature region in the target object under the second deformation condition, align the first position with the second position to estimate deformation from the first shape to the second shape using the model.

US2008/292164A1 describes that a method for joint analysis of non- concurrent magnetic resonance (MR) and diffuse optical tomography (DOT) images of the breast includes providing a digitized MR breast image volume comprising a plurality of intensities corresponding to a 3-dimensional (3D) grid of voxels, providing a digitized DOT breast dataset comprising a plurality of physiological values corresponding to a finite set of points, segmenting the breast MR image volume to separate tumorous tissue from non- tumorous tissue, registering a DOT breast dataset and the MR image volume and fusing said registered DOT and MR datasets, wherein said fused dataset is adapted for analysis.

SUMMARY OF THE INVENTION

It would be advantageous to have improved apparatus for detecting deformation of a body part.

The object of the present invention is solved with the subject matter of the independent claims, wherein further embodiments are incorporated in the dependent claims. It should be noted that the following described aspects and examples of the invention apply also for the apparatus for detecting deformation of a body part, medical system for detecting deformation of a body part, the method for detecting deformation of a body part, and for the computer program element and the computer readable medium.

In an aspect, there is provided an apparatus for detecting deformation of a body part as defined in appended claim 1. In another aspect, there is provided a medical system for detecting deformation of a body part as defined in appended claim 1 1. In another aspect, there is provided a method for detecting deformation of a body part as defined in appended claim 12.

In an example, there is provided an apparatus for detecting deformation of a body part, the apparatus comprising:

an input unit; and

a processing unit.

The input unit is configured to provide a magnetic resonance image of a patient's body part. The magnetic resonance image of the body part comprising structural information of the body part, wherein the structural information comprises large scale structural information. The input unit is configured to provide a model of the body part. The model of the body part comprising structural information of the body part, wherein the structural information comprises large scale structural information. The processing unit is configured to apply at least one transformation to the model of the body part and/or to the magnetic resonance image of the body part, wherein the at least one transformation comprises transformation on the basis of at least some of the large scale structural information. The processing unit is also configured to determine a fit between the model of the body part and the magnetic resonance image of the body part comprising application of the at least one transformation. After application of the at least one transformation, the processing unit is configured to determine a deviation between the model of the body part and the magnetic resonance image of the body part on the basis of the structural information in the magnetic resonance image and the structural information in the model of the body part. The processing unit is configured to provide information relating to the deviation.

Thus, in an example a model can be transformed towards imagery of the body part but done at such a scale that the model does not conform to local deviations in the imagery of the body part. Then, a comparison between the transformed model and the image of the body part can be used to determine deformation of the body part, through for example determining the location and quantifying the local deviations in the imagery with respect to the model. However, it is clear that the image of the body part could be transformed in the same (or similar) way that the model was transformed in the above example, where this transformation is done such that local deviations in the imagery are not washed out to conform to the model that does not exhibit such local deviations. Again, comparison between the model and now the transformed image of the body part can be used to determine deformation of the body part, through for example determining the location and quantifying the local deviations in the imagery with respect to the model. Similarly, both the image and the model could be transformed such that the local deviations can be identified.

In other words, using a common representation or model of the shape of a body part (e.g., a breast) that is fitted to a (e.g., breast) MR image using a limited set of transformations, the deviation of the imaged (breast) shape from the model can be measured. This measure can be used to assess the positioning quality, warn the user about bad positioning and be used to inform that different or better equipment, such as a different breast MR-coil, should be utilised.

To put this another way, the global shape of a model of a body part can be made to match that of a magnetic resonance image of a patient's body part, from which local deformations of the patient's body part can be determined. By applying a transformation on the basis of large scale structural information of the body part, relatively small scale deformations are maintained with respect to their local surroundings. If these deformations in the magnetic resonance image data, with respect to the model, are of a significant enough deviation then this indicates that the acquired image may not be fit for purposes. For example, it can indicate that the body part has suffered deformation to such an extent that this image is not suitable for comparison with a previously acquired image as part of a longitudinal study. Remedial action, to mitigate the deformation, can then be put in place. To put this another way, to enable tumour tracking applications a magnetic resonance image acquired now is commonly compared to a previously acquired magnetic resonance image, for example to determine if a tumour has grown or diminished in size. Registration approaches can be applied, registering one image onto the other, through for example stretching and translating one image until it matches the other. However, such registration can lead to an artificial scaling up or down of the imagery relating to the tumour, a situation that is exacerbated if one of the images, prior to registration, was acquired with the body part suffering a deformation that was not present when the other image was acquired. The present apparatus enables a determination to be made of those images that are suitable, or indeed not suitable, to be compared in this manner. Also, the magnetic resonance image can be part of a scout scan, prior to a lengthy and costly full MRI scan, and enables remedial action to be taken prior to the full scan if the body part is being deformed in some manner, thereby saving time and money. In an example, the scout scan can comprise utilisation of a 2D model, for example a super ellipse model.

In an example, the model of the body part is a generalised model of the body part.

In other words, a generalised model can be used as an input. And either, the model can be transformed toward the magnetic resonance image or the magnetic resonance image can be transformed toward the model, in order that a deformation of the body part can be detected.

By providing a generalised model, this can be easily applied to different patients and to the same patient attending different clinics.

In an example, the model of the body part comprises small scale structural information and/or the magnetic resonance image of the body part comprises small scale structural information, and wherein the at least one transformation substantially maintains a relative scaling of at least some of the small scale structural information.

Here, small scale structural information means small with respect to the large scale structural information. In this manner, the at least one transformation does not wash out, or introduce, small scale deviations in the magnetic resonance image or the model respectively. In this way, after determining a fit between the magnetic resonance image and the model, any deformation in the body part as shown in the magnetic resonance image will be identifiable with respect to the model of the body part, and a deviation determined.

In other words, a rather rigid of transformations is used such that the model does not adapt to the type of body part (e.g. breast) deformations that are seeking to be identified, where equally as described above the transformation could be applied to the image of the body part toward the model, where the deformations in the image are not being adapted to the model.

In an example, the at least one transformation comprises an affine transformation.

In this way, the well known affine transformation can be easily utilised as part of the transformation. By using the affine transformation, local features are not washed out in the magnetic resonance image or introduced into the model, such that the deformation can be detected and the deviation between image and model determined.

In an example, the at least one transformation is only applied to the model of the body part.

In this manner, the magnetic resonance image is not altered meaning that an image with significant structural content does not need to be transformed rather a model that can be relatively simple is transformed, leading to computational efficiency and expediency. Also, memory is saved as an updated magnetic resonance image does not have to be saved. Additionally, the magnetic resonance image can then be used for other purposes.

In an example, the determination of the deviation comprises at least one distance measurement between the structural information in the magnetic resonance image and the associated structural information in the model of the body part.

In this manner, a simple metric (a distanced or length measurement) that also have a clear intuitive meaning to a user, is used as part of the determination of deviation. This enables a clinician to more easily use their expertise to determine if the image is fit for purpose or whether remedial action is required. Also, computational efficiency is provided through the use of such a simple metric.

In an example, the at least one distance measurement is a plurality of distance measurements, and wherein the determination of the deviation comprises a local cluster of the distance measurements.

In this manner, size and extent of a deformation can be determined. Also, the statistical significance of a deformation can be determined and better enables real

deformation to be separated from noise and enables smaller deformations to be detected and acted upon.

In an example, the structural information of the body part comprises a shape of the body part, and wherein the deviation is determined on the basis of the shape of the body part in the model of the body part and the associated shape of the body part in the magnetic resonance image of the body part.

In this manner, the deviation of the imaged body part shape (e.g., breast) with respect to the model of the body part can be determined and measured.

In an example, the structural information of the body part comprises a contour of the body part, and wherein the deviation is determined on the basis of the contour in the model of the body part and the associated contour in the magnetic resonance image of the body part.

In this manner a contour in the image data, such as an outer boundary representing the skin of the patient (for example of a breast) can be compared to the outer boundary of a model of the body part (e.g., breast) from which the deformation of the breast in the imagery can be determined. Highly efficient and robust image analysis algorithms, such as edge detection, can then be utilised as part of this determination in a computationally efficient and robust manner.

In an example, determining a fit between the model of the body part and the magnetic resonance image of the body part comprises determining a 2D projection of the model of the body part and determining a 2D image of the magnetic resonance image.

In other words, the determination and characterisation of the deformation can be determined on the basis of 2D data, providing for computational efficiency and providing for data which can be output that is easily interpretable by a user.

In an example, determining a fit between the model of the body part and the magnetic resonance image of the body part comprises registration of the model of the body part to the magnetic resonance image of the body part.

In another aspect, there is provided a medical system for detecting deformation of a body part, the system comprising:

a magnetic resonance image acquisition unit;

an apparatus for detecting deformation of a body part according to the above described aspect and any of the associated examples; and

an output unit.

The magnetic resonance image acquisition unit is configured to provide the magnetic resonance image. The output unit is configured to output the information relating to the deviation.

In an example, there is provided a method for detecting deformation of a body part, the method comprising: a) providing a magnetic resonance image of a patient's body part, the magnetic resonance image of the body part comprising structural information of the body part, wherein the structural information comprises large scale structural information;

b) providing a model of the body part, the model of the body part comprising structural information of the body part, wherein the structural information comprises large scale structural information;

c) applying at least one transformation to the model of the body part and/or to the magnetic resonance image of the body part, wherein the at least one transformation comprises transformation on the basis of at least some of the large scale structural information;

d) determining a fit between the model of the body part and the magnetic resonance image of the body part comprising application of the at least one transformation;

e) determining a deviation between the model of the body part and the magnetic resonance image of the body part on the basis of the structural information in the magnetic resonance image and the structural information in the model of the body part, wherein step e) is carried out after step c); and

f) providing information relating to the deviation.

According to another aspect, there is provided a computer program element controlling apparatus as previously described which, in the computer program element is executed by processing unit, is adapted to perform the method steps as previously described.

According to another aspect, there is provided a computer readable medium having stored computer element as previously described.

Advantageously, the benefits provided by any of the above aspects equally apply to all of the other aspects and vice versa.

The above aspects and examples will become apparent from and be elucidated with reference to the embodiments described hereinafter.

BRIEF DESCRIPTION OF THE DRAWINGS

Exemplary embodiments will be described in the following with reference to the following drawings:

Fig. 1 shows a schematic set up of an example of an apparatus for detecting deformation of a body part;

Fig. 2 shows a schematic set up of an example of a medical system for detecting deformation of a body part; Fig. 3 shows an example of a method for detecting deformation of a body part; Fig. 4 shows example images of breast deformations in MR images;

Fig. 5 shows schematically the information shown in the images of Fig. 4; Fig. 6 shows schematically an example of an image of a breast and a model of a breast;

Fig. 7 shows an example of part of a process of model development;

Fig. 8 shows an example case of an image of a breast with a fitted model;

Fig. 9 shows an example 3D view of an adapted model;

Fig. 10 shows an example of image voxels with a fitted model;

Fig. 1 1 shows an example of image voxels, with a fitted model having been removed;

Fig. 12 shows an example of clustering;

Fig. 13 shows a further example of clustering;

DETAILED DESCRIPTION OF EMBODIMENTS

Fig. 1 shows as example of an apparatus 10 for detecting deformation of a body part. The apparatus 10 comprises an input unit 20 and a processing unit 30. The input unit 20 is configured to provide a magnetic resonance image of a patient's body part. The magnetic resonance image of the body part comprises structural information of the body part, wherein the structural information comprises large scale structural information. The input unit 20 is also configured to provide a model of the body part. The model of the body part comprises structural information of the body part, wherein the structural information comprises large scale structural information. The processing unit 30 is configured to apply at least one transformation to the model of the body part and/or to the magnetic resonance image of the body part. The at least one transformation comprises transformation on the basis of at least some of the large scale structural information. The processing unit 30 is also configured to determine a fit between the model of the body part and the magnetic resonance image of the body part comprising application of the at least one transformation. After application of the at least one transformation, the processing unit 30 is configured to determine a deviation between the model of the body part and the magnetic resonance image of the body part on the basis of the structural information in the magnetic resonance image and the structural information in the model of the body part. The processing unit 30 is configured to provide information relating to the deviation. The apparatus is able to detect unwanted deformations in the body part (e.g., a breast).

In an example, the magnetic resonance image comprises 3D data. In an example, the magnetic resonance image comprises a 2D slice of the body part, i.e., comprises a subset of 3D magnetic resonance image data.

In an example, the magnetic resonance image data comprises a 2D image. In an example, the body part is a breast, a leg, and arm, a face, a patient's back, or an internal part of the body such as a liver, a kidney.

In an example, the large scale structural information comprises the shape of the body part. In an example, the large scale structural information comprises the general or average shape of the body part. In an example, the large scale structural information comprises a general outline of the body part, such that small scale perturbations in that outline are not washed out during the transformation. In an example, the large scale structural information comprises a width and/or length (or height) of the body part.

In an example, the structural information comprises an outer extent of the body part. In other words, deformation of the body part leading to an associated deformation of an outer extent of the body part results in a magnetic resonance image that exhibits an outer deformed extent or surface. This deformation will not be present in the model, and the deviation between the model and the imagery can be determined. In an example, the structural information comprises the internal structure within the body part, such as the vascular structure or muscular structure. For example, a model can have this internal structure, and if the patient's body part is suffering deformation such that the vascular structure or muscular structure is not as shown in the model, then a deviation between the model and imagery can be determined. Thus, a body part can be being deformed in the imaging direction, such that a cross section is as it should be. In other words, the body part could be deformed in the viewing direction, through for example a cushion pushing the body part in the viewing direction rather than laterally to the viewing direction. An outer extent of the body part then appears to be correct, but the internal parts of the body part are suffering deformation, and this can be detected and characterised. Such deformation could also result from stitching of a wound for example, that is not leading to a noticeable deformation of the viewable outer extent of the body part, but is nevertheless still deforming the body part such that image data may not be suitable for comparison with previously acquired image data. In an example, a fit between the model and the magnetic resonance image of the body part comprises solving an optimisation problem comprising minimization of at least one distance between the image data and the model

In an example, a breast coil is leading to deformation of a breast. In an example, a cushion is leading to deformation of a body part such as a breast or leg/arm/chest. In an example, stitches in a body part are deforming that body part such that the body part is deformed. In this manner, a clinician can be made aware of this fact and take it into account when for example comparing a magnetic resonance image acquired now with that acquired previously.

In an example, an alarm is raised when the deviation exceeds a pre-set threshold. In an example, the threshold can be learnt from data that has been annotated and clustered with respect to the severity of observed deformations by clinical experts. In an example, the threshold can be defined as a multiple (or on the basis of) a mean distance of the model from a surface of the image data.

In an example, a mean model of the body part (e.g., breast) is built using principle component analysis (PC A).

In an example, a semi-super-ellipsoid is utilized in generating the model. In an example, a semi-super-ellipsoid having 5 internal parameters is utilized in generating the model.

In an example, a model can be generated using ground truth segmentations which are aligned (e.g. by an interactive closest point (ICP) algorithm) and averaged. In an example, utilization of ground truth segmentations can be used to generate a mean model.

In an example, the model of the body part is generated from image data of the body part from patients, where the body part for those patients is known not to be suffering deformation.

In an example, the model of the body part (e.g., breast) can be represented by a parametric shape.

According to an example, the model of the body part is a generalised model of the body part.

According to an example, the model of the body part comprises small scale structural information and/or the magnetic resonance image of the body part comprises small scale structural information, and wherein the at least one transformation substantially maintains a relative scaling of at least some of the small scale structural information. According to an example, the at least one transformation comprises an affine transformation.

According to an example, the at least one transformation is only applied to the model of the body part.

According to an example, the determination of the deviation comprises at least one distance measurement between the structural information in the magnetic resonance image and the associated structural information in the model of the body part.

In an example, a fit between the model and the magnetic resonance image of the body part comprises solving an optimisation problem comprising minimization of at least one distance between the structural information in the magnetic resonance image and the associated structural information in the model of the body part.

In an example, the at least one distance measurement comprises computing the Euclidean distance of the image points to the model. In an example, the at least one distance measurement comprises computing an algebraic distance of the image points to the model.

According to an example, the at least one distance measurement is a plurality of distance measurements, and wherein the determination of the deviation comprises a local cluster of the distance measurements.

In an example, a local cluster comprises the binning of points in pre-defined distance ranges. In an example, a deviation can be defined as a cluster-mean that is outside a pre-defined distance threshold.

In an example, the at least one distance measurement is a plurality of distance measurements, and wherein the determination of the deviation comprises an aggregate of the distance measurements.

According to an example, the structural information of the body part comprises a shape of the body part, and wherein the deviation is determined on the basis of the shape of the body part in the model of the body part and the associated shape of the body part in the magnetic resonance image of the body part.

According to an example, the structural information of the body part comprises a contour of the body part, and wherein the deviation is determined on the basis of the contour in the model of the body part and the associated contour in the magnetic resonance image of the body part.

In an example, the contour is an external contour of the body part. In an example, the deviation comprises a distance measurement between the contour in the magnetic resonance image and the associated contour in the model of the body part.

In an example, a deformation field of the model to the contour of the body part (e.g. surface of the image of a breast) is defined, providing information on the magnitude of displacement between the model and the contour of the body part for every point.

In other words a deviation, such as a distance measure, can be derived from a transformation that matches the outline of the breast to the model, but does so at a fidelity level such that relatively small scale features resulting from a deformation are still measureable. Limits in magnitude can be used as a quality measure to raise an alarm.

According to an example, determining a fit between the model of the body part and the magnetic resonance image of the body part comprises determining a 2D projection of the model of the body part and determining a 2D image of the magnetic resonance image.

In an example, the information relating to the deviation is useable to provide a warning.

In an example, the 2D image of the magnetic resonance data is a 2D slice through the magnetic resonance data.

According to an example, determining a fit between the model of the body part and the magnetic resonance image of the body part comprises registration of the model of the body part to the magnetic resonance image of the body part.

In an example, the registration comprises overlaying the model of the body part with the magnetic resonance image of the body part. In an example, the registration comprises overlaying a 2D projection of the model of the body part with 2D image data of the magnetic resonance image of the body part.

Fig. 2 shows an example of a medical system 100 for detecting deformation of a body part. The system 100 comprises a magnetic resonance image acquisition unit 1 10, an apparatus 10 for detecting deformation of a body part as described above with respect to any of the examples associated with Fig. 1, and an output unit 120. The magnetic resonance image acquisition unit 1 10 is configured to provide the magnetic resonance image. The output unit 120 is configured to output the information relating to the deviation.

In an example, the output unit is configured to output a representation of the magnetic resonance image. In an example, the output unit is configured to output a representation of the model of the body part. In an example, the output unit is configured to output a representation of the magnetic resonance image and a representation of the model of the body part. In other words, the magnetic resonance image of the body part can be presented with an indication of the model of the body part overlaid over the image, providing visual information of the deformation in the image data. In an example, the output unit is configured to output a representation of the magnetic resonance image and/or the model along with an indication of the deviation between the model and the magnetic resonance image.

Fig. 3 shows a method 200 for detecting deformation of a body part in its basic steps. The method comprises:

In a providing step 210, also referred to as step a), a magnetic resonance image of a patient's body part is provided, the magnetic resonance image of the body part comprising structural information of the body part, wherein the structural information comprises large scale structural information.

In a providing step 220, also referred to as step b), a model of the body part is provided, the model of the body part comprising structural information of the body part, wherein the structural information comprises large scale structural information;

In an applying step 230, also referred to as step c), at least one transformation is applied to the model of the body part and/or to the magnetic resonance image of the body part, wherein the at least one transformation comprises transformation on the basis of at least some of the large scale structural information.

In a determining step 240, also referred to as step d), a fit is determined between the model of the body part and the magnetic resonance image of the body part comprising application of the at least one transformation.

In a determining step 250, also referred to as step e), a deviation is determined between the model of the body part and the magnetic resonance image of the body part on the basis of the structural information in the magnetic resonance image and the structural information in the model of the body part, wherein step e) is carried out after step c).

In a providing step 260, also referred to as step f), information relating to the deviation is provided.

In an example, step e) comprises measuring at least one distance between the structural information in the magnetic resonance image and the associated structural information in the model of the body part.

In an example, step e) comprises measuring a plurality of distance measurements, and determining a local cluster of the distance measurements. In an example, the structural information of the body part comprises a shape of the body part, and wherein the deviation is determined on the basis of the shape of the body part in the model of the body part and the associated shape of the body part in the magnetic resonance image of the body part.

In an example, step d) comprises determining a 2D projection of the model of the body part and determining a 2D image of the magnetic resonance image.

In an example, step d) comprises registration of the model of the body part to the magnetic resonance image of the body part.

In other words, a method for automatically detecting breast (or other body part) deformation in MR images induced by breast coils (or cushions or other means that can lead to deformation of a body part) by using a common representation or model of the shape of the breast that is fitted to a body part (e.g. breast) MR image using a limited set of transformations. This enables the positioning quality to be assessed, facilitating the warning of a user about bad positioning, and prompting the suggestion that better equipment (breast MR-coil) for the imaged body part (e.g. breast) be used.

The apparatus, system and method for detecting deformation of a body part will now be described in more detail with reference to figs 4-13. The situation for breast deformation and breast imaging is presented, however applicability to the detection of the deformation of other body parts such as a leg, arm, back or internal body part such as liver is provided.

Figs. 4 and 5 show examples of breast deformations in MR images. The arrows indicate deformations of the breast shape caused by breast coils and/or cushions. MR breast imaging involves the usage of breast coils to ensure high signal and, thus, high image quality. Automatic assessment of the image quality in terms of positioning of the breast is an important task considering the restrictions on reading times. In applications like tumor tracking multiple MR images are taken in a longitudinal study to assess the response of a certain treatment with regards to the tumour size. The time between acquisitions can be in the order of months. Positioning of the breasts, even using the same model of breast coil, varies between studies. In addition, using unsuitable coils can alter shape of the breast significantly, as shown in Figs. 4 and 5. Furthermore, if the breast MR image is used in biomechanical simulations, deformations from breast coils result in undefined loading of the breast that cannot be accurately accounted for in numerical models.

The apparatus, system and method for detecting deformation of a body part addresses these issues, enabling deformation of the body part in the associated MR image to be determined. The clinician can then either take remedial action, such as repositioning items used in the imaging such as breast coils, or cushions, or make a note of the deformation such that it can be taken into account when comparing the presently acquired image to previously acquired images as part of a longitudinal study.

Fig. 6 schematically shows a representation of the whole breast on the left- hand side, to which a model has been fitted or adapted. On the right-hand side a magnified view is presented, within which distance measurements are shown between the fitted model and the image boundary. As shown in Fig. 6 a common representation or model of the shape of the breast is fitted to a breast MR image using a limited set of transformations. The deviation of the imaged breast shape from the model can then be measured. This measure can be used to assess the positioning quality, warn the user about bad positioning and suggest better equipment (breast MR-coil) for the imaged breast.

In more detail, the following relates to an example workflow for using the apparatus for detecting deformation of the body part:

• Build a mean model of the breast shape using techniques such as PCA, or define an

appropriate parametric model that represents the breast shape.

• Adapt the model representation to image data using for example affine transformation

• Measure distance of breast contour to the fitted model (see Fig. 6).

• Cluster distances locally or aggregate to a single measure to indicate deviation of the observed shape from the model representation.

· Raise an alarm when the deviation from the measured shape and the model representation exceeds a pre-set threshold.

• The assessment can be performed on a scout scan before running time-consuming MR sequences and especially before injection of an eventual contrast agent. This allows the technologist to reposition the patient before the actual MR study is performed.

· The distance measure can be derived from a transformation that matches the outline of the breast to the model. Limits in magnitude of the deformation field can be used as a quality measure to raise an alarm. Furthermore, the transformation can be applied to the image. The resulting image shows reduced breast coil deformation artifacts. • Therefore, not only can alarms be displayed to the user, depending on the amount of breast deformation, but also suggestions can be made to the user, to use for example more appropriate equipment (maybe directly referring to sales organization) via a user interface. Furthermore, the image can be overlaid with the local distance measure.

· Furthermore, the apparatus, system and method can be used in training environments.

As part of the apparatus, system and method a model of the body part is fitted to the magnetic resonance image of the body part. Development of the model and the fitting of the model image data is now described, where reference is made to Fig. 7, which shows how longitudinal correspondence is established. In the present approach, a geometric primitive (a semi-super ellipsoid) is fitted to patient data. Anatomical knowledge is incorporated by fixing the tip of the super ellipsoid to the Mammalia position and

constraining its centre-point to a reference plane defined by landmarks on the sternum. A coordinate system is then constructed by linearly scaling the fitted super ellipsoid, defining a unique set of parameters to each point in the image volume. By fitting such a coordinate system, positional correspondence can be generated. In other words, an automated reporting procedure is provided by fitting an anatomically oriented three-dimensional coordinate system to the image data. A fitted breast coordinate system enables comparison of longitudinal studies by establishing correspondence between different images acquired at different times.

For MR imaging of the breast, the breast hangs freely inside breast coils during acquisition which can deform the breast. A semi-super ellipsoid is developed in approximation of the shape, as described in the equation below.

This equation evaluates to F(x) 1 for coordinates X = (xi, x₂, x₃) that lie exactly on the surface of the super ellipsoid. For coordinates inside the super ellipsoid F(x) < 1 applies and for coordinates outside F(x) > 1 applies. The width of the super ellipsoid is controlled via w = (ω_χ1, ω_χ2, ω_χ3) . The exponents e = {<¾,%) control the shape the super ellipsoid. The roundness in the x₃ direction is controlled by ei, and e₂ controls the roundness of the shape in the xi-x₂ plane. For ei = e₂ = 1, the resulting shape is an ellipsoid.

In this way 5 parameters ω_χ2, ω_χ3, ,%) characterize the size and shape of the super ellipsoid, with 6 additional parameters introduced relating to translation along t = (_,t_xV t_x2, t_x3) and rotation of the super ellipsoid about the Euler angles r = (α, β, γ).

In the fitting procedure of the model to image data, an appropriate distance metric, e.g.

orthogonal distance, or normalized algebraic distance, di from a data point in the image to the super ellipsoid is used in the following objective function.

A semi-super ellipsoid is then fitted to image data. The extent of the breast is constrained, both in the anterior and posterior direction. In the anterior direction the breast skin is confining the breast. The peak of the super ellipsoid is fixed to the mammilla position. To constrain the shape in the posterior direction the following is undertaken. Two landmarks are set on the sternum, a plane is then constructed such that its normal vector is perpendicular to the line connecting the 2 landmarks and parallel to the transversal plane. The mammilla position is then projected onto this plane. The main axis of the super ellipsoid is then constrained to range from the mammilla to a point inside a circle described by the radius (r_max) around the projected mammilla position. Therefore ω_χ3 is a function of a radius r e \ , r_max \ and the angle around the projected mammilla position φ€ |0,2π| . Additionally, the super ellipsoid may rotate around its main axis during the fitting procedure by an angle

I ff Tt

— 7,7 . The surface with F(x) = 1 is then fitted to the skin surface of the breast, where the fitting parameters can be reduced to 7 (.ω_χί, ω_χ2, ω_χ2, {τ, φ),β_χ, β_ζ, ).

Figs 8-13 show fitting of the model to the breast, and the determination and quantification of deformations. In Fig. 8 an image of the breast is shown with the fitted model, with the 3-D view of the adapted model shown in Fig. 9. Fig. 10 shows image voxels, which are shown represented as boxes in 3-D, the position and colour of which encode the distance of the voxel to the model. In this manner, a deformation can be easily identified and quantified. In Fig. 1 1, equivalent data to that shown in fig. 10 is presented, except that the model is not shown in order to enhance visibility of the image voxels. In Fig. 12, image voxels have been clustered in order to better enable deformation of the breast to be identified and quantified. In this image, the lightest coloured voxels represent a distance of 0-5mm from the model, the next lightest represent a distance of 5- 10mm from the model, the next a distance of 10- 15mm from the model, with the darkest representing a distance of 15mm and above from model. In this way, for example an alarm can be raised when parts of the image, for example, are more than 10 mm away from the model. A constraint can be applied that a certain cluster size of such voxels be present in order for the alarm to be raised. In Fig. 13, K means clustering with K = 3 is shown. In this clustering, the lightest coloured voxels represent a deviation of 2.9mm, the next lightest 9.5mm deviation, and the darkest voxels a deviation of 21.7mm. Here the threshold for the mean should be said lower than 9.5 mm to detect the largest cluster. Furthermore, model-based knowledge could be used to further refine the method by only measuring the distance at positions of the model that are prone to additional deformation

In another exemplary embodiment, a computer program or computer program element is provided that is characterized by being configured to execute the method steps of the method according to one of the preceding embodiments, an appropriate system.

The computer program element might therefore be stored on a computer unit, which might also be part of an embodiment. This computing unit may be configured to perform or induce performing of the steps of the method described above. Moreover, it may be configured to operate the components of the above described apparatus. The computing unit can be configured to operate automatically and/or to execute the orders of a user. A computer program may be loaded into a working memory of a data processor. The data processor may thus be equipped to carry out the method according to one of the preceding embodiments.

This exemplary embodiment of the invention covers both, a computer program that right from the beginning uses the invention and computer program that by means of an update turns an existing program into a program that uses invention.

Further on, the computer program element might be able to provide all necessary steps to fulfill the procedure of an exemplary embodiment of the method as described above.

According to a further exemplary embodiment of the present invention, a computer readable medium, such as a CD-ROM, is presented wherein the computer readable medium has a computer program element stored on it which computer program element is described by the preceding section.

A computer program may be stored and/or 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 telecommunication systems. However, the computer program may also be presented over a network like the World Wide Web and can be downloaded into the working memory of a data processor from such a network. According to a further exemplary embodiment of the present invention, a medium for making a computer program element available for downloading is provided, which computer program element is arranged to perform a method according to one of the previously described embodiments of the invention.

It has to be noted that embodiments of the invention are described with reference to different subject matters. In particular, some embodiments are described with reference to method type claims whereas other embodiments are described with reference to the device type claims. However, a person skilled in the art will gather from the above and the following description that, unless otherwise notified, in addition to any combination of features belonging to one type of subject matter also any combination between features relating to different subject matters is considered to be disclosed with this application.

However, all features can be combined providing synergetic effects that are more than the simple summation of the features.

While the invention has been illustrated and described in detail in the drawings and foregoing description, such illustration and description are to be considered illustrative or exemplary and not restrictive. The invention is not limited to the disclosed embodiments. Other variations to the disclosed embodiments can be understood and effected by those skilled in the art in practicing a claimed invention, from a study of the drawings, the disclosure, and the dependent 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. A single processor or other unit may fulfill the functions of several items re- cited in the claims. The mere fact that certain measures are re-cited in mutually different dependent claims does not indicate that a combination of these measures cannot be used to advantage. Any reference signs in the claims should not be construed as limiting the scope.

Claims

CLAIMS:

1. An apparatus (10) for detecting deformation of a body part, the apparatus comprising:

an input unit (20); and

a processing unit (30);

wherein, the input unit is configured to provide a magnetic resonance image of a patient's body part, the magnetic resonance image of the body part comprising structural information of the body part, wherein the structural information comprises large scale structural information;

wherein, the input unit is configured to provide a model of the body part, the model of the body part comprising structural information of the body part, wherein the structural information comprises large scale structural information;

wherein, the processing unit is configured to apply at least one transformation to the model of the body part and/or to the magnetic resonance image of the body part, wherein the at least one transformation comprises transformation on the basis of at least some of the large scale structural information;

wherein, the processing unit is configured to determine a fit between the model of the body part and the magnetic resonance image of the body part comprising application of the at least one transformation;

wherein, after application of the at least one transformation, the processing unit is configured to determine a deviation between the model of the body part and the magnetic resonance image of the body part on the basis of the structural information in the magnetic resonance image and the structural information in the model of the body part, wherein the determination of the deviation comprises at least one distance measurement between the structural information in the magnetic resonance image and the associated structural information in the model of the body part; and

wherein, the processing unit is configured to provide information relating to the deviation.

2. Apparatus according to claim 1, wherein the model of the body part is a generalised model of the body part.

3. Apparatus according to any of claims 1 -2, wherein the model of the body part comprises small scale structural information and/or the magnetic resonance image of the body part comprises small scale structural information, and wherein the at least one transformation substantially maintains a relative scaling of at least some of the small scale structural information.

4. Apparatus according to any of claims 1-3, wherein the at least one

transformation comprises an affine transformation.

5. Apparatus according to any of claims 1-4, wherein the at least one

transformation is only applied to the model of the body part.

6. Apparatus according to claim 1, wherein the at least one distance measurement is a plurality of distance measurements, and wherein the determination of the deviation comprises a local cluster of the distance measurements.

7. Apparatus according to any of claims 1-6, wherein the structural information of the body part comprises a shape of the body part, and wherein the deviation is determined on the basis of the shape of the body part in the model of the body part and the associated shape of the body part in the magnetic resonance image of the body part.

8. Apparatus according to any of claims 1-7, wherein the structural information of the body part comprises a contour of the body part, and wherein the deviation is determined on the basis of the contour in the model of the body part and the associated contour in the magnetic resonance image of the body part.

9. Apparatus according to any of claims 1-8, wherein determining a fit between the model of the body part and the magnetic resonance image of the body part comprises determining a 2D projection of the model of the body part and determining a 2D image of the magnetic resonance image.

10. Apparatus according to any of claims 1-9, wherein determining a fit between the model of the body part and the magnetic resonance image of the body part comprises registration of the model of the body part to the magnetic resonance image of the body part.

1 1. A medical system (100) for detecting deformation of a body part, the system comprising:

a magnetic resonance image acquisition unit (1 10);

an apparatus (10) for detecting deformation of a body part according to any of the preceding claims; and

- an output unit (120);

wherein, the magnetic resonance image acquisition unit is configured to provide the magnetic resonance image; and

wherein, the output unit is configured to output the information relating to the deviation.

12. A method (200) for detecting deformation of a body part, the method comprising:

a) providing (210) a magnetic resonance image of a patient's body part, the magnetic resonance image of the body part comprising structural information of the body part, wherein the structural information comprises large scale structural information;

b) providing (220) a model of the body part, the model of the body part comprising structural information of the body part, wherein the structural information comprises large scale structural information;

c) applying (230) at least one transformation to the model of the body part and/or to the magnetic resonance image of the body part, wherein the at least one transformation comprises transformation on the basis of at least some of the large scale structural information;

d) determining (240) a fit between the model of the body part and the magnetic resonance image of the body part comprising application of the at least one transformation; e) determining (250) a deviation between the model of the body part and the magnetic resonance image of the body part on the basis of the structural information in the magnetic resonance image and the structural information in the model of the body part, comprising measuring at least one distance measurement between the structural information in the magnetic resonance image and the associated structural information in the model of the body part, wherein step e) is carried out after step c); and

f) providing (260) information relating to the deviation.

13. A computer program element for controlling an apparatus according to one of claims 1 to 1 1 , which when executed by a processor is configured to carry out the method of

14. A computer readable medium having stored the program element of claim 13.