Method for predicting the effects of an intracorporeal surgical intervention on the external appearance of the body segment

Abstract

A method for predicting the effects of an intracorporeal surgical intervention applied to a body segment comprising at least one organ on the external appearance of the body segment, the method comprising implementing, by means of data processing on a server, the following steps: - step (a) obtaining a biomechanical model of the body segment; - step (b) applying at least one principal deformation to at least one part of interest of a portion of the model relating to the organ, said principal deformation being representative of a step of the intracorporeal surgical intervention relating to the part of interest; - step (c) applying an incident deformation to the model, excluding the part of interest, said incident deformation being a function of the principal deformation. Figure 2 for the abstract.

Description

Title of the invention: Method for predicting the effects of an intracorporeal surgical intervention on the external appearance of the body segment. FIELD OF THE INVENTION

[0001] The present invention relates to the field of biomechanics. More specifically, it relates to a method for predicting the effects of an intracorporeal surgical intervention on the external appearance of the body segment. STATE OF THE ART

[0002] An intracorporeal surgical procedure performed on a body segment generally has effects on the external appearance of that body segment. For example, tibial osteotomy procedures, the purpose of which is to correct the alignment of the lower limb in order to relieve pressure on the knee joint surfaces, significantly alter the external appearance of the lower limb. Similarly, bunion surgery alters the alignment of the big toe and thereby changes the external appearance of the foot.

[0003] Currently, to enable a patient to imagine the future appearance of their body segment which is going to undergo surgery, surgeons show photos of body segments of other patients who have undergone a similar procedure.

[0004] However, patients cannot visualize what their own body segment will look like after the procedure.

[0005] However, this visualization could help in the decision between two types of surgical interventions, for example to choose between two cosmetic interventions having different repercussions on external appearance.

[0006] In addition, this visualization could allow patients to anticipate the effects of a surgical intervention on their body.

[0007] Artificial intelligence software is also known that can simulate the results of cosmetic surgery on a person based on a photograph of themselves. However, the results obtained are unreliable because they are not based on a faithful representation of the morphology of the person operated on. Description of the invention

[0008] One object of the invention is to predict the effects of an intracorporeal surgical intervention on the external appearance of the body segment.

[0009] According to a first aspect, a method for predicting the effects of an intracorporeal surgical intervention applied to a body segment comprising is proposed at least one organ on the external appearance of the body segment, the process comprising the implementation by data processing means of a server of the following steps:

[0010] - step (a) of obtaining a biomechanical model of the body segment;

[0011] - step (b) of applying at least one principal deformation to at least one part of interest of a part of the model relating to the organ, said principal deformation being representative of a stage of the intracorporeal surgical intervention relating to the part of interest;

[0012] - step (c) of applying an incident deformation to the model, except for the part of interest, the said incident deformation being a function of the principal deformation.

[0013] According to advantageous and non-limiting features, taken alone or in any combination:

[0014] - at least one principal deformation includes a principal deformation called rigid characterized by a matrix representing a translation and / or a rotation intended to be applied to the part of interest, step (b) comprising the application of the matrix to the part of interest;

[0015] - the incident deformation corresponding to said rigid principal deformation is characterized by said matrix and step (c) includes the application of the matrix to the model, except for the part of interest;

[0016] - at least one principal deformation includes a principal deformation called non-linear characterized by a principal vector field, said non-linear deformation being applied to the part of interest by implementing a step (bl) of application, at each point of the part of interest, of a principal vector of said principal vector field;

[0017] - the principal non-linear deformation is representative of an implantation of a prosthesis in the body segment and in which, to determine the principal vector field, the following steps are implemented:

[0018] - determination of the surface area of ​​the part of interest before the implantation of the prosthesis, also called initial surface,

[0019] - determination of a surface corresponding to the surface of the part of interest after the implantation of the prosthesis, known as the final surface,

[0020] - matching each point of the initial surface with points of the final surface, the difference in position between each point of the final surface and each corresponding point of the initial surface corresponding to the principal vector applied to each point of the initial surface;

[0021] - the incident deformation corresponding to said non-linear principal deformation is applied by implementing, for each point of the model except the part of interest, called the incident point, the following steps:

[0022] - step (cl) of calculating the orthogonal projection of the point of incidence on the part of interest and determination of the point of the part of interest closest to the orthogonal projection;

[0023] - step (c2) of applying to the point of incident the principal vector applied to the point of the part of interest determined at step (cl);

[0024] - the incident deformation is a function of an influence field associated with the part of interest, the field of influence expressing a level of influence of the main deformation at each point of the model;

[0025] - for each point of a set of points of the biomechanical model, the field influence includes an influence coefficient assigned to said point and in which the incident deformation is calculated by multiplying one or more characteristic vectors of the corresponding principal deformation by the influence coefficient at that point of the influence field;

[0026] - the process comprises:

[0027] - the implementation of step (b) for at least two parts of interest so as to apply, to each part of interest, at least one elementary principal deformation;

[0028] step (c) comprising the calculation of the incident deformation by summing elementary incident deformations, each being a function of an elementary principal deformation, each elementary incident deformation being calculated from the influence field associated with the corresponding part of interest;

[0029] - the method includes a step (c3) of calculating, for each part of interest, the field of influence of the interested party, and normalization of fields of influence;

[0030] - the method includes a step (e) of generating a rendering on an interface of a client equipment connected to said server via a network, said rendering illustrating the model to which the incidental deformation(s) was applied;

[0031] - the body segment includes at least one organ which is a bone;

[0032] - step (a) comprises the following steps:

[0033] - step (al) of obtaining, from a medical imaging device (2), at least one three-dimensional medical image of the body segment;

[0034] - step (a2) of constructing the biomechanical model of the body segment from said three-dimensional medical image of the body segment;

[0035] - the method includes a step (d) of applying a texture to a part of the model relating to a surface of the body segment, so as to represent for example the appearance of a scar or edema.

[0036] According to a second aspect, a server is proposed for predicting the effects of an intracorporeal surgical intervention applied to a body segment comprising at least one organ on the external appearance of the body segment, characterized in that it comprises data processing means configured to:

[0037] - obtain a biomechanical model of the body segment;

[0038] - apply at least one principal deformation to at least one part of interest of a part of the model relating to the organ, said main deformation being representative of a stage of the intracorporeal surgical intervention relating to the part of interest;

[0039] - apply an incidental deformation to the model, except for the part of interest, said incident deformation being a function of the principal deformation.

[0040] According to a third aspect, a system is proposed comprising a server as described above and a medical imaging device, the data processing means being further configured to:

[0041] - to obtain from said medical imaging device at least one medical image three-dimensional of the body segment;

[0042] - construct the biomechanical model of the body segment from said image three-dimensional medical body segment.

[0043] According to a fourth aspect, a computer program product is proposed comprising code instructions for the execution of a method as previously described for predicting the effects of an intracorporeal surgical intervention applied to a body segment comprising at least one organ on the external appearance of the body segment, when said program is executed on a computer.

[0044] According to a fifth aspect, a computer-readable storage means is proposed on which is stored a computer program product comprising code instructions for executing a method as previously described for predicting the effects of an intracorporeal surgical intervention applied to a body segment comprising at least one organ on the external appearance of the body segment, when said program is executed on a computer. DESCRIPTION OF FIGURES

[0045] Other features and advantages of the present invention will become apparent from the following description of a preferred embodiment. This description will be given with reference to the accompanying figures, including:

[0046] - Fig. 1 represents a system for predicting the effects of an intervention intracorporeal surgical procedure applied to a body segment;

[0047] - Figure 2 represents the steps in the process of predicting the effects of a intracorporeal surgical intervention on the external appearance of the body segment;

[0048] - Fig. 3 schematically illustrates the application to a body segment of a non- linear representation of the cross-section of an organ;

[0049] - Figure 4 illustrates part of a biomechanical model of an organ that has undergone cutout and an enlarged view of this part of the model;

[0050] - [Fig. 5] comprises figures 5(i), 5(ii) which illustrate respectively before and after surgical intervention a biomechanical model of a body segment with visible organs and figures 5(iii), 5(iv) which illustrate respectively before and after surgical intervention the biomechanical model without visible organs;

[0051] - Figure 6 illustrates a biomechanical model of a body segment comprising a scar. DETAILED DESCRIPTION OF THE INVENTION System

[0052] With reference to [Fig. 1], a system is proposed for predicting the effects of an intracorporeal surgical intervention applied to a body segment which includes at least one organ.

[0053] The system 100 comprises a server 1. The server 1 comprises data processing means 11 (typically a processor). The server 1 advantageously comprises data storage means 12 (memory, for example a hard drive). The server 1 advantageously comprises an interface 13 (for example a screen, a keyboard, an input port, etc.).

[0054] The data processing means 11 of server 1 are configured to:

[0055] - obtain a biomechanical model of the body segment;

[0056] - apply at least one principal deformation to at least one part of interest of a part of the model relating to the organ, said main deformation being representative of a stage of the intracorporeal surgical intervention relating to the part of interest;

[0057] - apply an incidental deformation to the model, except for the part of interest, said incident deformation being a function of the principal deformation.

[0058] The system 100 advantageously includes a medical imaging device 2 for acquiring medical images of the body segment.

[0059] In this advantageous case, the data processing means 11 of server 1 are further configured to:

[0060] - to obtain at least one medical image from the medical imaging device 2 three-dimensional of the body segment;

[0061] - construct the biomechanical model of the body segment from the image three-dimensional medical body segment.

[0062] The medical imaging device 2 can be directly or indirectly (for example via a network such as the internet) connected to the server 1 so that the latter is capable of receiving the medical images. The interface 13 of the server 1 can also serve as an interface for the medical imaging device 2 (to control it and obtain the acquired medical images).

[0063] These medical images are typically volumetric 3D images, possibly reconstructed from 2D slices, i.e., tomographies or "CT scans" (device 2 is typically an X-ray CT scanner). It should be noted that one is not limited to a particular technology and device 2 could be an MRI, an ultrasound scanner, a PET scanner, etc.

[0064] According to one embodiment, the system 100 comprises a client device 3, for example a laptop computer, including an interface 31. The client device 3 is connected to the server 1 via a network 300, which may be the internet. It will be seen that a rendering of a biomechanical model can be displayed by the interface 31 of the client device 31. It should be noted, however, that the client device 3 may be the server 1, and therefore that the interface 31 corresponds to the interface 13 of the server 1. In other words, the rendering of the biomechanical model can be displayed by the interface 13 of the server 1. Method

[0065] With reference to [Fig.2], a method is proposed for predicting the effects of an intracorporeal surgical intervention applied to a body segment on the external appearance of said body segment.

[0066] A body segment is the body or a limb / part of the body of a human or an animal. A body segment could, for example, be the leg of a human.

[0067] The body segment comprises at least one organ. By organ is meant a group of tissues working together to perform the same physiological function, i.e., a delimited part of said body segment that fulfills a particular function. The concept of organ will be interpreted broadly. The organ may be a bone, cartilage, a ligament, a muscle, a viscus such as a heart, a lung, etc. Preferably, said organ is a rigid organ (i.e., bone and / or cartilage), including, where appropriate, some soft structures considered to be integrally attached (i.e., anchored), such as blood vessels inside or on the surface.

[0068] The body segment may include a plurality of organs.

[0069] In the context of the present invention, the body segment will potentially undergo, in the future, an intracorporeal surgical procedure impacting said organ, such as the implantation of a prosthesis. It is desirable to be able to predict the external appearance of the body segment after the operation.

[0070] The surgical intervention is intracorporeal, that is to say, it is performed inside the body segment. By "inside," it is understood to mean inside the body envelope formed by the skin.

[0071] The surgical intervention may be for therapeutic and / or aesthetic purposes.

[0072] The surgical procedure may, for example, include: - the implantation of a prosthesis, - implantation of surgical hardware such as a plate (made of titanium, for example) or a screw, - a displacement (translation and / or rotation) of an organ (such as tibial osteotomy), - organ replacement (transplantation), - a partial or complete dissection of an organ, possibly involving the placement of filling material between the dissected parts of an organ to fill the void created by the dissection; it should also be noted that, following such a dissection, scar tissue can naturally fill the void as healing progresses, - the removal of a pathological part of the organ, such as a tumor.

[0073] The surgical intervention is not limited to the examples cited above.

[0074] The surgical intervention can cause effects on the external appearance of the body segment, for example, on its shape. It is these effects that we wish to predict.

[0075] The process includes a step (a), implemented by the data processing means 11 of server 1, of obtaining a biomechanical model of the body segment.

[0076] By "biomechanical model of the body segment," which may also be called a "digital twin," we mean a multidimensional object (in particular two-dimensional or three-dimensional, and preferably three-dimensional) representative of the body segment. The model optionally includes deformable parts (corresponding in particular to fat, muscles, or ligaments, which can be represented by volumetric meshes (for example, finite element discretizations)) and non-deformable parts (corresponding in particular to bones), which can be represented by 3D surface meshes. It should be noted that "non-deformable" is understood to mean in the absence of surgical intervention. Indeed, a bone can be deformed during a surgical procedure. In other words, the deformable parts can be said to be soft, as opposed to the non-deformable parts, which can be said to be hard.

[0077] The biomechanical model advantageously represents the body segment as a set of points. The biomechanical model may also include one or more meshes identifying surfaces, for example, the surface of the body segment (i.e., the skin) and / or the surfaces of the organs of the body segment. Figure 4 illustrates a mesh of the surface of an organ (a bone), and Figures 5(i) and 5(ii) illustrate a model M comprising a mesh of the surface of the body segment (the skin).

[0078] The model can be directly supplied to the data processing means 11. Alternatively, it is generated, preferably by the data processing means 11.

[0079] Advantageously, step (a) then includes a step (al) of obtaining at least one three-dimensional medical image of the body segment.

[0080] Step (al) may include the acquisition, by the medical imaging device 2, of the three-dimensional medical image of the body segment (and in particular a set of images, preferably representing the body segment in a plurality of postures or according to a plurality of viewpoints).

[0081] Advantageously, step (a) includes a step (a2) of constructing the biomechanical model from the three-dimensional medical image of the body segment.

[0082] The biomechanical model advantageously comprises a set of points.

[0083] According to a preferred embodiment, step (a2) includes the generation of one or more meshes representing one or more parts of the body segment.

[0084] Advantageously, each element of the body segment, i.e. each organ, muscle, fatty part, is modeled by a mesh. The mesh can be surface and / or volumetric.

[0085] A mesh consists of vertices forming polygonal facets, for example triangular, or polyhedral elements, for example cubes.

[0086] The mesh can identify only the surfaces of the body segment elements. For example, the identification of bone surfaces (including articular surfaces, i.e., the parts of bone surfaces covered by cartilage that come into contact with another bone to form a joint) can be performed in the three-dimensional medical image of the body segment. Having these articular and bone surfaces available is particularly useful, for example, in the case of a surgical procedure such as prosthesis implantation, since it is at these points that a prosthesis can be placed.

[0087] Advantageously, the mesh models the entire volume of the body segment elements (i.e. the surfaces and the interior of the body segment element, possibly of the entire body segment).

[0088] To generate a mesh from a three-dimensional medical image, a segmentation process can be implemented to generate the mesh. Preferably, the three-dimensional medical image is a binarized image, and this image therefore allows the mesh to be generated. A mesh normalization phase is advantageously implemented to eliminate topological artifacts (holes or inconsistencies related to the connections between the mesh vertices) or geometric artifacts (inconsistencies in the orientation of normals, facet inversions, or mesh edge crossings).

[0089] Thus, advantageously, certain points of the biomechanical model are vertices of the mesh. In other words, the vertices are specific points of the biomechanical model. We will see later that a deformation can be determined at all points of the biomechanical model only from the deformation applied to vertices.

[0090] The method includes a step (b), implemented by the data processing means 11 of server 1, of applying at least one principal deformation to at least one part of interest of a portion of the model relating to the organ. The principal deformation is representative of a step of the intracorporeal surgical intervention relating to the part of interest.

[0091] Step (b) aims to simulate the surgical procedure. The surgical procedure may comprise one or more steps, and each step causes a principal deformation of an organ of the body segment. In step (b), a principal deformation is applied to a part of the model relating to the organ to simulate the application of a step of a surgical procedure to a part of the organ of the body segment.

[0092] By "part of the model relating to the organ", we mean the part of the biomechanical model of the body segment that represents the organ. This part may be referred to as the "virtual organ" in the remainder of the description.

[0093] By "part of interest of a part of the model relating to the organ," we mean a part of the virtual organ to which a principal deformation is applied. The part of interest is representative of a part of the organ to which at least one step of the surgical intervention would be applied during the surgical procedure; i.e., it is a sub-part of the portion of the biomechanical model of the body segment that represents the organ. The part of interest may be the entire virtual organ. In other words, the part of interest may be representative of the entire virtual organ.

[0094] Applying a principal deformation to a part of a part of the model relating to the organ means that the principal deformation is applied to each point, and possibly vertex, of the part of interest of a part of the model relating to the organ.

[0095] Once the main deformation has been applied, in step (c), an incident deformation, which is a function of the main deformation, is applied to the model, excluding the part of interest. In step (c), the effects of the main deformation (which was applied to the part of interest) on the rest of the body segment model are determined. By "the rest of the body segment model," we mean the part of the model that is not the part of interest. In other words, "the rest of the body segment model" is, in the model, everything that is not the part of interest. The effect of the main deformation is called the incident deformation. A main deformation causes an incident deformation in the rest of the model. Indeed, surgical intervention on the organ part, in addition to deforming said organ part, has a remote effect on the rest of the organ and the rest of the body segment. The primary deformation simulates the deformation caused by the intervention on the part of the organ targeted by the intervention. The incidental deformation simulates the effect caused on the rest of the organ and body segment by the intervention on the part of the organ targeted by the intervention.

[0096] Advantageously, at each point of the biomechanical model, each incident deformation applied to that point depends on an influence coefficient. Indeed, when the part of interest undergoes a principal deformation, said deformation spreads, in the form of an incident deformation, throughout the rest of the model, excluding the part of interest. Generally, the diffusion of the principal deformation implies that the further one moves away from the part of interest, the smaller the incident deformation resulting from the principal deformation. To model this attenuation phenomenon, an influence field is preferably used, which represents the level of influence (i.e., the magnitude, the amplitude of the influence), at each point of the biomechanical model, of a principal deformation applied at a point of the model.

[0097] Each point of the biomechanical model is advantageously associated with an influence field, said influence field indicating the influence that a principal deformation applied at that point (called the "influencing point" for the purposes of this explanation) has on the other points (called the "influenced points") of the model. More generally, it is the part of interest that is associated with an influence field that indicates the influence that a principal deformation applied to the part of interest has on the rest of the model. The influence field is thus preferably a scalar field associating an influence coefficient with each influenced point in space (i.e., of the model). These coefficients are advantageously a rate between 0 and 1 or between 0 and 100 to express a percentage, expressing, for each influenced point, a level of diffusion, i.e., attenuation, at the influenced point of the principal deformation applied to the influencing point.

[0098] The influence field associated with a portion of interest comprises influence coefficients equal to a percentage of 100 (or a rate of 1) at the points of said portion of interest. Conversely, the influence field associated with one portion of interest comprises influence coefficients equal to a percentage of 0 (or a rate of 0) at the points of other portions of interest.

[0099] The closer the influenced point is to the influencing point, the higher the coefficient associated with an influenced point. Consequently, the greater the incident deformation at this influenced point, which results from the principal deformation applied to the influencing point, and the closer it will be to the principal deformation applied to the influencing point.

[0100] Conversely, the further the influenced point is from the influencing point, the lower the coefficient associated with an influenced point. Consequently, the greater the incident deformation at this The influenced point, which results from the main deformation applied to the influencing point, will be small and will approach zero.

[0101] There is, however, an exception to the two preceding paragraphs. Indeed, certain influenced points may be associated with a low influence coefficient, or even equal to 0, regardless of their distance from the part of interest. In fact, some points are called "anchor points." These points are points to which zero deformation is applied. This may, for example, correspond to a part of the body segment that is held in place by the surgeon while another part of the body segment is moved.

[0102] A person skilled in the art knows how to establish such a field of influence. For example, to determine the influence coefficients, one method is to apply a harmonic function. The harmonic function can, for example, be expressed as follows: . f R3 —» [0,1] [p = (x, y, z) u(p) u(x, y, z) d 2 ud 2 ud 2 u dx 2 dy 2 * dz 2 uIqj = l,ie,Vp € Oi,w(p) = 1 u|q(} = 0,iæ->Vp e QOî ^(p) = 0

[0103] The harmonic function 11 associates a continuous value between 0 and 1 to every point p in the space comprising the biomechanical model, with coordinates y. It is a solution to the Laplace equation concerning its second partial derivatives. The boundary conditions that uniquely define such a function are two in number: the function u must take the value 1 on the domain (i.e., the set of points) where the principal deformation is applied (domain denoted Ûj), and it must take the value 0 on a domain where the surgical intervention has no effect (domain denoted Qq).

[0104] A domain where the surgical intervention has no effect may be a domain very far removed from the domain where the main deformation is applied. A domain where the surgical intervention has no effect may also be a domain containing anchor points.

[0105] In summary, the field of influence makes it possible to simulate the attenuation of the effect, on the body segment, of a principal deformation applied to the part of interest. The The influence field is used to calculate incident deformations, as will be detailed later.

[0106] It will also be seen later that, in the case where different parts of interest are deformed, it is necessary to determine the influence of each elementary deformation.

[0107] There are several types of principal deformations. In particular, a distinction is made between so-called rigid deformations and so-called nonlinear deformations. The way in which a principal deformation is applied to a part of interest varies depending on the type of principal deformation. It follows that the way in which an incident deformation is applied varies depending on the type of principal deformation to which the incident deformation corresponds.

[0108] It should be noted that several different principal deformations can be applied to the same part of interest, for example, several rigid principal deformations and / or several nonlinear principal deformations. Consequently, several principal deformations can be applied to the same point in the model.

[0109] The case of the principal deformation, called rigid, and the incident deformation that results from it (called rigid incident deformation) will first be detailed, then the case of the principal deformation, called non-linear, and the incident deformation (called non-linear incident deformation) that results from it will be detailed.

[0110] The techniques for determining an incident deformation are extrapolation techniques which will be detailed later.

[0111] A rigid deformation represents a translation and / or a rotation of the part of interest. The rigid deformation can therefore represent a translation, a rotation, or a combination of a translation and a rotation. It should be noted that, in this case, the part of interest is a portion of the virtual organ that is distinct from other parts of the virtual organ; that is, it is completely detached and has no connection with the other part(s) of the virtual organ, for example, due to a partitioning of the virtual organ. The part of interest can also be the entire virtual organ.

[0112] Advantageously, the rigid deformation is characterized by a matrix representing a translation and / or a rotation intended to be applied to the part of interest.

[0113] Preferably, the matrix is ​​a 4x4 matrix in homogeneous coordinates. An example of such a matrix is ​​shown below: Rxx Ryx Rzx T -1 X Rxy Ryy Rzy Ty Rxz Ryz Rzz Tz 0 0 0 1

[0114] In the matrix defined above, the R components are the components of a spatial rotation. The first three columns of the matrix thus form a 3x3 spatial rotation matrix. The T components are the components of a spatial translation. The last column of the matrix thus forms a spatial translation vector.

[0115] Advantageously, step b) includes applying the matrix to the part of interest. In other words, the matrix is ​​applied to each point, possibly to each vertex, of the part of interest.

[0116] Advantageously, the rigid principal deformation is applied to all points of the region of interest by interpolation. More precisely, firstly, the matrix is ​​applied to each vertex of the mesh of the region of interest. Then, by interpolation, the deformation resulting from the application of the matrix to each vertex is determined at each point that is not a vertex. To do this, for each point of the region of interest, an interpolation, advantageously linear, is performed of the matrix applied to each vertex of the facet of which the point is a part. In the case of a rigid principal deformation, this amounts to applying the matrix to each point of the region of interest.

[0117] Applying the rigid principal deformation to the part of interest is representative of implementing a translation and / or rotation of the organ concerned during the surgical procedure. In other words, applying the rigid principal deformation to the part of interest of the model portion relating to the organ is equivalent to simulating a step of the surgical procedure applied to the organ, said step consisting of a translation and / or rotation.

[0118] Once the main rigid deformation has been applied (step (b)) we seek to apply to the rest of the model (i.e. to the model except for the part of interest) the incident deformation which is caused by this main deformation (step (c)).

[0119] The rigid incident deformation is advantageously characterized by the same matrix as that characterizing the corresponding rigid principal deformation, advantageously affected by an influence coefficient of the influence field. In other words, the rigid incident deformation is, for each point of the remainder of the model, the matrix characterizing the corresponding rigid principal deformation affected by the influence coefficient of said deformation on that point. The influence coefficient is part of the influence field associated with the points of the region of interest, i.e., the influence field characterizing the influence that the principal deformation, applied to the points of the region of interest, has on the points of the remainder of the model.

[0120] By "affected," it can be understood that the matrix is ​​multiplied by the influence coefficient. More generally, by "affected," it is understood that the matrix is ​​modified by the influence coefficient.

[0121] According to an advantageous embodiment, the incident deformation represented by the Mincidente matrix is ​​calculated as follows:

[0122] Here, the incident rigid strain matrix Mincidente is obtained from the principal rigid strain matrix Mprincipaie, weighted by the influence coefficient X. In the equation above, I is the identity matrix associated with the identity transformation, which leaves the space invariant. In this equation, we observe that the principal matrix is ​​indeed multiplied by the influence coefficient. The influence coefficient is also applied to the negative identity matrix.

[0123] The incident matrix, i.e., the rigid incident strain, can therefore be obtained by multiplying the principal matrix representing the rigid principal strain by the influence coefficient. The calculation may include other steps as shown in the figure above.

[0124] Therefore, advantageously, step (c) of applying the rigid incident strain corresponding to the rigid principal strain includes applying the matrix to the model, except for the part of interest (i.e., the matrix is ​​applied to the whole model except for the part of interest since the principal strain has been applied to it), affected by an influence field.

[0125] It is therefore considered that, when the part of interest of an organ undergoes a translation and / or a rotation, the rest of the body segment incidentally undergoes the same translation and / or rotation in different proportions, i.e. in different levels / degrees.

[0126] It should be noted that the rigid incident deformation is called "rigid" only because it is the incident deformation of a rigid principal deformation. Nevertheless, the rigid incident deformation, by virtue of the use of the influence field (which is non-linear), is itself non-linear.

[0127] The case of the principal deformation, called non-linear, and the incident deformation (called non-linear incident deformation) that results from it will now be detailed.

[0128] A nonlinear deformation can represent any other stage of a surgical procedure that is not a translation and / or a rotation. A nonlinear deformation can, for example, represent the implantation of a prosthesis, the implantation of surgical material such as a plate (made of titanium, for example) or a screw, or the cutting of the part of interest. A deformation Non-linear could be used to represent a translation and / or rotation, but applying a rigid deformation to simulate a translation and / or rotation is simpler to implement.

[0129] Advantageously, the nonlinear principal deformation is characterized by a vector field, hereafter referred to as principal vectors. Step (b) of applying the nonlinear principal deformation then comprises a step (bl) of applying, at each point of the part of interest, a principal vector of said principal vector field.

[0130] Different cases of non-linear principal deformations will be detailed.

[0131] The principal non-linear deformation may be representative of the implantation in the body segment of a plate or prosthesis.

[0132] In this case, the surface of the area of ​​interest is determined before implantation; this is called the initial surface. The initial surface can be characterized by points in the model, these points corresponding to the surface of the part of the organ on which a plate will be placed or which will be replaced by a prosthesis. Advantageously, the initial surface is characterized by a mesh comprising vertices and facets.

[0133] A surface corresponding to the surface of the part of interest after implantation of the plate or prosthesis, called the final surface, is also determined.

[0134] Next, a correspondence is established between the initial surface and the final surface. In other words, the points (or vertices) of the initial surface are matched with the points (or vertices) of the final surface. The difference in position between each point of the final surface and each corresponding point of the initial surface corresponds to the displacement vector of each point of the initial surface, called the principal vector. This yields the principal vector field.

[0135] According to an advantageous embodiment, the principal vector field is obtained at all points of the region of interest by interpolation. The principal vectors are first determined for each vertex of the region of interest. Then, the principal vector for each point of the region of interest, except for the vertices, is determined by interpolating the principal vectors applied to the vertices defining the facet in which the point is located.

[0136] The matching may be non-linear, particularly in the case of a prosthesis implantation, or may be linear, particularly in the case of a plate implantation.

[0137] In the case of linear matching, the matching consists of implementing a projection of the points of the initial surface onto the final surface. This embodiment is simpler than non-linear matching.

[0138] According to another embodiment, the principal non-linear deformation can be representative of a partial cutting of a part of the organ. By partial cutting, it is understood that the part of the organ is cut so as to form two sub-parts of the organ without these sub-parts being completely separated from each other (i.e., a physical link, a hinge, remains between these sub-parts).

[0139] Partial dissection of an organ is complex to simulate because it involves the following aspects: - At least one of the sub-parts resulting from the division moves. The organ generally includes and / or is in contact with soft tissue structures (nerves, veins, arteries, bone marrow) which, when the organ is cut, are not necessarily severed, and, when the sub-parts are moved, stretch and move relative to the organ's sub-parts. For example, in the case of a bone, blood vessels surround the bone, and during partial dissection, some are not cut and stretch as the two sub-parts of the bone move apart. These soft tissue structures, although they may be located on the surface of the organ, are considered to move with it as if they were part of the organ itself. - The cutting of an organ can be accompanied by the addition of filling material and / or lead (during the healing process) to the formation of scar tissue between the two sub-parts so that they are reconnected. For example, when a bone is partially cut, the bone can rebuild itself, i.e., heal, between the two sub-parts of the bone resulting from the cutting. Alternatively, the gap between the two sub-parts can be filled with surgical material such as a prosthesis, cement, or a bone substitute.

[0140] To simulate the displacement of sub-parts of the part of the organ intended to be cut and to simulate the stretching of soft structures, a progressive principal vector field is used.

[0141] As illustrated in [Fig. 3], a transition zone is defined in the biomechanical model of the organ, the transition zone T comprising or being near the cutting line. It is assumed that the part of interest in the model will be cut so as to form a sub-part A and a sub-part B. The transition zone T may, for example, represent a certain predetermined proportion of sub-part A and / or sub-part B.

[0142] The points of sub-part A of the part of interest, or advantageously the vertices of the mesh if the component is represented by a mesh, which are not part of the transition zone T, each undergo the same displacement. Respectively, the points of Sub-part B of the part of interest, or advantageously the vertices of the mesh if the component is represented by a mesh, which are not part of the transition zone T, each undergo the same displacement (possibly distinct from the displacement of sub-part A). Thus, a principal vector, which depends on the sub-part to which the point belongs, is applied to each point outside the transition zone T, advantageously to each vertex outside the transition zone.

[0143] The points within the transition zone undergo a gradual displacement, the extent of which depends on their proximity to the sub-parts. The closer a point is to sub-part A, the closer the principal vector applied to it will be to 100% of the principal vector associated with each point in sub-part A. Conversely, the closer a point is to sub-part B, the closer the principal vector applied to it will be to 100% of the principal vector associated with each point in sub-part B.

[0144] Part C in [Fig.3] corresponds to the added material which may be scar substance that forms due to the healing of the organ, a filling material applied by a surgeon or which may be surgical material such as a prosthesis.

[0145] As illustrated in [Fig. 3], it is understood that, to simulate a section of an organ, step (bl) consists of applying a nonlinear principal vector field CV to the points P of the region of interest. The nonlinear principal deformation is characterized by this CV field.

[0146] According to an advantageous embodiment, the principal vector field is obtained at all points of the region of interest by interpolation. Initially, the vectors of the nonlinear principal vector field can be applied only to the vertices of the mesh of the region of interest. Then, by interpolation, the vectors of the nonlinear principal vector field that must be applied to the points that are not vertices can be determined.

[0147] Advantageously, step (b) further includes a step (b2) for reconstructing the region of interest representative of the organ's healing at the cut. For this purpose, as illustrated in [Fig. 4], points of the region of interest are mapped on either side of the cutting line. A mesh can be constructed based on this mapping. [Fig. 4] illustrates a region of interest 5 of a portion of a biomechanical model representative of an organ, which is a bone. The region of interest 5 has been partially cut. [Fig. 4] illustrates the space 52 left by the cut and illustrates the hinge 51. [Fig. 4] also illustrates an enlarged view of the region of interest 5 at the cut, and the lines 53 schematically represent the mapping of points of the region of interest on either side of the cutting line.

[0148] The implementation of step (c) of applying a non-linear incident deformation corresponding to a non-linear principal deformation will be detailed.

[0149] The nonlinear incident deformation is applied to each point of the model except the part of interest, that is, to all points of the model except the points of the part of interest. These points (i.e., those that are not those of the part of interest) are hereafter referred to as "incident points".

[0150] Step (c) advantageously includes, for each incident point, a step (cl) of calculating the orthogonal projection of the incident point onto the part of interest and of determining the point in the part of interest closest to the orthogonal projection. The aim is thus to associate a point in the part of interest with an incident point.

[0151] Step (c) further advantageously includes a step (c2) of applying to the incident point the principal vector applied to the point of the part of interest determined in step (c1). In other words, the same principal vector that was applied, in step (b), to the associated point of the part of interest is applied to the incident point. The vector associated with an incident point is called the "incident vector." A vector is advantageously applied after having been assigned a corresponding influence coefficient. Step (c2) advantageously includes multiplying each incident vector by the corresponding influence coefficient, as described in the preceding section concerning rigid incident deformations. The nonlinear incident deformation is characterized by the set of incident vectors that forms an incident vector field, the incident vectors preferably being assigned, i.e., multiplied, by an influence coefficient.

[0152] In summary, each incident deformation, whether rigid or non-linear, is calculated from one or more vectors.

[0153] A rigid incident deformation is calculated from a matrix, which corresponds to a set of vectors. The rigid incident deformation is calculated by assigning an influence field to the matrix. Advantageously, by "assigning," it is understood that the rigid incident deformation is calculated by multiplying the matrix, i.e., the vectors, by an influence field. In other words, preferably, for each point in a set of influenced points of the model, the rigid incident deformation at that point is calculated by multiplying the matrix by the influence coefficient at that point of the influence field. The calculation to obtain the rigid incident deformation may include other steps in addition to said multiplication, as illustrated in the equation for the incident matrix Mincidente presented previously.

[0154] A nonlinear incident deformation is calculated from an incident vector field. The nonlinear incident deformation is calculated by assigning an influence field to the incident vector field. Advantageously, by "assigning," it is understood that the incident vector field is multiplied by the influence field. In other words, for each point in a set of influenced points of the model, the nonlinear incident deformation at that point is calculated by multiplying the vector incident at this point of the incident vector field by the influence coefficient at this point of the influence field.

[0155] Therefore, for each point of a set of points of the model, an incident deformation, whether rigid or non-linear, is advantageously calculated by multiplying one or more vectors determined from the corresponding principal deformation (i.e. characteristics of the corresponding principal deformation) by the influence coefficient at that point of the influence field.

[0156] Advantageously, the surgical intervention comprises at least two steps, and steps (b) and (c) of the procedure are applied to at least two parts of interest of the model. The parts of interest may be parts of interest of the same organ or of different organs. Each part of interest may correspond to an entire organ.

[0157] In this case, to determine the incident deformation, we seek to take into account the principal deformations, called elementary principal deformations, applied to each part of interest.

[0158] Therefore, the method includes carrying out step (b) for at least two parts of interest so as to apply, to each part of interest, at least one elementary principal deformation. In other words, step (b) is carried out several times and a principal (elementary) deformation is calculated at each implementation of said step (b).

[0159] Then, advantageously, in step (c), the incident strain is calculated by summing elementary incident strains, each of which is a function of an elementary principal strain, and each elementary incident strain is determined from an influence field associated with the corresponding part of interest, called the elementary influence field, expressing a level of influence of the elementary principal strain at each point of the model.

[0160] Advantageously, step (c) includes a step (c3) for calculating the influence field of each part of interest. In other words, the set of influence coefficients for each elementary influence field is calculated. Each elementary influence field characterizes the influence of an elementary principal deformation, applied to a part of interest, on the rest of the biomechanical model. Each elementary influence field is associated with a part of interest, i.e., with points of a part of interest, which has therefore undergone an elementary principal deformation. As explained previously, each elementary influence field is preferably a scalar field, with each elementary influence coefficient being a scalar associated with a point of the model.

[0161] The set of elementary influence coefficients obtained for the model and associated with a part of interest forms an influence field representative of the influence of the part of interest on the entire model. For the model, there are As many elementary influence fields are obtained as there are parts of interest. Similarly, we understand that we have as many elementary influence coefficients at each point of the model as there are parts of interest.

[0162] The elementary influence fields allow the elementary incident deformations to be determined. Consequently, the finally calculated incident deformation depends on the elementary incident deformations, which depend in part on the influence coefficients.

[0163] According to one embodiment, to calculate the elementary influence fields, in a first step, each part of interest is considered in isolation and the procedure is the same as for a single part of interest. In particular, the harmonic function described above can be applied so as to obtain the elementary influence coefficients, i.e. advantageously scalars between 0 and 1 (or between 0% and 100%).

[0164] Then, in a second step, normalization is implemented. More precisely, at each point of the model, each elementary influence coefficient is divided by the sum of all the elementary influence coefficients at that point so as to obtain a new scalar, always between 0 and 1 (or between 0% and 100%), for each part of interest. Consequently, each point of the model is associated with several normalized influence coefficients, each coefficient being relative to the influence of a part of interest on the point, each coefficient being between 0 and 1 (or between 0% and 100%), the sum of the coefficients at that point being equal to 1 (or 100%). In other words, the elementary influence fields are normalized.

[0165] It is understood that, when elementary principal deformations are applied to different parts of interest, to calculate the respective elementary incident deformations, elementary influence coefficients which have been normalized are advantageously used (i.e., at a point, the sum of the influence coefficients of the influence fields associated with different parts of interest which apply is equal to 1).

[0166] This makes it possible to calculate, at the same point, the combination of several incident deformations, possibly including both rigid incident deformations and non-linear incident deformations.

[0167] Advantageously, the method includes a step (d) of applying a texture to a portion of the model relating to a surface of the body segment (i.e., the surface being practically representative of the skin), so as to, for example, represent the appearance of a scar or edema. Figures 5(iii), 5(iv), and 6 illustrate a model with a texture enabling the representation of skin.

[0168] The term "application of a texture" means, for example, applying different colors to simulate a scar and / or edema. The application of a Texture can also include geometric modifications of the model's surface (which may be a mesh) by modifying points / vertices of the model to represent swelling characteristic of edema or a blister characteristic of a scar. Figure 6 illustrates a model M of a body segment including a scar Ci.

[0169] The scar trajectory can be obtained by identifying the start and end points of the incision, as well as the points where the incision line passes over the patient's skin. The complete trajectory is then calculated on the mesh as the discrete geodesic connecting these points. The deformation produced on the skin by the scar can then be simulated using the technique described above, by applying, as the main deformation function, a skin elevation field corresponding to the typical thickness of a scar for the surgical procedure under consideration. The effect on the tissues adjacent to the scar profile is then calculated using the extrapolation and interpolation techniques described above.

[0170] Another example of an application concerning the physical appearance of the operated body segment after the procedure is the simulation of postoperative edema. The tissue region concerned is defined for this purpose on a model of the operated body segment. The edema can then be modeled as a swelling (primary deformation) that propagates to the surrounding tissues (incident deformation).

[0171] Advantageously, the method includes a step (e) of generating a rendering on the interface 31 of the client equipment 3 connected to the server 1 via a network 300. The rendering illustrates the model M to which the incident deformation(s) have been applied. An example of a rendering is shown in [Fig. 5](iv). The rendering allows a patient to visualize the external appearance of their body segment after the hypothetical intracorporeal surgical intervention. [Fig. 5] illustrates a before / after of a body segment. [Fig. 5](iii) illustrates a model M of a body segment before surgery. [Fig. 5](i) allows visualization of the organs (bones) of model M. [Fig. 5](iv) illustrates model M of the body segment after surgery. [Fig. 5](ii) allows visualization of the organs (bones) of model M after surgery: it can be seen that one organ (bone) has been cut.

[0172] The rendering may include several images or may be a CAD (Computer-Aided Design) file that the patient can manipulate to view their body segment from multiple viewpoints. The rendering could also be a physical model of their body segment.

[0173] Product computer program and readable storage means

[0174] Also proposed is a computer program product comprising code instructions for the execution (on the data processing means 11 of server 1) of a method for predicting the effects of a surgical intervention intracorporeal applied to a body segment comprising at least one organ on the external appearance of the body segment, as well as computer-readable storage means (for example, data storage means 12 of server 1) on which this computer program product is found.

Claims

Demands

1. A method for predicting the effects of an intracorporeal surgical intervention applied to a body segment comprising at least one organ on the external appearance of the body segment, the method comprising implementing by data processing means of a server the following steps: - step (a) of obtaining a biomechanical model of the body segment; - step (b) of applying at least one principal deformation to at least one part of interest of a part of the model relating to the organ, said principal deformation being representative of a step of the intracorporeal surgical intervention relating to the part of interest; - step (c) of applying an incidental deformation to the model, except the part of interest, said incidental deformation being a function of the principal deformation.

2. A method according to claim 1, wherein at least one principal deformation comprises a principal deformation called rigid characterized by a matrix representing a translation and / or a rotation intended to be applied to the part of interest, step (b) comprising the application of the matrix to the part of interest.

3. A method according to any one of claims 1 and 2, wherein the incident strain corresponding to said rigid principal strain is characterized by said matrix and step (c) comprises applying the matrix to the model, except for the part of interest.

4. A method according to any one of claims 1 to 3, wherein at least one principal deformation comprises a principal deformation said to be nonlinear characterized by a principal vector field, said nonlinear deformation being applied to the part of interest by implementing a step (bl) of applying, at each point of the part of interest, a principal vector of said principal vector field.

5. A method according to claim 4, wherein the nonlinear principal deformation is representative of the implantation of a prosthesis in the body segment and wherein, to determine the principal vector field, the following steps are implemented: - determination of a surface of the part of interest before the implantation of the prosthesis, called the initial surface, - determination of a surface corresponding to the surface of the part of interest after the implantation of the prosthesis, called the final surface, - matching each point of the initial surface with points of the final surface, the difference in position between each point of the final surface and each corresponding point of the initial surface corresponding to the principal vector applied to each point of the initial surface.

6. A method according to any one of claims 4 and 5, wherein the incident strain corresponding to said non-linear principal strain is applied by implementing, for each point of the model except the part of interest, said incident point, the following steps: - step (c1) of calculating the orthogonal projection of the incident point onto the part of interest and determining the point of the part of interest closest to the orthogonal projection; - step (c2) of applying to the incident point the principal vector applied to the point of the part of interest determined in step (c1).

7. A method according to any one of claims 1 to 6, wherein the incident deformation is a function of an influence field associated with the part of interest, the influence field expressing a level of influence of the main deformation at each point of the model.

8. A method according to claim 7, wherein, for each point of a set of points of the biomechanical model, the influence field includes an influence coefficient assigned to said point and wherein the incident strain is calculated by multiplying one or more characteristic vectors of the corresponding principal strain by the influence coefficient at that point of the influence field.

9. A method according to any one of claims 7 and 8, comprising: - carrying out step (b) for at least two parts of interest so as to apply, to each part of interest, at least one elementary principal strain; step (c) comprising calculating the incident strain by summing elementary incident strains, each being a function of an elementary principal strain, each elementary incident deformation being calculated from the influence field associated with the corresponding part of interest.

10. Method according to claim 9, comprising a step (c3) of calculating, for each part of interest, the field of influence of the part of interest, and of normalizing the fields of influence.

11. A method according to any one of claims 1 to 10, comprising a step (e) of generating a rendering on an interface of a client device connected to said server by a network, said rendering illustrating the model to which the incidental deformation(s) has been applied.

12. A method according to any one of claims 1 to 11, wherein the body segment comprises at least one organ that is a

13. Uj. A method according to any one of claims 1 to 12, wherein step (a) comprises the following steps: - step (a1) of obtaining, from a medical imaging device (2), at least one three-dimensional medical image of the body segment; - step (a2) of constructing the biomechanical model of the body segment from said three-dimensional medical image of the body segment.

14. A method according to any one of claims 1 to 13, comprising a step (d) of applying a texture to a part of the model relating to a surface of the body segment, so as to, for example, represent the appearance of a scar or edema.

15. Server (1) for predicting the effects of an intracorporeal surgical intervention applied to a body segment comprising at least one organ on the external appearance of the body segment, characterized in that it comprises data processing means (11) configured to: - obtain a biomechanical model of the body segment; - apply at least one principal deformation to at least one part of interest of a part of the model relating to the organ, said principal deformation being representative of a stage of the intracorporeal surgical intervention relating to the part of interest; - apply an incident deformation to the model, except for the part of interest, said incident deformation being a function of the principal deformation.

16. System comprising a server (1) according to claim 15 and a medical imaging device (2), the data processing means (11) being further configured to: - obtain from said medical imaging device (2) at least one three-dimensional medical image of the body segment; - construct the biomechanical model of the body segment from said three-dimensional medical image of the body segment.

17. Product computer program comprising code instructions for performing a method according to any one of claims 1 to 14 of predicting the effects of an intracorporeal surgical intervention applied to a body segment comprising at least one organ on the external appearance of the body segment, when said program is executed on a computer.

18. Computer-readable storage means on which is stored a computer program product comprising code instructions for performing a method according to any one of claims 1 to 14 of predicting the effects of an intracorporeal surgical intervention applied to a body segment comprising at least one organ on the external appearance of the body segment, when said program is executed on a computer.

Images