Method and system for reconstructing three-dimensional medical images - Patents.com

Abstract

The method comprises a method for reconstructing a three-dimensional medical image of a region of interest from a set of two-dimensional X-ray images acquired by an X-ray imaging system including an X-ray source (S) and an image detector (D), the method comprising the steps of: The method includes computing n passes (n being an integer equal to or greater than 1) of the X-ray imaging system, the passes adapted to obtain a subset of n respective two-dimensional X-ray images that together form the set of two-dimensional X-ray images, and each image of the subset (I A , I B , I C ) is a projection cone (C) of each of the regions of interest onto the image detector such that, for the same position of the X-ray source (S) along the n paths, the projection cone of an image of any subset is continuous with the projection cone of an image of at least one other subset. A , C B , C C ), and the step of calculating the n paths includes: positioning a virtual isocenter (O) of the X-ray imaging system at a given position or a given series of positions relative to a center (C) of the region of interest; and defining the n projection cones such that for each position of the X-ray source along the n paths, the center of a height of an expansion cone formed by a combination of the n projection cones is the virtual isocenter (O); The method comprises: performing the calculated n passes to obtain n subsets of the image; reconstructing the three-dimensional medical image by intersecting the projection cones defined by each image of the n subsets.

Description

[Technical field]

[0001] The present disclosure relates to methods and systems for reconstructing three-dimensional medical images. [Background technology]

[0002] X-ray imaging systems are frequently used during medical surgical procedures to provide physicians with image-based information regarding the patient's anatomy and / or the position and orientation of surgical instruments.

[0003] Such x-ray imaging systems typically provide two-dimensional (2D) projection images with superimposition of different anatomical structures along the x-ray path.

[0004] A typical example of such a system used during surgery is the so-called C-arm, which consists of a platform on which a C-arm is mounted, with multiple intermediate joints that allow the C-arm to move in multiple degrees of freedom in space. The C-arm is provided with an X-ray source at one end and an image detector at the other end.

[0005] Because such two-dimensional images can provide limited information, the need for three-dimensional (3D) imaging techniques has emerged over the past few decades.

[0006] Computed tomography is an established type of stationary X-ray imaging system used for 3D reconstruction in radiology departments, but these systems are generally not available in the operating room.

[0007] In recent years, there has been growing interest in cross-sectional reconstruction techniques using two-dimensional detectors, also known as cone-beam computed tomography (CBCT).

[0008] Special efforts have been made to enable the above-mentioned C-arm to provide three-dimensional information by automatically acquiring a set of two-dimensional images and then reconstructing a three-dimensional image based on the cone-beam reconstruction technique.

[0009] A fundamental problem with this imaging modality is the limited size of the reconstructed 3D image. This reconstruction size is limited by the following factors: · Surface of the C-arm detection panel Characteristic dimensions of the C-arm (including in particular the distance between the X-ray source and the detector panel) -Movement range according to the degree of freedom of the C-arm The C-arm and / or its components (especially the detection panel) have a low degree of freedom of movement, which limits the complexity of the C-arm path during 2D image acquisition.

[0010] FIG. 1 is a schematic diagram of a trajectory path of a C-arm as typically performed in the prior art. Several positions of the X-ray source S and the detector D along half of the trajectory path (i.e. along an angular sector of 90°) are shown, with the dotted line connecting one position of the X-ray source S with the corresponding position of the detector D representing the rotation axis of the X-ray radiation cone of the X-ray source during the acquisition of a two-dimensional X-ray image. The patient P is lying on a table T. The path of the C-arm is circular, meaning that the C-arm rotates around an axis perpendicular to the plane of the C-arm (corresponding to the paper plane of FIG. 1) over an angular sector of 180°. In this plane, the center of rotation of the C-arm is the isocenter I of the C-arm, which is the center of the line segment connecting the X-ray source S to the center of the detector panel D. To reconstruct a three-dimensional image of the region of interest ROI of the patient, several two-dimensional X-ray images of the region of interest are acquired during the trajectory path of the C-arm.

[0011] However, in minimally invasive surgery, it is often desirable to make the reconstructed 3D image as large as possible, not only for the surgeon's comfort but also to obtain better output in certain procedures, e.g., interventions on multiple bones (e.g. multiple vertebrae) or procedures on obese patients. Summary of the Invention

[0012] It is therefore desirable to optimize the path of an X-ray imaging system during acquisition of a set of 2-dimensional images in order to generate a reconstructed 3-dimensional image of as large as possible of a patient's region of interest (ROI) within safety parameters.

[0013] Thus, in some embodiments, the present invention provides a method for reconstructing a three-dimensional medical image of a region of interest from a set of two-dimensional X-ray images acquired by an X-ray imaging system including an X-ray source and an image detector, comprising: The method includes the step of computing n passes of the X-ray imaging system, n being an integer equal to or greater than 1, the passes being adapted to obtain n respective subsets of two-dimensional X-ray images that together form a set of two-dimensional X-ray images, each image of a subset defining a respective cone of projection of the region of interest onto the image detector such that for an identical position of the X-ray source along the n passes, a cone of projection of an image of any subset is continuous with a cone of projection of an image of at least one other subset, the step of computing the n passes comprising: positioning a virtual isocenter of the X-ray imaging system at a given position or a given series of positions relative to a center of the region of interest; defining the n projection cones such that for each position of the x-ray source along the n paths, the center of a height of an expansion cone formed by a combination of the n projection cones is the virtual isocenter; The method comprises: performing the calculated n passes to obtain n subsets of the image; reconstructing the three-dimensional medical image by intersecting the projection cones defined by each image of the n subsets.

[0014] As used herein, "contiguous" means that two adjacent projection cones corresponding to the same position of the X-ray source along their respective paths touch or overlap.

[0015] Unlike conventional reconstruction techniques, where one pass of the X-ray imaging system is defined to acquire a set of two-dimensional X-ray images containing a sufficient number of images for a complete volume reconstruction, the method according to the invention includes at least two passes of the X-ray imaging system, where each subset of two-dimensional X-ray images acquired along each pass is insufficient by itself to reconstruct a three-dimensional image, in particular, the intersection between the projection cones of the images of each subset is smaller than the region of interest, while the intersection between the projection cones of the images of all subsets is generally larger than the region of interest.

[0016] In some embodiments, n is 2.

[0017] In another embodiment, n is 3.

[0018] In some embodiments, the given location of the virtual isocenter is substantially at the center of the region of interest.

[0019] In some embodiments, the given set of positions of the virtual isocenter is selected to minimize a distance between the virtual isocenter and a center of the region of interest for each position of the X-ray source along each of the paths.

[0020] Advantageously, at least two of the n passes are performed in opposite directions of an orbital rotation of the X-ray imaging system.

[0021] In general, the intersection between the projection cones of each subset of images is smaller than the region of interest.

[0022] In some embodiments, for the same position of the x-ray source along the n paths, images of at least two of the n subsets overlap.

[0023] In some cases, a calibration phantom including a radiopaque fiducial is detectable only in one first subset of images, and the method further includes registering at least one other subset of images to the first subset of images based on overlapping portions of the images.

[0024] In some embodiments, the method further comprises selectively reducing the projection cone of each two-dimensional x-ray image using a dynamic collimator.

[0025] In a preferred embodiment, the X-ray imaging system comprises a base, a C-shaped gantry supporting the X-ray source and the image detector, and a motorized arm connecting the C-shaped gantry to the base, the motorized arm having at least three axes of rotation aligned substantially along a common vertical direction.

[0026] Another object of the invention is a medical system for implementing the above method.

[0027] This system is an x-ray imaging system including an x-ray source and an image detector; a control unit configured to calculate n passes of the X-ray imaging system, n being an integer equal to or greater than 1, wherein the passes are adapted to obtain n respective subsets of two-dimensional X-ray images that together form a set of two-dimensional X-ray images, each image of a subset defining a respective cone of projection of the region of interest onto the image detector such that for an identical position of the X-ray source along the n passes, a cone of projection of an image of any subset is continuous with a cone of projection of an image of at least one other subset, and the calculation of the n passes comprises: positioning a virtual isocenter of the X-ray imaging system at a given position or a given series of positions relative to a center of the region of interest; defining the n projection cones such that for each position of the x-ray source along the n paths, the center of a height of an expansion cone formed by a combination of the n projection cones is the virtual isocenter; The control unit controlling a movement of the X-ray imaging system to perform the calculated n passes to acquire the n subsets of images; The apparatus is further configured to reconstruct the three-dimensional medical image by intersecting the projection cones defined by each image of the n subsets.

[0028] In a preferred embodiment, the X-ray imaging system comprises a base, a C-shaped gantry supporting the X-ray source and the image detector, and a motorized arm connecting the C-shaped gantry to the base, the motorized arm having at least three axes of rotation aligned substantially along a common vertical direction.

[0029] Further features and advantages of the present invention will be described in the following description, taken in conjunction with the accompanying drawings, in which: [Brief description of the drawings]

[0030] [Figure 1] FIG. 1 is a schematic diagram of a C-arm trajectory path commonly performed in the prior art. [Diagram 2] FIG. 1 is a schematic diagram of an X-ray imaging system that can be used in the present invention. [Diagram 3] FIG. 1 is a perspective view of an X-ray imaging system that can be used in the present invention. [Figure 4] FIG. 2 shows a schematic diagram of a magnification cone based on a union of two adjacent projection cones and a virtual isocenter of an X-ray imaging system for the magnification cone. [Diagram 5] FIG. 2 shows a schematic diagram of a first and second pass of an X-ray imaging system when a virtual isocenter of the X-ray imaging system is in a fixed position according to a first embodiment of the present invention. [Figure 6A] FIG. 2 is a diagram illustrating a schematic diagram of a first path of the X-ray imaging system in the first embodiment. [Figure 6B] FIG. 2 is a diagram illustrating a schematic diagram of a second path of the X-ray imaging system in the first embodiment. [Figure 7]FIG. 13 is a schematic diagram showing the translation of the C-arm to reduce the distance between the region of interest and the detector. [Figure 8] FIG. 2 is a schematic diagram showing the trajectory of the center of rotation of the orbital motion of the C-arm gantry according to the first embodiment of the present invention. [Figure 9] FIG. 11 is a schematic diagram showing the trajectory of the center of rotation of the orbital motion of a C-arm gantry according to a second embodiment of the present invention. [Figure 10A] FIG. 11 is a schematic diagram showing the trajectory of the center of rotation of the orbital motion of a C-arm gantry according to the third embodiment of the present invention. [Figure 10B] FIG. 13 is a schematic diagram showing the trajectory of the center of rotation of the orbital motion of a C-arm gantry according to a fourth embodiment of the present invention. [Figure 11] FIG. 10 is a schematic diagram illustrating a first and second pass of an X-ray imaging system when the trajectory of the virtual isocenter of the X-ray imaging system is elliptical in shape according to a second embodiment of the present invention. [Figure 12A] FIG. 10 is a diagram illustrating a schematic diagram of a first path of an X-ray imaging system in a second embodiment. [Figure 12B] FIG. 10 is a diagram illustrating a schematic diagram of a second path of an X-ray imaging system in a second embodiment. [Figure 13] FIG. 13 shows a schematic diagram of an expansion cone based on the union of two partially overlapping projection cones. [Figure 14] FIG. 2 shows a schematic diagram of a magnification cone based on the union of two partially overlapping projection cones, where the calibration phantom is visible in only one of the corresponding 2D X-ray images. [Figure 15] FIG. 13 shows a schematic diagram of an expansion cone based on the combination of two adjacent projection cones, where dynamic collimation is used to reduce the illuminated area. [Figure 16] FIG. 13 shows a schematic diagram of an expansion cone based on the union of three adjacent projection cones. [Figure 17] FIG. 1 illustrates an embodiment of the present invention that allows for increased three-dimensional image size along the patient's major axis. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0031] Please note that for drawing legibility, the figures are not necessarily drawn to scale. In the different figures, the same reference numbers indicate identical elements or elements with the same function.

[0032] The present invention may be implemented to reconstruct three-dimensional x-ray images in connection with a surgical procedure performed on one or more patient bones, such as, but not limited to, implantation of orthopedic implants such as pedicle screws into the spine, implantation of various orthopedic implants into bones, reduction and fixation of fractures during trauma procedures, placement of a guide or cannula at a desired location relative to a predetermined target, insertion of a catheter or stent during a cardiovascular or urological procedure, etc.

[0033] In this regard, the X-ray imaging system may be coupled to other surgical systems, such as a positioning system and / or a surgical robotic system.

[0034] The region of interest corresponds to a volume of the patient's body that is imaged by the X-ray imaging technique. The region of interest may include a bone, a portion of a bone, or adjacent bones.

[0035] (X-ray imaging system) 2, the X-ray imaging system includes at least one X-ray source S and at least one X-ray image detector D. The X-ray image detector may include a flat detector panel.

[0036] The X-ray source and the X-ray image detector are placed at either end of the C-shaped gantry G and are supported by the gantry. Due to the shape of the gantry, such imaging systems are usually called C-arms. As mentioned above, the center of the line segment connecting the center of the X-ray source and the center of the detector is called the isocenter of the C-arm.

[0037] In the present invention, the center of the orbital motion of the C-arm is the isocenter, as in the case of various conventional C-arms. Therefore, in this text, the center of the orbital motion and the isocenter refer to the same point I.

[0038] In a manner known per se, the X-ray imaging system is configured to generate at least one two-dimensional X-ray image that is the result of a cone-shaped projection of the region of interest onto the image detector, the apex of this cone being approximately the centre point of the X-ray source and the base being approximately the part of the image detector that is reached by the collimated X-ray beam of a given shape and orientation.

[0039] Conventional C-arms are designed such that the gantry can move relative to the base to move the x-ray source and detector relative to the patient while acquiring projection images of the patient positioned between the gantry's x-ray source and x-ray detector.

[0040] For example, the X-ray imaging system may be a conventional C-arm or any cone beam CT (CBCT) imaging system such as the SURGIVISIO device (Essential Robotics, Gieres, France), VISION FD VARIO 3D (Ziem), CIOS SPIN MOBILE 3D (Siemens), AIRO (Stryker), LOOP-X (Brainlab), or O-ARM (Medtronic).

[0041] To reconstruct a three-dimensional image, an imaging data set consisting of multiple two-dimensional X-ray images of the region of interest is acquired by an X-ray imaging system. The three-dimensional image of the region of interest can be reconstructed from the set of two-dimensional X-ray images acquired by the X-ray imaging system using tomography techniques. The three-dimensional image corresponds to the intersection of projection cones of the region of interest onto a detector corresponding to each two-dimensional X-ray image of the imaging data set.

[0042] In a preferred embodiment, the X-ray imaging system is motorized. In particular, the C-shaped gantry is equipped with motors that allow movement horizontally (X and Y directions), vertically (Z direction), and about the X direction (defined by angle α), allowing two-dimensional X-ray images of the patient to be generated from almost any angle. As shown in Figure 2, the X-axis is transverse to the operating table T on which the patient lies, and the Y-axis is parallel to the longitudinal axis of the operating table.

[0043] Each motor is associated with an encoder, allowing the relative position of the X-ray imaging system with respect to a reference position to be obtained at any time. When a 2D X-ray image is acquired, the corresponding position of the imaging system is recorded. Thus, each 2D image is recorded with reference to the imaging system.

[0044] The path of the motorized C-arm is determined by each 2D X-ray imaging acquisition position of the C-arm while performing the 3D image acquisition.

[0045] In some embodiments, the X-ray imaging system has a moving base that allows movement of the C-arm within the operating room. The C-shaped gantry can be slidably and / or rotatably mounted on the moving base. The motor of the C-arm can be located in the moving base and / or in the gantry.

[0046] FIG. 3 shows one embodiment of such an X-ray imaging system, comprising a moving base 10 , a motorized arm 20 , and a C-shaped gantry G attached to the moving base 10 by the motorized arm 20 .

[0047] The C-shaped gantry comprises an X-ray source S and an X-ray detector D.

[0048] The motorized arm 20 has at least three axes of rotation Z1, Z2, Z3 aligned substantially along a common vertical direction Z. An example of three such axes of rotation Z1, Z2, Z3 is shown in the exemplary embodiment of FIG.

[0049] The horizontal plane H is a plane perpendicular to the vertical direction Z. The horizontal plane H is defined by a longitudinal direction Y and a transverse direction X perpendicular to the longitudinal direction Y.

[0050] A table on which a person to be imaged lies may extend substantially in a horizontal plane H. A longitudinal direction Y extends along the length of the table. A lateral direction X extends along the width of the table. The floor of the room in which the X-ray imaging takes place may extend in the horizontal plane H.

[0051] The vertical plane V corresponds to the plane defined by the vertical direction Z and the lateral direction X.

[0052] The movement of the C-arm 30 is simple as it is governed by a single motorized arm 20.

[0053] The three rotation axes Z1, Z2, Z3, substantially along a common vertical direction Z, are substantially parallel to one another and articulate different segments of the motorized arm 20, allowing the motorized arm 20 to rapidly access a wide range of positions in a wide variety of orientations. In this way, the X-ray imaging system can follow complex paths, thereby enabling imaging of a wide range of regions of interest under a wide variety of constraints.

[0054] More specifically, the motorized arm 20 can move the C-shaped gantry G by rotation or a combination of rotations of different segments of the motorized arm 20 about one or some of the three rotation axes Z1, Z2, Z3 of the motorized arm 20.

[0055] The motorized arm 20 can move, for example, translate, the C-shaped gantry G in a lateral direction X, a vertical direction Y, or any combination of directions included in the horizontal plane H that is substantially perpendicular to the vertical direction Z.

[0056] The motorized arm 20 having three substantially parallel axes of rotation Z1, Z2, Z3 allows for high precision movement of the motorized arm 20 and therefore accurate positioning and orientation of the C-arm 30.

[0057] In this way, the X-ray imaging system is able to acquire successive imaging data sets of separate regions of interest by moving the C-shaped gantry G between successive acquisitions, with the movement commands being respected, and the effective position of the C-shaped gantry G substantially corresponding to the commanded position of the C-shaped gantry G. The positions of the X-ray source S and the X-ray detector D are therefore known for each successive acquisition of imaging data sets.

[0058] Thus, when multiple successive imaging data sets need to be acquired, the relative positions of the acquired images of a first data set with respect to the acquired images of a second data set can be accurately inferred from the commanded movements of the C-shaped gantry G. Thus, there is no need to recalibrate the position of the C-shaped gantry G by specific recalibration means for the acquired images of the different data sets.

[0059] The C-shaped gantry G can be positioned successively at different regions of interest, e.g., foot, knee, hip, multiple vertebrae, etc., as described below, and / or different data sets of images of the same region of interest can be acquired. The position of the C-shaped gantry G can be inferred from the commanded movements of the motorized arm 20 at any time as the commands are respected.

[0060] Finally, the base 10 is a mobile base. Thus, the base 10 can be moved according to the region of interest to be imaged and can be moved in coordination with the movement of the motorized arm 20. The mobile base 10 can be moved between successive acquisitions of different regions of interest. For example, the mobile base 10 can be placed as close as possible to the region of interest to be acquired so as not to interfere with the physician performing the X-ray imaging. This further reduces the complexity, weight and space required by the X-ray imaging system.

[0061] This method allows flexibility in the acquisition of image datasets without significantly impacting the weight or cost of the X-ray imaging system and without compromising accuracy.

[0062] (Control Unit) The X-ray imaging system is typically controlled by a control unit that includes a processor, data storage and communication devices, in particular the control unit controls the motors of the X-ray imaging system, i.e. the motorized C-arm motor and the motorized arm motor, if present.

[0063] The control unit may advantageously be embedded in the base of the X-ray imaging system, which may also include switches such as a power switch, an emergency button, etc.

[0064] Alternatively, the control unit may be housed on a separate cart having at least one interface with the C-arm, or may be remote, for example located in a separate control room or data center in the hospital.

[0065] In a preferred embodiment, the control unit is capable of controlling other surgical systems within the operating room.

[0066] The control unit is configured to implement appropriate algorithms for carrying out the workflow described below in order to reconstruct three-dimensional medical images.

[0067] (Workflow) A set of two-dimensional x-ray images used to reconstruct a three-dimensional medical image is divided into two or more subsets of two-dimensional x-ray images.

[0068] The acquisition of each subset of the 2D x-ray images is performed along a respective path of the x-ray imaging system, the determination of the path for each subset being described below.

[0069] Each subset contains multiple 2D x-ray images, the number of which varies depending on the size and nature of the region of interest, but each subset alone is insufficient to reconstruct a 3D image of the region of interest: that is, due to the specific design of the subsets described below, the intersection of the projection cones of the images of each subset may be suboptimal, small, or ineffective, or in any case smaller than the region of interest.

[0070] In contrast, the combination of all subsets contains a sufficient number of 2D X-ray images whose intersections are optimized at the level of all subsets so that a 3D image of the region of interest can be reconstructed. In particular, the intersection of the projection cones of the images of all subsets is calculated to be larger than the region of interest.

[0071] Preferably, each subset contains the same number of two-dimensional X-ray images. In particular, the method forms pairs, triplets, or more generally groups of n two-dimensional X-ray images (n being an integer equal to or greater than 1, the number of subsets), such that each two-dimensional X-ray image in a group belongs to a respective different subset. Alternatively, each image of a subset is associated with an image of another subset to form such a group. This association is only virtual, i.e., although the n images of a group meet the requirements described below, they are not physically combined, stitched, or otherwise linked together to reconstruct a three-dimensional image.

[0072] More precisely, each group corresponds to the same position of the X-ray source along the n passes performed to acquire the n subsets. The requirement that the images of each group must fulfill is that the projection cone of an image of any subset is contiguous with the projection cone of an image of at least one other subset. As mentioned above, "contiguous" means that two adjacent projection cones of a group are in contact or partially overlap.

[0073] A first example of such a situation, where n is 2, is shown diagrammatically in FIG.

[0074] The X-ray source S produces a pair of images I A , I B Each image I has a common position. A , I B belong to different subsets.

[0075] Image I A , I B are the respective projection cones C of X-rays generated by the X-ray source S projected onto the image detector in each pass of the X-ray imaging system. A , C B Corresponds to.

[0076] For example, as indicated by the arrow, image I A is acquired along a first pass of the X-ray imaging system following a clockwise revolution around the isocenter, and image I B is acquired along a second pass of the X-ray imaging system according to a counterclockwise revolution of the X-ray imaging system around the isocenter. Performing the first and second passes according to opposite rotational directions is advantageous for optimizing the time and overall pass of acquisition of a set of 2D X-ray images. In particular, performing the first and second passes in opposite directions avoids returning the C-arm to its starting position between the two passes, thus saving time. Furthermore, the movement of the C-arm is minimized, which is beneficial for reducing positioning errors of the projection cone due to inaccuracies in the mechanical parts of the X-ray imaging system.

[0077] In the illustrated embodiment, the projection cone C A , C B share a common boundary and are therefore contiguous.

[0078] Alternatively, the projection cones may overlap, which is considered to be continuous within the meaning of this disclosure. This situation is shown diagrammatically in FIG.

[0079] The X-ray source S produces a pair of images I A , I B Each image IA , I B belong to different subsets.

[0080] Image I A , I B is the respective projection cone C of X-rays emitted from the X-ray source S projected onto the image detector in each pass of the X-ray imaging system. A , C B Corresponds to.

[0081] For example, as indicated by the arrow, image I A is acquired along a first pass of the C-arm following a clockwise revolution around the isocenter, and image I B is acquired along a second pass of the C-arm following a counterclockwise revolution of the C-arm about the isocenter.

[0082] Projection Cone C A , C B partially overlap and share a common region corresponding to their intersection.

[0083] An advantage of such an overlap is that it avoids gaps between the projection cones that may occur due to imprecision in the mechanical parts of the X-ray imaging system, for example.

[0084] In addition, since a greater amount of X-rays pass through the region of interest in the overlapping area, the accuracy of the three-dimensional image reconstructed in this area is improved.

[0085] Another advantage of such duplication is that image I A , I B The advantage of this method is that it allows the user to register the images I at different times, based on the elements of the region of interest that are present in the intersections of the images. A , I B For example, the region of interest may move periodically due to the patient's breathing, and images I A , I BIf the overlapping areas of the images are not acquired, the location of the region of interest will not be the same in the two images. This can be corrected by the overlapping areas of the images registration method described above.

[0086] In a preferred embodiment, a phantom is used to calibrate the 3D geometry of an X-ray imaging system simultaneously with image acquisition. As shown in FIG. 14, the phantom 3 includes a number of radiopaque fiducials 30 arranged according to a known geometry. The radiopaque fiducials can be detected in a two-dimensional X-ray image of a portion of a patient's body that includes the phantom. However, the phantom is usually very small and compact, so it may only be visible in a portion of the two-dimensional X-ray image. In each group, an image in which the phantom is visible (e.g., image I in the figure) is selected. A ) using the overlapping parts of the images to create an image where the phantom is not visible (e.g., image I in the figure). B ) can be aligned.

[0087] In some embodiments, the X-ray imaging system may include a dynamic collimator adapted to reduce the area of ​​the X-ray beam generated by the X-ray source. In this manner, the dynamic collimator may be adjusted for each projection cone to protect areas of some of the patient's body that should not be or do not need to be irradiated. Such collimation is shown diagrammatically in FIG. 11, where cone C A and C B The grey areas in represent areas not illuminated by X-rays. By selecting dynamic collimation for each imaging position (which may be different for each position) and for two or more image acquisition passes of the X-ray imaging system, it is possible to reconstruct three-dimensional images that fit many desired geometries while avoiding irradiating selected areas of the patient's body with X-rays. Alternating clockwise and counterclockwise passes also makes it possible to generate these geometries in a reasonable time frame, since the X-ray imaging system is always moving at the desired speed to collect the required images.

[0088] Another example where n is 3 is shown diagrammatically in FIG.

[0089] The X-ray source S produces a set of images I A , I B , I C Each image I has a common position. A , I B , I C belong to different subsets. Image I A , I B , I C is the projection cone C of each X-ray generated by the X-ray source S projected onto the image detector in each pass of the X-ray imaging system. A , C B , C C For example, as shown by the arrow, image I A is acquired along a first pass of the C-arm following a clockwise revolution around the isocenter, and image I B is acquired along a second pass of the C-arm following a counterclockwise revolution of the C-arm around the isocenter, and image I C is acquired along a third pass of the C-arm following a clockwise revolution about the isocenter.

[0090] Image I A , I B , I C are represented as having a common boundary, but as mentioned above, image I A , I B , Image I B , I C There may be overlap between the three images. In this way, the three images can be aligned together based on their overlapping portions.

[0091] Of course, the number of subsets n may be greater than 3, but this would result in a greater number of passes through the X-ray imaging system.

[0092] The union of the n images of each group can be viewed as an expanded cone of X-rays projected onto a virtual image detector, the X-rays being generated by an X-ray source passing through the region of interest.

[0093] The intersection of all the magnification cones obtained in multiple passes of the X-ray imaging system is larger than the intersection of the projection cones of two-dimensional X-ray images obtained in one pass of the X-ray imaging system that do not form a group of consecutive images. Thus, the size of the reconstructed three-dimensional image is larger. As mentioned above, the reconstruction of a three-dimensional image does not require combining, stitching, or otherwise physically linking the two-dimensional X-ray images of each group. Conventionally, the reconstruction of a three-dimensional image requires the intersection of the projection cones corresponding to each two-dimensional X-ray image of the entire image set, but due to the requirement of continuity of the images of each group, the volume obtained from this intersection is larger than that of a conventional pass of the X-ray imaging system.

[0094] As shown in Figures 4, 9 and 12 above, the apex of the magnification cone resulting from the union of the projection cones of each group of two-dimensional X-ray images is the center point of the X-ray source S, and the base of the magnification cone defines the virtual image detector. Therefore, the center of the height of the magnification cone, i.e., the center of the distance between the X-ray source and the virtual image detector, can be considered to be the virtual isocenter O of the X-ray imaging system.

[0095] Considered individually, the paths of an X-ray imaging system have no clinical meaning and are not designed to provide sufficient intersection of the respective cones to reconstruct relevant 3D images. However, by using a virtual isocenter, a link can be established between the images acquired during the path and the patient.

[0096] To determine the path of the x-ray imaging system, a virtual isocenter is placed at a predetermined location or series of locations relative to the center of the region of interest.

[0097] Advantageously, in order to optimise the size of the three-dimensional image, the virtual isocentre should be located substantially at the centre C of the region of interest (see FIG. 5) or at the shortest distance from the centre C of the region of interest (see FIG. 11).

[0098] The location of the center of the region of interest relative to the X-ray imaging system can be determined or at least estimated in various ways. In some embodiments, a phantom containing a set of radiopaque fiducials of known spatial configuration is placed in the volume to be imaged such that the spatial relationship between the radiopaque fiducials (or any other fixed reference) and the center of the region of interest is known and fixed. Typically, the center of the region of interest relative to the patient's body is at least approximately known based on anatomical knowledge of the patient's body. Alternatively, the center of the region of interest can be defined by the physician's input on pre-operative or intra-operative images acquired prior to the acquisition of the path for reconstructing the three-dimensional image. For example, the three-dimensional positioning of the center of the region of interest can be performed by acquiring two two-dimensional X-ray images using the X-ray imaging system. For example, these two images can be a front view and a side view of the region of interest to be imaged. In other embodiments, the center of the region of interest can be automatically determined by image analysis. In any case, the location of the center of the region of interest is determined in the coordinate system of the X-ray imaging system. In this way, the control unit can move the C-arm relative to the center of the region of interest.

[0099] In this way, a path corresponding to each subset of the two-dimensional x-ray image can be calculated for each position of the x-ray source such that a virtual isocenter is located at or near the center of the region of interest.

[0100] FIG. 5 shows a schematic representation of a first and a second pass of an X-ray imaging system, for example where the virtual isocenter of the X-ray imaging system is at a fixed position corresponding to the position of the center C of the region of interest, when n is 2 (see FIG. 4). Alternatively, the virtual image detector has an orbital motion around the virtual isocenter. Each pass is designed to acquire a respective subset of two-dimensional X-ray images, such that each image of the subset corresponds to a projection cone that is successive to the projection cone of the image of the other subset. In each pass, the real isocenter of the C-arm (i.e. the center of the line segment connecting the X-ray source S to the center of the detector panel) moves to meet the above requirement for the position of the virtual isocenter.

[0101] point O A represents the actual isocenter position of the C-arm during the first pass, and point O B represents the actual isocenter position of the C-arm during the second pass.

[0102] Continuous projection cone C A , C B One group of projection cones along the first and second passes for other positions of the C-arm are shown with grey borders. The dotted contour surrounding the virtual isocenter O represents the reconstructed 3D image volume V, defined by the intersection of multiple cones.

[0103] d A and d B is each cone C A , C B This width is also the size of each 2D X-ray image. d is the base width of the cone C A , C B represents the width of the base of the expansion cone formed by the union of

[0104] 6A and 6B show schematic first and second trajectories of the actual isocenter of the C-arm for obtaining the volume V shown in FIG.

[0105] To reconstruct a volume V with diameter d of approximately 30 mm, the real isocenter of the C-arm needs to be moved by 7.5 cm in the x direction and 15 cm in the z direction for both passes. The angular displacement of the virtual image detector is 172.5° to −7.5° for the first pass and 187.5° to 7.5° for the second pass.

[0106] The orbit of the virtual image detector can be more complicated than a pure orbital motion.

[0107] WO 2022 / 101432 teaches a complex trajectory that may be implemented by the X-ray imaging system in the present invention to minimize the distance between the center of the region of interest and the virtual image detector. This complex trajectory combines a rotational motion about the isocenter and a translational motion along a central axis that extends between the X-ray source and the center of the virtual image detector and passes through the center of the region of interest. This complex trajectory also takes into account the position of the operating table and the patient to avoid collisions.

[0108] In some embodiments, the complex trajectory may also include a rotation about the X-axis (angle α). Such a rotation may reduce the effect of metal artifacts in the X-ray image. For example, the angle α with respect to the vertical plane XZ may vary between -5° and +5°, either continuously or discontinuously.

[0109] This optimized trajectory allows for a shorter distance between the patient's body and the virtual image detector compared to the simple trajectory, which provides at least two advantages: (1) Increase the size of 3D medical images. (2) Reducing the amount of x-rays patients receive.

[0110] This is best understood from FIG. 7, which shows one angular position of the C-arm for acquiring a two-dimensional x-ray image.

[0111] In the first step shown on the left, the operating table T, the patient P, the X-ray source S and the virtual image detector D' are represented during the orbital movement of the C-arm. The letter O denotes the orbital rotation centre of the C-arm (corresponding to the virtual isocenter of the C-arm) and the letter C denotes the centre of the region of interest. In the illustrated situation, the orbital rotation centre O is located slightly below the centre C of the region of interest. The axis A represents the central axis extending between the centres of the X-ray source and the virtual image detector and passing through the centre C of the region of interest.

[0112] In the second step, shown on the right side of FIG. 7, the translational motion T of the C-arm along axis A is A This translational motion T Ahas the effect of moving the virtual detector D' closer to the patient. As a result, the 2D X-ray image acquired when the C-arm is in this position is larger than in the situation shown on the left side of Figure 7. Also, because the X-ray source is further away from the patient, the X-ray dose received by the patient during this exposure is lower than in the situation shown on the left side of Figure 7.

[0113] A translational motion T is performed by the control unit to minimize the distance between the virtual detector D' and the center C of the region of interest. A is calculated.

[0114] The first and second steps may be carried out successively or simultaneously.

[0115] Although not shown, the C-arm can also be tilted by an angle α with respect to the vertical plane XZ. This additional operation can be performed simultaneously with the first and second steps, or can be performed as a third step following the first and second steps.

[0116] For each of the n passes, the optimized trajectory is composed of multiple positions of the C-arm relative to the patient, each position being determined as described above.

[0117] Of course, the control unit can also take into account the position of the patient and the operating table (and other surgical systems located in the vicinity of the patient) relative to the C-arm in order to avoid collisions. In such a case, the control unit determines the translational movement T of the C-arm sufficient to avoid collisions while minimizing the distance between the patient and the virtual detector. A Calculate.

[0118] To that end, a collision prevention system can be integrated into the X-ray imaging system. Such a collision prevention system may include a sensor, such as a proximity sensor, a telemeter, a tactile sensor, etc., and may be configured to detect the presence of an object at a defined distance from at least one moving part of the C-arm. The control unit may be configured to stop the movement of the C-arm if the sensor detects an object near the moving part of the C-arm.

[0119] For example, a collision avoidance system capable of detecting a collision risk with an external object along each translational motion calculated for optimizing imaging can execute a collision avoidance trajectory before the actual imaging trajectory, thereby ensuring that the calculated trajectory optimizes the translational motion and does not involve a collision risk (i.e., the maximum translational motion considering the collision risk is executed). If it is confirmed that there is no collision risk by this collision avoidance trajectory, the imaging trajectory can be executed accordingly.

[0120] Of course, the collision avoidance system may remain active during the imaging trajectory to avoid collisions if there are unexpected obstacles along the trajectory.

[0121] In some situations, the translational motion of the C-arm is limited by the maximum range of the motor: for example, the C-arm trajectory may not be able to completely minimize the distance between the region of interest and the detector because the motor reaches its maximum position and the C-arm cannot be translated further.

[0122] To avoid this situation, C-arm users are advised to start the trajectory with all C-arm motors centered in their range of motion.

[0123] Considering the trajectory of the virtual isocenter as a whole, it can be considered to be substantially U-shaped. By U-shaped is meant a planar shape consisting of three consecutive segments arranged substantially at 90° to each other. The segments may be substantially straight and may have a constant radius of curvature. Also, the connections between the segments may be at right angles and may have a small radius of curvature. According to a preferred embodiment, the first and third segments extend in the vertical direction and the second segment connecting the first and third segments extends in the horizontal direction. The second segment is advantageously connected to the lower ends of the first and third segments.

[0124] Such a trajectory has been demonstrated to be an optimal trade-off between maximizing volume and avoiding collision risk.

[0125] The U-shape may be angular or square, which is obtained when the C-arm gantry has no more than one degree of freedom (apart from the orbital motion) for translational motion along the X-ray central axis of a given imaging position.

[0126] Alternatively, the U-shape may be curved for more complex trajectories with two or more degrees of freedom of the C-arm gantry (apart from the orbital motion) that are optimized for both the size of the reconstructed volume and the dose received by the patient.

[0127] The degree of freedom of the C-arm gantry can be selected taking into account the imaging time and the mechanical capabilities of the C-arm.

[0128] In Fig. 8, multiple positions of the X-ray source and X-ray detector in a half orbit (90° angular rotation) of the C-arm are shown, with the dotted lines corresponding to the central axis extending between the center of the X-ray source and the center of the X-ray image detector at each angular position of the C-arm and passing through the center of the region of interest. The multiple positions of the center O of the orbital motion of the C-arm are arranged along half of a square U-shape around the center C of the region of interest.

[0129] 9 and 10 show an alternative U-shaped trajectory of the C-arm.

[0130] In the embodiment of FIG. 9, the translational motion of the C-arm is calculated to minimize the distance between the virtual detector and the center of the region of interest. The trajectory of the virtual isocenter is substantially in the shape of an angular U. In this case, the vertical translational motion of the C-arm is calculated to bring the virtual detector as close as possible to the patient's body, while the detector remains away from the patient's body if tilted relative to the vertical direction. In this way, the X-ray dose received by the patient is minimized. In this embodiment, the range of motion of the motor controlling the vertical movement of the C-arm must be increased. As mentioned above, this embodiment can also be realized using an operating table that can be moved along the vertical direction.

[0131] In the embodiment of Fig. 10A, the locus of the center of rotation O of the orbital motion of the C-arm has a curved U-shape. At each imaging position of the C-arm, the distance between the detector and the center of the region of interest is minimized.

[0132] The embodiment of Fig. 10B differs from the embodiment of Fig. 10A in that the patient is more obese and the center C of the region of interest is located further away from the operating table T. As a result, according to the optimized trajectory, the translational movement of the virtual detector D' towards the center C is not hindered by the operating table, allowing the virtual detector D' to be closer to the patient along the entire trajectory, especially laterally.

[0133] 11 and 12A-12B, the virtual isocenter O has an elliptical orbit adapted to minimize the distance between the virtual image detector and the center of the region of interest C for each position of the x-ray source. Such an orbit can increase the size of the reconstructed three-dimensional image.

[0134] For example, to reconstruct a volume V with diameter d of approximately 38 mm, the virtual isocenter needs to be moved 28 cm in the x direction and 15 cm in the z direction. The real isocenter of the C-arm needs to be moved 28 cm in the x direction on the first pass (30 cm on the second pass) and 24 cm in the z direction on both passes. The angular displacement of the virtual image detector is 172.5° to -7.5° on the first pass and 187.5° to -7.5° on the second pass.

[0135] Other types of paths can be defined as appropriate.

[0136] In particular, the method described above can be used to increase the size of the reconstructed image not only in the X and Z directions as described in the previous embodiment, but also in the Y direction.

[0137] Therefore, two passes of the X-ray imaging system can be calculated to obtain two subsets of two-dimensional X-ray images using a rotation of the C-arm about the angle α and a translation along the Y axis, as shown in Figure 17. For the same position of the X-ray source along both passes, the projection cone C of the first subset of images is D is the projection cone of the second subset of images C E As mentioned above, the first pass can be performed in a clockwise direction and the second pass can be performed in a counterclockwise direction.

[0138] Preferably, the above path is combined with the paths described above to increase the size of the reconstructed 3D image in X, Y and Z directions, making it possible to image particularly elongated anatomical structures such as the patient's spine.

[0139] The X-ray imaging system described with reference to FIG. 3 is particularly advantageous for carrying out the methods of the present disclosure due to (i) its high mechanical precision and (ii) the large amplitude of displacement provided by the motorized arm 20.

[0140] Of course, as mentioned above, more than two passes can be used, and the projection cones may overlap rather than simply sharing a boundary.

[0141] After the n passes have been calculated, the control unit controls the C-arm to perform each pass and acquire a subset of the two-dimensional x-ray image during each pass.

[0142] The control unit then calculates a reconstruction of the 3D image based on the n subsets of 2D X-ray images, for which purpose the control unit executes algorithms known to those skilled in the art.

[0143] As a result, the resulting three-dimensional image is larger in size than a three-dimensional image reconstructed from the imaging paths of a conventional X-ray imaging system.

[0144] It should be noted that the present disclosure is not limited to the illustrated embodiments described above: in particular, individual embodiments can be combined whenever technically feasible.

Claims

1. 1. A method for reconstructing a three-dimensional medical image of a region of interest from a set of two-dimensional X-ray images acquired by an X-ray imaging system including an X-ray source (S) and an image detector (D), comprising: The method includes computing n passes (n being an integer greater than or equal to 2) of the X-ray imaging system, the passes adapted to acquire a subset of n respective two-dimensional X-ray images that together form the set of two-dimensional X-ray images, and each image of the subset (I Ａ , I Ｂ , I Ｃ ) for the same position of the X-ray source (S) along the n paths, the projection cones (C) of each of the regions of interest onto the image detector are projected so that the projection cone of an image of any subset is continuous with the projection cone of an image of at least one other subset. Ａ , C Ｂ , C Ｃ ) and calculating the n paths positioning a virtual isocenter (O) of the X-ray imaging system at a given position or a given series of positions relative to a center (C) of the region of interest; and defining the n projection cones such that for each position of the X-ray source along the n paths, the center of a height of an expansion cone formed by combining the n projection cones is the virtual isocenter (O); The method comprises: performing the calculated n passes to obtain n subsets of the image; reconstructing the three-dimensional medical image by intersecting the projection cones defined by each image of the n subsets.

2. 2. The method of claim 1, wherein n is 2.

3. 2. The method of claim 1, wherein n is 3.

4. The method according to any one of claims 1 to 3, wherein the given position of the virtual isocenter (O) is substantially the center (C) of the region of interest.

5. 4. The method according to claim 1, wherein the given series of positions of the virtual isocenter (O) is selected to minimize the distance between the virtual isocenter (O) and the center of the region of interest (C) for each position of the X-ray source along each of the paths.

6. The method of any one of claims 1 to 3, wherein at least two of the n passes are performed in opposite directions of orbital rotation of the X-ray imaging system.

7. The method of any one of claims 1 to 3, wherein the intersection between the projection cones of each subset of images is smaller than the region of interest.

8. The method of any one of claims 1 to 3, wherein for the same position of the X-ray source along the n paths, images of at least two of the n subsets partially overlap.

9. 6. The method of claim 5, wherein a calibration phantom (3) including a radiopaque fiducial (3) is detectable only in one first subset of images, and the method further comprises registering at least one other subset of images to the first subset of images based on overlapping portions of the images.

10. The method of any one of claims 1 to 3, further comprising selectively reducing the projection cone of each two-dimensional X-ray image using a dynamic collimator.

11. The method according to any one of claims 1 to 3, wherein the X-ray imaging system comprises a base (10), a C-shaped gantry (G) supporting the X-ray source (S) and the image detector (D), and a motorized arm (20) connecting the C-shaped gantry (G) to the base (10), the motorized arm (20) having at least three rotation axes (Z1, Z2, Z3) along a substantially common vertical direction (Z).

12. 1. A medical system for reconstructing a three-dimensional medical image of a region of interest, comprising: an X-ray imaging system including an X-ray source (S) and an image detector (D); a control unit configured to calculate n passes (n being an integer greater than or equal to 2) of the X-ray imaging system, the passes adapted to acquire n respective subsets of two-dimensional X-ray images that together form a set of two-dimensional X-ray images, and each image of the subset (I Ａ , I Ｂ , I Ｃ ) for the same position of the X-ray source (S) along the n paths, the projection cones (C) of each of the regions of interest onto the image detector are projected so that the projection cone of an image of any subset is continuous with the projection cone of an image of at least one other subset. Ａ , C Ｂ , C Ｃ ) and the calculation of the n paths is positioning a virtual isocenter (O) of the X-ray imaging system at a given position or a given series of positions relative to a center (C) of the region of interest; and defining the n projection cones such that for each position of the X-ray source along the n paths, the center of a height of an expansion cone formed by combining the n projection cones is the virtual isocenter (O); The control unit controlling movement of the X-ray imaging system to perform the calculated n passes to acquire the n subsets of images; The medical system is further configured to reconstruct the three-dimensional medical image by intersecting the projection cones defined by each image of the n subsets.

13. 13. The medical system of claim 12, wherein the X-ray imaging system comprises a base (10), a C-shaped gantry (G) supporting the X-ray source (S) and the image detector (D), and a motorized arm (20) connecting the C-shaped gantry (G) to the base (10), the motorized arm (20) having at least three rotation axes (Z1, Z2, Z3) along a substantially common vertical direction (Z).

Images