MEDICAL ROBOT FOR PLACING MEDICAL INSTRUMENTS UNDER ULTRASOUND GUIDE

Description

Field of invention

[0001] The present application belongs to the field of robotic devices for assisting a practitioner during a minimally invasive medical intervention comprising the insertion of one or more medical instruments into an anatomy of interest of a patient. In particular, the invention relates to a medical robot configured to track the movement of a target point in a lesion within an anatomy of interest of a patient and to adjust in real time the position of an articulated arm of the robot to optimally guide a medical instrument towards the target point. The movement of the target point may in particular be generated by the patient's breathing, or by the insertion of the medical instrument. State of the art

[0002] To prepare for a minimally invasive procedure aimed at reaching a target anatomical area in a patient's anatomy of interest with a medical instrument, a practitioner generally plans the procedure based on a preoperative (a few days or weeks before the procedure) or pre-interventional (just before the procedure when the patient is installed on the operating table) medical image. The minimally invasive medical procedure may, in particular, aim to perform a biopsy or ablation of a tumor in an organ, to perform a vertebroplasty, a cementoplasty or to stimulate a particular anatomical area. The anatomy of interest may correspond, for example, to a lung, a kidney, the liver, the brain, the tibia, the knee, a vertebra, etc. The medical instrument may be a needle, a probe, a catheter, etc.

[0003] During this planning step, the practitioner defines a target point in a region of the anatomy of interest to be treated. The practitioner also defines an entry point for the medical instrument on the patient's skin. These two points then define a trajectory that the medical instrument must follow to perform the medical intervention. In the particular case of soft organs located in the thoracic region, the abdominal region or the pelvic region, movements related to the patient's breathing and / or local deformations of the organ due to the insertion of the medical instrument cause a displacement of the target point during the intervention. Preoperative or pre-interventional planning medical images do not allow this displacement of the target point during the intervention to be predicted.Thus, the position of the target point (i.e., the position of the region to be treated in the anatomy of interest) is usually different during the acquisition of the planning medical image and during the intervention. Therefore, when the insertion of the medical instrument is planned from the planning medical image, there is a risk that the target point will not be reached accurately by the medical instrument.

[0004] Furthermore, there is a risk that the medical instrument will bend during insertion and not reach the target point if the planned trajectory that the medical instrument is to follow is not adjusted accordingly.

[0005] To limit the displacement of the target point caused by the patient's breathing, it is possible, at the time of insertion of the medical instrument, to block the patient's breathing at a phase of the respiratory cycle corresponding to that at which the medical planning image was acquired. The blocking of breathing can be carried out voluntarily by the patient if the medical intervention takes place under local anesthesia, or in a controlled manner by the practitioner if the medical intervention takes place under general anesthesia (interruption of mechanical ventilation). This solution is however not always very precise because it is difficult to obtain an exact correspondence between the phase of the respiratory cycle at which the medical planning image was acquired and the phase of the respiratory cycle at which the patient's breathing is blocked during the intervention.Furthermore, this solution requires relatively rapid insertion of the medical instrument since this must be done while the patient's breathing is blocked.

[0006] It is also possible to take several medical planning images during a patient's respiratory cycle and determine the trajectory least subject to deformations and displacements of the anatomy of interest caused by breathing. However, there is again a risk that the target point will not be reached accurately by the medical instrument.

[0007] It is also possible to monitor the position of the target point throughout the procedure by regularly acquiring intra-interventional medical images (images acquired when the medical instrument is inserted into the patient's body). These medical images are generally acquired by computed tomography, X-rays or magnetic resonance imaging. In the case of computed tomography or X-rays, however, such a solution has the disadvantage of significantly irradiating the patient and the practitioner during the procedure. In the case of magnetic resonance imaging, it is necessary to use specific non-magnetic equipment, particularly for the anesthetic equipment, which is particularly restrictive. This solution also requires the use of bulky imaging devices throughout the procedure.

[0008] It is also known to track the position of a lesion within an anatomy of interest using ultrasound images. However, the lesion is not always visible on an ultrasound image, and existing solutions generally lack precision.

[0009] It therefore remains necessary to find a solution for inserting a medical instrument precisely at a target point in a region to be treated within a patient's anatomy of interest, particularly when movements related to the patient's breathing and / or local deformations of the anatomy of interest due to the insertion of the medical instrument cause the target point to move during the procedure.

[0010] Patent application US2008 / 287794A1 discloses a method for implementing an imaging and navigation system to perform a medical procedure such as cardiac ablation. The method includes using an ultrasound imaging device to provide a three-dimensional image of the patient in real time, identifying a target site on the three-dimensional image of the patient in real time, directing a medical instrument to the target site using a tracking system, and performing the medical procedure at the target site.

[0011] Patent application US2014 / 316247A1 discloses a method and apparatus for tracking a change in a region of interest in a subject as a function of respiration. A "respiration signal" is captured using ultrasound images. This is a time-dependent representation of the movement of the region of interest.

[0012] Patent application US2018 / 092628A1 describes an ultrasound diagnostic apparatus. Data from an ultrasound image acquired by an ultrasound probe are associated with the position of the probe. Image alignment between the ultrasound image and a medical image can be performed. Statement of the invention

[0013] The devices and methods disclosed in the present application aim to remedy all or part of the drawbacks of the prior art, in particular those set out above.

[0014] For this purpose, and according to a first aspect, a medical robot is proposed for assisting a practitioner during a medical intervention to treat a lesion in an anatomy of interest of a patient. The medical robot comprises a robotic arm comprising at a distal end a tool guide intended to guide a medical instrument. The medical robot also comprises a control unit configured to control the robotic arm. The medical robot is configured to cooperate with a navigation system and with an ultrasound probe intended to be positioned by the practitioner at the level of the anatomy of interest of the patient.The control unit is configured to be able to determine at any time, from information communicated by the navigation system, the position of a robot marker intended to be fixed on the medical robot, the position of a patient marker intended to be positioned on the patient near the anatomy of interest, and the position of a probe marker intended to be fixed on the ultrasound probe. During a modeling phase, the control unit is configured to receive a plurality of ultrasound images acquired by the ultrasound probe during at least one respiratory cycle of the patient and to generate, from said ultrasound images, a model making it possible to estimate the position of a target point at the lesion and the position of an entry point at the skin of the patient, as a function of the position of the patient marker, regardless of the moment considered in a respiratory cycle of the patient.During a guidance phase, the control unit is configured to control the robotic arm in real time based on the position of the patient marker to guide the medical instrument along a trajectory defined by the position of the target point and the position of the entry point associated with the position of the patient marker in the model.

[0015] In the present application, the term "position" must be understood in the broad sense as describing both the position and the orientation of an object in a three-dimensional reference frame (the term "pose" is sometimes used in the English literature). The positions of the markers (patient marker, robot marker and probe marker) as well as the position of the target point and the position of the entry point can be defined in a reference frame of the robot or in a reference frame of the navigation system.It should be noted that the robot reference frame can be defined relative to the navigation system reference frame because the position of the robot marker is known both in the navigation system reference frame and in the robot reference frame (each joint of the robotic arm includes, for example, an encoder making it possible to know the position of each articulated element of the robotic arm in the robot reference frame, and the position of the robot marker on the robot is known a priori by the control unit).

[0016] The modeling phase takes place at the beginning of the intervention, before approaching the robotic arm to the anatomy of interest and before inserting the medical instrument into the patient's body. The modeling phase allows modeling the movement of the target point and the movement of the entry point relative to the patient marker during a respiratory cycle. During the modeling phase, the positions of the target point and the positions of the entry point are correlated with the positions of the patient marker, thus it is possible to define the position of the target point and the position of the entry point as a function of the position of the patient marker. The resulting model therefore makes it possible to define, at any time during the respiratory cycle, the position of the target point and the position of the entry point that correspond to the trajectory that the medical instrument must follow to accurately reach the target point.

[0017] Once the position of the target point and the position of the entry point are modeled as a function of the position of the patient marker, it becomes possible, by tracking the position of the patient marker in real time using the navigation system, to define in real time the position that the robotic arm must take so that the medical instrument is guided along the trajectory defined by the position of the target point and the position of the entry point associated with the position of the patient marker. This real-time tracking can in particular take place before the insertion of the medical instrument.

[0018] With such arrangements, it becomes possible to block the patient's breathing at any time during the respiratory cycle to proceed with the insertion of the medical instrument. Indeed, regardless of the moment when the patient's breathing is blocked, the robotic arm will be correctly positioned to allow the insertion of the medical instrument along the desired trajectory.

[0019] Also, it is no longer even necessary to hold the patient's breathing during the procedure. In fact, the robotic arm is moved in real time so that the position of the robotic arm is constantly adjusted to guide the medical instrument along the desired trajectory.

[0020] The invention further makes it possible to minimize lateral readjustments of the trajectory after insertion of the medical instrument (such lateral readjustments of the trajectory are generally traumatic for the organ crossed by the medical instrument).

[0021] The medical instrument can thus be inserted into the region to be treated with great precision, regardless of the moment of insertion of the medical instrument during the respiratory cycle. The insertion of the medical instrument is generally implemented by the practitioner, the medical robot being intended to guide the insertion of the medical instrument by the practitioner. However, nothing would prevent the insertion of the medical instrument from being automated and controlled by the control unit.

[0022] Furthermore, since the modeling of the target point position and the entry point position during a respiratory cycle is performed from ultrasound images, the patient and the practitioner are not exposed to ionizing radiation during the procedure.

[0023] In particular embodiments, the invention may further comprise one or more of the following features, taken individually or in any technically possible combination.

[0024] In particular embodiments, during the modeling phase, the control unit is configured, for each ultrasound image received, to: determining the position of the patient marker and the position of the probe marker at the time at which said ultrasound image was acquired by the ultrasound probe, obtaining an analysis image, from the ultrasound image, on which the lesion is visible, determining, on the analysis image, a target point at the lesion and an entry point at the patient's skin, the target point and the entry point thus defining a trajectory to be followed for the medical instrument, determining the position of the target point and the position of the entry point from the position of the probe marker, establishing an association between the position of the patient marker, the position of the target point and the position of the entry point thus determined for said ultrasound image.

[0025] Furthermore, the control unit is configured to model, from the information thus obtained for the plurality of ultrasound images, the position of the target point and the position of the entry point, as a function of the position of the patient marker, regardless of the instant considered in a respiratory cycle of the patient.

[0026] In particular embodiments, during the guidance phase, upon insertion of the medical instrument, the control unit is configured to regularly receive new ultrasound images acquired by the ultrasound probe. The control unit is further configured to update the model from said new ultrasound images.

[0027] As soon as the practitioner begins to insert the medical instrument into the patient's body, the position of the entry point at the patient's skin is fixed and becomes a rotation pivot for the movements of the robot arm. However, it remains possible to monitor in real time the position of the target point, the position of the entry point, and the position of the patient marker with new ultrasound images acquired in real time during the insertion phase of the medical instrument. Such arrangements make it possible to take into account a possible displacement of the target point resulting from the insertion of the medical instrument. The target point can indeed move in the direction of the trajectory followed by the medical instrument during its insertion (this is particularly the case when the target point to be reached is in a lesion, for example a tumor, within a soft organ).Real-time determination of the position of the target point using ultrasound images makes it possible to update in real time the trajectory that the medical instrument must follow as well as the position of the robotic arm to guide the medical instrument along this trajectory.

[0028] In particular embodiments, during the guidance phase, upon insertion of the medical instrument, the control unit is further configured to determine, for each new ultrasound image received, the position of the medical instrument, and to adjust the real-time control of the robotic arm as a function of the position of the medical instrument.

[0029] Such arrangements make it possible to take into account the risk of the medical instrument bending during insertion and to adjust the real-time control of the robotic arm accordingly (the trajectory to be followed by the medical instrument is then no longer a straight line between the entry point and the target point).

[0030] In particular embodiments, the analysis image corresponds directly to the ultrasound image. This is the case in particular when the lesion is visible on the ultrasound image.

[0031] In particular embodiments, the control unit is configured to obtain the analysis image by merging the ultrasound image with a preoperative or pre-interventional reference image on which the lesion is visible.

[0032] Indeed, when the lesion is not visible on an ultrasound image (isoechoic lesion), it is not possible to determine the position of the target point directly on the ultrasound image. It is then necessary to register the ultrasound image with a reference image of a different modality on which the lesion is visible. This may include a preoperative or pre-interventional image of computed tomography, positron emission tomography or magnetic resonance imaging. The fusion of the ultrasound image with the reference image then gives an analysis image on which the lesion is visible. The registration can be global (registration on the entire anatomy of interest) or local (optimized registration on a particular area of ​​the anatomy of interest). The registration can be carried out rigidly (by translation and / or rotation) or non-rigidly (with deformation).Registration can notably be implemented by a machine learning algorithm based on the recognition of particular anatomical structures on the images to be merged.

[0033] In particular embodiments, a radiopaque element of the patient marker is visible on the reference image, and the control unit is configured to merge the ultrasound image with the reference image by registration based on the position of the patient marker relative to the position of the probe marker at the time at which the ultrasound image was acquired by the ultrasound probe.

[0034] In particular embodiments, the reference image is a computed tomography image, a positron emission tomography image, or a magnetic resonance imaging image.

[0035] In particular embodiments, the ultrasound images received from the ultrasound probe are B-mode ultrasound images.

[0036] In particular embodiments, the control unit is configured to receive and process ultrasound images acquired by the ultrasound probe at a rate of at least fifteen images per second.

[0037] Such arrangements ensure real-time monitoring of the position of the target point and therefore real-time adjustment of the position of the robotic arm so that the medical instrument is guided along the desired trajectory throughout the intervention.

[0038] In particular embodiments, the medical robot further comprises a user interface comprising a display screen allowing the practitioner to view the analysis images.

[0039] In particular embodiments, the user interface comprises input means allowing the practitioner to identify on an analysis image displayed on the display screen a target point and / or an entry point and / or an anatomical zone which must not be crossed by the medical instrument.

[0040] The practitioner can thus plan the intervention, using the user interface, on a pre-interventional image corresponding to an analysis image associated with a first ultrasound image acquired by the ultrasound probe (this can be directly the ultrasound image if the lesion is visible on the ultrasound image or, otherwise, a fusion image resulting from the registration of the ultrasound image with a preoperative image of a different modality on which the lesion is visible). The target point and the entry point defined by the practitioner on this first image are then automatically determined by the control unit on the following analysis images.Target point tracking can be implemented in particular by a motion tracking method in several successive ultrasound images, by a speckle deformation analysis or by an artificial intelligence algorithm (the speckle represents the set of small, rapidly fluctuating spots that appear in the instantaneous texture of an image and which give it a grainy appearance). When the lesion is not visible on the ultrasound image, to help track the target point on the analysis images, it is advantageous to track the movement of an anatomical structure close to the lesion visible on the ultrasound images (for example, a blood vessel).

[0041] In particular embodiments, the user interface includes an augmented reality device for overlaying the analysis images with real images of the patient's body on the display screen.

[0042] The augmented reality device allows the moving, three-dimensional lesion to be superimposed on the patient's body, as well as the progression of the medical instrument during its insertion. This could be, for example, a screen positioned on the operating table above the patient, or a mask, headset, or augmented reality glasses. This type of display facilitates the spatial representation of the patient's anatomy of interest by the practitioner.

[0043] In particular embodiments, the control unit is configured to compare an ultrasound image with a reference image on which the lesion is visible, and to provide an indication to the practitioner of the direction in which to move the ultrasound probe so that an ultrasound image acquired by the ultrasound probe includes an anatomical region in which the lesion is located.

[0044] The reference image is, for example, a preoperative or pre-interventional image of computed tomography, positron emission tomography or magnetic resonance imaging. The indication of the direction in which the ultrasound probe should be moved is, for example, signaled to the practitioner on the display screen of the user interface. According to other examples, the indication of the direction in which the ultrasound probe should be moved can be signaled to the practitioner by light signals or by haptic feedback (vibrations) at the ultrasound probe, or on an augmented reality display.

[0045] According to a second aspect, there is provided a medical device comprising a medical robot according to any one of the embodiments described above, as well as a navigation system and an ultrasound probe intended to cooperate with the medical robot. Presentation of figures

[0046] The invention will be better understood upon reading the following description, given by way of non-limiting example, and made with reference to the figures 1 to 15 which represent: [ Fig. 1 ] a schematic representation of a medical device according to the invention comprising a medical robot, a navigation system and a probe, [ Fig. 2 ] a schematic representation of the robotic arm of the medical robot, [ Fig. 3 ] a schematic representation of the tool guide intended to be fixed to the end of the robotic arm, [ Fig. 4 ] a representation of the tool guide highlighting a device for holding a medical instrument at the end of the tool guide, [ Fig. 5 ] a representation of the tool guide highlighting the positioning of the medical instrument on the tool guide as well as elements detectable by a navigation system forming a “robot marker”, [ Fig. 6] a schematic representation of a “patient marker” intended to be positioned on the patient near the anatomy of interest, [ Fig. 7 ] a schematic representation of a “probe marker” intended to be fixed on the ultrasound probe, [ Fig. 8 ] a schematic representation of the main steps of a process implemented by the control unit during a modeling phase (in the case where the lesion is visible on the ultrasound images) then during a real-time control phase of the robotic arm before the insertion of the medical instrument, [ Fig. 9 ] a schematic representation of a user interface allowing the practitioner to identify on an image a target point and / or an entry point and / or a risk zone, as well as treatment parameters, [ Fig. 10] a schematic representation of a preoperative or pre-interventional image (part a) of the figure), an ultrasound image (part b) of the figure), and a fusion image resulting from the registration of the preoperative or pre-interventional image and the ultrasound image (part c) of the figure), [ Fig. 11 ] a schematic representation of the main steps of a process implemented by the control unit during a planning phase based on a preoperative or pre-interventional image and then during the modeling phase (case where the lesion is not visible on the ultrasound images), [ Fig. 12 ] a schematic representation of an estimated movement of the patient marker during a patient breathing cycle, [ Fig. 13 ] a schematic representation of the tracking of the patient marker position over time, [ Fig. 14] a schematic representation of the real-time adjustment of the position of the robotic arm from the modeling of the position of the target point as a function of the position of the patient marker, [ Fig. 15 ] a schematic representation of the main steps of a process implemented by the control unit during an insertion phase of the medical instrument.

[0047] In these figures, identical references from one figure to another designate identical or similar elements. For reasons of clarity, the elements represented are not necessarily to the same scale, unless otherwise indicated. Detailed description of at least one embodiment of the invention

[0048] There figure 1 schematically represents a medical robot 10 according to the invention. The medical robot 10 is used to assist a practitioner during a medical intervention on an anatomy of interest of a patient 20 positioned on an intervention table 21.

[0049] An example is a minimally invasive or percutaneous medical procedure to treat a lesion within the patient's anatomy of interest. This type of procedure generally requires the practitioner to insert one or more medical instruments (e.g., a needle, a probe, a catheter, etc.) into the patient's body to a certain depth to reach a target anatomical area (a lesion, e.g., a tumor) in the anatomy of interest (e.g., in the liver, a lung, a kidney, etc.).

[0050] The medical robot 10 comprises a base 11. In the example considered, the base 11 of the medical robot 10 is equipped with motorized wheels, which allows the medical robot 10 to move in different directions by translational and / or rotational movements.

[0051] The medical robot 10 further comprises an articulated robotic arm 13, one end of which is connected to the base 11. At the other end of the robotic arm 13 is fixed a tool guide 14 intended to guide a medical instrument 15, such as for example a needle, a probe, a catheter, an electrode, etc. The medical robot 10 can then be used to help a practitioner position, hold, or guide the medical instrument 15 during the medical intervention. The medical robot 10 then plays the role of a third hand for the practitioner.

[0052] The medical robot 10 comprises a control unit 12 configured to control the movement of the robotic arm 13. The control unit 12 comprises, for example, one or more processors 122 and a memory 121 (magnetic hard disk, electronic memory, optical disk, etc.) in which a computer program product is stored, in the form of a set of program code instructions to be executed to implement the different steps of a method for positioning the robotic arm 13. The memory 121 also makes it possible to record the images and other information (in particular the navigation information) used to implement this method.

[0053] The medical robot 10 may also comprise a user interface 19 comprising a display screen allowing the practitioner to view ultrasound images acquired by the ultrasound probe 40 or other medical images (for example a preoperative or pre-interventional reference image of the anatomy of interest, or fusion images resulting from the registration of the ultrasound images with a reference image). The user interface may also comprise input means (keyboard, mouse, touch screen, etc.) allowing the practitioner to identify on an image displayed on the display screen a target point and / or an entry point and / or an anatomical zone which must not be crossed by the medical instrument 15.

[0054] In particular embodiments, the user interface may include an augmented reality device for superimposing the ultrasound images (or fusion images) with real images of the patient's body on the display screen. Such a device facilitates the spatial representation of the anatomy of interest for the practitioner.

[0055] The medical robot 10 is configured to cooperate with a navigation system 30 and with an ultrasound probe 40 intended to be positioned by the practitioner at the level of the anatomy of interest of the patient. The medical robot 10 comprises a communication module connected to the control unit 12 to exchange data with the navigation system 30 and with the ultrasound probe 40. The navigation system 30 and the ultrasound probe 40 also each comprise a communication module to exchange data with the control unit 12 of the medical robot 10. The communications established between the control unit 12, the navigation system 30 and the ultrasound probe may be wired communications or wireless communications. For the sake of simplification, the communication modules are not shown in the figure 1 .

[0056] In the example considered, the navigation system 30 is an optical navigation system. The navigation system 30 comprises two optical sensors 31 corresponding to two sensors of a stereoscopic camera operating in the infrared radiation range. In the example considered, the navigation system 30 further comprises a camera 32 operating in the visible light range.

[0057] The control unit 12 is configured to be able to determine at any time, from information communicated by the navigation system 30, the position of a robot marker 18 intended to be fixed on the medical robot 10, the position of a patient marker 22 intended to be positioned on the patient 20 near the anatomy of interest, and the position of a probe marker 41 intended to be fixed on the ultrasound probe 40.

[0058] In the present application, the term "position" corresponds to the combination of the position and orientation of an object in a given frame of reference which is generally a three-dimensional coordinate system. The term "pose" is used in English literature to represent this combination of the position and orientation of an object in space.

[0059] The control unit 12 is configured to receive from the ultrasound probe 40 ultrasound images acquired by the ultrasound probe 40.

[0060] The ultrasound images received from the ultrasound probe 40 and the information received from the navigation system 30 are temporally synchronized by the control unit 12 to be able to correlate at a given time the position of the lesion with the position of the patient marker 22.

[0061] Conventionally, the ultrasound probe 40 comprises one or more sound wave transmitter-receiver elements (piezoelectric materials, capacitive electronic transducers). The ultrasound probe produces ultrasonic waves by indirect piezoelectric effect. Each time a wave encounters an anatomical structure, a portion of this wave returns by reflection or diffusion (“speckle”) in the form of an echo. This echo is then transformed into an electric current by direct piezoelectric effect and then reconstructed into an image. The reconstruction of an ultrasound image depends mainly on the number, size and positions of the transmitter-receiver elements of the probe (lateral and longitudinal resolution), the duration of the transmission pulses and the echo times (axial and / or depth resolution). The energy of the received echo is then coded in gray levels. The higher the energy, the whiter the corresponding image portion (pixel).This grayscale coding is called "brightness" and the associated ultrasound mode is called "B-mode." The images produced by the ultrasound probe 40 may be two-dimensional images or three-dimensional images. Preferably, the ultrasound probe 40 is capable of generating images at a rate of at least fifteen images per second.

[0062] B-mode is particularly well suited when the anatomy of interest is the liver. However, it should be noted that the invention could also be applied with other ultrasound modes, such as elastography.

[0063] In the example considered, and as illustrated on the figure 2, the robotic arm 13 comprises six rotoid joints 131 to 136 providing six degrees of freedom allowing the medical instrument 15 to be positioned and / or moved in any position in the three-dimensional space. Advantageously, the joints 131 to 135 of the robotic arm 13 are not aligned and have an offset relative to each other, which allows a greater number of possible configurations of the robotic arm 13. Each joint comprises at least one encoder making it possible to know its angular position in real time. A configuration of the robotic arm 13 then corresponds to a set of parameter values ​​taken by the joints 131 to 136 (for example the value of an angle of rotation for each joint). The rotoid joint 136 corresponds to a rotation around the main axis of the tool guide 14.It should be noted that it is not necessary, however, to perform a rotation around the axis of symmetry of the medical instrument (five degrees of freedom are in fact sufficient to guide and release a medical instrument). This additional degree of freedom makes it possible to be in a situation of redundancy and to have an infinite number of possible configurations of the robotic arm 13 for a given position of the tool guide 14. This situation of redundancy is particularly useful for adapting to constraints linked to the position of the patient or to the configuration of the operating room. This situation of redundancy makes it possible in particular to adapt to the external envelope of the patient and to the position of the markers, for example if a configuration of the robotic arm hides one of the markers, it is possible to adopt another configuration of the robotic arm 13 while maintaining the same trajectory for the medical instrument 15.

[0064] In the example considered, and as illustrated on the figure 3 , the tool guide 14 is fixed to the robotic arm 13 by means of a flange 17. The tool guide comprises a main axis 145 shown in the figure 3 by a dotted line. The tool guide 14 is coupled to a force sensor 16 to enable the control unit 12 to determine a force exerted on the tool guide 14. This force may in particular be exerted by the practitioner when he manually moves the robotic arm 13. This force may also correspond to a force exerted on the tool guide 14 via the medical instrument 15 by the patient's body.

[0065] In the example considered, and as illustrated on the Figures 4 and 5, the tool guide 14 comprises a body 141 with a base 142 intended to be fixed to the flange 17 using screws 143, as well as a holding system 146 comprising two parts which are movable relative to each other. The holding system 146 is intended to hold the medical instrument 15 at the end of the body 141 of the tool guide 14 opposite the base 142. The two movable parts of the holding system 146 can be driven by a drive system such as a gear, a cam, a screw with reverse threads and / or a linear actuator, in order to block or release the medical instrument 15. The linear actuator can be reversible (the holding system 146 of the tool guide 14 can then be opened manually or automatically on command from the control unit 12) or non-reversible (the holding system 146 of the tool guide 14 can only be opened automatically on command from the control unit).The tool guide 14 makes it possible, for example, to guide medical instruments of different diameters. For example, such a guide makes it possible to guide medical instruments whose diameter is between 11 and 21 gauges. The gauge is a unit of measurement commonly used to define the external diameter of a medical instrument such as a needle, a probe or a catheter (11 gauges correspond to an external diameter of 2.946 mm; 21 gauges correspond to an external diameter of 0.812 mm).

[0066] As illustrated on the Figures 4 and 5, the tool guide 14 comprises studs 144 intended to receive optical markers 181. Advantageously, the tool guide 14 comprises at least three optical markers 181 so that the position of the tool guide 14 can be determined in the three spatial dimensions of the reference frame of the navigation system 30. The respective positions of the optical markers 181 of the tool guide 14 relative to each other are known a priori by the navigation device 30 and / or by the control unit 12. Advantageously, the geometric shape of each optical marker 181 can also be known a priori. In the example illustrated in Figure 5 , the optical markers 181 are spherical in shape.

[0067] All of the optical markers 181 present on the tool guide 14 correspond to the robot marker 18.

[0068] The use of at least three optical markers 181 makes it possible to define a plane and therefore a direct orthonormal three-dimensional reference frame with a z axis normal to the plane and x and y axes in the plane so that the reference frame is direct. This thus makes it possible to determine the position and orientation of the reference frame formed from the optical markers 181 which represent the tool guide 14. The three axes x, y and z make it possible to define six degrees of freedom, namely a translation along each of the x, y or z axes and a rotation around each of these axes.

[0069] The optical markers 181 may be passive or active. Passive optical markers reflect optical radiation emitted by another element, such as the navigation system 30. Passive optical markers may correspond, for example, to reflective spheres detectable by an infrared stereoscopic camera (this is what is used, for example, in the Polaris ®< navigation systems manufactured by Northern Digital Inc.), or to black and white patterns visible by a stereoscopic camera (this is what is used, for example, in the MicronTracker ®< navigation system from ClaroNav). Active optical markers themselves emit optical radiation, for example infrared radiation, detectable by the navigation system 30.

[0070] It should be noted, however, that a single optical marker with a characteristic three-dimensional geometric shape could be used instead of all 181 spherical optical markers.

[0071] There figure 6 schematically represents the patient marker 22 intended to be positioned on the patient 20 near the anatomy of interest. In the example considered, the patient marker 22 comprises four optical markers 23, such that the position of the patient marker 22 can be determined in the three spatial dimensions of the reference frame of the navigation system 30. The respective positions of the optical markers 23 of the patient marker 22 relative to each other are known a priori by the navigation system 30 and / or by the control unit 12. Advantageously, the geometric shape of each optical marker 23 can also be known a priori. In the example illustrated in figure 6, the optical markers 23 are spherical in shape. The spherical shape makes it possible to optimize the reflection of the optical radiation. What was mentioned previously for the active or passive type of the optical markers 181 of the tool guide 14 is also true for the optical markers 23 of the patient reference 22. Here again, it would be possible to use a single optical marker having a characteristic geometric shape in three dimensions instead of the four spherical optical markers 23.

[0072] Optionally, and as illustrated on the figure 6, the patient marker 22 may also comprise radiopaque markers 24 which are visible on a medical image acquired by a medical imaging device (for example by computed tomography, magnetic resonance, ultrasound, tomography, positron emission tomography, etc.). The respective positions of the radiopaque markers 24 relative to each other are known a priori by the navigation device 30 and / or by the control unit 12. Advantageously, the geometric shape of the radiopaque markers 24 may also be known a priori. In the example considered, the patient marker 22 comprises four radiopaque markers 24. The radiopaque markers 24 may be, for example, ceramic balls. It should be noted, however, that a single radiopaque marker having a characteristic geometric shape in three dimensions could be used instead of the four spherical radiopaque markers 24.

[0073] There figure 7 schematically represents the probe marker 41 intended to be fixed on the ultrasound probe 40 to allow the navigation system 30 to determine the position of the ultrasound probe 40. In the example considered, the probe marker 40 comprises three optical markers 42, such that the position of the probe marker 40 can be determined in the three spatial dimensions of the reference frame of the navigation system 30. The respective positions of the optical markers 42 of the probe marker 40 relative to each other are known a priori by the navigation system 30 and / or by the control unit 12. Advantageously, the geometric shape of each optical marker 42 can also be known a priori. In the example illustrated in figure 7, the optical markers 42 are spherical in shape. The spherical shape makes it possible to optimize the reflection of the optical radiation. What was mentioned previously for the active or passive type of the optical markers 181 of the tool guide 14 is also true for the optical markers 42 of the probe reference 40. Here again, it would be possible to use a single optical marker having a characteristic geometric shape in three dimensions instead of the three spherical optical markers 42.

[0074] In the remainder of the description, it is considered by way of non-limiting example that the optical sensors 31 of the navigation system 30 and the various optical markers 181, 23, 42 are designed to operate with infrared optical radiation. It is further considered that the optical markers 181, 23, 42 are passive markers. The optical sensors 31 are configured to emit infrared radiation. This infrared radiation is reflected by the various optical markers 181, 23, 42 towards the optical sensors 31. The optical sensors 31 are configured to receive this reflected infrared radiation. The navigation system 30 can then determine the distance between an optical marker 181, 23, 42 and an optical sensor 31 by measuring the time taken by an infrared ray to make the round trip between said optical sensor 31 and said optical marker 181, 23, 42.By knowing the distance between each optical marker 181, 23, 42 and each optical sensor 31, and by knowing a priori the arrangement of the optical markers 181, 23, 42 relative to each other on the robot marker 18, on the patient marker 22, and on the probe marker 41, it is possible to determine the position of the robot marker 18, of the patient marker 22, and of the probe marker 41 in the reference frame of the navigation system 30. It should be noted that infrared optical navigation is a well-known method in the field of surgical interventions assisted by a medical robot.

[0075] It should be noted that the invention is described using an optical navigation system. However, nothing would prevent the use, in a variant, of an electromagnetic navigation system instead of the optical navigation system. In this case, the different “markers” detectable by the navigation system (patient marker 22, robot marker 18, and probe marker 41) would then correspond to electromagnetic sensors whose position can be determined by the navigation system in a generated electromagnetic field.

[0076] In the example considered, the control unit 12 of the medical robot 10 is configured to receive from the navigation system 30 information on the current position of the robot marker 18 in the reference frame of the navigation system 30. However, the control unit 12 of the medical robot 10 knows the current position of the robot marker 18 in the reference frame of the medical robot 10 (via the encoders of the joints 131 to 136). The control unit 12 can therefore determine the transformation to be carried out to define a position in the reference frame of the medical robot 10 from a position in the reference frame of the navigation device 30.

[0077] The control unit 12 is also configured to receive from the navigation system 30 information on the position of the patient marker 22 and on the position of the probe marker 41 in the reference frame of the navigation system 30. The control unit 10 can then define the position of the patient marker 22 and the position of the probe marker 41 in the reference frame of the medical robot 10.

[0078] When the position of the probe marker 41 is known at a given instant, it is possible to determine the position of a visible element on an ultrasound image acquired by the ultrasound probe 40 at said instant. This visible element may in particular be a target point to be reached at the level of the lesion to be treated or an entry point of the medical instrument at the level of the patient's skin. The target point and the entry point define a trajectory that the medical instrument 15 must follow. When the position of the target point and the position of the entry point are known, that is to say when the trajectory that the medical instrument 15 must follow is defined, the control unit can automatically move the robotic arm 13 into a configuration allowing the tool guide 14 to guide the medical instrument 15 according to the defined trajectory.

[0079] However, and as explained previously, movements related to the patient's breathing may cause the target point to move. Thus, the trajectory that the medical instrument 15 must follow at a given moment in the patient's respiratory cycle is not the same at another moment in the respiratory cycle.

[0080] To solve this problem, the control unit 12 is configured to model, during a modeling phase which precedes the surgical procedure, the movement of the patient marker 22 during at least one respiratory cycle of the patient 20. During the modeling phase, the position of the target point and the position of the entry point are correlated with the position of the patient marker 22. It is thus possible to define the position of the target point and the position of the entry point as a function of the position of the patient marker. The modeling obtained therefore makes it possible to define, from the position of the patient marker 22, and at any time during the respiratory cycle, the position of the target point and the position of the entry point which correspond to the trajectory which the medical instrument 15 must follow to accurately reach the target point.

[0081] Once the position of the target point and the position of the entry point are modeled as a function of the position of the patient marker 22, it becomes possible, by tracking in real time the position of the patient marker 22 using the navigation system 30, to define in real time the position that the robotic arm 13 must take so that the medical instrument 15 is guided along the trajectory defined by the position of the target point and the position of the entry point associated with the position of the patient marker 22. This real-time tracking can in particular take place during a guidance phase, before the insertion of the medical instrument 15. The robotic arm 13 is moved in real time so that the position of the robotic arm is constantly adjusted to guide the medical instrument along the desired trajectory.

[0082] There figure 8schematically represents the main steps of a process implemented by the control unit 12 during a modeling phase, then during a real-time control phase of the robotic arm, before the insertion of the medical instrument 15.

[0083] During the modeling phase, the control unit 12 is configured to receive a plurality of ultrasound images acquired by the ultrasound probe 40 during at least one respiratory cycle of the patient 20. The following steps are therefore repeated throughout the modeling phase: acquisition 201 of an ultrasound image (the ultrasound image corresponds to an analysis image on which the lesion is visible), determination 202 of the position of the patient marker 22 and the position of the probe marker 41 at the instant at which said ultrasound image was acquired by the ultrasound probe 40, determination 204, on the analysis image, of a target point at the level of the lesion and an entry point at the level of the skin of the patient 20 (the target point and the entry point define a trajectory to be followed for the medical instrument 15), determination 205 of the position of the target point and the position of the entry point from the position of the probe marker 41, association 206 between the position of the patient marker 22, the position of the target point and the position of the entry point thus determined for said ultrasound image.

[0084] At the end of the modeling phase, the control unit 12 is configured to carry out, from the information thus obtained for the different ultrasound images, a modeling 207 of the position of the target point and the position of the entry point, as a function of the position of the patient marker 22, whatever the instant considered in a respiratory cycle of the patient.

[0085] Then, during the guidance phase, the control unit 12 is configured to control the robotic arm 13 in real time. For this, at each instant, the following steps are carried out by the control unit 12: determining 301 the position of the patient marker 22, determining 302, using the modeling, the position of the target point and the position of the entry point associated with the position of the patient marker 22, moving 303 the robotic arm 13 to adjust the position of the tool guide 14 so that the medical instrument 15 is guided by the tool guide 14 along the trajectory defined by the position of the target point and the position of the entry point thus determined.

[0086] An initial determination of a target point at the lesion and an entry point at the patient's skin is, for example, initially carried out on a first analysis image. Then, subsequently, a motion tracking algorithm in several successive analysis images can be implemented to determine the target point and the entry point on each new analysis image, for example using a deformation analysis of the "speckle" (set of fluctuating spots on the images due to the diffusion of waves) or by an artificial intelligence algorithm.

[0087] The first determination of a target point and an entry point can be carried out by the practitioner using the graphical interface 19. The figure 9illustrates a user interface allowing the practitioner to identify on an analysis image a target point 51 at the level of the region to be treated 50, and / or an entry point 52 at the level of the patient's skin, and / or a risk zone to be avoided (for example bones or blood vessels), as well as treatment parameters. This step can be facilitated by segmentation of certain anatomical regions (the anatomy of interest, the lesion to be treated, risk zones, etc.) by a machine learning algorithm.

[0088] Alternatively, the first determination of a target point and an entry point can be implemented automatically by an artificial intelligence algorithm.

[0089] However, it is possible that the lesion to be treated is not visible on an ultrasound image, for example because the nature of the lesion means that it is not (or is very little) visible on an ultrasound image. In such a case, a pre-interventional (i.e., an image acquired just before or at the beginning of the intervention) or pre-operative (i.e., an image acquired several days or weeks before the intervention) reference image should be used on which the lesion is visible. This may include a computed tomography image, a positron emission tomography image, or a magnetic resonance imaging image. The reference image may be a two-dimensional or a three-dimensional image. The first determination of a target point and an entry point can then be made on the reference image instead of on an initial ultrasound image.Again, the first determination of a target point and an entry point can be carried out by the practitioner (e.g. using the graphical interface 19 illustrated in the . figure 9 ) or automatically by an artificial intelligence algorithm. An ultrasound image acquired by the ultrasound probe 40 can then be registered with the reference image to form an analysis image corresponding to the fusion of the reference image with the ultrasound image. The target point and the entry point are then visible on the analysis image.

[0090] Alternatively, the first determination of a target point and an entry point may be performed on a first analysis image instead of being performed on the reference image.

[0091] There figure 10 schematically illustrates the use of a preoperative or pre-interventional reference image (part a) of the figure 10) to recalibrate an ultrasound image (part b) of the figure 10 ) in order to form an analysis image (part c) of the figure 10 ) resulting from the registration and fusion of the reference image with the ultrasound image. The lesion to be treated 50 and the target point 51 are visible on the reference image. In the example considered, the reference image is acquired by computed tomography. The lesion to be treated 50 is, however, almost not visible on the ultrasound image. The lesion to be treated 50 and the target point 51 become visible on the analysis image resulting from the registration of the reference image with the ultrasound image.

[0092] The registration is for example implemented by the control unit 12 using a machine learning algorithm trained to globally (registration on the entire anatomy) or locally (optimized on an area of ​​interest), rigidly (translation and rotation) or non-rigidly (deformation) a CT image of an anatomy of interest with an ultrasound image. The image resulting from this registration is called the analysis image or fusion image.

[0093] If radiopaque elements of the patient marker 22 are visible on the reference image, a rigid registration can also be based on the position of the patient marker 22 relative to the position of the probe marker 41 at the time at which the ultrasound image was acquired by the ultrasound probe 40.

[0094] There figure 11schematically represents the main steps implemented during a planning phase based on a preoperative or pre-interventional reference image, then during the modeling phase, in the case where the lesion is not visible on the ultrasound images.

[0095] The planning phase includes in particular a step 101 of acquiring a preoperative or pre-interventional reference image then a step 102 of determining the target point 51 and the entry point 52 on the reference image.

[0096] The modeling phase involves substantially the same steps as those described with reference to the figure 8. For each ultrasound image received during the modeling phase, the control unit 12 is further configured to implement a step 203 of generating an analysis image by registering the ultrasound image with the reference image. The determination 204 of the target point and the entry point is then carried out on this analysis image resulting from the registration of the ultrasound image with the reference image (whereas in the case of the figure 8 the analysis image corresponds directly to the ultrasound image).

[0097] When the lesion is not visible on the ultrasound image, to assist in tracking the target point on the analysis images, it may be advantageous to track the movement of an anatomical structure that is close to the lesion and visible on the ultrasound images (e.g., a blood vessel).

[0098] It should be noted that step 102 may be optional if the first determination of the target point and the entry point is performed on a first analysis image instead of being performed on the reference image.

[0099] Once the target point 51 and the entry point 52 have been determined on an analysis image, their respective position can be determined by rigid recalibration in the reference frame of the navigation system 30, or in the reference frame of the medical robot 10, thanks to the known position of the probe marker 41 relative to the position of the transmitter-receiver elements of the ultrasound probe 40.

[0100] THE figures 12 to 14 illustrate the modeling step 207 of the position of the target point 51 and the entry point 52 as a function of the position of the patient marker 22.

[0101] There figure 12represents, by way of example, a recording of the movement followed by the patient marker 22 during a recording period of a predetermined duration corresponding to several respiratory cycles of the patient. Each point corresponds to a position taken by the patient marker 22 over time in an XY plane of a coordinate system of the navigation system 30 (the movement of the patient marker 22 could also be represented in the reference frame of the medical robot 10). It can be seen, in this example, that the movement of the patient marker 22 takes place mainly along an axis 54.

[0102] The movement of the patient marker 22 is representative of the movement of the patient's rib cage caused by the patient's breathing. For a better interpretation of the marker's movement, and by analogy with the patient's respiratory cycle, it is preferable to obtain a one-dimensional curve, illustrating the oscillatory movement of the marker over time. There are different methods for obtaining this one-dimensional curve. For example, it is possible to consider that the movement of the marker is predominantly vertical, and therefore to consider only the Y axis. However, in this case, part of the amplitude of the marker's movement is lost. According to another example, it is possible to perform a principal component analysis of the marker positions. The marker positions can in particular be displayed according to a principal component corresponding to the main axis 54 of the marker's movement.

[0103] There figure 13 represents a curve 55 describing the movement of the patient marker 22 over time, during the recording period, along the main axis 54. The position (MP) of the patient marker 22 along the axis 54 is represented on the ordinate; the time (t) is represented on the abscissa. The recording period comprises several respiratory cycles of the patient 20.

[0104] As illustrated on the figure 13 , a high position of the rib cage of the patient 20 corresponds to the end of an inspiration in a breathing cycle. This also corresponds to a maximum of the curve 55 describing the position of the patient marker 22 over time. A low position of the rib cage of the patient 20 corresponds to the end of an expiration in a breathing cycle. This also corresponds to a minimum of the curve 55 describing the position of the patient marker 22 over time.

[0105] During the modeling phase, the control unit 12 determines for each analysis image the position of the target point, the position of the entry point and the position of the patient marker. These positions can then be correlated to model the position of the target point and the position of the entry point over time as a function of the position of the patient marker.

[0106] There figure 14 illustrates, by way of example, a model 56 of the position of the target point (C) as a function of the position of the patient marker (MP) during an inspiration period of a respiratory cycle.

[0107] By modeling the position of the target point and the position of the entry point over time as a function of the position of the patient marker, it becomes possible to determine in real time the configuration that the robotic arm 13 must take so that the tool guide 14 is constantly positioned such that the medical instrument 15 is guided by the tool guide 14 along the trajectory defined by the position of the target point and the position of the entry point. This real-time adjustment 303 of the position of the robotic arm 13 can be performed simply by tracking the position of the patient marker 22 in real time and using the model for estimating the position of the target point and the position of the entry point as a function of the position of the patient marker. It is then no longer necessary to acquire ultrasound images to adjust the position of the robotic arm 13 in real time during the guidance phase.

[0108] This makes it possible to block the patient's breathing at any time during the respiratory cycle to insert the medical instrument. In fact, regardless of when the patient's breathing is blocked, the robotic arm will be correctly positioned to allow the insertion of the medical instrument along the desired trajectory.

[0109] It is no longer even necessary to hold the patient's breathing during the procedure. Indeed, the phase of guiding the robotic arm using the model can be continued during the insertion of the medical instrument. The robotic arm is then controlled in real time so that the position of the robotic arm is constantly adjusted to guide the medical instrument along the desired trajectory. It should be noted, however, that during the insertion of the medical instrument, the position of the entry point at the patient's skin is fixed and becomes a rotation pivot for the robot arm's movements.

[0110] The guidance phase can be performed based entirely on the model generated during the modeling phase. Therefore, it is not necessary for ultrasound images to be acquired during the guidance phase.

[0111] However, it may be advantageous to take into account a possible displacement of the target point resulting from the insertion of the medical instrument. The target point may in fact move in the direction of the trajectory followed by the medical instrument during its insertion (this is particularly the case when the target point to be reached is in a lesion, for example a tumor, within a soft organ). It is then possible to update in real time the model making it possible to estimate the position of the target point in order to adjust the trajectory that the medical instrument must follow as well as the position of the robotic arm to guide the medical instrument along this trajectory.

[0112] There figure 15schematically represents the main steps of a method implemented by the control unit 12 during a guidance phase during the insertion of the medical instrument 15. The control unit 12 is configured to receive new ultrasound images acquired by the ultrasound probe 40. The following steps are therefore repeated for each new image: acquisition 401 of an ultrasound image, determination 402 of the position of the patient marker 22 and the position of the probe marker 41 at the time at which said ultrasound image was acquired by the ultrasound probe 40, generation 403 of an analysis image from the ultrasound image and the reference image (note that this step is optional if the lesion is directly visible on the ultrasound image), determination 404, from the analysis image and the position of the probe marker 41, of the position of the target point, the position of the entry point, and the position of the medical instrument 15, updating 405 of the model initially generated during the modeling phase, adjustment 406 of the position of the robotic arm 13 from the updated model.

[0113] Determining the position of the medical instrument for each new ultrasound image received also makes it possible to detect any curvature of the medical instrument during insertion, and if necessary to adjust the real-time control of the robotic arm accordingly.

[0114] If the ultrasound probe 40 is not correctly positioned, the lesion may not be within the field of view of the ultrasound probe. It is then appropriate to be able to provide the practitioner with information on a direction in which the ultrasound probe must be moved so that the lesion is within the field of view of the ultrasound probe.

[0115] For this purpose, the control unit 12 may be configured to compare an ultrasound image with a reference image on which the lesion is visible, and to give an indication to the practitioner of the direction in which the ultrasound probe 40 should be moved so that an ultrasound image acquired by the ultrasound probe 40 includes an anatomical region in which the lesion is located. This indication may for example be provided by light indicators or by a haptic feedback module of the ultrasound probe. Alternatively, this indication may be provided to the practitioner via the graphical interface 19.

[0116] The above description clearly illustrates that, through their various characteristics and advantages, the devices and methods presented achieve the set objectives.

[0117] In particular, the medical robot 10 uses the breathing information to position the tool guide 14 so as to follow the position of the lesion precisely and in real time, without specific management by the practitioner.

[0118] During insertion of the medical instrument 15, the lesion (particularly in soft organs) may move in the direction of the trajectory. Real-time determination of the position of the lesion on the ultrasound images and its correlation with the position of the patient marker 22 makes it possible to update in real time the trajectory of the medical instrument 15 inserted by the robotic arm. Lateral trajectory readjustments (which are generally traumatic for the anatomy of interest) are minimized. Monitoring the position of the medical instrument, during its insertion, makes it possible to compensate for any curvature of the medical instrument.

[0119] Real-time monitoring of the position of the robotic arm 13 is carried out using ultrasound images. The patient and the nursing staff are therefore not exposed to ionizing radiation during the procedure.

[0120] A lesion that is barely or not visible on ultrasound images can be detected by registration with the imaging modality that offers the best visibility.

Claims

1. A medical robot (10) for assisting a practitioner during a medical intervention for treating a lesion in an anatomy of interest of a patient (20), said medical robot (10) comprising a robotic arm (13) comprising, at a distal end, a tool guide (14) intended to guide a medical instrument (15), and a control unit (12) configured to control the robotic arm (13), the medical robot (10) being configured to cooperate with a navigation system (30) and with an ultrasound probe (40) to be positioned by the practitioner at the anatomy of interest of the patient, the control unit (12) being configured to be able to determine at any time, based on information communicated by the navigation system (30), the position of a robot marker (18) to be attached to the medical robot (10), the position of a patient marker (22) to be positioned on the patient (20) proximate to the anatomy of interest, and the position of a probe marker (41) to be attached to the ultrasound probe (40), said medical robot (10) being characterized in that: - during a modeling phase, the control unit (12) is configured to receive a plurality of ultrasound images acquired by the ultrasound probe (40) during at least one respiratory cycle of the patient and to generate, from said ultrasound images, a model for estimating the position of a target point at the lesion and the position of an entry point at the skin of the patient (20) irrespective of the moment considered in a subsequent respiratory cycle of the patient, based on the position of the patient marker (22) at said considered moment, without it being necessary to acquire an ultrasound image at said considered moment, - during a guidance phase, the control unit (12) is configured to control the robotic arm (13) in real time based on the position of the patient marker (22) such that the tool guide (14) allows guiding of the medical instrument (15) along a trajectory defined by the position of the target point and the position of the entry point associated with the position of the patient marker (22) in the model.

2. The medical robot (10) according to claim 1, wherein, during the modeling phase, the control unit (12) is configured to: - for each ultrasound image received: ∘ determine the position of the patient marker (22) and the position of the probe marker (41) at the moment that said ultrasound image was acquired by the ultrasound probe (40), ∘ obtain an analysis image, from the ultrasound image, on which the lesion is visible, ∘ determine, on the analysis image, a target point at the lesion and an entry point at the skin of the patient (20), the target point and the entry point thus defining a trajectory to be followed for the medical instrument (15), ∘ determine the position of the target point and the position of the entry point from the position of the probe marker (41), ∘ establish an association between the position of the patient marker (22), the position of the target point and the position of the entry point thus determined for said ultrasound image, - to model, from the information thus obtained for the plurality of ultrasound images, the position of the target point and the position of the entry point, based on the position of the patient marker (22), irrespective of the moment considered in a respiratory cycle of the patient.

3. The medical robot (10) according to any one of claims 1 to 2, wherein, during the guiding phase, upon insertion of the medical instrument (15), the control unit (12) is configured to regularly receive new ultrasound images acquired by the ultrasound probe (40) and to update the model based on said new ultrasound images.

4. The medical robot (10) according to claim 3, wherein, during the guiding phase, upon insertion of the medical instrument (15), the control unit (12) is configured to determine, for each new received ultrasound image, the position of the medical instrument (15), and to adjust the real-time control of the robotic arm (13) based on the position of the medical instrument (15).

5. The medical robot (10) according to any one of claims 2 to 4, wherein the analysis image is the ultrasound image, the lesion being visible on the ultrasound image.

6. The medical robot (10) according to any one of claims 2 to 4, wherein the control unit is configured to obtain the analysis image by merging the ultrasound image with a pre-operative or pre-interventional reference image on which the lesion is visible.

7. The medical robot (10) according to claim 6, wherein a radiopaque element (24) of the patient marker (22) is visible on the reference image, and the control unit (12) is configured to merge the ultrasound image with the reference image by registration based on the position of the patient marker (22) relative to the position of the probe marker (41) at the moment when the ultrasound image was acquired by the ultrasound probe (40).

8. The medical robot (10) according to any one of claims 6 to 7, wherein the reference image is a computed tomography image, a positron emission tomography image or a magnetic resonance imaging image.

9. The medical robot (10) according to any one of claims 1 to 8, wherein the ultrasound images received from the ultrasound probe (40) are B-mode ultrasound images.

10. The medical robot (10) according to any one of claims 1 to 9, wherein the control unit (12) is configured to receive and process ultrasound images acquired by the ultrasound probe (40) at a rate of at least fifteen images per second.

11. The medical robot (10) according to any one of claims 2 to 10, comprising a user interface (19) comprising a display screen enabling the practitioner to view the analysis images.

12. The medical robot (10) according to claim 11, wherein the user interface (19) comprises input means enabling the practitioner to identify, on an analysis image displayed on the display screen, a target point and / or an entry point and / or an anatomical region that is not to be traversed by the medical instrument (15).

13. The medical robot (10) according to any one of claims 11 to 12, wherein the user interface comprises an augmented reality device for superimposing the analysis images with actual images of the patient's body on the display screen.

14. The medical robot (10) according to any one of claims 1 to 13, wherein the control unit (12) is configured to compare an ultrasound image with a reference image on which the lesion is visible, and to give an indication to the practitioner of the direction in which the ultrasound probe (40) should be moved so that an ultrasound image acquired by the ultrasound probe (40) comprises an anatomical region wherein the lesion is located.

15. A medical device comprising a medical robot (10) according to one of claims 1 to 14, and a navigation system (30) and an ultrasound probe (40) for cooperating with the medical robot (10).