Medical instrument arrangement for minimally invasive examination and / or treatment of a subject, method for operating a medical instrument arrangement, computer program and electronically readable data carrier

Abstract

The invention relates to a medical instrument arrangement (1) comprising - a medical instrument (2) having a proximal operating section (3) for moving the instrument (2), - a robot (8) for moving and manipulating the medical instrument (2) at the operating section (3), - an operator acquisition device (6) for acquiring manually generated operator movement patterns of the operating section (3) of the medical instrument (2), and - a control device (7), wherein the control device (7) comprises: - a control unit (9) for controlling the robot (8) according to control commands to generate corresponding movements of the operating section (3), - a situation detection unit (12) for determining an intervention situation based on detected operator movement patterns, - a storage unit (11) for storing, in particular operator-related, recorded operator movement patterns with the associated intervention situation in a storage means (29) of the control device (7), and - a situation unit (17) for automatically selecting and / or generating a control movement pattern from stored operator movement patterns and / or evaluation movement patterns derived from these, depending on an existing intervention situation of the situation detection unit (12), and for automatically implementing the control movement pattern as control commands by means of the control unit (9).

Description

[0001] The invention relates to a medical instrument arrangement for minimally invasive examination and / or treatment of a subject, comprising - a medical instrument that can be inserted into the subject and has a proximal operating section for moving the instrument, and - a robot for moving manipulation of the medical instrument at the operating section.

[0002] The invention also relates to a method for operating such an instrument arrangement, a computer program and an electronically readable data carrier.

[0003] During the minimally invasive examination and / or treatment of a subject, in particular a patient, which will be referred to below as a minimally invasive procedure or simply as a procedure, a medical instrument, such as a catheter, a guidewire, or the like, is introduced into the subject, for example, into a blood vessel system. The medical instrument can, for example, comprise at least one sensor for examination and / or at least one treatment agent for treatment.

[0004] An example of such minimally invasive procedures is treatment and / or examination of a patient's vascular system using a catheter as a medical instrument, for example, to resolve complete vascular occlusions (CTOs), insert stents, and the like. The catheter or, more generally, the medical instrument, including a guidewire, is typically inserted through the patient's groin and from there guided to the actual procedure site, i.e., the examination site and / or treatment site. This has so far mostly been done manually, by rotating, translating (advancement / retraction), or bending the medical instrument proximally outside the subject at an operating section.For navigation, in addition to pre-interventional image datasets and positioning systems for the instrument, fluoroscopy images or other intra-interventional images are usually acquired, which can then be displayed, for example, superimposed on a pre-interventional three-dimensional image dataset. While navigation to the surgical site can prove challenging in itself, certain maneuvers that can be used for navigation, examination, and / or treatment depend particularly on the experience of the operator, i.e., the person performing the procedure. Complex movement patterns then result.

[0005] Examples of cases where complex movement patterns may occur to operate the medical instrument include navigation into small vascular branches far away from the entry point of the medical instrument into the subject, for example in neuroradiological procedures in the brain, and difficult navigation when opening total occlusions.

[0006] To simplify the performance of minimally invasive procedures, the prior art has already proposed the use of robots that take over the handling of the medical instrument at the operating section instead of the operator and manipulate the medical instrument. The robot, in turn, is controlled by the operator using appropriate control devices, such as a joystick, to generate rotational and / or translational movements or curvatures of the medical instrument. While any medical instrument can be rotated and translated, only appropriately trained medical instruments can be curved. Such control devices make it difficult to continue applying movement patterns for the medical instrument that have been learned and / or trained through experience over a long period of time.Although automated movement patterns for robots to move medical instruments have already been proposed, this only refers to a few simple movement patterns such as predefined translations or rotations of small magnitude.

[0007] In summary, it can be said that medical instruments, such as catheters, can now be controlled either directly manually or indirectly via robots, although in the latter case only a few standard movements can be used.

[0008] The invention is therefore based on the object of providing an improved, more extensive and, in particular, customizable usability of robots in minimally invasive procedures with medical instruments.

[0009] This object is achieved by a medical instrument arrangement having the features of claim 1, a method having the features of claim 13, a computer program having the features of claim 14 and an electronically readable data carrier having the features of claim 15. Advantageous further developments emerge from the subclaims.

[0010] According to the invention, a medical instrument arrangement for minimally invasive examination and / or treatment of a subject comprises: - a robot for moving manipulation of a medical instrument which can be inserted into the subject and has a proximal operating section for moving manipulation, on the operating section, - an operator detection device for detecting manually generated operator movement patterns of the operating section of the medical instrument, in particular directly at the operating section, and - a control device, the control device comprising: - a control unit for controlling the robot according to control commands to generate control movements of the operating section, - a situation detection unit for determining an intervention situation that can be assigned to recorded operator movement patterns, - a storage unit for storing, in particular operator-related, recorded operator movement patterns with the associated intervention situation in a storage means of the control device, and - a situation unit for automatically selecting and / or generating a control movement pattern from stored operator movement patterns depending on an intervention situation determined by the situation detection unit and for providing the control movement pattern as control commands for controlling the robot by means of the control unit.

[0011] The subject can be a human and / or animal patient, but also a phantom, in particular a training phantom. In particular, a different subject can be used at different times, for example, a training phantom in a first application phase of the medical instrument assembly according to the invention, namely a training phase. In particular, it can therefore be provided that the instrument assembly further comprises a phantom as a subject, on which at least some of the operator movement patterns are recorded.

[0012] The medical instrument may in particular be a catheter and / or a guide wire, although other medical instruments are of course also conceivable in order to carry out a minimally invasive procedure, i.e. an examination and / or treatment with the medical instrument.

[0013] The operator detection device, which can be understood as a recorder, is used in the aforementioned first application phase, namely the training phase, to record the manual movement performed by the operator, particularly directly at the operating section. It should be noted at this point that the operating section of the medical instrument does not have to be a fixed, immobile section of the instrument, but can change depending on the advance distance into the subject, for example, if the instrument can be gripped directly or by means of a movable gripping sleeve or the like. In particular, the operating section is therefore a portion of the medical instrument that is generally present for minimally invasive procedures and protrudes from the subject at the entry point, i.e., proximally.

[0014] The operator detection device provides detection data that describes operator movement patterns implemented by the operator. Since rotations and translations (advance / retraction) and / or curvatures of the medical instrument can typically be performed on the operating section of the medical instrument, in particular by the operator's fingers, an expedient development of the present invention provides that the detection data include a rotational path, a translational path, and / or a curvature path brought about by the operator's operating actions. Accordingly, it can be provided that the control commands include a rotational path, a translational path, and / or a curvature path for the instrument, in particular its operating section, to be implemented by the control unit.

[0015] Ultimately, and in general terms, the invention proposes transferring movement patterns of at least one operator during manual guidance of the medical instrument at the operating section to the control of the robot in such a way that these can be automatically provided depending on the situation. In other words, the robot movements are adapted to the movement patterns of the treating personnel for a specific intervention situation. The medical instrument arrangement according to the invention allows for the self-learning integration of specific movement patterns of an operator into a robotic minimally invasive intervention, thus expanding the movement repertoire of the robot. Specific movements of the at least one operator are learned together with the intervention situation and can be automatically retrieved in a corresponding intervention situation by the robot, or more precisely, the control device.

[0016] During a training phase, during manual operation of the medical instrument by an operator, particularly at the operating section itself, operator movement patterns can be learned by recording them using the operator recording device and stored using the memory unit of the control device. This not only learns the operator movement patterns, for example, a sequence of translations and rotations and / or curvatures, but also allows an intervention situation in which the operator movement pattern is used to be recorded and stored associated with the operator movement pattern. The training phase is expediently carried out without any use of the robot, but using the operator recording device (recorder device), which can be arranged, for example, at the instrument lock to the subject.

[0017] The operator detection device can be specifically designed in different ways. For example, it can have drivers, such as rollers, that move with the operating section of the medical instrument through friction or other interaction, the movement of which is measured. In a preferred embodiment, markers are provided on the operating section of the instrument, which can be detected by a corresponding, for example optical, sensor system of the operator detection device.

[0018] In a second application phase of the medical instrumentation system, which can be referred to as the recall phase, the presence of a specific intervention situation is automatically detected. A corresponding control movement pattern, which can be an operator movement pattern or derived from it, is retrieved and executed, particularly after confirmation by an operator.

[0019] In particular, the medical instrument arrangement can operate in a user-specific manner, for which purpose the operator, along with the operator movement pattern and the associated intervention situation, is also expediently stored. In particular, it can be provided that the situation unit is designed for the operator-related selection and / or generation of the control movement pattern based only on the stored operator movement patterns related to the current operator. In this way, the individual movement patterns of specific operators are learned, which the operator particularly trusts due to their own authorship and ultimately recognizes in the robot's working methods. The operator's specific experience and approach are then reflected in the robot's automatic operation according to the control movement pattern.

[0020] The learning / recording of specific operator movement patterns can be achieved in different ways. Particularly when using a phantom, explicit, targeted training can also be performed using the medical instrument arrangement. To do this, the operator can manually initiate movement recording, for example, by activating the operator detection device in a special training mode via the control device and consciously performing the corresponding movements of the operator movement pattern to be recorded, particularly using the phantom as the subject. In this context, the intervention situation can also be provided by the operator, for example, via a user interaction unit of the control device.

[0021] In preferred embodiments, however, minimally invasive interventions by the operator are recorded by the medical instrument arrangement, particularly during exercises on the phantom and / or actual minimally invasive procedures on a patient. Different approaches are also conceivable in this context.

[0022] Thus, in a first concrete approach to learning by observing the operator, it can be provided that the operator recording device is designed to continuously record the operating activity during an intervention on the subject, and the situation recording unit is designed for later situation assignment and corresponding selection of operator movement patterns from the operating activity. This means that the movements performed by the operator can be continuously recorded, stored in the storage means, and, in a later analysis, partially assigned to at least one intervention situation. In this analysis, for example, the operator can "review" the intervention and its implementation, whereby, in principle, an automatic analysis of such a recording of the operating activity can also take place during a minimally invasive intervention, particularly if additional information is stored simultaneously.

[0023] In a second specific approach to learning by observation, which is preferred over the first approach, it can be provided that the control device further comprises an activation unit for activating the operator detection device and / or for activating the storage of operator movement patterns upon fulfillment of an activation condition that evaluates a current intervention situation detected by the situation detection unit and / or a current operating pattern and / or instrument movement pattern of the instrument. This means that recording can be started automatically by the medical instrument arrangement when a corresponding activation condition exists, for example, when it has been determined that a specific intervention situation for which at least one operator movement pattern is to be recorded has occurred.Additionally or alternatively, it is also conceivable to specifically detect and store "unusual" behaviors, particularly those that deviate from the normal advance to the intervention site, particularly operating patterns or movement patterns. In this case, the operating and / or instrument movement pattern refers in particular to an operating period that ends at the current time.

[0024] In particular, if only the (final) storage is to be controlled, i.e. the operator recording device is generally active during the minimally invasive procedure, it may be expedient to use a ring buffer in which the most recently recorded recording data of the operator recording device can always be stored, for example to be evaluated by the activation condition and to be at least partially incorporated into the overall data set to be finally stored in the storage means (containing at least the operator movement pattern and the intervention situation).

[0025] However, variants are also conceivable in which the operator detection device does not have to be operated continuously, particularly if another means of recording the operator's behavior is made possible and / or at least the current intervention situation is continuously monitored. For example, an unusual movement can also be detected by imaging and the like, particularly if imaging monitoring of the minimally invasive procedure takes place, for example by fluoroscopy. Furthermore, it is also conceivable to use additional sensors, such as cameras, which monitor the minimally invasive procedure. In particular, the analysis of the operating and / or instrument movement pattern serves, as mentioned, to detect unusual behaviors that deviate from the operating and / or movement pattern for the usual advance and, if applicable, retraction.With regard to the intervention situation, the activation condition can, for example, monitor whether an intervention situation from a predefined list of intervention situations leading to activation exists, in particular its initial state.

[0026] In an expedient embodiment, it can be provided that the activation unit is designed to evaluate the activation condition by comparing the current operating and / or instrument movement pattern with at least one predetermined normal pattern, which in particular describes normal feed operation of the instrument, wherein the activation condition is fulfilled if the deviation exceeds a threshold value for each normal pattern. For example, correlation values ​​of the current operating and / or instrument movement pattern with the normal patterns can be determined, wherein the activation condition is fulfilled if a threshold value for the correlation value is undershot for all normal patterns. It is also conceivable to use a pattern classification function, in particular a trained one, which classifies an operating and / or instrument movement pattern as "normal" or "unusual".In general, it can be planned that the normal patterns are derived from training data. In other words, the operator's "normal" movements can be learned from training data, and a recording can be initiated when the current movements deviate from these.

[0027] It should be noted at this point that the situation detection unit can also use a classifier, especially a trained one, with regard to the intervention situation, which will be discussed in more detail below.

[0028] Preferably, the instrument arrangement can further comprise at least one sensor for acquiring sensor data to determine the current operating and / or instrument movement pattern by the activation unit and / or to determine the current intervention situation by the situation detection unit. For example, a camera directed at the operator's hands can be used as the sensor. However, other comparable sensor data can also be used. For example, the instrument movement pattern can also relate to the tip of the medical instrument, as visible, for example, in image data.

[0029] Finally, it should be noted that in some cases, it may also be possible for the situation detection unit to be configured to determine the intervention situation at least partially using operator input. This can be particularly useful when introducing new intervention situations and / or when determining the intervention situation is otherwise not possible. For example, a user can define a new intervention situation or intervention situation class and start recording for it manually, so to speak. This can be particularly useful for the aforementioned "dry run" on a phantom or the like.

[0030] Various approaches are also conceivable for assigning intervention situations to operator movement patterns. As already mentioned, explicit training by the operator is initially possible, whereby the operator explicitly assigns the operator movement pattern to an intervention situation.

[0031] It is also possible to observe operator procedures. For example, when activated by the activation unit, information describing the intervention situation, such as a position of the instrument and / or anatomical information, in particular a relative position of the instrument, in particular the instrument tip, to the surrounding anatomy, is already assigned to the recorded operator movement pattern. Other anatomical assignments are also possible, for example, by determining—possibly also subsequently, with the additional information recorded and assigned to the operator movement pattern—the anatomical background against which the medical instrument is located and / or whether a specific landmark, for example in an overlay, is reached.

[0032] A specific, preferred development in this context can, for example, provide that the situation detection unit is configured to detect the intervention situation for an operator movement pattern, to define a situation data set that describes at least the instrument position relative to the surrounding anatomy, in particular with respect to at least one landmark, and to determine the present intervention situation by comparing it with stored situation data sets. Overall, it is expedient if the situation detection unit is configured to use such situation data sets that define the intervention situation and describe at least the instrument position relative to the surrounding anatomy, in particular the position of the instrument tip relative to the surrounding anatomy.In particular, if the interventional situation is not yet known, meaning that no (sufficient) match is found during comparison, new interventional situations may arise. Such interventional situations may, for example, involve the tip of the medical instrument being located in front of an anatomical feature that requires more complex movement patterns, such as a bifurcation, a vascular branch, or a total occlusion.

[0033] Accordingly, general intervention situations and associated movement patterns can also include, for example, the opening of a total occlusion (CTO), navigation into a vascular branch, in particular palpation of a bifurcation and / or branch that is only imaged two-dimensionally, and navigation of small vessels (“fingertip technique”).

[0034] With regard to the situation data set, a useful further development provides that the situation detection unit is configured to store the situation data set with the associated operator movement pattern and / or the situation data set comprises image data showing the instrument in the anatomy and / or position data of a position detection device for the medical instrument. In this case, if a landmark is referenced, the landmark can be highlighted in the image data. Storing the situation data set, particularly in conjunction with the image data and / or position data, is particularly useful if the intervention situation is to be automatically determined in the second application phase (recall phase) in order to select and / or generate a corresponding control movement pattern.Then, for example, it can be provided that the medical instrument and / or the anatomical region and / or the reaching of the landmark are recognized in current image data, in order to be able to deduce the corresponding intervention situation by comparing it with stored intervention situations or, more specifically, the situation data sets. The control movement pattern can then be selected and, in particular, suggested.

[0035] It should be noted at this point that the position data of the medical instrument can preferably refer to the instrument tip as far as the intervention situation is concerned, but it is also conceivable additionally or alternatively that position data of the medical instrument include system positions, for example a current feed into the subject.

[0036] In addition to or as an alternative to the use of situation data sets and / or image analysis to determine an intervention situation, or implementing these variants, it can also be provided in an advantageous embodiment that the situation detection unit is designed to use at least one, in particular trained, classification function to recognize the intervention situation. The input data of such a trained classification function can comprise detection data from the operator detection device (for example, detection data stored in a ring buffer). However, they expediently comprise at least one piece of additional information, in particular position data from a position detection device for the medical instrument and / or the robot and / or image data from an imaging device and / or setting data of components of the instrument arrangement.In particular, the above situation data set can also be used to classify a current intervention situation. Accordingly, the output data of such a classification function generally includes classification into an intervention situation class. In this context, as in general, it may be expedient for at least one intervention situation and / or intervention situation class to be defined by an instrument position relative to the surrounding anatomy.

[0037] In general, a trained function replicates cognitive functions that people associate with other human brains. Through training based on training data (machine learning), the trained function is able to adapt to new circumstances and detect and extrapolate patterns.

[0038] Generally speaking, parameters of a trained function can be adjusted through training. In particular, supervised learning, semi-supervised learning, unsupervised learning, reinforcement learning, and / or active learning can be used. Representation learning (also known as feature learning) can also be used. The parameters of the trained function can be adjusted iteratively through multiple training steps.

[0039] For example, a trained function may include a neural network, a support vector machine (SVM), a decision tree, and / or a Bayesian network, and / or the trained function may be based on k-means clustering, Q-learning, genetic algorithms, and / or mapping rules. In particular, a neural network may be a deep neural network, a convolutional neural network (CNN), or a deep CNN. Furthermore, the neural network may be an adversarial network, a deep adversarial network, and / or a generative adversarial network (GAN).

[0040] Generally speaking, it is also possible for an intervention situation to be detected in the second application phase (recall phase) based on current movement patterns induced by the operator using the robot. Thus, the control device, in particular the situation detection unit, can detect, for example, whether the operator is attempting to execute one of their specific movement patterns, such as a "special sequence," with the robot. From this, the current intervention situation can be deduced and the control movement pattern can then be suggested.

[0041] For the concrete implementation of suitable automation upon detection of an intervention situation for which a control movement pattern can be selected and / or generated and, if necessary, suggested, a useful development of the present invention provides that the control device further comprises a trigger unit that triggers the selection and / or generation of the situation unit upon fulfillment of a trigger condition evaluating the present intervention situation of the situation detection unit. In particular, the trigger condition can check whether operator movement patterns and / or evaluation movement patterns are already stored in the storage means for the present intervention situation, to which a corresponding intervention situation is assigned.In summary, it is automatically monitored whether an intervention situation occurs for which a control movement pattern can be provided, whereby this provision is achieved automatically due to the trigger unit.

[0042] The control device can expediently comprise a user interaction unit for suggesting the control movement pattern to the operator, particularly upon triggering, and / or for receiving operator input, particularly for use in the selection process. A suggestion function that automatically opens up further, particularly customized, operating options for the robot based on the current intervention situation is particularly advantageous in this regard and also elicits confirmation from the operator.

[0043] In a particularly preferred development of the present invention, the control device can be provided with an evaluation unit for statistically evaluating multiple operator movement patterns present in the storage means for a similar intervention situation in order to determine a selectable evaluation movement pattern for this intervention situation. Thus, if multiple operator movement patterns, particularly those of the same operator, are present for an intervention situation, a statistical evaluation can be performed to derive the essence of the multiple detections, in particular the actually desired movement characteristics, and in particular to derive an optimal evaluation movement pattern that can then be selected by the situation unit for the intervention situation (and possibly the operator) or used generally.In particular, online generation is also conceivable, for example, directly before selection and / or generation in the situation unit. This allows, in particular, the determination of improved and also updateable control movement patterns, after learning from multiple processes relating to the same intervention situation.

[0044] The statistical evaluation can, for example, include averaging the duration and / or amplitude and / or course of the movement, particularly for translation and rotation and / or curvatures, for similar sections of the movement pattern. However, it is also possible to statistically determine partial movement patterns typical for the intervention situation and base the evaluation movement pattern on these.

[0045] In this context, it is particularly advantageous if the storage unit is further configured to store, in particular with a time delay, success information describing the success of the respective operator movement pattern, in particular as a continuous efficiency variable, for each operator movement pattern. The evaluation unit is configured to determine the evaluation movement pattern as an optimal movement pattern using the success information. In this embodiment, efficiency and / or treatment success, described by the success information, are taken into account, and a preferred evaluation movement pattern optimized for the respective intervention situation is created based on the success information.The success information allows the most efficient components of the operator movement patterns to be identified in the statistical evaluation, for example, partial movement patterns present or pronounced in particularly successful interventions, and to be made the subject of the evaluation movement pattern. Averaging processes within the framework of the statistical evaluation can also provide for weighting depending on the success information. Furthermore, operator movement patterns that lead to an insufficient success rate can also be excluded. In this case, the success information is particularly and preferably continuous or at least graded success information, for example, the greatest efficiency, and thus not binary success information. For example, a percentage efficiency or another score can be determined and used as success information, in particular an efficiency variable.

[0046] In this context, an expedient further development of the medical instrument arrangement provides that the evaluation unit is further configured to evaluate image data of the instrument and the surrounding anatomy recorded during and / or after execution of the operator movement pattern in order to at least partially automatically determine the success information. In other words, it is conceivable to automatically derive the success information from observation results for the procedure in which the operator movement pattern was used. An image-based evaluation is particularly suitable for this purpose. If, for example, the preferred application is an intervention in the vascular system of a patient, the image data can be used to assess, for example, whether a vessel was punctured and / or to what extent a total occlusion was removed.In principle, however, it is also conceivable to include user input and / or other information in such a postoperative evaluation.

[0047] For practical purposes, the analysis can also include evaluation movement patterns and their success information determined in a previous analysis. This is particularly useful for updates. However, success information for evaluation movement patterns used, in particular, as control movement patterns, can also be used to assess the quality of the statistical analysis and, if necessary, to improve it.

[0048] Preferably, the control device can further comprise an adaptation unit that adapts the operator movement patterns and / or evaluation movement patterns to be stored with respect to at least one movement restriction related to the robot and / or the protection of the subject. In this way, it is possible to also incorporate subject protection requirements and / or robot specifications into the learning process, so that the learned operator movement patterns and / or evaluation movement patterns can actually be executed by the robot. The control device can be "intelligent," so to speak, and take into account restrictions related to the robot and / or the subject.For example, if it is known that the robot cannot exceed a limit for the number of revolutions per second during rotation, any value of the operator movement pattern and / or evaluation movement pattern that exceeds this limit can be reduced to the limit. The same applies to feed / retract.

[0049] In addition to the medical instrument arrangement, a particularly computer-implemented method for operating a medical instrument arrangement for minimally invasive examination and / or treatment of a subject, comprising a medical instrument that can be inserted into the subject and has a proximal operating section for moving the instrument, is also proposed, wherein the method comprises the following steps: - Recording of manually generated operator movement patterns of the operating section of the medical instrument, in particular directly at the operating section, - Assigning an intervention situation to each recorded operator movement pattern, - in particular, operator-related storage of the recorded operator movement patterns with the associated intervention situation, and, - in the case of a further intervention with the medical instrument and a robot for moving manipulation of the medical instrument at the operating section, automatic selection and / or generation of a control movement pattern from stored operator movement patterns depending on the current intervention situation, and - automatic implementation of the control movement pattern as control commands for the robot.

[0050] All statements regarding the medical instrument arrangement according to the invention can be transferred analogously to the method and vice versa, so that the advantages already mentioned can also be obtained here.

[0051] A computer program according to the invention can be loaded directly into a storage means of a control device of a medical instrument arrangement and has program means such that, upon execution of the computer program in the control device, the control device is prompted to carry out the steps of a method described herein. The computer program can be stored on an electronically readable data carrier according to the invention, which thus comprises control information stored thereon, which comprises a computer program according to the invention and is configured such that, upon use of the data carrier in a control device of an instrument arrangement, the latter is configured to carry out a method described herein. The data carrier is, in particular, a non-transient data carrier, for example a CD-ROM.

[0052] Further advantages and details of the present invention will become apparent from the exemplary embodiments described below and from the drawings. In the drawings: Fig. 1 a schematic diagram of a medical instrument arrangement according to the invention, Fig. 2 the functional structure of a control device of the instrument arrangement of the Fig. 1, Fig. 3 is a flowchart of an embodiment of an operating method according to this disclosure, Fig. 4 an illustration of an intervention situation, Fig. 5 an exemplary operator movement pattern, and Fig. 6 schematically shows the implementation of the operator movement pattern as a control movement pattern for a robot.

[0053] Fig. Figure 1 shows a schematic diagram of a medical instrument assembly 1 according to the invention for performing minimally invasive procedures with a medical instrument 2, for example, a catheter or a guidewire. The medical instrument 2 has a proximal operating section 3, in particular the section of the instrument 2 located just outside a subject. At an entry point, the instrument 2 can be introduced into a subject, for example, a patient or a phantom 5, via an instrument lock 4.

[0054] The medical instrument 2 can be operated directly by an operator at the operating section 3, for example, for advancement / retraction (translation) and with regard to the rotation and / or curvature of the medical instrument 2. In order to be able to record such operator movement patterns induced by an operator, the medical instrument arrangement 1 further comprises an operator detection device 6, which has corresponding measuring means, for example, to detect markers on the instrument 2 or otherwise track the movement at the operating section 3. When the instrument 2 is manually guided at the operating section 3 by an operator, detection data describing the operator movement pattern can be recorded and forwarded to a control device 7.The Phantom 5, for example, can be used to specifically record operator movement patterns; it is of course also possible to record operator movement patterns during minimally invasive procedures on other subjects.

[0055] The medical instrument arrangement 1 also comprises a robot 8, in the present case, for example, a so-called vascular robot, which can be coupled to the operating section 3 in order to effect movements of the medical instrument 2 according to control commands which can be based on an operator input on an operating means or another source, in the present case again rotations and translations and / or curvatures.

[0056] Fig. Figure 2 shows the functional structure of the control device 7 in more detail. Fig. 2 first introduces the various functional units of the control device 7, before their interrelated operation is explained using the method of Fig. 3 will be explained in more detail.

[0057] The control device 7 initially has a control unit 9 which is designed to control the robot 8 according to control commands to generate corresponding movements, here translations and rotations and / or curvatures.

[0058] In the present exemplary embodiment, the control device 7 further comprises an activation unit 10, which controls the activity of various other functional units, in particular also the operator detection device 6 and a memory unit 11 of the control device 7, in particular depending on an activation condition. The activation unit 10 has access to various other functional units, for example, a situation detection unit 12, an interface 13 for receiving additional information, in particular image data from an imaging device for monitoring a minimally invasive procedure and / or position data from a positioning system for the medical instrument 2, and a user interaction unit 14.

[0059] The situation detection unit 12 just mentioned is configured to detect an intervention situation on various occasions, which, as will be explained in more detail below, can be based on various data or information. In particular, it is capable of assigning intervention situations to detected operator movement patterns.

[0060] The storage unit 11 is designed to store various, mutually associated information in a storage means 29 of the control device 7, in particular for storing a recorded operator movement pattern together with the associated intervention situation, in particular as a situation data record, and operator information describing the operator who performed the operator movement pattern. Such overall data records consisting of the operator movement pattern (in particular the corresponding recorded data), intervention situation (in particular the situation data record), and operator (in particular the operator information) can also be stored at later times by the storage unit 11 with success information describing the treatment success and / or the efficiency of the minimally invasive intervention in which the respective operator movement pattern was recorded.

[0061] Such success information can be provided, for example, by an evaluation unit 15, which can evaluate image data and / or other material (received via interface 13) relating to the minimally invasive procedure to determine the success information. Other sources for at least partially determining the success information are the user interaction unit 14 or directly via interface 13.

[0062] The evaluation unit 15 is further designed to statistically evaluate operator movement patterns assigned to the same intervention situation and the same operator, taking into account the success information, in order to determine evaluation movement patterns which are also stored in the storage means 29, in particular by the storage unit 11.

[0063] The control device 7 further comprises a trigger unit 16, which is active when the robot 8 is in use and evaluates intervention situations detected by the situation detection unit 12 to determine whether a trigger condition is met. The trigger condition ultimately indicates that an intervention situation exists, in particular the beginning of an intervention situation, in which a recorded operator movement pattern or an evaluation movement pattern could expediently be implemented by the robot 8 as a control movement pattern via control commands by the control unit 9. This control movement pattern is selected by a situation unit 17, which then forwards it accordingly to the control unit 9.

[0064] The interaction of the functional units is now examined with regard to the procedure of Fig. 3. Accordingly, the use of the medical instrument assembly 1 can be divided into two application phases 18, 19, namely a first application phase 18 (training phase) and a second application phase 19 (recall phase).

[0065] In the first application phase 18, in a step S0, during a minimally invasive procedure on a subject, for example, the Phantom 5 or another subject, the activation unit 10 monitors whether the activation condition is met. Several possible prerequisites exist for this. For example, it is possible for an operator to manually initiate a motion recording via the user interaction unit 14, for example, when working with the Phantom 5, which otherwise makes analysis difficult.

[0066] It is preferred to monitor the minimally invasive procedure in the activation unit 10. In a simple example, the activation unit 10 can, for example, check with regard to the activation condition whether a current intervention situation determined by the situation detection unit 12 is included in a list of intervention situations for which operator movement patterns are to be recorded. It is also possible, particularly if the operator detection device 6 provides detection data throughout the entire minimally invasive procedure, which can be stored, for example, in a ring buffer.It may then be provided to compare a current operating and / or instrument movement pattern with normal patterns that describe a normal feed operation of the instrument, for example by determining correlation values ​​for the normal patterns and the activation condition is met when a threshold value is undershot for all normal patterns.

[0067] However, current operating and / or instrument movement patterns can alternatively or additionally be recorded in other ways, particularly with special sensors, such as a camera aimed at the operator's hands. Image data relating to the movement of the instrument tip, as well as position data of the medical instrument, can also be evaluated to detect unusual operating and / or instrument movement patterns that deviate from normal patterns.

[0068] It should be noted that, in addition to the exemplary embodiment described here, it is also conceivable in principle to record the acquisition data continuously and later assign it to intervention situations and extract the operator movement patterns for corresponding time intervals. However, this is less preferred.

[0069] After the activation condition is met in step S0, the detection data for the corresponding operator movement pattern are recorded in step S1. In step S2, an intervention situation is then assigned to the operator movement pattern using the situation detection unit 12. It should be noted that the order of steps S1 and S2 can certainly be reversed, since, for example, the intervention situation present or how it should be described may already be known during activation.

[0070] While it is fundamentally conceivable to determine an intervention situation based on operator input, in particular using the user interaction unit 14, this is primarily expedient, as already described, for the use of the phantom 5. Preferably, the intervention situation is recorded automatically. In this context, it is particularly advantageous to compile a situation data set that describes the position of the medical instrument 2, in particular of its instrument tip, relative to the surrounding anatomy. This can, for example, explicitly be image data, for example image data used to monitor the minimally invasive procedure. It is therefore conceivable, for example, if the activation condition is met, to store image data and position data of the medical instrument 2 in the situation data set and to assign this to the operator movement pattern to be recorded.However, it is also conceivable to explicitly determine the anatomical background against which medical instrument 2 is located or whether a specific landmark, which can also be displayed in an overlay, is reached by the instrument. Such a landmark could be, for example, a complete vascular occlusion, a bifurcation, or a vascular branch.

[0071] Generally, it should be noted at this point that various variants for determining the specific intervention situation are conceivable, especially in parallel. For example, a trained classification function can be used and / or a comparison can be performed when situation data sets are used.

[0072] In a step S3, the storage unit 11 stores the operator movement pattern, in particular the corresponding detection data, the intervention situation, in particular the situation data record, and operator information describing the operator, since in the present case the work is carried out in an operator-specific manner, as a total data record in the storage means 29.

[0073] The steps S1 to S3 are explained schematically by the Fig. 4 and Fig. 5. Fig. Figure 4 shows the medical instrument 2 inserted into a subject 20 in an interventional situation, which, in this case, is navigated in a blood vessel system 21. In the operating section 3, a hand 22 of the operator is shown, which causes movements of the operating section 3 and thus of the medical instrument 2, in this case translations (see arrow 23) and rotations (see arrow 24). Fig. 5 shows, as an example, the curves 25, 26 of rotation (R(t)) and translation (T(t)) for an operator movement pattern.

[0074] Returning to Fig. 3, in a step S4, particularly after completion of the minimally invasive procedure, the evaluation unit 15 determines the success information, for example, by automatically evaluating image data and / or other additional information describing the state after execution of the operator movement pattern. The success information relating to the operator movement pattern describes its efficiency and / or treatment success, preferably continuously, for example, as a percentage / efficiency value. For example, the success information can indicate the extent to which a CTO was resolved and / or how quickly normal navigation could continue and / or whether or to what extent unwanted effects, such as punctures, occurred. In step S5, the storage unit stores the success information in the overall data set of the operator movement pattern.

[0075] As indicated by the dashed arrow 27, further operator movement patterns can now be recorded, in particular for the same intervention situation and the same operator.

[0076] In a step S6, the evaluation unit 15 then determines evaluation movement patterns for all intervention situations and operators for which multiple operator movement patterns exist. For this purpose, the multiple operator movement patterns are statistically evaluated, with the success information being used accordingly. In particular, an optimal movement pattern is determined for the intervention situation and the operator, which, from a statistical perspective, should produce the best treatment success and / or the best efficiency. The evaluation can include weighted averaging and / or the identification of efficient and / or characteristic partial movement patterns. In particular, when updating determined evaluation movement patterns after they have been used, these can also be included in the then updated evaluation along with corresponding success information.

[0077] In a step S7, the evaluation movement patterns with the associated intervention situations and operators are also stored in the storage means 29 by means of the storage unit 11.

[0078] In the second application phase 19, the so-called “recall phase”, in a step S8, during an intervention with the robot 8, the situation detection unit 12 continuously monitors which intervention situation is currently present based on various information, in particular image data and position data of the medical instrument 2.

[0079] This is used by the trigger unit 16 in step S9 to check, within the trigger condition, whether usable operator movement patterns or evaluation movement patterns for this intervention situation are present in the storage means 29, in particular those for the current operator. If this is the case, i.e., the trigger condition is met, in a step S10 the situation unit 17 selects the evaluation movement pattern appropriate for the intervention situation and the operator, or, if no evaluation movement pattern is present, the operator movement pattern as the control movement pattern.

[0080] In a step S11, the control movement pattern is suggested to the operator for execution via the user interaction unit 14. If the operator wishes to execute the movement pattern, the control unit 9 executes the control movement pattern using control commands in step S12. Of course, repetitions can also occur in the application mode 19, particularly within a single minimally invasive procedure.

[0081] This second application phase is in Fig. 6 is schematically illustrated, where now instead of the hand 22 the robot 8 carries out the translations (arrow 23) and the rotations (arrow 24), for example using the courses 25, 26.

[0082] Finally, it should be noted that the control device 7, compare Fig.2, may optionally also comprise an adaptation unit 28 which adapts operator movement patterns and / or evaluation movement patterns to restrictions regarding the robot 8 and to protect the subject 20.

[0083] Although the invention has been illustrated and described in detail by the preferred embodiment, the invention is not limited to the disclosed examples and other variations may be derived therefrom by those skilled in the art without departing from the scope of the invention.

[0084] Regardless of the grammatical gender of a particular term, persons with male, female or other gender identities are included.

Claims

[1] Medical instrument arrangement (1) for minimally invasive examination and / or treatment of a subject (20), comprising - a robot (8) for moving manipulation of a medical instrument (2) which can be introduced into the subject (20) and has a proximal operating section (3) for moving manipulation, on the operating section (3), - an operator detection device (6) for detecting manually generated operator movement patterns of the operating section (3) of the medical instrument (2), in particular directly on the operating section (3), and - a control device (7), wherein the control device (7) comprises: - a control unit (9) for controlling the robot (8) according to control commands to generate control movements of the operating section (3), - a situation detection unit (12) for determining an intervention situation to be assigned to detected operator movement patterns, - a storage unit (11) for storing, in particular operator-related, recorded operator movement patterns with the associated intervention situation in a storage means (29) of the control device (7), and - a situation unit (17) for automatically selecting and / or generating a control movement pattern from stored operator movement patterns as a function of an intervention situation determined by the situation detection unit (12) and for providing the control movement pattern as control commands for controlling the robot (8) by means of the control unit (9). [2] Medical instrument arrangement according to claim 1, characterized by that the control commands comprise a rotation course (25) and translation course (26) and / or curvature course for the medical instrument (2), in particular its operating section (3), to be implemented by the control unit (9). [3] Medical instrument arrangement according to claim 1 or 2, characterized by that the situation unit (17) is designed for operator-related selection and / or generation of the control movement pattern on the basis only of the stored operator movement patterns related to the current operator. [4] Medical instrument arrangement according to one of the preceding claims, characterized by that the control device (7) further comprises an activation unit (10) for activating the operator detection device (6) and / or for activating the storage of operator movement patterns upon fulfillment of an activation condition which evaluates a current intervention situation detected by the situation detection unit (12) and / or a current operating pattern and / or instrument movement pattern of the instrument (2). [5] Medical instrument arrangement according to claim 4, characterized byin that the activation unit (10) is designed to compare the current operating and / or instrument movement pattern with at least one predetermined normal pattern, which in particular describes a normal feed operation of the instrument (2), in order to evaluate the activation condition, wherein the activation condition is met for each normal pattern if the deviation exceeds a threshold value. [6] Medical instrument arrangement according to one of the preceding claims, characterized by that the situation detection unit (12) is designed to detect the intervention situation for an operator movement pattern for defining a situation data set which describes at least the instrument position relative to the surrounding anatomy, in particular with respect to at least one landmark, and to determine the present intervention situation by comparing it with stored situation data sets. [7] Medical instrument arrangement according to one of the preceding claims, characterized by that the situation detection unit (12) is designed to use at least one, in particular trained, classification function to detect the intervention situation. [8] Medical instrument arrangement according to one of the preceding claims, characterized by that the control device (7) further comprises a trigger unit (16) which triggers the selection and / or generation of the situation unit (17) when a trigger condition evaluating the present intervention situation of the situation detection unit (12) is fulfilled. [9] Medical instrument arrangement according to one of the preceding claims, characterized bythat the control device (7) has an evaluation unit (15) for the statistical evaluation of a plurality of operator movement patterns present in the storage means (29) for a same intervention situation in order to determine a selectable evaluation movement pattern for this intervention situation. [10] Medical instrument arrangement according to claim 9, characterized by that the storage unit (11) is further designed to store, in particular with a time delay, success information describing the success of the respective operator movement pattern, in particular as a continuous efficiency variable, for each operator movement pattern, wherein the evaluation unit (15) is designed to determine the evaluation movement pattern as an optimal movement pattern using the success information. [11] Medical instrument arrangement according to claim 10, characterized bythat the evaluation unit (15) is further designed to evaluate image data of the instrument (2) and the surrounding anatomy recorded during and / or after the execution of the operator movement pattern in order to at least partially automatically determine the success information. [12] Medical instrument arrangement according to one of the preceding claims, characterized by that the control device (7) further comprises an adaptation unit (28) which adapts operator movement patterns to be stored with regard to at least one restriction of movement related to the robot (8) and / or the protection of the subject (20). [13] Method for operating a medical instrument arrangement (1) for minimally invasive examination and / or treatment of a subject (20), comprising a medical instrument (2) which can be introduced into the subject (20) and has a proximal operating section (3) for moving the instrument (2), the method comprising the following steps: - detection of operator movement patterns of the operating section (3) of the medical instrument (2) generated manually, in particular directly on the operating section (3), - Assigning an intervention situation to each recorded operator movement pattern, - in particular, operator-related storage of the recorded operator movement patterns with the associated intervention situation, and, - in the case of a further intervention with the medical instrument (2) and a robot (8) for moving manipulation of the medical instrument (2) on the operating section (3), automatic selection and / or generation of a control movement pattern from stored operator movement patterns depending on a determined intervention situation and - Providing the control movement pattern as control commands for the robot (8). [14] Computer program which, when executed on a control device (7) of an instrument arrangement (1), causes the latter to carry out the steps of a method according to claim 13. [15] An electronically readable data carrier on which a computer program according to claim 14 is stored.

Images