DE102019214560A1 - Method for the automatic selection of medical structures and devices - Google Patents

Abstract

What is claimed is a method for the automatic selection of a medical structure intended for targeted registration from a planning image data set of an examination area of a patient with the following steps: provision of the planning image data set, segmentation of the planning image data set with regard to imaged medical structures in the event that there is no pre-segmentation , Automatic identification of marked landmarks in the planning image data set, comparison of the positions of the landmarks with the positions of the segmented medical structures and identification of structures to which landmarks are positionally assigned, selection of at least one segmented medical structure to which landmarks are assigned, and further use of the selected segmented medical structure, in particular for a registration with at least one further image data set, which was recorded by means of an X-ray device will be.

Description

The invention relates to a method for the automatic selection of a medical structure provided for a specific registration from a planning image data set of an examination area of a patient according to claim 1 and a device for performing such a method according to claim 11.

In order to plan certain surgical interventions under X-ray monitoring, so-called landmarks are set manually or automatically on previously recorded (e.g. two-dimensional or three-dimensional) planning X-ray images. Exemplary surgical interventions in which such (2D) planning are used are drilling holes in bones, e.g. the reconstruction of the medial patellofemoral ligament (MPFL), the replacement of the cruciate ligament, removal of bone parts or a drilling in the femur (thigh bone). If the planning information is to be transferred to intra-operative fluoroscopic x-ray images, a 2D / 2D registration of the planning x-ray images on the fluoroscopic images is generally necessary, since e.g. the patient has moved between the exposures or the planning X-ray images were taken with another device.

If there is a (e.g. 2D / 2D or 3D / 3D) registration of the complete images on top of each other, then the details and e.g. Inaccuracies can also occur in the affected medical structure intended for the procedure (e.g. bones), which means that the procedure may not be carried out optimally. In order to carry out a specific registration of the medical structure affected by the intervention, an operator must manually instruct a segmentation of the structure in question in order to be able to carry out the registration using this structure. To do this, the operator must know in advance which structure is exactly affected. Since this is not always the case in the clinical process, since different workflow steps can be carried out by different people, a general approach that excludes errors is desirable.

It is the object of the present invention to provide a method for the automatic selection of a medical structure provided for targeted registration from a planning image data set of an examination area of a patient; Furthermore, it is an object of the invention to provide a device suitable for carrying out the method.

The object is achieved according to the invention by a method for the automatic selection of a medical structure provided for a targeted registration from a planning image data set of an examination area of a patient according to claim 1 and by a device according to claim 11. Advantageous embodiments of the invention are each the subject of the associated Subclaims.

The method according to the invention for the automatic selection of a medical structure provided for targeted registration from a planning image data set of an examination area of a patient has the following steps: Provision of the planning image data set, segmentation of the planning image data set with regard to imaged medical structures in the event that there is no pre-segmentation , Automatic identification of marked landmarks in the planning image data set, comparison of the positions of the landmarks with the positions of the segmented medical structures and identification of structures to which landmarks are positionally assigned, selection of at least one segmented medical structure to which landmarks are assigned, and further use of the selected segmented medical structure, in particular for a registration with at least one further image data set, which is recorded by means of an x-ray device is taken. The method automatically selects one (or more) medical structure (s) provided for an intervention, without requiring manual input by an operator, in order to enable registration of the structure concerned. By comparing the landmarks set in the planning phase with the existing medical structures, it can be determined automatically with high reliability which structure will be relevant during an intervention. In this way, planning image data and other X-ray image data can be automatically registered to this relevant structure, quickly and with high quality, as soon as the procedure begins and further X-ray image data (e.g. live X-ray images) are recorded. The targeted registration of the image data with regard to the relevant medical structure ensures that the intervention can be carried out with the greatest possible precision.

According to one embodiment of the invention, that segmented medical structure is selected to which most of the landmarks are assigned. This can be particularly advantageous when there are a large number of landmarks. Several structures can also be selected. In order to be able to make a reliable selection, for example in the case of unclear assignments, additional information is used to make the selection. Such information can be, for example, information about previously performed selections, information about planned interventions, or user input can be used.

According to a further embodiment of the invention, the planning image data set and the at least one further image data set are formed from two-dimensional image data sets. In this case, a 2D / 2D registration takes place. While the planning image data are recorded by an X-ray device or another modality, the other image data can be live X-ray images (fluoroscopic images) e.g. a mobile or permanently installed C-arm X-ray machine.

According to a further embodiment of the invention, the planning image data set and the at least one further image data set are formed from three-dimensional image data sets. A 3D / 3D registration takes place.

The medical structures are expediently formed by bone structures, organs or hollow organs. The marked landmarks can e.g. be formed by planning points or crosses, drilling points, markers or other easily recognizable markings.

According to a further embodiment of the invention, the selected segmented medical structure is also used to optimize the image quality. In this case, the optimization is carried out in such a way that especially the selected structure is optimally displayed and the other structures are not taken into account, or only to a lesser extent. For such an optimization e.g. known methods of image processing are used, e.g. Noise reduction process or filtering. In addition, the segmented medical structure can be used to optimize the recording of the further X-ray image data, e.g. by choosing the exposure point or the dose.

According to a further embodiment of the invention, at least one pre-trained machine-learning algorithm is used to carry out the method. The latter can support the selection by e.g. is pre-trained for various surgical interventions, or which landmarks are to be assigned to which medical structures in the corresponding interventions.

The invention comprises a device for carrying out a method according to the invention, having an x-ray device for recording image data sets, an image processing unit for processing image data sets, a computing unit for performing arithmetic operations, a memory unit for storing data sets, a system controller for controlling the device and a display unit for Display of image data. The X-ray device can e.g. be designed as a C-arm x-ray device, in particular a mobile C-arm x-ray device with an equipment trolley.

• 1 eine Abfolge der Schritte des erfindungsgemäßen Verfahrens;
• 2 eine Ansicht eines Planungs-Röntgenbildes mit medizinischen Strukturen und Landmarken;
• 3 eine Ansicht eines Planungs-Röntgenbildes mit ausgewählter medizinischer Struktur; und
• 4 eine Ansicht einer Vorrichtung zur Durchführung des Verfahrens.

The invention and further advantageous configurations according to features of the subclaims are explained in more detail below with reference to exemplary embodiments shown schematically in the drawing, without thereby restricting the invention to these exemplary embodiments. Show it:

• 1 a sequence of the steps of the method according to the invention;
• 2 a view of a planning X-ray image with medical structures and landmarks;
• 3 a view of a planning X-ray image with a selected medical structure; and
• 4th a view of a device for performing the method.

In the 1 a sequence of the most important steps of the automatic method according to the invention is shown. This is particularly advantageous as preparation for a simple surgical procedure under X-ray monitoring, eg drilling in bones, reconstruction of the medial patellofemoral ligament (MPFL), replacement of the cruciate ligament, drilling in the femur (thigh bone). In a first step 10 a planning image data set with medical structures of a patient mapped thereon is first provided. The planning image data set can be, for example, a two-dimensional x-ray image previously recorded by means of an x-ray device, in which a doctor has noted landmarks for planning the operation. For example, the doctor can use a known planning tool to plan operations. The landmarks can be in various forms, for example as marked points, lines, circles or sections. The planning image data set can, for example, be retrieved from a storage unit or from a database and / or transmitted by a communication unit. The planning image data record can have been recorded by an X-ray device, alternatively also by another imaging device (MR, ultrasound device, etc.). Medical structures are to be understood as meaning, for example, bones, organs, hollow organs or other connected anatomical structures of the patient. In the 2 is an exemplary planning X-ray image 20th with illustrated medical structures 21st (e.g. bones) and marked landmarks 22nd shown in the form of drilling sites.

In a second step 11 In the event that there is no pre-segmentation, an automatic segmentation of the planning image data set is carried out with regard to imaged medical structures. Such segmentation methods are generally known and are often used for recognizing medical structures. For example, an image processing unit can be used to carry out the segmentation. If the planning image data set has already been pre-segmented, additional segmentation is generally not necessary. The segmentation results in segmented medical structures. For the in 2 Planning X-ray image shown 20th there is segmentation of the three bones shown.

In a third step 12 the landmarks previously marked, for example by a doctor, for an operation planning, become in the planning image data set 22nd automatically identified. The image processing unit, for example, can also be used for this. It can only be a landmark 22nd are present or even a large number of landmarks, depending on what type of surgical intervention is planned. In a fourth step 13 the positions of the landmarks are then automatically compared with the positions of the segmented medical structures and those medical structures are identified to which landmarks are positionally assigned. For example, the in 2 The planning image data set shown determines that the three landmarks 22nd in the area of the largest of the three structures 21st are located. It is possible that the landmarks only one med. Structure are assigned, but it is also possible that several medical structures are positionally assigned landmarks.

In a fifth step 14th at least one segmented medical structure to which landmarks are assigned is then selected automatically or at least semi-automatically. The selection is made according to one or more selection criteria. There are different options for selection criteria. The selection can be made, for example, by means of a computing unit. In the simplest case, a segmented medical structure is selected, specifically the one to which most of the landmarks are assigned. In the 3 e.g. fulfills the selected structure 23 (Bones) the selection criterion because all three existing landmarks 22nd are on this. Several segmented structures can also be selected, for example if landmarks are located in the area of several segmented structures. If it has previously been determined to select only a single segmented structure, it is often advantageous to select the segmented structure with the most landmarks. However, since there is also the problem that an image point in an image data set can be assigned to different regions, since X-ray images can contain superimpositions of structures along the beam direction, a different procedure for the selection can also be useful. For example, previously stored planning image data sets and the procedure used there can also be used by selecting the most likely segmented structure. This is possible because planning is often done in the same way and the landmarks and their arrangement are similar from plan to plan. In such a case, a pre-trained machine learning algorithm can also be used, for example, or a known database can also be used. It can also be provided that a user input of an operator is taken into account, for example in order to make a selection between two possible segmented structures.

The selected segmented structure is then used in a sixth step 15th further used, in particular for registration with a further image data set, such as a live fluoroscopic x-ray image. For example, during surgical interventions, a series of live fluoroscopic x-ray images recorded live with a C-arm x-ray device (eg mobile C-arm x-ray device) is created to monitor the procedure while the procedure is taking place. As soon as the first x-ray image has been recorded, a registration can take place between the planning data record and the x-ray image along the selected segmented structure, that is to say the structure on which the intervention takes place. The registration along the selected structure ensures an optimal overlay between the planning image data set and live image data sets. The targeted registration of the image data with regard to the structure relevant for the procedure ensures that the procedure can be carried out with the greatest possible precision.

In addition, the selected segmented structure can also be used to optimize the image quality of the planning image data set. E.g. known methods of image processing can be used, e.g. Process for noise minimization or filtering, which are optimized along the selected structure. In addition, the segmented medical structure can be used to optimize the recording of the further X-ray image data, e.g. by choosing the exposure point or the dose.

In the 4th is a device for performing the procedure with a C-arm X-ray machine 25th shown. It can be, for example, a mobile or a permanently installed C-arm X-ray device 25th act. The device has an image processing unit 27 and a computing unit 28 on. The image processing unit 27 can for example be designed for segmenting image data sets and for identifying landmarks. The arithmetic unit 28 can be designed to compare positions of the landmarks with the positions of the segmented medical structures and to identify structures to which landmarks are positionally assigned. In addition, the computing unit 28 be designed to select at least one segmented medical structure to which landmarks are assigned. The computing unit can, for example, have a processor with software or even a machine-learning algorithm. The device also has a memory unit 29 for the storage of data sets, e.g. planning image data sets. The device is controlled by a system controller 26th controlled. In addition, the device can have a display unit 30th (e.g. monitor or touchscreen) and an input unit 31 (Keyboard or touchscreen). The planning on the planning image data set can also have already been created by means of the device, for example with the aid of a planning tool (for example software). The planning image data set itself can previously be sent to the C-arm x-ray device 25th or another imaging device.

The invention can be briefly summarized as follows: What is claimed is a method for the automatic selection of a medical structure provided for a targeted registration from a planning image data set of an examination area of a patient with the following steps: provision of the planning image data set, segmentation of the planning image data set with regard to the mapped medical structures in the event that there is no pre-segmentation, automatic identification of marked landmarks in the planning image data set, comparison of the positions of the landmarks with the positions of the segmented medical structures and identification of structures to which landmarks are positionally assigned, selection of at least one segmented medical one Structure to which landmarks are assigned, and further use of the selected segmented medical structure, in particular for a registration with at least one wide ren image data set, which is recorded by means of an X-ray device.

Claims (12)

Method for the automatic selection of a medical structure provided for a specific registration from a planning image data set of an examination area of a patient with the following steps: • Provision of the planning image data set, • Segmentation of the planning image data set with regard to mapped medical structures in the event that there is no pre-segmentation, • Automatic identification of marked landmarks in the planning image data set, • comparing the positions of the landmarks with the positions of the segmented medical structures and identifying structures to which landmarks are positionally assigned, • Selection of at least one segmented medical structure to which landmarks are assigned, and • Further use of the selected segmented medical structure, in particular for a registration with at least one further image data set, which is recorded by means of an X-ray device. Procedure according to Claim 1 , the segmented medical structure to which the most landmarks are assigned is selected. Method according to one of the preceding claims, wherein the planning image data set and the at least one further image data set are formed from two-dimensional image data sets. Method according to one of the Claims 1 or 2 , wherein the planning image data set and the at least one further image data set are formed from three-dimensional image data sets. Method according to one of the preceding claims, wherein the medical structures are formed by bone structures, organs or hollow organs. Method according to one of the preceding claims, wherein the marked landmarks are formed by planning points, drilling locations, markers or other markings. Method according to one of the preceding claims, wherein the selected segmented medical structure is used to optimize the image quality. Method according to one of the preceding claims, wherein, for the selection of the at least one structure, information about previously made selections or planned interventions is additionally used. Method according to one of the preceding claims, wherein an accepted user input is additionally used to select the at least one structure. Method according to one of the preceding claims, wherein at least one pre-trained machine learning algorithm is used to carry out the method. Device for carrying out a method according to one of the Claims 1 to 10 , having an x-ray device (25) for recording image data sets, an image processing unit (27) for processing image data sets, a computing unit (28) for performing arithmetic operations, a memory unit (29) for storing data sets, a system controller (26) for controlling the Apparatus and a display unit (30) for displaying image data. Device according to Claim 11 , wherein the x-ray device is designed as a C-arm x-ray device, in particular a mobile C-arm x-ray device with an equipment trolley.

Images