KR20200068072A - Method and system for optimizing surgery planning and clinical application - Google Patents

Abstract

Provided is a system for optimizing a surgery plan and goal setting to set a surgery target portion based on deep learning, and a method thereof. According to one embodiment of the present invention, the system for optimizing a surgery plan and goal setting comprises one or more memories and a processor. The processor uses a deep learning algorithm to perform semantic segmentation of an anatomical structure and image selection, selects and targets an automatic target, which is the surgery goal portion, from a semantic segmentation result, calculates treatment-related factors, such as a distance between the automatic target and a good outcome target and the like, based on conformity assessment, and uses the factors, such as the calculated distance and the like, to train a semantic segmentation algorithm and a targeting algorithm based on post-surgery outcome optimization.

Description

Surgical planning and goal setting optimization system and method{METHOD AND SYSTEM FOR OPTIMIZING SURGERY PLANNING AND CLINICAL APPLICATION}

The present invention relates to an operation planning and goal setting optimization system and method, and more specifically, to set a clinical application authority for a plan set by artificial intelligence, and to set a surgical plan and goal to optimize the goal setting based on surgical prognosis It relates to an optimization system and method.

Machine learning algorithms can be divided into supervised learning and unsupervised learning depending on the learning method. In the supervised learning, labeled data is used for optimization of data, and in self-learning, unlabeled data is used.

When using machine learning algorithms for automatic analysis of electrophysiological signals such as existing image analysis, brain wave analysis, and other bio-derived data, in the case of supervised learning, the labeled electrophysiological data was determined by experts in the field. Since necessary results have been derived based on the results, the larger the amount of data becomes, the more time consuming and expensive the method based on expert judgment is. There were problems. According to the existing method, experts have provided data for supervising machine learning by directly marking pathological and non-pathological activities in brain waves and the like.

In the past, when surgical outcome data related to prognosis after surgery were collected, medical experts directly assigned an analysis method (or diagnosis method) and a detection threshold or criteria to be applied in the analysis method, and the specified analysis method and Since the surgical result data was analyzed according to the detection threshold, it was too dependent on expert judgment. In particular, it has been pointed out that according to the above method, since a diagnostic method for analyzing data is determined solely by the expert's opinion, it is fundamentally impossible to develop a method with a higher accuracy than the expert's opinion.

Outcome-based supervised learning without analyzing data by applying machine learning algorithms, etc., after data on surgical results or treatment results become expert labeled data If it is developed to determine the treatment plan, such as data analysis through human intervention, and surgery through this, it can not only prevent false diagnosis due to misdiagnosis by experts, but also accumulate relevant data, which is theoretically more accurate to experts. It will be possible to develop an analysis method with a higher accuracy than the judgment by.

An object of the present invention is to set a surgical target site based on deep learning.

Another object of the present invention is to optimize the surgical plan based on the prognosis of patients.

According to an embodiment of the present invention, in a surgical target area setting and surgical plan optimization system including one or more memories and processors, the processor performs semantic segmentation and image selection of an anatomical structure using a deep learning algorithm Then, targeting by selecting an automatic target that is a surgical target site from the semantic segmentation result, calculating a distance between the automatic target and a good prognosis target by conformity assessment, and calculating the calculated distance to the semantic segmentation algorithm and the Provided is a surgical target area setting and surgical plan optimization system used to train a targeting algorithm.

In the present invention, the training may be training to minimize the distance between the automatic target based on the deep learning algorithm and the good prognosis target.

In the present invention, the trained results can be used or verified in prognostic data sets of different patients.

In the present invention, the threshold for distinguishing the good prognosis may be different for each disease.

According to an embodiment of the present invention, a dip selector for selecting an image slice to which a target as a surgical target site belongs; A deep targeter targeting the automatic target that is the surgical target site based on the anatomical structure semantic segmentation using a deep learning algorithm; A deep trajector searching for a safe trajectory for surgery based on semantic segmentation information in the anatomical structure; An application unit that verifies the targeting result by user input and then establishes a surgical plan; Including, the applicator calculates the distance between the automatic target and the good prognosis target by conformity assessment, and uses the calculated distance to train the semantic segmentation algorithm and the targeting algorithm, the surgical plan A system for optimizing surgical target sites and optimizing surgical plans is provided.

In the present invention, the deep targeter grasps the boundary between the subthalmic nucleus and the red nucleus through semantic segmentation of an anatomical structure through a fully convolutional neural network, The surgical target site can be targeted.

In the present invention, the deep trajector can search the safe trajectory by semantic segmentation of anatomical structures such as sulci, gyrus, and ventricle.

According to an embodiment of the present invention, performing semantic segmentation and image selection of an anatomical structure using a deep learning algorithm; Selecting and targeting an automatic target that is a surgical target site from the semantic segmentation result; Calculating a distance between the automatic target and a good prognosis target by conformity assessment; Using the calculated distance to train the semantic segmentation algorithm and the targeting algorithm; A method for setting a surgical target site and optimizing a surgical plan is provided.

In the present invention, the training may be training to minimize the distance between the automatic target based on the deep learning algorithm and the good prognosis target.

In the present invention, the trained results can be used or verified in prognostic data sets of different patients.

In the present invention, the threshold for distinguishing the good prognosis may be different for each disease.

According to an embodiment of the present invention, the step of selecting an image slice (slice) to which the target, which is the surgical target site, belongs; Targeting the automatic target, which is the surgical target site, based on semantic segmentation of an anatomical structure using a deep learning algorithm; Searching for a safe trajectory for surgery based on semantic segmentation information in the anatomical structure; Establishing a surgical plan after verifying the targeting result by user input; The step of establishing the surgical plan includes calculating a distance between the automatic target and a good prognosis target by conformity assessment, and training the calculated distance to the semantic segmentation algorithm and the targeting algorithm. A method for optimizing the surgical plan using the target area and a surgical plan optimization method is provided.

In the present invention, the deep target is a fully convolutional neural network (fully convolutional neural network) or convolutional autoencoder (convolutional autoencoder) through the semantic segmentation of the anatomical structure of the hypothalamus (subthalmic nucleus) and red nucleus The boundaries of the red nucleus) may be identified, and the target region for surgery may be targeted.

In the present invention, the deep trajector can search the safe trajectory by semantic segmentation of anatomical structures such as sulci, gyrus, and ventricle.

In the present invention, the training is to minimize the distance between the automatic target based on the deep learning algorithm and the good prognosis target or to set the length, direction, area, and volume range of the treatment target and the treatment policy, order, size of external force, It may be training to optimize direction and torque according to the prognosis after treatment.

A computer-readable recording medium for recording a computer program for executing the method according to the present invention is provided.

According to the present invention, it is possible to set a surgical target region based on deep learning and optimize the surgical plan based on the prognosis of patients.

1 schematically shows a surgical target site setting system according to an embodiment of the present invention.
2 is a diagram showing the configuration of a server internal processor according to an embodiment of the present invention.
3 is a time series showing a surgical target site setting and surgical plan optimization method according to an embodiment of the present invention.
4 is a diagram illustrating an image slice according to an embodiment of the present invention.
Figure 5 illustrates a method for detecting the target of the nucleus and the hypothalamus according to an embodiment of the present invention.
6 shows a target detected according to an embodiment of the present invention.
7 is a comparison of the prognostic based surgical plan optimization method of the present invention and the existing method.
8 illustrates a prognosis based surgical plan optimization method in more detail according to an embodiment of the present invention.
9 illustrates an automatic target according to an embodiment of the present invention.

For a detailed description of the present invention, which will be described later, reference is made to the accompanying drawings that illustrate, by way of example, specific embodiments in which the present invention may be practiced. These embodiments are described in detail enough to enable those skilled in the art to practice the present invention. It should be understood that the various embodiments of the present invention are different, but need not be mutually exclusive. For example, specific shapes, structures, and characteristics described in this specification may be implemented by changing from one embodiment to another without departing from the spirit and scope of the present invention. In addition, it should be understood that the position or arrangement of individual components within each embodiment may be changed without departing from the spirit and scope of the present invention. Therefore, the following detailed description is not intended to be done in a limiting sense, and the scope of the present invention should be taken to cover the scope claimed by the claims of the claims and all equivalents thereto. In the drawings, similar reference numerals denote the same or similar components throughout several aspects.

Hereinafter, various embodiments of the present invention will be described in detail with reference to the accompanying drawings in order to enable those skilled in the art to easily implement the present invention.

1 is a diagram illustrating an example of a network environment according to an embodiment of the present invention.

The network environment of FIG. 1 shows an example including a plurality of user terminals 110, 120, 130, 140, a server 150, and a network 160. 1 is not limited to the number of user terminals or the number of servers as an example for describing the invention.

The plurality of user terminals 110, 120, 130, and 140 may be a fixed terminal or a mobile terminal implemented as a computer device. The plurality of user terminals 110, 120, 130 and 140 may be terminals of an administrator who controls the server 150. For example, a plurality of user terminals (110, 120, 130, 140), a smart phone (smart phone), mobile phones, navigation, computers, notebooks, digital broadcasting terminals, PDA (Personal Digital Assistants), PMP (Portable Multimedia Player) ), tablet PC, etc. For example, the user terminal 1 110 may communicate with other user terminals 120, 130, 140 and/or the server 150 through the network 160 using a wireless or wired communication method.

The communication method is not limited, and a communication method using a communication network (for example, a mobile communication network, a wired Internet, a wireless Internet, and a broadcasting network) that the network 160 may include may include short-range wireless communication between devices. For example, the network 160 includes a personal area network (PAN), a local area network (LAN), a campus area network (CAN), a metropolitan area network (MAN), a wide area network (WAN), and a broadband network (BBN). , Any one or more of the networks such as the Internet. In addition, the network 160 may include any one or more of a network topology including a bus network, a star network, a ring network, a mesh network, a star-bus network, a tree or a hierarchical network, etc. It is not limited.

The server 150 is a computer device or a plurality of computer devices that communicate with a plurality of user terminals 110, 120, 130, 140 through a network 160 to provide commands, codes, files, contents, services, etc. Can be implemented.

In one example, the server 150 may provide a file for installing the application to the user terminal 1 110 accessed through the network 160. In this case, the user terminal 1 110 may install an application using a file provided from the server 150. In addition, the service provided by the server 150 by accessing the server 150 under the control of an operating system (OS) included in the user terminal 1 110 and at least one program (for example, a browser or an installed application) or Content can be provided. As another example, the server 150 may establish a communication session for data transmission and reception, and route data transmission and reception between a plurality of user terminals 110, 120, 130, and 140 through the established communication session.

According to an embodiment of the present invention, the server 150 may automatically set a surgical target region based on deep learning, and optimize a surgical plan based on the prognosis of patients. The administrator terminal is a terminal capable of inputting a command to control the server 150, and may transmit a deep learning algorithm and surgical prognosis data to the server 150.

Using the server 150 of the present invention, the surgical target site setting and surgical plan optimization system and method analyzes image information such as MRI and CT, brain surgery microscopy, endoscopy image, surgical field of view image, and deep-based image-guided surgery Characterized in that the implementation of the goal setting. For example, the method of setting the target area for surgery and optimizing the surgical plan of the present invention can be applied to deep brain stimulation surgery, and in particular deep brain stimulation surgery, the hypothalamic nucleus and the red nucleus called the Bejjani target Surgical targets based on (red nucleus) or existing or new targets, such as posterior subthalamic areas, zona incerta or globus pallidus interna, can be used.

2 is a block diagram illustrating an internal configuration of a server in an embodiment of the present invention.

2, the internal configuration of the server 150 will be described. The server 150 may include a memory 221, a processor 222, a communication module 223, and an input/output interface 224. The memory 221 is a computer readable recording medium, and may include a non-destructive mass storage device such as random access memory (RAM), read only memory (ROM), and a disk drive. In addition, the memory 221 may store an operating system and at least one program code (for example, a code for a browser or the above-described application installed and driven in the user terminal 1 110 ). These software components may be loaded from a computer-readable recording medium separate from the memory 221 using a drive mechanism. Such a separate computer-readable recording medium may include a computer-readable recording medium such as a floppy drive, disk, tape, DVD/CD-ROM drive, and memory card. In other embodiments, software components may be loaded into memory 221 through communication module 223 rather than a computer-readable recording medium. For example, at least one program is a program installed by files provided by a file distribution system (for example, the server 150 described above) for distributing installation files of developers or applications through the network 160 (example) As described above, the memory 221 may be loaded.

The processor 222 may be configured to process instructions of a computer program by performing basic arithmetic, logic, and input/output operations. Instructions may be provided to processor 222 by memory 221 or communication module 223. For example, the processor 222 may be configured to execute a received command according to program code stored in a recording device such as the memory 221. The processor 222 may include a dip selector 111, a dip targeter 112, a dip trajector 113, and an application unit 114.

The communication module 223 may provide a function for the user terminal 1 110 and the server 150 to communicate with each other through the network 160, and other user terminals (eg, user terminal 2 120) or other A function for communicating with a server (eg, the server 150) may be provided. For example, a request generated by the processor of the user terminal 1 110 according to program code stored in a recording device such as a memory may be delivered to the server 150 through the network 160 under the control of the communication module. Conversely, control signals, commands, contents, files, etc. provided under the control of the processor 222 of the server 150 transmit the communication module of the user terminal 1 110 through the communication module 223 and the network 160. It may be received through the user terminal 1 (110). For example, control signals or commands of the server 150 received through the communication module may be transmitted to a processor or memory.

The input/output interface 224 may be a means for interfacing with the input/output device. For example, the input device may include a device such as a keyboard or mouse, and the output device may include a device such as a display for displaying a communication session of an application. As another example, the input/output interface 224 may be a means for an interface with a device in which functions for input and output are integrated into one, such as a touch screen. As a more specific example, the processor 222 of the server 150 inputs/outputs a service screen or content configured using data provided by the user terminal 2 120 in processing a command of a computer program loaded in the memory 221. It can be displayed on the display through the interface 224.

Further, in other embodiments, the server 150 may include more components than those in FIG. 2. However, there is no need to clearly show most prior art components. For example, the user terminal 1 110 is implemented to include at least a portion of the input/output device 215 described above, or other configuration such as a transceiver, global positioning system (GPS) module, camera, various sensors, database, etc. It may further include elements.

3 is a time series showing a surgical target site setting and surgical plan optimization method according to an embodiment of the present invention.

A method for setting a surgical target site and optimizing a surgical plan according to an embodiment of the present invention may follow the following process. In the specification of the following invention, the surgical target site may be referred to as a target, and the surgical target site setting may be referred to as targeting.

First, according to an embodiment of the present invention, a deep selector 111 selects an image slice to which a target as a surgical target site belongs (S1).

Next, the deep targetor 112 uses a deep learning algorithm to set and target a surgical target site based on the anatomical segmentation of the anatomical structure (S2).

Next, a deep trajector (113, a deep learning-based routing system) searches for a safe trajectory based on semantic segmentation information in an anatomical structure (S3). More specifically, the deep trajector 113 considers the entry point of the frontal lobe, which is commonly used for deep brain stimulation surgery on the skull, and the segmentation results of the brain's sulci, gyrus structure, cerebrovascular and ventricular structures. The path is determined so as not to cross the blood vessels and ventricles. In special cases, such as the anterior thalamic nucleus, the route can be set to pass through the ventricle. The deep targeter 112 and the deep trajector 113 can be applied to a single or multiple image slices detected by the deep selector, and the information from the multiple image slices is integrated to set the path in the deep trajector 113 You can do Alternatively, 3D semantic segmentation may be performed on 3D image information to set a path in 3D space through the information.

Finally, the application unit 114 establishes a surgical plan after verifying the targeting result by the user through a device or software (S4).

Hereinafter, the steps of S1 to S4 and each configuration will be described in more detail.

As described above, the deep selector 111 selects the image cell of the target. In more detail, the deep selector is capable of semantic segmentation such as a convolutional neural network or a convolutional autoencoder among deep learning algorithms for slices and areas for target setting among MRI images. Sort by algorithm.

4 is a diagram illustrating an image slice according to an embodiment of the present invention.

4 is an example in which the deep selector 111 determines the image slice (for depth and area) for targeting. In the embodiment of FIG. 4, an example of classifying and selecting an image using a convolutional neural network is expressed. The third picture in FIG. 4 is a result of determining the image depth and area for targeting.

Next, the deep targeter 112 performs targeting. For example, deep targeters identify the boundary between the subthalmic nucleus and the red nucleus through semantic segmentation of anatomical structures through a fully convolutional neural network, and auto-targeting You can do The anatomical structures to be detected are not limited to the hypothalamic and red nuclei, and various anatomical structures such as globus pallidus interna, optic tract, and anterior thalamic nucleus can be used depending on the surgical goal.

Figure 5 illustrates a method for detecting the target of the nucleus and the hypothalamus according to an embodiment of the present invention.

First, as can be seen in FIG. 5, the dip targeter 112 extracts semantic segmentation of the eukaryotic and hypothalamic nuclei (S21).

Next, the dip targeter 112 detects the medial margin of the hypothalamus nucleus and the anterior margin of the nucleus (S22). In addition, if there is a diffusion tensor image such as dentato-rubro-thalamic tract related to disease such as tremor, the target region setting can be corrected according to the brain nerve path by additional segmentation in the algorithm.

Next, the deep targetter 112 calculates a target based on the above-described semantic segmentation, inner margin, and leading edge information (S23).

6 shows a target detected according to an embodiment of the present invention.

6, an example in which the red core and the hypothalamus core are detected is shown. As can be seen in FIG. 6, a color for distinguishing a target or non-target may be added to facilitate identification of a user in an image. The target site shown in FIG. 6 is the posterior subthalamic area (zona incerta), in addition to the subthalamic nucleus and the anterior thalamic nucleus. It can be applied to a variety of target sites, such as the globus pallidus interna and ventral intermediate nucleus.

Next, the deep trajector (Deep trajector) based on the semantic segmentation (semantic segmentation) information in the anatomical structure to search for a safe trajectory and target (S3). For example, anatomical structures such as sulci, gyrus, ventricle, and cerebrovascular can be semantic segmentation to establish appropriate safety pathways.

Next, when the safe trajectory and target searched as described above are searched, the application unit 114 may use the determined position of the target to establish a surgical plan by a target site setting method based on deep learning. For example, by taking a target MRI, applying the deep learning-based automatic targeting of the present invention to an MRI image, and verifying the targeting result by a user through a device or software button or a software terminal command, so that it can be applied to surgery. Develop a surgical plan.

In addition, the applicator 114 calculates the distance between the automatic target and the good prognosis target by conformity assessment, and uses the calculated distance to perform the semantic segmentation and train the targeting algorithm to optimize the surgical plan. Can be.

The surgical target region (auto target) set by the surgical target region setting method according to an embodiment of the present invention described above is provided to the user through a physical button of a device asking whether to be used, a software selection window, or a command on the software. Verification of the target site that can be used for actual surgery may be required. According to an embodiment of the present invention, since the verification result sets the surgical target by deep learning is applied using a manual or automated robot of an actual medical practitioner, the user may be subject to a medical license holder of the corresponding country. have. The present invention can be verified whether the target can be used in actual surgery through manual verification of the actual user. In addition, the operation target area for which the verification has been completed can be established to apply to the actual operation after granting use rights to the operation plan established by deep learning afterwards.

In more detail, after the user's verification, the output of the stereotactic surgical equipment, other surgical tools, radiosurgery equipment, etc. is output with a sign indicating that the verified target has been verified in the surgical planning software or equipment. In the related equipment, the result of planning can be transmitted from the inside of the device, and it can be directly applied to surgery.

On the other hand, if the exception rate is very low in the process of authorization by manual verification or the medical and statistical basis for the applied exception is insufficient, or if there is very little or no need for manual exception setting according to the development of the algorithm, manual Treatment can be performed without verification, and judgment on this is made by experts in each field or other types of algorithms, and may include a separate button, a terminal input command, and a physical button for this process.

On the other hand, the deep learning-based automated target area setting method of the present invention can be optimized for the surgical plan based on direct prognosis without going through an existing method such as expert opinion or location modeling of anatomical structures. That is, the surgical target site setting and surgical plan optimization method of the present invention can use outcome-guided learning when establishing a surgical plan, unlike the conventional training method. In more detail, according to an embodiment of the present invention, a set target region (target) can be combined with a prognosis-based learning technique to optimize the surgical plan, and to optimize it from the existing prognosis-based learning (outcome We will name it -guided surgical planning optimization.

Outcome-guided surgical planning learning is a concept similar to, but different from, outcome-guided surgical planning. Prognosis-based learning uses a prognostic-based optimization concept for biosignal analysis such as imaging and brain waves. . The concept of prognosis-based optimization is a concept that is more relevant to feature selection of analysis signals such as EEG, and outcome-guided surgical planning learning will be more directly related to optimization of surgical target site setting. Can be.

7 is a comparison of the prognostic based surgical plan optimization method of the present invention and the existing method.

FIG. 7(a) is a block diagram showing a conventional method, and if there is an outcome, feedback training is performed, and an expert manually displays it. Machine learning was performed by using the training.

On the contrary, FIG. 7(b) is a block diagram showing a method for optimizing the prognosis-based surgical plan of the present invention, and the present invention can perform machine learning through prognosis-based learning without manual display by an expert.

8 illustrates a prognosis based surgical plan optimization method in more detail according to an embodiment of the present invention. The steps of FIG. 8 below may be performed by processor 222 of server 150.

Referring to FIG. 8, first, semantic segmentation and image selection of an anatomical structure are performed using deep learning (S81 ).

Next, 1. Anatomical structure and relative or absolute distance, 2. Artificial intelligence neural network or 3. Using a different machine learning algorithm, select and target an automatic target that is a surgical target site from the semantic segmentation result (S82). The algorithms are optimized through prognosis results after surgery.

Next, in the case of using the distance to the anatomical structure for optimization, to calculate the distance to the good prognosis target, evaluate the calculated distance with the fitness function described below to train the semantic segmentation algorithm or targeting algorithm Use (S83). At this time, the result of step S83 may be used to train steps S81 and S82.

Next, the training result after training is confirmed in the prognosis data set of different patients (S84). At this time, if the performance improvement no longer occurs, training may be stopped.

In optimizing the prognosis-based surgical plan of the present invention without using the distance to the target site, an artificial neural network or other machine learning algorithm may be optimized to minimize the distance between an automatic target based on deep learning and a target of a good prognosis patient.

9 illustrates an automatic target according to an embodiment of the present invention.

Referring to FIG. 9, segmentation results (a to d) and automatic targets (e) and electrode artifacts (f) after surgery are illustrated by image analysis. In addition, g represents the distance from the automatic target to the postoperative electrode artifact location of the good prognosis patient. At this time, the position of the electrode artifact after surgery of the good prognosis patient may be set as the target of the good prognosis patient. Optimizing prognosis-based surgical planning can minimize D _GO of a good prognostic patient group. At this time, the fitness function D _GO can be expressed as [Equation 1] below.

[Equation 1]

Where D _GO is the average distance between the electrode targets after surgery from the automatic target of the good prognosis patient group, N _GO is the number of patients in the good prognosis patient group, and D _i is from the automatic target of the i th patient in the good prognosis patient group. The distance between electrode artifacts after surgery.

According to an embodiment of the present invention, the optimization of the surgical plan is to minimize D _GO , and machine learning of various methods may be used for the optimization of the surgical plan. In this embodiment, since the target area for surgery is specified by one coordinate, this optimization method is used, but in other cases, the order of the treatment target area in the form of a line or face, volume, not a point, and a treatment procedure in other cases. And the external force (force), direction (direction), torque (torque), etc., which are physically applied to the therapeutic device and the robot.

As described above, the present invention can set a surgical target site by introducing various machine learning algorithms such as an artificial neural network or deep learning. In addition, in optimizing the relative/absolute distance factor or machine learning algorithm, the fitness function is optimized to minimize the distance between the post-operative electrode position and the automatic target of patients with good prognosis after surgery. It is possible. The fitness function of the present invention can be used to optimize various machine learning algorithms.

In a more specific example, the prognostic group after surgery may be determined by using a UPDRS scale as a threshold for a predetermined improvement rate for Parkinson's disease. In another embodiment, the TETRAS scale or The Burke-Fahn-Marsden dystonia rating scale may be used for patients with tremor or dystonia to determine the prognosis group after surgery, or patient's global impression of outcome (PGI) or clinician's global impression of outcome (CGI) and various postoperative test results can be used to evaluate various postoperative measures. In addition, in the case of epilepsy, ILAE or Engel seizure outcome scale, depression, and obsessive-compulsive diorder can be used for each disease scale.

According to an embodiment of the present invention, training of machine learning is possible even by classifying each subscore rather than the overall result score according to the type of prognostic indicator. In one embodiment, even in Parkinson's disease, scores are classified according to symptoms such as tremor, gait disorder, stiffness, and pain, and among them, each patient is identified by outcome-guided learning to identify a location that is important or interested in analyzing effects. It can be used to set a very important personalized target site.

In one embodiment of the present invention, a threshold for distinguishing a good prognosis may be selected at any point determined to be an appropriate level of prognosis for each disease. In other words, the threshold for distinguishing prognosis may be different for each disease. In addition, the distance from the electrode position after surgery in a group with poor prognosis can be considered as an additional factor by modifying the fitness function according to the disease.

In addition, after training the method for setting the surgical target site of the present invention, the target algorithm was recalculated in the patient group not included in the training for verification, and the average with the target that actually showed a good outcome. The distance D _GO can be verified by recalculation. That is, the effectiveness of the optimized algorithm can be verified by measuring the distance from the good prognosis data by re-running the algorithm for the patient group not included in the training even after the optimization.

In addition, the surgery is performed by setting the surgical target area optimized for the new patient, and the prognosis is compared with the existing surgical method.

Hereinafter, the advantages of the present invention will be described.

Existing atlas-based automatic surgical target site setting (automatic targeting) has a problem in that it is difficult to apply to each individual because the shape of the anatomical structure varies from person to person. In addition, since coretargeting-based automatic targeting also has a large difference in individual anatomical structures, there has been a problem in that interpersonal coregistration results are not suitable to be applied to actual surgery. In addition, algorithms that use segmentation-based or atlas and segmentation complexes also depend on low-level features such as color, brightness, and clustering characteristics of pixels, so that they can be used for higher-level abstract features. Stable segmentation of the anatomical structure was not achieved, and it was never applied to actual surgery.

In addition, in the case of semantic segmentation by deep learning, segmentation is performed according to the connectivity of high-level features of anatomical structures, so it is possible to adapt to various individual anatomical variations and the level practiced by physicians. As it is possible to determine the margin of the anatomical structure, it is judged to be a level applicable to actual surgery. However, in the case of surgical planning using machine learning such as deep learning or other types of artificial intelligence algorithms, verification by humans may be important due to the lack of comprehensive judgment of these algorithms. That is, the setting of the target area of surgery by the existing deep learning was applicable to establish a surgical plan only when appropriate by medical and social judgment through a manual verification process by a human. The present invention aims to replace this manual verification process with an automated surgical planning device, robot or software.

In the case of the deep learning algorithm, it can be applied to anatomical structures in various cases, and continuous anatomical margins can be detected by regularization characteristics, so it can be applied after manual verification.

Optimization of the target setting position relied on expert opinions and anatomical models in the existing automatic targeting method. For example, the automatic targeting position is set based on the opinion that the 1 mm lateral position in the medial margin of the subthalamic nucleus or the 3 mm lateral position in the lateral margin of the red nucleus is appropriate. Became. In addition, the automatic targeting result always judged the applicable/disabled ratio in light of expert opinion. However, the adequacy of expert judgment is different from that of experts, and it is difficult to prove with existing machine learning algorithms because margin evaluation is often difficult due to adaptability to anatomical diversity due to machine learning.

Accordingly, according to an embodiment of the present invention, by introducing an outcome guided-surgical planning optimization technique into a target setting site for deep brain stimulation surgery, deep learning based on more objective prognosis data and electrode location Based automated brain deep stimulation surgery target setting site can be optimized. In addition, in the process of optimizing the target setting program, data labeling and result determination process according to expert opinions are omitted, and thus, there is an advantage that the program can be developed and improved through a lower cost and simple process.

In addition, if you rely on the existing expert opinion, the final automatic targeting area becomes the same as the expert opinion and there is no possibility of improvement beyond the expert opinion. However, if you optimize directly with only the prognosis and surgical location data without going through expert opinion, it is possible that you will be better in the prognosis after surgery than the goal setting by the expert. There are configurable advantages.

After automatic targeting, the surgical planning can be optimized by verifying automatic targeting as a method of measuring the distance from the existing good outcome target. In the existing method, it was mainly determined whether expert opinion is appropriate, and in the verification method of the surgical target site setting method of the present invention, the error between the automatic target and a good prognosis target is measured as a distance value, and through this, according to the quality of the algorithm and data. The verification result can be measured numerically.

The embodiments according to the present invention described above may be implemented in the form of program instructions that can be executed through various computer components and can be recorded in a computer-readable recording medium. The computer-readable recording medium may include program instructions, data files, data structures, or the like alone or in combination. The program instructions recorded on the computer-readable recording medium may be specially designed and configured for the present invention or may be known and usable by those skilled in the computer software field. Examples of computer-readable recording media include magnetic media such as hard disks, floppy disks, and magnetic tapes, optical recording media such as CD-ROMs and DVDs, and magneto-optical media such as floptical disks. medium), and hardware devices specifically configured to store and execute program instructions, such as ROM, RAM, flash memory, and the like. Examples of program instructions include not only machine language codes produced by a compiler, but also high-level language codes executable by a computer using an interpreter or the like. The hardware device can be changed to one or more software modules to perform the processing according to the present invention, and vice versa.

In the above, the present invention has been described by specific matters such as specific components and limited examples and drawings, but it is provided to help a more comprehensive understanding of the present invention, and the present invention is not limited to the above embodiments, and Those skilled in the art to which the invention pertains may seek various modifications and changes from these descriptions.

Therefore, the spirit of the present invention is not limited to the above-described embodiments, and should not be determined, and the scope of the spirit of the present invention as well as the claims to be described later, as well as all ranges equivalent to or equivalently changed from the claims. Would belong to

Claims (16)

In the surgical target area setting and surgical plan optimization system comprising one or more memory and processor,
The processor,
Using deep learning algorithms, perform semantic segmentation and image selection of anatomical structures,
Targeting by selecting an automatic target that is a surgical target region from the semantic segmentation results,
The distance between the automatic target and the good prognosis target is calculated by conformity assessment,
A surgical target area setting and surgical plan optimization system that uses the calculated distance to train the semantic segmentation algorithm and the targeting algorithm. According to claim 1,
The training,
Trainer to minimize the distance between the automatic target based on the deep learning algorithm and the good prognosis target, surgical target site setting and surgical plan optimization system. According to claim 1,
A surgical target site setting and surgical plan optimization system that uses or verifies the trained results in prognostic data sets of different patients. According to claim 1,
The threshold for distinguishing the good prognosis is different for each disease, surgical target site setting and surgical plan optimization system. A deep selector for selecting an image slice to which the target, which is a surgical target site, belongs;
A deep targeter targeting the automatic target that is the surgical target site based on the anatomical structure semantic segmentation using a deep learning algorithm;
A deep trajector searching for a safe trajectory for surgery based on semantic segmentation information in the anatomical structure;
An application unit that verifies the targeting result by user input and then establishes a surgical plan;
Including,
The applicator calculates a distance between the automatic target and a good prognosis target by conformity assessment, and optimizes the surgical plan by using the calculated distance to train the semantic segmentation algorithm and the targeting algorithm, Surgical target area setting and surgical plan optimization system. The method of claim 5,
The deep targeter grasps the boundary between the subthalmic nucleus and the red nucleus through semantic segmentation of an anatomical structure through a fully convolutional neural network, and identifies the surgical target site. Targeting, surgical targeting and surgical planning optimization system. The method of claim 5,
The deep trajector semantic segmentation of anatomical structures such as sulci, gyrus, and ventricle to search for the safe trajectory, surgical target site setting and surgical plan optimization system. Performing semantic segmentation and image selection of an anatomical structure using a deep learning algorithm;
Selecting and targeting an automatic target that is a surgical target site from the semantic segmentation result;
Calculating a distance between the automatic target and a good prognosis target by conformity assessment;
Using the calculated distance to train the semantic segmentation algorithm and the targeting algorithm; Surgical target site setting and surgical plan optimization method comprising a. The method of claim 9,
The training,
A method of minimizing the distance between an automatic target based on the deep learning algorithm and the good prognosis target, a surgical target site setting and a surgical plan optimization method. The method of claim 9,
A method for setting a surgical target site and optimizing a surgical plan, using or verifying the trained results in a prognosis data set of different patients. The method of claim 9,
The threshold for distinguishing the good prognosis is different for each disease, surgical target site setting and surgical plan optimization method. Selecting an image slice to which the target that is the surgical target site belongs;
Targeting the automatic target, which is the surgical target site, based on semantic segmentation of an anatomical structure using a deep learning algorithm;
Searching for a safe trajectory for surgery based on semantic segmentation information in the anatomical structure;
Establishing a surgical plan after verifying the targeting result by user input;
Including,
The step of establishing the surgical plan includes calculating the distance between the automatic target and a good prognosis target by conformity assessment, and using the calculated distance to train the semantic segmentation algorithm and the targeting algorithm. How to optimize the plan, set surgical target areas and optimize the surgical plan. The method of claim 13,
The deep targeter is a boundary between the subthalmic nucleus and the red nucleus through semantic segmentation of an anatomical structure through a fully convolutional neural network or a convolutional autoencoder. Grasping, and targeting the surgical target site, surgical target site setting and surgical plan optimization method. The method of claim 13,
The deep trajector explores the safe trajectory by semantic segmentation of anatomical structures such as the sulci, gyrus, and ventricle. The method of claim 13,
The training,
Minimize the distance between the automatic target based on the deep learning algorithm and the good prognosis target, or set the length, direction, area, and volume range of the treatment target and optimize treatment policy, sequence, external force size, direction, and torque according to the prognosis after treatment Training, how to set surgical target areas and optimize surgical plans. A computer-readable recording medium recording a computer program for executing the method according to claim 9.

Images