Anthropomorphic stereotactic phantom: main components
The present disclosure of the subject matter of the technology presents below for consideration a detailed description of preferred embodiments with reference to the following figures, in which:
Fig. 2 shows a scheme of interactive medical phantom for training skills of functional stereotactic interventions with decision making unit and stereotactic tomography modification module.
FIG. 1 is a schematic representing a stereotactic operation using classical frame-based stereotactic system;
FIG. 2 is a schematic representing an interactive medical phantom for practicing functional stereotactic interventions according to the present principles;
FIG. 3A - schematic diagram representing an intracranial installation of a calibration rack;
FIG. 3B - schematic diagram representing preparation of training neuroimaging data;
FIG. 4 - block diagram representing the stages of modification of stereotactic phantom tomography;
FIG. 5 - flowchart of the sequence of operations of the method of simulation of intracranial space of the phantom with neuroimaging data of the human brain during the training stereotactic operation, according to the illustrative embodiment.
The phantom (10) has an anthropomorphic shape in the form of a life-size human head, neck and upper torso. The phantom is made of plastic by 3D printing technology. Neuroimaging data from magnetic resonance imaging and computer tomography of the human head are used as the basis for 3D modeling of the phantom, which are anonymized in advance, and facial features are removed. The person whose neuroimaging data are used to design the phantom and training examples provides informed consent. The shape of the phantom accurately reproduces the desired characteristics of the individual. The phantom's skin is modeled by the silicone material. Intracranial structures are removed from the phantom volume at the phantom design stage.
The intracranial space of the phantom is a cavity (11) whose volume corresponds to the volume of the human brain.
The system of non-contact estimation of stereotactic tool positioning (12) is located intracranially so that the working field (13) (area of space in which the object coordinate can be determined with required accuracy without contact) corresponds to the location of anatomical structures (targets) in the intended training stereotactic operation. The system of non-contact estimation of positioning can be a stereo pair of video cameras, depth cameras, magnetic positioning system and other devices that allow non-contact estimation of object coordinates in the studied space. In the presented embodiment, the non-contact positioning evaluation system is a stereo vision system, the optical cameras of which are located in the intracranial space - the region of the facial part of the skull, which allows to evaluate the accuracy of positioning of the stereotactic instrument in the region of the basal nuclei of the brain.
Non-contact positioning estimation system is connected to the decision unit (14) by wired or wireless way. The stereotactic imaging phantom used to plan the training operation is pre-processed by the stereotactic imaging phantom modification module (15), which is embedded in the tomograph software.
Two removable platforms (16) are placed parasagittally on the convexital surface of the skull, simulating the skull vault in the area of the trepanation hole at the time of the training operation. The platforms are made of bone-dense material to simulate the process of putting a trepanation hole with standard surgical instruments (electric torpedo, shovel, etc.). After the training operation, the pads are replaced with new ones. Depending on the clinical task, the location of the removable pads on the phantom may vary; in this version this location allows planning the trajectory of the stereotactic instruments diving into the basal ganglia region, for example, when practicing deep brain stimulation operations.
The stereotactic tool (17) for the phantom is a thin cylindrical sleeve 2 mm in diameter having two colored marks - one at the end of the sleeve and the other several centimeters away from the end of the sleeve. By determining the coordinates of the marks on the instrument (17) by the non-contact positioning evaluation system (12), the decision-making unit (14) virtually simulates different types of immersible stereotactic cannulas - cryoprobe, radiofrequency probe, deep intracranial electrodes and so on.
In the presented embodiment of the phantom, the contrast reference registration marks of the phantom with training neuroimaging data, hereafter registration marks (18), are spheres placed along the perimeter in the cranial space of the phantom. The position of the labels in the phantom's cranial space is known and is selected at the phantom design stage - Fig. 3B shows their virtual counterparts (26) in the neuroimaging data space. The labels are bimodal - made of contrast material for imaging by both computed tomography and magnetic resonance imaging methods. The minimum required number of reference marks (tie points) is three. The reference marks are used during phantom calibration as well as during the modification phase of stereotactic tomography of the phantom. Thus, when the phantom is tomographed, the labels in the phantom's cranial space will be clearly visualized on the tomograms - in Fig. 4 of the stereotactic tomography data import stage (28) an axial tomogram of the phantom is shown, on which the phantom intracranial space, its cranial part containing registration labels as well as the stereotactic frame fixed to the phantom head with localizer are visualized. Coordinates of the centers of the marker spheres are used as reference elements for alignment of the phantom head space with training neuroimaging data of different modalities by linear affine transformation (point-to-point alignment).
Calibrating the phantom and preparing clinical training tasks
Phantom calibration, neuroimaging training data and clinical examples (tasks) are prepared during phantom production.
Calibration refers to the process of linking the coordinate system of neuroimaging training data to the physical space of the phantom head and to the internal coordinate system of the non-contact evaluation system of stereotactic instrument positioning, with saving the obtained transformation functions into the decision-making block.
Figure 3A shows the phantom calibration step. Phantom calibration begins with the installation of an intracranial calibration rack (19), which is fixed to the base of the intracranial cavity (11). The position of the calibration post is predetermined at the phantom design stage. The calibration strut consists of cylindrical rods on a single base. At least three targets (20) in the form of spheres are placed on the rods. The rods are sized so that when the calibration rack is installed in the phantom, the targets (20) are positioned in the working field (13) of the non-contact positioning evaluation system. Coordinates of each of the rack targets in the intracranial space are known in advance relative to the registration reference marks (18) - determined at the phantom design stage. The non-contact positioning estimation system (12) determines the coordinates of each target. Target coordinate data (20) is transmitted to the decision unit, and the coordinate system of physical space of the phantom head (z, y, x) is defined by coordinates of three points in space. In this way the internal coordinate system of the positioning evaluation system and the coordinate system of the physical space of the phantom head are linked. The calibration rack is removed from the phantom.
Preparation of training neuroimaging data refers to the matching of all available neuroimaging modalities of the human head used in the phantom design, segmentation of anatomical structures and required landmarks, and classification of this data and saving it to the decision-making unit. Figure 3B shows various segmented training neuroimaging data in a single voxel space (k, i, j), the virtual space of the phantom head. The training data contain medical images of magnetic resonance imaging (21), computed tomography, tractography data (22), data of segmented large cerebral vessels (23), whole brain mask (24), whole head volume mask (25), registration reference marks (26) and other modalities (anatomical structures - targets) required to create the training task.
Thus, after phantom calibration and neuroimaging training data preparation there are two coordinate systems connected by transformation functions: coordinate system (z, y, x) of the phantom head physical space and coordinate system (k, i, j) of neuroimaging training data space (virtual space of the phantom head). The registration repertoires (18) here are the key element for linking these two spaces. Determining the coordinates of the object in the working field (13) of the non-contact positioning estimation system (12), the decision unit (14), using transformation functions, calculates its coordinates in the virtual space of the phantom head (human brain).
The decision unit
The decision unit (14) is a personal computer or any other electronic computing machine or processor with a display that is wired or wirelessly connected to the non-contact positioning evaluation system. The decision unit stores all human head neuroimaging data from the clinical example and transformation functions, as well as the clinical tasks created. Preparation of the clinical task consists of creating a series of logical checks of all the prepared classified data (stereotactic target, vessel, functional brain zone, brain ventricles, brain, tracts, stereotactic atlas data, and so on). Having determined the coordinate of the placed object in the working field of the non-contact positioning evaluation system (13), the decision unit (14) determines its position in relation to the classified training neuroimaging data. For example, it can determine whether this object is in the volume of the required anatomical target, how far from the required target, whether the trajectory of instrument plunge intersects with vessels and functionally significant areas of the brain, and so on.
In fact, the phantom has its own independent navigation system, hidden from the surgeon. This allows the phantom to work interactively - to evaluate in real-time the position of the stereotactic instrument in the neuroimaging training data space. Determining coordinates of marks on stereotactic tool (17) by non-contact estimation of positioning in intracranial phantom space the decision making unit (14) calculates equation of a line defined by two points in three-dimensional space. Thus the decision unit calculates the position of stereotactic tool and its trajectory in the training neuroimaging data space. Depending on the position of stereotactic tool (17) in the phantom's intracranial space, the decision making unit (14) provides information about possible clinical effects and side effects by evaluating the involvement of anatomical structures of the prepared neuroimaging training data in the simulated field of stereotactic exposure.
Algorithm of stereotactic tomography modification
The stereotactic tomography modification module (15) is a software module that can be integrated into the tomograph software, into the stereotactic equipment planning station software, into the decision-making unit, or can be a separate software product installed on a personal computer. Figure 4 shows a block diagram representing the phantom stereotactic tomography processing steps used in the stereotactic tomography modification module.
Input data for the module are stereotactic imaging phantom as well as training neuroimaging data from human head imaging.
In the initial step (27), the tomography operator performing stereotactic tomography on the phantom activates the stereotactic tomography modification module. Immediately after performing the phantom tomography, the tomography data are loaded into the memory of the stereotactic tomography modification module (28). The next step (29) determines the tomography modality used for the stereotactic tomography of the phantom. Then (30), by computer vision, patterns of intracranial registration marks are found and their coordinates are determined. The module pre-stores the coordinates of the corresponding labels in the space of prepared training neuroimaging data. Registration marks are thus a reference element for overlapping the space of localization tomography phantom and training neuroimaging data of human head by the method of affine transformations by points. The module calculates the transformation function (31). Then, using the transformation function, values of voxels of stereotactic tomography space corresponding to voxels of human head mask (32) are replaced by values of voxels of human brain tomography (33) of the same modality. Output data of the module is modified stereotactic tomography, where voxels of intracranial volume of the phantom are replaced by voxel values of human brain neuroimaging data, while extracranial localizer elements are preserved. The modified stereotactic imaging data of the phantom is exported (34) for further use in the planning of stereotactic surgery as a reference image. The operator deactivates (35) the modified stereotactic tomography module.
The key element of the phantom are the registration marks (18). They are reference elements linking the training neuroimaging data of the human brain, the physical space of the phantom head, the internal coordinate system of the non-contact positioning evaluation system, and the stereotactic tomography space of the phantom. This version of the phantom shows a particular case of using marks as registration reference elements, but it should be taken into account that modern algorithms of registration and processing of medical data allow working with image intensity, the shape of image contours and edges as well as surfaces of volume models constructed from neuroimaging data. In fact, the internal surface of the intracranial cavity of the phantom and other fixed intracranial elements of the phantom structure contrasted in the tomograms can become the reference element of the alignment in addition to the labels or instead of them.
Fig. 5 illustrates a method of simulating the intracranial space of the phantom with neuroimaging data of the human brain during a teaching stereotactic surgery, containing steps in which the surgeon receives a clinical task (36). A stereotactic frame with a localizer is attached to the phantom in a standard way (37). A stereotactic tomography is performed on the phantom with the localizer (38). The obtained localization tomogram is processed (39) in the stereotactic tomography modification module, where the cranial and intracranial volume of the phantom is replaced by the training neuroimaging data of the human brain with preservation of extracranial elements (reference landmarks) of the localizer hidden from the surgeon. Then the surgeon performs stereotactic planning of the operation (40) using standard software compatible with the stereotactic equipment used, where a modified stereotactic tomography is used as the reference image. Next, stereotactic operation (41) is performed according to the methodology of the stereotactic equipment used: after applying a trepanation hole at the planned entry point, using a manipulator, the stereotactic instrument is plunged along the planned trajectory into the desired target. At the next stage (42), real-time decision-making unit intraoperatively performs an assessment of the surgery efficiency by comparing the stereotactic tool position in the space of the human brain training neuroimaging data and gives a score at the end of the training session.
Example of method realization
Surgeon receives a clinical task: 53-year-old patient diagnosed with Parkinson's disease of mixed form, stage III according to Hen-Yar with pronounced motor fluctuations. Complaints: feeling of stiffness, hand tremor, intensifying outside the effect of Levodopa drugs.
Based on the task conditions, the surgeon decides that subthalamic nucleus is the optimal stereotactic target for therapeutic intervention, and the treatment method is chronic deep brain stimulation. The surgeon uses classical frame stereotactic system with an N-shaped localizer. Using a standard four-point fixation, the surgeon secures a stereotactic frame to the phantom's head. A localizer is secured to the stereotactic frame. Next, a CT scan is performed on the phantom with the localizer fixed. The CT operator activates the "phantom" mode, i.e. activates the stereotactic tomography modification module. Thereby, secretly from a surgeon, modification of stereotactic tomography is performed immediately after the examination. In fact, the surgeon receives a tomography of a person's head with the localizer attached.
The surgeon loads the data of the modified stereotactic tomography to the planning station included in the stereotactic frame set. Then he performs the usual planning actions of stereotactic surgery on the human brain - using standard planning tools based on stereotactic guidance methodology he plans the future trajectory of deep electrodes immersion into the stereotactic target and receives data which he has to expose on the scales of the stereotactic device in the operating room.
Next, stereotactic surgery is performed on the phantom. The surgeon makes a trepanation hole at the planned entry point, exposes the scales of the manipulator, and guides the stereotactic instrument into the planned target. As soon as the active end of the tool enters the working field of the non-contact positioning estimation system, the decision-making unit starts to evaluate the possible effect of exposure. So, let's say that it turns out that the active end of the tool has moved more laterally from the "ideal" target target for this task, and is close to the inner capsule of the brain. The decision unit alerts the surgeon that the patient's facial muscles are tight, and the optimal therapeutic effect (reduction of tremor and muscle stiffness) cannot be achieved. Taking these data into account, the surgeon decides to move the position of the stereotactic instrument 1 mm medially. At the end of the training operation the surgeon receives scores based on the correctness of stereotactic target selection, accuracy of hitting the planned target, adequacy of stereotactic trajectory construction. If the surgeon makes a completely wrong decision - planning the trajectory through a large blood vessel, the decision unit outputs information about fatal vital signs in the patient and the exam ends prematurely.
Conclusion
Stereotactic surgery is a borderline discipline which develops at the junction of neurophysiology, neuroanatomy, neurosurgery, neurology, neuroimaging methods and computer technology. Currently, there is no universal method for practicing the skills of stereotactic operations due to the specificity of such interventions and availability of many types of stereotactic devices. The principles of the invention presented in the current application allow approximating the methodology of training stereotactic surgery to the real human brain surgery. The phantom has anthropomorphic properties not only in terms of external features, but also in the intracerebral (virtual) space in which the surgeon plans the surgery. The phantom is compatible with any extracranial type of localizers, since the methodology of modification of stereotactic tomography of the phantom, involves substitution of only cranial-intracranial volume of neuroimaging data of the phantom.
The resulting modified phantom tomogram is compatible with the planning stations of stereotactic equipment. The presence of bimodal intracranial registration marks allows the phantom to perform both MRI and CT stereotactic markings. The availability of removable platforms for the application of trepanation holes (consumables), allows the surgeon to make a choice of the position of the hole in the skull each time. Each subsequent operation requires a new stage of fixation of the stereotactic frame, thereby redoing the stereotactic calculations. The phantom makes it possible to interactively simulate the intraoperative clinical picture depending on how correctly the surgeon selected the stereotactic target for exposure, planned the stereotactic trajectory and exactly hit the planned target, thus giving the surgeon the choice of further tactics. Virtual simulation of stereotactic exposure allows to create many clinical tasks from a variety of neurosurgical pathologies (deep brain stimulation surgery, stereotactic tumor biopsy, thermal destruction etc.) and depends on the amount of neuroimaging data prepared in advance for the phantom. The result of using the phantom is the training of skills in performing functional stereotactic interventions using different types of stereotactic devices and clinical tasks. The phantom can be used in stereotactic training centers.
The technical solution in this project was described in the Russian national patent application №2022115154 "Interactive medical phantom for practicing functional stereotactic interventions" Peskov V.A., Kholyavin A.I.