DE10109310A1 - Three-dimensional tracking of probe needles, biopsy needles or surgical instruments using a CT or MRT system with improved tracking provided by undertaking a calibration step using an infrared light source calibration frame - Google Patents

Abstract

Navigation system for probe needles, biopsy needles or surgical instruments using a CT or MRT system, a computer with image processing modules and 2 or more infrared sensors coupled spatially to the imaging device and a tracker device with 3 or more infrared LEDs. In addition the system comprises calibration means for the image recording sensors and means for extraction and labeling of selected anatomical tissue.

Description

For biopsy (taking tissue samples), radiation (brachytherapy), Laser therapy and cold therapy (cryotherapy) are probes under control a computer tomogram (CT) positioned in the tumor tissue. Both the right one Puncture channel (position and direction of the probes) as well as the probe distribution in the Tumor are crucial for the quality of the therapy.

At the moment, CT-supported probe positioning is usually only carried out within two dimensional section plane of a three-dimensional CT. The patient receives one for this CT examination in the region of the tumor, through which a three-dimensional Image data record is recorded. This data set turns into an inexpensive two-dim sought for a section through the tumor in which the probe is to be positioned. Then the probe is gradually positioned within this section, each step must be checked again with a CT scan.

Appreciation of the current state of the art

The three-dimensional position measurement in space is currently using mechanical, electromagnetic, infrared and laser-guided measuring systems in the computer-aided navigation performed [1]. Position measurement is often carried out of patients using external reference systems with a stereotactic framework. In the mechanical measurement, the position of a surgical instrument, e.g. B. a probe tip, determined by a measuring arm, the joints of which move a surgical instrument attached to the measuring arm in all 6 degrees of freedom enable [2, 3]. From the measurement data of the articulated position of the measuring arm, Submillimeter range on the position of the instrument. The problem with this measuring technique is both the limited freedom of movement  of the measuring arm as well as the retroactive forces from the mechanics. Both are through the construction specified.

An electromagnetic position measurement involves a transmitter and a sensor forces. The transmitter generates a dynamic or a static magnetic field. In the sensor there are three electromagnetic coils in an orthogonal arrangement. The induced coil currents can be calculated as a function of the position and orientation of the Represent sensors in relation to the magnetic field generated by the transmitter. If in Area around the transmitter and sensor no materials with magnetic Properties exist, the spatial position of the sensor can be changed Accuracy of approximately ± 1 mm can be determined. There are three different ones electromagnetic position measuring systems are available [4, 5, 6]. Because bigger magnetic components such as B. the operating table, the CT table or Construction elements of the buildings can lead to magnetic field distortion [7], is the use of these systems on special structural or furnishings Measures are bound that do not exist at first and will only be created have to.

Infrared measuring systems can measure the positions of infrared LEDs in the room. By attaching three diodes to the surgical instrument, their Positions of at least two infrared sensitive cameras in the Submillimeter range can be determined. This technique is mainly used in the Neurosurgery as well as in oral and facial surgery for navigation applied [8, 9, 10].

All three measurement methods described (mechanical, electromagnetic and infrared) are integrated in navigation systems that allow navigation outside the CT Recording room on the "virtual patient", d. H. on a three-dimensional CT or MR volume data set, enable. Precondition for this is an exact one Position correlation of the volume data set and the navigation system. This Correlation is usually done using anatomical landmarks, measure their positions on the patient and in the CT or MR data sets be registered. By knowing the spatial positions of the landmarks on A transformation rule can be created for patients and in the image data sets be, which allows further measured positions, e.g. B. those of surgical Instruments to represent in the three-dimensional image data sets. Currently is this process is carried out manually by the user using a sensor Measure landmarks on the patient as well as identify them in the image data sets and must mark.

The above-mentioned measuring methods are compared to the laser-assisted puncture Computer tomographs performed without "virtual patients". After several CT sections of the tumor region were created, now becomes a desired puncture channel specified in the CT sections. The system now provides the setting parameters for one Laser pointer that places a red dot (laser beam) on the patient's surface projected. This represents the puncture point of the probe. After positioning the Probe tip can be oriented so that the laser point too the probe end hits. Now the probe axis corresponds to the laser pointer and thus the Orientation of the desired puncture channel. The penetration depth must be done beforehand be drawn manually on the probe.

The disadvantage of this technique is that it can only be used to position a probe directly on a computer tomograph. The advantage here is that there is no correlation of Landmarks on the patient and their pixels must be made. Prerequisite for this  but is the constant position of the patient on the CT table throughout Procedure, which is usually not the case.

A mechanical arm with a further development of this technique Puncture probe from the Picker company, which is firmly connected to the computer tomograph is. First, a CT scan of the tumor region is made. Through the Position determination of the probe by arm adjustment and by knowing the performed CT cutting positions, which is defined by the table position, is at automatic probe positioning possible when the patient position is constant [11]. If the patient position is not guaranteed to be constant when puncturing, in the meantime new CT images to check the current probe position be performed.

task

The present invention has for its object a stereo optical and system free from magnetic or electromagnetic influences hands-free, computer-aided navigation of surgical instruments or probes spatially related to in CT or MR tomograms organs, vessels, other tissue structures that can be visualized and localized Risk structures of patients for easy and quick application under visual control of the Navigation movement and the guided instrument or probe direction largely interaction-free calibration of the recording system, the Guiding system of the instruments or probes, the instrument or probe length and the spatial assignment based on the patient situation.

The aim of this procedure is to support the doctor with the biopsy or Probe positioning in the tumor using a CT-based navigation system, for that the new measuring method is first developed. Before positioning the probe a volume data set of the tumor region is generated with a CT and the like captured image data via a defined interface from the CT to the Navigation system sent. The position is automatically registered of the tumor based on the measurement of the spatial position of landmarks on the fixed Patients using a new method with the presented system. After that the Navigation of the probes through the new stereographic measuring method of the Position of the probe tip and a display of the two-dimensional cutting plane by the volume image data set, which is defined by the probe position and direction is made possible. The disadvantages of the known methods mentioned, not free Navigation as with the mechanical version, no interference from magnetic Influencing building constructions or facilities as with electromagnetic method, manual determination of spatial positions as in the currently used infrared method, marking the penetration depth of the probes such as with laser-assisted procedures and the prerequisite for a constant patient position should be avoided.

Walkthrough

To solve the task specified above, a tracking sensor is used Stereo camera system consisting of at least two CCD video cameras. The cameras come with inclined optical axes for monitoring the Workspace used on the patient table. To suppress the for the  Cameras sensitive wavelength range with the exception of the near one The video cameras are equipped with appropriate optical infrared Long pass filters equipped. Due to the arrangement with inclined optical axes the calibration pattern does not include a right angle with the optical axes. Therefore, it is not a multiplane process but a single plane process for the Calibration of the camera parameters sufficient.

The calibration of the stereo camera system with regard to the position and orientation of the Cameras in the room, the image width of the lenses and a possible radial symmetry The distortion of the images caused by the lenses is determined by a calibration pattern carried out according to the single plane procedure [12-15].

For hands-free instrument or probe guidance, an instrument or probe holder ( Fig. 3 and 4) equipped with at least three light emitting diodes emitting in the infrared wavelength range ( Fig. 3 and 4) serves as a transmitter, hereinafter referred to as the tracker. The trackers are manufactured in two different versions. The different versions are used for the lateral ( Fig. 3) or dorsal / ventral ( Fig. 4) application of biopsy needles. The arrangement of the light emitting diodes must not contain any rotational symmetry with respect to the biopsy needle axes.

Definition of novelties

• 1. The exact positions of the light-emitting diodes are weighted from their images using a function of the brightness distribution minus a threshold value to be defined for segmenting the background using mathematical methods. The position of the light-emitting diode P _{LED is used} as the weighted focus
  with the weighting function w (x _i ) = (h (x _i ) - S) ²
  as a quadratic function of the brightness values h (x _i ), which are greater than a predefined threshold S. The automatic determination of the current position of the light spots of the light emitting diodes on the guide system by means of a mathematical model eliminates the time and device-consuming manual measurement of the current positions of the light spots of the diodes, which are not exactly known for reasons of production technology. The exact positions of the light spots are included in the accuracy of the three-dimensional position determination.
• 2. The calibration of the guidance system of the surgical instruments or probes takes place automatically in an iterative process. Starting from initial values that describe the approximate position of the LEDs on the tracker, but not the actual light spot, the correct correspondence is selected for all possible correspondence between the positions of the light spots in the right and left stereo image and the exact relative positions of the light spots with each other established. There is therefore no need for an exact mechanical measurement of the geometry of the guide system and the light spots of the light-emitting diodes.
  For this purpose, the tracker is modeled mathematically as an ordered set of n points in space: T = {T _i | T _i ∈ R ³ , i = 1,. , , n} Each of the points T _i corresponds to an IR LED on the tracker. This allows two properties to be assigned to the tracker:
  • 1. its geometry, ie the relative positions of the points T _i with each other regardless of their absolute positions, and
  • 2. its position in space, ie the absolute positions of the points T _i .
  For the sake of unambiguity in the position determination of the tracker, it must be assumed that there is no axis to which the points T _{i are} rotationally symmetrical. According to the first property mentioned, the tracker can generally be described in a defined position.
  Due to the mechanical structure of the tracker and due to the use of commercially available IR LEDs with an unknown internal structure, an exact measurement of the tracker is not possible. The trackers are therefore calibrated using the newly developed method. For this purpose, the approximate distances between the LED positions are defined.
  The specified number of n LED positions is extracted from the two stereo images (l ₁ ... _N for the left, r ₁ .. R _n for the right camera image). Since no correspondence of the LED positions in the stereo images is known, all combinations of assignments {l ₁ . , .l _n } → {r ₁ . , .r _n } and tries to calculate the three-dimensional world coordinates of the LED positions by triangulation. A number of these assignments do not result in an intersection or approximation of the projection beams of individual LED positions within a predetermined limit of 1 mm and are excluded from further consideration. All correspondences that are now potentially possible result in a set τ of tracker shapes and positions in space, which are determined by P (τ) = {P _i | P _i ∈ R ³ , i = 1,. , , n}. All combinations of assignments of P _i from P (τ) to T _i from T are formed and an error measure
  With
  d T | k = || T _k - T _{k + 1} || Distances between two LEDs with increasing index k from T,
  d P | k = || P _k - P _{k + 1} || Distances between two LEDs with increasing index k from P (τ)
  calculated. That P (τ) for which the error measure D is minimal describes the correct correspondence τ of the LEDs in the stereo images and maps the tracker optimally to the tracker model T.
  This process is automatically repeated several times and the respective position data of the LEDs are averaged and the positions of the LEDs which are considered to be calibrated are determined therefrom.
• 3. The lengths of the surgical instruments or probes, insofar as these can be determined by the distance of a defined point (eg the instrument or probe tip) from the guide system, are determined by tracking the guide system with the instrument or probe tip in a constant, fixed position automatically determined. The radius of a spherical surface, which is determined by the movement and by tracking of the guide system, is determined as the length of the instruments or probes. The orientation of the instruments or probes in space then results from a geometric modeling of the instruments, their known attachment points in the guide system and the now known position of the instrument or probe tip in relation to the guide system.
  For the modeling of the probe it is assumed that it can be described as a mathematical straight line g _N : x = N + λ m, where N is a base point on the straight line, m is the straight line direction and λ is a free parameter. The probe tip N _S can thus be regarded as a point on the probe line g _N.
  For the calibration of the probe, the tracker is moved with the probe tip fixed at one point and the tracker position is determined. The calculation for this is carried out according to the procedure described under point 2. In addition, in order to determine the position of the tracker, all individual LED positions are optimally transformed to the given tracker model T using the same transformation equation. As a result, when using more than 3 LEDs on the tracker, all available information is included in the position calculation of the tracker. Since the movement of the tracker defines a spherical surface for at least four positions that are not in one plane ( Fig. 5), the center point of this sphere can now be calculated by means of a compensation calculation of all recognized tracker positions and transformed into a coordinate system assigned to the tracker. This point defines a probe tip and lies on the probe line. Varying the position of the probe tip with respect to the tracker coordinate system, ie moving the probe in the holder on the tracker, and repeating the process described above, one obtains a number M of positions of the probe tip N _Si with i = 1. , , M, from which the probe line g _N is now represented by a compensating line x = No + λ m _N , for which the error sum of the Euclidean distances of the points N _{Si from} the line g _{N is} minimal. A time-consuming measurement of the length of the instrument or probe is omitted, there is no need to draw the penetration depth on the probes, deviations in the geometry of the instruments or probes from the specified shape are recorded by the type of calibration.
• 4. The correspondence of the surgical instruments or probes to the position of the patient is carried out by a further calibration. Infrared LEDs are used as landmarks that are located on the patient who is fixed in a defined position. The light-emitting diodes can be determined in the volume image data of the tomograms by image processing algorithms and their current position can be determined with the aid of the stereo camera system. This provides a correspondence between the position of medically anatomical tissue structures with regard to their current position and in the volume image data. This is a completely automatic process that is used for real-time registration.
  Now the surgical instrument or the probe can be navigated hands-free and represented in the section through the three-dimensional tomogram image data in whose plane the axis of the instrument or the probe moves. This enables direct visual control of the movement of the instruments or probes in their axial direction with regard to risk structures within the section shown.
• 5. The depth of penetration and the orientation of surgical instruments or probes can now automatically in tomogram image data in a virtual and real Navigation can be shown. This allows the optimal depth of penetration and a Danger from next to and behind the tissue structures to be treated horizontal risk structures can be visualized.
• 6. By infrared light diodes attached to the patient table, its position can be calibrated with respect to the image sensors. A connection of the stereo optical  Tracking procedure with a patient table control allows the patient to always do so positioned to hold that after movement of the instrument or probe CT section plane right through the tip of the surgical instrument or probe runs. The automatic positioning of the probe tip makes it quick Control of the exact position given relative to anatomical structures.
• 7. The applicability of the stereo-optical navigation and tracking procedures is can generally also be used in other volume image data-generating processes.

references

[1] R. Marmulla, M. Hilbert, H. Niederdellmann: Intraoperative Precision of Mechanical, Electromagnetic, Infrared and Laser-guided Navigation Systems in Computer-assisted Surgery, in: Oral and Maxillofacial Surgery, 2. Suppl. 1, p 145-148, May 1998
[2] RL Carrau, CH Snyderman, HD Curtin, IP Janecka, M. Stechison, JL Weissman: Computer-assisted Intraoperative Navigation During Skull Base Sur gery, in American Journal of Otolaryngol 17 (2), p. 95-101, March-April 1996
[3] S. Hassfeld, J. Muhling, J. Zoller: Intraoperative Navigation in Oral and Maxillo facial Surgery, in: International Journal of oral Maxillofacial Surgery 24 (1 Pt. 2), p. 111-119, Feb. 1995
[4] Fasttrack System, Polhemus Inc., Colchester VT, USA
[5] The Bird, Ascension Inc., Burlington VT, USA, discussed in: G. Straßmann, C. Koltas, R. Heyd. S. Walter, D. Baltas, Th. Martin, H. Vogt G. Ioannidis, G. Sakas, N. Zamboglou. Navigation System for Interstitial Brachytherapy, Radiother. Oncol., 2000
[6] Aurora, Mednetix, Villingen, Switzerland
[7] W. Birkfeller, F. Watzinger, F. Wanschitz et al .: Systematic Distortion in Magnetic Digitizer, Med. Phys. 25 (11), p. 2242-2248, 1998
[8] IM Germano, JV Queenan: Clinical Experience with Intracranial Brain Needle Biopsy Using Frameless Surgical Navigation, in Computational Aided Surgery 3 (1), p. 33-39, 1998
[9] R. Marmulla, H. Niederdellmann: Computer-assisted bone segment navigation, in: Jounal of Craniomaxillofacial Surgery 26 (6), p. 347-359, Dec. 1998
[10] T. Morioka, S. Nishio, K. Ikezaki, Y. Natori, T. Inamura, H. Muratani, M. Muraishi, K. Hisada, F. Mihara, T. Matsushima, M. Fukui: Clinical Experience of Image-guided Neurosurgery with a Framelass Navigation System, in: No Shinkei Geka 27 (1) p. 33-40, Jan. 1999
[11] PinPoint ™, Picker, USA
[12] R. Tsai: An Efficient and Accurate Camera Calibration Technique for 3 D Machine Vision, in Proc. Computer Vision and Pattern Recognition, IEEE, Miami Beach, 1986
[13] R. Lenz: Lens error-corrected calibration of semiconductor cameras with standard lenses for high-precision 3-D measurements in real time, in: Mustererkermung 1987, 9th DAGM Symposium, 1987
[14] S. Posch: Automatic depth determination from grayscale images, Deutscher Universitätsverlag, 1990
[15] D. Richter, W. Schick, S. Vormbrock. Verification of a robot base calibration with a stereo image processing system by evaluating the depth determination within the calibration volume, in: Pattern recognition 1996, 18th DAGM symposium, p. 493-501, 1996

Claims (8)

1. Methods and devices for navigation of probes, biopsy needles or surgical instruments to support surgical interventions, in particular in the field of radiology and brachytherapy, consisting of

• - a computer (PC) that is expanded with image processing modules,
• - at least 2 infrared-sensitive image recording sensors, which are coupled with a fixed spatial reference to a CT or MRI tomograph,
• - A navigation device (tracker) to which at least 3 light-emitting diodes emitting in the infrared wavelength range are firmly coupled and to which a probe, biopsy needles or surgical instruments can be mechanically attached, and
• - the probes, the biopsy needles and the surgical instruments,

characterized in that

• - Means for calibrating the image recording sensors are present and
• - Means for the extraction and identification of selected anatomical tissue structures are available as individual data sets with the help of digital image processing.

2. Navigation system according to claim 1, characterized in that with a Image processing program from three-dimensional volume image data operative Access routes to selectable anatomical tissue structures bypassing of risk structures can be calculated and visualized. 3. Navigation system according to claim 1 and 2, characterized in that on the Patients or subjects with the body structure firmly connected at least 3 in Infrared light emitting diodes are available with the tomograph can be recorded and using methods of digital Image processing can be extracted from the tomogram to the location anatomical structures of the patient, clearly related to the patient structure to be able to define. 4. Navigation system according to claim 1 to 3, characterized in that on the Patients or test persons with LEDs firmly connected to the body structure the imaging sensors which are firmly connected to the tomograph the location of the patient's extracted anatomical structures, to be able to clearly define it based on the current patient situation. 5. Navigation system according to claim 1 to 4, characterized in that on the Patient table of the tomograph with in the infrared wavelength range emitting LEDs are attached, its position in the room, based on the Camera positions with methods of digital image processing from the image data can be calculated. 6. Navigation system according to claim 1 to S. characterized in that with a Calculation rule from the positions of the LEDs from claim 3, 4 and 5 exact current position of anatomical structures calculated from claim 2 become. 7. Navigation system according to claim 1 to 6, characterized in that with the invention and under points 1 . and 2. the tracker is calibrated and the position of the tracker is calculated in the definition of novelties section. 8. Navigation system according to claim 1 to 6, characterized in that with the invention and under point 3 . The procedures described in the section Definition of new products calibrate the probes, biopsy needles or surgical instruments attached to the tracker with regard to the current position of the tracker and their position is calculated.