US 20050171557 A1
A miniature surgical robot and a method for using it are disclosed. The miniature surgical robot attaches directly with the bone of a patient. Two-dimensional X-ray images of the robot on the bone are registered with three-dimensional images of the bone. This locates the robot precisely on the bone of the patient. The robot is then directed to pre-operative determined positions based on a pre-operative plan by the surgeon. The robot then moves to the requested surgical site and aligns a sleeve through which the surgeon can insert a surgical tool.
1. A surgical system, comprising: a surgical robot for manipulating a surgical tool to a surgical site with precise positioning during a surgical procedure; an attachment member configured and dimensioned to attach the surgical robot to a bone associated with said surgical site; and a controller programmed prior to said surgical procedure to direct the robot to position the surgical tool at the surgical site.
2. The surgical system according to
3. The surgical system according to
4. The surgical system according to
5. The surgical system according to
6. The surgical system according to
7. A surgical system of
8. The surgical system according to
9. The surgical system according to
10. The surgical system according to
11. The surgical system according to
12. The surgical system according to
13. A surgical system for facilitating a surgical procedure at a surgical site, comprising: a surgical robot including a base member; two pairs of actuators extending outward from the base member at fixed angles, wherein said actuators each have first and second ends, said first ends of a pair being spaced apart on said base member and said second ends of a pair coming together to define a tool holding element; an attachment member removably securable to the robot base member and configured and dimensioned to attach the surgical robot to a bone associated with said surgical site; and a controller including a cpu and user interface communicating with said robot, said cpu containing a program for guiding the robot based on data generated from surgical site images created prior to said surgical procedure.
14. A method for facilitating a surgical procedure at a patient surgical site, comprising: generating an initial three dimensional image of a patient surgical site; planning the surgical procedure utilizing said initial three dimensional image prior to conducting said procedure to provide a surgical plan referencing said initial three dimensional image; attaching at least a support member of a surgical robot to a patient bone associated with the planned surgical procedure at the surgical site; generating at least two radiographic images of the surgical site and support member; correlating said radiographic images to form a pseudo three dimensional image including an accurate position of said support member attached to the bone; selecting a window from said pseudo three dimensional image and registering said window with a similarly chosen window from said initial three dimensional image; and correlating the pseudo three dimensional image with the initial three dimensional image such that the support member is located with respect to the surgical plan.
15. The method according to
16. The method according to
17. The method according to
18. The method according to
19. The method according to
20. The method according to
This Application is a continuation of co-pending U.S. application Ser. No. 09/912,687, which claims priority from U.S. Provisional Application Ser. No. 60/220,155, filed Jul. 24, 2000. The contents of both the above applications have been incorporated herein by reference.
The present invention relates to a robotic device. Particularly, this invention relates to a robotic device that attaches to a bone of the patient and aids or performs surgical procedures.
Generally, robots are used in many different industries for many different applications. One industry, for example, is the medical industry that uses robots in applications including assisting the surgeon during surgical procedures. Robots are especially suited for some surgical tasks because they can be constructed to be very steady, computer controlled, and precise in their movements. Characteristics such as these can be especially helpful during surgery on sensitive areas, such as, for example, the vertebral column but are applicable throughout the body.
Typical vertebral column surgical procedures include vertebral fusion, insertion of medical devices such as pedicle screws, discography, percutaneous discectomy, or the like. These procedures typically require a large invasive operation that exposes the patient to a high risk of infection, excessive trauma, fluid loss, post operative pain, scaring, and a lengthy recovery time. Some difficulties relating to surgery on the vertebral column include micro-movement of the vertebral column during the operation, inherently small target objects of the procedure such as the pedicles, extremely delicate nearby nerve tissue, and limited operating room space because large equipment is needed to aid in the procedure, such as C-arm X-ray devices. Furthermore, the patient and operating room staff are exposed to large doses of radiation because these procedures require repeated X-raying and/or fluoroscoping of the surgical site so the surgeon can view the position of surgical tools or implants relative to non-visible body parts.
A need exists for a device that can assist minimally invasive surgery with low radiation exposure while allowing the surgeon to precisely align and control or monitor the surgical procedure. Some prior art devices have attempted to accomplish this however, these devices are either too complicated, not sufficiently accurate, or consume too much operating room space.
One such device is disclosed in U.S. Pat. No. 6,226,548. This device combines a navigation system, a bone mounted apparatus, and surgical tools that communicate with the navigation system. This apparatus primarily consists of a clamp that attaches to the patient's spine and extends outward forming a reference arc bearing emitters or a tracking means. All the surgical tools used in this procedure are fitted with emitters or a tracking means similar to the reference arc. The surgical suite is fitted with a navigation system capable of recognizing the emitters or tracking means of the reference arc and surgical tools, a computer system for interpreting the location of the tools, and a video display for the surgeon. After surgically placing the clamp and reference arc on the patient a CT or MRI is taken creating a three-dimensional image of the patient with the attached device. When the patient is in place in the surgical suite with the attached reference arc the navigation system locates the arc and the surgical tools and displays them, relative to each other, on the three-dimensional CT scan.
While the device disclosed in the '548 patent offers some advantages in terms of accuracy and reduced trauma, the advantages of this type of prior art device are limited. The critical part of a surgical tool that must be monitored is the working end of the tool, whether that be a screwdriver or a drill bit or the like. These can not be tracked with such prior art systems. Transmitters or emitters cannot be attached to the working ends of tools so the computer must estimate the location of the working end by locating the tool generally and extrapolating. This causes inaccuracy and errors that can not be tolerated in spinal surgery or other high accuracy procedures where the smallest error can result in a serious and permanent outcome. Also, prior art devices such as these are hand held by the surgeon and thus, limited in accuracy to the surgeon's ability to hold and align the tool.
Furthermore, when using this system, the user must be cautious to not block the line-or-sight between the tool mounted emitters or receivers, the reference arc bearing emitters or receivers, and the navigation system. This can severely limit the ability of the surgeon or surgical team as the tool may actually limit their ability to aid the patient. Also, while such prior art systems do reduce the incision size, they complicate the surgical procedure. Usually a patient is brought into a surgical suite ready for a procedure, the procedure is performed, completed, and the patient leaves. However, the '548 patent system requires the patient to be put through a surgical procedure to affix the clamp and referencing arc, then the patient is transported to a CT or MRI, then transported back to the surgical suite in a non-sterile condition for the substantial portion of the procedure to commence. Finally, this system has many components, such as the navigation system and the computer output unit, that clutter up the already limited space in the surgical suite.
Therefore, there is a need in the art for a device with high precision and accuracy that can assist the surgeon in aligning the working end of the surgical tool such that delicate procedures can be preformed percutaneously with minimal radiation exposure to both the patient and the surgical staff.
The present invention is directed to a device and method for assisting in surgical procedures. According to the invention, a robot is disclosed that precisely positions a surgical tool with respect to a surgical site. The robot attaches to the bone of a patient with a clamp or with wires such as K-wires. Actuators extend from the robot base and move away from and toward the base member. This manipulates balls that rotate within spherical swivel joints that in turn align a sleeve. A surgical tool such as a screw driver or a drill bit is inserted through the sleeve and thus is precisely aligned with a site requiring surgery.
The present invention also includes a method for using the robot to assist in surgical procedures. Initially, three dimensional images are taken of the patient and the surgeon performs pre-operative planning of the procedure to be done on the images. This creates parameters that will later be used to direct the robot to the location where the surgical procedure is required. The robot is then attached to the patient by the clamp or the k-wire. C-arm images are taken of the patient with the attached clamp and these images are co-registered and calibrated such that a precise image of the bone with the robot attached is generated. This image is then registered, or matched, with the three dimensional image. This is accomplished in a highly efficient and accurate manner by taking small windows of the images where the surgery is to take place and registering these small portions. The small windows are chosen off the images by locating the bone attached clamp and selecting a window according to pre-operative calculation of the bone-robot attachment location. After these windows are chosen and registered, the remaining bone is registered by aligning the registered windows. At this point the robot is located precisely on the bone of the patient in the three dimensional image and can be manipulated by the surgeon to a pre-operative planned location for percutaneous insertion of surgical tools, medical devices, or implants.
For a better understanding of the nature, objects, and function of the present invention, reference should be made to the following detailed description in conjunction with the accompanying drawings, in which:
Like reference numerals refer to corresponding elements throughout the several drawings.
Referring to the illustrations and particularly to
In a preferred embodiment of the invention, robot 30 includes base 35 that sits vertically on clamp adaptor 45. At least two pair of actuators 80 extend from base 35. The actuators 80 extend from the base 35 forming a fixed angle 85 between base 35 and actuator 80. This angle is generally between about 15-.pi.degrees and more preferably about 45 degrees. In one preferred embodiment, the points of attachment of actuators 80 are spaced apart by about 50 mm in the Z direction and about 50 mm in the Y direction. Each actuator 80 is capable of operating independently from the other actuators 80. Actuator 80 is similar to known linear actuators and includes a housing, a motor, a lead screw, an electrical connection, and a position sensor such as an optical encoder, an LVDT, or the like. In a preferred embodiment each actuator is approximately 5 mm in diameter and approximately 35 mm in length.
The end of actuator 80 that is not fixedly attached to base 35 contains hinge joint 90. Hinge joint 90 links actuator 80 to rigid member 100. In a preferred embodiment member 100 is about 4 mm in diameter and 40 mm in length. Hinge joint 90 permits member 100 to freely rotate through about 270 degrees on an axis that runs parallel to base 35. The other end of the rigid member 100 is fixed with solid connection 105 to ring member 110. There is no movement between rigid member 100 and ring member 110 at solid connection 105.
Upper ring member 110A and lower ring member 110B, solidly connected to individual rigid members 100, come together at spherical swivel joint 120. Each ring member 110 forms one half of an outer race of spherical swivel joint 120. Ring members 110 are free to rotate with respect to one another, but are held fixedly from separating in the Z axis direction. Contained between upper ring member 110A and lower ring member 110B, and free to swivel, is ball 130. Passing through ball 130 is sleeve 60. Sleeve 60 passes through both upper and lower balls 130, forming an aligning axis through which surgical tool 70 is passed. As actuators 80 extend and retract, hinge joints 90 freely rotate about the Z axis and balls 130 swivel in the spherical swivel joints 120 formed by upper and lower ring members 110. A hollow axis is formed by the sleeve passing through each of upper and lower balls 130 such that a surgical tool 70 can be inserted through and be accurately aligned with the working location.
According to the present invention the above described robot 30 is just one example of a robot configured for surgical assistance that may be utilized with the system according to the present invention. Other robot configurations that could satisfy the same tasks include, for example, a parallel robot constructed to the required dimensions, such as that described in Simaan, N., Glozman, D., and Shoham, M.: “Design Considerations of new types of Six-Degrees-of-Freedom Parallel Manipulators,” IEEE International Conference on Robotics and Automation, Belgium, 1998, which is incorporated by reference herein.
In a preferred embodiment of the invention, robot 30 is attached with the bone of a patient by clamp 40. As shown in
With reference specifically to
Above described clamp 40 is an example of one embodiment according to the invention by which a robot may be attached to a bone for assisting in a surgical procedure. Other attachment devices can also be incorporated with a robot such as, for example, K-wire connections.
With reference now to
Next, the patient is brought into the operating room, a small incision is made according to standard surgical practice at the site where clamp 40 is to be attached, and the clamp is attached to the selected bone using handles as described above, step 462,
In another embodiment of the present invention the secondary calibration step, step 465B, can be accomplished by attaching the robot 30 to the clamp and taking multiple C-arm images. By knowing the dimensions, or by placing referencing markers on robot 30 and knowing the distance and angle between the referencing markers the C-arm images can be calibrated in a secondary calibration step, step 465B.
The next step of the process is co-registration, step 500. The C-arm images are transferred into the control unit 10 as data, step 502. At each location an image is taken from, the position of the C-arm is recorded, step 504, into the control unit 10. The data of the images, step 502, and the position of the C-arm, step 504, are correlated by knowing the position from which each images was taken, step 504, and by aligning the referencing markers 810 (
According to a preferred embodiment of the invention, bone to bone registration next occurs in step 600. Step 600 is a process of estimating and matching the true surface contours or the objects in the images. Registration methods are either based on geometry or intensity of the image. Geometric based registration is achieved by finding features in the 2D fluoroscopic images and matching these features with corresponding features in the 3D image, acquired, for example, from a CT scan dataset, MRI image, ultrasound image or from a CAD model. The features can be known landmarks (anatomical landmarks or implanted fiducials), or contour points in the fluoroscopic image, matched with the registered object's surface. An algorithm that may be used to compute the transformation is the Iterative Closest Point (ICP) algorithm. This algorithm is described, for example in Besl, P. J. and McKay, N. D., “A Method for Registration of 3D Shapes”, IEEE Trans. on Pattern Analysis and Machine Intelligence, 1992, 14(2), 239-255, which is incorporated herein by reference. The input to the algorithm are sets of back-projected rays from the fluoroscopic images, and a model of the registered object. The algorithm iteratively computes a transformation that approximates the ray sets to the model. For landmark registration, a match between each ray and the corresponding landmark is defined before searching for the transformation. Contour registration selects a new surface point to match with each ray on every iteration.
Preferably, the registration process uses two or more fluoroscopic images, as described in greater detail, for example, in Hamadeh, et al., “Towards automatic registration between CT and X-ray images: cooperation between 3D/2D registration and 2D edge detection”, Medical robotics and computer assisted surgery, 1995, Wiley 39-46, and Hamadeh, et al., “Automated 3-Dimensional Computed Tomographic and Fluoroscopic Image Registration”, Computer Aided Surgery, 1998, 3, which are incorporated herein by reference. According to this method, anatomical landmarks in the images are detected and matched manually. Based on this match, an approximated initial guess is computed, with ray intersections, which are 3D points in the registration environment, being matched with the model's landmarks. Then, the object's contour in the 2D image is registered with the model's surface. A likelihood estimator is used to remove outliers, or pixels not in the contour, from the sample point set. A signed distance function is defined to overcome any internal contours problems. The overall in-vitro accuracy of this method can be better than 2 mm.
In one alternative, a single fluoroscopic image may be used for registration, achieving an accuracy of about 3 mm. This technique is based on a combinatorial search among matches of three points and three rays. The match with minimal average distance for the registration is then selected. This alternative is described in Tang, “Method for Intensity-based Registration with CT Images,” Masters Thesis: Department of Computer Science, Queen University, Ontario Canada, 1999, which is incorporated herein by reference.
In a further alternative according to the invention, intensity-based registration is achieved by comparing fluoroscopic images with simulated X-rays (digitally reconstructed radiographs, or DRR's) from an estimated position. Such a technique is generally described in Lemieux et al., “Patient-to computed-tomography image registration method based digitally reconstructed radiographs”, Medical Physics, 21, 1994, 1749-1760 and Murphy, M. “An automatic six-degree-of freedom image registration algorithm for image-guided frameless stereotactic surgery”, Medical Physics, 24(6), June 1997, which are incorporated by reference herein.
When the camera position guess and the actual position are very close, the original and reconstructed image are very similar. Pixel intensity information is used to define a measure of similarity between the datasets. The similarity measure can include intensity values, cross-correlation, histogram correlation, and mutual information. The algorithm proceeds in three steps. The input is a CT data set, intrinsic camera parameters, one or more fluoroscopic images and an initial camera position estimate for each image. In the first step, the algorithm generates one DRR for each given camera position. In the second step, a dissimilarity measure is computed between the real and reconstructed image. In the third step, new camera poses are computed that best reduce the dissimilarity between the images. The process is repeated until convergence is reached. The parametric space of camera positions in then searched incrementally from an initial configuration. The space is six-dimensional (three rotations and three translations). The advantages of this technique is that no segmentation is necessary. However, the search space is six-dimensional, and can contain may local minima.
A benefit of the present invention is that it can utilize either of the above described registration methods. By utilizing the dimensions of the bone attached robot and its attachment location, the initial location of the window is a very good guess of the location and therefore the intensity based method can be utilized. Thus, according to the present invention, a faster and more accurate registration process is accomplished as between the fluorscopic and 3D images. This is done in step 600, and occurs very quickly and with a high degree of accuracy because the registration process is performed on small windows of the images, rather than the images as a whole. Preferably windows are selected that specifically relate to the known location of the robot and/or its support member. Windows of about 20 mm by 20 mm located approximately adjacent to the clamp location, according to pre-operative calculation of the bone-robot attachment location, are selected from the C-arm (fluoroscopic) image data, step 610. For example, these windows may be selected as the area above the attached clamp 40 in the C-arm image and the tip of the transverse process of the vertebra covering the area where the surgical procedure is to take place. Generally, the same windows are chosen from both the pseudo three-dimensional hybrid C-arm image, step 510, and also from the CT image (3D image), step 410. The small windows chosen from the C-arm images and the CT scan image are then laid over each other and matched or registered by the control unit, step 620, as described above. Focusing only on a small window of the C-arm image rather than looking for a matching anatomical landmark in the entire image, makes the process occur very fast and with the high degree of accuracy needed for precise procedures such as vertebra surgery.
Next, the remaining portion of the CT and C-arm image of the bones are overlaid, the registration windows are aligned, and the remaining bone is registered, step 630. Since the windows have already been accurately registered this step occurs quickly and also with a high degree of accuracy. Now clamp 40 is located precisely on the bone, step 640, of the CT image. Next, the user attaches robot 30 to clamp 40 and thus, robot 30 is located precisely with respect to the bone, step 645.
After robot 30 is co-registered 500 and registered 600, its position is known relative to the patient's bone and therefore can move to align with the pre-operatively picked location such that the operation can virtually take place on the control unit. The user selects a pre-operatively planned location and task from step 420 by use of a joystick, mouse, touch screen, or the like, step 710. The Robot 30 responds and moves sleeve 60 into position, step 720, such that when the user inserts a surgical tool 70 through the opening in the sleeve 60 the surgical tool 70 will be precisely aligned with the location requiring the surgical procedure, step 730. The surgeon can then insert a selected surgical tool 70 and operate without opening the surgical site to see the placement of the surgical tool because the surgeon can verify the positioning of the surgical tool 70 on the control unit 10 and display 20. Thus operating percutaneously or in general open procedures, with a high degree of accuracy, low trauma, small incisions, low chance of infection, and minimal exposure to radiation. A further benefit of this system is that because the robot is miniature it can be freely attached to the bone of a patient and move with the body. Therefore, the robot system does not need a dynamic referencing device to maintain orientation with the body once it is registered. This creates a more precise and less complicated system that is versatile and user friendly as the surgeon can manipulate the patient into different surgical positions without disturbing the robot system.
The present invention is illustrated herein by reference to a spinal vertebra attachment. However, it will be appreciated by those in the art that the teachings of the present invention are equally applicable to other bone attachments.