1

Image-guided Telerobot for Percutaneous Instrument Distal Tip Repositioning

Conor James Walsh, Student Member, IEEE, Jeremy Franklin, Alexander H. Slocum, Rajiv Gupta
and ablations used for the diagnosis and treatment of disease respectively. These procedures end up being quite iterative due to the fact that accurately repositioning the distal tip of the instrument is difficult, if not impossible, once it is inside the body because of forces from the tissue that resist its pivoting motion. The need to reposition the distal tip of the instrument arises from correcting for targeting errors caused by instrument misalignment or deflection and also from a desire to sample/treat multiple adjacent points in a volume. To reposition the distal tip, radiologists currently have two options; 1) overcompensate when realigning the instrument, or 2) retract it and attempt to re-insert it along the correct trajectory. Both of these approaches lead to trauma-inducing insertions that can be uncomfortable for the patient and lead to a variety of complications. Percutaneous treatment of disease is a rapidly growing field with brachytherapy, where small radioactive seeds are deployed through a needle and implanted into the tumor, and ablation, where thermal energy is applied to destroy the tumor, being the most common. Initially providing only palliative care, Radio Frequency Ablation (RFA) is now used for complete eradication of tumors in a variety of organs (e.g. lung, liver and kidney). While there are clear benefits to locally treating disease, the number of procedures that can be performed in this way is limited in part due to the fact that current probes only provide a static volume of coagulative necrosis. They are deployed at a single location to achieve a pre-determined (typically cylindrical or ellipsoidal) burn volume that includes the tumor. Due to the difficulty in placing the instrument precisely, the physician is often forced to take large treatment margins and thus this technique cannot be used near sensitive anatomy. To generate large ablation volumes RF ablation systems have been developed that have three or more electrodes in parallel attached to a single probe handle (Cool-tipTM RF Cluster Electrode, Mansfield, MA, USA) as well as an umbrella like flock of electrodes that deploys from the distal tip of the needle (LeVeen CoAccess Electrode System, Boston Scientific, Natick, MA, USA). However, in the first, it is difficult to keep all electrodes correctly spaced relative to each other and these systems can cause the tissue or tumor to be pushed out of the way as opposed to being punctured. In the second system charring of the electrodes or tynes can result in uneven energy deposition in the tissue and make it difficult to retract the tynes fully. Further, both of these are limited to a fixed burn volume that limits their use near sensitive regions of
Abstract Manually performed image-guided percutaneous procedures are limited by targeting errors due to instrument misalignment, deflection and an inability to reposition the distal tip of the instrument after it has been percutaneously inserted. These limitations result in suboptimal probe deployment, improper tissue sampling, and/or inadequate treatment for patients as well as excessive procedure times and radiation doses. Hence we are developing a telerobot capable of steering the distal tip of a medical instrument in soft-tissue or free space that is based on the concept of deploying a flexible pre-curved stylet from a concentric straight cannula. The proximal end of the cannula is attached to the distal end of a screw-spline that enables it to be translated and rotated with respect to the casing. Translation of the stylet relative to the cannula is achieved with a second threaded screw with a splined groove. The device is made of mostly plastic components and actuation is achieved using micro-stepper motors. Measurements of the maximum axial force for the cannula screw-spline and stylet screw were found to match those predicted. The sub-millimeter and sub-degree translation and angular accuracy enabled the distal tip of the stylet to be accurately positioned within a volume of 113 cm3. Index Terms percutaneous, needle, steerable, screw-spline, telerobot.

I. INTRODUCTION inimally invasive percutaneous procedures are routinely performed for both the treatment and diagnosis of disease. In medical parlance, percutaneous needle insertion refers to any medical procedure where the skin is punctured with a rigid needle to access to inner organs or other tissue as opposed to an open surgical approach. Typically the procedures are performed under image-guidance such as computed tomography (CT), fluoroscopy, ultrasound and magnetic resonance imaging (MRI) that provide high resolution images of the patient anatomy. After a target is identified in the body with the aid of medical imaging, a needle insertion point is chosen so as to avoid obstructions (such as ribs, nerves, and blood vessels) and the needle is then manually inserted towards the target. Example procedures are biopsies

Manuscript received October 9, 2009. This work was supported in part by the Center for Integration of Medicine and Innovative Technology and was part of course 2.752 taught by Professor Alex Slocum and Julio Guerrero. C. J. Walsh and A. H. Slocum are with the Massachusetts Institute of Technology, Cambridge, MA 02139 USA (corresponding author phone: 617780-9915; e-mail: walshcj@mit.edu). J. Franklin was with the Massachusetts Institute of Technology, Cambridge, MA 02139 USA and is now with Apple, Cupertino, CA 95014. R. Gupta is with the Department of Radiology, Massachusetts General Hospital, Boston, MA 02114.

2 the anatomy. To generate even larger or irregularly shaped ablation volumes multiple skin punctures are made with multiple probes to insert electrodes into different locations of the tumor so as to burn the entire extent of it. A. Previous Work in Needle Steering Some physicians do attempt to steer standard instruments by bending part of them with their hand so they take a curved trajectory when inserted or by exploiting the asymmetric bevel tip to cause the instrument to glide to one side. Mathis and Yankelevitz et al. developed a steerable needle that enables the radiologist to exert a curved shape on the needle in a more controlled manner [1]. The mechanism consists of a pivoting handle on the proximal end of the needle that is attached to its distal tip via four small steel bands. Salcudean et al. developed a device that enables active needle steering during insertion [2, 3]. The device employs a stylet that is longer than the cannula so that up to 2 cm of the stylet tip (with a mild curve) can be selectively exposed. The extended curve essentially acts as an adjustable bevel on the tip of the needle. Motors provide actuation and the steering direction is selected by rotating the stylet and the steering rate is selected by extending the stylet and exposing the curve. Other approaches that have been taken use bevel [4-6] and external [4] forces on the needle to cause it to bend so as to steer through the tissue. In [4], DiMaio and Salcudean formulated a needle Jacobian that described tip motion due to needle base motion and a tissue finite element model. They assumed that the needle was rigid compared to the tissue and that it was redirected by pulling on and angling the needle shaft outside the body. Another approach by Webster et al. considered a system where the needle was flexible relative to the tissue [5]; needle steering resulted from the asymmetric forces on the needle tip due to the bevel. A similar system is presented in [6] that extends this approach to needle steering by applying rotation duty cycling [7]. By rotating the needle in a spin-stop-spin-stop manner, the time the bevel spent in a particular orientation could be controlled with longer stop intervals creating steeper curvature and longer spin intervals creating straighter trajectories. However, continuous spinning of a beveled needle may lead to tearing of tissue. These projects have shown the promise of actively steering standard medical needles. While offering improvements over unintuitive manual control, these approaches rely on a reaction force from the tissue to achieve steering and further knowledge of the mechanical properties of the tissue would be required for trajectory planning. In general, tissues are inhomogeneous, and vary with patients as well as across tissue layers. Another approach that has been used to steer thin medical instruments is the use of pre-curved concentric tubes. Two commercially available, passive devices for spinal procedures, such as vetebroplasty, are based on this concept [8, 9] and consist of a superelastic pre-curved stylet with a stiffer concentric outer cannula that can achieve distal tip steering as illustrated in Fig. 1. These systems, however, require manual operation and do not lend themselves to accurately targeting points in a volume.

Fig. 1 Illustration of the concept of achieving steering with a stylet with a precurved distal portion and a concentric outer cannula. When the stylet is inside the cannula it is substantially straightened but once it is deployed it will deflect laterally and assume its pre-curved shape (assuming no yielding of the material occurs).

To address this limitation, there has been recent exciting work to actively control devices based on concentric precurved tubes; both to control the needle curvature through overlapping pre-curved cannulas [10] and to steer around obstacles in the body [11, 12]. B. Contribution In this paper we present a telerobot that can accurately reposition the distal tip of a percutaneous instrument to a near by point as illustrated in Fig. 2. Such a device would be useful for correcting for targeting errors, reaching multiple points in a volume or for directing the instrument tip around obstacles when a straight line trajectory can not be taken.

Fig. 2 Concept of adjusting the distal tip of a medical instrument through a single needle insertion.

Such a system could be used with existing medical robots that have been developed. These robots are mounted on the CT scanner bed [13-15] or the patient [16-18] and provide some method for remote needle orientation and insertion. The majority of these manipulators provide a remote center of rotation so that the needle can pivot about the skin surface [1921]. All of these systems use some combination of intraoperative CT images [13], pre-operative 3D imaging [13], static real-time fluoroscopy [13, 22], or tracking systems [810] for procedure planning and execution. The doctor typically controls them via joystick [13, 22] or a point-and-click user interface that directly incorporates the medical images [2, 11].

3 II. STEERING MECHANISM DESIGN The first step in the design process was to choose a mechanism that is capable of steering inside the body. The functional requirements the mechanism had to meet included: It should be deployed from standard size cannula for percutaneous procedures (0.5 mm < D < 2.5 mm). It should function independent of the tissue properties i.e. work in tissue and free space It should be able to position the distal tip of the stylet in a sufficiently large work volume (V = 100 cm3). A. Steering Concept These functional requirements led to a design based on concentric devices with an inner curved compliant stylet and a rigid, outer cannula as previously illustrated in Fig. 1. When inside the cannula, the stylet is substantially straightened and upon deployment of the stylet from the distal tip of the cannula, the stylet will then take its preformed shape. The included angle between a line tangent to the tip of the stylet and the axis of the cannula is defined as and is related to the stylet curvature and displacement by = zs R (1) The total pre-formed bend angle of a stylet is defined as curve with the total arc length of the curve given by (2) This is the limit for stylet displacement relative to the cannula.
lcurve = curve * R

Forward Kinematics From Fig. 3 it can be seen that in the -z plane the position of the tip of the stylet is a function of its radius of curvature, R, the amount it is extended from the cannula, zs, and the axial position of the cannula with respect to the casing, zc: = R(1 cos zs R ) (3) z z = zc + R sin s (4) R B. Kinematic Equations In order to reposition the distal tip of the stylet within a The angle between the positive x-axis and the -z plane is the volume, (1) the cannula needs to be translated and rotated angle of rotation between the cannula and the casing. relative to the casing, and (2) the stylet be translated relative to = c the cannula. An illustration of the kinematics of the cannula (5) and stylet is shown in Fig. 3 with the distal tip of the stylet Inverse Kinematics defined in standard Cartesian and cylindrical coordinate Rearranging (3), (4) and (5) the relation between the systems. Position variables to represent the motions of the actuated degrees of freedom to a desired end-point in cannula and stylet are also shown. From the geometry, the cylindrical coordinates can be obtained. necessary equations for determining the appropriate stylet and cannula motions to position the stylet tip to a point in zs = R cos 1 1 (6) R cylindrical and Cartesian (CT) coordinate systems can be derived, where the radius of curvature of the distal portion of z zc = z R sin s ) the stylet, R, is assumed to remain constant when outside of the R (7) cannula.

(8) If a desired end-point of the stylet is first specified in Cartesian coordinates, then this point can first be transformed to cylindrical coordinates before the cannula and stylet positions are calculated. C. Stylet Design and Manufacture The important dimensions pertaining to the design of the steering system are the stylet total bend angle, stylet radius of curvature and diameters of the stylet and cannula. These dimensions determine the working volume that can be targeted, the necessary material constraints, the forces necessary to move the stylet relative to the cannula and the stiffness of the stylet. It was found that due to the large strains in the cannula when they were retracted inside the cannula that the stylets need to be constructed from a superelastic material such as Nitinol. Nitinol was obtained in straight wire form (Forte Wayne Metals, IN, USA) and heat treatment was chosen as the means of imparting a curve into the distal end. The process for heat setting Nitinol has been previously reported [13] and involves evenly heating the material to an annealing temperature of 550C where it is maintained for 3 to 15 minutes until internal

c =

Fig. 3 Coordinate system and position variables for cannula and stylet. In order to position the distal tip of the stylet in a volume then three degrees of freedom have to be controlled; zc/g, the axial position of the cannula with respect to ground (i.e. the casing); c/g, the angle of rotation between the cannula and the casing and zs/c, the axial position of the stylet with respect to the cannula.

4 A. Telerobot Concept stresses have been relieved, followed by a quenching operation to maintain the material in the desired Austenitic Phase (Af). The concept chosen was for a telerobot that could be either The fixture, shown in Fig. 4, was designed and built for this attached to an access cannula that is first placed using a purpose and provided minimal thermal resistance to ensure patient-mounted robot such as [18] or could be placed using a rapid quenching. robotic manipulator arm (e.g. Neuromate Neuro Surgical Robot, Renishaw, UK) as illustrated in Fig. 5. The connection to the access cannula would be via a standard medical leurlock connector. This approach could enable the system to be used within the size constraints of a medical imaging system as the overall height requirement is reduced by first placing the access cannula. However, if placed manually, a means of supporting the access cannula and the mass of the telerobot will be required.

Fig. 4 Nitinol quench fixture components used for heat setting nitinol wire. The outline and vent plates maintain the Nitinol in its final desired curved shape while allowing water to pass through for quenching. The stiffening frames prevent the thinner plates from bowing outward from the thermal shock of the water quench.

Stylets with bend radii ranging from 10-40 mm and diameters of 0.508-0.990 mm were tested with 14, 16, 18 and 20 gauge cannulae in the experimental setup described in [23]. The ranges for the deployment and retraction forces were found to be 0.2-7.0 N and 0.3-14.0 N respectively. III. TELEROBOT DESIGN In order to enable the distal tip of the stylet to be accurately and reproducibly positioned within a working volume, a telerobot capable of axially advancing and rotating a cannula and axially advancing a stylet through the cannula was designed and built. As previously mentioned, the data collected in [23] found a maximum force of 14 N to retract a curved Nitinol stylet into a cannula. Previously reported results of inserting needles into tissue range from 2.3 N to 15.6 N [24, 25] and depend of the size and type of needle used; thus, these maximum values were used as design specifications. The functional requirements were that the telerobot should have sufficient force to insert the cannula into tissue (15.6 N), provide sufficient force to move the stylet relative to the cannula (14 N), actuate the appropriate degrees of freedom (3) with 1 mm translation and 1 degree angular accuracy, be compatible with CT machines (i.e. fit within the CT bore and not substantially degrade images with metal artifacts), and be sufficiently lightweight so that it can be mounted directly to the patient if required.

Fig. 5 Concepts for application of the needle steering system. In A), the telerobot is attached to a previously inserted access cannula. In B), the needle steering system is attached to a robot arm.

B. Mechanism Design The mechanism design, largely of plastic components for CT compatibility, is shown in Fig 6. The device has a protruding cannula and a stylet with a curved distal tip preassembled inside. The proximal end of the cannula is attached to the distal end of a hollow screw-spline and the proximal end of the stylet is attached to the distal end of a screw that fits inside the screw-spline. Each attachment is achieved via plastic threaded inserts that are bonded to the proximal end of the shafts. The length of the cannula and stylet are chosen so as to be positioned at the distal tip of a 10 cm access cannula when the parts are connected via a standard medical leur-lock. The screw-spline is a custom plastic ACME threaded screw that also has a splined groove along its length. It is functionally similar to the ball-screw spline that is produced by THK that has been used in SCARA robots (e.g. EPSON RS3-Series, EPSON Robots, CA 90746) and other robotic applications where the combination of translation and rotation are required in a compact design [26, 27]. Fig. 7 shows nuts 1 and 2 engage the screw threads and spline respectively. Nut 1 has a bore that is threaded to match the lead of the screw and nut 2 had a slot broached into the inside diameter that allowed a small plastic 1.5 mm wide key to be inserted. This key then engaged the splined groove on the cannula screw-spline.

5
threaded nut engages the screw that rides inside the screw-spline. The nuts are bonded to the inside race of ceramic ball bearings and have teeth that engage spur gears on the corresponding gearhead shafts. The cannula and stylet have threaded rod bonded to their proximal ends to connect to the screw-spline and screw respectively.

By appropriate control of the two nuts, three modes of operation of the screw-spline, and hence cannula, with respect to the casing could be obtained as shown in Fig. 8.

Fig. 8 The cannula axial and rotational position is controlled by the screwspline. Through appropriate control of the two nuts three modes of operation can be obtained. The helix mode is not required for this application.

Fig. 6 CAD image of the telerobotic needle steering system. The cannula is attached to a screw-spline (turquoise) that can be translated and rotated through control of two stepper motors grounded to the casing (brown). The stylet is attached to a screw (red) that can be translated with respect to the screw-spline through control of a motor that is grounded to a stage that is fixed to the screwspline.

Fig. 7 Section view of the drive mechanism. A spline nut and screw nut engage the spline and threads of the screw-spline respectively. Another

Translation of the stylet is achieved in a similar way in that it is a second ACME threaded screw with a splined groove. A keyed feature on the inside of the screw-spline mates with the splined groove to constrain it from rotating with respect to it as shown in Fig. 7. The axial travel of both the cannula screwspline and stylet screw were chosen to be 40 mm and the cannula screw-spline was capable of 360 degrees of rotation. This resulted in a total volume of 113 cm3 that could be targeted by the telerobot. The stepper motors were clamped to their respective mounting plates so to aid with gear alignment and any necessary repairs. The design was such that the same plastic clamp could be used for each of the motors. The stepper motors and planetary gearheads used for actuation will result in shadowing artifacts; however, placement of them away from the central axis of the cannula and stylet, minimized distortion to the images. A cylindrical precision ground tube (McMaster-Carr) was used to house all of the components. The tube has an internal tolerance of 0.0005 and so provides the secondary purpose of a guide way for the flanges of the cannula screw-spline assembly. In order to reduce friction between the flanges and inside of the casing, 4-48 spring plungers (McMaster) were threaded into the side of the flanges to provide spring-loaded rolling contact. A standard -24 plastic leur-lock (McMaster) was threaded into base plate for attaching to standard medical needles for access to the target site. The system height and diameter are 15 cm and 5 cm respectively with a weight of approximately 180 grams. An image of the prototype is shown in Fig. 9. We have included supplementary video files which demonstrate both the screw-spline ability to translate pure translation and rotation as well as the telerobot to reposition the distal tip of the stylet to multiple points in a volume. This information will be available at http://ieeexplore.ieee.org.

d m sec

(10)

Thus, by choosing an appropriate lead and diameter for the screws the cannula can retain its axial position when the stylet is being translated and vice versa. Other important considerations when sizing the transmission are its efficiency and the maximum stress it will experience. If T0 is the torque achievable assuming no frictional losses due to sliding contact between the threads the efficiency, e, for power transmission with a screw is given by T Fl e= 0 = (11) T 2 T The equivalent von Mises stress in the screws is calculated Fig. 9 Prototype of the telerobot shown attached to an access cannula via a by combining the axial and shear stresses. The axial stress in a standard medical leur-lock. The stylet is shown in its deployed position and the screw due to a load, F, is given by cannula is protruding about 1 cm from the distal tip of the access cannula. 4F C. Mechanism Actuation and Transmission Selection = (12) (d r di ) 2 To simplify control, it is desirable to use micro stepper motors to actuate the nuts of the screw-spline and screw. Each where dr and di are the root and inside diameter of the of the nuts had an integral flange with spur gear teeth that leadscrew respectively. For a solid screw, the nominal shear mated with a spur gear attached to the gearhead shaft. stress in torsion, T, in the leadscrew is given by Although not CT compatible themselves due to their metallic 16T = (13) construction, two of the motors (for the screw-spline) were dr 3 positioned so as to yield a metallic free scan plane. The third motor for actuating the stylet screw was also positioned away Assuming a stress concentration of two, the von Mises from the center of the device; however, depending on the equivalent stress can be calculated from rotational position of the screw-spline, it will occasionally (14) equivelent = 2* 2 + 3 2 show up in the CT image. Equations (11) and (14) are plotted in Fig. 10 that shows a A design specification for the device was to change the smaller diameter and lead results in rapidly increasing stress angle of the cannula with respect to the casing in degree increments. Micro stepper motors have relatively large step levels in the screw for a desired axial force of 29.6 N (force to angles (typically 15-18 degrees) and so it was necessary to move stylet relative to cannula and force into tissue). A larger have a gear reduction between the motor shaft and the screw- diameter screw or lead will reduce the stress in the screw (the spline. Use of a gear reduction between the motor output and latter by reducing the torque requirement); however, it will screw and spline nuts also meant that a stepper motor with a also result in a reduced efficiency for the screw transmission. lower holding torque could be used to constrain the screw- Given this trade-off, a lead of 1/16 inch (1.5875mm) was spline from rotating during a commanded pure translation. chosen for the stylet screw and cannula screw-spline. A stylet As well as the torque required to exert sufficient axial force, diameter of 6 mm was chosen for the stylet so that it would be it is also important to determine the necessary torque to sufficiently strong for the loads involved while maximizing overcome frictional forces due to sliding contact between the efficiency. threads and any thrust bearings. Neglecting friction forces arising from bearings, the general equation for calculating the torque to raise a load is Fd m l + d m sec T= (9) 2 d m l sec Where F is the desired maximum force, dm is the pitch diameter of the lead screw, l is the lead, is the coefficient of friction between the threads and is the ACME thread angle (i.e. 29). Using a simple sliding test, the coefficient of sliding friction of Acetal on Acetal was found to be 0.2. During operation of the device the cannula and stylet will be activated independently. An ACME screw is not backdrivable if the effective coefficient of thread friction is equal or greater than the tangent of the helix angle, i.e.
Fig. 10 A plot of maximum equivalent von Mises stress and transmission efficiency as a function of screw diameter and lead.

7
2T As can be seen from Fig. 10, the stylet was well below the (16) Fradial = dp maximum allowable stress; however a reduced stiffness would have made prototyping the screw difficult on a lathe. The Further, a pressure angle , between the gear teeth generates a screw-spline diameter was chosen to be 11 mm so that it just force that acts to spread the gears apart allowed the stylet screw to concentrically nest inside it. A (17) Fspread = Fradial sin summary of the results of the design calculations for the The radial load on the gearhead shaft was calculated to be cannula screw-spline and stylet screw are shown in Table 1. 13 N with a force of 4.8 N tending to spread the gears apart. TABLE 1 Thus, to ensure that the gearhead shafts were not overloaded, a SUMMARY OF THE TRANSMISSION CHARACTERISTICS FOR THE CANNULA hubbed spur gear was used so that the gear teeth were placed SCREW-SPLINE AND STYLET SCREW TO ACHIEVE THE DESIRED FORCES OF 15 N in line with the sintered brass bearings in the planetary AND 29.6 N RESPECTIVELY. gearhead as shown in Fig. 7. Ceramic bearings were used for Screw-Spline Screw attaching each of the nuts to their respective mounting plates Torque Required [Nmm] 23.9 28.1 Maximum Stress [MPa] 0.7 6.9 (VXB Bearing Inc.). These bearings provided a compact nonEfficiency [%] 15.9 26.6 metallic package with sufficient axial and radial stiffness to Backdrivability NO NO ensure that the gear teeth always remained engaged. Ceramic The stepper motors and gearheads selected for this ball bearings also offer the advantages of low friction (so no application were 10 mm diameter (AM1020, MicroMo lubrication is required) and are significantly lighter and quieter Electronics (Faulhaber Group), FL). The motors have a than metal bearings. Plastic ball bearings were also considered holding torque of 1.6 Nmm and a two-stage planetary gearhead but they have significantly larger height and outside diameter with 16:1 reduction was chosen to meet the torque for a given internal diameter as well as approximately 1 mm of requirement. There was also a further gear reduction between axial play. the gearhead shaft and the nuts for the cannula screw-spline E. Control (2:1) and stylet screw (1.875:1). The nuts for the screw and The current system is designed to be controlled based on screw-spline were manufactured from off-the-shelf Acetal spur gears (SDP-SI, NY); 24 mm pitch diameter for the cannula open-loop command signals sent to the stepper motors. The screw-spline nuts and 22.5 mm pitch diameter for the stylet appropriate kinematic equations were used to relate the motor screw. The bending stress in the gear teeth was calculated angular positions to the distal tip of the stylet. A simple user using the basic Lewis equation. The maximum value was 6.9 interface was developed that allows the desired position within kN/m2; almost four orders of magnitude below the maximum a volume to be specified. User inputs may be in cylindrical or allowable stress for Acetal. Stainless Steel spur gears with a Cartesian coordinates that are then converted into desired pitch diameter of 12 mm (SDP-SI, NY) were used to transmit position and speed commands sent to the controller. The drive power from the planetary gearhead to the plastic spur-gear electronics for the system are contained in a control box that nuts. The efficiency of the gearbox and spur gears were would be located away from the patient on the CT bed. It is approximately 80% and 90% respectively. The stepper motor plugged into a standard 120 V wall outlet and connected via a had a step angle of 18 degrees and so this gear reduction also USB cable to a laptop. Inside the box are offtheshelf yielded a minimum step angle for the screw-spline of just over components; a USB stepper motor controller (PMX-4CX-SA, half a degree (18/32), fulfilling the design specifications. ARCUS Technology Inc., CA, USA), power supply (S-100F, The keyed feature for each of the splined grooves on the stylet and cannula screws was sized to ensure that it would be Astrodyne, MA, USA) and four stepper motor drivers (RDable to withstand the shear stresses generated due to torque 021M8, RORZE, CA, USA).

transmission which is given by 2T shear = dk A

IV. SYSTEM EVALUATION (15) Before evaluating the ability of the device to reposition the distal tip of the stylet within a volume, experiments were performed to compare the force and accuracy performance to the design specifications. An experimental rig was constructed that provided a solid mounting platform for the mechanism. The rig consisted of a boxed aluminum frame with movable walls for mounting the mechanism and either a load cell or potentiometer.

Where dk is the diameter where the key engages to transmit the torque and A is the area. Thus, based on the maximum motor torque, a key 1.4 mm wide and 8 mm long will experience a shear stress of 2 MPa which is well below the maximum allowable stress in Acetal.

D. Bearings The axial and radial loads on the nut and gearhead bearings A. Force and Torque Measurements were also calculated to ensure that the rated load capacities of An image-of the setup is shown in Fig. 11 with a load cell the bearings were not exceeded. The torque transmitted at a (LCM200, Futek, CA, USA) mounted for measuring the pitch diameter, dp, results in a radial load, Fradial, on the shaft maximum force that can be achieved by the screw-spline. An on which the gear is mounted, given by amplifier (CSG105, Futek, CA, USA) was used to amplify the signal before it was read into a data acquisition board (NI

8 USB-6009, National Instruments, TX, USA) at a sampling rate of 500 Hz. The data was filtered with a 10th order digital butterworth filter in Matlab with a frequency of 10Hz and the average over a 5 second period was calculated. The load cell and potentiometer were both mounted concentrically with the screw-spline so as to be able to measure pure axial forces and rotations respectively. commanding the screw-spline to rotate in these increments while recording data from the potentiometer. The data obtained with the calipers and potentiometer is summarized in Table 3 showing that the mechanism has sub-millimeter and subdegree accuracy and repeatability.
TABLE 3 COMPARISON OF AXIAL AND ANGULAR MOVEMENTS IN RESPONSE TO RESPECTIVE COMMANDS. FOR EACH RESULT, THE AVERAGE OF 10 MEASUREMENTS WAS CALCULATED.

Commanded Cannula Screw-Spline Translation [mm] Rotation [deg] Rotation [deg] Rotation [deg] Stylet Screw Translation [mm] Translation [mm] Translation [mm] 5 5 10 20 5 20 25

Measured 5.01 0.02 4.97 0.22 10.15 0.36 19.84 0.47 4.96 0.06 20.02 0.06 24.97 0.08

Fig. 11 Experimental rig that was used to mount the telerobot mechanism. The rig allowed a potentiometer of load cell to be mounted concentrically with the cannula screw-spline or stylet screw.

The load cell was calibrated using weights of known increments. A spring scale and a lever were used for measuring the maximum torque that could be transmitted by the screwspline. The moment arm of the lever was 10 mm and the spring scale could measure axial force in increments of 40 grams. A summary of the experimental measurements compared to the predicted values is shown in Table 2.
TABLE 2 COMPARISON OF THE MAXIMUM FORCE AND TORQUE CAPACITIES OF THE CANNULA SCREW-SPLINE AND STYLET SCREW TO THEIR PREDICTED VALUES.

C. Translation and Rotation Coupling of Screw-Spline The screw-spline couples the rotation and translation of the cannula and so the degree to which axial position was effected by commanded rotation of the screw spline was evaluated. This was accomplished by first measuring the current axial position of the screw-spline and then by commanding the screw-spline to rotate a set number of times before measuring the position again. For each axial position measurement the average of 10 readings was taken to account for human error. Fig. 12 illustrates that deviation of the axial position of the screw-spline is less than 0.2 mm for up to 15000 rotations.

Predicted Cannula Screw-Spline Axial Force [N] Drive Torque [Nmm] Holding Torque [Nmm] Stylet Screw Axial Force [N] 28 37 51 43

Measured 26 25 59 25

B. Translation and Rotation Accuracy A linear caliper (CD-6CS, Mitutoyo) was used to measure the position of the stylet screw and cannula screw-spline after they were commanded to move to certain positions. The typical resolution and repeatability of the caliper is 0.01 mm. The tail end of the calipers fitted into the slots of both the screw and screw spline and so was able to be repeatably positioned for each measurement. A total of 10 measurements were recorded for each position and the average value was taken. The typical standard deviation of the 10 readings for any one position was 0.05 mm. A rotary potentiometer was used to measure the rotational accuracy of the screw spline. The potentiometer was calibrated using a protractor. The potentiometer was powered using 10 V and the data was sampled at a frequency of 500 Hz and then filtered with a 10th order digital butterworth filter in Matlab with a cut-off frequency of 10 Hz. The ability of the screw-spline to command rotations of 5, 10 and 20 degrees was evaluated by

Fig. 12 Axial position of the screw-spline as a function of pure rotational commands. Ten measurements were taken for each data point and the standard deviation for each is plotted.

V. CONCLUSIONS AND FUTURE WORK In this paper we have presented the design of a telerobot capable of repositioning the distal tip of a percutaneous instrument, after it has been inserted into the body. The steering mechanism consisted of deploying a flexible stylet with a pre-curved distal tip from a concentric straight cannula. Experimental measurements of the accuracy and force for the cannula screw-spline and stylet screw mechanisms were found to be those predicted in the design process and were sufficient

9 for actuating the necessary motions of the cannula and stylet so that the distal tip of the stylet could be accurately positioned in a volume. For biopsy, an increased targeting capability and an ability to accurately target multiple locations within a single lesion would allow for higher diagnostic rates as well as a reduction in procedure time and radiation dose for the patient. As well as correcting for targeting errors due to instrument deflection or tissue deformation, for the instrument could greatly simplify local treatments such as radiofrequency ablation. Instead of a large diameter probe, or a probe with multiple electrodes, a thin probe with a small burn volume could be inserted into the tumor and then robotically steered so as to raster-scan through the tumor after a single needle insertion through the skin. Such a system might enable treatment to be applied to multiple small but overlapping volumes allowing conformation to the tumor morphology, while avoiding damaging any delicate neighboring tissue. Finer treatment margins as well as the ability to treat multiple tumors, satellite lesions or tumors that cannot be accessed along a straight path are significant advantages of such an approach. The current system was controlled using a custom interface; ultimately this device will be controlled by an interface that directly incorporates the images in a similar manner to [26] that allows a physician to configure the target points in a pointand-click manner. ACKNOWLEDGMENT The authors wish to express their sincere appreciation to Julio Geurrero, Dave Custer, Rick Slocum and all members of the MIT Precision Engineering Research Group (PERG) for their feedback during the design process. REFERENCES
M. Mathis, D. Thompson, B. Addis, and D. Yankelevitz, "Steerable Device for Accessing a Target Site and Methods," vol. US 2006/0167416. [2] R. Ebrahimi, S. Okazawa, R. Rohling, and S. E. Salcudean, "Hand-Held Steerable Needle Device," in Medical Image Computing and ComputerAssisted Intervention - MICCAI 2003, 2003, pp. 223-230. [3] S. Okazawa, R. Ebrahimi, J. Chuang, S. E. Salcudean, and R. Rohling, "Hand-held steerable needle device," Mechatronics, IEEE/ASME Transactions on, vol. 10, pp. 285-296, 2005. [4] S. P. DiMaio and S. E. Salcudean, "Needle insertion modelling and simulation," in Robotics and Automation, 2002. Proceedings. ICRA '02. IEEE International Conference on, 2002, pp. 2098-2105 vol.2. [5] R. J. Webster, III, J. S. Kim, N. J. Cowan, G. S. Chirikjian, and A. M. Okamura, "Nonholonomic Modeling of Needle Steering," Int. J. Rob. Res., vol. 25, pp. 509-525, 2006. [6] J. A. Engh, G. Podnar, S. Y. Khoo, and C. N. Riviere, "Flexible Needle Steering System for Percutaneous Access to Deep Zones of the Brain," in Bioengineering Conference, 2006. Proceedings of the IEEE 32nd Annual Northeast, 2006, pp. 103-104. [7] D. S. Minhas, J. A. Engh, M. M. Fenske, and C. N. Riviere, "Modeling of Needle Steering via Duty-Cycled Spinning," in Engineering in Medicine and Biology Society, 2007. EMBS 2007. 29th Annual International Conference of the IEEE, 2007, pp. 2756-2759. [8] R. L. Pakter and E. M. Morris, "Hollow, curved, superlastic medical needle," vol. US Patent No. 6,592,559 B1. [9] W. R. Daum, "Deflectable Needle Assembly," vol. US Patent No. 6,572,593. [10] M. Loser, "A new robotic system for visually controlled percutaneous [1] interventions under X-ray or CT-fluoroscopy." vol. M.S. Thesis Freiburg, Germany: Albert-Ludwig-Univ.. 2002. R. J. Webster, J. M. Romano, and N. J. Cowan, "Mechanics of PrecurvedTube Continuum Robots," Robotics, IEEE Transactions on, vol. 25, pp. 67-78, 2009. P. Sears and P. Dupont, "A Steerable Needle Technology Using Curved Concentric Tubes," in Intelligent Robots and Systems, 2006 IEEE/RSJ International Conference on, 2006, pp. 2850-2856. A. Bzostek, A. C. Barnes, R. Kumar, J. H. Anderson, and R. H. Taylor, "A Testbed System for Robotically Assisted Percutaneous Pattern Therapy," Medical Image Computing and Computer-Assisted Intervention MICCAI99, vol. 1679/1999, pp. 1098-1107, 1999. D. Stoianovici, K. Cleary, A. Patriciu, D. Mazilu, A. Stanimir, N. Craciunoiu, V. Watson, and L. Kavoussi, "AcuBot: a robot for radiological interventions," Robotics and Automation, IEEE Transactions on, vol. 19, pp. 927-930, 2003. M. Rasmus, S. Dziergwa, T. Haas, P. Madoerin, H. R, D. Bilecen, and A. Jacob, "Preliminary clinical results with the MRI-compatible guiding system INNOMOTION," Int J CARS vol. 2 pp. S138-S145, 2007. E. Taillant, J.-C. Avila-Vilchis, C. Allegrini, I. Bricault, and P. Cinquin, "CT and MR Compatible Light Puncture Robot: Architectural Design and First Experiments," in Medical Image Computing and Computer-Assisted Intervention MICCAI 2004, 2004, pp. 145-152. B. Maurin, B. Bayle, J. Gangloff, P. Zanne, M. de Mathelin, and O. Piccin, "A robotized positioning platform guided by computed tomography: practical issues and evaluation," in Robotics and Automation, 2006. ICRA 2006. Proceedings 2006 IEEE International Conference on, 2006, pp. 251-256. C. Walsh, N. Hanumara, A. Slocum, J.-A. Shepard, and R. Gupta, "A Patient-Mounted, Telerobotic Tool for CT-Guided Percutaneous Interventions," ASME Journal of Medical Devices, vol. In Press, 2008. R. H. Taylor, R. H. Taylor, J. Funda, B. Eldridge, S. A. G. S. Gomory, K. A. G. K. Gruben, D. A. L. D. LaRose, M. A. T. M. Talamini, L. A. K. L. Kavoussi, and J. A. A. J. Anderson, "A telerobotic assistant for laparoscopic surgery," Engineering in Medicine and Biology Magazine, IEEE, vol. 14, pp. 279-288, 1995. N. Hata, N. Hata, R. Hashimoto, J. Tokuda, and S. A. M. S. Morikawa, "Needle Guiding Robot for MR-guided Microwave Thermotherapy of Liver Tumor using Motorized Remote-Center-of-Motion Constraint," in Robotics and Automation, 2005. ICRA 2005. Proceedings of the 2005 IEEE International Conference on, 2005, pp. 1652-1656. E. Boctor, R. r. Webster, H. Mathieu, A. Okamura, and G. Fichtinger, "Virtual remote center of motion control for needle placement robots.," Comput Aided Surg, vol. 9, pp. 175-83, 2004. K. Cleary, B. Zigmund, F. Banovac, C. White, and D. Stoianovici, "Robotically assisted lung biopsy under CT fluoroscopy: Lung cancer screening and phantom study," International Congress Series, vol. 1281, pp. 740-745, 2005. J. Franklin, "Characterization of a pre-curved needle for use in Distal Tip Manipulation Mechanism," in Mechanical Engineering. vol. S.B Thesis Cambridge: MIT, 2009. A. M. Okamura, C. Simone, and M. D. O'Leary, "Force modeling for needle insertion into soft tissue," Biomedical Engineering, IEEE Transactions on, vol. 51, pp. 1707-1716, 2004. T. Podder, D. Clark, J. Sherman, D. Fuller, E. Messing, D. Rubens, J. Strang, R. Brasacchio, L. Liao, W.-S. Ng, and Y. Yu, "In vivo motion and force measurement of surgical needle intervention during prostate brachytherapy," Medical Physics, vol. 33, pp. 2915-2922, 2006. A. Seitel, C. J. Walsh, N. C. Hanumara, J.-A. Shepard, A. H. Slocum, H.P. Meinzer, R. Gupta, and L. Maier-Hein, "Development and evaluation of a new image-based user interface for robot-assisted needle placements with the Robopsy system," in Medical Imaging 2009: Visualization, Image-Guided Procedures, and Modeling, Lake Buena Vista, FL, USA, 2009, pp. 72610X-9. Conor James Walsh received his B.A.I and B.A. degrees in Mechanical and Manufacturing engineering from Trinity College Dublin, Ireland in 2003 and an M.S. degree in Mechanical Engineering from the Massachusetts Institute of Technology in 2006. He is currently a Ph.D. candidate in the Mechanical Engineering Dept. with a minor in

[11]

[12]

[13]

[14]

[15]

[16]

[17]

[18]

[19]

[20]

[21]

[22]

[23]

[24]

[25]

[26]

10
entrepreneurship through the Sloan School of Management and also a Certificate in Medical Science through the Harvard-MIT Division of Health Sciences and Technology. During his time at MIT, he has received the Whitaker Health Sciences Fund Fellowship as well as numerous other design, entrepreneurship and mentoring awards. His career focus is on research efforts at the intersection of science, engineering and medicine with a focus on smart medical devices. Jeremy C. Franklin received his B.S. degree in Mechanical Engineering from the Massachusetts Institute of Technology in 2009. He is currently a Product Design Engineer at Apple, Inc. in Cupertino, CA doing design, development and manufacturing work for desktop computer products. During his time at MIT, his studies focused on mechanical design in robotics, automotive and consumer products. He is a member of the MIT Gordon Engineering Leadership Program, has received awards in mechanical design and presentation competitions, and has worked in academic and commercial engineering programs in Europe, Asia, and Africa. Alexander H. Slocum received his B.S., M.S. and Ph.D in Mechanical Engineering from the Massachusetts Institute of Technology in 1982, 1983 and 1985 respectively. He is the Pappalardo Professor of Mechanical Engineering at MIT. He has written two books on machine design Precision Machine Design and FUNdaMENTALs of Design (free download on http://web.mit.edu/2.75), published more than 150 papers, and has seven dozen+ patents issued/pending. He regularly works with companies on the development of new products and has been significantly involved with the invention and development of 11 products that have been awarded R&D 100 awards. He is a Fellow of the ASME and the recipient of the Society of Manufacturing Engineers Frederick W. Taylor Research Medal, ASME Leonardo daVinci Award, and the ASME Machine Design Award. Dr. Rajiv Gupta received his BE (Hons) and MSc (Hons) degrees in Electrical Engineering and Physics from the Birla Institute of Technology and Science, and his MS and Ph.D. in Computer Science from the State University of New York and his MD from Cornell University in 2001. He worked for 12 years in academia and industry (University of Southern California, Los Angeles and GE Global Research Center, Niskayuna), before returning to school and receiving his MD in 2001 from Cornell University. He completed his residency and fellowships in Cardiac and Neuro Radiology at the Massachusetts General Hospital/Harvard Medical School. He is currently an Assistant Radiologist at MGH, and is the Director of the Ultra-high Resolution Volume CT Lab.