Intraoperative computer-assisted prediction of intraarticular contact pressures in the knee during high tibial osteotomy.

OBJECTIVES: To develop an accurate intraoperative method to estimate changes in intraarticular contact pressures during high tibial osteotomy (HTO). METHODS: Changes in knee alignment and pressure were monitored in real time in seven cadaver specimens that received HTO. Intraarticular contact pressure (N/mm2 ) in each knee compartment was estimated based on extraarticularly acquired data (leg alignment, correction, and ankle tilt) and based on the application of an axial force of half bodyweight (400-450 N). RESULTS: Contact pressure estimation was more accurate in the lateral compartment (R2  = 0.940) than in the medial compartment of the knee (R2  = 0.835). The optimism-corrected R2 was 0.936 for the lateral compartment and 0.821 for the medial compartment. CONCLUSIONS: We have established a framework for estimating the change in intraarticular contact pressures based on extraarticular data. This research could be helpful in generating appropriate algorithms to estimate joint alignment changes based on applied loads.

Biomechanical and histological analyses of the fracture healing process after direct or prolonged reduction.

BACKGROUND: Reduction of femoral shaft fractures remains a challenging problem in orthopaedic surgery. Robot-assisted reduction might ease reduction and fracture treatment. However, the influence of different reduction pathways on patients' physiology is not fully known yet. Therefore, the aim of this study was to examine the biomechanics and histology of fracture healing after direct and prolonged robot-assisted reduction in an in vivo rat model. METHODS: 144 male CD® rats were randomly assigned to 12 groups. Each animal received an external fixator and an osteotomy on the left femoral shaft. On the fourth postoperative day, the 1× reduction groups received a single reduction maneuver, whereas the 10× reduction groups received the same reduction pathway with ten repetitions. The control groups did not undergo any reduction maneuvers. Animals were killed after 1, 2, 3 and 4 weeks, respectively, and the composition of the fracture gap was analyzed by µCT and non-decalcified histology. Biomechanical properties were investigated by a three-point bending test, and the bone turnover markers PINP, bCTx, OPG, sRANKL, TRACP-5b, BALP, and OT/BGP were measured. RESULTS: One week after the reduction maneuver, µCT analysis showed a higher cortical bone volume in the 1× reduction group compared to the 10× reduction group. Biomechanically, the control group showed higher maximum force values measured by three-point bending test compared to both reduction groups. Furthermore, less collagen I formation was examined in the 10× reduction group compared to the control group after 1 week of fracture healing. PINP concentration was decreased in 10× reduction group after 1 week compared to control group. The same trend was seen after 3 weeks. CONCLUSION: A single reduction maneuver has a beneficial effect in the early phase of the fracture healing process compared to repeated reduction maneuvers. In the later phase of fracture healing, no differences were found between the groups.

Improving the human-robot interface for telemanipulated robotic long bone fracture reduction: Joystick device vs. haptic manipulator.

OBJECTIVES: Intramedullary nailing is the treatment of choice for femoral shaft fractures. However, there are several problems associated with the technique, e.g. high radiation exposure and rotational malalignment. Experimental robotic assistance has been introduced to improve the quality of the reduction and to reduce the incidence of rotational malalignment. In the current study, we compare two devices for control of the fracture fragments during telemanipulated reduction. METHODS: Ten male and ten female subjects were asked to participate as examiners in this experiment. A computer program was developed to render and manipulate CT-based renderings of femur fracture bone fragments. The user could manipulate the fragments using either a simple joystick device or a haptic manipulator. Each examiner performed telemanipulated reduction of 10 virtual fracture models of varying difficulty with each device (five in a 'training phase' and five in a 'testing phase'). Mixed models were used to test whether using the haptic device improved alignment accuracy and improved reduction times compared to using a joystick. RESULTS: Reduction accuracy was not significantly different between devices in either the training phase or the testing phase (P > 0.05). Reduction time was significantly higher for the Phantom device than for the Joystick in the training phase (P < 0.0001), but it was no different in the testing phase (P = 0.865). High spatial ability with electronics had a significant effect on the alignment of fracture reduction and time to reduction. CONCLUSIONS: The Joystick and the Phantom devices resulted in similarly accurate reductions, with the Joystick having an easier learning curve. The Phantom device offered no advantage over the Joystick for fracture telemanipulation. Considering the high cost of the Phantom device and the lack of a demonstrable advantage over the Joystick, its use is not justified for implementation in a fracture telemanipulation workflow. The Joystick remains as a low-cost and effective device for developing 3D fracture telemanipulation techniques.

Comparison of algorithms for automated femur fracture reduction.

PURPOSE: We designed an experiment to determine the comparative effectiveness of computer algorithms for performing automated long bone fracture reduction. METHODS: Automated reduction of 10 3D fracture models was performed using two computer algorithms, random sample matching (RANSAM) and Z-buffering (Z-Buffer), and one of five options of post-processing: none; iterative closest point algorithm (ICP); ICP-X1; ICP-X2; and ICP-X3. We measured the final alignment between the two fragments for each algorithm and post-processing option. RESULTS: The RANSAM algorithm combined with postprocessing algorithm ICP-X1 or ICP-X3 resulted in the most accurate fracture reduction in the translational plane. No discernible difference was observed in the rotational plane. Automated reduction had more accurate translational displacement than telemanipulated manual reductions. CONCLUSION: This study supports the use of the RANSAM algorithm for automated fracture reduction procedures. The use of ICP algorithms provides further optimization of the initial reduction.

Repeatability and reproducibility of a telemanipulated fracture reduction system.

We evaluate the inter- and intraobserver variability of a telemanipulated femur fracture reduction system using a joystick device. Five examiners performed virtual reduction of 3D femur fracture models on two separate occasions. We assessed the inter- and intraobserver variability for the final alignment and reduction. The average difference between testing rounds was only 0.3 mm for overall displacement and 0.5° for overall rotation. There was an average time reduction between rounds of 11.7 s. The mean differences in overall displacement between examiners ranged between 0.2 and 0.9 mm; between 0.2° and 3.2° for overall rotation; and between 9 and 82 s for time to reduction. The time required to complete the reduction did not have a significant effect on the overall displacement or rotation of the final model. Telemanipulated fracture reduction is a reliable and reproducible technique, which does not require extensive training.

Robotic technique improves entry point alignment for intramedullary nailing of femur fractures compared to the conventional technique: a cadaveric study.

We aimed to test whether a robotic technique would offer more accurate access to the proximal femoral medullary cavity for insertion of an intramedullary nail compared to the conventional manual technique. The medullary cavity of ten femur specimens was accessed in a conventional fashion using fluoroscopic control. In ten additional femur specimens, ISO-C 3D scans were obtained and a computer program calculated the ideal location of the cavity opening based on the trajectory of the medullary canal. In both techniques, the surgeon opened the cavity using a drill and inserted a radiopaque tube that matched the diameter of the cavity. The mean difference in angle between the proximal opening and the medullary canal in the shaft of the femur was calculated for both groups. Robotic cavity opening was more accurate than the manual technique, with a mean difference in trajectory between the proximal opening and the shaft canal of 2.0° (95% CI 0.6°-3.5°) compared to a mean difference of 4.3° (95% CI 2.11°-6.48°) using the manual technique (P = 0.0218). The robotic technique was more accurate than the manual procedure for identifying the optimal location for opening the medullary canal for insertion of an intramedullary nail. Additional advantages may include a reduction in total radiation exposure, as only one ISO-C 3D scan is needed, as opposed to multiple radiographs when using the manual technique.

Robot-assisted endoscope guidance versus manual endoscope guidance in functional endonasal sinus surgery (FESS).

BACKGROUND: Having one hand occupied with the endoscope is the major disadvantage for the surgeon when it comes to functional endoscopic sinus surgery (FESS). Only the other hand is free to use the surgical instruments. Tiredness or frequent instrument changes can thus lead to shaky endoscopic images. METHODS: We collected the pose data (position and orientation) of the rigid 0° endoscope and all the instruments used in 16 FESS procedures with manual endoscope guidance as well as robot-assisted endoscope guidance. In combination with the DICOM CT data, we tracked the endoscope poses and workspaces using self-developed tracking markers. RESULTS: All surgeries were performed once with the robot and once with the surgeon holding the endoscope. Looking at the durations required, we observed a decrease in the operating time because one surgeon doing all the procedures and so a learning curve occurred what we expected. The visual inspection of the specimens showed no damages to any of the structures outside the paranasal sinuses. CONCLUSION: Robot-assisted endoscope guidance in sinus surgery is possible. Further CT data, however, are desirable for the surgical analysis of a tracker-based navigation within the anatomic borders. Our marker-based tracking of the endoscope as well as the instruments makes an automated endoscope guidance feasible. On the subjective side, we see that RASS brings a relief for the surgeon.

Robotic distal locking of intramedullary nailing: Technical description and cadaveric testing.

Interlocked intramedullary nailing is the treatment of choice for femoral shaft fractures. However, distal locking is a technically challenging part of the procedure that can result in distal femoral malrotation and high radiation exposure. We have tested a robotic procedure for robotic distal locking based on the computation of a drilling trajectory on two calibrated fluoroscopic images. Twenty distal holes were attempted in ten cadaveric femur specimens. Successful screw hole drilling was achieved at the first attempt in each of the ten specimens (20 drill holes in total). No failures were recorded. The average total number of images needed was 6.5 +/- 3.6. The average computation time was 16.5+/- 16.0 seconds. Robotic distal locking was feasible in this test and can be integrated into a fully robotic intramedullary nailing procedure.

The effect of distal tibial rotation during high tibial osteotomy on the contact pressures in the knee and ankle joints.

PURPOSE: Intraoperative fracture of the lateral cortex fractures of the tibia is a potential complication of high tibial osteotomy (HTO), which may result in inadequate rotational alignment of the distal tibia. Our aim was to determine how rotational malalignment of the distal tibial segment distal would affect intraarticular contact pressure distribution in the knee and ankle joints. METHODS: A medial, L-shaped opening-wedge HTO was performed on seven human lower body specimens. A stainless steel device with integrated load cell was used to axially load the leg. Pressure-sensitive sensors were used to measure intraarticular contact pressures. Intraoperative changes in alignment were monitored in real time using computer navigation. Measurements were performed in the native knee alignment, after 10° and 15° of alignment correction and with the distal tibia fixed at 15° of external rotation. RESULTS: Moderate-to-large alignment changes after medial opening-wedge HTO resulted in a shift in intraarticular contact pressures from the medial compartment of the knee towards the lateral compartment. However, fixation of the distal tibial segment at 15° of external rotation neutralized this intended beneficial effect. In the ankle, external rotation of the distal tibia also caused a reduction in contact pressures and tibiotalar contact area. CONCLUSION: Malrotation of the distal tibial fragment negates the intended effect of offloading the diseased compartment of the knee, with the contact pressures remaining similar to those of the native knee. Furthermore, malrotation leads to abnormal ankle contact pressures. Care should be taken to ensure appropriate rotational alignment of the distal tibial segment during intraoperative fixation of HTO procedures.

Repetitive reduction lead to significant elevated IL-6 and decreased IL-10 levels in femoral osteotomies: A quantitative analysis of a robot-assisted reduction process in a rat model.

INTRODUCTION: The field of robot-assisted fracture reduction has been developed by several research groups over more than one decade by now, with the main goals of increasing the fracture reduction accuracy. However, the influence of different reduction paths to patients' physiology is not fully known yet. The aim of our study was to compare the impacts of a robot-assisted direct reduction path versus an artificially prolonged reduction path by measuring the cytokine responses in an in vivo rat model. MATERIALS AND METHODS: Thirty-six male CD(©) rats were assigned into three groups with an external fixator and osteotomy on the left femur. Seven days later, the robot was attached and one group was reduced in a single attempt, while the other group underwent 10 attempts by the robot. The third group was the control group without reduction. Before, and as well as 6, 24 and 48h after the reduction process blood samples were collected. IL-1, IL-6, IL-10, IL-17, and MCP-1 concentrations where analysed via ELISA or cytometric bead assay. Muscle biopsies in the osteotomy area were collected 48h after the reduction process for histological analyses. Statistical significance was set at p≤0.05. RESULTS: Analysis of the cytokines showed that the pro-inflammatory cytokine IL-6 of the Ten-Attempts reduction group significantly increased 6h after reduction compared to the control group. IL-6 further showed markedly elevated levels 6h after surgery in the Ten-Attempts reduction group compared to the Single-Attempt reduction group. On the anti-inflammatory side, IL-10 showed a significant decrease in the Ten-Attempts reduction group 6h after reduction compared to the Single-Attempt reduction and control group. Muscle biopsies showed a significant increase of pathological changes in both reduction groups and an increase in the severity of bleedings of the Ten-Attempts reduction group compared to the Single-Attempt reduction and control group. CONCLUSION: A direct and gentle reduction procedure as feasible by the aid of a robot is preferable over a prolonged reduction in terms of cytokine response and tissue changes.

Workspace and pivot point for robot-assisted endoscope guidance in functional endonasal sinus surgery (FESS).

BACKGROUND: For the further development of robot-assisted endoscope guidance in functional endoscopic sinus surgery (FESS), ground data about the workspaces and endoscope movements in conventional FESS are needed. METHODS: Applying a self-developed marker-based tracking system, we collected the pose data (position and orientation) of the endoscope and all other instruments used in five real sinus surgeries. RESULTS: The automated segmentation of the endoscope poses shows the shape of a hourglass, with a pivot region or pivot point at the 'waistline' of the hourglass, close to the nasal entrance in the nasal dome. CONCLUSION: We were able to identify a pivot point at the waistline of the segmented endoscope poses. The size of the pivot point corresponds with the diameter of the 4 mm endoscope. Because of the reduction to four degrees of freedom for endoscope motions (three rotations and one translation), easier and safer robot-assisted endoscope guidance becomes feasible.

Assessment of the accuracy of infrared and electromagnetic navigation using an industrial robot: Which factors are influencing the accuracy of navigation?

Our objectives were to detect factors that influence the accuracy of surgical navigation (magnitude of deformity, plane of deformity, position of the navigation bases) and compare the accuracy of infrared with electromagnetic navigation. Human cadaveric femora were used. A robot connected with a computer moved one of the bony fragments in a desired direction. The bases of the infrared navigation (BrainLab) and the receivers of the electromagnetic device (Fastrak-Pohlemus) were attached to the proximal and distal parts of the bone. For the first part of the study, deformities were classified in eight groups (e.g., 0 to 5(°)). For the second part, the bases were initially placed near the osteotomy and then far away. The mean absolute differences between both navigation system measurements and the robotic angles were significantly affected by the magnitude of angulation with better accuracy for smaller angulations (p < 0.001). The accuracy of infrared navigation was significantly better in the frontal and sagittal plane. Changing the position of the navigation bases near and far away from the deformity apex had no significant effect on the accuracy of infrared navigation; however, it influenced the accuracy of electromagnetic navigation in the frontal plane (p < 0.001). In conclusion, the use of infrared navigation systems for corrections of small angulation-deformities in the frontal or sagittal plane provides the most accurate results, irrespectively from the positioning of the navigation bases.

A standardized fracture reduction model for long bones--implication and evaluation in the femur.

Fractures of the femoral bone are frequent injuries with a wide range of affected individuals. New treatment strategies and technologies are being explored permanently. Their quality is biomechanically judged by the accuracy of the anatomical reduction. Malalignment of the fragments would have an eminent impact on the overall outcome and rehabilitation. To establish a method for investigations of the reduction results of femoral fractures, we developed a model, using a navigation system for taking measurement. The dynamic reference bases (DRBs) were mounted to the intact femoral bone and registered as the reference position. A special construction allowed removal and reattachment of the DRBs without provoking change in the DRB-bone system. The model was evaluated in its constancy. Translational deviations remained below 0.9 mm and rotational deviations below 0.3° after 40 repetitive reattachments. The model could prove to be valid and reliable. An application in long-bone trauma research is reasonable.

Robotized access to the medullary cavity for intramedullary nailing of the femur.

INTRODUCTION: The insertion site for an antegrade femoral intramedullary nail in the treatment of a femoral shaft fracture has traditionally been performed using a free-hand technique. An inappropriate starting point can result in suboptimal nail insertion leading to malreduction, or iatrogenic fracture. Furthermore, repeated attempts to establish the optimal starting point can cause additional soft tissue trauma and radiation exposure. In the following study we compared a robot-guided technique with the standard free-hand technique for establishing the entry point of an antegrade femoral nail. We hypothesized that the robot-guided technique is more reliable and efficient. METHODS: A custom-made drill-guide was mounted onto the arm of an industrial robot. Two orthogonal fluoroscopic images were acquired from the proximal femur of five cadaveric human specimens. Images were processed with a special software in order to create an enhanced contour-recognition map from which the bone axes were automatically calculated. The drilling trajectory was computed along the extension of the bone-axis. The robot then moved the drill-guide on this trajectory toward the entry point. The drilling was then performed by the surgeon. In the control group, five cadaveric human femora were utilized to manually establish the starting point using the free-hand technique. RESULTS: 100% of the intramedullary cavities were successfully accessed with both the robot-guided and the manual techniques. In the manual technique repositioning of the drill was necessary in three out of five cases. The mean number of acquired fluoroscopic images was significantly reduced from 11.6 (manual) to 4 (robot-guided). CONCLUSION: Robot-assisted drilling of the entry-point in antegrade femoral nailing is more reliable and requires fewer radiographic images than the free hand technique. Yet, based on economical and logistical considerations, its application will probably only be accepted when a concomitant application for fracture reduction is available.

Hands-on robotic distal interlocking in intramedullary nail fixation of femoral shaft fractures.

INTRODUCTION: Intramedullary nailing has become the gold standard in the treatment of femoral shaft fractures. This procedure involves the placement of distal interlocking bolts using the freehand technique. Accurate placement of distal interlocks can be a challenging task, especially in inexperienced hands. Misplacement of distal interlocking bolts can lead to iatrogenic fracture, instability of the bone-implant construct, or even malalignment of the extremity. Repeated drilling attempts increase radiation exposure and can cause additional bony and soft tissue trauma. We hypothesize that robot-guided placement of distal interlocks is more accurate, precise, and efficient than the freehand technique. METHODS: A custom-designed drill guide was mounted onto the arm of an industrial robot. We developed a special device to secure a generic block (Synbone, Malans, Switzerland) into which an intramedullary nail could be inserted in a standardized way. A metric scale allowed later measurements of the drillings. Digital images were taken from each side of the block for analysis of the drilling trajectories. The fluoroscope was adjusted to obtain perfect circles of the distal interlocking holes. The number of images necessary to achieve this was recorded. The axis was recognized automatically by using the differences in contrast between the matrix of the generic bone and the implant (intramedullary nail). The drill trajectories were then computed. The robot with the mounted drill-guide automatically moved onto the calculated trajectory. The surgeon then executed the drilling. We performed 40 robot assisted drillings in generic blocks. Freehand drilling served as our control group. RESULTS: Analysis of the digital images revealed a mean deviation of 0.94 mm and 2.7° off the ideal trajectory using robotic assistance. In 100% of the cases (n = 40), the distal locking hole was hit. A mean of 8.8 images was acquired. After manual drilling, 92.5% of the distal interlocks were hit. A mean deviation of 3.66 mm and 10.36° was measured. A mean of 23.4 fluoroscopic images were needed. The differences between the two methods were statistically significant. CONCLUSION: Robot-guided drilling increases the accuracy and precision of distal interlocking while reducing irradiation. Considering economical and logistical aspects, this application should be integrated with robot-guided fracture reduction.

3D visualized robot assisted reduction of femoral shaft fractures: evaluation in exposed cadaveric bones.

The main problems in intra-medullary nailing of femoral shaft-fractures are leg-length discrepancies and rotational differences with an incidence of 2-18% and 20-40% respectively. These may lead to severe postoperative sequelae such as additional correctional operations and difficult rehabilitation. Insufficient visualization can be considered the main reason for these complications. Finally, retention of the fragments in the correct alignment before nail insertion is difficult. To overcome these problems we established a robotic telemanipulator system to support the reduction process. It was evaluated in 30 fractures of embalmed human femora. Specially programmed software used an image-dataset which was acquired by an isocentric 3D fluoroscope. For visualization, a surface projection was generated. Localization and tracking of the fragments and the robot-arm as well as accuracy measurement was performed by using an optical navigation system. Manipulation was controlled via a force-feedback joystick. This way, collisions of the fragments were transmitted back to the surgeon. At the end of the reduction the robot could rigidly retain the fragments' position.

A rat model for evaluating physiological responses to femoral shaft fracture reduction using a surgical robot.

The first step in treatment of displaced femoral shaft fractures is adequate reduction of the fracture fragments. Manually performed, reduction can be challenging, and is frequently associated with soft tissue damage, especially when repeated reduction attempts are made. The magnitude of local and systemic inflammatory responses caused by prolonged and repeated reduction maneuvers has not been fully established. We devised an operative technique utilizing a robotic reduction device for use in a rat. A femoral fracture was simulated by means of an osteotomy. The robot enabled reproduction of both manual and guided precision reductions, performed in a single path movement. An external fixator was designed specifically to manipulate the rat femur and also for fixation of the osteotomy region. First, reduction accuracy was assessed in eight femurs, then the quality of fixator placement and reduction accuracy was analyzed in 22 femurs. In the first case, 100% of the femurs were accurately reduced. In the second case, 91% had successful stable fixation and an accurate reduction was achieved in 86% of the specimens. We demonstrated the feasibility of a model of robot-assisted fracture reduction that could be used to analyze the effects of reduction on the surrounding soft tissue via biochemical and histopathological means. A future aspect will be to evaluate whether the robot confers an advantage in fracture reduction versus the conventional technique, which would have significant implications for the use of robotic devices in orthopaedic surgery.

Ex vivo evaluation of force data and tissue elasticity for robot-assisted FESS.

The objective of this study was to describe and evaluate soft tissue and bone properties of nasal cavity and paranasal sinuses in ex vivo preparations for a safe robot-assisted endoscope movement during functional endoscopic sinus surgery (FESS). In a first experiment we measured forces exerted by the endoscope during FESS with a force/torque sensor. In a second experiment we used a purpose built device to exert forces on chosen tissue structures. The experiment was monitored by a custom software, which records force of the endoscope and the deformation and the breaking point of tissue. All tests were performed on five formalin fixed cadaver heads. In the first experiment we found that the average force during FESS is 2.21 N and the maximal force is 7.96 N. The force-way-ratio measurement shows highest elasticity for the ethmoidal bulla, followed by the lamina papyracea; however, they break at low forces (> or =6 N). Furthermore the carotid canal seems to have the lowest elasticity but it can tolerate forces up to 30 N. Based on these measurements force thresholds can be defined for robot-assisted endoscope guidance. All thresholds have to be assigned to subregions of the nasal cavity and paranasal sinuses.