This paper describes the simulation of real-time catheter navigation using an electro-mechanical device in our interactive surgical simulation system. The device consists of a position/rotation measurement system, a mechanical system and a micro-controller. The mechanical system houses a set of simulated catheter and guidewire whose feel resembles their actual counterparts. Using optical encoders, the measurement system detects the rotations and displacements of the simulated catheter and guidewire. The movement signals are then analyzed and accumulated by the micro-controller and sent to the workstation host via serial link. The device driver on the host computer further processes these signals and turns them into correctly scaled displacement and rotation values. By manipulating the simulated catheter on the device, the virtual catheter and guidewire are advanced, retracted or twisted in the virtual blood vessels. Accurate navigation of the catheter and guidewire is achieved through the use of efficient finite element methods. The images of the virtual catheter/guidewire and the virtual blood vessels are computed based on the user-specified zoom and pan factors. These images are superimposed on the corresponding fluoroscopic image and displayed in real-time on the computer screen.

Endovascular techniques play a significant role in the management of cerebral AVMs. Currently, flow-guided microcatheters are used for the transfemoral embolization of intracranial AVMs. These catheters are carried by blood flow from their initial position in the neck into the intracranial circulation to the point of greatest flow, which is usually into the feeders of the AVM. Despite this state-of-the-art technology, a significant limitation includes difficulty reaching inaccessible branches secondary to suboptimal placement of the microcatheter. In this report we describe a new device concept to overcome the current limitations of low-guided transfemoral embolization of cerebral AVMs. It involves a magnetic surgery system (MSS) that magnetically manipulates the tip of the microcatheter. The feasibility of this concept was successfully tested using multiple transparent glass intracranial vascular models of the anterior cerebral circulation.

Purpose: To evaluate measurement accuracy of 3D volumetric medical imaging from Spiral CT for craniofacial surgical planing. Material and methods: The study population consisted of 5 cadaver heads that were imaged on a spiral CT scanner with volumetric technique high-resolution contiguous axial slices 3mm thickness and 2mm/sec table feed, with 120Kvp and 200 mA. The archived CT data were stored on optical disks to allow full retrospective review of any image. The data sets were transferred to a networked computer workstation, to generated 3D volumetric images for subsequent manipulation and analyses. The computer graphics workstation allowed to do measurements, based on conventional craniometric anatomic landmarks, by 2 observers with 2 sessions each. The specimens were then submitted to a dynamic blunt force, in an effort to simulate craniofacial fractures, scanned and measured again. The soft tissues were then partially subsequently removed and the measurements were repeated by electromagnetic digitizer. Statistical analysis was done using analysis of variance. Results: Measurements from 3D spiral CT scans can be precise with high repeatability and sufficient accuracy for surgical planing. Conclusion: 3D computer graphics by spiral CT allowed, in vitro, sufficient precision for assessment of surgical management. Digital volumetric spiral CT imaging is valid quantitatively and qualitatively for craniofacial surgical planning and evaluation.

In every surgical procedure there are key steps and skills that, if performed incorrectly, can lead to complications. In conjunction with efforts, based on task and error analysis, in the Videoscopic Training Center at UCSF to identify these key elements in laparoscopic surgical procedures, the authors are developing virtual environments and modeling methods to train the elements. Laparoscopic surgery is particularly demanding of the surgeon's spatial skills, requiring the ability to create 3D mental models and plans while viewing a 2D image. For example, operating a laparoscope with the objective lens angled from the scope axis is a skill that some surgeons have difficulty mastering, even after using the instrument in many procedures. Virtual environments are a promising medium for teaching spatial skills. A kinematically accurate model of an angled laparoscope in an environment of simple targets is being tested in courses for novice and experienced surgeons. Errors in surgery are often due to a misinterpretation of local anatomy compounded with inadequate procedural knowledge. Methods to avoid bile duct injuries in cholecystectomy are being integrated into a deformable environment consisting of the liver, gallbladder, and biliary tree. Novel deformable tissue modeling algorithms based on finite element methods will be used to improve the response of the anatomical models.

Deep nerve regional anesthesiology procedures, such as the celiac plexus block, are challenging to learn. The current training process primarily involves studying anatomy and practicing needle insertion is cadavers. Unfortunately, the training often continues on the first few patients subjected to the care of the new resident. To augment the training, we have developed a virtual reality surgical simulation designed to provide an immersive environment in which an understanding of the complex 3D relationships among the anatomic structures involved can be obtained and the mechanics of the celiac block procedure practiced under realistic conditions. Study of the relevant anatomy is provided by interactive 3D visualization of patient specific data nd the practice simulated using a head mounted display, a 6 degree of freedom tracker, and a haptic feedback device simulating the needle insertion. By training in a controlled environment, the resident may practice procedures repeatedly without the risks associated with actual patient procedures, and may become more adept and confident in the ability to perform nerve blocks. The resident may select a variety of different nerve block procedures to practice, and may place the virtual patient in any desired position and orientation. The preliminary anatomic models used in the simulation have been computed from the Visible Human Male; however, patient specific models may be generated from patient image data, allowing the physician to evaluate, plan, and practice difficult blocks and/or understand variations in anatomy before attempting the procedure on any specific patient.

Retinal laser photocoagulation is a current practice in many eye diseases therapy. Its mastering requires a specific training usually made on actual patients with some risks. The authors present a new device aimed to deliver a complete training separated from the therapeutic practice. This training simulator is built around the actual instrument to comply with the required realism. The instrumental functionalities of the device give the residents the same operating conditions as in the actual practice. The eye fundus visualization is simulated by virtual images, based on actual fundus pictures. They are computed at the rate of 10-12 frames/second according to the adjustments and manipulations of the 3-mirror lens made by the operator. All the pictures are combined in a fundus database planned to collect a wide variety of pathologies. The pedagogical functionalities are gathered in the user's interface. The two major guidelines of the developed software was to achieve an easy to use interface and to enforce no 'school dependent' rules of valuation. This new pedagogical instrument runs on PC micro-computers which allows a low- cost technology and could provide a practical training to retinal photocoagulation without the patient. A clinical validation of its pedagogical efficiency is submitted in another abstract.

Virtual reality is one of these recent technologies which can provide an efficient help in the field of surgical apprenticeship. We achieved an original training simulator for retinal photocoagulation destined to the residents of the ophthalmological department. This paper describes the comparison between this new training tool and the conventional practice. Two groups of residents, randomly selected, were trained exclusively by one of these methods. These two groups were under the responsibility of two distinct experts. A final evaluation was made by a third and different expert, ignoring the training mode practiced by each of the residents. The study lasted six months. The results show that this new training mode is at least as efficient as the current one in terms of elapsed time and efficiency. It may even reduce the training duration. These results confirm that a pedagogical simulator could give a new approach in the medical teaching, particularly in its management. Such a device may solve the problems of practitioner's lack of disponibility and of patients' safety and comfort during a conventional training. Furthermore, it could bring an objective way to value the students; practical ability. On the other hand, this preliminary study emphasizes the difficulties in introducing a new modality in a traditional teaching environment.

Iatrogenic complications in laparoscopic surgery, as in any field, stem from human error. In recent years, cognitive psychologists have developed theories for understanding and analyzing human error, and the application of these principles has decreased error rates in the aviation and nuclear power industries. The purpose of this study was to apply error analysis to laparoscopic surgery and evaluate its potential for preventing complications. Our approach is based on James Reason's framework using a classification of errors according to three performance levels: at the skill- based performance level, slips are caused by attention failures, and lapses result form memory failures. Rule-based mistakes constitute the second level. Knowledge-based mistakes occur at the highest performance level and are caused by shortcomings in conscious processing. These errors committed by the performer 'at the sharp end' occur in typical situations which often times are brought about by already built-in latent system failures. We present a series of case studies in laparoscopic surgery in which errors are classified and the influence of intrinsic failures and extrinsic system flaws are evaluated. Most serious technical errors in lap surgery stem from a rule-based or knowledge- based mistake triggered by cognitive underspecification due to incomplete or illusory visual input information. Error analysis in laparoscopic surgery should be able to improve human performance, and it should detect and help eliminate system flaws. Complication rates in laparoscopic surgery due to technical errors can thus be considerably reduced.

Using image guidance for stereotactic surgery has been widely adopted in neurosurgery, orthopedic surgery and other surgery operations. Careful, precise and robust implementation of image-guidance can offer surgeon accurate intra-operative information that traditional techniques can not reach. Weak design, careless utilization, and dilemma in quality assurance protocol may result in severe scenarios. It is because that introducing image guidance into the operating room involves high precise technologies, delicate instruments and sophisticated processes. These can offer precision as well as space for human errors. A method based on the 'failure modes and effects analysis' is introduced to systematically study human errors in the image-guided surgery field. The paper presented the fundamental steps and architectures of the method. For better understanding of the method, a simple example is also provided. Analyzing human errors with the 'failure mode and effects analysis' benefits the development life cycle of the image-guided surgery system. It also helps for designing the clinical quality assurance process and the training courses for surgeons.

A surgical simulator for needle biopsy of the spine is being developed in the Radiology Department at Georgetown University Medical Center to assist in learning the procedure as well as to maximize accuracy and efficiency. Spine biopsies are often done under computed tomography (CT) guidance and while this technique is effective, it can be time consuming since the biopsy needle must be advanced slowly and its position checked several times to ensure vital organs are not damaged. Quantifying performance during simulation will allow accurate feedback tot eh surgeon as well as the design engineers. Quantifying performance during simulation will allow accurate feedback to the surgeon as well as the design engineers. Performance measures are also important to determine transfer of simulator training skills to actual surgical skills. The simulation protocol is in advanced development, and the steps include selecting the best CT slice for viewing the lesion, choosing the skin entry point, and advancing the needle to the biopsy location. Our methods for developing the system include the following: 1) A task analysis, which produces a detailed list of tasks needed to complete a goal, their order, and time to completion, 2) A function allocation assessment, which identifies critical task components with the goal of relieving the human workload by a reallocation of system functions, and 3) A simulator evaluation through subjective ratings and objective human performance measures.

Cyclic Olefin Polymer (COP)-coated silver hollow glass waveguides were proposed and fabricated for IR laser light transmission such as Er:YAG, CO, and CO₂ lasers. Silver mirror reaction and liquid-flow coating technique were employed to form the silver and dielectric layers inside the fused glass capillary tube for the simplicity, lower cost and potentiality of mass-production. Hollow waveguides with the length of 2m, and the inner diameters of 700 μm and 1mm have been successfully fabricated. It is shown that COP- coated silver hollow glass waveguides can be used to transmit Er:YAG, CO, and CO₂ laser light with low losses when the thickness of COP layer is properly chosen. The measured spectra in the visible and mid-IR regions also show that uniform films were deposited inside the tube. Comparisons were also made among the waveguides that were inner-coated with other dielectric materials used before.

Hollow glass waveguides are an attractive fibers delivery system for a broad range of IR wavelengths, including the 3 μm Er:YAG and 10.6 μm CO₂ lasers. The losses for these waveguides are as low as 0.2 dB/m at the 10.6 wavelength for waveguides with a 700 μm bore. At the shorter wavelengths, like that of the Er:YAG laser, losses are higher than those predicted theoretically. This is shown to be a result of the increasing effect of surface roughness of the inner coatings. Variation of attenuation from waveguide to waveguide is examined, and the variation in the dielectric layer is evaluated.

Recently, hollow wave guides (HWG) have become available enabling CO₂ beam delivery for treatments in less accessible parts of the human body. In this study, an IR fiber delivery system was designed which can easily be attached to and removed from an existing medical laser without interference with its normal functionality. The system accepts SMA terminated HWG with an inner diameter as small as 320 μm . Measurements were performed with 320- 1000 micrometers diameter HWG. The coupling losses were around 10 percent. Transmission losses were up to 2.5 dB/m and depended on the bending radius of the fiber. The fibers tolerated pulse energies up to 400 mJ with peak powers of 800 W and an average power of 10 W. The beam quality was preserved during transmission. Deposited the losses, the output energy easily exceeded the threshold for tissue ablation. The IR fiber delivery system was successfully used in a clinical setting. The flexible 320 μm fiber was introduced through the 1 mm working channel of a 2.3 mm fiber-endoscope with a 90 degree bendable tip. Tissue was effectively ablated with pulses up to 300 mJ out of the fiber. To prevent damage due to pollution of the coating inside the HWG, it will be necessary to close the distal end of the fiber by an IR transmitting material. The IR fiber delivery system has shown to be practical in clinical use without interference with the normal functionality of the CO₂ laser. The system enables endoscopic application of CW and pulsed CO₂ laser light.

Hollow fibers for transmitting high-power CO₂ laser light are fabricated by metal organic chemical vapor deposition method. A dielectric film of metal oxides, such as Y₂O₃, ZrO₂ and Cu₂O deposited inside a Ag-coated glass capillary by using metal acetylacetonate as a precursor. The waveguide coated by Cu₂O with the bore diameter of 700 μm showed the lowest attenuation of 0.8 dB/m for CO₂ laser light. The metal oxide films deposited by MOCVD have high chemical and heat resistivity. Therefore the hollow fiber coated with these materials is suitable for high power laser applications under a severe environment.

Coiled hollow waveguide gas absorption cells were designed and fabricated using polymer and glass tubing. These coiled waveguides were found to be acceptable for the detection of gases in the IR region of the spectrum. Through the in-depth investigation of straight and bent losses in hollow waveguides, an equation was developed to calculate the loss of the coiled system. By setting the straight loss to zero, it was found that additional loss on bending was approximately 1 dB per turn, independent of the bore size or bending radius.

The Free Electron Loss (FEL) at Vanderbilt University is tunable form 2 μm to 9 μm in the mid-IR spectrum, which is capable of controlling predicted laser-tissue interaction by selecting a specific wavelength. However, delivery of this laser into the internal portion of the eye is difficult because of strong water absorption at this spectrum range and the high peak power of the FEL. We used a metallic coated hollow waveguide with a 530 μm inner diameter and 1 meter in length, and delivered the FEL beam to an autoclaved surgical probe. The prove tip was an 18 gauge canula with a mini CaF₂ window fixed in front of it to protect the waveguide from contacting water. Human and animal cadaver eyes were used to perform an open sky retinal cutting procedure. The system was able to deliver 60 percent of FEL energy to the intraocular tissues. Up to 6 X 10⁵w peak power was reached without damage to the waveguide or the surgical probe at the spectrum range of 2.94 μm to 7.7 μm. In conclusion, the hollow waveguide is suitable for delivering the IR FEL for intraocular microsurgical procedures.

Tellurium Iodide based glass fiber preparation has been optimized in order to respond to several IR waveguide technological demands. Three directions have been examined: 1) realization of a two fiber systems for radiometry in the room temperature range, 2) chemical polishing of a core-clad fiber for evanescent wave remote IR spectroscopy, 3) attempts for single mode fiber operating in the 10 μm region for planetary spectroscopy.

Two years ago, UV-improved (UVI) fibers were described for the first time. These fibers show significantly reduced UV- absorption around 215 nm during the transmission of UV-light from different UV-sources like deuterium-lamps, excimer- lasers and frequency-doubled tunable dye lasers. This improvement has been realized due to a passivation of the generated UV-defects by gas-doping. Depending on the ambient temperature and the fiber diameter, the lifetime of UV- stability was quite different in these first generation UVI- fibers. This long-term behavior has been studied with different illuminations, temperatures and fiber diameters; there is a good agreement between experimental and theoretical data. Additional modifications have lead to a further improvement in the long-term stability. The predicted lifetime of these UVI²-fibers of the second generation at room temperature is longer than 1 year. In addition, a short-term behavior was observed using diode- array spectrometer. In this case, the influence of temperature on the transient induced losses is significant and will be described.

An innovative spectroscopic diagnostic method has been developed for investigation of different regions of normal human skin tissue. This new method is a combination of Fourier transform IR fiberoptic evanescent wave (FTIR-FEW) spectroscopy and fiber optic techniques for the middle IR (MIR) wavelength range. The fiber optical sensors we have used are characterized by low optical losses and high flexibility for remote analysis. Our fiber optical accessories and method allows for direct interaction of the skin tissue with the fiber probe and can be utilized with a diversity of standard commercial Fourier transform spectrometers. The FTIR-FEW technique, using nontoxic unclad fibers in the attenuated total reflection regime, is suitable for noninvasive, fast, sensitive investigations of normal skin in vivo for various medical diagnostics applications including studies of acupuncture points. Here we present the first data on IR spectra of skin tissue in vivo for normal skin and several acupuncture points in the range of 1300 to 1800 cm^-1 and 2600 to 4000 cm^-1.

A new optical improvement for image-guided surgical navigation employs coded apertures to measure with submillimeter accuracy the 3D locations of surgical instruments equipped with point light sources relative to a patient. The system consists of multiple angular sensors to measure the position of each light source in several angular dimensions. Each sensor includes a coded aperture to image the light from the point source onto a linear charge-coupled device (CCD) detector array. A coded aperture consists of spatially redundant slits of various widths. This replaces the cylindrical optics or single precision slit which has been used heretofore to focus light from the point source into a linear image crossing the CCD. Although cylindrical lenses have adequate light-gathering capability and tolerate dust, off-axis focus is a major problem. A slit has excellent depth of field and focus but forms a dim image, and dust affects accuracy. A coded aperture combines the advantages of slits and lenses, but does require computing the correlation function using the wide image and a corresponding kernel. The peak in the correlation determines the angular position of the point light source. Sub-pixel accuracy is achieved by interpolation near the correlation peak. Image magnification problems are solved by sing scaled kernels.

Minimally invasive spine procedures can spare the patient the trauma associated with open surgery. However, these procedures can be difficult to learn and require extensive training for proficiency. At Georgetown University Medical Center, spine biopsies are often done under computed tomography (CT) guidance. While this technique is effective, it is time consuming since the biopsy needle must be advanced slowly and its position checked several times to ensure vital organs are not damaged. A research project is being conducted to develop a computer-guided, image-based minimally invasive system for therapy and surgical techniques. As an initial step, a needle biopsy simulator for training is being developed. In the next phase, this simulator could also be used for preoperative planning. The simulator consists of two major modules: a visual module to display the medical images and biopsy tools and a haptic module to provide force feedback based on the needle position. The haptic module incorporates a robotic device that provides force feedback in three translational directions. In the future, it is anticipated that semi- autonomous robotic systems, in which the human controls some degrees of freedom and the robot the other degrees of freedom, will be developed for interventional tasks such as needle spine biopsy. The simulator described here can then be used as a 'master arm' to control the robotic system that actually performs the intervention.

Surgical navigation embodies the use of landmarks to guide a surgeon along a specified course. Often this course is chosen through the careful analysis of the patient's anatomy using accurate images obtained before surgery begins. The very nature of an invasive surgical procedure involves the alteration of the patients anatomy during the course of the procedure and as a result the preoperative images do not accurately reflect the state of the patient that the surgeon is attempting to navigate through once the procedure has begun. We have developed a system to augment the preoperative imaging data with that of imaging data acquired intraoperatively. This provides a more accurate reflection of the surgeon's current position and orientation with respect to the patient's anatomy and enables a more aggressive approach to the clinical procedure.

Experience with image guided, frameless stereotactic neurosurgery shows that intraoperative brain position shifts can be large enough to be problematic, and can occur in different directions at different directions at different stages of an operation. An understanding of the behavior of shifts will allow the surgeon to make the most appropriate use of the image guidance by first minimizing the shift itself, and then anticipating and compensating for any influence the remaining shift will have on the accuracy of the guidance. Three types of shift are described. Type I shift is a local outward bulging that occurs after the skull and dura are opened but before a mass lesion is resected. Type II shift is a local collapse of the brain tissue into the space previously occupied by the tumor. Type III shift is related to loss of cerebrospinal fluid or brain dehydration and is a generalized, more symmetric loss of brain volume. Strategies to minimize these types of shift include appropriate use of medical measures to reduce brain swelling early in the procedure without producing so much brain dehydration that Type II shift is accentuated later in the procedure. Other strategies include mechanical stabilization of brain position with retractors. Anticipating shift, the neurosurgeon should use the guidance as far as possible to map key boundaries early in the procedure before shift becomes more pronounced. Ultimately, however, the correction for the problem of intraoperative brain shift will require the ability to update the imaging data during the surgery.

Image-guided surgery has evolved over the past 15 years from stereotactic planning, where the surgeon planned approaches to intracranial targets on the basis of 2D images presented on a simple workstation, to the use of sophisticated multi- modality 3D image integration in the operating room, with guidance being provided by mechanically, optically or electro-magnetically tracked probes or microscopes. In addition, sophisticated procedures such as thalamotomies and pallidotomies to relieve the symptoms of Parkinson's disease, are performed with the aid of volumetric atlases integrated with the 3D image data. Operations that are performed stereotactically, that is to say via a small burr- hole in the skull, are able to assume that the information contained in the pre-operative imaging study, accurately represents the brain morphology during the surgical procedure. On the other hand, preforming a procedure via an open craniotomy presents a problem. Not only does tissue shift when the operation begins, even the act of opening the skull can cause significant shift of the brain tissue due to the relief of intra-cranial pressure, or the effect of drugs. Means of tracking and correcting such shifts from an important part of the work in the field of image-guided surgery today. One approach has ben through the development of intra-operative MRI imaging systems. We describe an alternative approach which integrates intra-operative ultrasound with pre-operative MRI to track such changes in tissue morphology.

Stereotactic surgery for movement disorders is typically performed using both imaging and physiologic guidance. However, different neurosurgical centers vary in the emphasis placed on either the imaging or the physiological mapping used to locate the target in the brain. The relative ease with which imaging data is acquired currently and the relative complexity and invasiveness associated with physiologic mapping prompted an evaluation of a method that seeks to maximize the imaging component of the guidance in order to minimize the need for the physiologic mapping. The evaluation was carried out in 37 consecutive stereotactic procedures for movement disorders in 28 patients. Imaging was performed with the patients in a stereotactic head frame. Imaging data from MRI in three planes, CT and positive contrast ventriculography was all referenced to this headframe and combined in a stereotactic planning computer. Physiologic definition of the target was performed by macroelectrode stimulation. Any discrepancy between the coordinates of the imaging predicted target and physiologically defined target was measured. The imaging- predicted target coordinates allowed the physiologically defined target to be reached on the first electrode penetration in 70% of procedures and within two penetrations in 92%. The mean error between imaging predicted and physiologically defined target position was 1.24 mm. Lesion location was confirmed by postoperative MRI. There were no permanent complications in this series. Functional outcomes were comparable to those achieved by centers mapping with multiple microelectrode penetrations. The findings suggest that while physiologic guidance remains necessary, the extent to which it is needed can be reduced by acquiring as much imaging information as possible in the initial stages of the procedure. These data can be combined and prioritized in a stereotactic planning computer such that the surgeon can take full advantage of the most reliable information from each imaging modality.

The possibility of the high-speed curvilinear trajectory reconstruction algorithm to be applied to the real-time optical imaging is demonstrated. The method for the calculation of the photon mean path (PMP) in strongly scattering media was used for the mathematical simulation of the absorptive macroinhomogeneity distribution reconstruction for the cases of the flat slab cylindrical body. The relative shadow representation by the integrals along the PMP corresponds to the projection procedures used in conventional tomography algorithms. The results of the reconstruction are in a good agreement with the given distribution. As opposed to the FEM method, the employed PMP method for tomography reconstruction is shown to enable the computing time to be many times decreased and become close to that for conventional projection algorithms.

Minimally-invasive surgical techniques reduce the size of the access corridor and affected zones resulting in limited real-time perceptual information available to the practitioners. A real-time feedback system is required to offset deficiencies in perceptual information. This feedback system acquires data from multiple sensors and fuses these data to extract pertinent information within defined time windows. To perform this task, a set of computing components interact with each other resulting in a discrete event dynamic system. In this work, a new discrete event requirements model for sensor fusion has been proposed to ensure logical and temporal correctness of the operation of the real-time diagnostic feedback system. This proposed scheme models system requirements as a Petri net based discrete event dynamic machine. The graphical representation and quantitative analysis of this model has been developed. Having a natural graphical property, this Petri net based model enables the requirements engineer to communicate intuitively with the client to avoid faults in the early phase of the development process. The quantitative analysis helps justify the logical and temporal correctness of the operation of the system. It has been shown that this model can be analyzed to check the presence of deadlock, reachability, and repetitiveness of the operation of the sensor fusion system. This proposed novel technique to model the requirements of sensor fusion as a discrete event dynamic system has the potential to realize highly reliable real-time diagnostic feedback system for many applications, such as minimally invasive instrumentation.

In current practice, microsurgery involves dexterous manipulations on small tissues viewed through a stereo microscope. Surgeons hold special grasping and cutting instruments in a pencil-like grip, with their palms supported, to optimize fine motor control and minimize hand tremor and fatigue. Telepresence-based microsurgery has the potential to provide the surgeon with a magnified workspace in which he can comfortably work with his hands on full-size instrument handles, using normal hand motions and experiencing the feel he would expect from the magnified tissues that he sees. To address the needs for performing microsurgical procedures, the SRI telepresence surgery workstation has been combined with a pair of micromanipulator arms. The prototype microsurgery system has been tested with ex-vivo tasks similar to those required for surgical procedures, such as cutting, grasping, suturing, and knot tying. Initial animal testing has been done on a rat model in which end-to-end anastomosis of the femoral artery was completed with 10 rats, and 100% patency was obtained.

Robotic telesurgery is a promising application of robotics to medicine, aiming to enhance the dexterity and sensation of minimally invasive surgery through millimeter-scale manipulators under control of the surgeon. With appropriate communication links, it would also be possible to perform remote surgery for care in rural areas where specialty care is unavailable, or to provide emergency care en route to a hospital. The UC Berkeley/Endorobotics/UCSF Telesurgical Workstation is a master-slave telerobotic system, with two 6 degree of freedom (DOF) robotic manipulators, designed for laparoscopic surgery. The slave robotic has a 2 DOF wrist inside the body to allow high dexterity manipulation in addition to the 4 DOF of motion possible through the entry port, which are actuated by an external gross motion platform. The kinematics and the controller of the system are designed to accommodate the force and movement requirements of complex tasks, including suturing and knot tying. The system has force feedback in 4 axes to improve the sensation of telesurgery. In this paper, the telesurgical system will be introduced with discussion of kinematic and control issues and presentation of in vitro test results.

The advances in radiographic imaging techniques that have occurred within the past decades have significantly enhanced our abilities to display anatomy as well as pathology. Although image acquisition commonly generates three-dimensional datasets, limitation in user interfaces generally requires that this information be presented clinically as a series of two dimensional images. Consequently, during preoperative planning, surgeons are required to mentally transform a wealth of two dimensional hard copy images qualitatively into three-dimensional concepts that are used as a road map to surgery. The sugeron's success is dependent on the accurate mental transfer of Computer Tomographic X-ray (CT) and Magnetic Resonance (MR) imaging information to the operative site to assist direct visual perception of the procedure. Thus, the surgical procedure is performed with the surgeon intuitively transferring radiographic information to the surgical site. Neurosurgeons, for example, are especially hampered because of limited ability to visually distinguish tumor tissue, peritumoral edema, and normal brain parenchyma. This limitation at least partially accounts for the relatively high incidence of subtotal tumor excision. Based on the need to assit the surgeon transform the preoperative scans to the operative site, stereotactic systems were developed and would evolve into frameless stereotaxy with the advance of various sensors.

As an alternative to the 2,94 μm Er:YAG laser, new laser crystals have been developed to convert near IR laser light into a 3 μm laser beam. With this concept the 3 μm laser light is created directly at the position where the applications occur, either inside the body or on the tissue surface. The problem of transmitting the 3 μm laser light through fibers consequently no longer exists. The emission line of the laser-converter (L-C) crystal corresponds to a transition of the ⁵I₆ to the ⁵I₇ energy niveau of holmium ions, e.g. in a BaYb₂F₈ crystal. Pumping with the Nd:YAG laser line of 1.12 μm leads to direct excitation of holmium ions from the ⁵I₈ niveau to the upper laser level. This direct pumping is more efficient than pumping of the Yb ions at 1.047 μm. With this L-C very small systems can be realized for endoscopic applications or for handheld lasers. The advantages of this converter crystals are high conversion efficiency of 25% and low dependence of conversion efficiency on Ho and Yb concentrations, pump density and pump geometry. Pulsed system for laser ablation of soft and hard tissue and also cw-applications of the 3 μm radiation are possible.

A Fourier transform IR spectrometer and IR transmitting AgClBr fibers were used for fiberoptic evanescent wave spectroscopy (FTIR-FEWS) of cancer. Malignant and healthy tissue samples were extracted from patients at the Meir Hospital in Israel, placed on a Silver Halide fiber, and measured using the FTIR-FEWS system. The IR spectra were analyzed and compared by taking the ratio of absorption of the active functional groups of Amide I at 1642 cm^-1 and Amide II at 1545 cm^-1. Clear differences appeared between the two types of tissue. When placing the tissue samples on bare fiber the reproducibility of the result was not satisfactory due to chemical interaction between the tissue and the fibers. This problem was solved by applying Polyethylene coating of thickness 1-2 μm on the fiber, leading to reproducible results. The results of these preliminary studies indicate that eh FTIR-FEWS technique can be used for cancer diagnostics. Combined with endoscopy this technique could be used to analyze tissues inside the body in vivo and in real time.

There is an interest in using IR spectroscopy techniques for blood analysis, and in particular for glucose measurements. Glucose concentration in aqueous solutions were measured using fiberoptic evanescent wave spectroscopy. A new spectrometer consisting of tunable CO₂ laser as the radiation source, and a silver halide fiber was built. Several solutions were measured for calibration with concentrations in the range of 70-2000 mg/dl. After the calibration, solution were measured and predicted according to the calibration parameters. The predicted concentrations were in agreement with the reference ones. The predicted errors for concentrations range up to 200 mg/dl, which is the clinical range, was less then 18 mg/dl (a relative error of about 12%). The AgClBr fibers are non-toxic and therefore this measurement can be carried out in real-time, for in-situ measurements of glucose in blood.

The luminescence of silver bromide crystals doped with rare earth ions (Nd³⁺) was investigated in the visible and near IR spectral ranges. The emission, excitation, and absorption spectra, as well as the kinetic parameters, were measured over a broad temperature range. Crystal doping was produced by growing in the melt and by a diffusion method. A novel method for measuring the diffusion profile of rare earth ions in AgBr crystals, based on the luminescence distribution of the dopant, was proposed. The luminescence parameters of AgBr:Nd crystals doped by the diffusion method were compared with these for crystals doped by adding Nd³⁺ to the melt. The spectroscopy parameters of Pr³⁺ and Er³⁺ ions in AgBr crystals doped in the melt were also investigated. The Judd-Ofelt analysis was applied to the rare-earth doped crystals, and transition rates, branching ratios, and quantum efficiencies were calculated. Good agreement between theory and experiment was obtained.

The appearance of intrinsic defects, especially in the UV spectral range, has been investigated in step-index fibers having undoped silica core and a fluorine doped cladding; the concentration of the dominant impurities ranged from 2 to 50 ppm and from 400 to 600 ppm. The optical attenuation from 190 to 500 nm was not only determined in fibers, but also in the preforms and core rods. The absorption bands at 248 nm, at 340 nm and at 210 nm were observed. While the preform process has no significant influence, the fiber drawing process influences greatly the defects' concentration: the 210 nm band increases slightly, the 248 nm band increases significantly, the 340 nm band decreases. No influences of the fluorine cladding was determined with one exception when the chlorine content has been increased in the cladding of special species. The possible reactions and related defects will be described for the tested low-OH material. Especially the influence of chlorine will be discussed.

A most radical resection is of great importance in the treatment of brain tumors, however they can hardly be differentiated from normal brain parenchyma by the naked eye of the neurosurgeon. Photosensitizers are highly selective taken up into malignant tissues, therefore the fluorescence properties of photosensitizers could be utilized for optical differentiation of normal and malignant tissue. Ten patients presenting with brain malignancies were sensitized for photodynamic diagnosis (PDD) and photodynamic treatment (PDT) with 0.15 mg/kg b.w. m-THPC. On day 4 intraoperative PDD and fluorescence guided tumor resection (FGR) was performed, followed by intraoperative PDT. The fluorescence was induced by a xenon lamp at an excitation wavelength ranging from 370 to 440 nm. A sensitive CCD camera was employed for imaging, equipped with a long pass filter to shut off the excitation wavelength and to improve the signal to noise ratio. The pictures are converted digitally by a standard frame grabber and processed in real time and calculated for the tissue auto fluorescence in the emission band of m-THPC at 652 nm. Out of 10 0bservations, two were false negative and 2 were false positive. Our preliminary results indicate that fluorescence guided surgery is feasible and proved to be of significant help in delineating tumor margins and in resection of residual tumor that could not be detected by the surgeon, however the sensitivity and specificity needs to be further improved.

Sapphire fibers grown using the Saphikon EFG technique have proven to be effective delivery systems for Erbium:YAG and Erbium:YSGG laser energy. A special emphasis has been given to the hard tissue dental application following the FDA approval of this procedure in May of 1997. A range of sensor applications also utilize these EFG sapphire fibers. Through the latter half of 1997, Saphikon has successfully transitioned the process from a research/pilot scale level to high volume production. Fiber losses as low as 0.2 dB/meter with an average of 1.5 dB/meter, which were first demonstrated on a research level in which tens of meters were produced, have now been sustained over thousands of meters in full scale production. Use of sapphire fibers in non-medical applications, erbium laser dentistry, and other medical procedures is discussed.