Computer-Integrated Surgery

COMPUTER ASSISTED SURGERY Computer assisted surgery is such a recent development on the orthopaedic scene that it is too early to identify a classic article. Rather, what is needed is a brief exposition of what the relationship between the computer and the surgeon promises for the future. The application of computer science to aid in the performance of surgical procedures holds real promise, especially for orthopaedic surgery. The following selection is from the introduction of a recent book devoted to the subject.1 The authors represent an international collaboration involving various disciplines. The authors are as follows: Russell H. Taylor, professor of computer science, Johns Hopkins University, Baltimore, MD, and former manager of Computer Assisted Surgery Research, IBM J. Watson Research Center; Stephane Lavallée, PhD, investigator for the Computer Assisted Surgery Group, Techniques de l'Imagerie, de la Modelisation et de la Cognition (TIMC) Laboratory, Genoble, France; Grigore C. Burdea. MD, assistant professor of computer engineering, Rutgers, The State University of New Jersey, and Ralph Mosges, head of Clinical Research, Department of Otolaryngology, Aachen Technical University, Aschen, Germany. Leonard F. Peltier, MD, PhD INTRODUCTION An emerging partnership For clinicians, this human-machine partnership is important because it offers the possibility both of significantly improving the efficacy, safety, and cost-effectiveness of existing clinical procedures and of developing new procedures that cannot be performed at all otherwise. For technologists, this partnership offers real, challenging applications with articulate end users. Furthermore, incremental progress is possible. Even relatively simple uses of new technology can make significant differences clinically, and the lessons learned can then be applied to harder problems. In exploring a partnership involving complementary capabilities, it is useful to consider the strengths and weaknesses of each party. Human surgeons have, of course, many capabilities. They are very dexterous, quite strong, and fast, and are highly trained to exploit a variety of tactile, visual, and other cues. They are adaptable and can exercise these skills over a surprisingly wide range of geometric scales. "Judgmentally" controlled, they understand what is going on in the surgical procedure and use their dexterity, senses, and experience to execute the procedure. They can analyze their own performance and apply the lessons learned-that is, they can improve with practice. However, surgeons do have limitations. They are not geometrically accurate. In other words, they cannot easily place an instrument at an exact, numerically defined location relative to the patient and then move it through a defined trajectory, nor are they very good at exerting exactly a predefined force in a particular direction. They do not tolerate ionizing radiation well and are understandably not eager to be exposed to it on a daily basis. They get clumsy if forced to work in very confined spaces or over long periods of time. They may have small hand tremors that limit their ability to operate on very delicate structures. They get tired and make mistakes. They get old and lose some of their skill. Unfortunately, many of these limitations affect the efficacy of certain surgical procedures, especially in cases where great geometric accuracy is required or in which the surgeon's direct use of his or her senses or manual dexterity are impaired. Fortunately, machines have complementary capabilities that can remedy some of these defects. Machines are very precise and untiring. They can be equipped with any number of sensory feedback devices and can measure and position instruments very accurately in six degrees of freedom. Numerically controlled machines can move a surgical instrument through an exactly defined trajectory with precisely controlled forces. Potentially, they can be miniaturized to function in very confined spaces and can be hardened to withstand significant doses of ionizing radiation. When properly maintained, they are reliable, and their every move can be recorded for subsequent analysis. Viewed in this light, even very sophisticated computer-controlled machines such as robots are simply better surgical tools. As such, they do not pose any significant threat to replace surgeons, but they can help surgeons to work better. Nonetheless, the computer control required to achieve many of these advantages of machines introduces important questions. Because surgeons are responsible for everything that goes on in the operating theater, they are naturally wary of any entity-be it a scrub nurse, intern, resident, or machine- that may not respond to their wishes in exactly the desired manner and at exactly the right time. Surgeons have learned (more or less) to trust human assistants, who can be trained to respond to verbal and nonverbal commands and to exercise some judgment in carrying out assigned tasks. Machines, on the other hand, are very literal-minded, possessing a very limited ability to interpret human language. Furthermore, much of the information needed to exploit the geometric precision of machines is not easily described linguistically. How can surgeons tell a machine what it is supposed to do? How can they be sure that the machine "understands"? Theymust rely on the machine to provide precision or access to otherwise inaccessible parts of the patient's anatomy. How can they keep track of what it is doing when their own senses are inadequate? How can they trust it not to harm the patient? How can the machine help them to perform a tricky, precise maneuver or possibly warn them when human error may hurt the patient? None of these questions are easily answered. Indeed, all are the subjects of ongoing research. We hope that the work reported in this book will give the reader a better understanding both of the underlying issues and of potential areas for improvement. The synergy between planning and execution A related theme is the significant synergy that can be achieved between computer methods for presurgical planning and an enhanced ability actually to execute the plans developed. This book has many examples of systems that extract information from medical images to model anatomic structures in live patients and to help clinicians use this information in diagnosis and treatment planning. For these systems to have value, it is essential that surgeons actually be able to carry out the surgical plans developed. If a plan is primarily a qualitative preoperative simulation, it may be possible for a surgeon simply to rehearse ahead of time and then carry the results into the operating room in his or her head. However, the usefulness of such a simulation may be greatly enhanced if the system is capable of following the progress of the actual surgery and of providing the surgeon with an interactive real-time display as a reference. If the plan involves quantitative information, such as the shape and position of a tumor or desired positions and orientations of osteotomy fragments, then the ability to achieve the necessary geometric accuracies intraoperatively becomes even more important. Again, many of the chapters in this book [not reprinted here] describe systems designed to provide appropriate real-time information about surgical plans or to execute precise geometric plans. The availability of such systems, in turn, makes the computational effort required to plan the surgery much more worthwhile. Furthermore, the consistency of computer-assisted surgical execution means that comparative studies involving several subjects become more meaningful. A structuring architecture The figure below illustrates, in a general way, our view of the key architectural components required to support such human-machine partnerships in surgery. In a sense, it represents our view of what a comprehensive future system of computer-integrated surgery might include.FigureFigure