Man-machine interfaces in computer assisted surgery.

Abstract

The clinical potential of computer assisted surgery (CAS) has been more and more widely acknowledged since CAS systems have been introduced into the operating room (OR) theater. Especially the improvements in safety and accuracy are remarkable and strengthen the ties between surgeons and engineers. Tumor stereotaxis was introduced to neurological surgery in the early 1980s, and currently systems with and without robotic navigation are in use for specific medical indications. Recently, solutions for computer assisted orthopedic surgery were developed and applied to various anatomical regions. However, with the establishment of CAS in vivo, a new complex of problems, which was not present in the laboratory setup, was introduced: the man-machine interface. Currently, the complexity of available CAS systems requires the presence of at least one system engineer (often called the "operator") in the OR. As a consequence, there is no possibility for direct communication between the surgeon and the machine or software. Most of the program steps involved in CAS and choices to be made intraoperatively have to be transferred to the software by means of communication of the surgeon with the operator. Particularly, the establishment of a relation between the virtual object (i.e., a medical image) and the surgical object (i.e., the patient), often denoted as "matching" or "skeletal registration," requires intensive interaction of the surgeon with the computer. A literature survey revealed that no CAS system in clinical use exists without a system engineer or a comparable person, and our clinical experience indicated that the matching process is a weak point in most systems. Because it appears to be contradictory to cost-reduction efforts in health care to have a highly paid specialist in the OR, this research evaluates strategies to facilitate the man-machine interface with the final goal of establishing a direct control of the system by the surgeon or the medical personnel traditionally present at surgery. Options to be investigated include 1) a CAS control panel (virtual keyboard) as an integrated component of the existing navigation system and 2) introduction of a commercial voice-recognition system. The implementation of these strategies into the existing CAS setup at the Department of Orthopaedic Surgery at the Inselspital (University of Bern) and clinical experience gained are reported.