Thomas A. Tami

Technological advances in otolaryngology–head and neck surgery continue to change the way we approach diagnosis and treatment. Although this serves as a primer for current diagnostic methods in otolaryngology, diagnostic approaches will undoubtedly continue to evolve. The evolution in genetically based diagnosis, innovations in audiology, and imaging breakthroughs for head and neck surgery have been touched upon previously in this section. Computer-aided differential diagnosis utilizing artificial intelligence, and electronic olfaction are but two of the many other areas of contemporary technology that may soon be impacting our specialty. This chapter offers a very brief overview of these cutting edge technologies.

Artificial Intelligence in Medicine

Starting with the initial description of artificial intelligence (AI) by Isaac Asimov in his classic 1950 novel I, Robot, the attempts to achieve this elusive goal have been met with only limited success and acceptance. From a diagnostic standpoint AI in medicine (AIM) could be a tremendous tool for both medical training as well as clinical practice. Already, from financial systems, billing and coding programs, diagnostic laboratory reporting, to electronic medical records, computers have been playing an increasing role in the medical arena. The combination of decision models, flow charts, clinical histories, clinical laboratory, and procedural data with the mathematics of decision theory can result in systems to augment the decision-making activities of practitioners. This knowledge can be expressed in the form of simple rules, or as a decision tree. A classic example of this type of system is KARDIO, which was a system developed to interpret electrocardiograms (ECGs).

AIM can be extended to explore poorly understood areas of medicine. Investigators are now discussing “data mining” processes and “knowledge discovery” systems. It is now possible, using patient data, to automatically construct pathophysiological models that describe the functional relationships among clinical datasets. For example, in 1993, Hau and Coiera described a learning system that used patient data obtained during cardiac bypass surgery to create models of normal and abnormal cardiac physiology. These models could be used to detect real-time changes in a patient’s clinical status. Alternatively, in a research setting, these models could help establish initial hypotheses to drive further experimentation.