Machine Intelligence in Neurologic and Head and Neck Imaging

Published:September 17, 2020DOI:https://doi.org/10.1016/j.nic.2020.08.006
      Figure thumbnail fx1
      Reza Forghani, MD, PhD, FRCP(C), DABR, Editor
      Interest in intelligent machines is not new, and serious attempts at building machines that can solve problems and make decisions simulating what humans do were already underway in the 1950s. However, the field of artificial intelligence (AI) has clearly seen a revolutionary revival in the last decade. This is at least in part fueled by developments in application of deep neural networks for complex analytic and image analysis tasks, made possible by the impressive progress and achievements in hardware development and computing power. AI is already incorporated in our daily lives. As a testament, one must look no further than at smart phones, commonly used search engines, social media, or smart home devices. There is also increasing adoption in various major industries. Not surprisingly, there is a lot of interest in health care applications, especially in imaging, where the core data (ie, medical images) have been in digital format for many years and therefore optimally stored and available for computerized analysis and AI applications.
      There is no doubt that this disruptive technology has the potential to reshape health care in the long term. However, at least in the foreseeable future, AI’s application potential is not unlimited. Extreme scenarios of AI representing a be-all and end-all solution and replacement for everything not only are unrealistic but also do disservice to the technology by setting unrealistic expectations that could inevitably lead to disappointment, in addition to hindering engagement by the very professionals with the domain knowledge needed to successfully implement AI in the clinical setting. At the same time, most experts would agree that ignoring this technology would be at one’s own peril, and it is imperative for physicians to leverage this technology to improve health care processes. The purpose of this issue is to familiarize radiologists and other health care professionals, who are interested in neurologic and head and neck imaging, with the strengths and limitations of AI technology and its potential positive impact in improving the health care enterprise by augmenting human intelligence.
      This issue consists of a collection of review articles written by experts across North America that includes general technical reviews, reviews of various clinical applications in neurologic and head and neck imaging, and reviews related to the use of AI for process improvement and incorporation in smart devices. In addition to discussing current proposed image-based applications, this collection is meant to provide a glimpse into the future, as this disruptive technology is likely to blur traditional specialty boundaries and integrate different information in order to enable precise personalized therapy as part of a wholistic patient-centric care pathway. It is my hope that the content can serve as a resource and catalyst both for those unfamiliar and for those already entrenched in AI research or clinical implementation.
      I conclude by expressing my gratitude to Dr Suresh K. Mukherji for the opportunity to guest edit this issue of Neuroimaging Clinics and by thanking all the authors for their fantastic work and contributions to this issue. This issue would not have been possible without the support and patience of John Vassallo, Associate Publisher, and Casey Potter and Nicholas Henderson, Developmental Editors at Elsevier. I hope that our readers will find the issue both informative and inspiring.