Computer‐aided detection and diagnosis/radiomics/machine learning/deep learning in medical imaging

Abstract Over the years, medical physicists have played important roles in the establishment and advancement of computer‐aided detection/diagnosis in medical image interpretation. Note that the benefit of a medical imaging exam in presenting an accurate diagnosis depends on both the quality of the image (i.e., the physical image acquisition system) and the quality of the interpretation (e.g., conducted by a radiologist). The interpretation of medical images by radiologists is limited by the human eye‐brain system, including various search and perception problems, potentially resulting in missed detections, inaccurate diagnoses, and clinical decision‐making errors. Having computer vision Artificial Intelligence (AI) systems yield “interpretations” was seen as a method to effectively and efficiently improve the medical image interpretation process. In this essay, I will try to convey, through my eyes, the exciting times during the past 35 years that led up to AI in medical imaging as we experience today. In the mid‐1980s, I was part of a University of Chicago team of imaging scientists/medical physicists and radiologists (within the Kurt Rossman Laboratories) who conducted early research in the field of CAD [computer‐aided detection (CADe) and computer‐aided diagnosis (CADx)], where, as the name infers, the goal of the computer output is to be used as an aid to radiologists (i.e., as a second opinion as opposed to a replacement of the radiologist). Kunio Doi, Heang‐Ping Chan, and I submitted the seminal patent on CAD that was granted on 6 March 1990, as well as co‐authored various papers in MEDICAL PHYSICS along with others on our interdisciplinary team, including Heber MacMahon, Robert A. Schmidt, Carl J. Vyborny, Charles E. Metz, Shigehiko Katsuragawa, Robert M. Nishikawa, and Wei Zhang.