Artificial Intelligence, or AI, is a term at the forefront of many industries, radiology foremost among them. Of course, advanced technologies and radiology have always been interrelated. The benefits provided by computer-assisted imaging, storage and sharing, for example, are innumerable and ever-advancing. The deep learning made possible through AI algorithms holds tremendous promise now and going forward in the areas of information gathering, integration and application to patient care. While some postulate that AI will eventually replace medical professionals, including radiologists, most in the industry agree this will not the be case. “I don’t envision a future where doctors aren’t required for optimal patient care,” says Diagnostic Radiology subspecialist and Vascular & Interventional radiologist Dr. Caleb Rivera. “What AI is particularly good at is establishing patterns in massive amounts of data in relatively little time. AI utilizes speed and tirelessness to help medical professionals apply our skills, experience and compassion. I don’t see these as mutually exclusive, but instead, logically integrated strengths.”
AI has already begun to improve tests such as Magnetic Resonance Imaging, or MRI, which is excellent at imaging all areas of the body, particularly soft tissues such as organs, muscles and fat. Despite its preferability for many exams, one of the factors that make MRI prohibitive is that it can take up to 90 minutes to perform and requires the patient to remain still throughout. Remaining immobile in an enclosed environment can make having an MRI feel claustrophobic to some patients, and, in cases such as stroke, it is too slow to be optimal.
"AI is already speeding up time-consuming imaging tests like MRI, but it’s only the beginning," says Dr. Rivera. “Going forward, we see AI being able to compile the information from thousands or millions of exams so the systems will become ever-faster at capturing, quantifying and analyzing data. Given what AI can do currently, it’s easy to extrapolate that tests and diagnoses will soon take only a fraction of the time they do now.”
AI is already able to delineate fine details in imaging exams like computed tomography (CT), identifying the different chambers of the heart, brain hemorrhages, and stroke damage. By working in partnership with radiologists, who have many years of advanced education and training, AI strengthens accuracy by spotting patterns in dense information without being subject to tedium, exhaustion or time constraints.
“AI can detect certain problems in seconds instead of minutes, and in acute situations, that time savings can be monumental to the patient,” says Dr. Rivera. “So, I see AI as providing the speed, while we educated mortals provide the real-world experience, training and understanding. To me, it’s an outstanding collaboration.”
