The Future of Imaging: Assessing the early impacts of COVID-19 and the path to innovation through Artificial Intelligence (AI) A conversation with Dr. Geoff Rubin
Dr. Geoff Rubin is a globally recognized Radiologist and thought leader. He has spent the majority of his career building teams at some of the leading education medical centers across the country, including Stanford and Duke Universities. He has been on the forefront of new technologies throughout his career, always believing there is an engineered solution allowing a view into the body without invading the body. As such, he was a pioneer in the use of 3D images to advance diagnosis. Healthcare Performance Insider sat down with Dr. Rubin to gain his insights on the impact of COVID-19 on the Imaging department and the future of innovation in Radiology (specifically AI) given the pandemic.
COVID-19 Impact on Radiology
This has been an absolutely profound period of time in terms of the impact that it has had on our delivery of healthcare. One of the many uncertainties is the impact that we have not yet realized. Dr. Rubin hones in on this topic, highlighting several areas he foresees our focus quickly shifting to.
Millions of imaging studies that would have been performed have not been over the last several months. Some of those imaging studies would have been performed on patients undergoing treatment that have serious diseases like cancer. Imaging is used to assess their status and that information has not been available. As we start to open up and those patients come back, hopefully we do not find an overwhelming amount of advanced disease that we didn’t anticipate. Hopefully we don’t find that there are patients who have irreversibly suffered because they were unable to have the benefits of medical imaging.
Fear of coming to the ED, even in an emergency. There has been an increase in deaths from myocardial infarction and stroke during the course of the COVID-19 pandemic most likely because patients that would have come to the Emergency Department did not come despite concerning symptoms of stroke or heart attack. Patients have concern about being exposed to COVID-19 when seeking health care – an issue that is going to continue to be with us for some time.
Recovery of substantial financial losses across the spectrum of health care providers. Be it hospitals, physician groups, and other providers in the healthcare value chain, all will continue to feel this strain. It is going to lead to a lot of reflection around how funds are spent moving forward in the pursuit of delivering health care and in advancing our healthcare institutions. There will be a belt tightening and renewed understanding of priorities to provide quality healthcare in this post-COVID environment.
The new normal is going to be dynamic. Like many, the Imaging environment has a large set of considerations and we are learning as we go. Policies and procedures are critical to assure patient and staff safety as we re-open. Identifying the right configuration of space and technology, personnel availability to ensure social distancing, while bringing the full measure of clinical expertise to the interpretation of imaging studies and performance of image guided procedures will be a new normal.
We need to be open minded as we reconfigure our departments and workflows. Profitability will come back through the performance of radiology and medical imaging. The pathway for getting there however is to create safe and effective environments so patients can have the imaging procedures that are medically necessary.
Innovation in a COVID World
In a pre-COVID world, two letters consumed 90% of any discussion around the future of medical imaging: AI. Physicians and non-clinical consumers alike followed the latest available technologies, trends, and early advancements. Research and investment were synonymous with success and the best wanted to be at the top of both heaps. Enter COVID-19. The work is still there but the focus has shifted – for now. We asked Dr. Rubin to comment on this transition.
The transformation of imaging through AI
The transformation of radiology has already begun. From our initial encounter with the patient, to our provision of information to our referring physicians, to the way that information is used, AI offers the opportunity to better understand the value of imaging before imaging is performed. By integrating information from past imaging tests with other health data elements AI may predict the value that performing an imaging test will bring to a specific patient. Additionally, AI will soon be able to integrate image findings with other clinical characteristics to personalize treatment suggestions and predict how a patient might respond to a specific drug or operation in the course of treatment.
Less really is more
We already see AI’s remarkable impact in the reconstruction of the images that are being acquired. CT, PET, and MR images have been profoundly influenced by new techniques that allow the images to be reconstructed with a much sparser input of data. These advancements may allow a hospital or imaging center to perform more studies in a shorter period of time, with less patient time on a scanner, potentially less exposure to radiation, and the ability to scan more patients per day.
AI makes it our way
New advancements allow for the acquisition of images in a manner that is customized to the specific indications for the imaging, personalizing how the machine is set up – e.g., the parameters for image acquisition, the pulse sequences that are used (in MR) – all depending on the patient’s physical characteristics and health status. AI can separate or “segment” organs and tumors, determining change in characteristics (over time) with greater consistency and higher quality. AI also offers the promise of helping us interpret the acquired images, making preliminary observations to detect urgent conditions and prioritize which exams should be examined first and ensuring that we do not miss subtle findings.
Given all this, we had to ask: What about the Radiologist?
Does this all mean that interpreters of medical images will soon be out of a job? Absolutely not. I don’t believe so for a second. I have spent a fair bit of time examining various manifestations of AI, researching it with our teams at Duke. I have an appreciation of the strengths, but also, the weaknesses of leaving decisions to computer models. There is no AI on the horizon that fully replaces human reasoning, our integration and understanding of the nuances in manifestations of disease and associations to uncommon conditions that we encounter. I do think however that AI is going to empower Radiologists to interpret more data within the same amount of time, or quicker, and also provide valuable inferences and observations that are unavailable today.
Dr. Rubin left us with one more profound thought: Radiologists serve patients through the information that they convey to the physicians who then directly interact with, and treat, the patient. The Radiologist’s responsibility is to empower those physicians, with meaningful imagery, to maximize the value they bring in caring for that patient. It is a significant role in the diagnostic process, using some of the most advanced technology in healthcare today, operating behind the proverbial scenes. We thank Dr. Rubin for bringing his valuable insights into the spotlight here.