Guest Column | November 30, 2022

How Can AI-Powered Advanced Imaging Battle Long COVID?

By Rajarshi Banerjee, Ph.D., Perspectum


Throughout my time building Perspectum, the medical imaging technology company I cofounded in 2012, I’ve had the opportunity to meet individuals across the U.S. and U.K. and hear about their personal struggles with chronic disease. Something that emerged with force over the past almost-three years is long COVID. Like other chronic diseases, long COVID is isolating, confusing, and variable, but also like other chronic diseases, those suffering with it deserve more information about their condition. Improving the state of long COVID care requires leaning into innovative technologies that can close the gap on data collection, diagnosis, and treatment.

In the U.S., according to the National Center for Health Statistics (NCHS), between Sept. 14 and Sept. 26, 2022, one in four adults with long COVID reported significant limitations on day-to-day activities, with 40% of respondents reporting as Black, Latino, or disabled. It’s not a rare disease — over 100 million worldwide likely have it — so why do we still lack the proper infrastructure to tackle and understand long COVID? This is because there are two key areas that delay progress on this — data collection and symptom interpretation — and I believe both can be assisted by AI-enhanced technologies.

Data Collection

Especially in the U.S., data collection around long COVID has been segmented and decentralized, resulting in less information for researchers to work with. Although the U.K. has fared better due to our centralized healthcare system, across both countries (and the globe) there is a need to invest in more clinical studies to compile data. Additional U.S.-funded studies have been announced, including a $500 million study in the U.S., but the learnings from this data will be too far in the future to help the 23 million already suffering.

AI-powered imaging technology can provide a report of dozens of metrics to physicians from one scan of a person’s organs. It can be a vital tool in the effort to understand multiorgan pathology. And, with a public health lens, leveraging the benefits of AI-powered advanced imaging technology means accessing richer data that can help the broader healthcare community better understand long COVID. In all areas of life, artificial intelligence is used to identify, track, and analyze data trends quickly and efficiently – far better than humans by design. Identifying trends and patterns is how the medical community (specifically clinicians) goes about treating long COVID patients right now, and this work and processing are also what support the research designs for long COVID. A new U.K.-based cross-disciplinary study called Stimulate-ICP, which is currently recruiting 4,000+ people with long COVID, is utilizing Perspectum’s CoverScan assessment tool to inform new integrated care pathways for people with long COVID. This way, we will capture more data that researchers can use to help us understand the disease and potential treatment pathways.

Evaluating Symptoms And Diagnosing Long COVID

Another complicating factor with long COVID is that diagnosis is challenging. There are more than 200 symptoms associated with long COVID, making identification difficult as it can present differently in patients based on their demographic group and comorbidities. Symptoms can be entirely debilitating, as they are for the millions forced out of work due to chronic pain or fatigue, or they can be innocuous at times, while worse at other times. Sometimes, symptoms are not mentally connected or registered by a patient.

Due to the lack of disease definitions and classifications, most U.S. long COVID patients are currently not able to receive the diagnostic, monitoring, treatment, and/or rehabilitation services needed to provide care. As healthcare providers struggle to give long COVID patients the much-needed reassurance that they will receive appropriate care, treatment, and hospital admission, scalable solutions for efficient and effective stratification of patients from primary care are urgently needed to get ahead on the long COVID treatment curve.

Because of the range of potential symptoms and effects of long COVID, AI-advanced imaging can give physicians a boost when it comes to detection and diagnosis. AI-advanced imaging can provide quantifiable metrics to the physician faster and more accurately than imaging by hand alone. This is particularly important when it comes to conditions like long COVID involving many symptoms which would otherwise be investigated in sequence one by one. AI-powered technology can enable parallel investigations so multiple symptoms are addressed in one go by enabling quantitative assessments to aid in the identification of impairment or morbidity (heart, lungs, liver, spleen, pancreas, and kidney) associated with the effects of long COVID. This can aid physicians in detecting organ damage to inform clinical management decisions, with:

  • Actionable data (e.g., cardiac pathology that may merit further monitoring and/or pharmacological treatment;
  • Reassurance and knowledge of affected organs; and
  • Potential prevention of mortality and morbidity with the early detection of at-risk organs (e.g., diabetes from pancreatic damage).

Most importantly from a patient perspective, AI-powered technology can lead to improved health-associated results. This means more metrics, more data, more transparency, fewer sequential imaging and referrals to multiple specialists, less time needed to arrange for additional transportation to medical appointments, and less taking time off work.

Moving Forward In Advanced Imaging Medtech

The U.K., in partnership with myriad stakeholders vested in long COVID care, has begun developing frameworks to provide advanced medical technology AI-based services like this to scale, and the U.S. would be wise to follow suit.

If we seek to understand why patients do not seek care, research finds many reasons — fear of discrimination, discomfort and intimidation with medical procedures, cost, and many more. Understanding how innovative technology can meet patients’ needs – and mitigate some of their concerns — needs to be a critical goal for those of us in the field of medical technology. By offering a patient experience that can at the start be communicated as noninvasive (in more ways than just in the body, such as time away from work, etc.), patients may be less hesitant to seek care.

In medicine, data is power, and we know that there are increasingly more advanced ways to use data to deliver powerful images (and more) that can be interpreted to support physician and patient decision-making. As we continue to battle existing long COVID cases with undoubtedly more variants on the way, we must use every tool and resource at our disposal to help those managing the disease. Therefore, it is abundantly clear to me why we need to accelerate the adoption of AI-driven solutions in the field of medical imaging and beyond.

The range of applications for AI-powered technology is limitless. AI-powered technology can change the future of healthcare for the better, helping physicians deliver higher quality and accessible care that gives patients more power over their health.

About The Author:

Rajarshi Banerjee, Ph.D., has served as CEO and a member of the board of directors of Perspectum since November 2012. Previously, Banerjee was a research fellow at the University of Oxford and developed the magnetic resonance imaging techniques for rapid noninvasive liver assessment and commercialized the method as LiverMultiScan, which has received 510(k) clearance from the FDA and is used in over 400 sites. Banerjee continues to work as a consultant physician with the Oxford University Hospitals NHS Foundation Trust, with research into the phenotyping of liver disease. Banerjee received a BM. BCh. in medicine and a Ph.D. in cardiovascular health from the University of Oxford, as well as an MSc. in public health from the London School of Hygiene and Tropical Medicine.