The American Board of Imaging Informatics (ABII), eloquently describes informatics as the ‘intersection of information science, computer science and healthcare.” It’s the ability to access, understand and use information in healthcare in ways that are far more effective and that enhance services and patient care. It’s also an area that has seen significant innovation and development, specifically in artificial intelligence (AI) solutions. Informatics radiology solutions are leading the way when it comes to capability and invention.
In a recent review written for Academic Radiology and entitled ‘The Importance of Imaging Informatics and Informaticists in the Implementation of AI’, the author points out that imaging informatics is critical to the success of AI implementation in radiology because the informaticist introduces an understanding of the data that’s critical to the success of the AI solution.
Ultimately, the value of the algorithm lies in its ability to extrapolate from the data. The value of informatics radiology lies in the data that populates the AI solution. Together, they allow for a more comprehensive and reliable implementation of any AI platform in radiology. As Cook says, “An imaging informaticist is a unique individual who sits at the intersection of clinical radiology, data science and information technology.’ These multiple layers of expertise ensure that AI deployment has value and meaning beyond just the application of an algorithm.
The value of data in informatics radiology and AI implementations has long been recognized by the medical profession. The Digital Imaging Adoption Model (DIAM) was developed as a strategic roadmap to digital imaging security that allows for the profession to better assess its digital strategies so that any investment would ultimately result in improved outcomes for patients. Developed as a collaborative effort between the Society of Imaging Informatics in Medicine, the European Society of Radiology, and the European Society of Medical Imaging Informatics, DIAM is focused on providing clear guidelines into digital, health and strategic planning.
Advanced radiology imaging informatics
Informatics in radiology has always been advanced, but the introduction of AI has taken this to the next level. AI in radiology is increasingly prevalent as more and more organizations recognize the value that it brings and the support that it can offer radiologists themselves. AI, however, requires significant integration to ensure that it’s capable of getting the results that it has promised. For example, many of the organization’s that develop AI solutions for radiology are not au fait with imaging informatics along with some of the other more complex aspects of the industry. This means there is sometimes a gap between what the AI can do and what the AI is capable of doing.
In the paper mentioned earlier, Cook points out that this is where the informaticist role is extremely important. Through collaboration, informatics radiology and AI can populate the system and develop the data to ensure that the overall system is more cohesive and holistic. This takes the conversation back to the data – not many AI algorithms have the right data so they need to be immersed into the clinical environment and undergo extensive testing and implementation to ensure that they deliver the right results.
In a recent analysis into how organizations should implement AI-based solutions into radiology, the founder of Aidoc, Elad Walach, pointed out clinical validation is one of the most important factors to consider before investing into any AI application or partnerships. AI vendors have to demonstrate data accuracy from the outset but this is only one part of the whole equation. The solution has to show measurable results within the clinical context; have a proven track record in live clinical environments; and show competence across results, workflows and organizations. And, of course, it should have a comprehensive roadmap that outlines the pathologies it is planning to undertake over the next five to ten years. The futureproof capabilities of an AI radiology vendor are a critical factor.
Informatics in radiology
In the past, AI solutions for radiology received some suspicions that they’re set to steal their jobs. The reality is that informatics radiology and the radiologist are essential to the ongoing success of any AI application and implementation. It’s the informatics in radiology that cement the value of the data – something that is of critical importance to ensure the right results – and it’s the radiologist that ensures the human edge to analysis that makes all the difference in patient care and diagnosis. AI definitely changes the shape of the role – speeding up diagnosis, flagging concerns early, supporting overburdened professionals and streamlining workflows – but it’s only one part of the whole equation.
Informatics radiology is also another part of this equation and the evolving role of the radiologist. In the past, the job is far different from what it is today. The profession has evolved alongside the technology to make the role far more dynamic and impactful than ever before. What the radiologist brings to the table in terms of decision making, insight and understanding is essential and powered by informatics radiology, AI and innovations in medical technology.
Aidoc has been working with radiologists in more than 300 live installations across the world. The solution has a solid reputation and regulatory approval across multiple modalities and has been described by radiologists as the ‘team member that never sleeps’. Aidoc has also developed a clearly defined roadmap. Initially, this focused on acute pathologies and covered around 80% of CT scans but, since 2019, the company has moved into non-acute pathologies for CT such as oncology.