What role AI is playing in the field of medical diagnosis and how are radiologists benefiting from it?
Artificial Intelligence (AI) is rapidly entering healthcare and supporting functions from robotic surgeries to streamlining routine workflows in medical practices to managing patients and medical resources. However, when we look at Artificial Intelligence (AI) within the field of medical imaging field, it has come a long way in aiding radiologists for better decision support systems. From providing better insights during diagnosis to improving the accuracy of predicting tumor malignancy, AI is supporting radiologists in various niche and subdomains of radiology. It is improving the computational powers of traditional radio-imaging techniques such as MRI, and CAT-scans through complex algorithms such as neural network and deep learning. These higher computational capabilities, in turn continue to improve the radiograph quality, and results in a faster and improved diagnosis.How mature is the field of AI application in medical diagnosis?
Though application of AI is progressing at a fast rate, it is still relatively new when compared with some of the other verticals where AI has been implemented. However, when we look at the growth of AI in medical imaging and radio-diagnosis, we clearly see unprecedented rate of the sector. By a report published by Signify Research- a leading market research firm within healthcare and AI realm, the world market for machine learning in medical imaging, is set for a very rapid growth and forecasted to cross the $2 billion by 2023, up from $1.2 billion market in 2017.
But are the regulatory bodies ready for the future of medical imaging?
I would be lying if I say that regulatory bodies like ICMR, FDA, EMA have everything under control when it comes to automations and AI, especially in medical diagnosis. However, these regulatory bodies are taking necessary steps to incorporate these AI advancements and understand their unique preposition in the bigger context of healthcare improvement. From establishing good AI standards to reviewing pre-submissions to increasing transparency in use of real-world performance reporting, the regulatory bodies are making strides to accommodate these newer AI guided diagnostic approach.
Could you elaborate on some of the applications of AI as it applies to the field of medical imaging?
Artificial intelligence and data analytics technologies are currently being implemented in a variety of medical use cases. Some of the most common applications include:
- Radiograph Review: Leveraging AI and machine learning to detect anomalies and diseases based on algorithms to better assist physicians.
- Early screening and diagnosis: Many studies have shown that early markers in diseases such as ALS, Alzheimer can be picked up through AI mediated computer imaging
- Real World Image Analysis: AI through deep mining can support physicians by analysis similar records through data repositories of the past diagnosis. This layered with AI enabled literature review could provide great outcome driven results for the physicians.
What do you see as some of the biggest issues in AI mediated medical imaging?
As with any other technology, there are some challenges that inhibit not only the growth and adoption of AI but also possess serious threats to a healthy function of the society. Regulatory concerns (AI algorithms, as a part of the applications are not subjected to same scrutiny as those faced by similar purpose medical devices), ethical concerns (replacing physician’s subjective judgment to diagnose medical conditions) and data privacy concerns (data ownership need to reside within enterprise than with the patients to build and deploy robust AI models) would top my list of concerns when it comes to AI application in medical diagnosis.
Where do you see growth of AI in medical imaging growing?
Despite strong investments in AI-based medical imaging, the growth and AI adoption has been rather gradual, primarily driven by regulatory challenges and physician skepticism. However, I feel that with ever-improving algorithms, slow but gradual openness by physicians and new regulatory guidelines getting drafted by the respective authorizing agencies, AI will see an increased adoption, more so in the field of optimizing radiology workflow, image triage and clinician decision support system.