Can India take AI road to excellence in radiology?

February 5, 2019 0 By FM

Artificial intelligence (AI) has made several strides in recent years, allowing machines to interpret complex data leading to advances like self-driving vehicles to natural language processing.
AI is the much-needed tool to make healthcare reach all segments in India as it can reduce costs, improve reach and clinical outcomes.
New applications like big data analysis in DNA and RNA sequencing will help improve healthcare and India’s computational capabilities can certainly make it a hub of AI in healthcare. Imaging in fields like pathology, dermatology and ophthalmology have already started benefiting through the implementation of AI. The application of AI in radiology is called radiomics, and it must be prioritised in India due to the fact that we have a skewed doctor-to-patient ratio and an ever-growing population.
There is an urgent need to leverage technology to face this huge task of adoption of digital healthcare in day-to-day clinical decision-making, which will make patient care more affordable. By combining teleradiology and AI in India, we can provide better approaches in reaching out to rural areas. According to a recent publication by CIS (The Centre for Internet and Society), AI could potentially add $957 billion (15% of current gross value added) to the Indian economy by 2035.
One of the main obstacles in implementing AI in Indian healthcare is not technological, but related to data access. Other obstacles include consent and ensuring clean, uniform and digitized data. As hospitals are typically only contact centres for illness, and care is often delivered through multiple physicians, it makes AI less penetrative despite India being relatively data-dense. Furthermore, even though India started adopting electronic healthcare records (EHR) policy, the same is not robust enough and is in fact highly inconsistent and not harmonized, leading to several difficulties. Along with the above problems, radiologists are overloaded with work and left with no time to do research on incorporating AI into their practice, which in turn results in industries not having enough backing and support from clinical practitioners. Another aspect of the adoption of AI in radiology is that startup companies are often forced to show proof of concept in the form of clinical trials/studies and this takes enormous investments of both money and time. A possible solution to this problem could be for the industry to have strong collaborations or partnerships with radiologists to perform clinical trials and studies.
In order to enable true implementation of AI in radiology, it is important to develop a multi-stakeholder plan incorporating all relevant sectors to formulate comprehensive guidelines. It is also imperative to encourage digitization and open data systems, and set standards for data collection, privacy and safety. This can encourage research and development in AI applications, while promoting public-private partnerships. India is presently in a unique position to be in the driver seat in radiomics and digital healthcare in general with a flourishing startup community and a realisation by clinicians, scientists, policymakers and the general population about the importance of AI. We have joined the revolution of AI and will have to clear each obstacle as we move forward. This can be sustainable only with robust skill development and the establishment of an ecosystem of interdisciplinary research continuum of AI in healthcare.