Dr Raju Kalidindi, CEO and Medical Director, Apollo Radiology International elucidates on the advantages of simultaneously observing and rapidly processing almost limitless number of inputs with AI, unlike what single radiologist can do
With technological advances transforming all aspects of life, would medicine be far behind? Among the many ways in which technology has changed the practice of medicine is the use of artificial intelligence (AI) and machine learning (ML) in diagnostics, drug development, personalisation of treatment and the futuristic arena of gene editing. November 8 being International Day of Radiology, let us take a look at how AI is making inroads in the practice of radiology.
Advances in AI led by developments in image-recognition and interpretation of sensory information, have resulted in applications that have brought a paradigm shift to the practice of radiology through tools designed to interpret complex radiological data. AI provides techniques that uncover complex associations in a similar fashion to the human brain. While allowing AI to interpret data just as a radiologist might, AI has the advantage that it can simultaneously observe and rapidly process an almost limitless number of inputs, unlike what a single radiologist can do.
Traditionally, radiologists visually assess medical images and report findings to detect, characterise and monitor diseases. Such an assessment is based on training and experience, and this leads to conclusion that are often subjective. This is where AI has the upper hand. AI excels at automatically recognising and rapidly analysing complex patterns in imaging data. This provides an objective assessment in an automated fashion.
The advantages of AI in radiology are many. Today, radiologists face a challenge with a huge increase in radiology investigations and shortage of radiologists. In such a situation, where human expertise is scarce, a single AI system can help support a large population, helping radiologists to become more efficient and effective. By virtue of its ability to find patterns in data that humans cannot see, AI can take over time-consuming monotonous tasks, freeing radiologists for more qualitative decision making. Integrating AI into a clinical workflow can free radiologists from tedious tasks.
AI can be used in preventive and personalised health checks. AI can help in detection of subtle tumours, which the human eye may miss. In life threatening conditions such as bleeding in the brain, AI can help in immediate diagnosis. AI can do things that are beyond human capabilities such as predicting the risk of breast cancer even from normal mammograms, and in the future, help in detection of molecular markers in tumours.
Tuberculosis is a major problem in India, and tackling it requires a major push in early diagnosis. In remote health centres, there may be a lack of radiologists. However, x-rays can be uploaded from these centres and sent to a single central system where they can be interpreted using AI. Another example of the use of AI in radiology can be in lung cancer, where AI can be used to identify pulmonary nodules as well as to categorise them as benign or malignant. In screening mammography, which is challenging to interpret, AI can help in identifying micro-calcifications and reduce false diagnosis of malignancy. In case of liver lesions, brain tumours and colonic polyps, AI can again help in categorising the lesions as benign or malignant and prioritising follow-up evaluation. In radiation oncology, AI can provide assessment of treatment response fast and accurately.
In the near future, AI in radiology will help radiologists in characterising and detecting disease, and standardising the reporting. AI can play a major part in predicting and monitoring disease and empower radiologists to make superior clinical decisions.