Written by Deepak Kumar Nayak and Aryan Pratyush Nayak and Guided by Dr. Anand Agarwal.
Heart diseases and their treatment predominantly rely on the time at which they are detected. The sooner the detection, the quicker the treatment is, and the better are the results. But there lies the problem. Early detection of heart attacks and other heart-associated ailments has still been restricted to theory and has been very marginally executed in the practical world. This leaves us in an uncomfortable void for necessary technology that could help us get the work done. Thankfully Artificial intelligence not just only fills that void but also aims at having more profound effects.
WHAT IS ARTIFICIAL INTELLIGENCE?
Artificial intelligence is the intelligence shown by machines and is unlike the intelligence shown by organisms. The intelligence showed by humans, and other animals is affected mainly by emotional consciousness. However, on the other hand, artificial intelligence runs purely on algorithms and information provided to it, so there are very slight chances of the decisions taken by it being clouded under any influence. There are many possible uses of this technology in the real world, and its specific prophecy in medical advancements is what we think will attract many.
- The standard clinical tests used by cardiologists have nearly outlived their lifespan and have become somewhat, if not rudimentary. Artificial intelligence aims to break the mould by predicting exact results that would remain significant for a longer period as digesting and synthesizing vast volumes of information is what it does best.
- A study from “circulation” published in February 2014 shows that the AI program can reliably predict heart attacks and strokes with high precision. Kristopher Knott, a research fellow at the British Heart Foundation, and his team conducted the largest study yet involving cardiovascular magnetic resonance imaging (CMR) and AI. CMR is a scan that measures blood flow to the heart and accordingly decides the possible amounts of blockages in the heart vessels.
- However, reading those scans is time-consuming, and it also gives space for human error on the part of the doctor analyzing it. To counter this, Knott and his colleagues trained an AI model to read scans and to detect signs of compromised blood flow. The model was tested on 1000 people, and it showed perfect results on the people who were more likely to have a heart attack or die from one. The tests gave in quantitative output, which was specific instead of qualitative result which the doctors had to still break and synthesize into helpful information.
- We could use AI technology to pick out people at risk and employ on general people who could be helped with the information as a precautionary tale. For example, the test could be conducted on any youngster and detection of irregularities. They could be advised to take the necessary steps to get a better blood flow in the heart vessels, which would induce a healthy lifestyle that will profit for times to come.
THE CHALLENGES THAT LIE AHEAD
- No matter how revolutionary technology is, at the end of the day, it still is reduced to its practical effectiveness in the real world. For AI to be effective in medical science, it needs to be cleverly and adequately integrated into the existing workforce.
- Tina Manoharan, Global Lead Data science and AI center of excellence and Digital research division at Phillips, said that they need to understand the workflow of health care professionals first and design a solution with human-AI collaboration in mind. It becomes highly essential that proper training and education be given to doctors and health care workers before they start implementing AI models into practicality.
- The next challenge is to integrate consolidated health data into the AI model. AI models work on carefully sets of curated data. In real life, however, the data provided by the healthcare field is messy, fragmented, and hard to access. It isn’t easy to exchange, analyze, and interpret, so it is essential to provide one integrated model that offers seamless solutions. This can only be achieved by concerted efforts between health care organizations, vendors, and hospitals.
- For the model to work, it needs a data reservoir, and the data reservoir is essentially all of an individual’s experiences tracked 24/7, their medical history, primary health care records, and something as essential as their daily home routine. This may develop a new set of legal challenges, and medical data is susceptible and is subjected to privacy regulations. These regulations change as per jurisdictions and hence can create roadblocks.
WHAT DOES THE FUTURE HOLD?
There are many possible AI solutions in the field of healthcare, like Artificial Intelligence-Clinical Decision Support System (AI-CDSS) that aims to assist physicians with analysis to diagnostic results or Analytics for Life machine learning algorithm that aims at detecting chances of coronary artery disease. However, we should still be looking forward to a much optimal and proper model that would seek to solve most problems in the future. For instance, the present solutions that we have proven to be effective, but accessibility to such solutions is restrictive. Hence the direction of approach should be towards solutions that address these problems. With further discovery down the path, we probably would be introduced to newer challenges and newer questions whose answers we don’t have yet. Hopefully, with different advancements in the field, we can probably clarify the subject, which would be beneficial to medical science as a whole.