Recruiting within Data Science and having a partner who is a Surgeon got me thinking about the advances in medicine, and after a lengthy discussion I thought I’d share some of my learnings.
There is no doubt that advancement in science and technology has affected every aspect of human endeavor. Processes and procedures are being replaced with software, apps and the likes at different levels and medical practice is not left out.
It is believed that we are in a sort of revolution that is characterized by new technologies that are fusing the physical, digital and biological worlds, these have, in turn, impacted all disciplines, economies, and industries. And it is most certain that one of the major catalysts for the revolution in the medical practice is going to be artificial intelligence.
Years back, who would have imagined the use of artificial intelligence in the medical field? But today, it’s a different ball game as artificial intelligence (AI) algorithms are showing promises in performing medical procedures.
For example, Artificial Intelligence (AI) have been able to diagnose the presence or absence of tuberculosis (TB) in chest x-ray images with amazing accuracy. This feat was possible because researchers first “trained” the AIs with hundreds of x-ray images of patients with or without TB after which they tested the AIs with 150 fresh x-rays result. AI achieved and impressive 96% accuracy rate – a result that was found to be more accurate than may human radiologist. In fact, the researchers are of the opinion that they can improve upon this result with a more advanced learning model for the AI algorithm and more training cases. The authors of the above study noted that automated detection of TB at chest radiography might go a very long way in facilitating screening and evaluation efforts in TB-prevalent areas who have limited access to a radiologist.
What about other areas of medicine? That too! Similar AI algorithms have shown a level of success in other branches of medicine like pathology, ophthalmology and cardiology.
Interestingly, Researchers at Google were able to train an AI to detect the spread of breast cancer into lymph node tissue on microscopic specimen images with a greater accuracy when compared to a human pathologist. Looking out for tiny deposits of cancer on a sample slide can be challenging, while the human pathologist may suffer from fatigue or inattention, the AI, on the other hand, will be able to effortlessly process large images of these tiny deposits of cancerous cells on a specimen slide and provide a more accurate result.
Also, similar AI have shown greater successes at detecting changes in diabetes in the images of patient’s retina better than human physicians.
Most intriguing is the possibility of AI identifying new associations and correlations that are yet to be detected by humans. For example, UK researchers turned over the data of about 295,000 patients to AI, to allow them to correlate medical history with the rate of heart attacks. After that, the AI was given another record of 82,000 patients whose history of heart attacks were already known for the AI to predict the ones that are most likely to have a heart attack. The result of the AI when compared to the predictions based on current “best practice” American College of Cardiology/American Heart Association (ACC/AHA) guidelines, which include patient age, smoking history, cholesterol levels, diabetes history, etc. the AI beat the human’s hands down.
Finally, with time, AIs are more likely to displace many practitioners in the medical practice. And the fact remains that, the potential benefits of AIs being deployed in medical practice far outweigh the short-term cost.
What do you think the future holds with regards to AI and the medical field?
I’d love your thoughts on this as well as the article!