This AI Could Revolutionize Cancer Treatment

Monday, 30 October 2017 - 11:01AM
Medical Tech
Monday, 30 October 2017 - 11:01AM
This AI Could Revolutionize Cancer Treatment
< >
Image Credit: Pixabay

Scientists may have found the perfect method for identifying cancer with pinpoint accuracy.

A new study from Showa University in Japan has analyzed the efficacy of using deep-learning AI to identify early signs of cancer.

The study involved showing the AI photos of colorectal polyps, which may turn into full-blown dangerous tumors, or end up proving to be benign.

Using machine-learning algorithms, the AI then cross-referenced these images with a databank of over 30,000 similar photos and their ultimate outcomes in order to make an instantaneous prediction about whether the polyps in the photos would ultimately develop into a form of colorectal cancer.

The outcome? The AI was able to predict the eventual manifestation of these polyps 86 percent of the time, giving doctors an incredibly advanced warning about whether or not the patients in their care would go on to need treatment.

According to the study:

Opening quote
"Overall, 306 polyps were assessed real-time by using the AI-assisted system, providing a sensitivity of 94 percent specificity of 79 percent, accuracy of 86 percent, and positive and negative predictive values of 79 percent and 93 percent respectively, in identifying neoplastic changes."
Closing quote

While these results sound incredible, it's important to note that this is only an initial study and one that doesn't have a huge body of data.

Before these results can be verified, they'll need to be achieved in a series of different repeat trials to prove that the findings of Showa's study are accurate.

That said, it's exciting to think of the potential benefits of using AI for tumor detection.

Similar technology is used in a lot of different deep-learning programs, from those that are designed to play board games, to those used to translate different languages within Facebook.

As programs like AlphaGo continue to grow in sophistication, they improve their success rate at making these kinds of calculated predictions.

What may be an 86 percent accuracy rate today will likely only increase in the future as AI programs are able to pool their findings, combining the results from hundreds of thousands of test cases in order to learn all the telltale signs of medical issues.

We've all used the internet to diagnose our own aches, pains, and ailments. While in the past the results of this kind of fact-finding have been laughably inaccurate, that time is coming to an end.

As computers improve their capabilities to learn about the world around them, their ability to make accurate diagnoses for a variety of illnesses will only increase as time goes on.

This doesn't come without risks for humans, though.

AI programs are already replacing humans as insurance brokers, making calculations about the safety of offering insurance policies. If a computer can accurately detect customers with potential illnesses, these machines may be responsible for denying medical insurance to those who are most in need of coverage.

If diseases are being caught earlier, in theory, lengthy medical treatments should be less common, but this is a two-edged sword: robots help us stay healthy while simultaneously potentially collecting more information than we'd like.

Either way, one thing is certain.

In years to come, the vast majority of cancer cases will be caught far earlier than they are today. Robots are fighting cancer, and humanity has a lot to gain from helping our metal friends to keep us healthy.

Artificial Intelligence
Medical Tech
This AI Could Transform Cancer Treatment in Future