If you were to walk up to someone you know and mention the phrase “artificial intelligence,” chances are pretty good their mind would immediately make the leap to the silver screen. For decades, we’ve seen autonomous robotic minds both help and hinder mankind from achieving its goals, but what we often see on the TV and in the movie theater is nowhere close to its real-world applications.
Artificial intelligence, also known as AI, is probably in your living room or kitchen right now. A growing number of homes are making use of smart personal assistants, so if you’ve ever asked Alexa or Siri to help you find a Chinese restaurant, you’ve used AI. If you’ve ever taken a flight on an airplane, marveled at a self-driving car or even taken a peek at your email spam filter, each of those processes makes use of artificial intelligence and machine learning.
But what about situations where decisions may mean the difference between life and death? Artificial intelligence and machine learning are beginning to make a massive impact in how we provide and use healthcare, especially when it comes to how we treat and diagnose diseases in patients. While the technology may not be perfect, healthcare AI has a wide variety of uses and will likely change the way of healthcare as we know it in the near future.
Diagnosing Diseases Using AI
According to the National Cancer Institute, about 9.2 million person-years of life were lost to cancer deaths in 2012, resulting in a loss of nearly 16 years per person. While AI isn’t yet capable of taking on every task associated with diagnosing and treating cancer, it is making the job easier for the doctors and specialists serving on the front lines.
It’s highly unlikely that a computer will perform your physical examination in the next few years, but doctors’ offices are slowly beginning to adopt some of the technologies highlighted by artificial intelligence. For example, physicians are starting to make more use of AI’s capabilities when it comes to monotonous jobs that eat up valuable time. From data entry to X-ray and CT scans, AI is coming in handy, saving time and performing tasks with a higher degree of accuracy than their human counterparts.
However, when it comes to diagnosing and treating a disease as awful and difficult as cancer, AI is only scratching the surface of what it’s capable of. For years, researchers and developers have promised the technology as the wave of the future, but it’s largely been a series of small breakthroughs with an indeterminate impact on patient health outcomes. Still, AI has shown some early success in diagnosing and identifying several types of cancer, including those of the breast and colon.
According to research coming out of the Netherlands, scientists wanted to see if deep learning algorithms could fare better than a doctor in determining the spread of breast cancer to the lymph nodes. In the study, the algorithms were directly compared to 11 pathologists in both timed and non-timed simulations. In some cases, artificial intelligence performed as well or better than the pathologists. The tests need to be run again in a clinical setting to reaffirm the current results, but researchers believe AI could be used to assist pathologists save some time and prevent them from burning out in the long-run.
Meanwhile, researchers in South Korea recently performed a study using AI to identify tumors in 30 colorectal cancer patients using convolutional neural networks and histology images. Despite some limitations with the study, including the size of the images used, machine learning was able to reach a high point of 93 percent accuracy and 92 percent specificity.
These breakthroughs are providing researchers and patients with much needed hope. This is especially true when it comes to diagnosing, treating, and eventually curing rare diseases, including sarcomas, lymphoma and preventable diseases like mesothelioma.
In early August, an organization in Edinburgh, Scotland, was awarded £140,000 to study and develop AI to help better fight malignant pleural mesothelioma. Across all types, mesothelioma was responsible for nearly 2,500 deaths in the United Kingdom in 2015. The hope is that AI can play a vital role in reducing the amount of time spent diagnosing diseases like mesothelioma, but also lower the costs associated with research, which would provide more efficient and effective clinical trials for patients. In that same vein, drug companies can use AI to reduce the cost of developing new treatments.
AI’s Next Steps
Artificial intelligence is still a very green technology, but its ever-evolving nature opens the door for improvement across a wide spectrum of fields, including healthcare. The promise it provides researchers, physicians, patients and drug developers is undeniable, despite its growing pains. AI is gaining momentum, and while we won’t be treated to a virtual visit with a robotic doctor in the near future, physicians will soon be capable of spending less time on repetitive tasks and focusing more on what matters most; making people well.
Written by Anna Suarez