AI in drug development and personalized medicine
Medicine and other scientific areas have always used computing power wherever they could find it-- to help modeling go faster and arrive at viable drugs more quickly.
But when we apply the most recent advancements in artificial intelligence to the most advanced drug development programs, we get something else entirely: truly "personalized medicine." But what's personalized about it, and how does AI play a role?
What's New in AI and Drug Development?
Multiple studies have been published recently describing practical ways of using artificial intelligence to achieve personalized healthcare. Researchers who've published in SLAS Technology and Science Translational Medicine outline the use of AI-powered "quadratic phenotypic optimization platform," or QPOP.
This is a fancy name for something scientists of all kinds have long coveted: a computer powerful enough, and an algorithm complex enough, to help churn out and test as many viable new drug formulas as possible, to either:
- Render a viable course of treatment for a patient whose physiology made conventional treatments ineffective or their condition impossible to diagnose.
- Safely computer-model as many potential new drug formulas as possible to beat competitors to market or improve on existing formulas.
In one of these "commentary" pieces about the "QPOP" method, the researchers used multiple myeloma as an example of a stubbornly drug-resistant condition. With the AI platform "crunching the numbers," as it were, small pools of data were tested in rapid succession and extremely efficiently. More efficiently than any team of human researchers could have managed.
The result was that the AI uncovered a much larger pool of potential patients. More specifically, the AI accounted for the many complex "signaling pathways" that cancers use to spread through the human body. Since cancer and other serious ailments progress a little differently for each patient based on their unique physiology, researchers needed a tool that could create as many viable solutions as possible, for as many patients as possible.
They found that tool in AI. And with it, many more potential positive outlooks for patients in need of relief.
Other Applications for AI in Bespoke Medicine
If it sounds like the "drug combinations" here just involved mixing a few ingredients together in different proportions, know that it's a lot more complicated than that. This AI platform was dealt the task of tailoring compounds for "weaponized" use against molecular-level targets. No easy feat -- but this combined research, and the detailed commentary available on it, has shown a way forward.
In fact, this is only one of quite a few tentative forays into using AI to "solve" for the personalized ways that disease impacts the body. There are several entirely different class of applications for AI, too. One of the earliest has been in the development of the Industrial Internet of Things -- and especially the IIoT as it applies to clean rooms, cold chain storage, pharmaceutical and vaccine shipments, and more.
For example, it's a huge and costly problem that about 4 percent of pharmaceuticals go to waste because they don't arrive in usable condition. The availability of more sophisticated sensors for drug shipments, not to mention lower barriers of entry for wireless connectivity, means that when pharmaceutical shipments are subjected to temperature variances in transit, automated climate control equipment can make adjustments as needed.
Artificial intelligence is everywhere else in drug development and healthcare, too. And it should help out in a number of ways in our quest for a more accessible, social, transparent, accurate and efficient healthcare system. A look at the entrepreneurial scene will turn up many dozens of young companies who've identified AI as one of the most significant areas of innovation in the healthcare fields in a very long time.
Here's just a brief rundown of AI-driven technology coming down the healthcare pipeline. Expect machine learning to soon play a vital role in:
- Building leaner and more automated administrative and bureaucratic functions in healthcare systems and small practices
- Matching applicants with clinical trials that suit their condition -- and for more specialized trials, that match any distinctive patient markers
- Sifting through available histories and health data for red flags that humans might miss
- Validating chemical testing results of all kinds much more quickly than with previous methods
- Perform protein modeling to better understand the mechanics of disease and reveal novel treatment methods
So it's not just about creating new drug combinations for patients who are already sick. In some cases, it's a matter of allowing AI platforms to access the data already accumulated on a patient -- such as relevant keyword phrases in medical records, immunization and medication histories, genetic markers, data from wearables and heart monitors -- and discovering how well it can predict serious conditions before they become untreatable or worse.
In cases like these, personalized medicine becomes proactive rather than reactive. Both are extremely promising avenues of ongoing research and innovation.
Kayla Matthews is a senior writer at MakeUseOf and a freelance writer for Digital Trends. To read more from Kayla, visit her website productivitybytes.com.