AI Targets Dangerous SuperBugs; How Healthtech is Saving Lives

AI Targets Dangerous SuperBugs; How Healthtech is Saving Lives

In 2014, the BBC published an article stating that antibiotic-resistant strains of bacteria—so-called superbugs—could be expected to kill more people than cancer annually by the year 2050. Currently, superbugs are responsible for $20 billion in extra healthcare costs, $35 billion in societal costs, 8 million additional hospital days, and nearly 50,000 deaths in the US alone.

On February 20th, a group of MIT scientists published a study in Cell that detailed the results of an effort to identify new antibiotic compounds using a Deep Learning computer model that can screen over ‘a hundred million chemical compounds in a matter of days’. After training the model on roughly 2,500 molecules, scientists tested its effectiveness using the Broad Institute’s Drug Repurposing Hub and the 6,000 compounds it contains.

The result was the discovery of a molecule predicted to have powerful antibiotic properties and a different chemical structure than any currently known antibiotic. The researchers named it ‘Halicin’ after the AI featured in 2001: A Space Odyssey.

Halicin was further tested by a different computer model to ensure that its toxicity was low for human cells, then tested on a collection of dangerous bacterial strains, including Acinetobacter baumannii—a drug-resistant organism known to infect soldiers in Iraq and Afghanistan through their wounds. An Halicin topical ointment eradicated the Acinetobacter baumannii from colonized skin and infected wounds within 24 hours.

In addition to Halicin, the computer model identified 23 other molecules predicted to be effective in the treatment of bacterial infections, searching over 100 million molecules from the ZINC15 database in 3 days.

To put this into perspective, the Deloitte Center for Health Solutions estimates that the total average cost of launching a new drug is over $2 billion and that 1/3 of that cost is spent in the pretrial R&D aimed at discovering new molecules—a task that MIT’s AI model took 3 days to complete.

Deloittle concluded that the potential of AI went beyond just the identification of new molecules for drug treatments, arguing that the cost reduction alone could revitalize the pharmaceutical industry.

The average forecast peak sales per late-stage asset in the drug pipeline declined to US$407 million in 2018, less than half the 2010 value of $816 million. As a result, the expected return on investment from drug development has declined steadily from 10.1 percent in 2010 to 1.9 percent in 2018. Finding ways of improving the efficiency and cost-effectiveness of bringing new drugs to the market is imperative for the industry.

BenchSci, a market watch organization focused on AI-assisted drug development, reported that, as of February 5, 2020, there are currently 203 tech startups in the US and Europe using machine learning to discover new drugs and advance other aspects of pharmaceutical healthtech, such as identifying patient markers.

Investment in AI-powered research is growing, and this represents a revolutionary step forward in the healthtech industry. The next generation of effective drug treatments may be brought to market in record time and at a vastly reduced R&D cost, allowing pharmaceutical companies to offer life-saving treatments at competitive prices while seeing a greater ROI. That is in everyone’s interest.

Would you like to know more about how we help companies create AI solutions that will improve the world? Do you have questions about how AI can power your business or fulfill a great new vision for your company? Send us an email, or leave a comment below. We love to discuss the future of AI technology and how we can help startups and established businesses to benefit from these game-changing solutions.

Submit a Comment

Your email address will not be published. Required fields are marked *