Using speech recognition to order rideshare trips with a smartphone app, filtering out spam, and conducting background checks are everyday examples of Artificial Intelligence (AI) at work. AI, which makes it possible for machines to learn from experience, is powering innovations that are transforming industries such as finance, manufacturing, retail, and transportation. While AI-powered self-driving cars and cashier-free stores may be generating the most media buzz now, some major AI innovations of the not-so-distant future promise to be in medicine.
AI is expected to revolutionize drug discovery; not by replacing researchers, but by complementing their efforts. Today, scientists across the globe are working to develop more than 8,000 medicines for disorders such as cancer, diabetes, HIV/AIDS, and cardiovascular diseases. Yet, only 12 percent of potential treatments that enter Phase I clinical trials ever make it to patients. Due to the high failure rate, taking a new medicine from initial drug discovery through regulatory approval takes an average of 10 years and an estimated $2.6 billion.
Using AI could help speed scientific progress, which has the potential to dramatically cut the R&D challenges associated with drug development. By nature, the path to scientific discovery is long, winding, and filled with trial and error. Myriad chemical compounds exist inside laboratories, each with a different potential to affect biology. Historically, perhaps one of the greatest challenges for scientists has been trying to develop those substances to target human genetics, which affect how our bodies react to a medicine.
AI-powered innovations could help researchers make those connections. For example, in 2018, an AI program known as “Eve” helped scientists discover that triclosan, an ingredient found in many toothpastes, could potentially treat drug-resistant malaria, a growing public health issue in places such as Africa and Southeast Asia.
To make their discovery, researchers created genetically-altered strains of yeast that included either genes from parasites that cause malaria or human genes. Eve screened thousands of compounds, eventually identifying triclosan as a compound that stopped or slowed the growth of the parasitic genes, but not the human genes.
Researchers previously suspected that triclosan—which became a toothpaste ingredient due to its anti-microbial properties—could help treat malaria. So, Eve didn’t invent a new medicine (although debates are underway in the biopharmaceutical industry as to whether this could become a reality). Instead, Eve merely helped scientists zero in on the most likely compound.
A common application of AI in medicine is to help monitor vast amounts of medication data so doctors and researchers can spot potential negative effects of treatments. AI algorithms are only as good as the data on which they rely; therefore, quality evidence is necessary to draw accurate conclusions. In recent years, biopharmaceutical companies and clinical researchers have begun to rely on the growing body of real-world evidence (RWE), which measures data on clinical delivery and outcomes, pooled from health records, billing claims, and other relevant sources. RWE also is generated through the growing use of wearable devices, such as the mobile health trackers found in many smartphones.
RWE already is changing drug discovery and development by improving biopharmaceutical companies ability to reduce barriers, such as in clinical trials. Recent data shows 100 percent of PhRMA (Pharmaceutical Research and Manufacturers of America) member companies are already incorporating some form of RWE into their drug development programs. The U.S. Food and Drug Administration (FDA) currently uses RWE to study drugs once they are in the market. Moreover, clinical trials participants can now report their experience through mobile apps, meaning people will have to spend less time and money traveling to clinical trial sites, and scientists will be better able to predict earlier when a patient isn’t adhering to his or her treatment protocol.
AI already is an integral part of everyday life and will continue to transform the world. For biopharmaceutical researchers, in particular, AI is a welcome—and potentially ground-breaking—tool they can use in the effort to develop new medical treatments, while still ensuring accuracy, effectiveness, and patient safety.