Artificial Intelligence Already Revolutionizing Pharma

Jan 10, 2018

As the pharma industry convened at J.P. Morgan Healthcare Conference in San Francisco this month, we are looking back at the WuXi Global Forum, now one of the events featured in the week. Running for the past six years, the forum has become a hub for the industry to come together with 2,000 executives, focussing on the most transformational models and solutions that will deliver tomorrow’s healthcare innovations.

One of the key topics in 2018 is how artificial intelligence (AI) is drawing plenty of attention in the pharmaceutical industry, based on its remarkable feats in other industries—namely, successfully training machines to learn how to recognize faces, how to talk, how to drive cars, how to play games, and how to compose music. Uptake in pharma has been relatively slow, but at last it is on the way. Now, what everyone wants to know is how the technology works, its benefits and challenges, and how it will affect drug discovery and development—especially costs—in the long-term.

Using computer algorithms, AI can teach machines how to unravel raw complex data through the detection of patterns. This makes AI and its subsets, machine learning and deep learning, a natural fit for mining and relating the mountainous genotypic and phenotypic data being collected worldwide in public and private databases, in hospitals and doctors’ offices, in academic research journals and in individual wearable health monitoring devices.

Though no AI-driven drug has acquired regulatory approval yet, experts across the board anticipate that implementing AI will soon be necessary to compete in the industry. In the next ten years, they say AI will be universally integrated into pharma R&D operations. The chief reason for this is that AI seems to be able to solve the problems that have perpetually plagued the field, notably time for drug discovery and clinical trial success rate. With 90% of drug candidates failing to reach approval and the expense of clinical trial failure estimated to be up to $1.4 billion (over half of the average total new drug cost), pharma companies are very much incentivized by the prospect of improved drug development success.

What usually makes drug discovery time-consuming and inefficient is the lack of understanding of the biological intricacies of disease, and it is this that AI could impact on—particularly in helping to identify true causal genes and pathways for complex diseases, which still eludes scientists despite advances in genomic sequencing. By having a more comprehensive understanding of the biological basis of disease through AI, the R&D process will be more streamlined with scientists in a better position to develop therapies from the offset, without so much trial and error. Additionally, it is hoped that AI will also help eliminate the element of “luck” that is observed in whether a drug is successful or not in clinical trials by accurately identifying the subset of patients who will benefit. Reducing the failure rate will in turn save the industry billions of dollars.

Not only that, if a drug is able to get to the market quicker, there will be more years of patent-protected market exclusivity for pharma innovators to achieve a profit from the initial investment. Drug patents last 20 years, but the period of initial filing to regulatory approval leaves on average only seven years of exclusivity, often inadequate to regain costs. Both savings and gains would presumably get carried over to the consumers in the form of cheaper medicines – a divergence from the current trend of rising drug prices.

However, not to worry—human scientists will not be replaced by AI anytime soon, but rather they will be expected to work alongside the technology through asking the appropriate questions and providing a sufficient amount of data in order to compute the algorithms. Though, even that will bring about its own challenges, as high-quality, accessible data is often unavailable or scattered across multiple institutes.

But first, there will have to be a change in attitude away from the current R&D culture. The nature of AI is to see signals that humans cannot detect, so its value and scope will be limited for scientists that insist on understanding exactly how predictions are derived, instead of aiming to validate them.

native1_300x100
lorem ipsum