Artificial Intelligence (AI) has been defined as the use of intelligent machines to work and react like humans. The technology is familiar to most of us through everyday things like voice and face recognition devices. Much has also been written about its potential to lower costs and improve outcomes in healthcare with it already being used in areas such as administration, diagnostics and surgery. However, for those of us who work in pharma, perhaps the most interesting question is whether it can help us better meet the needs of our customers?
Currently, it is estimated that less than 12% of all drug development programmes succeed, whilst the cost of developing a new drug averages $2.6 billion and takes at least 10 years*. There is therefore huge scope for improved efficiency with all the ensuing benefits for healthcare systems and their patients. AI is seen as an enabler because of its ability to process mountains of data, find patterns and test conclusions. Consequently, many new companies have arisen that work in pharma drug discovery. Two eye-catching examples are:
• Berg who identify possible treatments on the basis of a much better insight into the biological causes of disease. One ongoing application is the development of a drug for pancreatic cancer based on understanding derived from the first model of how the disease functions and what allows it to grow
• BenevolentBio who have used data to develop a platform with a billion known and inferred relationships between genes, symptoms, diseases, proteins, tissues, species, candidate drugs and more. This was applied to finding treatments for motor neurone disease and highlighted around 100 existing compounds with the possible potential to do so
Optimism abounds about the contribution AI can make. In the future some believe that it is likely that drug development and assessment in clinical trials will be driven entirely by highly sophisticated pattern recognition. One result could be the pinpointing of previously unknown causes of disease, thereby accelerating the move towards personalised medicine. Ultimately, AI may even bring about a complete understanding of human biology and the means to fully address human disease!
Despite this optimism, a significant percentage of scientists remain unaware of the capabilities of AI, even though wet lab skills may be made obsolete within a decade. There is also scepticism. Some compare the hype to that which surrounded computer aided drug design, a development which has failed to halt the decline in R&D productivity since the mid-90s.
From a pharma point of view, the key question is whether AI will facilitate the development of better drugs at more affordable prices. Until they are launched, we won’t know the answer but the good news is that we won’t need to wait long to find out!
*The challenge of developing new treatments and cures August 26th2015, www. phrma.org the answer