AI set to speed up glacial pace of drug development

Experts say artificial intelligence (AI) has the potential to one day personalise treatment for individuals in near real time but as its use in medicine grows it is already helping healthcare professionals to make sense of the massive data sets from an individual’s physiology and the effects on individual health outcomes from environment, lifestyle and diet.

Barry Heavey, who leads Accenture’s life sciences practice in Ireland, says AI will ultimately help clinicians make sense of the huge swathes of data — even for a single patient — currently available to them.

“When we think about how AI could be used in the healthcare and life sciences space we have to think about what kind of data that the actors in the healthcare system are dealing with. If you think about a doctor dealing with an elderly patient, they are dealing with the three V’s of data — volume, velocity and variety,” says Heavey.

“For example, if a patient is diagnosed with cancer the doctor will have a high volume of data to process on that patient — the patient may have done many genetic tests, there may have been diagnostics carried out on their blood or their tissue or their cancer, and there may be a high volume of data from medical imaging such as CT scans, MRIs or PET scans.”

He says this data tsunami is also coming at them at speed. “That’s why we talk about velocity — doctors are dealing with data coming at them very quickly.”

Ongoing monitoring means doctors will have to analyse a constant stream of data such as heart activity sensors or blood sugar sensors. This variety means the treating clinician must make a decision based on a multiplicity of data — from images, scans and diagnostics — and consider this in the context of currently available treatment options but also more targeted treatments hitting the market.

“New and highly precise treatments could be suitable based on the analysis of all this data. It will allow doctors to select the optimal, most highly personalised treatment for that individual patient based on all the available data,” Heavey explains. Considering this avalanche of data for even just one patient creates a challenging environment for the doctor to operate in, he says, “He or she may benefit from AI and machine-learning support to enable them to make better decisions faster.

“The doctor ultimately still has the autonomy and the responsibility to make the call on what’s best for the patient but they may benefit from the assistance in analysing all the different types of data to find the precise combination of personalised treatment for that particular patient.”

AI is also being used to speed up the historically glacial pace of drug development. The drug discovery process is, at its core, a costly and highly inefficient one. Thousands of potential molecules go through a lengthy and laborious sequence of testing and analysis, only for the best candidates to often fail at the final hurdles of efficacy and safety. AI is now being used to streamline the more cumbersome elements of R&D but experts warn it may take years before AI is fully integrated and even longer before it will significantly improve the odds of drug discovery. Right now, pharmaceutical companies — and patients — are still beginning to benefit from AI as it is increasingly employed to streamline production processes and enhance complex logistic chains.

AI will undoubtedly be an invaluable tool as its use grows in the preclinical phases of drug development, but also right through to the clinical phases and regulatory approval, as well as production and distribution, says Matt Moran, director of BioPharmaChem Ireland.

“Pharmaceutical companies are dealing with huge amounts of data, from combinatorial chemistry right through to checking a biological process to see if it works. They are trying to simulate it in the lab but this spits out tons of data and there is a very definite application for AI there,” he explains. “This has always been a difficult nut for the pharma industry to crack — they produce so much data but what to do with it has been the problem.” He sees AI becoming an integral part of its processes within the next two to five years.

Personalised medicine is a particular area where the science is rapidly evolving, Moran says. For example, CAR-T cell therapy is a new, customised treatment for certain cancers whereby the patient’s cells are collected, treated and delivered back to them. This process is highly regulated and time sensitive and the supply chain is of critical importance. Moran says AI will play a crucial role in enabling the safe and efficient delivery of this therapy.