Big Pharma has long touted the game-changing potential of artificial intelligence to improve drug development. But what does it mean in practice?
According to Sanofi’s Matt Truppo, there is a clear ambition: “Can we cut the timeline from portfolio entry to delivering a molecule of interest to the clinic by several years? That’s the goal.”
Truppo should know. As the French pharma’s global head of research platforms, he oversees the team “responsible for the invention of both small- and large-molecule drugs across all therapeutic areas of Sanofi, including oncology, rare diseases, neurology and immunology.”
“Ultimately, we’re responsible for designing, engineering and delivering new molecular entities to take into the clinic,” he explains.
“If you’re put off by failure, this is not the industry for you.” — Sanofi’s Matt Truppo
It’s a journey that increasingly involves AI and machine learning. Sanofi has secured a number of intriguing partnerships in this space in recent months, including expanding on its five-year relationship with AI drug discovery shop Exscientia in January to the tune of $100 million.
The aim is to tackle the pharma industry’s “longstanding challenge” with the time and cost of delivering drug candidates into the clinic. “As an example, it routinely takes four to five years from a project entering the portfolio to the delivery of a candidate molecule,” Truppo said in an interview at the Fierce Biotech Next Gen virtual event. “And this is before you ever enter clinical trials.”
“In the case of small molecules, this may mean synthesizing over 5,000 molecules before you arrive at the candidate that you want to move forward,” he adds. “So if you’re put off by failure, this is not the industry for you.”
Trimming the timeline
With a background in bioengineering and chemistry leading to a more than 20-year career at Merck & Co., Johnson & Johnson’s Janssen unit and most recently Sanofi, it’s perhaps no surprise that Truppo is so enthusiastic about the potential for digitizing the drug discovery process.
By applying AI machine learning models, so the thinking goes, pharmas can reduce the timelines and the “sheer number of compounds we need to synthesize in the real world” by doing much of the analysis on a computer.
Sanofi hopes that the collaboration with Exscientia alone will deliver drug candidates across 15 targets in oncology and immunology, Truppo says. “Can we start to predict with very high fidelity the biophysical properties—from how they bind in the target to how they’ll be differentiated in different patient populations?”
The Exscientia deal wasn’t the only time over the past year when Sanofi has opened its checkbook to gain a bigger footprint in the AI space. In November, the pharma committed $270 million in an equity investment and payments over three years to Owkin as part of a collaboration to bring its digital clinical research platform to bear on the drugmaker’s core oncology efforts in four different cancers.
“The way that this was done is really exciting,” says Truppo of the partnership. To accelerate Sanofi’s medical research while preserving the privacy of patients’ medical data, scientists will train their AI and machine learning models locally without putting their data in a centralized repository. The parameters from these local models are then shared to the benefit of everyone taking part in the research, without the data at risk of getting into the wrong hands.
“This is where some of the AI tools, data processing and data correlation are really starting to come home now with how we can leverage these very large data sets, but do it in an ethical way that preserves patient privacy,” Truppo says.
Hitting the accelerator
To transform Sanofi into a digital powerhouse, the company isn’t only looking externally for talent. The drugmaker’s digital accelerator launched in June with a mission to upskill its workforce as well as recruit top talent in digital product management, data science and computer system development.
The accelerator’s first program will be developing an integrated platform and data solution to help European healthcare professionals better engage with atopic dermatitis patients. But Truppo says this is “just the latest step in an ongoing transformation” that will lead to myriad advances such as accelerating the company’s digital pathology—a term for the ability to analyze information generated from digitized specimen slides.
When Truppo was invited by one of his teams to see a live demo of this souped-up process, he was taken aback.
“It’s a remarkable thing to watch firsthand, to give me an idea of the type of applications that can be achieved with AI image analysis,” he recalls. “This is something that for digital pathology used to take weeks, and it can now be done in minutes.”
From the patient perspective, AI doesn’t just offer a way to improve how their data are used while strengthening privacy. It can also improve clinical trial efficiency by reducing the number of patients required as well as “leveraging and using real world evidence and remote data submission to make it easier for patients to enroll in and participate in trials.”
With Sanofi diving headfirst into a more digital future, what are the practical implications for the company’s pipeline? “The goal is to cut [drug development] by a few years,” Truppo says. “That in turn reduces the cost.
“If you can get to a clinical trial and a proof of concept faster—beyond that, what’s really exciting is the idea of getting the right patient the right drug at the right time.”