Bristol Myers Squibb CEO Visits Stevens With a Message: Tech Is Supercharging Drug Development

Artificial intelligence and other technologies are not just the future of the traditionally cautious biopharmaceutical industry, they are saving lives now.

And students entering the workforce with technical skills will play a critical role in the effort.

That was the message Bristol Myers Squibb (BMS) CEO Giovanni Caforio brought to an appreciative Stevens audience gathered in the university’s Tech Flex Auditorium October 4, speaking in the latest installment of the President’s Distinguished Lecture Series. (A full recorded video of the lecture appears at the bottom of this story.)

“With a background as a physician,” said Stevens President Nariman Farvardin, introducing the prestigious speaker to the assembled guests, “Dr. Caforio has helped strengthen the company’s patient-focused culture.”

“Fasten your seat belts.”

Global AI partnerships, speeding the process

“I have tremendous respect for Stevens,” began Caforio, noting the large number of Stevens graduates who have joined BMS, as he introduced the company’s mission in direct terms.

“Our mission — and our purpose — is to find better medicines and to get those right medicines to the right patients as fast as possible. It’s as simple as that.”

Medical knowledge has certainly leapt ahead during the past several decades. Lung cancer, Caforio noted, is now understood to actually represent a constellation of nine separate diseases. Immuno-oncology, once scoffed at, is now a respected field as scientists race to discover new ways of leveraging the immune system to battle dangerous cancers. New cellular therapies are also revolutionizing disease treatment, as well.

But as the pace of medical discovery and knowledge accelerates exponentially, he continued, the speed of developing and rolling out new medicines to patients has not kept up.

It costs approximately $5 billion to bring a new medicine to market, Caforio pointed out; the eventual success rate of promising Phase I drugs is only 12% — and it still takes nearly a decade to get successful medicines from conception to approval.

How to improve those numbers?

“The answer is: technology,” he said. “We can best match the speed of medicine delivery to the pace of innovation through what I call digital computational medicine.”

Technology, he went on, allows biopharma companies to bypass the traditional ‘try and fail’ model by more quickly understanding the biology of diseases — then matching potentially more effective drugs with those diseases much more quickly.

But, Caforio added, “we won’t get there unless we fundamentally disrupt the industry.”

He then took the audience through BMS’ process of creating new medicines in some detail, breaking that process down into three phases: drug discovery (identifying a target and synthesizing a molecule to attack that target), drug development (testing the most promising candidate medicines in live trials) and final delivery to patients.

With regards to the lengthy initial discovery phase, “machine learning and AI enable a breakthrough” that can shortcut the laborious trial and error currently required.

Instead of optimizing for one characteristic of a prospective new medicine at a time, AI can enable all characteristics of a molecule to be optimized simultaneously, speeding the process of finding the right molecule for the task greatly.

BMS is now partnering with German tech company Immatics, Caforio explained, to narrow the universe of possibilities down more quickly.

“This new technology promises to cut years from the time required to get to a new medicine,” he emphasized.

AI can also help optimize — and make more equitable — human clinical trials, which are critical to developing a final and effective product.

Drug studies have traditionally been conducted largely on Americans, living in large cities near major teaching hospitals, working with subjects who are not necessarily particularly diverse.

BMS is working in partnership with the French AI firm Owkin to change those biases, using technology to mine large datasets and understand where patients affected with specific diseases are located.

“We learned a lot during COVID” when clinical trials worldwide paused, noted Caforio, including enhanced methods of collecting and analyzing electronic health records for new insights.

Fully 58% of BMS’ current clinical sites are now located in areas that have been traditionally undeserved by trials, he pointed out — and the company’s effort to diversify trials will continue.

Regarding the final phase of the process, the delivery of medications to the patients who most need them, Caforio offered the example of the BMS-developed drug Camzyos. The recently approved medicine is used to treat patients with a dangerous heart disorder known as obstructive HCM that often mimics other health conditions.

“HCM is difficult to diagnose,” he said, noting it typically takes two years and six to seven incorrect diagnoses by doctors before the condition is correctly spotted and treated.

“How do we actually find patients with HCM and help them?”

Once again, he explained, the answer is AI. BMS is now working closely with the Israeli company Viz.ai, which uses AI to flag biomarkers that can identify likely HCM patients.

Caforio concluded his lecture by making an appeal for students in AI, computational and data fields to consider working in his industry.

“We need new talents in biopharma: scientists that are as adept with programming as they are with pipettes,” he said. “We need technologists that are excited by the computing challenge and the reward.”

“And we must give people who could work at Google or Meta or Amazon good reasons to work in biopharma.”

Big bets, risk-reward; a Presidential Medal

During a Q&A session following the talk, Caforio took a number of detailed questions on topics including the cost of drug development, the prioritization of challenges and the need for earlier detection and diagnosis.

He addressed the risk and reward of seeking therapies for expensive, high-risk problems such as medicines for rare or difficult-to-treat diseases — once again using the example of immuno-oncology as a success story that defied the odds and is now transforming lives.

BMS, Caforio said, determined to back early research in the immuno-oncology field even though it had internally assessed there was only a slim chance that treatments derived from the work would ever prove viable and effective.

“We reached a conclusion that there was a 5% probability it would work, but if it worked, it would be transformational.”

“We had a 95% probability of being wrong. But in that case, it worked.”

Cellular therapies represent a similar high priority (and high-risk bet) for BMS now, he continued.

“Just five years ago, nobody believed in these, and now everybody is seeing the results,” he explained, noting that the company must task hundreds of workers with modifying and improving the cells of just 25 patients at a time (per FDA regulations), at tremendous expense.

Now the company is seeing significant response from lymphoma patients in advanced stages, who had little hope of surviving their cancer, after receiving the therapy. That has already made the bet on cellular therapies worthwhile.

Caforio finished his time on the podium by again referencing the abundant career opportunities at BMS for graduating computer science and other students at Stevens.

President Farvardin then announced a new Stevens scholarship named for Caforio — the inaugural scholarship will be granted to an incoming first-year student in fall of 2024 — and presenting Caforio with the eighth President’s Medal for his many contributions to science and medicine.

For more on the President’s Distinguished Lecture Series, visit this website.