AstraZeneca’s Bob Li says just 7% of cancer patients access clinical trials, and explains how AI could change that.
Clinical trials are widely considered the best available care for cancer patients. Yet in the U.S., only 7% of patients ever get access to one.
That statistic sits at the center of a conversation with Bob Li, AstraZeneca’s Senior Vice President and Global Head of Medical Affairs, Oncology. Li spent more than 20 years in clinical practice, including over a decade at Memorial Sloan Kettering Cancer Center (MSK) in New York, before moving into industry. He argues that access, not scientific discovery, is now the biggest barrier standing between cancer patients and a cure.
A Career Shaped by Patients Who Couldn’t Get In
Li’s conviction comes from firsthand experience. As a physician-scientist treating patients across Australia, the U.S., and through international outreach in Asia and Latin America, he repeatedly watched patients struggle just to reach the care that was already sitting on the shelf.
One story stands out: a patient in rural Minnesota with a HER2 mutation and metastatic lung cancer who couldn’t get an appointment at the Mayo Clinic. Li stepped in with phone calls to colleagues and a remote telemedicine consultation from MSK, during the early days of the COVID-19 pandemic when telemedicine was just becoming normalized.
“Without that remote access, the patient likely would still be in the dark,” Li says. He came to realize that the well-informed, well-connected patients he saw regularly at MSK were actually the minority. Most patients, in the U.S. and internationally, face the same struggle to reach the right care.
Why the Bottleneck Isn’t Money
It would be easy to assume clinical trials stay scarce because of funding. Li says that’s not the real problem.
According to a paper he co-authored in Cancer Discovery, the industry spent more than $80 billion on clinical trial drug development in 2022, a figure he estimates has likely surpassed $100 billion today. That spending already outpaces the research budgets of the National Institutes of Health and the National Cancer Institute combined.
The real bottleneck, Li says, is the drug development process itself: long, complex, and expensive to navigate, regardless of how much money is behind it. Bringing a new medicine to market through the traditional path still takes ten years or more.
Li points to a concrete example from his own work developing KRAS inhibitors during the pandemic, when international flights were grounded. Through coordination with Ministries of Health, his team shipped medicine directly to a patient in the Middle East rather than requiring the patient to travel to the U.S. “Those are proof-of-concept experiences that taught me these tools are available today,” he says.
AI Is Already Cutting Research Time from Years to a Single Day
One of the clearest changes Li points to is how AI has compressed the time it takes to turn raw patient data into usable research findings.
Early in his career, Li manually pooled data from electronic medical records, a process that could take one to two years to produce results. His team later co-authored a paper in Nature (with Justin Jee as first author) demonstrating that natural language processing could automate that same pooling process.
“Today, you can ask a question and use AI to generate forest plots and Kaplan-Meier curves within 30 minutes,” Li says. “You ask a question in the morning, and you have a draft analysis by afternoon.” He’s careful to note that a human still needs to review and quality-check every result, but the underlying pace of research has shifted from incremental improvements every few years to daily progress.
The MARS Mission: Four Pillars to Scale Access
AstraZeneca’s effort to close the access gap is organized around what Li calls the MARS mission, short for Medical Affairs Revolution Spirit. It rests on four pillars.
The first is next-generation evidence generation: decentralized clinical trials that move beyond major academic cancer centers, with modernized inclusion criteria and digitally assessed trial quality, rather than the older model of physically shipping imaging discs between sites.
The second and third pillars focus on community education and what Li calls “Building Cancer Networks,” connecting academic institutions with local community practices. He traces the current comprehensive cancer center model back to President Nixon’s 1971 National Cancer Act. That model has succeeded, Li says, but it was only ever built to serve a minority of patients.
The fourth pillar is international acceleration: getting countries to align on shared drug development standards instead of operating in separate regulatory silos.
Where Li Draws the Line Between Hope and Hype
Asked directly where AI’s real impact ends and hype begins, Li doesn’t dismiss the skepticism. AI-assisted early cancer detection, he says, is genuinely promising but not yet the standard of care. Emerging models can help predict which nodules found on a chest X-ray are likely to become malignant months later, but that capability is still being validated, not yet deployed at scale.
AstraZeneca’s MILTON initiative, a partnership with the UK’s National Health Service published in Nature Genetics, is one proof point Li cites for multimodal disease prediction that could eventually extend beyond cancer to conditions like COPD and tuberculosis. “There is always hype around new technology, so you have to distinguish hope from hype,” he says, “but with scientific rigor, we can isolate the real clinical value.”
A Shared Mission, Not a Company Slogan
Li is candid that some former academic colleagues still ask him what it’s like on “the dark side” of industry. He rejects the framing entirely. In his view, pharmaceutical companies bring international infrastructure, data networks, and scaling capacity that individual academic institutions and cooperative research groups simply don’t have on their own.
His closing point is a claim about timeline, not certainty: if decentralized trials, AI-assisted research, and international collaboration scale together, Li believes cancer could stop being a leading cause of death within his own lifetime. It’s an optimistic, forward-looking claim from someone with a direct stake in the outcome, not a settled scientific conclusion.