The Billionaire Who Keeps Promising to Cure Pancreatic Cancer
STAT News drops a major investigation into billionaire Patrick Soon-Shiong's decade of bold pancreatic cancer claims. The data behind his "breakthrough" therapies tell a very different story than his public promises.
A Pattern That Keeps Repeating
Imagine someone told you they could fix your house. They showed up with a toolbox, gave a great speech, took your money, and then… the roof still leaks. Now imagine they did this for a decade, with different houses, and kept getting invited back.
That's the uncomfortable picture painted by a sweeping STAT News investigation published this week. The target: billionaire Patrick Soon-Shiong and his long history of bold claims about conquering pancreatic cancer. The finding: a wide gap between what he says his drugs can do and what they've actually done.
Pancreatic cancer is one of the cruelest diagnoses in medicine. Only about 13% of patients survive five years. For those with metastatic disease (cancer that has spread), that number drops to roughly 3%. Patients are desperate. Families are desperate. And into that desperation, STAT argues, Soon-Shiong has repeatedly walked with promises his data can't support.
The Moonshot That Missed
Soon-Shiong isn't a random guy with a lab coat and a dream. He invented Abraxane, a blockbuster cancer drug used in breast, lung, and pancreatic cancer. He's worth billions. He owns the Los Angeles Times. When he talks, people listen.
The problem, according to STAT, is what happens after they listen.
Back in the mid-2010s, Soon-Shiong launched "Cancer Breakthroughs 2020," an initiative that aimed to help cure cancer by the end of the decade. STAT's earlier reporting found the program produced "very little scientific progress" and functioned mostly as a marketing vehicle for his GPS Cancer diagnostic test. The initiative claimed partnerships with companies like Pfizer and Merck KGaA that those firms couldn't verify. It even took credit for clinical trials that started before the program existed.
The targeted enrollment goal was 20,000 patients in vaccine-based immunotherapy trials. By 2020, the actual number was closer to 5,000. The central promise of an effective, broadly applicable cancer vaccine? Not achieved.
STAT's new investigation argues that pancreatic cancer is now the latest stage for this same playbook: bold narratives, sparse data, and financial structures that funnel attention (and resources) back toward his own companies.
Get tomorrow's biotech intelligence before your competitors.
Join thousands of biotech professionals who start their day with our free, daily briefing.
What the Pancreatic Cancer Data Actually Show
Soon-Shiong's main pancreatic cancer program runs through ImmunityBio, the company he founded and controls. The flagship effort is a multi-drug cocktail called the "Nant Cancer Vaccine," which combines chemotherapy, an immune-boosting molecule called N-803 (branded Anktiva), and engineered natural killer cells designed to hunt tumors.
The company's marquee claim: this combination "more than doubles median overall survival" in patients who've already failed multiple rounds of treatment.
Sounds incredible, right? The reality is more complicated.
The key data come from QUILT-88, a phase 2 trial with approximately 80-86 heavily pretreated patients. In the third-line group (patients who'd already tried two other treatments), median survival was 6.3 months, compared to a historical benchmark of roughly 3 months. For the broader group, it was 5.8 months.
Numerically, yes, that's roughly double. But there's a catch: there was no control group. The comparison is to old data from different trials, with different patients, different supportive care, and different selection criteria. That's like claiming you're a faster runner than your friend because you timed yourself on a downhill track and compared it to their uphill time from three years ago.
Most patients' tumors still grew within about 2.4 months of starting treatment. And median survival of six months, while slightly better than the historical bar, is still devastating.
One Patient Does Not Make a Cure
ImmunityBio has also spotlighted individual success stories. One patient with recurrent metastatic pancreatic cancer has been in remission for over six years. A five-patient case series showed survival ranging from about 5 months to nearly 27 months.
These stories are genuinely remarkable. They suggest something biologically interesting is happening. But a single long-term survivor is not proof that a treatment works for everyone. In oncology, anecdotes are the appetizer; randomized trials are the main course. And that main course has never arrived for the Nant pancreatic cancer program.
As of 2026, there is no FDA approval for any Nant-branded pancreatic cancer therapy. One key trial, QUILT-3.088, was actually withdrawn before it even started. The program remains entirely investigational.
The FDA Steps In
The credibility concerns go beyond data interpretation. In March 2026, the FDA sent ImmunityBio a blistering warning letter about how it was promoting Anktiva. On a podcast, Soon-Shiong had called Anktiva "the most important molecule that could cure cancer" and claimed it "actually can treat all cancers."
The FDA's response was blunt: those statements were "false" and "misleading." The agency said the promotional materials created a misleading impression of efficacy and failed to mention key side effects like urinary tract infections, pain, chills, and fever.
For a company whose founder positions himself as a revolutionary cancer fighter, getting publicly rebuked by the FDA for overclaiming is a significant blow.
Meanwhile, Real Progress Looks Different
The STAT investigation carries extra weight because of what's happening elsewhere in pancreatic cancer research. At ASCO 2026, Revolution Medicines presented phase 3 data for daraxonrasib, a pill targeting KRAS mutations (found in over 90% of pancreatic cancers). In a randomized trial of 500 patients, the drug nearly doubled median survival: 13.2 months versus 6.7 months on standard chemo.
The contrast is striking. Daraxonrasib's data came from a large, controlled study with a clear comparator. ASCO's chief medical officer reportedly called it a "grand slam." That's what rigorous, practice-changing evidence looks like.
Soon-Shiong's pancreatic cancer data, by comparison, come from small, uncontrolled studies that have never been tested head-to-head against standard treatments.
Follow the Money
STAT's reporting also raises uncomfortable questions about where the money flows. Previous investigations found that a $12 million donation Soon-Shiong made to the University of Utah was structured so that $10 million went right back to his company NantHealth for sequencing work. The arrangement also gave NantHealth access to patient data and, according to STAT, allowed the company to inflate reported order volumes for its GPS Cancer test by roughly one-third to investors.
Independent tax and nonprofit experts reviewed the structure and called it likely "indirect self-dealing." Soon-Shiong has called that characterization "maliciously false."
But the pattern is what matters here. When charitable giving, clinical research, and corporate revenue all feed into the same ecosystem controlled by one person, it becomes very hard for patients and investors to separate genuine scientific progress from self-promotion.
The Bottom Line for Patients
Pancreatic cancer patients deserve honesty. They deserve to know the difference between a six-month median survival in a small, uncontrolled study and a transformative cure. They deserve leaders who present their data with appropriate caveats, not as marketing ammunition.
The science behind ImmunityBio's approach might eventually prove meaningful. Rare complete responses are real, and the biology is worth studying. But studying is different from selling, and the gap between Soon-Shiong's rhetoric and his results has been wide for over a decade.
As one oncology observer previously told STAT about his earlier cancer initiatives: the findings were "the most minuscule and vague," and the field had heard "virtual radio silence" in terms of peer-reviewed results.
When you're talking to desperate patients about the deadliest common cancer in America, radio silence isn't good enough. And neither are promises that keep arriving without proof.
5 min read
The Lung Cancer Drug Where Half the Patients Are Still Winning After 7 Years
Seven years into a clinical trial, more than half of metastatic lung cancer patients on Pfizer's lorlatinib still haven't progressed. Oncologists are starting to whisper the word "cure," and the data might actually justify it.
