The Discovery That Took 15 Years to See
A statistical pattern emerged from a cancer trial that nobody expected: patients with follicular lymphoma stopped relapsing. Between year 5 and year 15, the relapse rate dropped from 6.8% to 0.6%. After 15.5 years of tracking 531 patients, researchers at Fred Hutch Cancer Center, SWOG, and the University of Rochester Medical Center published findings in JAMA Oncology showing a 42% cure rate for a disease that medical textbooks still call incurable.
The word "cure" has never been used for follicular lymphoma, a slow-growing blood cancer that accounts for up to 30% of all non-Hodgkin lymphoma cases. About 15,000 people receive this diagnosis annually in the United States and Europe. Standard practice has been to tell patients they'll likely need multiple rounds of treatment over their lifetime because the cancer almost always comes back.
But the S0016 trial, which opened in 2001 and published its primary results in 2013, kept following patients long after the official study ended. What became visible in that extended timeline challenges how we measure success in cancer treatment.
What Cure Looks Like in the Data
The trial enrolled patients with untreated advanced-stage follicular lymphoma and gave them chemoimmunotherapy: either CHOP chemotherapy combined with rituximab (R-CHOP) or CHOP with radioimmunotherapy (CHOP-RIT). Both arms targeted CD20, a protein on the surface of the cancer cells.
At 15 years, approximately 70% of patients remained alive. The two treatment approaches showed no statistically significant difference in overall survival rates. What caught researchers' attention was the shape of the relapse curve over time.
Michael LeBlanc, a biostatistician at Fred Hutch, used cure modeling to analyze the data. This statistical method incorporates background mortality rates to estimate what fraction of patients can be considered cured rather than simply in remission. The analysis required more than a decade of follow-up because you need to watch long enough to distinguish between a very long remission and an actual cure.
The 42% cure rate represents a threshold that only became measurable because researchers chose to keep tracking patients years after the trial's primary endpoint.
The Timeline Gap
FDA drug approvals typically rely on 2 to 5 years of clinical trial data. The S0016 trial published its main findings at year 12. The cure analysis required year 15.
This creates a mismatch between regulatory timelines and biological reality. Mazyar Shadman, who directs the Bezos Family Immunotherapy Clinic at Fred Hutch and served as first author on the JAMA Oncology paper, and Jonathan Friedberg, who directs the Wilmot Cancer Institute at University of Rochester Medical Center and was senior author, led a study that accidentally proved something transformative because they kept watching.
The 531 patients lived in the space between "incurable" and "possibly cured" for over a decade. Nearly 90% of follicular lymphoma patients survive at least five years past diagnosis, but historical patterns showed most would relapse eventually. The S0016 data suggests that for a substantial fraction, "eventually" never comes.
Nothing in the trial design required this extended follow-up. The infrastructure for long-term tracking exists because individual research networks like SWOG maintain it, not because the regulatory system demands proof of cure.
The Next Wave Measured in Months
While the S0016 analysis was revealing cure rates over 15 years, a new class of treatments entered clinical trials. Bispecific antibodies target both CD3 on T cells and CD20 on lymphoma cells, essentially recruiting the immune system to attack the cancer. Early trials show response rates around 80%.
Dana-Farber investigators are leading phase 2 trials combining these bispecific antibodies with monoclonal antibodies. One trial pairs rituximab with epcoritamab; another combines obinutuzumab with glofitamab. The treatments show particular promise for patients whose disease relapsed within 24 months of initial treatment, historically the hardest cases to manage.
But success is being measured in months and early-year response rates. The bispecific antibodies can trigger cytokine release syndrome, requiring careful dose escalation over the first three weeks. Researchers are optimizing for safety and initial efficacy because that's what the approval timeline rewards.
Whether these treatments cure anyone won't be known until 2040.
What the System Selects For
The S0016 follow-up happened during a period when long-term research infrastructure was relatively stable. FDA recently moved toward accepting single trials for approval rather than requiring two confirmatory studies. NIH narrowed its clinical trial registry requirements. These shifts prioritize speed.
Speed has value. Patients with aggressive disease need treatments now, not in 15 years. But the S0016 results expose what gets lost in acceleration: the ability to answer whether we're actually curing people or just extending the interval between relapses.
Follicular lymphoma is slow-growing, which made the long follow-up feasible. Patients stayed in contact with research networks. The cancer's pace meant researchers could track outcomes without losing people to rapid disease progression. The biology cooperated with the timeline required to see cure.
For faster-moving cancers, or in a regulatory environment that doesn't fund decade-long observation, the question of cure becomes unanswerable by design.
The Uncomfortable Math
The 42% cure rate means 58% of patients weren't cured. Some died from their disease. Others from unrelated causes after living cancer-free for years. The cure modeling accounts for background mortality, but it can't change the fact that more than half the patients needed the word "incurable" to remain accurate.
For newly diagnosed patients today, the S0016 results change the conversation in the oncologist's office, but not cleanly. The standard treatment remains chemoimmunotherapy targeting CD20. The data now suggests that for some patients, one round of treatment will be the only one they ever need. But there's no way to know which patients at diagnosis.
The trial enrolled people with untreated advanced-stage disease between 2001 and 2008. Treatment protocols have evolved. Rituximab is now standard. Chemotherapy regimens have been refined to reduce side effects, particularly for older patients who struggle with toxicity. Whether current approaches produce the same cure rates won't be confirmed until the 2040s.
What Requires 15 Years
The S0016 extended follow-up wasn't mandated by FDA, funded by a specific grant for long-term observation, or required by the trial's original protocol. It happened because researchers at Fred Hutch, SWOG, and Rochester maintained contact with patients and kept analyzing the data.
That continuation depended on institutional stability, sustained funding for cancer research networks, and individual scientists choosing to prioritize a question that wouldn't yield results for over a decade. The current regulatory environment is optimizing for different incentives.
The most important question in cancer treatment takes 15 years to answer. The system is increasingly structured around timelines measured in months.