Surgeons at Brigham and Women’s Hospital in Boston have injected a genetically engineered virus directly into the brain tumors of 41 patients with recurrent glioblastoma, the most lethal form of brain cancer. The therapy, called CAN-3110, is a modified version of herpes simplex virus type 1 that is designed to invade cancer cells, replicate inside them until they burst, and then recruit the patient’s own immune system to hunt down whatever tumor remains. Results from the first-in-human trial, published in Nature, suggest the approach is safe and that the immune response it triggers may be linked to how long patients survive.
For a disease that kills most people within 15 to 18 months of diagnosis, and for which recurrence is nearly inevitable, even a credible signal of benefit is significant. But what makes CAN-3110 especially interesting is a finding that upends a long-standing worry in the field: patients who already carried antibodies against herpes, the vast majority of adults, appeared to mount a stronger anti-tumor immune response after injection, not a weaker one.
How CAN-3110 attacks a tumor on two fronts
CAN-3110, formally designated rQNestin34.5v.2, belongs to a class of treatments called oncolytic viruses. The concept is straightforward: engineer a virus so that it can replicate freely inside cancer cells but is locked out of healthy tissue. When the virus copies itself enough times, the cancer cell ruptures. That destruction scatters tumor-specific proteins into the surrounding brain tissue, effectively waving a flag for the immune system.
This two-pronged attack sets oncolytic virotherapy apart from chemotherapy and radiation, which tend to suppress immune function. CAN-3110 is meant to do the opposite: kill tumor cells directly and then convert the debris into a kind of personalized vaccine that teaches T cells to recognize the cancer.
Because glioblastoma sits behind the blood-brain barrier, a network of tightly packed cells that blocks most drugs from reaching the brain, CAN-3110 cannot simply be infused through an IV. Instead, neurosurgeons inject it straight into the tumor during an operation, using real-time imaging to guide the needle. The Phase 1 trial, registered as NCT03152318, tested escalating doses delivered as a single stereotactic injection at the time of surgery for recurrent disease.
What the Phase 1 trial found
The trial, led by neuro-oncologist E. Antonio Chiocca at Harvard Medical School, enrolled 41 adults whose glioblastoma or high-grade glioma had returned after standard treatment. Its primary goals were safety, dose-finding, and early signs of immune activation.
On the safety front, the data were reassuring. No dose-limiting toxicities halted the study, and there was no signal of catastrophic brain inflammation or uncontrolled viral spread. The registry record confirms that investigators tracked neurological complications, systemic side effects, and viral shedding at each dose level.
The more provocative finding involved immunity. Patients who received CAN-3110 showed measurable oncolytic immunoactivation, and the strength of that immune response correlated with survival outcomes. In a 41-person, uncontrolled study, correlation is not proof of cause and effect. But it provides biological plausibility: the virus appears to do more than just pop open cancer cells. It seems to wake up an immune system that glioblastoma has spent months suppressing.
The herpes paradox: why prior infection may actually help
Roughly two-thirds of the global population carries HSV-1, and the percentage is even higher among older adults. For years, researchers developing herpes-based cancer therapies worried that pre-existing antibodies would neutralize the therapeutic virus before it could reach the tumor. CAN-3110’s early data suggest the opposite may be true.
Patients who were seropositive for HSV-1, meaning their blood already contained antibodies from a past infection, showed signs of stronger immune activation after the injection. If that finding holds up in larger studies, it would flip a core assumption in the oncolytic virus field and could mean that the majority of glioblastoma patients are, paradoxically, better candidates for this therapy because of an infection they picked up decades ago.
The hypothesis is biologically plausible. A primed immune system may recognize fragments of the engineered virus faster, arrive at the tumor site sooner, and encounter tumor antigens released by oncolysis while already in an activated state. But plausible is not proven. Confirming the effect would require a larger trial that stratifies patients by HSV-1 antibody status from the start and compares outcomes between groups.
Building on earlier herpes-virus trials
CAN-3110 is not the first engineered herpes virus tested against glioblastoma. A related agent called G47-delta, developed in Japan, underwent a Phase 2 trial in patients with residual or recurrent glioblastoma. That study, published in 2022, showed that repeated stereotactic injections were feasible and tolerable, and it identified long-term survivors among treated patients. In 2021, Japanese regulators granted G47-delta conditional approval for malignant glioma, making it the world’s first oncolytic virus approved for brain cancer.
The two viruses share a platform (both are engineered from HSV-1 and both are injected directly into tumors) but carry different genetic modifications and were tested in different patient populations. No head-to-head comparison exists, so it is impossible to say whether one outperforms the other. What the G47-delta experience does provide is proof of concept: herpes-based oncolytic virotherapy can reach the clinic, clear regulatory review, and produce durable responses in at least some patients.
Other oncolytic platforms are also in clinical testing for glioblastoma, including adenovirus-based agents and a modified poliovirus (PVSRIPO) that generated attention after early trials at Duke University. The field is small but growing, and CAN-3110 enters a competitive landscape where multiple viral strategies are vying to prove they can crack one of oncology’s hardest problems.
What the data cannot tell us yet
Phase 1 trials are built to answer a narrow question: is this therapy safe enough to keep testing? They are not designed to prove that a treatment extends life, and their small size makes them vulnerable to statistical noise. The survival signals in the CAN-3110 study are genuinely encouraging, but 41 patients without a control group cannot establish a survival advantage. Patients in early-phase oncology trials are often younger, fitter, and more motivated than the broader population, which can skew results in ways that disappear in larger studies.
Several specific uncertainties stand out:
- Durability. A single injection may trigger a burst of immune activity that fades within weeks. Whether one dose is enough for lasting tumor control, or whether patients will need repeated injections, remains unknown.
- Combination therapy. Checkpoint inhibitors like pembrolizumab have transformed treatment for some cancers but have largely failed in glioblastoma on their own. Combining CAN-3110 with a checkpoint inhibitor is a logical next step, but no data from such a combination exist yet.
- Long-term safety. The Phase 1 data cover a limited follow-up window. Delayed inflammation, viral reactivation outside the tumor, or unexpected interactions with future treatments have not been fully characterized. Years of monitoring, not months, will be needed.
- Patient selection. If HSV-1 serostatus truly predicts response, it could become a biomarker for selecting patients most likely to benefit. But acting on that hypothesis before it is validated in a controlled trial would be premature.
Where CAN-3110 goes from here
As of mid-2026, the Phase 1 results have been published and the scientific community is digesting them. The next logical step is a Phase 2 trial with a larger patient cohort, a control arm, and pre-specified stratification by immune markers including HSV-1 serostatus. Whether that trial is already in planning has not been publicly confirmed, but the strength of the Phase 1 signal and the growing institutional interest in oncolytic virotherapy make continued development likely.
For patients and families facing a glioblastoma diagnosis today, CAN-3110 is not yet an available treatment. It remains investigational, accessible only through clinical trials. But the data so far represent something that has been rare in glioblastoma research: a mechanistically coherent therapy that appears safe, activates the immune system in a measurable way, and produces survival signals worth chasing in a rigorous, larger study. That is not a cure. It is a reason to keep going.
More from Morning Overview
*This article was researched with the help of AI, with human editors creating the final content.
