A cancer drug's defining feature is usually what it binds to. Gotistobart's is also when it lets go. The antibody is designed to release its target inside a cell, allowing an important immune-control protein to return to the surface. That is an unusually precise idea to test against a devastating disease.
On September 14, BioNTech and OncoC4 reported updated results in metastatic squamous non-small cell lung cancer that had progressed after immunotherapy and chemotherapy. Among 87 randomized patients, median overall survival was 18.5 months with gotistobart and 10.0 months with the chemotherapy docetaxel. The difference between those medians is 8.5 months. These are findings from the exploratory, non-pivotal first stage of PRESERVE-003, a phase 3 trial whose confirmatory stage is still ongoing. The companies presented the update at the World Conference on Lung Cancer.
The result earns attention because it connects an intricate engineering choice to an outcome people can understand: time. It also presents a useful test of how we judge medical progress. An attractive molecular explanation and an encouraging clinical signal belong together. They still have different jobs. The explanation tells us why an approach might work; the trial must establish how reliably it does.
The body needs the checkpoint
Immune checkpoints help prevent defensive cells from damaging healthy tissue. Cancer can exploit those restraints. Checkpoint inhibitors interfere with the signals that suppress an immune attack, giving T cells a better opportunity to kill tumor cells. CTLA-4 is one such checkpoint; PD-1 and its partner PD-L1 are others. This strategy has become an established part of cancer treatment. The National Cancer Institute explains the underlying biology here.
The same biology creates a difficult engineering problem. Healthy organs also depend on immune restraint. Checkpoint treatment can cause inflammation in the colon, liver, lungs and other tissues. A more forceful immune response is therefore an incomplete definition of a better cancer drug. Where that response occurs, and what remains protected, matter enormously.
Gotistobart's developers are trying to exploit that distinction. Their candidate targets CTLA-4 on regulatory T cells, which can suppress immune activity around a tumor. After the antibody and receptor move inside a cell, changing acidity prompts the antibody to detach. CTLA-4 can then recycle to the surface instead of being destroyed. The intended effect is stronger depletion of tumor-associated regulatory T cells while preserving immune restraint in peripheral tissues. This is the developer's account of the design, rather than proof that healthy organs are protected in every patient.
Timing becomes part of the medicine
A 2025 PNAS study shows why the timing deserves scrutiny. Researchers engineered variants of the established CTLA-4 antibody ipilimumab and tested them in cells and mice. One variant detached at about pH 6, characteristic of an early intracellular compartment called an endosome. It produced greater tumor regression and less severe immune toxicity in the mouse experiments. Another variant, which needed more acidic conditions to release its target, failed to show the same improvement. Read the original experiments.
This was research on engineered antibody variants, not a human comparison between gotistobart and ipilimumab. Its contribution is a testable principle: an antibody's route through a cell can matter as much as its initial attachment. Releasing a receptor after it has already entered a destructive compartment may be too late. The chemical setting of the release becomes part of the drug's behavior.
That is what makes this story more interesting than a contest between two survival numbers. Drug design is reaching into the sequence of events after binding. A successful medicine might need to recognize a target, accompany it for part of its journey, and then disengage. The ambition is control at a finer scale. Whether that control translates into a worthwhile benefit remains a clinical question.
What the survival result means
The latest comparison assigned 45 patients to gotistobart and 42 to docetaxel. The data cutoff was July 17, with median follow-up of 25.4 months. The reported hazard ratio for death was 0.56, with a nominal p-value of 0.0295. The announcement provides no updated confidence interval for that estimate.
A median describes the point at which the estimated survival curve reaches 50%. It does not tell each patient how long they will live. Subtracting the two medians describes a difference between groups; it does not mean every treated person gained that amount of time. NCI's definition of median survival.
A hazard ratio answers a different question, comparing the rate of an event between groups over time. Here, 0.56 corresponds to an estimated 44% lower hazard of death. It is neither a cure rate nor a promise of 44% longer life. Reading it alongside the survival curve and uncertainty interval is more informative than treating it as a standalone score. NCI's definition of hazard ratio.
The earlier analysis, published in Nature Medicine on March 27, reported a hazard ratio of 0.46, with a 95% confidence interval of 0.25–0.84. Gotistobart's median survival had not yet been reached. The September update adds follow-up to that cohort; it is not independent replication. The peer-reviewed paper and methods.
A phase 3 label needs its fine print
PRESERVE-003 was designed in two stages. The first explored dose selection and preliminary activity; the second tests the selected regimen against docetaxel in squamous disease. Its principal outcome is overall survival. OncoC4 described that structure in its September 2025 trial-design announcement. The distinction is crucial: the phase number names the overall program, while the stage determines what this particular result can establish.
The study excluded specified actionable genetic alterations. The program initially included squamous and nonsquamous cancers and a lower-dose arm. Poor outcomes led to termination of the lower-dose cohort after ten patients and a pause in nonsquamous development. Docetaxel patients could not cross over to gotistobart. The first-stage analysis was exploratory, and the p-values were nominal rather than formal hypothesis tests.
Those decisions are part of learning which patients and dose to take forward. They also make confirmation especially valuable. A result observed while a program is being refined needs to survive a trial designed around the resulting hypothesis. Watching the same participants for longer makes an estimate more mature. It does not create another group of patients in whom the finding has held up.
Precision still carries a cost
The September announcement reports grade 3 or worse treatment-related adverse events in 20 of 45 gotistobart recipients, or 44.4%. The candidate remains investigational.
The March paper reported grade 3 or worse immune-related events in 33.3% of gotistobart patients, versus 4.9% with docetaxel. Adverse events prompted discontinuation in 22.2% versus 4.9%. It reported no fatal treatment-related events. Median progression-free survival was 2.4 versus 2.6 months.
An antibody with a sophisticated mechanism can still impose substantial harm. The survival finding deserves to be evaluated alongside that burden, and the mixed pattern across endpoints deserves explanation. It would be premature to turn a molecular design objective into a claim of clinically proven safety superiority. The relevant question is whether the full balance becomes meaningfully better for the patients being studied.
The next result that matters
Overall survival is a strong endpoint precisely because it counts deaths from any cause. FDA guidance emphasizes randomized comparisons: historical survival figures can differ because of patient selection, imaging or supportive care. Survival analysis still needs adequate follow-up and attention to subsequent treatment. Tumor response and progression measures provide additional information, but they cannot automatically stand in for living longer. FDA's guidance on cancer-trial endpoints.
For the pivotal result, the useful questions are concrete. Does the benefit persist in the confirmatory population? How wide is the uncertainty around its size? What do later treatments and missing follow-up do to the interpretation? How many people discontinue treatment, require intensive management of immune toxicity, or experience a sustained improvement in daily life? A convincing answer needs more than a favorable headline number.
Gotistobart now has a reason to be watched closely. Its broader promise is the possibility of designing immune medicines around what a cell preserves, recycles and destroys. If that approach survives the next clinical test, the advance will be measured in patients' outcomes. The molecular ingenuity will have earned its place there.
