FDA grants accelerated approval to GSK dostarlimab as first immunotherapy for rectal cancer

For decades, the standard course of treatment for locally advanced rectal cancer has run through a fixed sequence: chemotherapy, then radiation, then surgery to remove part of the rectum. The regimen helps many patients, but it also leaves a lasting mark, with changes to bowel, urinary and sexual function, risks of secondary cancers and the possibility of infertility. A small subset of tumors carries a biomarker that has now opened a different route entirely, one that in most trial patients required none of those three interventions.

Those tumors are described as mismatch repair deficient (dMMR) or microsatellite instability-high (MSI-H). Their DNA repair defects leave them carrying an unusually heavy load of mutations, which makes them visible to the immune system. Dostarlimab, which GSK sells as Jemperli, is a PD-1 inhibitor that lifts a brake on T cells so they can attack the tumor.

On October 19, 2026, the US Food and Drug Administration granted accelerated approval to Jemperli (dostarlimab-gxly) as a single agent for adults with previously untreated stage II or stage III locally advanced rectal cancer that is dMMR or MSI-H, as confirmed by an FDA-authorized test. The agency framed the announcement around what the drug may let patients avoid: chemotherapy, radiation and surgery. It is the first immunotherapy approved specifically for this biomarker-defined rectal cancer population.

Accelerated approval is a conditional pathway. It allows the FDA to clear a drug on the basis of an intermediate or surrogate endpoint when the disease is serious and the need is unmet, on the understanding that confirmatory data will follow. For patients and clinicians, that distinction matters as much as the headline.

Key Facts

The evidence comes from AZUR-1 (NCT05723562), a multicenter, single-arm, open-label phase 2 study in 154 patients with previously untreated, histologically confirmed MSI-H or dMMR stage II or III locally advanced rectal cancer. Everyone enrolled received dostarlimab at 500 mg intravenously every three weeks for nine cycles, roughly six months of therapy. Tumor response was checked by endoscopy, rectal MRI and CT scans at week 6, week 12 and about three weeks after the last dose. Patients who reached a clinical complete response moved into non-operative management, with surveillance every four months for two years and every six months after that.

The disposition numbers carry much of the story. The ASCO Post reported on October 19 that 110 of 154 patients (71%) received dostarlimab alone, needed no additional treatment and did not progress. Twenty-six patients (17%) went on to surgery, 13 (8%) received additional systemic therapy or radiation, one (0.6%) had progressive disease, and four (2.5%) withdrew or were lost to follow-up.

Response rates depend on who measures them, and the agency published both versions. Investigator-assessed sustained clinical complete response at 12 months, known as cCR12, was 68% (98 of 144 evaluable patients, 95% CI 60 to 76), with an objective response rate of 82% (95% CI 75 to 88). Independent central review put cCR12 at 47% (54 of 116 patients, 95% CI 27 to 56) and the objective response rate at 79% (95% CI 71 to 85). Another 38 patients (33%) had a clinical complete response with less than 12 months of follow-up, 44 (38%) had a near-complete response, and of 15 patients who had surgery after dostarlimab alone, 12 had a pathological complete response. Medscape reported on October 19 that no disease recurrences were seen among responders after a median follow-up of more than two years.

Safety data in the label set the boundaries of the trade. Serious adverse reactions occurred in 20% of patients. The most common reactions of any grade were fatigue (32%), rash (22%), diarrhea (18%) and pruritus (17%), and 3% of patients discontinued permanently. The label carries warnings for immune-mediated adverse reactions, infusion-related reactions, complications of stem cell transplant and embryo-fetal toxicity, noting that immune-mediated events can be severe or fatal.

The FDA said on October 19 that the application had Priority Review and Breakthrough Therapy designation, that the review finished nearly four months ahead of its goal date, and that the action was the 11th approval under the Commissioner's National Priority Voucher pilot. The review ran under Project Orbis alongside Health Canada, Israel's Ministry of Health, Swissmedic and the UK MHRA. On prevalence, the agency cited published estimates that roughly 3% of locally advanced rectal cancers are dMMR or MSI-H. GSK announced on October 19 that the biomarker is found in around 5% to 10% of rectal cancers.

Analysis

The first thing to notice is that this approval rests on a single-arm trial with no randomized control group. Every patient in AZUR-1 received dostarlimab, so the 71% figure describes what happened to people who took the drug, not a head to head comparison against chemoradiation followed by surgery. In a disease where the standard of care is long established, that is a real evidentiary gap, and the accelerated pathway is the mechanism the FDA uses to bridge it later.

The second thing is the measurement problem. A 68% cCR12 by investigator assessment and a 47% cCR12 by independent central review describe the same trial, yet they differ by 21 percentage points. Central review is the more conservative and, for regulatory purposes, the more defensible number. Some of the spread is timing: 38 patients had a clinical complete response but had been followed for less than 12 months, and responses can deepen over time. Still, the gap is a reminder that organ preservation depends entirely on the quality of the endoscopy, MRI and biopsy that decide whether a tumor is really gone.

What this really means is that the endpoint of this approval is not survival but the avoidance of three treatments and the organ damage they cause. That is a legitimate and patient-centered goal, and the surveillance protocol attached to it is the price of admission. The earlier Memorial Sloan Kettering Cancer Center study (NCT04165772) in 50 patients went further: no patient received chemoradiation or rectal surgery, cCR12 was 93% and the objective response rate was 100%, and among 27 patients followed at least three years, one had a local recurrence treated with lymphadenectomy. That single recurrence is both reassuring and cautionary.

The bigger picture here is that the drug is not the bottleneck; testing and surveillance are. A biomarker present in roughly 3% of locally advanced rectal cancers, or 5% to 10% of rectal cancers by GSK's count, can only direct therapy if tumors are tested at diagnosis. Every eligible patient who is never tested receives the old sequence by default, and every patient who chooses non-operative management depends on follow-up that many centers are not set up to deliver.

Why It Matters

For patients, the practical shift is substantial. Rectal cancer surgery often means a temporary or permanent ostomy and a measurable change in quality of life, and radiation brings its own late effects. A pathway in which most trial patients finished with six months of intravenous therapy and no operation is a genuine alternative for a defined group, not a marginal tweak to an existing regimen.

For the health system, the approval creates a new dependency on non-operative management. Patients who decline surgery are not finished with care; they enter years of surveillance, and the credibility of the whole approach rests on catching the minority whose disease returns. That requires endoscopists, radiologists and oncologists working to a shared protocol, and that capacity is unevenly distributed.

For GSK, the decision adds a third approved use to a drug already cleared in endometrial cancer and in dMMR recurrent or advanced solid tumors, and it arrives with a first and only claim in this setting. It also arrives with conditions attached, since continued approval may depend on confirmatory trial results.

Next Up

Full interim AZUR-1 data are due to be presented at the ESMO Congress in October 2026, which should give the oncology community a clearer look at durability, at the central review numbers as they mature, and at what happens to the 38 patients whose responses were still short of the 12 month mark.

The supplement is also under Project Orbis review in other countries, so decisions from Health Canada, Swissmedic, the UK MHRA and Israel's Ministry of Health are the next regulatory milestones to watch. The confirmatory program is the longer test: accelerated approval is a promise to keep gathering evidence, and the recurrence-free follow-up of these 154 patients will decide whether avoiding surgery becomes a standard option or remains an experiment.