FDA Approves First Immunotherapy for Rectal Cancer

JemperliWhat to know

30-second read

For Oncology · General Surgery · Gastroenterology

Who

Adults with previously untreated, stage II/III rectal cancer confirmed dMMR or MSI-H by an FDA-authorized test — about 5–10% of rectal cancers.

What’s new

First immunotherapy approved for curative-intent, non-metastatic rectal cancer; every earlier dMMR checkpoint-inhibitor approval targeted metastatic disease.

Evidence

Single-arm AZUR-1 (n=154): 47% sustained complete response at 12 months (54/116 evaluable); no recurrences among responders beyond 2 years’ median follow-up.

Big advantage

A complete responder now has a real question: can total mesorectal excision — and a possible permanent colostomy — be deferred for close surveillance instead?

Key consideration

FDA approved the drug, not surgery omission. About half of treated patients won’t reach complete response, and omitting surgery remains unproven outside trial settings.

Availability

Approved 19 October 2026, four months early via the Commissioner’s National Priority Voucher; dosed 500 mg IV every 3 weeks for up to 9 cycles.

Practice impactConsider in a multidisciplinary center

A genuine option for a narrow, biomarker-defined group, managed by a team equipped for intensive post-treatment surveillance. Most locally advanced rectal cancer patients are MMR-proficient and still need standard chemoradiation and surgery.

Read the full physician analysis

Full analysis · 10 min readGSK accelerated-approval announcement (19 Oct 2026) · AZUR-1 trial · FDA accelerated-approval pathway

FDA granted accelerated approval to GSK’s Jemperli (dostarlimab-gxly) on 19 October 2026 for adults with previously untreated, stage II or III mismatch repair-deficient (dMMR) or microsatellite instability-high (MSI-H) locally advanced rectal cancer — the first immunotherapy ever approved in a curative-intent, non-metastatic rectal cancer population. Regulatory News covered the approval itself and its unusual regulatory path on the day it happened. This is the clinical layer underneath that: who actually qualifies, what the trial did and didn’t show, and the surgical question this approval raises without answering.

A narrower approval than the headline suggests

Checkpoint inhibitors have treated dMMR and MSI-H colorectal cancer since 2017, but always in the metastatic setting, where surgery is no longer curative and a drug’s job is to control disease. Stage II and III locally advanced rectal cancer is different: these are, by definition, patients being treated with curative intent, historically through chemoradiation followed by total mesorectal excision (TME) — an operation that can leave a permanent colostomy and carries real rates of bowel, bladder and sexual dysfunction. Dostarlimab is now approved to treat that population directly, which is the substantive clinical story here, independent of how quickly the application moved through FDA.

The approval is also narrow by biomarker. Mismatch repair deficiency or microsatellite instability-high status shows up in an estimated 5% to 10% of rectal cancers — most locally advanced rectal cancer is MMR-proficient and gets no benefit from this drug. The label requires dMMR or MSI-H status confirmed by an FDA-authorized test for solid tumors before starting treatment, which makes universal biomarker testing at diagnosis, not just in metastatic disease, the first practical change this approval creates. A patient who isn’t tested at diagnosis is a patient who can’t be identified for this option later without going back for it.

What AZUR-1 actually measured

The approval rests on AZUR-1, a single-arm, open-label Phase 2 trial that enrolled 154 previously untreated patients and dosed dostarlimab 500 mg intravenously every three weeks for up to nine cycles — about six months of treatment, with no chemotherapy, radiation or surgery given alongside it. The primary endpoint was sustained clinical complete response at 12 months (cCR12): GSK reported 47% among 116 evaluable patients (95% CI: 37% to 56%), with 38 participants still inside their 12-month follow-up window at the time of reporting. Among responders, GSK said no disease recurrences had been reported at a median follow-up beyond two years. This outlet found a higher figure, 68%, circulating in some secondary trade coverage of AZUR-1; absent a clear explanation of what population produced that number, GSK’s own primary-source figure of 47% is the one used throughout this story.

On safety, GSK reported serious adverse reactions in 20% of patients, with fatigue (32%), rash (22%), diarrhea (18%) and pruritus (17%) the most common adverse reactions overall. That profile is consistent with dostarlimab’s existing label in endometrial and other dMMR/MSI-H cancers — immune-mediated reactions that can be severe or fatal, monitored with baseline and periodic liver, renal and thyroid testing, managed with corticosteroids and treatment interruption or discontinuation by severity. There is nothing rectal-cancer-specific about the toxicity signal; the new information here is who tolerates roughly six months of single-agent immunotherapy before any decision about surgery is made, not a new safety category.

The decision AZUR-1 didn’t make for surgeons

FDA’s approval authorizes dostarlimab as a treatment. It does not authorize omitting surgery, and the agency’s own accelerated-approval structure underlines why: cCR12 is a surrogate endpoint, and GSK must verify clinical benefit — durable disease-free survival, not just a complete-response rate — in confirmatory trials or the indication can be withdrawn. Whether a given complete responder should proceed to TME anyway, or be observed instead, is a clinical judgment made by an oncologist, a colorectal surgeon and the patient together, informed by evidence that is still accumulating outside what AZUR-1 itself was designed to answer. The trial measured response; it was not built to prove organ preservation is safe.

That judgment call has a precedent, though not from immunotherapy. The OPRA trial, a chemoradiation-based (not immunotherapy-based) watch-and-wait study, found that roughly half of patients who reached a clinical complete response after chemoradiation avoided surgery entirely over several years of surveillance, with a smaller but real share of those patients eventually needing salvage surgery after local regrowth. AZUR-1’s own extension is studying the analogous question for dostarlimab alone: among patients reaching complete response, how many can avoid TME and a permanent colostomy through structured surveillance, and how many need salvage treatment later. That data is not yet mature. A physician reading this approval as license to skip surgery in every complete responder is reading past what FDA actually authorized.

Three ways to curative-intent rectal cancer, in sequence

	Standard chemoradiation + TME	OPRA (chemoradiation, watch-and-wait)	AZUR-1 (dostarlimab alone)
Eligible population	All stage II/III rectal cancer, any MMR status	Stage II/III rectal cancer reaching complete response after chemoradiation	dMMR/MSI-H stage II/III rectal cancer only — 5–10% of cases
Treatment given	Chemoradiation, then total mesorectal excision	Chemoradiation, then active surveillance if complete response	Dostarlimab alone, 500 mg IV every 3 weeks, up to 9 cycles
Surgery avoidance	Not applicable — surgery is the treatment	About half of complete responders avoided surgery over several years in OPRA	Under separate study; not an outcome AZUR-1’s approval itself measured
Maturity of evidence	Decades of outcomes data	Multi-year, published randomized-design cohort data	Single-arm, 12-month primary endpoint; confirmatory trial still required

OPRA figures from the published trial literature on chemoradiation-based nonoperative management, not from AZUR-1; the two strategies are not interchangeable and this table is not a treatment recommendation.

What surveillance for a complete responder actually requires

Nonoperative management, wherever it is pursued, is not a lighter-touch alternative to surgery — it is a different, intensive commitment. Programs built around the OPRA-style watch-and-wait model combine digital rectal exam, endoscopy and pelvic MRI on a schedule that starts every few months and continues for years, specifically because local regrowth after a clinical complete response is detectable and, caught early, often still salvageable with surgery. A practice considering this pathway for a dostarlimab complete responder needs the same infrastructure: a radiologist experienced in reading post-treatment rectal MRI, an endoscopist available on a tight schedule, and a patient who understands that surveillance is the price of avoiding surgery, not a reason to skip follow-up visits. None of that is new to colorectal oncology, but a practice that has never run a nonoperative pathway for chemoradiation patients should not assume it can improvise one for immunotherapy patients without the same infrastructure in place first.

• Test for dMMR/MSI-H at diagnosis, not just when metastatic disease is suspected — the biomarker has to be known before a patient can be offered this option at all.
• Route eligible patients to a multidisciplinary discussion before treatment starts, not after a complete response is already apparent on imaging.
• Set expectations early that roughly half of patients will not reach complete response, and that reaching one does not automatically mean surgery is off the table.
• Build or confirm surveillance capacity — MRI, endoscopy and exam on a tight cadence — before offering nonoperative management as a real option, not a theoretical one.
• Document informed consent around the maturity of the evidence: a 12-month surrogate endpoint from a single-arm trial, with a confirmatory trial still outstanding.

Access and what is still unsettled

Dostarlimab is already a marketed drug with established billing and distribution channels from its endometrial and other dMMR/MSI-H indications, so the access questions here are narrower than for a brand-new molecule: existing J-code and infusion-billing pathways should extend to the rectal cancer indication, and payers who already cover Jemperli for other dMMR/MSI-H cancers have less reason to treat this as a novel coverage decision. What remains genuinely unsettled is the clinical pathway, not the drug’s reimbursement: GSK is scheduled to present the full AZUR-1 interim analysis at the ESMO Congress in Madrid, 23 to 27 October 2026, and the confirmatory trial data that accelerated approval requires has not yet read out. A tumor board weighing this option today is weighing it on the evidence available now, not on evidence that is still a trial or two away.

From a 2022 case series to an approved indication

1. 2022

   First published cohort

   A small NEJM case series reports complete responses in every evaluable dMMR rectal cancer patient treated with dostarlimab alone, drawing wide attention to the approach.

2. 2024–2025

   AZUR-1 enrolls and reads out

   The registrational single-arm trial enrolls 154 patients; GSK reports the cCR12 primary endpoint met in 2026.

3. 24 Aug 2026

   Priority review accepted

   FDA accepts the supplemental filing with a PDUFA date in February 2027.

4. 19 Oct 2026

   Accelerated approval

   FDA clears dostarlimab for stage II/III dMMR/MSI-H locally advanced rectal cancer, four months early via the Commissioner’s National Priority Voucher.

5. 23–27 Oct 2026

   Full data at ESMO

   GSK presents the complete AZUR-1 interim analysis in Madrid.

Frequently asked questions

Does this approval mean a patient can skip surgery?

Not by itself. FDA approved the drug for treating dMMR/MSI-H locally advanced rectal cancer, not a nonoperative strategy. Whether a complete responder forgoes total mesorectal excision and enters surveillance instead is a separate clinical decision, made case by case in a multidisciplinary setting, with evidence still maturing.

How common is dMMR or MSI-H rectal cancer?

Roughly 5% to 10% of rectal cancers carry mismatch repair deficiency or microsatellite instability-high status, so most locally advanced rectal cancer patients are not eligible. Universal MMR/MSI testing at diagnosis is the only way to find the ones who are.

What happens to a patient who doesn't reach a complete response?

AZUR-1 enrolled 154 patients and reported a 47% sustained complete response rate at 12 months among 116 evaluable patients. Patients who don't reach a complete response return to the standard discussion of chemoradiation and surgery; the trial did not test dostarlimab as a replacement for that pathway in non-responders.