HEALTH
For decades, two people with stage III colon cancer have walked out of surgery and into the same chemotherapy clinic, whatever the genetic fingerprint of their tumours. On 9 October 2026, that changed. The US Food and Drug Administration approved Roche's immunotherapy Tecentriq (atezolizumab), in combination with chemotherapy, as adjuvant treatment, the treatment given after surgery to mop up any remaining cancer, for stage III colon cancer with deficient mismatch repair, known as dMMR.
It is the first approved adjuvant immunotherapy regimen for this setting. It is also Tecentriq's twelfth US indication. The approval covers both the intravenous form and Tecentriq Hybreza, a version combined with hyaluronidase that is given as an injection under the skin, along with a fluoropyrimidine and oxaliplatin chemotherapy backbone.
Behind the regulatory language is a simple, striking result: adding the immunotherapy cut the risk of cancer coming back, or of death, by half.
The trial that changed the standard of care
The approval rests on ATOMIC, a phase III randomised, open-label, multicentre trial, also known as Alliance A021502 and registered as NCT02912559. Results were published in The New England Journal of Medicine in 2026 (volume 394, pages 1155 to 1166).
ATOMIC enrolled 712 patients, randomly split in a 1 to 1 ratio. One group received the standard chemotherapy regimen mFOLFOX6 for 12 cycles, roughly six months. The other group received the same chemotherapy plus Tecentriq for 12 cycles, then Tecentriq on its own for a further 13 cycles, around another six months. The main goal, or primary endpoint, was disease-free survival, meaning the length of time patients lived without the cancer returning.
The headline numbers are hard to ignore. Adding Tecentriq reduced the risk of recurrence or death by 50 per cent. At 36 months, 86 per cent of patients on the combination were alive and free of disease, compared with 76 per cent of those on chemotherapy alone. That ten-percentage-point gap, in a disease where relapse is the great fear, is the kind of margin oncologists describe as practice-changing. The safety profile was consistent with previous studies of Tecentriq and mFOLFOX6, meaning no new unexpected side effects emerged.
The trial was sponsored by the US National Cancer Institute and led by the Alliance for Clinical Trials in Oncology, with Genentech, a member of the Roche Group, supporting it under a cooperative research agreement. Dr Frank A. Sinicrope of the Mayo Clinic led the study globally, and German participation was coordinated by AIO-Studien-gGmbH under Professor Anke Reinacher-Schick.
Why dMMR tumours are different
Our cells constantly copy their DNA, and typos creep in. A repair system, the mismatch repair machinery, normally proofreads and fixes them. In some tumours that system is broken, which is what dMMR means. Such tumours accumulate huge numbers of mutations, which makes them look strange to the immune system, and, crucially, makes them vulnerable to drugs that release the brakes on immune cells.
Tecentriq is a checkpoint inhibitor. It blocks a protein called PD-L1, which many cancers use as a kind of invisibility cloak against the body's T cells. Take away the cloak, and the immune system can recognise and attack the tumour. In dMMR cancers, with their abundant mutations, the response can be dramatic.
Roughly 15 per cent of colon cancer patients have dMMR or microsatellite-instability-high tumours, according to Roche. Patients in ATOMIC were selected using an immunohistochemistry test, such as the VENTANA MMR RxDx Panel, which checks tumour tissue for the missing repair proteins. That means a simple lab test on tissue already removed at surgery now determines which treatment path a patient is offered.
The scale of the problem
Colon cancer is one of the most common cancers on the planet. Over one million people are diagnosed with it worldwide every year, according to Roche. Nearly a quarter of cases are found at stage III, when the cancer has spread to nearby lymph nodes but not to distant organs. Surgery plus chemotherapy aims to cure those patients, yet nearly 30 per cent of people with stage III disease see their cancer return within five years despite treatment.
Every one of those recurrences means more treatment, more fear and often a much poorer prognosis. A treatment that meaningfully cuts relapse, even for a subset of patients, therefore matters enormously, and for the roughly one in seven patients whose tumours are dMMR, the ATOMIC data show a clear route to doing exactly that.
What the experts and the company are saying
Dr Levi Garraway, Roche's chief medical officer, said the approval provides a new adjuvant treatment option for people with stage III dMMR colon cancer who face a high risk of disease recurrence.
Dr Sinicrope put the shift in historical context: until now, he said, the adjuvant standard of care for stage III colon cancer has not differed based on mismatch repair status. In other words, for the first time doctors will choose post-surgery therapy by the molecular profile of the tumour in the way they long have in breast and lung cancer.
Roche says it is pursuing further regulatory filings, including with the European Medicines Agency, so patients outside the United States will have to wait for their own regulators to rule.
Questions patients will be asking
Does everyone with colon cancer qualify? No. The approval is specific to stage III disease with dMMR tumours. The great majority of patients, those with proficient mismatch repair, are not covered by this decision, and ATOMIC does not show that they would benefit.
How do I know if my tumour is dMMR? The test is done on tumour tissue removed at surgery. If you or a relative have had stage III colon cancer, ask your oncology team whether mismatch repair or microsatellite instability testing was done and what the result was.
What about side effects? Immunotherapy can cause the immune system to attack healthy tissue, leading to inflammation of organs such as the bowel, skin, liver or thyroid. The ATOMIC safety profile matched what was known from earlier studies, but the treatment adds roughly six months of drug exposure beyond chemotherapy, and doctors will need to monitor patients closely.
Will it be covered? Insurance coverage, cost and timing will vary by country and plan, and have not been detailed in the approval announcement. Patients should ask their care team and insurer.
A decade in the making
ATOMIC did not appear overnight. The trial is registered as NCT02912559, a number that dates back to 2016, which means it took roughly a decade of patient recruitment and follow-up to reach the point where doctors could say with confidence that the benefit was real. That is the unglamorous reality of adjuvant cancer trials: because the question is whether cancer stays away, researchers must wait years to count the relapses that did not happen.
The collaboration behind it is also notable. A publicly funded cancer institute, an academic cooperative group, a major hospital and a pharmaceutical company all pulled in the same direction, and German investigators joined the effort. Public sponsorship matters because it tends to produce trials that answer questions of practical importance to patients, such as whether a drug already used in other cancers helps at an earlier stage, rather than only questions that serve commercial strategy.
The approval also strengthens a wider trend in oncology: treating tumours by what drives them rather than by where they began. Mismatch repair status has already shaped treatment choices in advanced disease. Moving that logic to the curative, post-surgery setting, where the stakes are highest because a cure is still possible, is a significant step, and it will likely encourage more trials of immunotherapy earlier in the course of other cancers.
Doctors will still have questions. ATOMIC was an open-label study, meaning patients and doctors knew which treatment was being given, and longer follow-up will show how durable the benefit is beyond three years. But a 50 per cent reduction in the risk of recurrence or death is not a marginal signal, and it was strong enough to convince regulators.
What this means for you
If you have been diagnosed with stage III colon cancer, ask whether your tumour has had mismatch repair testing. That one question could decide whether immunotherapy belongs in your plan. If you are a surviving patient who is still within the treatment window, talk to your oncologist rather than assuming the new option applies; timing after surgery matters, and this decision should be individual.
If you are simply worried about colon cancer, the best protection remains screening. Many national guidelines now recommend starting routine screening at 45, and people with a family history, persistent changes in bowel habit, blood in the stool, unexplained weight loss or ongoing abdominal pain should see a doctor rather than wait. Catching the disease before it reaches the lymph nodes avoids the need for this kind of treatment altogether.
People with a strong family history of colon or related cancers should also ask about genetic counselling, because dMMR can sometimes be inherited, as in Lynch syndrome. Tumour testing can be the first clue.
The bottom line
Cancer drug approvals can feel like a steady drumbeat, but some are genuine turning points. The ATOMIC result and the FDA decision on 9 October give patients with one genetically defined slice of stage III colon cancer something they have never had: a post-surgery treatment shown to halve the risk of relapse or death compared with chemotherapy alone, with 36-month disease-free survival rising from 76 to 86 per cent.
The challenge now is practical. Every eligible patient needs to be tested, every oncology team needs to know the new standard, and health systems around the world need to follow the US lead. For the thousands of people who will hear the words stage III colon cancer this year, the right test at the right time could now change the ending of the story.
