Cancer gone in every patient. No surgery. No chemo.

Researchers at Memorial Sloan Kettering gave patients with a subtype of rectal cancer one drug for six months; in every one, the tumor disappeared with no surgery, radiation or chemotherapy.

A tumour that the immune system can see

In 2015, Dung Le, Luis Diaz and colleagues at Johns Hopkins published a phase 2 study in the New England Journal of Medicine built on a single hypothesis: tumours that carry very large numbers of mutations, because their DNA mismatch repair machinery is broken, should present enough abnormal proteins to be recognised by T cells once the brake on those cells is released. They gave the PD-1 antibody pembrolizumab to 41 patients with progressive metastatic cancer. Among patients with mismatch repair-deficient colorectal cancer, 4 of 10 responded. Among those with mismatch repair-proficient colorectal cancer, 0 of 18 did. Sequencing found a mean of 1,782 somatic mutations per tumour in the deficient group and 73 in the proficient group. Mismatch repair status predicted who would benefit, and a biomarker rather than an organ became the way to choose patients for the drug.

What rectal cancer treatment cost

Locally advanced rectal cancer is normally treated with chemotherapy, radiation and surgical removal of the rectum. Andrea Cercek and colleagues at Memorial Sloan Kettering Cancer Center, writing in 2022, noted that this sequence produces a pathological complete response in up to a quarter of patients and causes bowel, urinary and sexual dysfunction, infertility and altered quality of life in a substantial proportion. Memorial Sloan Kettering puts the mismatch repair-deficient share of rectal cancer at 5 to 10 percent, and these tumours respond poorly to standard chemotherapy. The question was whether the 2015 result in metastatic disease could be moved to the front of treatment, before surgery, in a tumour that was still confined to the pelvis.

The design

The Memorial Sloan Kettering trial enrolled adults with mismatch repair-deficient stage II or III rectal adenocarcinoma. Each received dostarlimab, a PD-1 blocking antibody, at 500 milligrams intravenously every three weeks for six months, nine cycles in all. The protocol then called for standard chemoradiotherapy and total mesorectal excision, with one exception: a patient with a clinical complete response after the antibody alone would skip both and be followed without surgery. A clinical complete response required no residual disease on digital and endoscopic rectal examination and none on rectal magnetic resonance imaging, with assessments at 6 weeks, 3 months and 6 months and every 4 months thereafter, supplemented by PET and CT. The statistical design assumed that an overall response rate below 25 percent would be a failure.

Twelve of twelve

At the time of the June 2022 report, 16 patients had been enrolled and treated, and 12 had completed the six months and at least six months of follow-up. The median age was 54, with a range of 26 to 78; 62 percent were women; 15 of 16 had stage III disease. All 12 who completed treatment had a clinical complete response: 100 percent, with a 95 percent confidence interval of 74 to 100. Tumours were undetectable on MRI, on PET, on endoscopy, on digital examination and on biopsy. No patient had received chemoradiotherapy or undergone surgery. No patient had progressed or recurred during a median follow-up of 12 months, with a range of 6 to 25. No adverse event of grade 3 or higher had occurred.

The speed was as notable as the completeness. Biopsies showed no viable tumour in most patients as early as six weeks into treatment. PET uptake fell to background in every patient by three months. Germline sequencing found pathogenic alterations associated with Lynch syndrome in 8 of 14 patients tested, though none had a known family history of it. The authors wrote that longer follow-up was needed to assess the duration of response, and the durability end point, a sustained clinical complete response at 12 months, had at that stage been reached by 4 patients.

Forty-two, then forty-nine

The count grew without a failure. By June 2024, when updated data were presented, Memorial Sloan Kettering reported that 42 patients had completed treatment and all 42 had no evidence of disease, with the earliest participants cancer-free for up to four years. On December 16, 2024, the U.S. Food and Drug Administration granted dostarlimab Breakthrough Therapy designation for mismatch repair-deficient locally advanced rectal cancer.

The full accounting appeared in the New England Journal of Medicine in June 2025, now covering two cohorts and 117 patients. In cohort 1, rectal cancer, 49 patients had completed treatment and all 49 had a clinical complete response and chose nonoperative management; 37 had sustained it for at least 12 months, meeting the trial's second co-primary end point, and two-year recurrence-free survival was 96 percent. Cohort 2 extended the same six months of dostarlimab to mismatch repair-deficient tumours of the oesophagus, stomach, colon, liver and bile ducts, urinary tract and gynaecologic organs. There, 35 of 54 patients who completed treatment had a clinical complete response and 33 chose to forgo surgery. Across both cohorts, 84 of 103 who completed treatment had a complete response and 82 avoided surgery; two-year recurrence-free survival was 92 percent; 5 patients recurred, 4 of them only in lymph nodes. Ninety-five percent of patients had either no adverse events or reversible grade 1 or 2 events. In no patient was the option of curative surgery lost during the long course of treatment.

Why one drug was enough

Dostarlimab does not act on the tumour. It binds PD-1 on T cells and prevents the tumour from switching them off. The 2015 and 2022 papers together explain why that suffices in this subtype and not in others: a mismatch repair-deficient tumour carries on the order of 1,800 mutations, each a potential foreign flag, so an immune system freed of the PD-1 brake has an abundance of targets. In the 2022 cohort, PD-L1 and CD8 T cells rose in tumour and surrounding tissue within six weeks of the first dose, and immune aggregates in the rectal wall reached 6 to 10 times baseline after treatment ended. The tumour was removed by the patient's own cells.

What the story does and does not show

The 2022 paper showed a 100 percent clinical complete response in 12 consecutive patients with mismatch repair-deficient locally advanced rectal cancer after six months of a single PD-1 antibody, with no surgery, radiation or chemotherapy and no severe adverse events. The 2025 paper showed the same in 49 of 49 rectal patients, with 96 percent recurrence-free at two years, and a lower but substantial response in other organs.

The subtype is 5 to 10 percent of rectal cancer; the result does not apply to the other 90 percent, whose tumours lack the mutational load that makes PD-1 blockade work. The study was single-arm and single-centre, and its authors state that a larger study is needed to confirm long-term benefit, particularly for the non-rectal cohort, whose median follow-up for recurrence was 14 months. Clinical complete response is an imaging and endoscopic judgement, not a pathological one, because no tissue was removed; 5 recurrences did occur. The slide text's statement that a 100 percent complete response had never before been reported in oncology is not a claim made in either paper and is not repeated here. What the papers show is that, for one molecularly defined group, the standard of care changed on the strength of a trial that began with twelve people.

Forty-two patients. Forty-two tumors gone.