Thirty-two people walked into a UK hospital with bowel cancer that had not yet spread beyond the gut. They received three infusions of an immunotherapy drug over nine weeks, then had surgery. Nearly three years later, not one of them has relapsed. Zero out of 32.
The trial behind those numbers, called NEOPRISM-CRC, is small and still awaiting full peer-reviewed publication. But its results, updated in April 2026 by University College London, have landed squarely in a field already buzzing with evidence that immunotherapy can do remarkable things to a specific type of colorectal tumor. The question the trial now forces is whether a brief course of a single drug, given before surgery, could eventually replace the grueling months of chemotherapy that many of these patients currently face.
The patients and the treatment
Every participant in NEOPRISM-CRC had stage II or III colon or rectal cancer with a molecular feature known as mismatch-repair deficiency with high microsatellite instability (dMMR/MSI-H). In plain terms, their tumors had a broken DNA-repair system, which caused mutations to pile up at an unusually high rate. That flood of mutations makes the cancer cells look foreign to the immune system, almost like a neon sign saying “attack here.” Roughly 10 to 15 percent of early-stage colorectal cancers carry this profile, according to estimates in NCCN clinical guidelines.
The drug used was pembrolizumab, a PD-1 checkpoint inhibitor already approved for multiple cancer types. It works by releasing a brake on T cells, allowing them to recognize and kill tumor cells they would otherwise ignore. Each patient received up to three doses before going to the operating room, a strategy oncologists call neoadjuvant therapy.
The trial, formally registered as NCT05197322, was designed with an unusual layer of precision. Investigators used the FoundationOne CDx companion diagnostic to measure each tumor’s mutation burden, the number of mutations per megabase of DNA. That data was used to stratify patients into subgroups, with the goal of identifying who would respond most powerfully to a short immunotherapy course.
What the results show
After completing pembrolizumab and undergoing surgery, 59 percent of the 32 patients had no detectable cancer remaining, according to the UCL update. The remaining patients still had residual disease in their surgical specimens but had their tumors fully removed. At a median follow-up of 33 months, no patient in either group had experienced a recurrence.
For context, stage III colorectal cancer treated with surgery and adjuvant chemotherapy still recurs in roughly 20 to 30 percent of patients within three to five years, depending on substage. Even among dMMR/MSI-H tumors, which tend to carry a somewhat better prognosis in early-stage disease, recurrence is not rare. A zero-relapse rate at nearly three years, even in a small cohort, is an attention-grabbing number.
Kaikeen Shiu, a consultant oncologist at University College London Hospitals and a lead investigator on the trial, has described the findings as evidence that treatment duration can be safely shortened while still achieving durable disease control. Marnix Jansen, the trial’s translational lead at UCL, has pointed to the biomarker stratification strategy as central to selecting the patients most likely to respond.
The biological logic is straightforward. By giving pembrolizumab while the tumor is still intact, the immune system has a large, mutation-rich target to train against. T cells activated during that window may continue circulating after surgery, hunting down microscopic cancer deposits that might otherwise seed a relapse months or years later. If that mechanism holds, a short pre-surgical course could provide lasting protection without the prolonged chemotherapy that currently follows most operations for stage III disease.
How it fits with other evidence
NEOPRISM-CRC did not emerge from nowhere. The KEYNOTE-177 trial, published in the New England Journal of Medicine, had already established that pembrolizumab outperformed chemotherapy as a first-line treatment for MSI-H metastatic colorectal cancer, where the disease has spread to distant organs. That randomized study proved the drug’s activity against this molecular subtype in advanced settings. NEOPRISM-CRC tested whether the same mechanism could work even earlier, in patients whose cancer was still localized.
Another important reference point is the NICHE-2 trial from the Netherlands, whose results were published in the New England Journal of Medicine in 2024. That study gave patients with dMMR colon cancer a combination of two immunotherapy drugs, nivolumab and ipilimumab, before surgery and reported a pathologic complete response rate of 67 percent with strong relapse-free survival. NEOPRISM-CRC’s 59 percent clearance rate with a single agent over a shorter course invites direct comparison, though differences in patient populations, drug regimens, and follow-up length make head-to-head conclusions premature.
Taken together, these trials are building a case that neoadjuvant immunotherapy can produce deep, lasting responses in dMMR/MSI-H colorectal cancer. The field is now actively asking a bolder question: if the tumor disappears entirely after immunotherapy, can some patients skip surgery altogether? Organ-preservation strategies, sometimes called “watch and wait,” are already being explored in rectal cancer after chemoradiation. Extending that concept to immunotherapy responders is a logical next step, though no completed trial has yet validated it for this population.
What remains uncertain
A 32-patient trial with no control arm cannot, on its own, change clinical practice. The zero-relapse finding is striking, but the sample is small enough that confidence intervals remain wide. Without a comparison group receiving surgery alone or surgery plus chemotherapy, it is impossible to isolate precisely how much of the benefit came from pembrolizumab versus the favorable natural history of dMMR/MSI-H tumors.
The full dataset has not yet been published in a peer-reviewed journal. As of June 2026, the primary source for the 33-month follow-up is an institutional summary from UCL. That means detailed safety tables, surgical complication rates, immune-related adverse event profiles, and patient-level biomarker correlations have not been independently scrutinized. Pembrolizumab is generally better tolerated than chemotherapy, but immune-related side effects, including inflammation of the lungs, liver, or colon, can be serious. Whether the neoadjuvant approach caused surgical delays, wound-healing problems, or other complications is not yet publicly documented.
The 59 percent figure for patients with no detectable disease also needs clarification. The UCL summary does not specify whether this represents a true pathologic complete response, meaning zero viable cancer cells found in the surgical specimen under a microscope, or a broader clinical and imaging-based assessment. The distinction matters because pathologic complete response is a stronger predictor of long-term survival and is the standard endpoint in neoadjuvant trials.
The tumor mutation burden data collected through FoundationOne CDx could prove to be the trial’s most valuable contribution, but individual scores and their correlation with outcomes have not been released. If high-TMB patients account for most of the complete responses, that biomarker could become a tool for selecting patients who benefit most. Without the published breakdown, the stratification strategy remains a promising design feature rather than a validated selection method.
Finally, generalizability is limited by design. NEOPRISM-CRC enrolled only patients with dMMR/MSI-H tumors. For the 85 to 90 percent of colorectal cancer patients whose tumors are mismatch-repair proficient and microsatellite stable, PD-1 blockade alone has shown little activity. Nothing in this dataset suggests the approach would work for that larger group.
Where the field goes from here
Larger, randomized trials will need to test neoadjuvant pembrolizumab against existing treatment regimens, refine the biomarker thresholds that predict deep responses, and determine whether some patients can safely forgo chemotherapy or even surgery. Several such studies are in various stages of planning or enrollment across the UK, Europe, and the United States.
For patients diagnosed today with dMMR/MSI-H bowel cancer, the practical takeaway is that molecular testing at diagnosis is more important than ever. Knowing whether a tumor carries mismatch-repair deficiency can open the door to clinical trials and, increasingly, to treatment strategies that look very different from the chemotherapy-first playbook that dominated colorectal oncology for decades.
NEOPRISM-CRC is not yet proof that three doses of immunotherapy can cure bowel cancer. It is 32 patients, one drug, no relapses, and a set of unanswered questions that the oncology community is now racing to resolve. But for a disease that kills more than 16,000 people a year in the UK alone, according to Cancer Research UK, even a preliminary signal this clean demands serious attention.
More from Morning Overview
*This article was researched with the help of AI, with human editors creating the final content.
