NEOPRISM-CRC: A 33-Month Cancer-Free Run Challenges Standard Chemo

TL;DR

• What: In the UK-led NEOPRISM-CRC trial, 32 patients with stage 2–3 MMR-deficient / MSI-high colorectal cancer received nine weeks of neoadjuvant pembrolizumab before surgery. After 33 months of follow-up, none have relapsed.
• Contrast: Standard care (surgery followed by months of chemotherapy) sees roughly 25% recurrence within three years in this population.
• Mechanism insight: Personalized circulating tumour DNA (ctDNA) tests tracked molecular response; when tumour DNA disappeared from blood, long-term remission followed.
• Scope limit: This applies to the MMRd/MSI-high subtype only—about 10–15% of colorectal cancers, or roughly 2,000–3,000 patients per year in the UK.
• Status: Presented at AACR 2026; not yet a peer-reviewed publication or regulatory label expansion. The evidence is encouraging but not definitive.

What Happened

In February 2023, a 73-year-old patient named Christopher Burston was diagnosed with stage 3 bowel cancer after routine screening detected blood in his stool. He enrolled in a trial led by University College London (UCL) and University College London Hospitals (UCLH) and received three doses of pembrolizumab over nine weeks before undergoing surgery in May 2023. The tumour, in his surgeon's words, "melted away." Nearly three years later, he remains cancer-free and has returned to normal life.¹

Burston is one of 32 participants in the NEOPRISM-CRC study. All patients had stage 2 or 3 colorectal cancer with a specific genetic signature: mismatch repair deficient (MMRd) or microsatellite instability-high (MSI-high). This subtype is characterised by a high mutational burden and is already known to respond well to immune checkpoint inhibitors in metastatic disease. What is new here is the timing and duration—a short neoadjuvant burst rather than months of adjuvant chemotherapy.

Early pathological results showed that 59% of patients had no detectable cancer remaining by the time of surgery. The update presented at the American Association for Cancer Research (AACR) Annual Meeting 2026 in San Diego confirmed that after 33 months of follow-up, no patient—whether they had a complete pathological response or residual disease—has experienced a relapse.²

What It Actually Means

The numbers carry the argument

Three years is the critical window for recurrence in high-risk colorectal cancer. A 0% relapse rate at 33 months in a population where 25% would typically recur is not a marginal improvement; it is a categorical difference. The standard of care in this subgroup has been surgery followed by adjuvant chemotherapy—usually FOLFOX or CAPOX for three to six months. That regimen is toxic, associated with peripheral neuropathy, fatigue, and diarrhoea, and a meaningful fraction of patients cannot complete it.

Replacing months of chemotherapy with nine weeks of immunotherapy—if the durability holds—would spare patients toxicity while potentially improving disease-free survival. That is the case the data make. The trial does not yet prove overall survival benefit; the sample size is small; and the follow-up, while substantial, is not yet mature enough to declare cure.

Why this subtype responds

MMRd/MSI-high tumours produce large numbers of neoantigens. Pembrolizumab, an anti-PD-1 antibody, removes the brake on T-cell recognition of these antigens. In the metastatic setting, this mechanism is well established. In the neoadjuvant setting—treating a primary tumour before surgery—there is emerging evidence that the intact tumour microenvironment and higher antigen load generate a more robust systemic immune response. The NEOPRISM data add a long-term durability signal to that biological hypothesis.

Hype Deconstruction: What This Is Not

This is not a cure for all colon cancer. It is not even a cure for all stage 2–3 colorectal cancer. The MMRd/MSI-high subtype represents 10–15% of cases. In the UK, that is roughly 2,000–3,000 patients per year; in the US, perhaps 15,000–18,000. The remaining 85–90% of colorectal cancers—MSS/pMMR tumours—do not respond to single-agent checkpoint inhibitors in the same way. Extending this approach to the broader population will require combination trials, different agents, or biomarker-driven selection that is still under investigation.

This is also not a randomised controlled trial. NEOPRISM-CRC is a phase 2 single-arm study with 32 patients. The comparison to historical 25% recurrence rates is suggestive, not conclusive. A phase 3 randomised trial against standard adjuvant chemotherapy is the next required step. Until that readout, this protocol should be considered promising investigational therapy, not a new standard.

The "melted away" language that appears in patient testimonials and press coverage is emotionally accurate but scientifically imprecise. Pathological complete response means no viable tumour cells were found in the resected specimen under microscopy. It does not mean the cancer was "destroyed" by the immune system in every patient; some had residual disease that remained dormant. The biological distinction matters for how we think about long-term monitoring.

Stakeholder Landscape

Cross-Layer Implications

Diagnostics before therapeutics. This story is as much about biomarker testing as it is about pembrolizumab. If neoadjuvant immunotherapy becomes standard for MMRd/MSI-high tumours, every colorectal cancer diagnosis will need rapid molecular profiling to determine eligibility. Health systems that still batch-test post-surgery will need to move to upfront parallel testing. That infrastructure shift has workforce, reimbursement, and turnaround-time implications.

Circulating tumour DNA as a surveillance tool. The trial used personalised ctDNA blood tests (developed with biotech company Personalis) to track molecular response. When ctDNA cleared, long-term remission followed. If validated, ctDNA could replace frequent imaging and CEA monitoring in this population, reducing patient burden and imaging costs—but only if payers reimburse it.

Regulatory precedent. The FDA and EMA have already approved neoadjuvant pembrolizumab in other tumour types (e.g., triple-negative breast cancer, melanoma). A successful phase 3 in colorectal cancer would accelerate the trend toward "immunotherapy first, surgery second" in solid tumours, reshaping surgical oncology training and clinical trial design.

Recommendations

For patients and families:

• If you or a relative has been diagnosed with stage 2–3 colorectal cancer, ask your oncology team: Has the tumour been tested for MMR/MSI status? If not, request it. This biomarker determines whether you might be eligible for neoadjuvant immunotherapy trials or, eventually, standard protocols.
• Do not stop or delay standard chemotherapy based on this news if you are already in treatment. The evidence is promising but not yet practice-changing outside a clinical trial.

For clinicians:

• Review the AACR 2026 abstract for patient-selection criteria and adverse-event profile. If your practice treats colorectal cancer, begin internal discussions about how to sequence MMR/MSI testing so results are available before surgery, not after.
• For eligible patients, consider referral to ongoing or upcoming neoadjuvant immunotherapy trials. In the UK, the trial network is UCL-led; in the US, NCT registration should be checked for open sites.

For health system administrators and insurers:

• Run a budget-impact model for replacing adjuvant chemo with neoadjuvant immunotherapy in the MMRd/MSI-high subgroup. Include avoided toxicity costs (neuropathy management, hospitalisations for neutropenic fever) and potential downstream savings from reduced recurrence.
• Evaluate ctDNA test coverage policies. If ctDNA-guided surveillance proves predictive, early coverage decisions will affect patient access and system cost.

Uncertainty Ledger

Bottom Line

A 33-month, zero-relapse signal in 32 patients with a specific colorectal cancer subtype is not a miracle—it is strong, early evidence that neoadjuvant immunotherapy may outperform months of chemotherapy in MMRd/MSI-high disease. The biological mechanism is understood, the biomarker is testable, and the patient burden is lower. But the population is narrow, the data are not yet from a randomised trial, and the broader 85% of colorectal cancer patients should not extrapolate from these results. If you have stage 2–3 colorectal cancer, get your tumour tested. If you are a clinician, prepare your diagnostic pipeline. If you are a policymaker, start modelling cost now—because if the phase 3 confirms this, adoption will be rapid.

Sources & Tiering