Neoadjuvant immunotherapy has emerged as a promising treatment strategy for patients with locally advanced mismatch repair-deficient (dMMR) colon cancer. Although previous studies have demonstrated high pathological response rates with immune checkpoint blockade, long-term survival data have remained limited.
Mismatch repair deficiency is present in approximately 10–15% of non-metastatic colon cancers. Despite the sensitivity of these tumours to immunotherapy, surgery remains the principal treatment for localised disease, with adjuvant chemotherapy generally recommended for stage III disease. Patients with high-risk stage II or stage III dMMR colon cancer can still face a substantial risk of recurrence following conventional treatment.
The phase II NICHE-2 trial previously demonstrated high pathological response rates following short-course neoadjuvant nivolumab plus ipilimumab. The updated analysis now reports long-term survival outcomes and circulating tumour DNA (ctDNA) dynamics following treatment.
The long-term results of the phase II NICHE-2 trial appeared in the October 31, 2026, issue of The Lancet under the title: “Disease-free survival and overall survival after neoadjuvant immunotherapy in mismatch repair-deficient colon cancer in the Netherlands: results of the single-arm, multicentre, phase 2, NICHE-2 trial.”
Authors: Myriam Chalabi, Lauren D. W. van den Dungen, Peter G. M. de Gooyer, Sara Balduzzi, Arend G. J. Aalbers, Emile D. Kerver, Cecile Grootscholten, Simone Dokter, Johanne G. Bloemen, Eddy R. Hendriks, Tjeerd S. Aukema, Steven J. Oosterling, Wieke H. M. Verbeek, Emile E. Voest, Niels F. M. Kok, Marnix H. Geukes Foppen, Regina G. Beets-Tan, Antony Tin, Adham Jurdi, Ton N. Schumacher, Monique E. van Leerdam, José G. van den Berg, and John B. A. G. Haanen.
Study Design and Treatment
NICHE-2 (NCT03026140) was an investigator-initiated, single-arm, multicentre phase II trial conducted across six hospitals in the Netherlands. The study enrolled adults with previously untreated, resectable, clinical stage II or III dMMR colon adenocarcinoma.
Patients received a short course of dual immune checkpoint blockade:
- Ipilimumab: 1 mg/kg intravenously on day 1.
- Nivolumab: 3 mg/kg intravenously on days 1 and 15.
- Surgery: Scheduled within six weeks of study enrolment.
Adjuvant chemotherapy was permitted at the treating physician’s discretion in patients with pathologically confirmed lymph node involvement. The trial had two independent primary endpoints: safety and 3-year disease-free survival (DFS). Safety and pathological response outcomes had been reported previously.
The current publication focuses on 3-year DFS, the secondary endpoint of ctDNA analysis, and an exploratory post-hoc assessment of 5-year overall survival (OS). DFS was measured from surgery until recurrence or death from any cause, whereas OS was calculated from study enrolment.
Patient Characteristics and Pathological Responses
Between July 4, 2017, and July 18, 2022, 123 patients were screened, 115 received treatment, and 111 were included in the per-protocol efficacy population.
The median age was 60 years, and 32% of enrolled patients had Lynch syndrome. Baseline imaging identified clinically node-positive disease in 67% and cT4 tumours in 64%, indicating that the study included a substantial proportion of patients with high-risk disease. All 111 patients in the per-protocol population completed both immunotherapy cycles and underwent surgery, achieving R0 resections.
Previously reported pathological findings demonstrated:
- Pathological response: 98% (109/111).
- Major pathological response: 95% (105/111).
- Pathological complete response (pCR): 68% (75/111).
Among 14 patients with pathologically confirmed lymph node involvement after surgery, three received adjuvant chemotherapy. These findings demonstrated the substantial pathological activity of short-course neoadjuvant nivolumab plus ipilimumab. The updated analysis assessed whether these responses were associated with durable disease control.
Long-Term Survival: No Recurrences in the Per-Protocol Population
At the June 19, 2026, data cutoff, the median postoperative follow-up was 58.7 months (IQR 49.0–72.9). All 111 patients in the per-protocol population remained alive and disease-free at their last follow-up. The 3-year DFS was 100% (97.5% CI 100–100), exceeding the prespecified success threshold of 93% and meeting the primary efficacy endpoint (one-sample log-rank p<0.0001).
Overall survival results were similarly encouraging. After a median follow-up of 60.4 months (IQR 50.6–74.1), calculated from study enrolment, the exploratory 5-year OS was 100% (95% CI 100–100). A post-hoc intention-to-treat analysis, including all 115 treated patients, showed a 3-year DFS of 99% (97.5% CI 97–100) and a 5-year OS of 100%.
Among the four patients excluded from the per-protocol analysis, one developed recurrent peritoneal disease after peritoneal metastases were identified during surgery. This patient subsequently became disease-free following cytoreductive surgery and hyperthermic intraperitoneal chemotherapy. The survival results indicate that the previously reported pathological responses were accompanied by durable disease control.
ctDNA Dynamics During Neoadjuvant Immunotherapy
The study also examined whether circulating tumour DNA could provide an early indication of treatment response and residual disease. Plasma samples were analysed using the personalised, tumour-informed Signatera assay (Natera) at baseline, week 2, before surgery, and approximately three weeks after surgery.
At baseline, ctDNA was detectable in 99 of 108 evaluable patients (92%). Among patients with detectable ctDNA at baseline and available follow-up samples, molecular clearance occurred rapidly:
- After the first cycle, at week 2: 45% (44/98) had ctDNA clearance.
- After the second cycle, before surgery: 82% (80/97) had ctDNA clearance.
ctDNA concentrations decreased significantly after each immunotherapy cycle (adjusted p<0.001). The investigators also observed differences in ctDNA clearance according to pathological response. At week 2, ctDNA clearance occurred in 58% of patients who subsequently achieved pCR, compared with 19% of those with a near-pathological complete response.
Before surgery, clearance was observed in 92% (60/65) of patients with pCR and 70% (19/27) with near-pathological complete response. Baseline ctDNA concentrations did not significantly differ between these two response groups (p=0.59). However, ctDNA concentrations after the first and second treatment cycles were significantly lower among patients who ultimately achieved pCR (p=0.004 and p=0.002, respectively). These findings suggest that early ctDNA dynamics may help characterise treatment response.
Can ctDNA Identify Residual Disease Before Surgery?
Despite the high rates of molecular clearance, presurgical ctDNA testing did not reliably exclude residual pathological disease.
For detecting residual disease, presurgical ctDNA demonstrated:
- Specificity: 93% (95% CI 83–97).
- Sensitivity: 35% (95% CI 19–55).
- Positive predictive value: 69% (95% CI 41–89).
- Negative predictive value: 75% (95% CI 64–84).
Notably, six patients with presurgical ctDNA clearance still had viable lymph node metastases in their resection specimens. At the postoperative minimal residual disease assessment, approximately three weeks after surgery, ctDNA was undetectable in all evaluated patients. Although these results demonstrate the potential value of ctDNA for monitoring treatment response, its limited presurgical sensitivity indicates that molecular clearance alone cannot establish the absence of residual disease.
Safety and Treatment Tolerability
No additional treatment-related adverse events or new safety signals were identified in the updated analysis. As previously reported, grade 3–4 treatment-related adverse events occurred in 4% of patients receiving neoadjuvant nivolumab plus ipilimumab. Importantly, the treatment involved only two immunotherapy cycles before surgery, rather than a prolonged course of systemic therapy.
Clinical Implications and Remaining Considerations
The updated results add long-term survival data to the evidence supporting neoadjuvant immune checkpoint blockade in locally advanced dMMR colon cancer. The combination of high pathological response rates, favourable long-term survival, and a short treatment duration is particularly relevant for patients with clinically high-risk disease.
However, the trial’s single-arm design remains an important limitation. Without a randomised control group, the observed outcomes cannot establish superiority over standard treatment or determine the contribution of each treatment component.
The investigators also discussed the phase III ATOMIC trial, in which adding atezolizumab to adjuvant FOLFOX improved 3-year DFS in stage III dMMR colon cancer. Nevertheless, differences in study design, staging methods, and patient characteristics prevent direct comparisons between ATOMIC and NICHE-2. Another consideration is whether response assessment could eventually support organ-preserving approaches in selected patients.
Although the pathological complete response rate is encouraging, the authors emphasised that non-operative management in colon cancer requires further prospective investigation. In particular, ctDNA findings support the need for combined molecular, radiological, and endoscopic assessment rather than reliance on a single test.
Further studies are also needed to clarify optimal treatment duration, patient selection, and the role of adjuvant therapy following neoadjuvant immunotherapy.
Takeaway
The long-term NICHE-2 analysis demonstrated 100% 3-year disease-free survival and an exploratory 5-year overall survival estimate of 100% in the per-protocol population following short-course neoadjuvant nivolumab plus ipilimumab and surgery for locally advanced dMMR colon cancer.
Alongside previously reported pathological responses, the ctDNA findings show rapid molecular clearance during treatment and ctDNA negativity in all patients evaluated at the postoperative assessment.
While the absence of a randomised comparator limits definitive conclusions about comparative efficacy, these results provide compelling evidence for the durability of response and support consideration of neoadjuvant immune checkpoint blockade as part of the standard of care for appropriately selected patients with locally advanced dMMR colon cancer.
