Background: The randomized METIMMOX trial evaluated short-course oxaliplatin-based chemotherapy alternating with the immune checkpoint inhibitor nivolumab in patients with previously untreated, unresectable abdominal metastases from microsatellite-stable/mismatch repair-proficient colorectal cancer. A subgroup receiving this experimental treatment showed remarkably improved outcome compared to control arm patients who received chemotherapy alone and had median progression-free survival (PFS) of 9.3 months. We examined whether the kinetics of serum carcinoembryonic antigen (CEA) could provide insights into responsiveness to the METIMMOX regimen.
Methods: Patients were randomly assigned to the control arm (oxaliplatin with bolus 5-fluorouracil/folinic acid, Q2W) or the experimental arm (alternating two cycles of chemotherapy Q2W and nivolumab Q2W). In this post hoc biomarker analysis, CEA levels were measured at baseline, at each treatment administration, and every two months during prespecified treatment breaks. Patients were categorized according to CEA kinetics-Flare (transient ≥20% increase above baseline), Non-responding (gradual increase), and Stable/Responding (values around baseline or gradually decreasing).
Results: Flare comprised 16 out of 71 patients. According to Log rank test, experimental-arm Flare (n = 6) had significantly improved PFS (median 34.9 months) compared to Stable/Responding (n = 21; median PFS 9.9 months; p = 0.005) and Non-responding (n = 9; median PFS 2.1 months; p = 0.007). Flare was not associated with improved PFS for control-arm subjects. Adjusted Cox regression analysis indicated that experimental-arm Flare exhibited a lower progression risk than Non-responding for up to 6.2 months, creating an observation period for a potential decrease following the initial increase.
Conclusion: CEA-Flare may indicate immune checkpoint inhibitor responsiveness following short-course oxaliplatin-based chemotherapy in this patient population; however, this is an exploratory finding that requires prospective validation.
Keywords: biomarker; carcinoembryonic antigen; colorectal cancer; immune checkpoint inhibition.
This study examined whether changes in the blood marker CEA can help identify patients who benefit from a new treatment approach for large bowel cancer. The METIMMOX study tested alternating standard chemotherapy and immunotherapy in patients whose cancer had spread to abdominal organs and could not be surgically removed. All patients had a common type of bowel cancer that usually does not respond well to immunotherapy. Some patients who received the combined treatment performed much better than expected, with their cancer staying under control longer than those who received chemotherapy alone. The researchers studied changes in CEA levels over time during treatment. Based on these changes, patients were grouped into the following three patterns: CEA flare: a temporary increase (at least 20%) followed by a decreaseNon-responding: a steady increase in CEAStable/responding: stable or gradually decreasing CEA levels Among patients who received alternating chemotherapy and immunotherapy, those who showed a CEA flare had much better outcomes. Their cancer remained under control for a median time of almost 3 years, compared with approximately 10 months for patients with stable/responding CEA levels and only 2 months for those which non-responding CEA levels. This benefit was not seen in the patients who received chemotherapy alone, meaning that a CEA flare may signal a positive immune response after chemotherapy. In summary, a short-term increase in CEA (CEA flare) may be a helpful sign that patients with large bowel cancer spread to other organs respond well to a combination of chemotherapy and immunotherapy.
© 2026 Meltzer et al.