Immunotherapy has transformed the treatment of mismatch repair-deficient (dMMR) colorectal cancer, but its role in proficient mismatch repair (pMMR) disease remains limited. While neoadjuvant immune checkpoint inhibition has demonstrated substantial activity in dMMR colon cancer, identifying effective immunotherapeutic approaches for pMMR tumors remains an important area of investigation.
New findings from the phase 2 UNICORN trial provide further evidence supporting the investigation of neoadjuvant botensilimab plus balstilimab in both molecular subtypes, alongside detailed analyses of the biological changes associated with treatment response.
OncoDaily spoke with Filippo Ghelardi, MD, first author of the study, from the Department of Medical Oncology, Fondazione IRCCS Istituto Nazionale dei Tumori, Milan, Italy. He shared his insights into the clinical significance of the findings, the biological characteristics associated with treatment response, and the potential role of circulating tumor DNA (ctDNA) in assessing neoadjuvant treatment activity.
On November 2, 2026, Nature Medicine published the study titled “Neoadjuvant botensilimab with or without balstilimab in pMMR and dMMR colon cancer: the phase 2 UNICORN trial.”
The study was authored by Filippo Ghelardi, Margherita Ambrosini, Alessandra Raimondi, Andrei Tyshevich, Federica Palermo, Francesca Bergamo, Vincenzo Nasca, Paolo Manca, Ilaria Mulas, Lisa Derosa, Gina Brown, Claudia Chiodoni, Massimo Di Maio, Samantha Di Donato, Anna Chiaramonte, Lorenzo Castagnoli, Stephanie Ni, Darryl Nousome, Erik Spickard, Antony Tin, Robert Lentz, Adham Jurdi, Giovanna Sabella, Daniele Spada, Stefano Tamberi, Alessandro Passardi, Federica Morano, Emiliano Tamburini, Chiara Rossi, Luca Azzolin, Francesco Cambuli, Antonia Martinetti, Elisa Sottotetti, Andrew Allan Almonte, Valerio Iebba, Chiara Ratti, Riccardo Cerantola, Özlem Kutlu, Eleonora Cristarella, Serenella M. Pupa, Cecilia Villa, Eleonora Perissinotto, Margherita Galassi, Andrea Barina, Fabrizio Vittadello, Marcello Guaglio, Gianluca Baiocchi, Andrea Sartore-Bianchi, Riccardo Rosati, Chiara Cremolini, Isacco Montroni, Sara Lonardi, Andrea Sottoriva, and Filippo Pietrantonio.
About the UNICORN Trial
UNICORN (NCT05845450) is an ongoing, multicenter, open-label, non-randomized phase 2 platform trial evaluating short-course preoperative treatments in patients with molecularly selected, resectable colorectal cancer. The published analysis included 56 patients with radiologically staged rT3–T4, N0/N+, M0 colon adenocarcinoma enrolled between June 2023 and July 2024.
Patients were assigned sequentially to four cohorts according to mismatch repair status and treatment:
- pMMR: Botensilimab (BOT) alone (n=14) or botensilimab plus balstilimab (BOT+BAL; n=14).
- dMMR: BOT alone (n=14) or BOT+BAL (n=14).
Botensilimab, a multifunctional Fc-enhanced anti-CTLA-4 antibody, was administered at 1 mg/kg on day 1. Patients assigned to combination treatment also received balstilimab, an anti-PD-1 antibody, at 3 mg/kg on days 1 and 15. Curative-intent surgery was planned for day 35 ± 5 days.
The primary endpoint was the centrally assessed pathological major response (pMR) rate, defined as 10% or less residual viable tumor at surgery. Secondary and exploratory assessments included safety, quality of life, event-free survival (EFS), circulating tumor DNA (ctDNA) dynamics, and extensive translational analyses.
Pathological Responses Across Treatment Cohorts
The trial met its primary endpoint in three of the four cohorts, with the exception of pMMR patients receiving botensilimab monotherapy. In pMMR colon cancer, BOT+BAL achieved a pMR rate of 36% (5/14; 95% CI: 12.8–64.9%), including four patients (29%) with a pathological complete response (pCR). Another five patients (36%) achieved a pathological partial response, resulting in an overall pathological response rate of 71%.
In the pMMR BOT monotherapy cohort, the pMR rate was 0% (95% CI: 0–23.2%), although six patients (43%) experienced a pathological partial response. In dMMR colon cancer, BOT+BAL achieved a 100% pMR rate (14/14; 95% CI: 76.8–100%), including 13 patients (93%) with a pCR. Among patients receiving BOT monotherapy, the pMR rate was 36% (5/14; 95% CI: 12.8–64.9%), including four patients (29%) with a pCR.
These findings demonstrate pathological activity with the combination in both molecular subtypes, including pMMR colon cancer, where conventional checkpoint inhibitors have historically shown limited efficacy. However, the small, sequentially enrolled, non-randomized cohorts do not establish comparative superiority between treatment regimens.
Commenting to OncoDaily on the clinical significance of the findings, Filippo Ghelardi highlighted the potential role of the combination in localized colon cancer:
“UNICORN provides preliminary proof of evidence that a short course preoperative treatment with the combination of botensilimab and balstilimab is safe and active in pMMR and dMMR localised colon cancer. This is especially significant for patients with pMMR disease, which represents the most common molecular subtype in CC and warrants further investigation of the role of this combination in the neoadjuvant setting.”
Event-Free Survival and Surgical Feasibility
At the February 9, 2026, data cutoff, median follow-up was 19.2 months for pMMR patients receiving BOT+BAL and 16.7 months for dMMR patients receiving the combination. No EFS events were observed in either combination cohort, with 12-month EFS rates of 100%.
In the BOT monotherapy cohorts, 12-month EFS was 93% in pMMR disease and 79% in dMMR disease. Six EFS events were reported, three in each monotherapy cohort. Surgical feasibility was maintained across the study. All 56 patients underwent resection, and surgery was timely according to the study definition in 98% of patients.
Adverse events occurred in 30 patients (54%), including immune-related adverse events in 23 (41%). Grade 3–4 immune-related adverse events occurred in one patient (2%). Immune-related adverse events were reported in eight of 28 patients (29%) receiving BOT alone and 15 of 28 (54%) receiving BOT+BAL.
One patient experienced grade 3 hyperthyroidism associated with a delay in surgery. Two patients did not receive their second balstilimab dose because of immune-related adverse events. No adverse event resulted in death or unresolved sequelae. Quality-of-life assessments also identified a significant decline in global health status among pMMR patients receiving BOT+BAL, from a mean baseline score of 77.8 to 61.1 at the end of treatment (P=0.01).
Translational Findings: Understanding Treatment Response
A major component of UNICORN was its translational research program, which analyzed matched pretreatment biopsies and surgical specimens from all 56 patients using single-cell spatial transcriptomics. The analyses revealed distinct immune features associated with pathological response in pMMR and dMMR tumors.
In pMMR disease, responders demonstrated significantly higher baseline expression of CD274, the gene encoding PD-L1, in adjacent and distant stromal compartments, but not within the tumor compartment. Responders also showed numerically higher regulatory T-cell (Treg) signatures at baseline and significantly lower signatures at surgery compared with non-responders.
In dMMR disease, pathological responders showed higher baseline expression of the activating Fc gamma receptor genes FCGR2A and FCGR3A in intratumoral macrophages, suggesting a potential association between botensilimab activity and tumor-resident myeloid cells. These exploratory findings identify immune features associated with treatment response but do not establish causal mechanisms or validated predictive biomarkers.
Discussing the translational findings with OncoDaily, Ghelardi emphasized the biological features associated with treatment response:
“We were able to leverage the incredibly rich translational platform to identify biological correlates of treatment response. In my opinion, the most interesting feature is the significantly higher expression of CD274 (the gene encoding PD-L1) in the baseline stromal compartment of pMMR responders. Furthermore, pMMR responders showed a dynamical regulatory T-cell compartment, with higher T-regs at baseline and a significant depletion after treatment, which has been consistently observed with botensilimab both clinically and pre clinically.”
ctDNA Clearance and Pathological Response
Longitudinal ctDNA analysis provided another important component of the study. Among patients with pMMR disease, ctDNA data were available for 27 of 28 patients. All evaluable patients were ctDNA-positive at baseline, while 78% remained positive at the end of treatment (P=0.023).
Six of 27 patients (22%) achieved ctDNA clearance before surgery, including all five evaluable patients with a pathological major response and one patient with a pathological partial response. All six had received BOT+BAL. In dMMR disease, ctDNA data were available for 24 of 28 patients. ctDNA positivity decreased from 88% at baseline to 38% at the end of treatment (P<0.001). Clearance was observed in 74% of dMMR pathological responders (P<0.001). Of the 19 pathological responders, three had no evaluable plasma samples, and one had been ctDNA-negative since baseline and was therefore excluded from the clearance analysis.
Importantly, no disease-free survival events occurred among patients who achieved ctDNA clearance before surgery at the data cutoff. These findings suggest that longitudinal ctDNA monitoring could provide a minimally invasive measure of response during neoadjuvant immunotherapy. However, its ability to guide treatment adaptation, surveillance, or postoperative chemotherapy decisions requires prospective validation.
Highlighting the potential future clinical application of these findings, Ghelardi commented to OncoDaily:
“Another very important finding is that all patients with pMMR disease had positive ctDNA at baseline and all responders experienced ctDNA clearance at the pre-surgical time point, allowing to hypothesise that in the future we may be able to see the incorporation of liquid biopsies as non invasive assessment of preoperative treatment success.”
What Comes Next?
The UNICORN findings contribute to a growing body of evidence investigating botensilimab plus balstilimab in the neoadjuvant treatment of colon cancer, alongside findings from the phase 2 NEST program. Following the initial results, UNICORN was amended to investigate an extended eight-week course of BOT+BAL in a larger cohort of 29 patients with locally advanced pMMR colon cancer.
These findings are expected to inform a planned global randomized phase 3 trial evaluating neoadjuvant BOT+BAL versus upfront surgery in patients with locally advanced pMMR colon cancer.
While the results are encouraging, the phase 2 UNICORN study was designed to identify signals of activity rather than establish long-term clinical benefit. Its small cohort sizes, non-randomized design, and relatively limited and unequal follow-up periods remain important considerations.
Takeaway
The phase 2 UNICORN trial provides preliminary evidence of pathological activity with short-course neoadjuvant botensilimab plus balstilimab in both pMMR and dMMR localized colon cancer.
The activity observed in pMMR disease is particularly relevant, given the historical limitations of conventional immunotherapy in this population. The accompanying spatial transcriptomic and ctDNA analyses offer additional insights into the biological and molecular features associated with response.
Larger randomized trials and longer follow-up will be needed to determine whether these pathological and molecular responses translate into improved long-term outcomes and whether ctDNA can reliably inform treatment decisions in the neoadjuvant setting.
The full article is available in Nature Medicine.
