A meta-analysis of six randomized controlled trials found that adding PD-1 or PD-L1 checkpoint inhibitors to chemotherapy, with or without anti-VEGF treatment, was associated with longer progression-free and overall survival in patients with proficient mismatch repair metastatic colorectal cancer. Across 1,165 participants, the pooled progression-free survival hazard ratio was 0.84 and the overall survival hazard ratio was 0.83, indicating lower risks of progression or death and death, respectively, with combination treatment. The analysis was reported in Frontiers in Oncology in 2026, but most included trials were early-phase, and the reviewers rated the certainty of evidence for both outcomes as low.
Colorectal cancer is common worldwide, with approximately two million new cases and 900,000 deaths in 2022, according to the study’s background. Around one-third of patients have metastatic disease when the cancer is first diagnosed, making effective systemic treatment crucial. Mismatch repair status strongly influences treatment. Approximately 15% of colorectal cancers are mismatch repair-deficient or microsatellite instability-high, a group in which PD-1/PD-L1 inhibitor monotherapy or dual immunotherapy has shown substantial activity. The other 85% are mismatch repair-proficient, or pMMR, and these tumors have generally shown objective response rates below 5% with checkpoint-inhibitor monotherapy alone.
The rationale for combination therapy is that chemotherapy and anti-VEGF agents may alter a pMMR tumor’s immune environment, potentially allowing checkpoint blockade to become effective where it is ineffective by itself. Chemotherapy can promote immunogenic tumor-cell death, release tumor-associated antigens and enhance the activity of natural killer cells and CD8-positive T cells. Anti-VEGF treatment may normalize abnormal tumor blood vessels, improve oxygen delivery, reduce interstitial pressure and favor the recruitment of immune cells into tumors. These effects could help reverse immune exclusion and restore T-cell function, but the meta-analysis was designed to evaluate clinical outcomes rather than directly test these proposed mechanisms.
The researchers searched PubMed, Web of Science, Embase, the Cochrane Library and ClinicalTrials.gov from database inception through May 2026. Their analysis followed PRISMA guidelines and was prospectively registered in PROSPERO. Two reviewers independently screened and extracted data and assessed randomized trials comparing checkpoint inhibitors plus chemotherapy, with or without targeted therapy, against the corresponding chemotherapy regimen without a checkpoint inhibitor. Progression-free survival and overall survival were analyzed as hazard ratios with 95% confidence intervals using a random-effects model. One study supplied only age-subgroup data because its progression-free and overall survival results were not mature, so the main pooled analyses used five studies.
The six trials enrolled patients in Italy, the United States, Norway, China and multiple countries between 2015 and 2022. Five were phase II studies and one was phase II/III; five were open-label and one was double-blind. Five evaluated first-line treatment and one examined second-line or later therapy. Most experimental regimens combined a PD-1/PD-L1 inhibitor with chemotherapy and bevacizumab or another VEGF-targeted agent, while one used checkpoint inhibition with chemotherapy alone. The trials included 525 patients in experimental groups and 640 in control groups, and 88.5% of the population had pMMR tumors, consistent with the review’s focus but also indicating that the analysis was not restricted entirely to a uniformly defined pMMR population.
Both primary outcomes favored combination treatment. The pooled progression-free survival hazard ratio was 0.84, with a 95% confidence interval of 0.73 to 0.95 and P = 0.007, while overall survival was associated with a hazard ratio of 0.83, a 95% confidence interval of 0.73 to 0.94 and P = 0.004. Statistical heterogeneity was zero for both outcomes, and fixed-effect analyses produced identical hazard ratios. Egger’s tests did not identify significant publication bias for progression-free survival, P = 0.182, or overall survival, P = 0.467. Sequential exclusion of individual studies also left the pooled estimates within relatively narrow ranges, although this sensitivity exercise cannot remove the limitations of a small evidence base.
The trial-level risk of bias assessment identified five studies with clearly described randomization and low risk for allocation concealment, but five were open-label and therefore rated at high risk for blinding of participants and investigators. Four reported blinded outcome assessment, four had complete outcome data and all six were judged low risk for selective reporting. Using the GRADE framework, the reviewers downgraded the initial high certainty for both progression-free and overall survival by two levels to low, primarily because most studies did not blind participants and investigators. This low rating is important: the pooled estimates are statistically significant, but the strength and reliability of the underlying clinical evidence remains limited.
Exploratory progression-free survival subgroup analyses suggested possible larger benefits in several groups, including men, patients with an Eastern Cooperative Oncology Group performance status of 0, those with liver metastases, those without prior adjuvant treatment, patients with synchronous metastases, patients whose PD-L1 combined positive score was below 1 and those with KRAS mutations. The corresponding hazard ratios were 0.77, 0.73, 0.82, 0.81, 0.79, 0.53 and 0.49, respectively. No significant difference was detected within the complementary female, lower-performance-status, higher-CPS, KRAS wild-type or metachronous-metastasis groups. These contrasts should not be interpreted as established treatment selectors, however, because the investigators did not perform interaction tests and most subgroup estimates drew on only two to five studies.
The authors proposed biological interpretations for the exploratory findings, while emphasizing that these remain hypotheses. Anti-VEGF treatment might improve immune access to liver metastases through vascular normalization and immune sensitization, and patients without previous adjuvant chemotherapy may retain more intact immune function. A KRAS-mutated subgroup result was especially uncertain because it was based on an indirect comparison and conflicts with some prior evidence suggesting that KRAS mutations can contribute to an immunosuppressive microenvironment and resistance to checkpoint-inhibitor monotherapy. Similarly, the apparent advantage in patients with PD-L1 CPS below 1 could reflect an association with KRAS status, dynamic changes in PD-L1 expression during treatment or a statistical artifact. Prospective studies are needed before any of these subgroup signals can guide selection.
The review identifies checkpoint inhibitors plus chemotherapy, with or without anti-VEGF treatment, as a potential first-line option for pMMR metastatic colorectal cancer, but the clinical magnitude and implementation remain unresolved. A progression-free survival hazard ratio of 0.84 corresponds to a 16% relative reduction in the instantaneous risk of progression or death, and the meta-analysis did not perform a pooled safety analysis. The authors also note that the incremental value over established first-line chemotherapy plus bevacizumab remains uncertain, while checkpoint inhibitors add immune-related toxicity and financial burden. With only six randomized trials, mostly phase II, immature overall-survival data and low GRADE certainty, the findings provide a signal for larger phase III confirmation rather than definitive evidence for routine adoption.
Subject of Research: Efficacy of PD-1/PD-L1 inhibitor combinations in pMMR metastatic colorectal cancer
Article Title: PD-1/PD-L1 inhibitors in the treatment of pMMR mCRC: a meta-analysis based on randomized controlled trials
Article References: Lv, S., Ma, C., Tang, Z., Zhang, X., Shi, W., Zhang, S., Ni, M., Gu, L., & Chen, B. (2026). PD-1/PD-L1 inhibitors in the treatment of pMMR mCRC: a meta-analysis based on randomized controlled trials. Frontiers in Oncology, 16, Article 1969514. https://doi.org/10.3389/fonc.2026.1969514
Image Credits: AI Generated
DOI: 10.3389/fonc.2026.1969514
Keywords: pMMR colorectal cancer, metastatic colorectal cancer, PD-1 inhibitors, PD-L1 inhibitors, chemotherapy, anti-VEGF therapy, randomized trials, meta-analysis, progression-free survival, overall survival, PD-1, PD-L1
