Immunotherapy for Metastatic Mismatch Repair–Deficient Colorectal Cancer: Game-Changer for Small Group of Patients

Get Permission

PATIENTS WITH DNA mismatch repair–deficient (dMMR) metastatic colorectal cancer display a high level of microsatellite instability (MSI-H)1 and demonstrate poor chemosensitivity and shorter overall survival than patients with mismatch repair–proficient (pMMR) metastatic metastatic colorectal cancer.2 However, immune checkpoint blockade has elicited high responses in this subgroup, characterized by high tumor mutation burden.3,4 

Most recently, the programmed cell death protein 1 (PD-1)–blocking antibody nivolumab (Opdivo) has shown durable responses and disease control in patients with dMMR colorectal cancer in the open-label, single-arm, multicenter phase II CheckMate 142 study, leading to accelerated U.S. Food and Drug Administration approval for this indication.5 Nivolumab thus joins pembrolizumab (Keytruda) as a second weapon in the immunotherapy armamentarium for patients with previously treated dMMR colorectal cancer. 

“Nivolumab thus joins pembrolizumab as a second weapon in the immunotherapy armamentarium for patients with previously treated dMMR colorectal cancer.”
— Karuna Ganesh, MD, PhD

Tweet this quote

Clinical Trial Findings 

AS REVIEWED in this issue of The ASCO Post, Overman et al5 reported on the efficacy of 3 mg/kg of nivolumab given every 2 weeks in 74 patients with locally confirmed dMMR recurrent or metastatic colorectal cancer who had progressed on or were intolerant of at least one previous line of therapy, including a fluoropyrimidine and oxaliplatin or irinotecan (CheckMate 142). 

Similarly, Le et al reported on previously treated patients administered pembrolizumab at 10 mg/kg given every 2 weeks. This study included 86 patients with dMMR metastatic solid tumors regardless of site of tumor origin, and included 40 patients with colorectal cancer and 46 patients with 11 other histologies, expanding on a previously reported cohort of 11 patients with colorectal cancer.4 The investigator-assessed objective response rates and 12-month disease control rate in the nivolumab study and in the pembrolizumab study were comparable. Median overall survival had not yet been reached in either study, and median progression-free survival was just reached in the nivolumab study. 

Acquired resistance was uncommon. Although low-grade toxicities were common, grade 3 or 4 drug-related adverse events were reported in 20% of patients in both studies, with pancreatic enzyme elevation and colitis reported most frequently. Overall, the studies demonstrate the safety and efficacy of PD-1 blockade and highlight the promise of immunotherapy for achieving long-term durable responses in patients with previously treated metastatic colorectal cancer. 

One notable difference between the two studies was that 18 of 86 patients (20.9%) treated with pembrolizumab, including 5 (12%) with colorectal cancer, achieved a complete response, whereas no complete responses were achieved with nivolumab based on investigator-assessed objective response, the primary endpoint. Secondary blinded independent review did identify two complete responders in the nivolumab-treated population. On the other hand, the median time to response was 2.8 months with nivolumab in colorectal cancer, but 21 weeks in the pembrolizumab study across all tumor types. These differences may be attributable to specific clinical, histopathologic, or genetic differences in the patients, and larger studies are needed to evaluate whether they persist. In another study arm that was not reported by Overman et al, CheckMate 142 is currently assessing combination treatment with nivolumab and ipilimumab (Yervoy) in patients with dMMR/MSI-H colorectal cancer. 

Predictive Biomarkers of Response to PD-1 Inhibition 

EVEN WITHIN THE BIOMARKER-DEFINED dMMR/MSI-H population, about 25% of patients in both studies had primary progressive disease. Thus far, exploratory analyses have failed to identify biomarkers that predict response to PD-1 inhibition within the dMMR population. Overman et al found that PD-1 and programmed cell death ligand 1 (PD-L1) expression levels and BRAF- and KRAS-mutation status did not differ between responders and nonresponders. 

“The challenge before the oncology community is to expand the efficacy of immunotherapy to the 95% majority of patients with MMR-proficient tumors.”
— Luis A. Diaz, Jr

Tweet this quote

Le et al identified mutations of the B2M gene, a componet of the antigen-presenting major histocompatibility complex (MHC) I, which is crucial for cancer cell recognition by T cells, in four of five sequenced tumors that developed acquired resistance. However, no B2M mutations were identified in tumors that were primarily resistant to pembrolizumab. 

Lynch syndrome status also did not differentiate responders from nonresponders in either study. Using deep sequencing of T cell receptors, Le et al identified specific neoepitopes that elicited antitumor T-cell expansion and activity. Future studies may identify shared neoantigen sequences, which could be used to prospectively identify patients with tumors that are likely to respond to PD-1 blockade. 

Immunotherapy for MMR-Proficient Disease 

CHECKMATE 142 underscores the importance of MMR deficiency as a biomarker of response to immunotherapy in colorectal cancer. With two PD-1–inhibiting antibodies granted accelerated FDA approval, patients with metastatic dMMR/MSI-H colorectal cancer now have an expanding array of treatment options, with the potential for durable response. Although this is a game-changer for these patients, dMMR tumors represent only 5% of metastatic colorectal cancer. The challenge before the oncology community is to expand the efficacy of immunotherapy to the 95% majority of patients with pMMR tumors. 

The identification of cooperating signaling pathways that dampen antitumor immunity may be one approach to overcoming the resistance of pMMR colorectal cancer to immune checkpoint blockade. For example, in metastatic melanoma, Shin et al have identified tumor Janus kinase (JAK)1/2 mutations as mediators of primary resistance to PD-1 blockade.6 In a tantalizing advance for pMMR colorectal cancer, a phase Ib trial of combination MEK and PD-L1 inhibition showed enhanced T-cell infiltration, MHC I and PD-L1 expression, and an objective response rate of 17% at 3.78 months.7 The results of further clinical studies employing these approaches are eagerly awaited. 

DISCLOSURE: Drs. Ganesh and Diaz reported no conflicts of interest. 


1. Boland CR, Goel A: Gastroenterology 138:2073-2087.e3, 2010. 

2. Venderbosch S, Nagtegaal ID, Maughan TS, et al: Clin Cancer Res 20:5322- 5330, 2014. 

3. Le DT, Uram JN, Wang H, et al: N Engl J Med 372:2509-2520, 2015. 

4. Le DT, Durham JN, Smith KN, et al: Science 357:409-413, 2017. 

5. Overman MJ, McDermott R, Leach JL, et al: Lancet Oncol 18:1182-1191, 2017. 

6. Shin DS, Zaretsky JM, Escuin-Ordinas H, et al: Cancer Discov 7:188-201, 2017. 

7. Bendell JC, Kim TW, Goh BC, et al: 2016 ASCO Annual Meeting. Abstract 3502.

Related Articles

Nivolumab in Advanced DNA Mismatch Repair–Deficient or Microsatellite Instability– High Colorectal Cancer

Michael J. Overman, MD

Michael J. Overman, MD

Scott Kopetz, MD

Scott Kopetz, MD

Thierry André, MD

Thierry André, MD

THE PHASE II CHECKMATE 142 TRIAL has shown that nivolumab (Opdivo) produces durable responses in previously treated recurrent or metastatic DNA mismatch repair–deficient (dMMR)/ microsatellite instability–high (MSI-H)...