Neoadjuvant Nivolumab/Ipilimumab Combination Shows Activity in Early-Stage dMMR Colon Cancer

According to findings from the first neoadjuvant study to test ipilimumab (Yervoy) plus nivolumab (Opdivo) in early-stage mismatch repair deficient tumors and MMR proficient colon cancers, the combination induced major pathologic responses from all 7 patients with dMMR tumors. 

Myriam Chalabi, MD

According to findings from the first neoadjuvant study to test ipilimumab (Yervoy) plus nivolumab (Opdivo) in early-stage mismatch repair deficient (dMMR) tumors and MMR proficient (pMMR) colon cancers, the combination induced major pathologic responses from all 7 patients with dMMR tumors. The short course of preoperative treatment was safe; surgery was not delayed in any patient, said Myriam Chalabi, MD, at the 2018 ESMO Congress.

Among the patients with dMMR tumors, response could not be predicted by pretreatment CD3 infiltration or immune gene signatures, said Chalabi, from The Netherlands Cancer Institute, Amsterdam.

Studies of PD-1 and CTLA-4 blockade in colorectal cancer “have focused mainly on patients with metastatic disease, and mostly heavily pretreated patients,” said Chalabi. “In that population, we know that MMR deficient tumors have very high response rates to both monotherapy and combination treatment, and these treatments are also now FDA approved. However, in MMR proficient tumors, no responses are seen, and even though dMMR tumors comprise only about 4% of metastatic colorectal cancers, 15% of the early-stage colon tumors are MMR deficient. However, so far we have no data on immune checkpoint inhibition in this patient population.”

For these reasons, researchers hypothesized that immune checkpoint inhibition would work in the neoadjuvant setting in primary colorectal cancer, as well.

In early-stage colon cancer, a much higher proportion of tumors have high T cell infiltration compared with the metastatic setting. Pre-existing T-cell response has been predictive of the response to immune checkpoint inhibition in nonmetastatic and metastatic colorectal cancer.

All of the data together led Chalabi’s group to hypothesize that the probability of response to immune checkpoint inhibition in early-stage colorectal cancer would be higher in early-stage dMMR and pMMR tumors.

The study included 19 patients with resectable, early-stage colon cancer who were treated with ipilimumab at 1 mg/kg on day 1 and nivolumab at 3 mg/kg on days 1 and 15. Patients then underwent surgery within a planned maximum of 6 weeks. Fourteen patients were evaluable for the analysis shared at ESMO. The adaptive design of the study will allow the addition of other combinations, especially for the pMMR tumors.

There were 8 pMMR tumors and 7 dMMR tumors (1 patient had a double tumor) in patients with histologically confirmed colon cancer (no rectal cancers were allowed) without distant metastases and no signs of perforation or clinical bowel obstruction. As expected, the tumor mutation burden was much higher in the dMMR tumors compared with the pMMR tumors (median, 1795 vs 103).

Treatment was well tolerated and all patients underwent radical resection without delays in surgery. Three patients had grade 3 events postoperatively: 1 case each of abdominal infection, anastomotic leak, and pneumonia. “None of these could be attributed to the immunotherapy,” said Chalabi.

Major pathological responses (<10% residual vital tumor) were observed in 7 of 7 (100%) dMMR tumors, with 4 (57%) complete responses. No major pathologic responses were observed in pMMR tumors, with little to no tumor regression. “However, we did see changes in the tumor microenvironment of these tumors as well,” Chalabi said.

Pretreatment CD3 infiltration was hypothesized to be much higher in the dMMR tumors, but this was not the case, said Chalabi, with no significant difference (P= .662) observed between the dMMR and pMMR tumors. Posttreatment, there was little increase (P= .461) in CD3 infiltration in the pMMR tumors and a significant increase (P= .031) in the dMMR tumors.

dMMR tumors had a significantly (P= .027) higher number of CD8+ cells at baseline. There was a median 4.8-fold increase in CD8+ T cells postoperatively in the dMMR tumors (P= .0009). “What was striking here was when we looked at the posttreatment samples, we saw that there was a significant (P= .018) 2.4-fold increase of CD8+ T cells also in these pMMR tumors that do not seem to be responding clinically,” said Chalabi.

In looking at the clonality of the T cell infiltration, there was no significant difference between dMMR and pMMR tumors preoperatively and a trend toward a more clonal T cell repertoire not only in the dMMR tumors, but in the pMMR tumors, as well.

Immune gene signatures were not associated with response, she said, but posttreatment IFNᵧ signatures, which increased significantly in the dMMR tumors (P= .036) but not in the pMMR tumors (P= .08), improved the ability to distinguish responders (dMMR) from nonresponders (pMMR).

“We believe that our data warrant independent validation of neoadjuvant immunotherapy in dMMR tumors in larger trials and that they have the potential to be practice changing,” Chalabi said.


Chalabi M, Fanchi L, van den Berg J, et al. Neoadjuvant ipilimumab plus nivolumab in early stage colon cancer. Presented at: 2018 ESMO; October 19-23; Munich, Germany. LBA37_PR.

Impressive response rates have been seen with neoadjuvant treatment in patients with melanoma and lung cancer, “which seem to be higher than what we know in the metastatic disease setting,” she said.