Mismatch Repair a Potential Biomarker for PD-1 Inhibition

Treatment with the PD-1 inhibitor pembrolizumab demonstrated high response rates in patients with heavily pretreated colorectal cancer who harbored genetic defects in mismatch repair.

Dung T. Le, MD

Treatment with the PD-1 inhibitor pembrolizumab (Keytruda) demonstrated high response rates in patients with heavily pretreated colorectal cancer who harbored genetic defects in mismatch repair (MMR), according to findings from an ongoing phase II study presented at the 2015 ASCO Annual Meeting.

In patients with MMR deficiencies the objective response rate (ORR) was 62% compared with 0% in patients with MMR-proficient tumors. Median progression-free survival (PFS) and overall survival (OS) were not reached, with many patients responding to treatment for longer than 12 months in the MMR-deficient arm.

"This is the first study to use genetics in a prospective manner to guide immunotherapy," lead author Dung T. Le, MD, an assistant professor of oncology at Johns Hopkins Kimmel Cancer Center, said during a press briefing. "Mismatch repair deficient tumors are highly responsive to checkpoint blockade with anti-PD-1."

Findings from the analysis were simultaneously published in The New England Journal of Medicine (NEJM); however, data at the ASCO meeting were from a more up-to-date analysis that was conducted on May 8, 2015.

In the 3-arm study, pembrolizumab was administered at 10 mg/kg every 2 weeks to patients with colorectal cancer (CRC) who were MMR-deficient (n = 13) and MMR-proficient (n = 25). Additionally, a separate arm looked at pembrolizumab in patients with MMR-deficient non-CRC malignancies (n = 10). MMR and microsatellite instability testing was conducted using PCR and IHC, which are standard tests conducted for patients with CRC in order to detect Lynch syndrome.

"Mismatch repair deficiency is easily determined using an existing commercially available test," Le said. "In terms of cost, the tests run in the hundred of dollars, not the thousands that panels cost."

The primary endpoint of the study was immune-related PFS and response rate at 20 weeks. Secondary endpoints focused on OS, PFS, and disease control rate (DCR; complete response, partial response, plus stable disease). Response and survival were assessed by RECIST criteria in addition to immune-related criteria.

In the first 48 patients analyzed from the study, those with MMR-deficient CRC experienced an ORR of 62% and a DCR of 92%. The ORR was 0% and the DCR was 16% in MMR-proficient tumors. After a median treatment duration of 5.9 months, no patients in the MMR-deficient group who responded had progressed.

OS and PFS were not reached in the MMR-deficient group versus a median PFS of 2.3 months (HR = 0.103; 95% CI, 0.029-0.373; P <.001) and an OS of 7.6 months in the MMR-proficient group (HR = 0.216; 95% CI, 0.047-1.000; P = .05). In patients with MMR-deficient non-CRC tumors, the ORR was 60% and the DCR was 70%.

The adverse events seen in the study were consistent with other studies of pembrolizumab. The most common side effects were rash/pruritus (17%), pancreatitis (15%), and thyroiditis/hypothyroidism (10%).

"Responses were durable in a treatment-refractory patient population and many of these responses were ongoing for over a year," said Le. "These data suggest that genomics are more influential than histology for mismatch repair deficient tumors treated with anti-PD-1."

Defects in MMR commonly lead to microsatellite instability, which can be found in most cancers, including a majority of patients with hereditary nonpolyposis CRC (Lynch syndrome). Without this repair mechanism, the mutational burden is generally higher, suggesting a higher likelihood of developing cancer. In total, more than 80% of patients in the MMR-deficient arm were positive for Lynch syndrome.

"Mismatch repair deficiency is represented in approximately 4% to 5% of many tumor types, so it has broad applicability," Le said. "We saw responses in colorectal cancer, endometrial cancer, stomach cancer, small bowel cancer, and bile duct cancers."

Interestingly, patients with Lynch syndrome (n = 11) were less likely to respond compared with those with other forms of MMR, according to the data published in NEJM. In those with Lynch syndrome, the ORR was 27% with pembrolizumab compared with 100% in those with MMR that was unrelated to Lynch syndrome (n = 6).

In total, 1782 somatic mutations were identified per patient in the MMR-deficient arm compared with 73 in those with MMR-proficient tumors. Predominately, these tumors were found to alter amino acids (63%) and 578 of the somatic mutations in the deficient arm were associated with the immune system.

Overall, membranous PD-L1 expression was only identified in patients with MMR-deficient tumors. Additionally, tumors with MMR-deficiencies were more likely to contain a greater density of CD8+ lymphoid cells. However, the researchers noted that neither PD-L1 nor CD8 were significantly associated with PFS and OS.

"PD-L1 was done on all patients. There was a trend for PD-L1 expression but nothing was stronger than the mutational burden," senior author Luis A. Diaz, Jr, MD, associate professor of Oncology, Johns Hopkins Sidney Kimmel Comprehensive Cancer Center, said in an interview with OncLive. "Follow-up studies are looking at the mutations. We did have 4 non-responders, and so we're trying to understand who those non-responders were and what their mutational profiles were."

Pembrolizumab, which was the first PD-1 inhibitor to gain FDA approval in the United States, is the topic of multiple presentations at the ASCO meeting. More than 40 abstracts being presented at the meeting mention the highly selective humanized monoclonal IgG4 antibody. An application for pembrolizumab is pending in non-small cell lung cancer. It is currently approved for patients with advanced or unresectable melanoma following progression on prior therapies.

Merck, the developer of pembrolizumab, plans to launch a phase II study (KEYNOTE-164) to evaluate the effectiveness of MMR as a predictive marker in patients with locally advanced unresectable or metastatic colorectal cancers. The study is expected to begin enrolling patients later this year.

1. Le DT, Uram JN, Wang H, et al. PD-1 blockade in tumors with mismatch repair deficiency. J Clin Oncol. 2015;(suppl; abstr LBA100).

2. Le DT, Uram JN, Wang H, et al. PD-1 blockade in tumors with mismatch-repair deficiency. N Engl J Med. 2015. doi:10.1056/NEJMoa1500596


View more from the 2015 ASCO Annual Meeting