Background: Tumors with deficient mismatch repair (dMMR) have a favorable immunological phenotype permitting exploitation by immunotherapies. We aimed to assess our institutional experience of dMMR advanced gastrointestinal (GI) cancers treated with the PD-1 inhibitor pembrolizumab. Materials and Methods: We conducted an observational cohort study of a clinical series of patients with dMMR metastatic GI cancers treated with pembrolizumab from 2015 to 2017. Patients were assessed for best response, time to and reason for discontinuation, and adverse events. Results: A total of 13 patients received at least one dose of pembrolizumab. Median age was 62 years (range 33–74 years). Diagnoses included colorectal (colorectal cancer [CRC], n = 7); extrahepatic and intrahepatic cholangiocarcinoma (EHCC,n = 2;n = 1); pancreatic (pancreatic ductal adenocarcinoma [PDAC], n = 2); and adenocarcinoma of the appendix (n = 1). Five patients received concurrent chemotherapy (FOLFOX or capecitabine) with pembrolizumab (200 mg intravenous [IV] q 2 weeks with FOLFOX or 2 mg/kg IV q 3 weeks with capecitabine). Pembrolizumab was administered 2 mg/kg IV q 3 weeks to all patients who received single-agent treatment. Eleven patients were evaluable for response assessment. Three patients had a complete response (CRC and two EHCC) and one of these patients received concomitant pembrolizumab and FOLFOX. Two patients had a partial response, one with PDAC (−88% per RECIST, continues on treatment after 15.7 months) and the other with CRC (−45% per RECIST, continues after 14.6 months), both patients received concomitant pembrolizumab and FOLFOX and are now maintained on single-agent pembrolizumab. The objective response rate was 42%. Three patients experienced immune-related adverse events requiring discontinuation. Conclusions: This single-institution case series confirms the activity of pembrolizumab in various GI cancers harboring dMMR. Future studies are warranted to determine the role of combinatorial treatment with chemotherapy and/or novel immunotherapies in this population.
PD-1 inhibitors were recently approved in the United States for the treatment of advanced mismatch repair deficient (deficient mismatch repair [dMMR]) or microsatellite instability-high (MSI-high) colorectal cancer (CRC) (nivolumab) or all solid tumors harboring these genetic aberrations (pembrolizumab). The approval of pembrolizumab represents the first cancer indication based solely on a genetic alteration without regard to the site/tissue of origin.
DNA mismatch repair is a highly conserved cellular process responsible for identifying and repairing mismatched base pairs or short insertions and deletions that occur during DNA replication and recombination. When this process develops a functional error or defect, it results in dMMR leading to a high mutational burden and MSI. Lynch syndrome is an autosomal dominant disorder in which patients develop dMMR tumors that arise through inheritance of a germline mutation in MLH1, MSH2, MSH6, or PMS2, or altered EPCAM (TACSTD1 gene). dMMR also occurs in sporadic colon cancer secondary to somatic mutations, most commonly hypermethylation of the MLH1 promoter.
Lynch syndrome is responsible for 1%–5% of all colon cancer whereas sporadic dMMR accounts for 10%–20% of colon cancer.  dMMR in colon cancer has a distinct clinical phenotype including favorable prognosis, proximal origin, lymphocytic infiltration, inflammatory state, poorly differentiated morphology, and mucinous or signet ring differentiation. A pooled analysis of four large prospective phase III studies revealed a 5% prevalence of dMMR in metastatic colon cancer. Due to this lower dMMR prevalence in metastatic disease, it is believed that there is reduced metastatic potential in early stage dMMR disease. However, dMMR in metastatic disease is associated with inferior survival, potentially due to cooccurring BRAF mutations. Outside of colon cancer, other gastrointestinal (GI) cancers demonstrate dMMR including gastric (9%–19%), rectal (3%), hepatocellular (3%–4%), pancreatic (~2%–3%), and cholangiocarcinoma (~2%–3%).[5, 6]
Tumors with dMMR have a favorable immunological phenotype permitting by novel immunotherapies. The high mutational burden and substantially increased mutation-associated neoantigens generated in dMMR tumors pose the immune system for activation.[7–9] These tumors are host to increased numbers of tumor-infiltrating lymphocytes, theorizing the ability for an increased immune response.[10, 11] dMMR exploitation tumors also demonstrate increased expression of immune checkpoint ligands including PD-1, PD-L1, cytotoxic T-lymphocyte associated protein 4 (CTLA-4), LAG-3, and IDO. The advent of immunotherapies which target these ligands, such as the CTLA-4, programmed death-1 (PD-1), and PD-1 ligand (PD-L1) inhibitors have resulted in remarkable clinical activity in tumors harboring dMMR.[7, 13–15]
Based on the clinical success of PD-1 inhibitors in this patient population, we aimed to assess our institutional experience of dMMR advanced GI cancers treated with the PD-1 inhibitor pembrolizumab to characterize response rates and adverse events.
Materials and Methods
We conducted a retrospective cohort study of a series of consecutive patients with dMMR metastatic GI malignancies treated with pembrolizumab for at least one dose between July 15, 2015 and August 9, 2017. Patients were not excluded if they received concomitant chemotherapy and pembrolizumab and/or received treatment while enrolled in an institutional, investigator-initiated clinical trial assessing pembrolizumab combined with mFOLFOX6 in patients with advanced GI cancers (NCT02268825, IRB_00076239). MSI status was assessed from pathology reports from a local CLIA certified laboratory using immunohistochemistry (IHC) or polymerase chain reaction-based tests or from MSI testing included from FoundationOne testing from Foundation Medicine. Patients were assessed for best response (RECIST v1.1), time to and reason for discontinuation, and adverse events (IRB_00010924).
A total of 13 patients were identified that received at least 1 dose of pembrolizumab. Median age was 62 years (range 33–74 years), [ Table 1]. Three patients had germline mutations (MSH2, MLH1 and PMS2, respectively). dMMR tumors ( n = 10) were identified by IHC (7 of 10) and comprehensive genomic profiling (3 of 10). The most common variants were loss or mutation of MLH1 (5 of 10) and PMS2 (5 of 10). Nine subjects had tumor mutational burden (TMB) reported the median score was 36.5 mutations/megabase (muts/mb) and 2 (22%%) subjects had a score <20 muts/mb, the cut point for high tumor mutation burden. Table 2 for individual subject dMMR loss or alterations and TMB. Diagnoses included colorectal (CRC, n = 7), extrahepatic cholangiocarcinoma (EHCC, n = 2), intrahepatic cholangiocarcinoma (IHCC, n = 1) pancreatic ductal adenocarcinoma (PDAC, n = 2), and adenocarcinoma of the appendix (n = 1). Five patients received concurrent cytotoxic chemotherapy (FOLFOX or capecitabine) with pembrolizumab (200 mg intravenous [IV] every 2 weeks with FOLFOX or 2 mg/kg IV every 3 weeks with capecitabine). Pembrolizumab was administered 2 mg/kg IV every 3 weeks to all patients who received single-agent treatment. Pembrolizumab was administered a median of 12 doses (range 1–24).
Twelve patients were evaluable for response assessment [ Figure 1]. One patient was unevaluable and died before the first scan due to a bowel perforation not related to treatment (pembrolizumab + FOLFOX). Three patients had a complete response (CRC and two EHCC) and one of these patients received concomitant pembrolizumab and FOLFOX. Two patients had a partial response, one with PDAC (−88% per RECIST, continues on treatment after 15.7 months), [ Figure 2] and the other with CRC (−45% per RECIST, continues on treatment after 14.6 months), both patients received concomitant pembrolizumab and FOLFOX and are now maintained on single-agent pembrolizumab. Six patients had stable disease; 1 patient with PDAC (−28% per RECIST, continues on treatment after 10.6 months), 4 patients with CRC (two −21%, −27%, and +6% per RECIST), and 1 patient with adenocarcinoma of the appendix (4% per RECIST, time to discontinuation 2.8 months). One patient had progressive disease as best response to treatment (IHCC, +21% per RECIST, time to progression 4.2 months) with spinal cord compression and transitioned to hospice.
The objective response rate in the evaluable population was 42%. The median time to first response was 3.3 months. The median time to first response was similar between those with germline and somatic dMMR alterations (3.3 vs. 3.0 months, P = 0.12) and with and without concomitant FOLFOX/capecitabine (3.1 vs. 2.4 months, P = 0.33). In patients with CRC, the objective response rate was 33% (2 of 6 patients) and in those with non-CRC GI malignancies, the objective response rate was 50% (3 of 6 patients). The objective response rate in those who received concomitant FOLFOX or capecitabine was 60% (3 of 5 patients), compared to 28% (2 of 7 patients) in those who received single-agent pembrolizumab. In those who received concomitant pembrolizumab + FOLFOX prior treatment for metastatic included gemcitabine + nab-paclitaxel (2 patients), FOLFIRI + bevacizumab (1 patient), and two received no previous treatment.
One patient with (CRC-6) died from upper GI bleeding related to disease progression. Three patients experienced immune-related adverse events requiring discontinuation of pembrolizumab. The median time to discontinuation of pembrolizumab for immune-related adverse events was 6 months (range 4.8–10.2 months). One patient treated with concomitant FOLFOX discontinued pembrolizumab after experiencing Grade 3 hyperthyroidism and adrenal insufficiency (EHCC-1). A second patient treated with concomitant capecitabine discontinued pembrolizumab after an incidental finding of pneumonitis with enlarged mediastinal lymph nodes on restaging CT scan and subsequent biopsy showing non-necrotizing granulomas consistent with sarcoidosis (Appendix Adeno). A third treated with pembrolizumab monotherapy patient also discontinued for pneumonitis (EHCC-2). Seven patients remain on pembrolizumab at time of the last follow-up (August 2017) and one was lost to follow-up (CRC-3).
The results from our small cohort confirm the activity of the PD-1 inhibitor pembrolizumab in advanced GI tumors with mismatch repair deficiency. These results are particularly relevant given the small number patients per cancer type forming the basis of the FDA approval for this indication (≤11 patients per non-GI cancer type).[16, 17] The previous reports have demonstrated similar activity in this patient population with an objective response rate of 53% treated with pembrolizumab (10 mg/kg IV every 2 weeks) compared to an objective response rate of 42% in our study.[6, 7] In an observational report presented in abstract and poster form, an objective response rate of 52% was observed with pembrolizumab (mostly 2 mg/kg or 200 mg flat dose) in dMMR CRCs. It is notable that the previously published reports also demonstrate a similar response rate in the CRC and non-CRC populations stemming from the increased mutant neoantigens present in dMMR cancers irrespective of cancer type.[6, 7]
In our study, responses were observed in patients treated with and without concomitant cytotoxic chemotherapy. Concomitant cytotoxic chemotherapy did not appear to prevent sensitivity to PD-1 inhibition through chemotherapy-induced immunosuppression. Indeed, it has been hypothesized that concomitant cytotoxics (chemotherapy or radiation therapy) may induce release of antigens and prime response to immunotherapy. In our cohort, we observed a trend for increased response rates in those treated with immunotherapy combined with chemotherapy (60% response rate vs. 28% with single-agent pembrolizumab); however, this needs to be formally assessed prospectively with larger sample sizes. Prospective and randomized studies are planned to evaluate the role of checkpoint inhibitors in combination with chemotherapy in patients with MSI-High CRC. In the adjuvant setting, the Alliance 021502 (NCT02912559) will randomize patients with Stage III MSI-High CRC to 6 months of adjuvant FOLFOX versus FOLFOX plus atezolizumab (PD-L1 inhibitor). The IDEA collaboration recently showed that 3 months of adjuvant chemotherapy was not clinically inferior to 6 months in patients with low risk (T3N1) CRC. The investigators of IDEA have not yet reported outcomes for MSI-High CRC. These results are critical to the design of Alliance 021502 as 6 months of chemotherapy may no longer be the standard duration of treatment for these patients. For patients with metastatic MSI-High CRC, the ongoing KEYNOTE-177 randomizes patients to standard FOLFOX or FOLFIRI ± EGFR or VEGF monoclonal antibody (per investigator choice) versus pembrolizumab. The primary end-point of KEYNOTE-177 is overall survival. The SWOG1610 study is being planned and will randomize patients to standard chemotherapy with FOLFOX plus bevacizumab versus the same regimen with the addition of atezolizumab versus atezolizumab single agent (NCT02997228).
The previous studies utilized a higher dose intensity of pembrolizumab (10 mg/kg IV every 2 weeks) compared to this study (200 mg IV every 2 weeks or 2 mg/kg IV every 3 weeks).[6, 7] Our data demonstrates similar activity at a lower dose and frequency and supports the FDA labeling decision of 200 mg IV every 3 weeks, which is also substantiated by Leal et al.[ 16, 17] In our case series, we did observe a higher than expected rate of immune-related adverse events, with two patients experiencing pneumonitis and one case of hyperthyroidism. It is unknown if this higher rate of immune-related adverse events is due to random chance or related to use of concomitant cytotoxic chemotherapy or the patient population studied. Further, study is warranted to more accurate assess rates of immune-related adverse events in this population.
This study is limited by the small sample size, retrospective response assessment, and the different treatments utilized.
This case series confirms the activity of pembrolizumab in various GI malignancies harboring dMMR. Future studies are warranted to determine the mechanisms of primary and acquired resistance to PD-1 inhibition and the role of combinatorial treatment with standard chemotherapy and/or novel immunotherapies (IDO or other checkpoint inhibitors) in this population.
Financial support and sponsorship
The authors declared no funding related to this study.
Conflicts of interest
The authors disclosed no conflicts of interest.
For reprints contact:email@example.com