Search

PURPOSE To characterize the relationship between tumor-infiltrating lymphocytes (TIL), tumor mutational burden (TMB), and genetic alterations in microsatellite stable (MSS), microsatellite instability (MSI), or mutant POLE/POLD1 colon cancer. MATERIALS AND METHODS Four hundred ninety-nine resected stage I-III colon tumors treated between 2014 and 2019 were assessed for TIL; somatic mutations, copy number alterations, and structural changes in > 400 oncogenes; and MSI status. RESULTS Of the 499 tumors analyzed, 313 were MSS, 175 were MSI, and 11 had POLE/POLD1 pathogenic mutations. Both the percentage of tumors with a high level of TIL (≥ 4 lymphocytes per high-power field) and the median TMB differed significantly between the three phenotypes: MSS, 4.5% and 6 mutations/Mb; MSI, 68% and 54 mutations/Mb; POLE/POLD1, 82% and 158 mutations/Mb ( P < .05). Within each phenotype, TMB did not vary significantly with TIL level. Among MSI tumors, the median number of frameshift indels was significantly higher in tumors with high levels of TIL (20 v 17; P = .018). In the MSS group, significantly higher proportions of tumors with high levels of TIL had mutations in the transforming growth factor-β (36% v 12%; P = .01), RAS (86% v 54%; P = .02), and Hippo (7% v 1%; P = .046) pathways; in contrast, TP53 alterations were associated with low levels of TIL (74% v 43%; P = .01). CONCLUSION The association between TIL, TMB, and genetic alterations varies significantly between MSI, MSS, and mutant POLE/POLD1 colon tumors. These differences may help explain tumoral immunity and lead to predictors of response to immunotherapy.

3584 Background: Patients (pts) with microsatellite stable (MSS) metastatic colorectal cancer (mCRC) that progressed on ≥2 prior therapies have limited treatment options, with median OS ranging from 6-9 months (mo). In the dose-escalation phase of this first-in-human multicohort study (NCT02720068), the anti-lymphocyte activation gene (LAG)-3 antibody favezelimab (fave) was well tolerated alone and with pembrolizumab (pembro) across all dose levels (Lakhani, SITC, 2018, abstract O26). Here, we evaluate the safety and efficacy of fave alone or in combination with pembro in pts with advanced MSS CRC from the dose confirmation phase. Methods: Eligible pts with MSS PD-1/PD-L1-treatment-naïve mCRC that progressed on prior standard-of-care (3L+) were enrolled (cohort A) to receive the RP2D of 800 mg fave alone (Arm 1), 800 mg fave + 200 mg pembro (Arm 2C), or 800 mg fave + 200 mg pembro (MK-4280A) co-formulation (Arm 5), all Q3W. Treatment continued for 35 cycles or until progression, unacceptable toxicity, or investigator/pt decision. Pts with confirmed progression per irRECIST v1.1 on fave alone could crossover to 800 mg fave + pembro. Safety was assessed in all treated pts; efficacy in the full analysis set (FAS) of all treated pts with baseline scan. Objectives included safety (primary), ORR (RECIST v1.1 by investigator [secondary]), and DOR, PFS, and OS (exploratory). Interim analysis data cut-off was: Oct. 23, 2020. Results: A total of 20 pts received fave (Arm 1); 89 pts (including 9 crossover) received fave + pembro (Arms 2C+5); 12 pts (Arm 1) and 36 pts (Arms 2C+5), had PD-L1 CPS ≥1 tumors. At data cut-off, median follow-up was 5.8 months (mo) in Arm 1 and 6.2 mo in Arms 2C+5. Treatment-related adverse events (TRAEs) were 65% with fave (Arm 1) and 65.2% with fave + pembro (Arms 2C+5). Grade ≥3 TRAEs were 15% (Arm 1), and 20% [Arms 2C+5]). No grade 5 TRAEs were reported. Common TRAEs (≥15%) included fatigue (20.0%), nausea (15%) with fave, and fatigue (16.9%) with fave + pembro. Confirmed ORR was 6.3% (4PR, 1CR) with fave + pembro (Arms 2C+5). No pt receiving fave alone responded. In Arms 2C+5, median DOR was 10.6 mo (range, 5.6-12.7). ORR, OS and PFS by PD-L1 status are reported in the Table. Conclusions: Favezelimab alone or in combination with pembrolizumab had a manageable safety profile, with no treatment-related deaths. Promising antitumor activity was observed with combination therapy, including with MK-4280A, compared with monotherapy most notably in pts with PD-L1 CPS ≥1 tumors. Clinical trial information: NCT02720068. [Table: see text]

e15560 Background: PD-1 blockade is particularly ineffective in patients with microsatellite stable (MSS) or mismatch repair (MMR)-proficient colorectal cancer (CRC). Regorafenib (R) has been shown to modulate anti-tumor immunity through different mechanisms, including the reduction of tumor-associated macrophages (TAMs) and lymphocyte activation or immunosuppressive cells, such as forkhead box P3 (Foxp3)+ CD25+ regulatory T cells (Tregs). Synergy between R and anti–PD-1/PD-L1 antibodies has been demonstrated in pre-clinical models compared to that of either treatment alone. Accordingly, this study attempted to evaluate the combined activity of regorafenib with an immune checkpoint inhibitor. Methods: This trial was a prospective, open-label, monocentric, nonrandomized, single arm study. This study included patients with non MSI-H mCRC who received R (80 mg QD 3weeks/4) and sintilimab (S)(200 mg every 3 weeks).The primary endpoint was the confirmed objective response rate (ORR).The secondary endpoints were progression free survival (PFS),overall survival (OS),disease control rate (DCR) and safety. Results: As of January 8, 2021, 24 patients (median age 59 years) with mCRC were enrolled and received treatment with R+S. RAS wild type was 50% of all patients. In addition, 83.3% of patients received two prior lines of treatment, while 58.3% had liver metastases at enrollment. Of the 24 patients, the best response was observed to be stable disease(SD) in six patients (25%) and progressive disease (PD)in four patients (16.7%). Another nine (37.5%) patients have yet to be evaluated. In the 15 evaluated patients, the ORR(RECIST version 1.1) was 33.3% , the DCR was 73.3%. Additionally, among the 10 evaluated patients with liver metastases, the ORR was 30% , the DCR was 80%. Among the nine evaluated patients with wild-type populations of k-ras, the ORR and DCR was found to be 44.4% and 66.7%, which was higher than the mutation patients. The median PFS was 4.2 (95% CI, 2.5, NA) months, the median OS was not reached. The most common grade 3/4 adverse events were palmar-plantar erythro-dysesthesia syndrome (4.2%), erythra (4.2%). No death was related to the treatment. Moreover, four (16.7%) patients were subject to at least 1 dose modification or treatment interruption. Conclusions: In this study, the combination of R+S achieved a similar ORR with REGONIVO. Furthermore, this combination was well tolerated and had a manageable safety profile. However, due to the limited sample size, some biases may be present. As a result, we will continue to expand the sample size for future verification. Clinical trial information: NCT04745130.

5581 Background: Traditional platinum-based systemic chemotherapy continue to be the SOC for aEC in the first line. Phase 2 clinical trials of chemotherapy (GOG 129 series) and some targeted therapies (229 series) for second line advanced endometrial cancer (aEC) have proved disappointing. Recently the treatment landscape for aEC patients has significantly changed with newer targeted therapies focusing on the microsatellite instability (MSI) status of endometrial tumors. The objective of the ECHO study was to describe real-world treatment patterns and outcomes in non-MSI-high or DNA mismatch repair proficient (pMMR) aEC patients in clinical practice in the United States (US) prior to 2019. Methods: The ECHO study is a multicenter, retrospective chart review study in women diagnosed with aEC in the US. Data were obtained from medical records of adult women (≥18 years) diagnosed with advanced or inoperable aEC (stages III or IV) with known MSI status, who had received at least one prior systemic therapy and progressed between July 1, 2016 – June 30, 2019. De-identified patient data extracted by treating oncologists included patient demographics, clinical and treatment characteristics, and clinical outcomes. Kaplan-Meier analyses were performed to estimate real-world progression-free survival (rwPFS) and overall survival (OS). Results: A total of 124 non-MSI-high or pMMR aEC patients who had progression following first line therapy were included in this interim analysis. Average age was 63 years, 62.9% White/Caucasian, 16.9% Hispanic/Latino, and 86% had ECOG ≤1. Metastases were observed in 70% of patients at diagnosis, with the most common metastatic sites being lung (47.6%), liver (32.3%), and distant lymph nodes (29%). As 2nd line therapy, 69% of patients received mono or combination chemotherapy (primarily with doxorubicin), 13% hormonal therapy, and 18% targeted therapy ± chemotherapy. Median duration of 2nd line therapy was 4 months. The majority (86.3%) discontinued 2nd line therapy, with disease progression the most common reason (66.4%). A quarter (26.6%) of patients initiated an additional line of therapy. Median rwPFS from initiation of 2nd line therapy was 5 months (95% confidence interval [CI]: 4-9). Median OS from initiation of 2nd line therapy was 12 months (95%CI: 9-18). Estimated OS rates from initiation of 2nd line therapy at 6, 12, and 24 months were 66%, 47%, and 30%, respectively. Conclusions: In this retrospective, chart review study, patients with non-MSI-high/pMMR aEC in the US who failed at least one systemic therapy had poor prognosis on subsequent therapies. There continues to be a significant unmet need in this group of women. Novel therapies are needed that delay progression and/or improve overall survival and further research is indicated to explore this.

5062 Background: Loss-of-function genomic alterations (GAs) in MSH6 have been associated with a unique subtype of hypermutated mCRPC that is often microsatellite stable (MSS) and may occur in either a sporadic or familial Lynch Syndrome-like clinical setting. Methods: 5,617 mCRPC cases were sequenced to evaluate all classes of GA using a hybrid capture-based FDA-approved comprehensive genomic profiling (CGP) assay. Tumor mutational burden (TMB) was determined on 0.8 Mb of sequenced DNA and microsatellite instability high (MSI-High) was determined on 95 loci. MSI-low status was not assessed. Results: 78 (1.4%) mCRPC were MSH6mut (Table). MSH6mut mCRPC included 73.1% short variant mutations, 23.1% biallelic deletions, 2.6% genomic rearrangements, and 1.3% multiple GAs/sample. Co-mutation of MSH2 was found in 28% of MSH6mut cases vs. 2% in MSH6wt cases (P mut mCRPC, which was significantly greater than the 2% seen in MSH6wt cases (P mut cases, compared to 3% MSH6wt cases (P mut cases with neither MSI-High nor MMR mutational signature, 87% did not have biallelic loss of MSH6 or any other MMR gene, confirming that monoallelic pathogenic mutations are insufficient to cause the MMR-D phenotype. For subjects whose variants could be classified, 45% (19/42) of pathogenic MSH6 alleles were germline; of these, 58% (11/19) had neither MSI-High nor an MMR single nucleotide signature. MSH6mut cases had fewer TMPRSS2: ERG fusions (P =.01), but harbored significantly higher frequencies of GAs in AR (P =.0002), ATM (P =.04), PIK3CA (P =.0003), APC (P =.005), ERBB2 (P =.001), and CDK6 (p =.046), likely at least partially attributable to the higher TMB in MSH6mutcases (P mut mCRPC is a unique disease that often features a hypermutated genomic signature, although only 46% of cases exhibited MSI-high status. This complex phenotype highlights the potential utility of multiple rather than single biomarkers to understand tumor biology and determine patients who may benefit from immunotherapy.[Table: see text]

e15585 Background: A previous phase 1b trial has shown encouraging efficacy of regorafenib plus nivolumab in patients with microsatellite stable/mismatch repair proficient (MSS/pMMR) metastatic colorectal cancer (mCRC). We aimed to evaluate the efficacy and safety of this regimen in Chinese patients in the real world. Methods: We retrospectively identified patients with MSS/pMMR mCRC who received at least one dose of programmed cell death-1 (PD-1) inhibitors plus regorafenib from 5/2019 to 2/2021 in 10 Chinese medical centers. The primary endpoint was overall survival (OS). Secondary endpoints included progression-free survival (PFS), objective response rate (ORR), disease control rate (DCR) and the safety. Results: Fifty-two patients were identified. Liver metastases were presented in 35 patients (67%). A total of 48 patients (92%) received regorafenib plus a PD-1 inhibitor as the third or later line treatment. At the data cut-off, 11 patients (21%) were still on treatment. Other patients terminated treatment because of progressive disease (45%), treatment-related adverse events (TRAEs) (14%) or treatment-unrelated deaths (6%). The median treatment cycle was 3 (range, 1-18). At a median follow-up of 4.9 months, the median OS was 17.3 months (95% CI, 10.2-NR) and the median PFS was 3.1 months (95%CI, 2.5-6.0). Baseline liver metastases were associated with inferior PFS (2.7 versus 6.3 months, p

3573 Background: Microsatellite stable metastatic colorectal cancer (MSS mCRC) represents a high unmet need since there are currently no approved immunotherapy options. Since the inhibition of the transforming growth factor-β (TGF-β) pathway is known to contribute to the enhancement of immunotherapy efficacy, here, we report the results of vactosertib, a potent and selective TGF-β receptor I kinase inhibitor, combined with pembrolizumab in patients with MSS mCRC. Methods: In this phase 2, open label trial, patients have received vactosertib (300 mg BID, 5 days on / 2 days off) and pembrolizumab (200 mg, every 3 weeks) until confirmed progressive disease (PD), unacceptable toxicity or consent withdrawal. Patients with histologically confirmed mCRC who had disease progression after treatment with all available therapies including fluoropyrimidine and oxaliplatin or irinotecan were enrolled. Eligible patients were ≥19 years old, had ECOG status ≤1, and had no prior exposure to immunotherapy. The objectives were to evaluate the safety and efficacy (objective response rate (ORR) per RECIST v1.1). Results: Thirty-three patients with MSS mCRC were enrolled. Median age was 60 (range 33-72), 55% were male, median number of previous lines of chemotherapy was 3 (range 1-7), and 82% were consensus molecular subtype (CMS) 4. The ORR was 15.2% including 5 partial responses (PRs), 7 stable diseases, and 17 PDs as best overall responses; 12 patients remain on treatment. Among 5 patients with PR, 3 patients had confirmed PR and median duration of response was not reached yet. As of 04 Jan 2021, the most common treatment related adverse events (AEs) were increased amylase (21.2%), pruritus (21.2%), rash (21.2%), and increased lipase (18.2%). There were 3 treatment-related SAEs reported; drug induced pneumonitis (3%), nausea (3%), and vomiting (3%). Conclusions: The combination treatment with vactosertib and pembrolizumab showed favorable safety profile with promising efficacy in patients with MSS mCRC. Updated data including pharmacodynamic markers will be presented at the meeting. Clinical trial information: NCT03724851.

3561 Background: BRAF V600E+, MSS mCRC patients comprise up to 10% of advanced CRC. They have a poor prognosis with median survivals typically

3559 Background: The MORPHEUS platform consists of multiple, global, open-label, randomized Phase Ib/II trials designed to identify early efficacy and safety signals of treatment (tx) combinations across cancers. Here, atezo (anti-PD-L1) was tested with Imprime and bev (anti-VEGF) for MSS mCRC, a poorly immunogenic cancer generally resistant to checkpoint inhibitors. Imprime acts as a pathogen-associated molecular pattern that, when bound to anti-β glucan antibodies (ABA), activates the innate immune system with the potential to 1) promote priming and expansion of tumor-specific T cells, 2) promote M2-M1 macrophage polarization and 3) enhance the immunomodulatory effects of atezo and bev. Therefore, we hypothesized that atezo + Imprime + bev would induce an antitumor response beyond that of rego, a standard-of-care multikinase inhibitor, in patients (pts) with MSS mCRC. Methods: Pts with MSS mCRC unselected for the Imprime-specific biomarker (ABA) and refractory to 1-2 prior lines of standard therapy received atezo (1200 mg IV every 3 weeks [q3w]) + Imprime (4 mg/kg IV on Days 1, 8, 15) + bev (7.5 mg/kg IV q3w) or control tx with rego (160 mg orally days 1-21; dose escalation to 160 mg during Cycle 1 allowed per institutional guidelines). The primary endpoint was objective response rate (ORR; investigator-assessed RECIST 1.1); secondary endpoints included disease control rate (DCR; response or stable disease ≥ 12 weeks), progression-free survival (PFS), overall survival (OS) and safety. Results: Pts were followed-up for ≥18 wk. 15 pts received atezo + Imprime + bev and 13 received rego. Grade (Gr) 3/4 tx-related adverse events (TRAEs) were seen in 13% of atezo + Imprime + bev and 62% of rego pts. No Gr 5 AEs occurred in atezo + Imprime + bev pts and 1 (8%) was reported in a rego pt. One pt in each arm (7% vs 8%, respectively) withdrew from tx due to a TRAE. No radiological responses were seen in either arm. Five pts (33%) receiving atezo + Imprime + bev and 8 (62%) receiving rego had stable disease as best response. DCR was 13% with atezo + Imprime + bev and 23% with rego. Median PFS was 1.5 mo (95% CI: 1.4, 2.8) and 2.8 mo (95% CI: 1.6, 3.1), and median OS was 5.7 mo (95% CI: 4.4, 10.5) and 10.2 mo (95% CI: 4.8, NE) with atezo + Imprime + bev and rego, respectively. There was no apparent correlation between baseline PD-L1 expression or CD8+ lymphocyte tumor infiltration and clinical benefit. Further, the systemic exposure of atezo, Imprime and bev and immunogenicity of atezo and bev are in line with previous clinical experience. Additional biomarker, pharmacokinetics and anti-drug antibody data will be shown. Conclusions: Atezo + Imprime + bev was well tolerated; toxicities were consistent with the safety profiles of the individual agents. No efficacy signal was identified with atezo + Imprime + bev in pts with MSS refractory mCRC. Clinical trial information: NCT03555149.

TPS5611 Background: Folate receptor-alpha (FRα) is expressed on endometrial cancer (EC) cells and is associated with poor prognosis. Mirvetuximab soravtansine (ImmunoGen), an antibody drug conjugate (ADC) comprising a FRα-binding antibody, cleavable linker, and the tubulin-disrupting maytansinoid DM4, showed tolerability and single agent activity in a Phase 1 study with dose expansion in FRα+ advanced/recurrent EC (NCT01609556) and also when combined with chemotherapy, bevacizumab as well as pembrolizumab (NCT02606305). In addition to having direct target-mediated cytotoxicity, ADCs also stimulate the local tumor immune microenvironment. Mirvetuximab soravtansine has been shown to activate monocytes and promote phagocytosis of mirvetuximab-treated FRα-positive tumor cells through a mechanism of Fc-FcγR interaction. Further, the combination of ADCs with immune checkpoint inhibitors (ICI) can overcome primary resistance to immunotherapy in murine models. Given the low response of MSS endometrial cancers to PD-1 blockade, we hypothesized that addition of mirvetuximab may enhance response of these tumors to immunotherapy. Methods: This is a Phase 2, single cohort study of mirvetuximab soravtansine with pembrolizumab in recurrent or persistent EC. Patients must have advanced or recurrent MSS serous endometrial cancer with at least 1 and up to 3 prior lines of therapy. Confirmation of FRα expression (with PS2+ staining intensity in ≥ 50% of cells, performed centrally at Ventana Medical Systems, Inc) is required. Prior receipt of ICI is excluded. Patients will receive the combination of mirvetuximab soravtansine 6 mg/kg AIBW IV and pembrolizumab 200 mg IV administered every 21 days. The co-primary endpoint is progression-free survival at 6 months (PFS6) and objective response rate (ORR) by RECIST 1.1. Translational objectives include assessment of tumor infiltrating lymphocytes (TILs), expression of immune checkpoint markers, and whole exome sequencing (WES) for DNA repair pathway mutations, neoantigens, and polymorphisms in immunologically relevant genes. Statistical considerations are for a Simon two-stage optimal design with 16 patients in Stage 1 and 19 patients in Stage 2, to a total of 35. Prespecified activity for the first stage of accrual was met, and second stage accrual began November 2020. Clinical trial information: NCT03835819.

3560 Background: The role of immunotherapy in the treatment of pMMR/MSS metastatic CRC is not established. A Japanese phase 1b trial in this setting showed the combination of regorafenib (multikinase inhibitor with immunomodulatory activity) plus nivolumab (anti PD-1) had encouraging activity and manageable safety (Fukuoka, 2020). This study further assessed the safety and efficacy of this combination. Methods: Patients (pts) from the US aged ≥18 years who progressed on/were intolerant to standard chemotherapy were enrolled. Regorafenib was given orally, once daily in 28-day (D) cycles (21D on/7D off) plus IV nivolumab 480 mg on D1. Regorafenib starting dose was 80 mg; if well tolerated, it could be escalated to 120 mg in Cycle 2. Primary endpoint was overall response rate (ORR; RECIST 1.1); secondary aims included disease control rate (DCR), overall survival (OS), progression-free survival (PFS), and safety (NCI-CTCAE v5.0 grade). Biomarker analysis was exploratory. Results: 70 pts (59% male) started treatment. At baseline, median age was 57 years (range 34–85), ECOG PS 0/1 was 51%/49%, 67% had liver metastases (mets), and the primary tumor site was right-sided colon in 36% and rectum in 17%. Median number of cycles was 3.0 (range 1–13); 41% of pts escalated regorafenib to 120 mg. Five pts (7.1%) had a partial response (PR) lasting ≥16 weeks (wks) and 22 (31.4%) had stable disease (SD); pts without liver mets had a higher ORR (21.7%). In pts with tumor samples (n = 40), higher baseline expression (IHC) of cytotoxic T cells (CD3+/CD8+/GranzymeB+), Tregs (FoxP3+), and macrophages (CD68+) trended with clinical benefit (PR/SD ≥16 wks/PFS); pts with liver mets had lower expression. Lower plasma levels of biomarkers of vascular biology (e.g. VEGF-D, Ang-2, VWF) trended with longer PFS. Grade (Gr) 3 treatment-emergent adverse events (TEAEs) occurred in 53% of pts and Gr 4 in 10%. Three pts had a Gr 5 TEAE: n = 1 related to the combination (sepsis); n = 1 related to nivolumab only by investigator (sepsis); n = 1 unrelated to treatment (respiratory failure). Most common Gr 3/4 TEAEs: maculopapular rash (14%), fatigue (7%), pneumonia (6%), increased bilirubin (6%). Conclusions: Combination treatment with regorafenib (up to 120 mg/day) and nivolumab (480 mg every 28D) has manageable safety. Efficacy of this combination in the North American population did not emulate results in the Japanese population. Absence of liver mets and expression of specific biomarkers indicate a better response and may warrant further analysis. Clinical trial information: NCT04126733. [Table: see text]

124 Background: Microsatellite stable (MSS) colorectal cancer and CRCLM are highly resistant to current immunotherapeutic approaches. Understanding the TME and the impact of standard chemotherapeutics is of utmost importance in developing optimal therapeutic strategies. Data is limited on the impact of NACT in the TME. In this study, we sought to explore the effects of chemotherapy on the proportions of different immune cell populations [Total leukohematopoetic cells, CD45+, T cell markers, CD3+, CD8+, and macrophage (M) markers, CD163+, CD68+)], as well as to assess their prognostic capacity. Methods: CRCLM pts who underwent liver resection were divided in 2 groups: pts who had NACT within 4 months of liver resection [NACT (+)] and no NACT [NACT (-)]. Whole slide staining of metastatectomy specimen for CD3, CD8, CD68, CD163, and CD45 was done. Cellular density within representative tumor core sections was analyzed. Retrospective chart review was done to obtain clinical data. Disease free (DFS), relapse free (RFS), and overall survival (OS) were analyzed using Kaplan Meir methods. Association between markers and outcomes was evaluated using Cox Regression Models. Results: A total of 43 pts were studied [NACT (+), n= 14; NACT (-), n=29). There were no significant differences in baseline characteristics between the two groups, including age, primary site, T stage, N stage, and grade. Median OS was 48 months. Outcomes were not significantly different with use of NACT. Within the NACT (+) CRCLM significant increases were seen in densities of CD3, CD8, CD45 and CD163 cells. Densities of CD8 cells were strongly correlated to CD163 cell densities ([r] 0.77, p < .0001). The NACT (+) cohort saw significantly increased ratios of T-cells/macrophages as compared to NACT (-) (CD3/CD68, CD3/CD163, CD8/CD163). High CD45 density was associated with improved OS (HR = 0.28, p = 0.006). CD3, CD8, CD68, and CD163 were not independently associated with OS. However, both the ratio of CD3/CD68 and CD3/163 were associated with improved OS (HR 0.44, p=0.03 and HR 0.27, p=0.03 respectively). Conversely, CD68/45 and CD163/CD45 ratios were associated with a worse OS (HR 1.42, p=0.05 and HR 1.43, p=0.03 respectively). No difference was seen in cellular populations based on the type of NACT received: oxaliplatin vs irinotecan. Conclusions: Our study highlights that the administration of NACT is associated with favorable changes in LM TME. While there is a general increase in leukocytes (CD45), the density of CD3+ and CD8+ T cells is particularly increased. Further, this increase is proportionately greater than the concurrent increases on monocytoid populations, reflected by increased CD3/CD163 and CD8/CD163 ratios. Ratios of T-cells/ macrophages are predictive of survival. This study was partly supported by the American Cancer Society Grant, 126771-IRG-14-194-11-IRG.

72 Background: Microsatellite stable (MSS) metastatic colorectal cancer (mCRC) has been associated with immunotherapy resistance. However, the impact of metastatic diseases sites on clinical outcome with checkpoint inhibition has not be adequately investigated. Methods: Following IRB approval (14361), we performed a retrospective study to assess the response and progression-free survival (PFS) to PD-1 or PD-L1 based therapy in patients with MSS mCRC who progressed following standard chemotherapy and targeted therapy. Patients were included if they received an FDA approved anti-PD1 or anti-PD-L1 agent, alone or in combination with other investigational agents. Exclusion criteria included concurrent cytotoxic therapy. Response and PFS were determined by RECIST guidelines. Results: 97 patients with refractory MSS mCRC satisfied the inclusion criteria were evaluable for analysis. 57, 17, 13, and 10 received nivolumab, atezolizumab, pembrolizumab, and durvalumab, respectively. 47, 8, 10, 10, 17 received concurrent VEGFR, MEK, CTLA-4, radiation, or other targeted therapies. Among tested variables, including age, gender, primary tumor location, ECOG status, RAS, BRAF, APC, TP53, TMB, and sites of metastasis, only ECOG status (0 vs. 1) and metastatic disease to the liver were associated with disease control (partial response and stable disease; p = 0.005 and < 0.001 respectively). Disease control rates in patients without liver metastases (n = 43) was 56% (95% CI: 40-71%), compared with 2% (95% CI: 0.1-10%) in patients with liver metastases (n = 54) (p < 0.001). Also, liver involvement was predictive for PFS based on multivariate Cox regression model (p < 0.001). The median PFS in patients with liver involvement was 1.5 vs 4.5 months for no liver involvement (HR = 4.41, p = < 0.001). Amongst patients without liver metastases, 35% (15/43) remained without progression at 6 months. These patients were characterized by lung metastases (n = 8) and/or lymph node metastases (n = 4). Conclusions: The presence of liver metastases is MSS mCRC is predictive for lack of benefit from PD-1/PD-L1 inhibitors. Patients without liver metastases can derive a meaningful clinical benefit, with 35% being progression-free at 6 months. Future PD-1/PD-L1 drug development should consider metastatic sites of disease in study design and development.

26 Background: Microsatellite instable (MSI) metastatic colorectal cancer (CRC) tumors have been found to be responsive to immune checkpoint inhibitors. Most MSI tumors are in consensus molecular subtype CMS1. Here, in a search for microsatellite stable (MSS) samples that may respond to checkpoint inhibitors, we used the ColoType CMS1-score to identify microsatellite stable (MSS) CRC tumors that have immune activity profiles similar to MSI tumors. Methods: This study was performed in Affymetrix cohort (n = 1,888) of primary CRC tumor samples assayed with hgu133plus2 Affymetrix microarrays. Degrees of infiltration of populations of immune cells were assessed with MCPcounter. Gene Set Enrichment Analysis (GSEA) using the Hallmark gene sets was used to compare biological features between subsets of samples. An 18-gene genomic score (PembroSig) of Ayers, et al, was used to predict response to Pembrolizumab. ColoType 40-gene signature identifies the CMS of a CRC tumor using values of continuous scores, one score for each of CMS1-4. CMS1+ denotes the samples with CMS1-score sufficiently high to predict it is in CMS1. A genomic score was used to predict MSI status for all samples in Affymetrix cohort; MSI predictions using said score had 97% agreement with clinically determined MSI status for 749 Affymetrix cohort samples. Results: The MSI score predicted 240 MSI samples in Affymetrix cohort, and 211 MSS samples in CMS1+ (CMS1+/MSS). The union of MSI and CMS1+/MSS compared to other samples exhibited elevated infiltration of cytotoxic lymphocytes (p < 0.0001) and CD8 T-cells (p < 0.0001), activity of interferon-gamma (p = 0.006) and interferon-alpha signaling (p = 0.012), inflammatory response (p = 0.012), and elevated PembroSig score (p < 0.0001). CMS1+/MSS compared to MSI showed no significance difference in any of these measures. None of the immune regulatory genes PD-1, PD-L1, PD-L2, CTLA4, LAG3, and IDO1 were differentially expressed between CMS1+/MSS and MSI. Of note, the only Hallmark gene sets enriched in MSI compared to CMS1+/MSS were DNA repair (p = 0.002) and MYC targets V1 (p = 0.009) and V2 (p = 0.013). We further analyzed the biological differences between MSI samples in CMS1+ (CMS1+/MSI, n = 190) and MSI samples not in CMS1+ (CMS1-/MSI, n = 50). We found that CMS1+/MSI compared to CMS1-/MSI exhibited elevated cytotoxic lymphocyte infiltration (p < 0.0001), interferon-gamma activity (p = 0.027), and PembroSig score (p < 0.0001), highlighting the relationship between immune activity and CMS1-score. Conclusions: ColoType CMS1-score identified MSS samples in Affymetrix cohort (11%) with similar immune activity as MSI samples (13%) and equivalently high values of a genomic score predictive of Pembrolizumab response.

77 Background: MSS mCRC rarely responds to pembrolizumab monotherapy, but capecitabine and bevacizumab may induce immune-stimulatory effects. This study evaluates the safety, tolerability and preliminary efficacy of pembrolizumab in combination with capecitabine and bevacizumab in MSS mCRC. We present results at the planned interim analysis. Methods: Design:single-arm, open-label, single-site phase 2 trial with a safety lead-in to confirm the recommended phase 2 dose (RP2D) for capecitabine and expansion cohorts. Per the Simon’s 2-stage design, ≤1 response in 29 patients (pts) requires trial suspension. Key eligibility criteria: MSS mCRC with stable disease (SD) or progressive disease (PD) on prior fluoropyrimidine-based therapy. Treatment: RP2D PO capecitabine on days 1-14 plus 200 mg IV pembrolizumab and 7.5 mg/kg IV bevacizumab on day 1 in 21-day cycles. Pts are followed for toxicity and radiographic response. Results: From 04/2018-09/2020, 29 pts were enrolled, of whom 15 (52%) were female; 21 (72%) white; and median age was 55 years (range 36-77 years). Prior therapies: 2 (7%) pts had SD and 27 (93%) pts had PD on fluoropyrimidine-containing regimens; 24 (83%) pts had prior exposure to bevacizumab. The RP2D for capecitabine was 1000 mg/m2 PO BID, with no dose limiting toxicities observed. Complete toxicity data are available for 25 off-treatment pts. The most common related adverse events (AEs) were palmar-plantar erythrodysesthesia (PPE) (64%) and fatigue (68%). Grade ≥3 related AEs occurred in 9 (36%) pts, including immune-related AEs of Grade 3 dyspnea, hypophosphatemia, and pancreatitis in 1 pt each. Treatment related AEs leading to dose interruptions, reductions, or delays occurred in 15 (60%) pts, most commonly PPE in 13 (52%) pts. No pt had a related AE leading to treatment discontinuation or death. Disposition: of 29 pts enrolled, 24 were removed for PD and 1 was removed for an unrelated AE. Best response by RECIST 1.1 in 23 evaluable pts: partial response (PR) in 2 (9%); SD in 14 (61%); PD in 7 (30%). Median time on treatment was 6 months (range 2-26 months). Conclusions: Combination of pembrolizumab with capecitabine and bevacizumab was found to be tolerable with an expected toxicity profile in MSS mCRC pts. With 2 responses, the study met interim analysis criteria to continue accrual. Tissue and blood-based immune correlatives are planned. Clinical trial information: NCT03396926.

95 Background: Preclinical data suggests metformin can improve immune exhaustion of tumor infiltrating lymphocytes and potentiate the effects of PD-1 blockade. By normalizing the hypoxic TME, metformin was shown to improve cytotoxic T cell function and efficacy of anti-PD-1 antibody in highly aggressive B16 melanoma and MC38 colon adenocarcinoma tumor models. Based on this preclinical rationale we conducted a phase II study with nivolumab and metformin combination in treatment refractory MSS metastatic colorectal cancer (mCRC). Methods: Nivolumab 480 mg IV every 4 weeks and Metformin 1000 mg po twice daily was administered in 28-day cycles following a 14-day metformin only lead-in phase.Eligible patients included stage IV metastatic treatment refractory MSS mCRC (patients must have received oxaliplatin, irinotecan, and fluoropyrimidine), age ≥18 years, ECOG PS 0-1, adequate organ function, no prior anti PD-1 agent. The primary endpoint was overall response rate (ORR). Secondary endpoints included overall survival (OS) and progression free survival (PFS). Simon’s two-stage Minimax design was employed (H0: ORR =4%; H1: ORR=15%; alpha = 0.1; power =80%). If ≥1 objective response was observed in the first evaluable 18 patients, 10 additional patients would be included in the cohort. ≥3 objective responders in 28 patients would be required to be considered positive study. Pre-treatment and on-treatment research biopsies and correlative peripheral blood specimens were collected. Results: A total of 24 patients were enrolled, 6 patients were replaced per protocol, and 18 patients had evaluable disease. Of the 18 evaluable patients 11/18 (61%) were female, median age 58 [IQR 50-67]. 2 patients had prolonged stable disease (4 and 10 cycles). No patients had objective response based on RECIST 1.1. Median OS and PFS was 5.1 months [95% CI (2-11.7)] and 2.3 months [95% CI (1.7-2.4)], respectively. Most common grade 3 and 4 toxicities were anemia (n=2) and diarrhea (n=2). Conclusions: In treatment refractory MSS mCRCnivolumab and metformin combination was well tolerated. Two patients achieved stable disease, but no objective response was seen; therefore, the study did not proceed with the second stage of enrollment. Immunologic correlative analysis of this study is ongoing. Clinical trial information: NCT03800602.

126 Background: In microsatellite stable (MSS) colorectal cancer (CRC) cases, frequent amplifications in 13q and 20q regions have been reported with limited understanding of their potential role for prognosis and treatment. Based on an internal research molecular tumor board case analysis of a KRAS WT, MSS left-sided (LS), metastatic CRC patient, we noted 50% of the 11 listed variants of unknown significance were on 20q. This study aims to investigate the potential prognostic role for gain of 20q chromosomal regions relative to other chromosomal imbalances using a retrospective cohort. Methods: We performed a genome-wide CN cluster analysis using the TCGA data with resulting patient clusters compared based on Kaplan-Meier estimated disease-free interval (DFI) in months (mos) using a log-rank. We examined associations between CN-derived clusters and CRC phenotypes. Results: Using TCGA data, we identified four CN-derived patient clusters of which two had significantly different DFI’s. One longer DFI cluster (n = 83; median DFI = 28mos) was defined by a signature of copy neutral genes along Chr8p and 18q regions. The shorter DFI cluster (n = 46; median DFI = 20mos) was defined by a signature of Chr8p and 18q regions of loss. The two clusters significantly differed in terms of site (85% right-sided in the shorter vs. 85% LS in the longer DFI cluster) and MS status (81% MSS in the shorter vs. 72% MSI in the longer DFI cluster) with no significant differences in age (mean of 70 years at diagnosis) and KRAS status (around 50% WT) between clusters. The shorter DFI was able to be further differentiated based on gain of chromosomal regions 20q (n = 26; median DFI = 24mos) and 13q (n = 20; median DFI = 18mos). No significant differences in mean age, tumor stage, site, and MS status between the shorter DFI subgroups were noted. Enriched pathways in the Chr20q gain sub-cluster include cell differentiation (Hippo, Rap1, VEGF), whereas metabolic signaling pathways via AGE-RAGE and AMPK, as well as cell growth and malignant transformation via RAS are enriched in the Chr13q gain sub-cluster. Conclusions: Clinical sequencing is able to identify regions of gain in Chr20q and Chr13q as part of VUS’s. Gain of Chr20q has been reported with better survival in mCRC patients. Gain of chromosomal region 20q and loss of 18q has been reported to discriminate between Lynch Syndrome and familial colorectal cancer. Herein, we show that gain of the 20q region with 18q and 8p regions of loss is associated with shorter DFI and gain of 13q region with loss of 8p even worse prognosis.

4100 Background: Chemoradiotherapy (CRT) followed by radical surgery (S) is standard therapy for patients (pts) with locally advanced rectal cancer (LARC). Sequential use of an anti-PD-1 antibody after radiation demonstrates synergistic effects in in vivo models, and an anti-PD-1 antibody is effective in pts with microsatellite instability-high (MSI-H) metastatic colorectal cancer (mCRC). We studied nivolumab (nivo) and radical S following CRT (50.4 Gy with capecitabine 1,650 mg/m2) in T3–4 NanyM0 LARC. Methods: After the quality-assured CRT, 240 mg q2 weeks x 5 cycles of nivo and radical S were investigated. In cohort A-1, for pts with microsatellite stable (MSS) LARC, the primary endpoint was a centrally confirmed pathological complete response (pCR) rate using AJCC tumor regression grading. The estimated required sample size assuming null and alternative hypotheses pCR = 10% and 30% was 37 pts, with a 1-sided alpha of 5% and power of 90%. In Cohort A-2, 5 pts with MSI-H LARC were included in an exploratory manner. Results: From Jan/2017 to Oct/2019, a targeted number of pts was included and assessed. In cohort A-1, 30% (11/37; 90% CI 18-44%) of pCR (AJCC grade (gr) 0) rate and 38% (14/37) of major pathological response (MPR) (AJCC gr 0+1) rate were observed. Clinical CR was observed in one additional pt (3%) refusing S after nivo. In cohort A-2, 60% (3/5) of pCR rate and 60% (3/5) of MPR rate were observed. As of Jan/2020, only 2 pts (1 local and 1 metastatic) in cohort A-1 and none in cohort A-2 recurred. Immune-related severe adverse events were observed in 3 pts (gr 3 myasthenia, gr 3 interstitial nephritis, and gr 2 peripheral motor neuropathy); all fully recovered and received radical S. During the follow-up period, one additional pt with gr 2 colitis was observed. No treatment-related deaths were observed. Conclusions: Promising pCR rates of 30% and 60%, with mild toxicities, were shown in MSS and MSI-H LARC pts treated with nivo plus radical S after CRT, suggesting the candidate therapy for the future non-surgical approach. Clinical trial information: NCT02948348 .

3006 Background: Single agent PD-1/L1 inhibition are not effective in metastatic colorectal cancer (MCRC) with microsatellite stable tumors. However, signal of efficacy was shown using combo therapy of anti-PD-L1 and anti CTLA-4 in multitreated patients. FOLFOX regimen could induced immunogenic cell death and elimination of myeloid derived suppressor cells (MDSCs) thus leading to a potential positive effect on antitumor immune response. Methods: This is a single arm exploratory investigator-initiated trial planned to include 57 pts to receive mFOLFOX6 (6cycles) in combination with durvalumab (150mg/q2W) and tremelimumab (75mg/q4W). After 6 cycles of chemotherapy, patients are treated with durvalumab untils progression. Primary endpoint is 6 months’ progression-free survival rate. Secondary endpoints are response rate, tolerability and translational research evaluating tissue and blood immune parameter. Upon Simon’s design an efficacy intermediate analysis was planned after the 16th patient has passed 6 months. Results: As of 1st of January 2020 the 55/57 pts were enrolled and treated with the MEDETREME regimen at 8 French sites. The intermediate efficacy analysis was conducted on the 1st of January after a median of 13.4 months of treatment. The following adverse events were noted: asthenia (81.25%), neuropathy (87.5%), diarrhea (56.25%) and neutropenia (62.5%). Notable grade 3/4 (CTC AE 4.03) include asthenia (18.75%) diarrhea (12.5%) neutropenia (50%) and elevated blood pressure (25%). Most of adverse events were related to chemotherapy. It has been noted one cytolysis grade 3, one thyroid dysfunction grade 3 and one hypophysitis grade 3 related to immunotherapy. Median PFS was not reached. Progression free survival at 6 months was observed in 10/16 pts (62,5%) given 5 CR, 5 PR and 4 SD. Updated translational data will be presented at the meeting. Conclusions: The interim safety analysis has supported safety and efficacy of the MEDITREME regimen in first- line MCRC. Finally, results will be presented after maturation of follow up. Clinical trial information: NCT03202758 .

3010 Background: Checkpoint inhibition using PD-1/PD-L1 inhibitors does not show clinically relevant activity in MSS/pMMR (Mismatch Repair Proficient) colorectal cancer. Previous work showed that inhibition of CCR5 (C-C chemokine receptor type 5) leads to a macrophage re-polarization towards M1 macrophages within the tumor microenvironment which directly affects immune cell infiltrates. The current phase I trial explores a combined modification of the innate immune system (by CCR 5 blockade) and the adaptive immune system (by PD-1 inhibition) in the treatment of MSS CRC. Methods: 20 patients with metastatic MSS/pMMR colorectal cancer with failure of fluoropyrimidines, oxaliplatin, irinotecan, VEGF antibodies and EGFR antibodies (in ras WT patients) received pembrolizumab 200 mg q21d and maraviroc 300 mg bid cont. for 8 cycles, followed by pembrolizumab monotherapy for a maximum of 24 additional cycles. Imaging was performed every nine weeks (RECIST and irRECIST criteria). Primary endpoint was the feasibility rate (rate of patients receiving the protocol treatment during the core treatment without special event: treatment-related Grade ≥ 3 immune-related abnormalities, treatment-related Grade ≥ 4 AEs or any toxicity-related premature withdrawal of treatment). Secondary endpoints included safety/toxicity, ORR, PFS and OS. Results: 20 patients were enrolled. The median number of applied cycles was 3.5 for pembrolizumab and 3.5 for maraviroc. Two patients completed the core treatment period with pembrolizumab and started maintenance treatment. The feasibility rate was 94.7% (90% CI 77.4 to 99.7%), with one patient experiencing a special event. Except this grade 4 event (hyperglycemia) no ≥ 3 treatment-related toxicities were observed. According to irRECIST criteria one patient showed a partial response and one a stable disease as best response, resulting in an irDCR of 10.5%. Median PFS according to irRECIST was 2 months (CI 95%, 2 to 3), median OS 9 months (CI 95%, 6 to 20). Conclusions: Therapy with pembrolizumab and maraviroc was feasible and showed a beneficial toxicity pattern. Clinical activity in MSS CRC patients was limited, however prolonged disease stabilizations were observed in single patients and overall survival was higher than expected in this heavily pretreated population. Clinical trial information: NCT03274804 .

4076 Background: Patients with microsatellite instability-high (MSI-H) colonic adenocarcinoma (COAD) are always immunotherapy-sensitive for the use immune checkpoint inhibitors, however, the vast majority of COAD patients (85%) are microsatellite stable (MSS). Homologous recombination deficiency (HRD) is demonstrated to be a response predictor to immunotherapies in gynecologic cancers, while limited studies were reported in colon cancer. We focus herein on the mutational pattern of HRR related genes in a large Chinese COAD cohort and further analyze the relationship between HRR-gene mutations and clinical response to immunotherapy in MSS COAD. Methods: The genomic profiling of Genecast cohort which is consisted of 406 Chinese patients with COAD were analyzed in a panel of 543 cancer related genes via next generation sequencing (NGS). The correlation between HRR-gene mutations and tumor immunity or clinical outcome using two COAD genomics datasets (TCGA and MSK-COAD) by the bioinformatic approach. Results: In Genecast Cohort, seventy of 406 (17.2%) patients were identified genomic alterations in HRR-gene, the most frequently mutated genes were ATM (9%), BRCA2 (4%), ATR (3%), RAD50 (3%) and BRIP1 (3%). In MSK-COAD cohort (treated with immune checkpoint inhibitors), HRR-mut group (n = 34) had a significantly better OS than HRR-wt group (n = 50) (log-rank test, P = 0.0087). From the analysis of TCGA cohort, we found that mutations of HRR-gene could increase immune activity in MSS COAD, including increased cytotoxic cells infiltration ( P = 0.035), increased exhausted CD8 T cells infiltration ( P = 0.0098) and higher IFN-g score ( P = 0.03). In contrary, similar results were not found in MSI-H COAD (all P > 0.05). Conclusions: Mutations of HRR-gene could significantly increase immune activity in patients with MSS COAD, implying the feasibility of using HRR-mut as a response predictor of immunotherapy in MSS-COAD.

152 Background: Monotherapy with immune checkpoint blockade (ICB) is ineffective for patients (pts) with MSS mCRC. Novel approaches to modulate the tumor microenvironment (TME) are needed. Here, we investigate whether the combination of trametinib (T) with durvalumab (D) can alter the immune TME by successfully priming and activating T-cells. Methods: An open-label, single center phase II trial with primary endpoint of immune-related response rate for T+D in refractory MSS mCRC pts (NCT03428126). T is 2mg/day orally starting 1 week prior to D, which is given 1500mg intravenously every 4 weeks. Dose de-escalation strategy performed to identify maximum tolerated dose (MTD). Simon 2-stage design utilized with plans to enroll 29 pts into the first stage, requiring response in 2 or more pts to proceed to stage 2 (n = 15). Results: Demographics for 29 treated pts: 48% female, median age 48 years (range 28-75), and median prior therapies was 2 (range 1-5). No grade (G) 4 treatment-related adverse events (TRAE). The most common G3 TRAE included autoimmune hepatitis (14%) and acneiform rash (10%). G1/2 TRAE included acneiform rash (69%), fatigue (24%) and anemia (21%). No fatal TRAE and 4 pts discontinued treatment due to TRAE. 1 of 29 pts had confirmed partial response (PR) lasting 9.3 months (mo) for an ORR of 3.4%. This pt had an ATM E221fs*14 mutation. 7 pts had stable disease (SD) with median time to progression (TTP) of 5.4 mo (range 3.9-9.3 mo). 1 pt remains on active therapy with SD ( > 10 mo). 5 pts (1 PR, 4 SD) demonstrated decrease in total CEA ng/mL (best percentage reduction: 94%, 95%, 42%, 34% and 21.6% respectively). Median TTP for the entire cohort was 3.2 mo (range 1.1-9.3 mo). Consensus molecular subtypes (CMS) were performed on the primary CRC in 23 pts: 12 CMS2, 2 CMS3, and 9 CMS4. 4 SD pts were CMS2, 1 SD pt was CMS4 and the CMS status of the pt with a PR was unknown. Conclusions: The combination of T+D did not meet efficacy criteria to proceed to the second stage of the study. Analysis of 15 paired on-treatment biopsies is ongoing and will be presented. Utilization of CMS characterization in mCRC clinical trials is feasible and may provide an improved biological understanding of treatment activity. Clinical trial information: NCT03428126.

215 Background: AT-rich interactive domain 1A (ARID1A) is a chromatin regulator mutated in human cancers, frequently resulting in truncation and loss of expression of this protein. ARID1A recruits MSH2 during DNA replication to perform mismatch-repair. ARID1A deficiency has been shown to increase mutational load and immune activation in preclinical models (Shen J, Nat Med 2018) but its role in colorectal cancer (CRC) patients (pts) is being explored. Methods: The DNA sequencing and gene expression profiling of microsatellite-stable (MSS) CRC pts were extracted from TCGA and MD Anderson Cancer Center databases. The expression levels were normalized according to the mean values of each dataset. The mutational burden and expression signatures for IFN-γ and various immune markers were compared according to the ARID1A mutational status. Results: Among 417 pts with MSS CRC, 28 pts (6.7%) had a non-silent mutation in ARID1A. Out of the 58 genes most commonly mutated in CRC, non-silent mutation in ARID1A had the strongest association with the frame-shift mutation rate in MSS cases (8-fold increase, p< .001). ARID1A mutation had also a strong correlation with the IFN-γ expression signature in MSS CRC (Δz score +1.91, p< .001) . Compared with ARID1A wild-type pts, higher expression signatures for cytotoxic T cell function, NK cells, and immune checkpoints were observed in MSS ARID1A mutated cases. ARID1A mutant cases showed higher expressions of various immune checkpoint genes (CD274, CTLA4, HAVCR2, IDO1, LAG3, PDCD1, and PDCD1LG2) compared to wild-type cases (all p < .05). All findings were observed independently in both datasets. Conclusions: In MSS CRC, ARID1A mutation is associated with a high expression of IFN-γ pathway and immune signatures (such as cytotoxic T cell function and immune checkpoint markers). The immunogenicity of ARID1A mutant cases is likely due to the increased level of neoantigens resulting from the increased rate of frame-shift mutations. Tumors with ARID1A mutation may be more susceptible to immune therapy-based treatment strategies and should likely be recognized as a unique molecular subgroup in future immune therapy trials.

TPS262 Background: The majority of mCRC pts are microsatellite stable (MSS), have poor intratumoral CD8+ T cell infiltration, and no clinical response to immunotherapy checkpoint inhibitors. Preclinical studies suggest that chemotherapy may synergize with anti-PD-1. In non-small cell lung cancer (NSCLC), the combination of pembrolizumab (PemB), pemetrexed (PemT), + carboplatin demonstrated synergistic activity. This study will combine PemB with PemT, then that combination + oxaliplatin (Ox). The rationale for addition of Ox to PemT is that enhanced immunogenic cell death may induce CD8+ T cell infiltration into CRC tumors and model the mechanism of cytotoxicity seen in NSCLC. Thus, the combination of PemB+ PemT + Ox may induce synergistic antitumor immune activity. Methods: This multi-center phase Ib trial is actively enrolling pts with incurable mCRC with prior treatment for mCRC including fluoropyrimidine-, oxaliplatin-, and irinotecan-based chemotherapy, and if RAS wild-type, anti-EGFR therapy. Measurable disease by imaging (RECIST 1.1) is required. Standard ineligibility includes active infections, systemic steroid use, or other conditions contraindicating immunotherapy. Cohort 1 will receive PemB + PemT; Cohort 2 will receive PemB + PemT + dose-escalated Ox. Imaging will be performed every 6 wks. The primary aim of Cohort 1: to evaluate for safety and efficacy using doses of PemB and PemT that have been studied in NSCLC. The primary aim of Cohort 2: to evaluate the safety, tolerability, and efficacy of PemB in combination with PemT + Ox. The RP2D of the 3-drug combination will be at the MTD taking into account toxicity profiles of study therapy agents. Secondary aims: to evaluate the clinical benefit rate of the doublet and triplet combinations in pts with chemo-refractory MSS mCRC and to estimate progression-free survival and overall survival in pts with MSS mCRC treated with these combinations. The cohorts will be analyzed separately with descriptive intent only. Maximum enrollment is 33 pts. Support: Merck; Lilly; NSABP Foundation, Inc. Clinical trial information: NCT03626922.

160 Background: Checkpoint inhibitors have demonstrated significant clinical impact in colorectal cancer with deficient mismatch repair (MMR) and high microsatellite instability. However, most patients have disease with stable microsatellites and typically respond poorly to most immunotherapies. The purpose of this study was to quantify the immune responses induced by mono versus dual immune checkpoint blockade (DICB) in a MMR proficient model of colorectal cancer. Methods: A MMR proficient model of colorectal cancer was established in 30 BALB/c mice via bilateral subcutaneous flank injections of 0.5x106 CT26.WT murine colorectal cancer cells (CRL- 2638; ATCC, Manassas, VA). Mice were assigned to receive either sham antibodies (InVivoMAb IgG controls), monotherapy with anti-PD-1 antibodies (Clone J43; BioXcell, West Lebanon, NH, USA), or DICB with anti-PD-1 and anti-CTLA-1 antibodies (Clone 9D9; BioXcell). Tumor growth was monitored over time and mice were sacrificed at either 7 or 14 days post-treatment. Single cell suspensions of harvested tumors and spleens were analyzed using FACS. The number of CD8+ and CD4+ T cells as well as the expression of co-inhibitory surface molecules PD-1, LAG3, and TIM3 was quantified in each sample. Results: DICB was associated with a reduction in tumor volume as compared to either mono PD-1 inhibition or control (p < 0.05). Neither monotherapy nor DICB significantly affected tumor infiltration by lymphocytes. Tumor infiltrating CD8+ T cells in the DICB treatment group demonstrated less expression of PD-1 and LAG3 compared to the control and monotherapy groups (PD-1: 1723±90.5 [Mean fluorescence intensity ± SEM] vs 2489±119.9 & 2404±117.9; p < 0.05) (LAG3: 453.8±21.58 vs 664.1±53.21 & 719.7±14.25; p < 0.05). The DICB group also showed increased expression of TIM3 (11,779±1271 vs 2832±380.9 & 4081±426.6; p < 0.05). Conclusions: These results suggest dual therapy with anti-CTLA and anti-PD-1 antibodies inhibits the growth of stable microsatellite colorectal cancer by suppressing key immunosuppressive checkpoints. Upregulation of TIM3 represents a potential escape mechanism that could be a target for future combination immunotherapies.

TPS266 Background: Patients (pts) with MSI-H/dMMR mCRC treated with CT have poorer outcomes than pts with microsatellite stable/MMR proficient mCRC. NIVO (anti–programmed death [PD]-1) and IPI (anti–cytotoxic T lymphocyte antigen-4 [CTLA-4]) are immune checkpoint inhibitors that act synergistically and promote antitumor immune response by complementary mechanisms. NIVO±IPI received accelerated US FDA approval for MSI-H/dMMR mCRC that progressed after fluoropyrimidine, oxaliplatin, and irinotecan treatment based on the phase 2, non-randomized, multicohort CheckMate 142 study. Indirect comparisons suggest that NIVO (3 mg/kg) + low-dose IPI (1 mg/kg) provides improved clinical benefit vs NIVO (investigator-assessed [INV] objective response rate [ORR] 55% vs 31%; 12-month [mo] INV progression-free survival [PFS] rate 71% vs 50%; 12-mo overall survival [OS] rate 85% vs 73%) with a favorable benefit-risk profile for previously treated MSI-H/dMMR mCRC (Overman et al. JCO 2018). NIVO+low-dose IPI also demonstrated robust and durable clinical benefit in first-line MSI-H/dMMR mCRC (INV ORR 64%; 12-mo INV PFS rate 77%; 12-mo OS rate 84%; Lenz et al. ASCO 2019, #3521). To date, no prospective phase 3 studies have reported results for anti–PD-1, anti–PD-1 + anti–CTLA-4, or CT in MSI-H/dMMR mCRC; these treatments will be evaluated in the international, multicenter, open-label, randomized, phase 3 CheckMate 8HW (NCT04008030) study. Methods: Approximately 494 pts aged ≥18 years with histologically confirmed recurrent or mCRC, irrespective of prior treatment with CT and/or targeted agents, not amenable to surgery, and with known tumor MSI-H or dMMR status, and Eastern Cooperative Oncology Group performance status ≤1 will be randomized to receive NIVO, NIVO+IPI, or investigator’s choice CT (pts in the CT arm can receive NIVO+IPI upon progression). The primary endpoint is PFS, assessed by blinded independent central review (BICR). Secondary endpoints include PFS by INV, ORR and disease control rate by BICR and INV, OS, time to and duration of response. Exploratory endpoints include safety. Clinical trial information: NCT04008030.

TPS264 Background: Clinical improvements for locally advanced rectal cancer have been relatively static over the past few decades. While immunotherapy shows no benefit in microsatellite stable (MSS) colorectal cancer, preclinical models suggest that radiotherapy (RT) can enhance neoantigen presentation, modulate the microenvironment, and improve the likelihood of anti-tumor activity with checkpoint inhibitor use. Using a “window-of-opportunity” study design, this prospective phase II trial will determine the safety and activity of this approach with the anti-PD-L1 agent durvalumab (MEDI4736). Methods: This multi-center phase II trial is currently enrolling patients (pts) with rectal cancer who are undergoing standard NCCN guideline-compliant neoadjuvant chemoradiotherapy (CRT). Eligibility includes pts with MSS stage II-IV rectal cancer with adequate organ function and pre-treatment diagnostic tumor available for profiling with intent to proceed to surgical resection after CRT. Stage IV disease must be limited such that the primary pelvic tumor requires definitive management. Standard ineligibility criteria include active infections, systemic steroid use, or other conditions making immunotherapy use unsafe. Treatment includes durvalumab (750mg IV infusion once every 2 wks) for 4 total doses beginning within 3-7 days after CRT completion. Surgery must be within 8-12 wks of the final CRT dose. Primary endpoint is a demonstrated improvement in Neoadjuvant Rectal Cancer (NAR) score compared to historical controls targeting a 20% relative risk reduction in DFS and 3-4% absolute OS improvement. Secondary endpoints include OS, DFS, toxicity, pCR, cCR, therapy completion, negative surgical margins, sphincter preservation, off-target “abscopal” effects for the subset of stage IV pts, and exploratory assessments of tumor infiltrating lymphocytes, tumor Immunoscore, circulating immunologic profiles, and molecular predictors of response. A safety run-in phase has completed as a precedent to full enrollment. Enrollment now continues to 47 total pts to achieve 41 surgically evaluable pts. Support: AstraZeneca-Medimmune, NSABP Foundation. Clinical trial information: NCT03102047.

e15048 Background: PD-1/PD-L1 inhibitor-based monotherapies and combination therapies are being investigated to challenge microsatellite stable (MSS) mCRC. This NMA aimed to assess the efficacy of anti-PD-1/PD-L1-based therapies against active comparators (regorafenib, trifiuridine+tipiracil, and fruquintinib [R/T/F]) and negative controls (placebo and best supportive care [P/BSC]) in patients with refractory MSI-unselected mCRC, which are predominantly MSS. Methods: We searched PubMed, EMBASE, the Cochrane Central Register of Controlled Trials, and major oncology websites for relevant RCTs comparing the above therapies with each other up to Feb 15, 2019. We used the frequentist NMA to evaluate hazard ratios (HRs) for pairwise treatment comparisons in terms of overall survival (OS). Results: Seven RCTs were included, totaling nine comparisons between five therapies (durvalumab+tremelimumab [anti-PD-L1+Tre] vs. P/BSC [n = 1], atezolizumab+cobimetinib [anti-PD-L1+Cobi] vs. R/T/F [n = 1], atezolizumab [anti-PD-L1-only] vs. R/T/F [n = 1], anti-PD-L1+Cobi vs. anti-PD-L1-only [n = 1], and R/T/F vs. P/BSC [n = 5]). No significant heterogeneity was detected in the NMA (P = 0.340 in Q test; I2= 11.6%). Compared to P/BSC, OS was significantly improved with anti-PD-L1+Tre (fixed-effects model HR, 0.72 [95% confidence interval, 0.54-0.96]; P = 0.014) and anti-PD-L1+Cobi (HR, 0.70 [0.50-0.98]; P = 0.018), but not with anti-PD-L1-only (HR, 0.83 [0.57-1.21]; P = 0.172). OS with anti-PD-L1+Tre and anti-PD-L1+Cobi did not significantly differ from that with R/T/F (P = 0.565 and P = 0.500, respectively). Conclusions: In refractory MSI-unselected mCRC, anti-PD-L1 monotherapy was ineffective but anti-PD-L1+Tre and anti-PD-L1+Cobi therapies exhibited efficacies comparable to that of R/T/F.

3606 Background: Chemoradiotherapy (CRT) with surgery (S) is standard for patients (pts) with locally-advanced rectal cancer (LARC), and nivolumab (nivo) is active in microsatellite instability-high (MSI-H) metastatic colorectal cancer (mCRC). We studied nivo and radical S following CRT (50.4 Gy with capecitabine 1,650 mg/m2) in T3–4 NanyM0 LARC. Methods: Phase I included testing of a recommended phase II dosing schedule (RP2S). Efficacy and safety were studied in phase II pts and those given RP2S in phase I. In Cohort A-1, for microsatellite stable (MSS) LARC pts, the primary endpoint was centrally confirmed pathologic complete response (pCR) rate using AJCC tumor regression grading. The estimated required sample size assuming null and alternative hypotheses pCR = 10% and 30% was 37 pts, with a 1-sided alpha of 5% and power of 90%. Cohort A-2 was exploratory and included a maximum of 5 MSI-H pts. Results: Nivo 240 mg q2 weeks x 5 cycles, following CRT but pre-S, was the RP2S. From 1/17 to 6/18, 37 pts were enrolled in Cohort A-1. Eleven pts (30%; 90% CI 18-44%) showed pCR (AJCC grade (gr) 0). Including the 3 pts (8%) graded AJCC 1, 14 (38%) had major pathologic responses. In addition, clinical CR was observed in 1 pt (3%) refusing S after nivo. Both MSI-H LARC Cohort A-2 pts showed pCRs. Immune-related severe adverse events were observed in 2 pts (gr 3 myasthenia and gr 2 interstitial nephritis); both fully recovered and had S. No treatment-related deaths were observed. pCR rates of 60% (6/10) and 19% (5/27) (p = 0.038, Fisher exact test) were seen in pts with tumor cells with PD-L1 ≥1% and < 1% IHC staining, respectively, performed on biopsy samples taken pre-CRT. Rates of 62% (8/13) and 10% (1/10) (p = 0.029) were seen in 23 pts with samples analyzable by flow cytometry, according to CD8+ lymphocyte /regulatory T cell (CD8/Treg) ratios ≥2 and < 2, respectively. Conclusions: A promising pCR rate of 30%, with mild toxicity, was shown in MSS LARC pts treated with nivo plus radical S. PD-L1 expression and elevated CD8/Treg ratio may be better predictors of nivo benefit, warranting further study in a larger cohort. Clinical trial information: NCT02948348.

3577 Background: Transforming growth factor- β pathway (TGF-β) has an established role in promoting growth, invasion, metastasis as well as epithelial to mesenchymal (EMT) transition. Among 4 different described molecular subtypes of colorectal cancer (CRC), consensus molecular subtype 4 (CMS4) comprises up to 25% of CRC pts, distinguished by activation of this pathway, and is associated with higher relapse rate and poor prognosis. Recently, it has also been proposed that TGF-β activation drives immune evasion in murine models, but these findings have not been clinically validated. Methods: Using multi-gene RNA expression profiling, fresh-frozen paraffin-embedded samples of 35 patients with CRC were analyzed to determine TGF-β and EMT expression levels. Multiplexed IHC staining was performed on FFPE tumor blocks by using the Opal 7-Color fIHC Kit and the stained slides were scanned by a Vectra multispectral microscope (PerkinElmer) to measure infiltration of immune cells (i.e., T lymphocytes, cytotoxic T lymphocytes (CTL), T cell antigen-experienced, macrophages, etc.) in the tumor, stroma, and both components. TGF-β and EMT expression levels – as continuous variables - were compared with the infiltration of various immune cells using Spearman’s rank correlation analysis. Results: Among 35 pts, 28 pts had non-CMS1/MSS CRC. TGF-β RNA expression in the tumor microenvironment of these samples was inversely associated with the infiltration of CTL into the tumor (r=-0.43, p= 0.022). In contrast, there was no association of TGF-β with non-cytotoxic T-cells or macrophage infiltration. The tumor and stromal CTL infiltration differed substantially by CMS ( p=0.04, p=0.02, respectively) with tumor infiltration lowest in CMS4 (n=7). Consistent with this, EMT gene signature, which includes TGF-β expression, showed a similar inverse correlation with CTL infiltration (r=-0.48, p=0.009). Conclusions: TGF-β and EMT gene signatures have important roles in the exclusion of CTL in the tumor microenvironment of CRC pts. Inhibiting TGF-β pathway can potentially increase the intratumoral infiltration of CTL, which is a necessary (but not sufficient) step for immunotherapy response in MSS CRC. Clinical trials evaluating this hypothesis are currently ongoing (NCT03436563).

e15047 Background: Adjunctive therapies are essential to enhance the effect of anti-PD1 therapies for the treatment of microsatellite stable (MSS) colorectal cancer. SBRT is utilized to treat liver metastatic CRC, causing an increase in immunogenic intra-tumoral and intra-lymphatic antigen release and a rapid influx of responding immune cells. We hypothesize that radiation enhances the immunogenicity of MSS CRC and potentiates the effectiveness of PD-1 blockade. This phase Ib study examines the safety and efficacy of the sequential combination of SBRT and Pem in patients for whom the goal is to resect all sites of known disease. Methods: Key eligibility criteria include MSS CRC with liver-confined metastatic disease with the therapeutic goal of resection of all radiographic disease with one operation. Subjects must be a candidate for SBRT to 1-3 liver metastases. Prior surgery and systemic chemotherapy are allowed. Subjects receive sequential SBRT and cycle 1 of Pem prior to operative management. Postoperatively, patients complete cycles 2-9 of Pem followed by scheduled surveillance with imaging every 12 weeks. The primary objectives are to determine the safety/tolerability of this regimen and the recurrence rate at 1 year following clearance of metastatic disease. Secondary objectives include time to recurrence, DFS, and OS. Results: 10 patients (median age 61.5 [range 39-69]) have been enrolled and completed the intended neoadjuvant therapy, operative management and at least 4 adjuvant cycles of Pem. All patients had received prior FOLFOX. SBRT median dose was 50 Gy (40-60 Gy) to a single lesion targeted in all patients. Mean gross tumor size targeted was 19.1cc (2.3-80.4). Any-grade post-operative AEs (>20%) attributable to Pem include diarrhea (30%), hyperbilirubinemia (20%%), and leukopenia (20%). Post-operative AEs included one case of biliary tract injury and biloma, not related to immunotherapy. No grade 3/4 immunotherapy AEs have occurred. Conclusions: The combination of SBRT with Pem and surgical resection is well tolerated with no signal of increased immunotherapy-related toxicity in the post-operative setting. Clinical trial information: NCT02837263.

3604 Background: PI3K pathway alterations, PIK3CA mut or PTEN loss, are known to modulate anti-tumor immune microenvironment. MSS CRC pts respond infrequently to immunotherapy, suggesting the presence of a rare MSS immunogenic subset. We investigated the immune repertoire and outcomes on immunotherapy trials in PIK3CA mut MSS CRC pts. Methods: Immune infiltrates and checkpoints were evaluated using quantitative immuno-histochemistry (IHC) on primary CRC. Mutations were assessed by next generation sequencing. PIK3CA mut neoepitopes and HLA allele affinities were predicted using NetMHC 4.0 Server. Outcomes of MSS CRC pts enrolled in 7 immunotherapy trials were assessed. Clinical benefit (CB) was defined as CR, PR or SD of 24 weeks. Time to progression (TTP) was calculated using Kaplan-Meier Method. PIK3CA mut vs wild type (wt) groups were compared using Mann-Whitney U, Fisher’s exact, or Log-Rank tests as appropriate. Results: PI3K alterations were present in 14/40 MSS CRC pts in IHC cohort (7 PIK3CA mut, 33 wt; 7 PTEN loss, 33 intact). The center of PIK3CA mut MSS CRC had higher median densities of CD3+ cells [1112 (IQ range 865-1421) vs 435 (300-744) cells/mm3; P=0.037] and CD8+ cells [554 (331-1200) vs 185 (60-473) cells/mm3; P=0.037] as compared to PIK3CA wt tumors. Intratumoral immune infiltrates did not differ by PTEN IHC staining in MSS CRC. PD-L1 H-scores were also higher in PIK3CA mut MSS CRC [85 (34-114) vs 29 (11-60); P=0.01]. Several activating PIK3CA mut (E542K, E545K, H1047R) were predicted to generate true neoepitopes with high binding affinity to common HLA types. Indeed, among MSS CRC pts enrolled in 7 immunotherapy trials, half (4/8) of PIK3CA mut pts derived CB as compared to 3/35 (8.6%) PIK3CA wt pts (P=0.015). PIK3CA mut pts had trend towards longer TTP (3.8 months in mut vs 2.1 months in wt; P=0.08). CB or TTP did not differ by colon sidedness, monotherapy / combination therapy, number of mut, or mut in other key genes ( APC, SMAD4, TP53, KRAS, NRAS or BRAF). Conclusions: PIK3CA mut MSS CRC are associated with increased cytotoxic T cell infiltration, higher PD-L1 expression, and greater clinical benefit from immunotherapy. Further investigation of immunotherapy outcomes in the context of neoepitope-HLA allele interaction may help identify a subset of PIK3CA mut MSS CRC pts who are likely to benefit from immunotherapy.

e14143 Background: NOX-A12 is an inhibitor of the chemokine CXCL12 which prevents receptor engagement and blocks the formation of chemotactic CXCL12 concentration gradients. The OPERA study (NCT03168139) is a Phase 1/2 study to evaluate pharmacodynamic effects and safety of monotherapy with NOX-A12 as well as safety and efficacy of a combination with pembrolizumab in metastatic microsatellite-stable (MSS) colorectal (CRC) and pancreatic (PaC) cancer where PD-1 inhibition alone has not shown clinical benefit. Methods: Patients received 300 mg NOX-A12 twice weekly during a 2-week monotherapy phase. Biopsies were taken from liver metastases before and after NOX-A12 monotherapy for analysis of immune cell infiltration and cytokine signature. In a combination phase, patients received repeated 21-day cycles of 300 mg NOX-A12 and 200 mg pembrolizumab until progression or intolerable toxicity. Results: 11 of the patients had CRC and 9 PaC. All patients were heavily pretreated with 5 (CRC) and 3 median lines (PaC) of prior treatment. Best responses to last prior treatment was progressive disease for 95% of patients. The AE profile was comparable with the pembrolizumab profile or typical for the underlying diseases. At baseline, mean T cell density at the invasive margin was 327 cells/mm², clearly below the 600 cells/mm² that are predictive for a good prognosis. 2 weeks of NOX-A12 monotherapy induced Th1 cytokines (IFNγ, IL-2, IL-16) in approx. half of the patients. Although there were no objective responses, 25% of patients achieved stable disease; 7 patients showed prolonged time on treatment vs. the prior line of therapy. Updated figures for progression-free survival and overall survival for the whole patient cohort will be presented. Conclusions: In patients with MSS metastatic PaC and CRC cancer with impaired immune systems and a high tumor load that have failed multiple prior lines of therapy, NOX-A12 plus pembrolizumab shows induction of immune response, stable disease in 25% of patients, and prolonged time on treatment vs. prior therapy for 35% of patients. The safety profile of the combination therapy was consistent with that of pembrolizumab in advanced cancer patients. Clinical trial information: NCT03168139.

e15060 Background: Microsatellite stable (MSI-S) RS mCRC patients (pts) have a worse prognosis relative to left sided tumors for overall survival (OS). The present analysis aims to test the hypothesis that a triplet-regimen is superior compared to doublet-regimens (DR; FOLFOX or FOLFIRI) for OS. Methods: Pts with treatment-naive RS mCRC at MD Anderson Cancer Center between January/2011 to December/2018 were selected. We compared the progression-free survival (PFS) and OS of mCRC pts treated with FOLFOXIRI versus DR. Pts treated with anti-EGFR therapy were excluded. Results: A total of 37 pts were treated with FOLFOXIRI and 111 pts with DR. There were no statistical difference between groups regarding gender, KRAS and BRAF mutations, peritoneal metastasis and bevacizumab use. There were statistical difference in age (median: 46y vs 59y) and metastasectomy rates (14% vs 32%) (p < 0.001). KRAS mutation was found in 65% of the population. Median follow-up was 55.3m. Median PFS was 6.5m vs 11.2m (HR: 1.30 95% CI 0.85 – 1.99) and median OS was 17.0m vs 26.3m (HR: 1.01 95% CI 0.60 – 1.68). By univariate analysis, pts who have undergone metastasectomy had superior PFS (14.9m vs 9.2m; p

3514 Background: mCRC remains a lethal cancer and immunotherapy in MSS mCRC has yet to show significant activity. In preclinical models, radiation induced cellular damage may increase responsiveness to immunotherapy via the abscopal effect, with evidence for synergy between radiation therapy (RT) and dual checkpoint blockade. In this study, we assessed dual blockade of CTLA-4 and PD-1 combined with RT as a strategy to stimulate an immune response for patients with MSS mCRC. Methods: In this open-label, single arm phase-2 study, we enrolled 40 MSS mCRC patients (pts). Eligible pts had histologically-confirmed MSS mCRC, ECOG PS 0-1, and progression on at least two lines of therapy. Treatment (Tx) consisted of ipilimumab (1mg/kg q6 weeks), nivolumab (240 mg q2 weeks) and 3 fractions of 8 Gy of RT at cycle 2 every other day. Tx continued until progression, discontinuation or withdrawal. The primary endpoint was Disease Control Rate (DCR), with radiological evaluations every 3 months. Exploratory endpoints included ORR, PFS, OS and safety. Response was defined as disease control outside the radiation field. We obtained serial tumor biopsies pre-tx, during checkpoint blockade alone (cycle 1) and 2 weeks after initiation of radiation. Intention-to-treat analysis (ITT) includes all pts receiving at least one dose of study agent. Results: 40 pts (median age 59 years (26-83), 58% male) enrolled and started treatment from 7/2017 to 12/2018. DCR was 17.5% (7/40) with a 7.5% (3/40) ORR by ITT. Median duration of disease control was 77 days in the ITT; 252 days for those with disease control (n=7) based on the first re-staging scans at 3 months or censored (n=3) and 70.5 days for pts with PD (n=17) or who did not receive RT due to clinical progression (n=13). In the modified ITT (all pts receiving RT and restaged), N=24 pts, excluding 1 pt pending 1st scans post-RT, DCR was 29.2% (7/24) and ORR 12.5% (3/24). Median duration of disease control in mITT was 77.5 days: 252 days for those with disease control and 77 days for those with PD. TRAEs were reported in 22/40 (55%). 20/40 (50%) with grade ≥3 toxicities, with fatigue, nausea, vomiting, diarrhea, infusion-related reaction and dyspnea being the most common. 1(2%) pt died of respiratory failure possibly related to tx. Conclusions: Dual blockade of CTLA-4 and PD-1 with RT is feasible and demonstrates durable activity in pts with MSS mCRC. There are 3 pts who have not completed RT or had their post-RT re-staging. We will report updated efficacy data and outcomes from correlative serial tumor biopsies upon trial completion. Clinical trial information: NCT03104439.

48 Background: Patients (pts) with colorectal cancer (CRC) with microsatellite instability (MSI) have recently been shown to respond to anti–programmed death (PD)-1 drugs. Preclinical data suggest that trifluridine/tipiracil (FTD/TPI) treatment converts MSS CRC cells to MSI, sensitizing them to the activity of anti–PD-1 drugs. The aim of this phase 2 study was to evaluate this hypothesis. Methods: This was a multicenter, single-arm, safety lead-in, phase 2 study that used Simon’s 2-stage design to evaluate the safety and efficacy of FTD/TPI (35 mg/m2 twice daily on days 1-5 and 8-12 every 28 days) plus nivolumab (3 mg/kg every 2 weeks) in pts with heavily pretreated MSS mCRC. Pts had histologically proven metastatic or locally advanced colorectal adenocarcinoma that was MSS (assessed by a local laboratory based on either previous or fresh biopsy), ≥1 measurable lesion for RECIST and immune-related response criteria (irRC) assessment, and failure of ≥2 previous lines of chemotherapy. Six pts were to be enrolled in the safety lead-in, and in order to proceed to Simon’s stage 2, ≥2 of the first 15 pts had to demonstrate a partial or complete response within 6 months based on irRC assessment. Results: The first 6 dose-limiting toxicity (DLT)-evaluable pts enrolled tolerated dosing with no DLTs. A total of 18 pretreated pts enrolled in the first stage (50% males; median age 56.5 yr), among whom 72% had colon cancer and 100% had MSS disease, 56% with RAS mutations. Pts received a median of 2.5 cycles of study therapy (range 1-8). The most common grade 3/4 adverse events (AEs) were neutropenia (28%); diarrhea (17%); and nausea, abdominal pain, fatigue, and anemia (11% each). No pts discontinued treatment due to AEs. No pts achieved a tumor response (either per RECIST or irRC), and the study did not progress to the second stage. Median (6-month) progression-free survival was 2.8 months (21%) per RECIST and 2.2 months (30%) per irRC. Conclusions: The combination of FTD/TPI and nivolumab was feasible and tolerated at the full dose of both compounds. Adding nivolumab to FTD/TPI did not provide additional clinical benefit in pts with previously treated MSS mCRC. Clinical trial information: NCT02860546.

680 Background: Adjunctive therapies are essential to enhance the effect of anti-PD1 therapies for the treatment of microsatellite stable (MSS) colorectal cancer. SBRT is utilized to treat liver metastatic CRC, causing an increase in immunogenic intratumoral and a rapid influx of responding immune cells. We hypothesize that radiation enhances immunogenicity of MSS CRC and potentiates effectiveness of PD-1 blockade. This phase Ib study examines the safety and efficacy of the sequential combination of SBRT and Pem in patients for whom the goal is to resect all sites of known disease. Methods: Key eligibility criteria include MSS CRC with liver-confined metastatic disease with the therapeutic goal of resection of all radiographic disease with one operation. Subjects must be a candidate for SBRT to 1-3 liver metastases. Prior surgery and systemic chemotherapy are allowed. Subjects receive sequential SBRT and cycle 1 of Pem prior to operative management. Postoperatively, patients complete cycles 2-9 of Pem followed by scheduled surveillance with imaging every 12 weeks. The primary objectives are to determine the safety of this regimen and the recurrence rate at one year following clearance of metastatic disease. Secondary objectives include time to recurrence, DFS, and OS. Results: Nine patients (median age 61.5 [range 39-69]) have completed the intended neoadjuvant therapy, operative management and at least one adjuvant cycle of Pem. All patients received prior FOLFOX. Any-grade AEs (> 20%) through cycle 2 of Pem attributable to SBRT include fatigue (44%) and nausea (22%). Any-grade AEs related to Pem include lymphopenia (25%). Postoperative AEs included one case of biliary tract injury and biloma, not related to immunotherapy. One patient developed a rash following SBRT and Pem which may be an immunotherapy-related toxicity. No grade 3/4 immunotherapy AEs have occurred. Conclusions: The combination of SBRT, Pem, and surgical resection is well tolerated with no signal of increased immunotherapy-related toxicity. Clinical trial information: NCT02837263.

e15597Background: Responses to anti-PD1 is ~35% in microsatellite unstable (MSI-h) and near 0% in microsatellite stable (MSS) Colorectal cancer (CRC). Understanding the composition of immune cell p...

3551Background: Pts with MSS all-RAS and BRAFV600E wild-type met CRC are more likely to respond to anti-EGFR drugs, irrespective of treatment (tx) line, particularly if left-sided primaries (prim)....

3540Background: Monalizumab (IPH2201) is a humanized IgG4 antibody targeting CD94/NKG2a to suppress inhibitory signaling by tumors on NK cells. The combination of monalizumab and durvalumab (anti-P...

3557Background: In studies reported to date, no objective responses have been observed with PD-1 inhibitors in advanced/metastatic CRC (mCRC) pts with microsatellite stable (MSS)/mismatch repair pr...

560 Background: The majority of pts with mCRC have microsatellite stable (MSS) tumors with minimal response to PD-L1/PD-1 blockade. MEK inhibition upregulates tumor major histocompatibility complex-I expression, promotes intratumoral T-cell accumulation and improves anti–PD-L1 responses (Ebert, Immunity. 2016), supporting clinical evaluation of combined anti–PD-L1 (A) plus MEKi (C) in pts with mCRC. Methods: Pts with chemotherapy-refractory or locally advanced mCRC were evaluated. A was administered IV q2w at 800 mg. C was dosed PO daily at 20 to 60 mg during dose escalation and 60 mg during dose expansion (14/14 or 21/7 d on/off schedule). Primary endpoints were safety and tolerability. Investigator-assessed ORR and PFS by RECIST v1.1 and OS were secondary endpoints. Microsatellite instability (MSI) status was locally reported and centrally confirmed by NGS-based scoring. Results: As of May 10, 2017, 84 pts (57 KRASmt, 25 KRASwt, 2 unknown) were enrolled and evaluable. 66 pts had received 5+ prior therapies; 38 had known MSI status at baseline (MSS [n = 29], MSI-low [n = 8], MSI-H [n = 1]). A + C combination was tolerable. Treatment-related Gr 1-2 and Gr 3-4 AEs occurred at 60% and 37%, respectively. The most frequent related Gr 3-4 AEs were increased blood CPK, rash, diarrhea and fatigue (5% each). No treatment-related Gr 5 AEs occurred. 13% and 24% of pts experienced AEs leading to A and C withdrawal, respectively. Median follow-up was 14.3 mo (range, 0.5-29.8). Confirmed PR was observed in 7 pts (8%; locally reported MSS [n = 3], MSI-low [n = 1], MSI unknown [n = 3]). Median DOR was 14.8 mo (95% CI: 6.0, NE); DCR (PR + SD) was 31%. Median PFS was 1.9 mo (95% CI: 1.8, 2.3); median OS was 10.0 mo (95% CI: 6.2, 14.1). The 6-mo and 1-y OS rates were 65% (95% CI: 54, 76) and 46% (95% CI: 34, 58), respectively. Molecular determinants and immune biomarkers of response will be presented. Conclusions: A + C demonstrated a tolerable safety profile and improvements in OS vs those reported with SOC in heavily pretreated pts with mCRC. Objective responses observed in pts with MSS/MSI-low mCRC were durable, suggesting benefit of this novel combination in a pt population refractory to immune therapies. Clinical trial information: NCT01988896.

TPS870 Background: Approximately 96% of CRCs have an MSS phenotype, which results in more immunologically quiescent tumors for which immunotherapies are largely ineffective (Lee et al. 2015; Overman et al. 2016). Pts with MSS CRC and activating RAS mutation (35%–45% of CRCs) have treatment options limited still further because anti-EGFR monoclonal antibodies (eg, cetuximab) are ineffective owing to dominant activation of RAS in the MAPK pathway (Douillard et al. 2014). However, preclinical and preliminary clinical data suggest that MAPK pathway inhibition enhances antigen presentation and T-cell cytotoxicity to positively modulate the efficacy of checkpoint inhibitors (Brea et al. 2016; Bendell et al. 2014). The main objective of this open-label multicenter phase 1b/2 study is to evaluate whether the potential positive modulation of NIVO or NIVO plus IPI, when combined with BINI, translates into clinically meaningful overall response in pts with MSS mCRC and RAS mutation. Methods: The study will enroll ~90 previously treated pts (1 or 2 prior regimens), ~42 in phase 1b and ~48 in phase 2. The primary objective of phase 1b will be to determine the recommended phase 2 dose (RP2D) of BINI in combination with NIVO ± IPI. Dose finding in the doublet arm will begin with BINI 45 mg BID + NIVO 480 mg Q4W; the triplet arm will begin with the BINI RP2D from the doublet arm + NIVO 480 mg Q4W + IPI 1 mg/kg Q8W. In phase 2, pts will be randomized 1:1 to doublet or triplet arms, incorporating the BINI RP2Ds found in phase 1b; treatment will continue in 28-day cycles until disease progression, unacceptable toxicity, withdrawal of consent, initiation of subsequent anticancer therapy, loss to follow-up, or death. The primary objective for phase 2 will be to assess response by RECIST version 1.1. The study will also characterize safety and PK. CT.gov Identifier: Clinical trial information: NCT03271047.

TPS5615 Background: The Cancer Genome Atlas project identified 2 groups of hypermutated endometrial cancers (ECs): an ultramutated group that harbored mutations in the exonuclease domain of polymerase e ( POLE), and a hypermutated group with microsatellite instability (MSI), the majority of which harbored MLH1 promoter methylation. We (Howitt, JAMA Onc 2015) and others have shown that POLE and MSI ECs are associated with higher number of predicted neoepitopes and tumor infiltrating lymphocytes, which is counterbalanced by overexpression of PD-1/PD-L1, suggesting that they may be excellent candidates for PD-1/PD-L1 blockade. Anti-PD-1 therapy has also demonstrated promising activity in mismatch repair deficient colorectal cancers and collectively in non-colorectal cancers (Le, NEJM 2015). Methods: This is an open-label, two-cohort, two-stage, phase 2 trial, of avelumab, a fully human IgG1 antibody directed against PD-L1, in two cohorts: i) a MSI/ POLE cohort including ECs with immunohistochemical (IHC) complete loss of expression of at least one of the mismatch repair (MMR) proteins and/or documented mutation in the exonuclease domain of POLE and ii) a MSS cohort including ECs with normal IHC expression of all MMR proteins. Key eligibility criteria include measurable disease, no upper limit of prior therapies, and any EC histology. Co-primary objectives include objective response rate and rate of progression-free survival at 6 months. Avelumab is administered at 10 mg/kg as 1-hour IV infusion every 2 weeks until disease progression or unacceptable toxicity; therapy may continue at the investigator’s discretion while awaiting radiologic confirmation of disease progression 4 weeks later. Maximum target enrollment is 70 patients (35 for each cohort). In the first stage, 16 patients will be enrolled in each cohort; if there are at least two objective responses or two patients progression-free at 6 months, accrual will continue to the second stage where 19 more patients will be enrolled for each cohort. Thus far, 16 patients have been enrolled, 13 on the MSS cohort and 3 on the MSI/ POLE cohort. Clinical trial information: NCT02912572.

3534 Background: Metastatic colorectal cancers (mCRCs) with a right-sided primary site are associated with shorter survival and insensitivity to EGFR inhibitors compared to those originating in the left side of the colon or rectum. Methods: We performed targeted gene sequencing of 928 consecutive MSS mCRCs. Primary tumor site was divided into right-sided for cecum to distal transverse colon (n = 242), left-sided for splenic flexure to rectum (n = 673), or unknown colonic location (n = 13). Histologic subtypes were conventional (adenocarcinoma not otherwise specified), conventional with mucinous features ( < 50% mucinous), mucinous, signet ring, and poorly differentiated. To evaluate receptor tyrosine kinase (RTK) signaling, we analyzed ligand mRNA expression in TCGA. Results: Overall survival from time of metastasis was shorter for right-sided than left-sided primary site (survival at 5 years: 45% v 67%, p < 0.001). Right-sided tumors had more mutations (5.60 v 4.62 per MB, p < 0.001) but fewer copy-number changes (0.18 v 0.22 fraction genome altered, p = 0.001) compared to left-sided tumors. Alterations of KRAS, BRAF, PIK3CA, PTEN, AKT1, RNF43, and SMAD4 were significantly enriched in right-sided tumors, and of APC and TP53 in left-sided tumors. In a multivariate model, APC (HR = 0.7, p = 0.03), BRAF (HR = 3.7, p < 0.001), and KRAS (HR = 1.7, p < 0.01) alterations predicted for survival, but primary site did not (HR = 0.74, p = 0.07). Amphiregulin, epiregulin, neuregulin, and HGF expression was significantly higher in left-sided tumors. We found a higher proportion of conventional histology (83% v 63%) and moderate differentiation (82% v 69%) for left versus right-sided cases. Conclusions: We find that within MSS mCRC there are significantly more oncogenic mutations in right-sided tumors, and the difference in survival between right- and left-sided mCRC is primarily driven by differences in mutations. Left-sided tumors more commonly exhibit a “simpler” conventional histology that is lower grade and may rely on native RTK signaling, such as EGFR, for growth, providing a potential mechanism for the differential sensitivity to EGFR inhibitors seen by primary tumor site.

3054 Background: Microsatellite stable (MSS) metastatic colorectal cancer (mCRC) has relatively poor tumoral infiltration of CD8+ T cells and is resistant to pembrolizumab (Pembro) when compared to MSI-H mCRC. DNA hypomethylating agent induces epigenetic expression of multiple genes including cancer-testis antigens in CRC, which are recognized by cytotoxic CD8+ T cells in vitro and in vivo. This trial tested whether concurrent treatment with azacitidine (Aza) could enhance the anti-tumor activity of Pembro. Methods: This is a phase 2 trial to evaluate anti-tumor activity and safety of Pembro plus Aza in patients (pts) with previously treated mCRC without any further standard chemotherapy option. Pts received Pembro 200 mg IV on day 1 of each cycle Q3W and Aza 100 mg daily SQ injection on days 1-5 of each cycle Q3W. Primary endpoint was response rate (ORR) using RECIST v1.1. Secondary endpoints included progression-free survival (PFS) and overall survival (OS). Tumor tissues were collected for correlative studies. Results: Thirty-one pts were enrolled [median age, 61 years (range, 30-79); 17 M/14 F; ECOG PS 0/1 (58%/42%); 30 pts with MSS mCRC]. Pts received at least 2 lines of prior systemic chemotherapy for mCRC (median, 3; range, 1-5). Thirty pts received at least one dose of study therapy (median, 3 cycles; range, 1-8). Ten pts could not complete the first 3 cycles due to rapid symptomatic tumor progression. One pt with MSS mCRC achieved PR and 3 pts had SD as best response. The ORR was 3% (1/30; 95% CI, 0.1-17%). Seven pts with PD at the end of cycle 3 continued on study therapy, and 2 pts had stabilization of tumor progression. Median PFS was 2.1 months (95% CI, 1.8-2.8), and median OS was 6.2 months (95% CI, 3.5-8.7). While treatment-related adverse events (TRAEs) were reported in 63% of pts, most of the TRAEs were Gr 1/2 (96%). Frequent TRAEs possibly related to Aza were anemia (n = 5), constipation (n = 5), and leukopenia (n = 4); and possibly related to both Aza and Pembro were nausea (n = 5) and fatigue (n = 5). Gr 3 TRAEs included anemia (n = 1), ALT elevation (n = 1), and alkaline phosphatase elevation (n = 1). Conclusions: Pembro plus Aza is feasible with a tolerable safety profile but appears to have minimal anti-tumor effect for MSS mCRC. Clinical trial information: NCT02260440.

627 Background: Microsatellite instability (MSI) can predict response to pembrolizumab in metastatic colorectal cancer (mCRC). Benefit to PD-1 blockade is limited in microsatellite stable (MSS) mCRC yet MSS tumors represent >95% of all mCRC cases. Evidence suggests that high mutational load remains an important predictor of benefit to checkpoint inhibition across several tumors. We aimed to explore associations between tumor mutational burden (TMB) and various patient factors and mutations of interest to identify features of patients (pts) with MSS mCRC of potential significance to prognosis and PD-1 targeting. Methods: We conducted a retrospective study based on next-generation sequencing (NGS) of MSS metastatic colorectal tumors at our single institution using FoundationOne. TMB was categorized as high ≥ 20 mutations/megabase (Mb), intermediate 6-19 mutations/Mb, and low ≤ 5 mutations/Mb. TMB frequencies were analyzed according to age (≤50, >50, >65, and >70 years old), gender, race, location of primary tumor, and KRAS, BRAF, and PIK3CA mutations. Results: Among 194 MSS tumors with available TMB data, frequencies of tumors with intermediate TMB were significantly different in pts ≤50 years (24.6%, n=69) than in pts >65 (48.9%, n=47, Fisher’s exact (two-tailed) p=0.0096) and >70 (53.1%, n=32, p=0.0067). Frequencies of tumors with low TMB were significantly different in pts ≤50 years (75.4%, n=69) than in pts >65 (46.8%, n=47, p=0.0029) and >70 (43.8%, n=32, p=0.0032). Frequencies of high TMB tumors by age were 0% (≤50 years), 1.6% (>50), 4.3% (>65), and 3.1% (>70) and did not reach significance. Frequencies of high, intermediate, or low TMB tumors were not significantly different based on gender, race, and location of primary tumor. Frequencies of KRAS, BRAF, and PIK3CA mutations were similar across all TMB categories (p=0.7995, 0.8774, and 0.053, respectively). Conclusions: Older individuals with mCRC are less likely to have a low mutation burden by FoundationOne assay. Further validation of these findings in larger cohorts may indicate a potential higher likelihood for elderly colorectal cancer patients to benefit from immune-based strategies.

TPS3626Background: Pembrolizumab has been shown to result in compelling responses and progression free survival in advanced MMR-d CRC. However, the vast majority (~95%) of patients with advanced CR...

1523Background: MSI status is predictive of anti-PD-1/L1 therapy efficacy independent of tumor histology. This study describes the development and validation of a novel computational method to assess MSI status. In addition, we sought to determine the MSI prevalence and mutation characteristics in the cohort by applying the method to 11,573 clinical solid tumor cases previously assayed with CGP. Methods: To determine MSI status, 114 intronic homopolymer repeat loci with adequate coverage on the CGP panel were analyzed for length variability and compiled into an overall MSI score via principal components analysis. Translation of the MSI score to MSI-H (microsatellite high) or MSS (microsatellite stable) was established using training data. A matched-normal sample is not required, and CGP is accurate when tumor content is at least 20%. Results: Our methodology was validated in 69 samples with MSI status (20 MSI-H, 49 MSS) previously determined by standard clinical testing (PCR or immunohistochemistry). Conc...