Search

BackgroundNowadays, immunotherapy targeting immune checkpoint receptors is one of the cornerstones of systemic treatment in melanoma. Homologous recombination repair (HRR) is one of the DNA damage response (DDR) pathways, which has been proved to correlate with the efficacy of platinum-based chemotherapy, PARP inhibitor therapy, and immunotherapy in a variety of cancers. However, their predictive value of HRR remained unknown in patients with advanced melanoma.MethodsData of advanced melanoma patients from an independent cohort (Samstein2018) were used to analyze the correlation with immunogenic markers and the prognostic effect of HRR on immunotherapy, and another four cohorts (pooled cohort: Miao2018, Allen 2015, Hugo2016, and Synder2014) were used for validation. Immune infiltration cell scores analyzed by TCGA-SKCM cohort were used to explore potential mechanisms related to the immune microenvironment.ResultsCompared to patients with an HRR wild type (HRRwt), those with HRR mutations (HRRmut) in anti-CTLA-4-treated patients of the Samstein2018 cohort had higher tumor mutation burden (TMB; P = 0.0041) and longer median overall survival (mOS; P = 0.0094). In terms of results validation, it was also confirmed that the mOS (P = 0.0014) of HRRmut patients receiving anti-CTLA-4 therapy was significantly better than that of HRRwt patients in the pooled cohort, and objective response rates (ORR; P = 0.0053) were also found to be significant. However, there was no significant difference in mOS between HRRmut patients who received anti-PD-1/L1 therapy and HRRwt patients in either the discovery (Samstein2018 cohort, P = 0.94) or validation (pooled cohort, P = 0.96) set. Exploratory analysis found that although HRRmut patients showed no significant difference in mOS between anti-CTLA-4 and anti-PD-1/L1 therapy (P = 0.79), the mOS value of the anti-CTLA-4 therapy group (31.7 months) in HRRmut patients was numerically superior to the anti-PD-1/L1 therapy group (27.5 months). In contrast, the mOS of the anti-CTLA-4 therapy group was significantly lower than that of the anti-PD-1/L1 therapy group (12.4 vs. 32.0 months) in HRRwt patients. In addition, transcriptome profiling analysis revealed that the 29 (65.9%)-gene mutation of the HRR pathway associated with reshaping of the immunological microenvironment in melanoma.ConclusionsHRR mutations were associated with a higher TMB level, and better anti-CTLA-4 therapy outcomes. HRR may serve as an independent predictor of anti-CTLA-4 therapy efficacy in patients with advanced melanoma and their clinical value warrants further investigation.

BackgroundBiomarkers based on immune context may guide prognosis prediction. T-cell inactivation, exclusion, or dysfunction could cause unfavorable tumor microenvironments, which affect immunotherapy and prognosis. However, none of the immuno-biomarkers reported to date can differentiate colorectal-cancer (CRC) patients. Thus, we aimed to classify CRC patients according to the levels of T-cell activation, exclusion, and dysfunction in the tumor microenvironment.MethodsRNAseq data of 618 CRC patients from The Cancer Genome Atlas and microarray data of 316 CRC patients from Gene Expression Omnibus were analysed using the Tumor Immune Dysfunction and Exclusion algorithm. Unsupervised clustering was used to classify patients.ResultsBased on the expression signatures of myeloid-derived suppressor cells, cancer-associated fibroblasts, M2-like tumor-associated macrophages, cytotoxic T-lymphocytes, and PD-L1, all patients were clustered into four subtypes: cluster 1 had a high level of immune dysfunction, cluster 2 had a low level of immune activation, cluster 3 had intense immune exclusion, and cluster 4 had a high level of immune activation and a moderate level of both dysfunction and exclusion signatures. Compared with cluster 1, the hazard ratios and 95% confidential intervals for overall survival were 0.63 (0.35–1.13) for cluster 2, 0.55 (0.29–1.03) for cluster 3, and 0.30 (0.14–0.64) for cluster 4 in multivariate Cox regression. Similar immune clustering and prognosis patterns were obtained upon validation in the GSE39582 cohort. In subgroup analysis, immune clustering was significantly associated with overall survival among stage I/II patients, microsatellite stable/instability-low patients, and patients not treated with adjuvant therapy.ConclusionsOur findings demonstrated that classifying CRC patients into different immune subtypes serves as a reliable prognosis predictor and may help to refine patient selection for personalized cancer immunotherapy.

Purpose: NOTCH signaling is associated with tumorigenesis, mutagenesis, and immune tolerance in non–small cell lung cancer (NSCLC), indicating its association with the clinical benefit of immune checkpoint inhibitors (ICI). We hypothesized that NOTCH mutation in NSCLC might be a robust predictor of immunotherapeutic efficacy. Experimental Design: Multiple-dimensional data including genomic, transcriptomic, and clinical data from cohorts of NSCLC internal and public cohorts involving immunotherapeutic patients were analyzed. Polymorphism Phenotyping v2 (PolyPhen-2) system was performed to determine deleterious NOTCH mutation (del-NOTCHmut). Further investigation on molecular mechanism was performed in The Cancer Genome Atlas (TCGA) data via CIBERSORT and gene set enrichment analysis. Results: Our 3DMed cohort (n = 58) and other four cohorts (Rizvi, POPLAR/OAK, Van Allen, and MSKCC; n = 1,499) uncovered marked correlation between NOTCH1/2/3 mutation and better ICI outcomes in EGFR/ALKWT population, including objective response rate (2.20-fold, P = 0.001), progression-free survival [HR, 0.61; 95% confidence interval (CI), 0.46–0.81; P = 0.001], and overall survival (HR, 0.56; 95% CI, 0.32–0.96; P = 0.035). Del-NOTCHmut exhibited better predictive function than non-deleterious NOTCH mutation, potentially via greater transcription of genes related to DNA damage response and immune activation. Del-NOTCHmut was not linked with prognosis in TCGA cohorts and chemotherapeutic response, but was independently associated with immunotherapeutic benefit, delineating the predictive, but not prognostic, utility of del-NOTCHmut. Conclusions: This work distinguishes del-NOTCHmut as a potential predictor to favorable ICI response in NSCLC, highlighting the importance of genomic profiling in immunotherapy. More importantly, our results unravel a possibility of personalized combination immunotherapy as adding NOTCH inhibitor to ICI regimen in NSCLC, for the optimization of ICI treatment in clinical practice.

Blood-based tumor mutational burden (bTMB) has been studied to identify patients with NSCLC who would benefit from anti-programmed cell death protein 1 (anti-PD-1) or anti-programmed death ligand 1 (anti-PD-L1) therapies. However, it failed to predict overall survival (OS) benefits, which warrants further exploration.Three independent cohorts of patients with NSCLC treated with immunotherapy were used in this study. A new bTMB algorithm was first developed in the two independent cohorts (POPLAR, N = 211, and OAK, N = 462) and further validated in the third National Cancer Center (NCC) cohort (N = 64).bTMB-H (bTMB ≥ cutoff point) was not associated with favorable OS after immunotherapy regardless of the cutoff points in either the POPLAR and OAK or the NCC cohorts (p0.05) owing to its correlation with the amount of circulating tumor DNA, which was associated with poor OS. In the POPLAR and OAK cohorts, with allele frequency (AF) adjustment, a high AF bTMB (HAF-bTMB, mutation counts with an AF5%) was strongly correlated with the amount of circulating tumor DNA (Pearson r = 0.65), whereas a low AF bTMB (LAF-bTMB, mutation counts with an AF ≤ 5%) was not (Pearson r = 0.09). LAF-bTMB-H was associated with favorable OS (hazard ratio [HR] = 0.70, 95% confidence interval [CI]: 0.52-0.95, p = 0.02), progression-free survival (PFS; HR = 0.62, 95% CI: 0.47-0.80, p0.001), and objective response rate (ORR) (p 0.001) after immunotherapy but not chemotherapy, with a cutoff point of 12 trained in the POPLAR cohort and validated in the OAK cohort. The LAF-bTMB algorithm was further validated in the NCC cohort in which LAF-bTMB-H was associated with OS (HR = 0.20, 95% CI: 0.05-0.84, p = 0.02), PFS (HR = 0.30, 95% CI: 0.13-0.70, p = 0.003), and ORR (p = 0.001).We developed and validated a new LAF-bTMB algorithm as a feasible predictor of OS, PFS, and ORR after anti-PD-(L)1 therapies in patients with NSCLC, which needs to be prospectively validated.

PurposeThe KEAP1-NFE2L2 (Kelch-like ECH-associated protein 1 (KEAP1)-Nuclear factor (erythroid-derived 2)-like 2 (NFE2L2)) mutations are associated with resistance to chemotherapy or immunotherapy in non-small cell lung cancer (NSCLC). Conversely, it has been reported that NFE2L2 mutations potentiate improved clinical outcome with immunotherapy. However, therapeutic benefits for patients with KEAP1/NFE2L2 mutations remain unclear. The purpose of this study was to investigate the association between KEAP1/NFE2L2 and NSCLC prognosis, and to explore whether immunotherapy can improve prognosis in populations with KEAP1/NFE2L2 mutations.Experimental DesignThe impact of KEAP1/NFE2L2 mutations on survival outcomes in NSCLC patients received immunotherapy and chemotherapy was verified in the randomized phase II/III POPLAR/OAK trials (blood-based sequencing, bNGS cohort, POPLAR (n = 211) and OAK (n = 642)). The Cancer Genome Atlas (TCGA) NSCLC cohort (n=998) and an in-house Chinese NSCLC cohort (n=733) was used For the analysis of immune-related markers.ResultsCompared with KEAP1/NFE2L2 wild-type, patients with KEAP1/NFE2L2 mutations were significantly associated with poorer overall survival (OS, HR = 1.97, 95% CI 1.48–2.63, P < 0.001) on atezolizumab and docetaxel (HR = 1.66, 95% CI 1.28–2.16, P < 0.001). In KEAP1/NFE2L2 mutant group, there was no significant difference in median OS between atezolizumab and docetaxel (HR 0.74, 95% CI 0.53–1.03, P = 0.07). NFE2L2/KEAP1 mutations were significantly associated with higher TMB values and PD-L1 expression in the OAK/POPLAR and in-house Chinese NSCLC cohorts. GSEA revealed that KEAP1/NFE2L2mutant subgroup was associated with deficient infiltration of CD4+ T cells, NK T cells and natural Treg cells, and lower expression of DNA damage response genes in TCGA NSCLC cohort.ConclusionsOur study revealed that patients with KEAP1/NFE2L2 mutations have a worse prognosis than wild-type patients, both on immunotherapy and chemotherapy. In addition, in patients with KEAP1/NFE2L2 mutations, immunotherapy did not significantly improve prognosis compared to chemotherapy.

Background Immunotherapy is effective in treating unresectable esophageal squamous cell carcinoma (ESCC), but little is known about its role in the preoperative setting. The aim of this study was to evaluate the safety, feasibility and efficacy of neoadjuvant treatment with camrelizumab plus chemotherapy in locally advanced ESCC. Methods Patients diagnosed with locally advanced ESCC were retrospectively included if they had received neoadjuvant camrelizumab plus nab-paclitaxel and S1 capsule followed by radical esophagectomy between November, 2019 and June, 2020 at Sun Yat-sen University Cancer Center. Primary endpoints were safety and feasibility. In addition, pathological response and the relationship between tumor immune microenvironment (TIME)/tumor mutational burden (TMB) and treatment response were also investigated. Results Twelve patients were included and they all received three courses of preoperative treatment with camrelizumab plus nab-paclitaxel/S1. No grade 3 or higher toxicities occurred. No surgical delay or perioperative death was reported. Nine patients (75%) responded to the treatment, four with a complete pathological response (pCR) and five with a major pathological response (MPR). Neither programmed death-ligand 1 (PD-L1) expression nor TMB was correlated with treatment response. TIME analysis revealed that a higher abundance of CD56dim natural killer cells was associated with better pathological response in the primary tumor, while lower density of M2-tumor-associated macrophages was associated with better pathological response in the lymph nodes (LNs). Conclusions Neoadjuvant camrelizumab plus nab-paclitaxel and S1 is safe, feasible and effective in locally advanced ESCC and is worth further investigation.

Background: Epithelial-mesenchymal transition (EMT) has been reported to play a pivotal role in tumor immunosuppression and immune evasion. In multiple solid tumors, PD-L1 expression was much higher in “mesenchymal” compared with that in “epithelial”, indicating patients with mesenchymal phenotypes are more likely to benefit from PD-1/PD-L1 immunotherapy. ECM2 encodes a putative extracellular matrix protein, which involves in cell-cell and/or cell ECM recognition processes, thereby may affecting EMT. However, the predictive value of ECM2 in cancer immunotherapy has not been revealed. Method: Seven independent public cohorts of NSCLC, melanoma, and pan-cancer, including Rizvi, Van Allen, MSKCC, OAK/POPLAR, Miao, and Samstein cohorts were used to investigate the correlation between ECM2 mutation and clinical events, such as objective response rate (ORR), overall survival (OS), and progression free survival (PFS). Result: Across all the analyzed cancer types, ECM2 mutation has a strong connection with better outcomes to immune checkpoint inhibitors only in melanoma cancer. ECM2-mutated group was significantly associated with higher ORR (75.0% vs 13.9%, P = 0.013), longer OS (35.9 months vs 8.5 months, 95% CI, 0.02-1.09, P = 0.032) than ECM2-wide-type group. PFS between these two groups was comparable (8.15 months vs 2.8 months, 95% CI, P = 0.25). Furthermore, melanoma patients harboring ECM2 mutation exhibited significantly higher TMB compared with ECM2-wide-type patients in Van Allen cohort (1279.5 mut/Mb vs 125.5 mut/Mb, P = 0.0056). Conclusion: This study demonstrated that ECM2 mutation can serve as a potential predictor for a favorable response to ICIs in melanoma patients. Citation Format: Qitao Yu, Aiping Zeng, Yun Zhao, Xiaochun Huang, Wenzhuan Xie, Mengli Huang, Ting Bei. ECM2 mutation can serve as a potential efficacious predictor of immunotherapy in melanoma patients [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2021; 2021 Apr 10-15 and May 17-21. Philadelphia (PA): AACR; Cancer Res 2021;81(13_Suppl):Abstract nr 397.

e21547 Background: Gene mutations in DNA damage repair (DDR) pathway were reported to affect the clinical response to immune checkpoint inhibitors (ICIs) by driving mutagenicity. Previous studies showed that RECQL5 is involved in RAD51-mediated strand invasion for homologous recombination DNA damage repair (HR-DDR). RECQL5 encoding RecQ protein-like 5 (RECQL5) is a member of RecQ helicase family and was discovered to play a tumor-suppressive or oncogenic role in various cancers. However, the association between RECQL5 mutation and ICI efficacy has not been revealed. Methods: Data of nine publicly independent cohorts of NSCLC, melanoma, and pan-cancer were retrieved (Rizvi, MSKCC, OAK/POPLAR, Van Allen, Hugo, Synder, Miao and Samstein cohorts) to investigate the correlation between RECQL5 mutations and clinical events including overall survival (OS), progression-free survival (PFS), objective response rate (ORR), and clinical benefit rate (CBR). Wilcoxon test was used for comparing tumor mutational burden (TMB) between RECQL5-mutated patients and their wild-type counterparts. In addition, the correlation between infiltration of immune cells and RECQL5 mutation status was also analyzed by QUANTISEO using data of SKCM cohort (n=466) from TCGA database. Statistical significance was set at P = 0.05. Results: RECQL5 mutations were most commonly seen in melanoma patients and the detection rate was 4%-8% in melanoma patients from cohort Van Allen, Hugo, Synder, and Miao. RECQL5 mutation was not detected in other cancers except for one lung cancer case. In the three melanoma cohorts (Van Allen, Synder, Miao) with OS data available, the OS of the RECQL5-mutated patients was 2 to 3 times longer than that of the RECQL5-wt patients. In Synder cohort, the difference in OS between these two subsets was statistically significant (53.9 months vs 25.0 months, P = 0.045). In addition, the occurrence of RECQL5 mutations was correlated with higher CBR both in Miao ( P = 0.017) and Synder cohorts ( P = 0.011). No difference was observed in the PFS or ORR between these two subsets. Notably, RECQL5 mutations were associated with higher TMB levels both in Miao (32.6 mut/Mb vs 7.9 mut/Mb, P = 0.016) and Synder cohorts (1124 mut/Mb vs 358 mut/Mb, P = 0.033). Moreover, RECQL5 mutations were found to be correlated with increased infiltration of CD4+ T cells (no-regulatory) ( P = 0.037), B cells ( P = 0.13) and decreased infiltration of regulatory CD4+ T cells ( P = 0.02) in RECQL5-mutated tumors over RECQL5-wt tumors. Conclusions: RECQL5 mutations can serve as a potential predictor for a durable response to ICIs in melanoma. Moreover, the occurrence of RECQL5 mutations was correlated with higher TMB level as well as infiltration of immune cells, indicating the underlying mechanism of its predictive effect.

e21537 Background: PALB2, a gene in the homologous recombination repair (HRR) pathway, has been shown to be associated with the efficacy of platinum based chemotherapy, and PARP inhibitor therapy in breast, prostate, ovarian, and pancreatic cancers. However, their predictive value of PALB2 remained unknown in patients with advanced melanoma. Methods: Five independent cohorts (Miao2018, Samstein2018, Allen 2015, Hugo2016, and Synder2014. study cohort) with data from 672 patients with advanced melanoma were used to analyze the correlation with immunogenic marker, and the prognostic effect of PALB2 on immunotherapy. Results: A pooled analysis of five independent cohorts of 672 advanced patients melanoma show that 31 (4.6%) harbored PALB2 mutation ( PALB2mut). PALB2mut (72.63Muts/Mb) was associated with higher tumor mutation burden (TMB) (P < 0.001) than PALB wild-type ( PALB2wt) (19.71Muts/Mb). The same phenomenon has also been observed in TNB, PALB2mut (1983 counts) was associated with higher tumor neoantigen burden (TNB) than PALB2wt (603.5 counts) (P = 0.0147). The objective response rate (ORR) of immunotherapy was 53.33% for the patients with PALB2mut and 24% for the PALB2wt subgroup (P = 0.027). The PALB2mut patients had significantly improved median overall survival (mOS) than the PALB2wt group (Not reach versus 29 months, hazard ratio (HR) = 0.38, 95%CI 0.19−0.73, P = 0.003). A subgroup analysis of CTLA4 inhibitor treatment found that PALB2mut was associated with better ORR (50% versus 18.3%, P = 0.016) on immunotherapy, and mOS in the PALB2mut group was significantly better than that in the PALB2wt group (Not reach versus 21 months, HR = 0.3, 95%CI 0.13−0.68, P = 0.002). However, in the subgroup receiving PD-L1 inhibitor treatment, no ORR benefit was found in the PALB2mut group (66.67% versus 56.76%, P = 1), and there was no difference on mOS between PALB2mut and PALB2wt (Not reach versus 31 months, HR = 0.45, 95%CI 0.11−1.8, P = 0.254). Conclusions: PALB2 mutations was associated with a higher TMB and TNB level. PALB2 may serve as a positive predictor of immunotherapy (CTLA4 inhibitor therapy) in patients with advanced melanoma and their clinical value warrants further investigation.

e16076 Background: Immunotherapy is effective in treating advanced esophageal squamous cell carcinoma (ESCC), but little is known about its role in the setting of neoadjuvant therapy. Methods: Data from a retrospective cohort and a prospective cohort of locally advanced ESCC patients were analyzed. All patients included had received neoadjuvant camrelizumab plus nab-paclitaxel and S1 capsule followed by radical esophagectomy. The main purpose of this study was to evaluate the safety and feasibility of the neoadjuvant treatment. In addition, pathological response and the relationship between tumor immune microenvironment features (TIME)/tumor mutational burden (TMB) and treatment response were also explored. Results: A total of 25 locally advanced ESCC patients from both cohorts were included, of which 12 from the retrospective cohort and 13 from the prospective cohort. Only two patients experienced grade 3 or 4 adverse events, including onehad anemia and the other (4%) had limb numbness. No surgical delay or perioperative death was reported. Sixteen patients (64%) responded to the treatment, eight (32%) with a complete pathological response and eight (32%) with a major pathological response. Neither programmed death-ligand 1 expression nor TMB was correlated with treatment response. In terms of TIME analysis, patients with better pathological response in the primary tumor had significantly higher abundance of CD56dim natural killer cells (109 vs 32, P= 0.01), while lower density of M2-tumor-associated macrophages was associated with numerically better pathological response in the lymph nodes (47 vs 163.1, P= 0.55). Conclusions: Neoadjuvant camrelizumab plus nab-paclitaxel and S1 is safe, feasible and effective, and the efficacy could be predicted by TIME, which is worthy of further investigation. Clinical trial information: ChiCTR2000029807.

e21587 Background: Immunotherapies targeting immune checkpoint receptors have become the cornerstone of systemic treatment options for malignant melanoma. The response to these immunotherapies may correlate with driver mutations. MAP2K1/2 genes are mutated in approximately 10% of melanomas, however, the impact of MAP2K1/2 gene alterations on the efficiency of immunotherapy has not been clarified. Methods: Six metastatic melanoma clinical cohorts treated with ICIs were included to investigate the association between clinical efficacy of immunotherapy and MAP2K1/2 mutations. Survival analyses were conducted in cohorts receiving two kinds of ICB agents, namely anti-CTLA-4 or anti-PD-1. RNA expression profiling from these cohorts and from the TCGA melanoma cohort were used to explore the potential mechanism related to immune activation. Results: In an independent anti-CTLA-4-treated cohort (n = 110), we found that MAP2K1/2 mutations are predictive of high objective response rate (17.6% vs 1.3%, p = 0.0185) and long progression-free survival [median OS, 49.2 months vs 8.3 months; hazard ratio (HR) = 0.37; 95% CI, 0.15–0.91; p = 0.0307] and overall survival (median PFS, 19.4 months vs 2.8 months; HR = 0.2; 95% CI, 0.05–0.83; p = 0.0262). This predictive value was further validated in a pooled anti-CTLA-4-treated cohort (n = 235) in terms of overall survival (median OS, 49.3 months vs 22.0 months; HR = 0.44; 95% CI, 0.22–0.91; p = 0.0255). However, no correlation between MAP2K1/2 mutations and overall survival was observed in the anti-PD-1-treated cohort (n = 285). Subgroup Cox regression analysis indicated that MAP2K-mutated patients receive less benefit from the anti-PD-1 monotherapy than from the anti-CTLA-4 treatment (median OS, 27.0 months vs 49.3 months; HR = 3.26; 95% CI, 1.18–9.02; p = 0.0225), which was contrary to the result obtained for the total population. Furthermore, transcriptome profiling analysis revealed that MAP2K-mutated tumors are enriched in CD8+ T cells, B cells, and neutrophil cells and also express high levels of CD33 and IL10, which might be the underlying mechanism for melanoma patients with MAP2K1/2-mutated benefit more from anti-CTLA-4 treatment. Conclusions: We identified mutations in MAP2K1/2 genes as the independent predictive factors for anti-CTLA-4 therapy in melanoma patients and found that anti-CTLA-4 treatment in patient harbouring MAP2K1/2 mutations might be more effective than the anti-PD-1 therapy.

Background: Immune checkpoint inhibitors (ICIs) of programmed cell death 1 (PD-1) and its ligand (PD-L1) have led to a paradigm shift in cancer treatment. Here we conducted a systematic review and meta-analysis to compare the efficacy and safety profile of anti-PD-1 and anti-PD-L1 in pan-cancer patients. Methods: We systematically searched PubMed, CENTRAL, and Embase from January, 2000 to March, 2019 for randomized controlled trials that compared anti-PD-1 and anti-PD-L1 with standard treatment in patients with solid tumors. We also reviewed abstracts and presentations from all major conference proceedings. Studies with anti-PD-1 and anti-PD-L1 were screened and paired upon the matching of clinical characteristics as mirror groups. For the main results, effect size for each mirror group was derived from the frequentist approach with R package netmeta, and pooled using a random-effects model. The primary outcome was the difference in overall survival (OS) between anti-PD-1 and anti-PD-L1. Findings: 19 randomized controlled trials involving 11,500 patients were included for the meta-analysis. Overall, anti-PD-1 exhibited superior OS (HR 0.75, 95% CI 0.65-0.86i¼ŒP

IMPORTANCE: Immune checkpoint inhibitors of programmed cell death 1 (PD-1) and its ligand (PD-L1) have led to a paradigm shift in cancer treatment. Understanding the clinical efficacy and safety profile of these drugs is necessary for treatment strategy in clinical practice. OBJECTIVE: To assess the differences between anti–PD-1 and anti–PD-L1 regarding efficacy and safety shown in randomized clinical trials across various tumor types. DATA SOURCES: Systematic searches of PubMed, Cochrane CENTRAL, and Embase were conducted from January 1, 2000, to March 1, 2019. In addition, abstracts and presentations from all major conference proceedings were reviewed. STUDY SELECTION: All randomized clinical trials that compared anti–PD-1 and anti–PD-L1 with standard treatment in patients with cancer were selected as candidates. Retrospective studies, single-arm phase 1/2 studies, and trials comparing anti–PD-1 and anti–PD-L1 with other immunotherapies were excluded. Studies of anti–PD-1 and anti–PD-L1 therapy were screened and paired by the matching of clinical characteristics as mirror groups. DATA EXTRACTION AND SYNTHESIS: Three investigators independently extracted data from each study following the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-analyses) guideline. Trial names, first author, year of publication, study design, National Clinical Trial identifier number, blinding status, study phase, pathologic characteristics, number of patients, patients’ age and sex distribution, Eastern Cooperative Oncology Group Performance Status, lines of treatment, study drugs, biomarker status, follow-up time, incidence of adverse events, and hazard ratios (HRs) with 95% CIs for overall survival and progression-free survival were extracted. A random-effects model was applied for data analysis. MAIN OUTCOMES AND MEASURES: Differences in OS between anti–PD-1 and anti–PD-L1 across different cancer types were assessed. An effect size was derived from each mirror group and then pooled across all groups using a random-effects model. RESULTS: Nineteen randomized clinical trials involving 11 379 patients were included in the meta-analysis. Overall, anti–PD-1 exhibited superior overall survival (HR, 0.75; 95% CI, 0.65-0.86; P

e15078 Background: Immune checkpoint blockades (ICBs) have been approved for colorectal cancer (CRC) with microsatellite instability high (MSI-H). However, 50% to 60% MSI-H and almost 100% microsatellite stability (MSS) CRC patients can’t benefit from ICBs treatment. How to warm these “cold tumor” for boosting response to ICBs is still a problem. We aimed to integrate the expression signatures of T cell activation, exclusion and dysfunction to generate immune microenvironment subtypes which may guide precision immunotherapy. Methods: RNA-Seq of 596 CRC patients from The Cancer Genome Atlas were analyzed. T cell activation, dysfunction and exclusion signatures were calculated from TIDE model algorithm. Unsupervised clustering was used to classify CRC into various immune subtypes. Mann Whitney U test and Kruskal test were used to compare the difference among two or more than two groups. COX regression multivariate analysis was applied to analyze the association of subtypes with overall survival (OS). Results: Compared with 543 patients with MSS, 83 patients had significantly higher expression of T cell activation related genes (including CD8A, CXCL9, CXCL10, GZMB, IFNG et al) and immune checkpoint genes (including PD-L1, PD-1, LAG3 et al.), but lower signatures scores of M2-like tumor-associated macrophage (TAM) and T cell exclusion ( all P value