Search

BackgroundThe human leukocyte antigen class I (HLA-I) genotype has been linked with differential immune responses to infectious disease and cancer. However, the clinical relevance of germline HLA-mediated immunity in gastrointestinal (GI) cancer remains elusive.MethodsThis study retrospectively analyzed the genomic profiling data from 84 metastatic GI cancer patients treated with immune checkpoint blockade (ICB) recruited from Peking University Cancer Hospital (PUCH). A publicly available dataset from the Memorial Sloan Kettering (MSK) Cancer Center (MSK GI cohort) was employed as the validation cohort. For the PUCH cohort, we performed HLA genotyping by whole exome sequencing (WES) analysis on the peripheral blood samples from all patients. Tumor tissues from 76 patients were subjected to WES analysis and immune oncology-related RNA profiling. We studied the associations of two parameters of germline HLA as heterozygosity and evolutionary divergence (HED, a quantifiable measure of HLA-I evolution) with the clinical outcomes of patients in both cohorts.ResultsOur data showed that neither HLA heterozygosity nor HED at the HLA-A/HLA-C locus correlated with the overall survival (OS) in the PUCH cohort. Interestingly, in both the PUCH and MSK GI cohorts, patients with high HLA-B HED showed a better OS compared with low HLA-B HED subgroup. Of note, a combinatorial biomarker of HLA-B HED and tumor mutational burden (TMB) may better stratify potential responders. Furthermore, patients with high HLA-B HED were characterized with a decreased prevalence of multiple driver gene mutations and an immune-inflamed phenotype.ConclusionsOur results unveil how HLA-B evolutionary divergence influences the ICB response in patients with GI cancers, supporting its potential utility as a combinatorial biomarker together with TMB for patient stratification in the future.

The association between genetic variations and immunotherapy benefit has been widely recognized, while such evidence in gastrointestinal cancer remains limited. We analyzed the genomic profile of 227 immunotherapeutic gastrointestinal cancer patients treated with immunotherapy, from the Memorial Sloan Kettering (MSK) Cancer Center cohort. A gastrointestinal immune prognostic signature (GIPS) was constructed using LASSO Cox regression. Based on this signature, patients were classified into two subgroups with distinctive prognoses (p N = 54) and Peking University Cancer Hospital cohort (N = 92). Multivariate analysis revealed that the GIPS was an independent prognostic biomarker. Notably, the GIPS-high tumor was indicative of a T-cell-inflamed phenotype and immune activation. The findings demonstrated that GIPS was a powerful predictor of immunotherapeutic survival in gastrointestinal cancer and may serve as a potential biomarker guiding immunotherapy treatment decisions.

We report on a comprehensive analysis of the gut microbiomes of patients with gastrointestinal (GI) cancer receiving anti–PD-1/PD-L1 treatment. The human gut microbiota has been associated with clinical responses to anti–PD-1/PD-L1 immunotherapy in melanoma, non–small cell lung cancer, and renal cell carcinoma. We aimed to investigate this association in GI cancers. We also identified bacterial taxa with patient stratification potential. We recruited 74 patients with advanced-stage GI cancer receiving anti–PD-1/PD-L1 treatment and collected their fecal samples prior to and during immunotherapy, along with clinical evaluations. Our 16S rRNA taxonomy survey on the fecal samples revealed an elevation of the Prevotella/Bacteroides ratio in patients, with a preferred response to anti–PD-1/PD-L1 treatment, and a particular subgroup of responders harboring a significantly higher abundance of Prevotella, Ruminococcaceae, and Lachnospiraceae. The shotgun metagenomes of the same samples showed that patients exhibiting different responses had differential abundance of pathways related to nucleoside and nucleotide biosynthesis, lipid biosynthesis, sugar metabolism, and fermentation to short-chain fatty acids (SCFA). Gut bacteria that were capable of SCFA production, including Eubacterium, Lactobacillus, and Streptococcus, were positively associated with anti–PD-1/PD-L1 response across different GI cancer types. We further demonstrated that the identified bacterial taxa were predictive of patient stratification in both our cohort and melanoma patients from two previously published studies. Our results thus highlight the impact of gut microbiomes on anti–PD-1/PD-L1 outcomes, at least in a subset of patients with GI cancer, and suggest the potential of the microbiome as a marker for immune-checkpoint blockade responses. See articles by Tomita et al., p. 1236, and Hakozaki et al., p. 1243

Objective: Immune checkpoint inhibitors (ICIs) have achieved remarkable results in cancer treatments. However, there is no effectivepredictive biomarker for gastrointestinal (GI) cancer. Methods: We conducted integrative analyses of the genomic and survival data of ICI-treated GI cancer patients from the MemorialSloan Kettering Cancer Center cohort (MSK-GI, n = 227), the Janjigian cohort (n = 40), and the Peking University Cancer Hospital& Institute cohort (PUCH, n = 80) to determine the possible associations between DNA damage response and repair (DDR) genemutations and clinical outcomes. Data from The Cancer Genome Atlas database were analyzed to determine the possible correlationsbetween DDR gene mutations and the tumor microenvironment. Results: In the MSK cohort, the presence of ≥ 2 DDR gene mutations was correlated with prolonged overall survival (OS). TheJanjigian and PUCH cohorts further confirmed that subgroups with ≥ 2 DDR gene mutations displayed a prolonged OS and ahigher durable clinical benefit. Furthermore, the DDR gene mutation load could be considered as an independent prognostic factor,and exhibited a potential predictive value for survival in GI cancer patients treated with ICIs. Mechanistically, we showed that thepresence of ≥ 2 DDR gene mutations was correlated with higher levels of tumor mutation burden, neoantigen, and T cell infiltration. Conclusions: The DDR gene mutation status was correlated with favorable clinical outcomes in GI cancer patients receiving ICIs,which could serve as a potential biomarker to guide patient selection for immunotherapy.

BackgroundDespite the great achievements made in immune-checkpoint-blockade (ICB) in cancer therapy, there are no effective predictive biomarkers in gastrointestinal (GI) cancer.MethodsThis study included 93 metastatic GI patients treated with ICBs. The first cohort comprising 73 GI cancer patients were randomly assigned into discovery (n=44) and validation (n=29) cohorts. Comprehensive genomic profiling was performed on all samples to determine tumor mutational burden (TMB) and copy-number alterations (CNAs). A subset of samples was collected for RNA immune oncology (IO) panel sequencing, microsatellite instability (MSI)/mismatch repair and program death ligand 1 (PD-L1) expression evaluation. In addition, 20 gastric cancer (GC) patients were recruited as the second validation cohort.ResultsIn the first cohort of 73 GI cancer patients, a lower burden of CNA was observed in patients with durable clinical benefit (DCB). In both the discovery (n=44) and validation (n=29) subsets, lower burden of CNA was associated with an improved clinical benefit and better overall survival (OS). Efficacy also correlated with a higher TMB. Of note, a combinatorial biomarker of TMB and CNA may better stratify DCB patients from ICB treatment, which was further confirmed in the second validation cohort of 20 GC patients. Finally, patients with lower burden of CNA revealed increased immune signatures in our cohort and The Cancer Genome Atlas data sets as well.ConclusionsOur results suggest that the burden of CNA may have superior predictive value compared with other signatures, including PD-L1, MSI and TMB. The joint biomarker of CNA burden and TMB may better stratify DCB patients, thereby providing a rational choice for GI patients treated with ICBs.

e16101 Background: To date, although a number of biomarker-related investigations have focused on the intrinsic properties of tumor cells and immune microenvironment, how germline genetics influences efficacy of immune checkpoint inhibitors (ICIs) immunotherapy in gastrointestinal (GI) cancer is scarcely understood. Methods: Our investigation enrolled 94 metastatic GI cancer patients treated with ICIs recruited from Peking University Cancer Hospital (PUCH) between August 1, 2015, and May 24, 2019. A publicly available dataset from the Memorial Sloan Kettering (MSK) Cancer Center (MSK GI cohort) was used for validation. For the PUCH cohort, we performed HLA genotyping by whole exome sequencing (WES) analysis on the peripheral blood mononuclear cell (PBMC) from all 94 patients. Tumor tissues from 86 patients were subjected to WES analysis and immune oncology-related RNA profiling. Results: We assessed the clinic relevance of germline HLA heterozygosity and evolutionary divergence (HED, a quantifiable measure of HLA-I evolution) to immunotherapy in patients with advanced GI cancers from the PUCH cohort. Our data showed that neither HLA heterozygosity nor mean HED correlated with the overall survival (OS) in the PUCH cohort. However, patients with high HLA-B HED showed a better OS and durable clinic benefit (DCB) rate compared with the lower subgroup (p < 0.05 for all comparison). The prognostic value of HLA-B HED was consistently validated in the MSK GI cohort (N = 84). Notably, a combinatorial biomarker based on HLA-B HED and tumor mutation burden (TMB) could better stratify potential responders (86 patients with tumor samples). Survival analysis performed on the PUCH and MSK GI cancer datasets further demonstrated the potential joint utility of HLA-B HED and TMB for prognosis stratification. Moreover, HLA-B HED high subgroup was characterized with a lower prevalence of TP53 mutation and enrichment of multiple immune related pathways, indicating an immune-inflamed phenotype of this subgroup. Conclusions: Taken together, our data create an intriguing argument for the germline HLA-B sequence divergence and TMB as complementary biomarkers in guiding patient selection for GI cancer immunotherapy. Further investigation revealed a potential HLA-B restricted mutational pattern of TP53, and can be indicative of tumor immune microenvironment in GI cancer.

Background Peritoneal dissemination (PD) is a type of metastasis frequently found in gastrointestinal (GI) cancer patients, resulting in the production of massive malignant ascites which associate with poor prognosis. However, the question of how the expression of tumor biomarkers on cancer cells in malignant ascites correlates with PD is yet to be answered. This study aims to investigate the clinical significance of PD-L1 and HER2 expression on malignant ascites disseminated tumor cells (maDTCs) in GI carcinoma patients. Patients and Methods Paired blood and malignant ascites specimens were respectively collected from 24 of metastatic GI subjects, including 15 gastric, 2 pancreatic, 1 esophageal, and 6 colorectal cancer patients. Non-hematologic CTCs and maDTCs were enriched by subtraction enrichment (SE), followed by immunostaining-FISH (iFISH) to co-examine PD-L1 and HER2 expression on aneuploid maDTCs and CTCs. In addition, CD4+, CD8+, Treg and naive T cells in the same cohort of samples were quantified by flow cytometry. Results In situ karyotypic and phenotypic characterization performed by iFISH showed that the average number of maDTCs in the 24 patients was 32755/ml ascites, which was significantly higher than the average number of CTCs detected (3.5/ml blood) (P Conclusions Our study suggested that a higher proportion of Tregs in malignant ascites may contribute towards fostering an immunosuppressive microenvironment.PD-L1+/HER2+ maDTCs, in such an immunosuppressive microenvironment, might be able to evade immune system attack. Moreover, PD-L1 and HER2 identified on maDTCs could serve as therapy targets with respect to inhibition of peritoneal metastasis, potentially leading to an additional treatment strategy for GI cancer patients with malignant ascites. Citation Format: Yang Chen, Changsong Qi, JiaJia Yuan, Dan Liu, Zhi Peng, Xiaotian Zhang, Jian Li, Pin Peter Lin, Yilin Li, Lin Shen. Elevated PD-L1 and HER2 expression on aneuploid carcinoma cells in malignant ascites accelerate stumor cell immune evasion in gastrointestinal cancer patients [abstract]. In: Proceedings of the Annual Meeting of the American Association for Cancer Research 2020; 2020 Apr 27-28 and Jun 22-24. Philadelphia (PA): AACR; Cancer Res 2020;80(16 Suppl):Abstract nr 5381.

We report the first and largest cohort, to the best of our knowledge, on the gut microbiome of 126 gastrointestinal (GI) cancer patients in China receiving the immune checkpoint inhibitors (ICI) targeting PD-1. The cohort included 31 patients with colon cancer, 23 with esophageal carcinoma, 47 with gastric cancer, and 25 with other cancer types. 43 of the patients had samples collected before treatment. 35 of patients were assessed as microsatellite-instability-High (MSI-High) and 71 as non-MSI-High. 408 fecal samples from the patients were collected. We performed 16s rRNA amplicon sequencing (V4 region) of the patients' samples to investigate the relationship between microbial profiles and patients' immunotherapy treatment response (RECIST 1.1). Principal coordinate analysis (PCoA) showed a subgroup of responders exhibiting a microbiome profile which was enriched in Prevotella. Hierarchical clustering of patients' microbiome profiles also suggested that for each gastric and colon cancer, patients with the same response tend to share similar microbiome profiles. By comparing the microbiome profiles between R and NR group using samples collected before treatment or within 3 weeks since the start of treatment, we identified 16s rRNA V4 sequences from Prevotella enriched in R group for each tumor type, which suggests positive association between Prevotella and tumor response. Our data also suggested that MSI as good marker for tumor response, although more data is required to achieve the significance level. Besides, no significant difference in alpha diversity was observed between R and NR group for each tumor type. In sum, we studied the microbiome profiles of 126 GI cancer patients and evaluated their association with anti-PD-1 immunotherapy. Patients with the same tumor type and response tend to share similar microbiome profiles, where Prevotella and a few other taxa contributed to the pattern. Citation Format: Zhi Peng, Siyuan Cheng, Quan Zhang, Yan Kou, Ziqi Wang, Rong Jin, Han Hu, Xiaotian Zhang, Jifang Gong, Jian Li, Ming Lu, Xicheng Wang, Jun Zhou, Zhihao Lu, Dongtao Bi, Yan Tan, Lin Shen. The gut microbiome is associated with clinical responses to anti-PD-1 immunotherapy in gastrointestinal cancer patients in China [abstract]. In: Proceedings of the Annual Meeting of the American Association for Cancer Research 2020; 2020 Apr 27-28 and Jun 22-24. Philadelphia (PA): AACR; Cancer Res 2020;80(16 Suppl):Abstract nr 4298.

Objective: Recent clinical trials considered Claudin18 splice variants 2 (CLDN18.2) as the promising potential target for gastrointestinal(GI) cancer treatment. CLDN18.2 specifically express in normal stomach and gastric cancer, while ectopic activation can also be found among esophageal, pancreatic and bile duct cancers. However, high-consistency of CLDN18.1 and CLDN18.2 sequences makes it challenging to explore CLDN18.2-specific antibody. Therefore, detecting CLDN18.2 RNA expression may be another choice. Circulating tumor cells (CTCs) are regarded as biomarkers guiding treatment and tools investigated drug-resistant mechanisms. We aim to establish a CLDN18.2 RNA detection method in CTCs(CLDN18.2+ CTCs)for further application in monitoring anti-CLDN18.2 treatment of GI tumor. Methods: We established a CLDN18.2 RNA detection method based on designing a CLDN18.2-specific RNA molecular beacon (MB). 12 late stage GI cancer patients(pts)were enrolled. Paired blood (6ml) and tumor tissue specimens were respectively collected from 2 pancreatic and 10 gastric cancer pts. We evaluated CLDN18.2 expression on tissue by immunohistochemical (IHC) results. Then we analyzed CLDN18.2 expression on CTCs and tumor tissue. Results: Molecular and cell-line validation indicated that CLDN18.2 MB could specifically hybridized with CLDN18.2 RNA. In our study, Every pt had an evaluable CLDN18.2 IHC result. CTCs were identified in all of the pts (CTCs numbers ranging from 2 to 25). 9 of 12 pts had detectable CLDN18.2+CTCs, among which the CLDN18.2+CTCs ratio (CLDN18.2+CTCs/ Total number of CTCs) fluctuated from 14% to 100%. 96 CTCs fluorescence intensity were calculated statistically and classified by Agglomerative Clustering arithmetic. Consequently, we classified 3 clusters with significant differences (p Conclusions: CLDN18.2 RNA positive CTCs ratio was consistent with CLDN18.2 expression level in GI cancer pts. It is warranted to further exploring the phenotypes and distinct mechanisms of CLDN18.2 RNA positive CTCs dynamic changes during anti-CLDN18.2 therapy. CLDN18.2 Expression in 12 GI cancer ptsPtsCancerCLDN18.2 Expression on Tissue(IHC)CLDN18.2 Expression in CTCs(CLDN18.2 MB)CLDN18.2-CLDN18.21+CLDN18.22+CTCs SumCLDN18.2+CTC ratio1GC020020%2GC0010010100%3GC040040%4GC030030%5GC2+07714100%6GC2+0257100%7PC2+410520%8PC2+0628100%9GC2+710814%10GC3+023225100%11GC3+221560%12GC3+0145100%*GC: Gastric cancer ; PC: Pancreatic cancer Citation Format: Xiaoyi Chong, Linyang Fan, Changsong Qi, Jifang Gong, Zhi Peng, Dan Liu, Yang Chen, Chenhui Xu, Yanyan Li, Congcong Ji, Qin Li, Zhiyuan Hu, Xiaotian Zhang, Lin Shen. In situ detection of CLDN18.2 RNA in circulating tumor cells in gastrointestinal cancer [abstract]. In: Proceedings of the Annual Meeting of the American Association for Cancer Research 2020; 2020 Apr 27-28 and Jun 22-24. Philadelphia (PA): AACR; Cancer Res 2020;80(16 Suppl):Abstract nr 5386.

Purpose: Predictive biomarkers for immune checkpoint inhibitors (ICIs) are urgently needed in advanced gastrointestinal (GI) carcinoma. Methods: Peripheral blood lymphocyte, lymphocyte subset and early dynamic changes of them were retrospectively analyzed in patients with GI cancer treated with ICIs. Cox regression and Kaplan-Meier analyses were conducted to analyze survival. Results: Eighty patients were enrolled. According to whether tumor progression occurred by 6 months after the initial ICI, patients were divided into two cohorts, 6m-progressive disease (PD) (N=40) and 6m-non-PD (nPD) (N=40).BaselineCD4+/CD8+ T cell (CD4/CD8) ratio was higher in the 6m-nPD cohort than in the 6m-PD cohort. Lymphocyte subsets were obtained both at baseline and before the second dose of ICIs in 61 patients. In total, 54.1% (33/61) of patients experienced tumor progression within 6 months of the initial ICI treatment, and these patients showed significant declines in CD4+ T cell counts (CD4-C1-decline) (p=0.0002) and the CD8+ T cell counts (CD8-C1-decline) (p=0.045) after the first dose of ICIs. In multivariate analyses, CD4-C1-decline (HR=12.66; 95% CI: 2.17-73.91; p=0.005) was an independent prognostic factor for OS. Furthermore, Stratified analyses indicated that CD4-C1-decline was a poor prognostic factor for tumor progression in patients with a deficient mismatch repair (dMMR)/microsatellite instability-high (MSI-H). Conclusion: Early change in CD4+ T cell counts in peripheral blood samples may help as prognostic biomarkers for GI cancer patients treated with ICIs. Further prospective studies are needed to validate this conclusion. Citation Format: Zhi-Hao Lu, Chang Liu, Yan-ni Wang, Shuang Li, Xi Jiao, Jian-ling Zou, Zheng-hang Wang, Chang-song Qi, Xiao-tian Zhang, Jian Li, Lin Shen. Early change in peripheral blood CD4+T cells associated with the clinical outcome in gastrointestinal cancer patients receiving immune checkpoint inhibitors [abstract]. In: Proceedings of the Annual Meeting of the American Association for Cancer Research 2020; 2020 Apr 27-28 and Jun 22-24. Philadelphia (PA): AACR; Cancer Res 2020;80(16 Suppl):Abstract nr 3162.

Immune checkpoint inhibitors (ICIs) have revolutionized the therapeutic landscape of gastrointestinal cancer. However, biomarkers correlated with the efficacy of ICIs in gastrointestinal cancer are still lacking. In this study, we performed 395-plex immune oncology (IO)-related gene target sequencing in tumor samples from 96 patients with metastatic gastrointestinal cancer patients treated with ICIs, and a linear support vector machine learning strategy was applied to construct a predictive model. ResultsAll 96 patients were randomly assigned into the discovery (n=72) and validation (n=24) cohorts. A 24-gene RNA signature (termed the IO-score) was constructed from 395 immune-related gene expression profiling using a machine learning strategy to identify patients who might benefit from ICIs. The durable clinical benefit rate was higher in patients with a high IO-score than in patients with a low IO-score (discovery cohort: 92.0% vs 4.3%, p

e16551 Background: Chowell et al. showed that HLA (human leukocyte antigen) class I (HLA-I) heterozygosity may avail to survival benefit from immune checkpoint blockade (ICB) in advanced cancer patients. Among their cohorts of melanoma patients, patients with the HLA-B44 supertype had improved survival, while the individuals with HLA-B62 supertype had worse outcome. However, these observations were not supported by another study of NSCLC patients treated with immune checkpoint inhibitors (ICIs). Now whether HLA-I genotype can serve as predictive marker for ICB treatment of other malignancies such as gastrointestinal cancer remains unclear. Methods: Tumor tissue samples before immunotherapy as well as related clinical data were collected from 91 Chinese gastrointestinal cancer patients receiving anti-PD-1/PD-L1 inhibitors and/or anti-CTLA-4 antibodies. Genomic features of tumor specimens were characterized by whole exome sequencing (WES) and HLA-I genotyping of normal DNA from peripheral blood cells were performed using WES data. Tumor immune microenvironment signatures were examined by RNA-sequencing targeting 395 immune related genes. Results: We observed that the individuals with HLA-B62 supertype (n = 15) had a significantly shorter PFS (median PFS, 1.83 months versus 3.53 months; HR, 2.3; 95% CI, 1.03-5.0; p= 0.009) than that of those without HLA-B62 supertype (n = 35) in a subgroup of patients with microsatellite stability (MSS, n = 50) and it was not so in other patients with microsatellite instability-high (MSI-H, n = 17) or unknown MSI status (n = 24). Moreover, the patients with B62 supertype and TP53 mutations (n = 15) exhibited more markedly shortened PFS (mPFS, 1.8 months versus 5.57 months; HR, 3.9; 95% CI, 1.5-9.6; p< 0.0001) and OS (mOS, 4.2 months versus 17.3 months; HR, 4.5; 95% CI, 1.5-13.2; p< 0.0001) compared to those patients with non-B62 supertype and TP53 wildtype (n = 26). Furthermore, gene expression levels of certain immune activation related pathways or markers (IFN-γ signature, cytolytic marker, etc.) in the patients with B62 supertype and without B44 supertype (n = 23) were statistically lower ( p< 0.05) than in those patients without B62 supertype (n = 62). Chi-square test revealed no correlation between HLA-B62 supertype and TP53 mutation, MSI status, tumor mutation burden (TMB) or PD-L1 expression. Conclusions: HLA-B62 supertype in combination with TP53 mutations may define a subset of gastrointestinal cancer patients who probably fail to benefit from ICIs, especially in MSS patients.