Construction of a gene model related to the prognosis of patients with gastric cancer receiving immunotherapy and exploration of COX7A1 gene function.

BACKGROUND: GC is a highly heterogeneous tumor with different responses to immunotherapy, and the positive response depends on the unique interaction between the tumor and the tumor microenvironment (TME). However, the currently available methods for prognostic prediction are not satisfactory. Therefore, this study aims to construct a novel model that integrates relevant gene sets to predict the clinical efficacy of immunotherapy and the prognosis of GC patients based on machine learning. METHODS: Seven GC datasets were collected from the Gene Expression Omnibus (GEO) database, The Cancer Genome Atlas (TCGA) database and literature sources. Based on the immunotherapy cohort, we first obtained a list of immunotherapy related genes through differential expression analysis. Then, Cox regression analysis was applied to divide these genes with prognostic significancy into protective and risky types. Then, the Single Sample Gene Set Enrichment Analysis (ssGSEA) algorithm was used to score the two categories of gene sets separately, and the scores differences between the two gene sets were used as the basis for constructing the prognostic model. Subsequently, Weighted Correlation Network Analysis (WGCNA) and Cytoscape were applied to further screen the gene sets of the constructed model, and finally COX7A1 was selected for the exploration and prediction of the relationship between the clinical efficacy of immunotherapy for GC. The correlation between COX7A1 and immune cell infiltration, drug sensitivity scoring, and immunohistochemical staining were performed to initially understand the potential role of COX7A1 in the development and progression of GC. Finally, the differential expression of COX7A1 was verified in those GC patients receiving immunotherapy. RESULTS: First, 47 protective genes and 408 risky genes were obtained, and the ssGSEA algorithm was applied for model construction, showing good prognostic discrimination ability. In addition, the patients with high model scores showed higher TMB and MSI levels, and lower tumor heterogeneity scores. Then, it is found that the COX7A1 expressions in GC tissues were significantly lower than those in their corresponding paracancerous tissues. Meanwhile, the patients with high COX7A1 expression showed higher probability of cancer invasion, worse clinical efficacy of immunotherapy, worse overall survival (OS) and worse disease-free survival (DFS). CONCLUSIONS: The ssGSEA score we constructed can serve as a biomarker for GC patients and provide important guidance for individualized treatment. In addition, the COX7A1 gene can accurately distinguish the prognosis of GC patients and predict the clinical efficacy of immunotherapy for GC patients.

Novel staging for gastric neuroendocrine neoplasms by incorporating the WHO grading into the TNM staging system.

BACKGROUND: The 8th tumor-node-metastasis (TNM) classification of the American Joint Committee on Cancer (AJCC) can be used to estimate the prognosis of gastric neuroendocrine tumor (gNET) and gastric neuroendocrine carcinoma (gNEC) patients but not gastric neuroendocrine neoplasms (gNENs). METHODS: First, in the SEER (training) dataset, a TNMG system was built by combining the WHO G grade (G1-4; NEC grouped into G4) with the 8th AJCC T (T1-4), N (N0-1), and M (M0-1) stage, which was then validated in a Chinese (validation) cohort. RESULTS: In all, 2245 gNENs cases from the training dataset and 280 cases from the validation dataset were eligible. The T stage, M stage, and G grade were independent prognostic factors for OS in both datasets (all p < 0.05). The TNMG staging system demonstrated better C-index for predicting OS than the 8th AJCC TNM staging system in both the training (0.87, 95%CI: 0.86-0.88 vs. 0.79, 95%CI: 0.77-0.81) and validation (0.77, 95%CI: 0.73-0.80 vs. 0.75, 95%CI: 0.71-0.79) datasets. The AUC of the 3-year OS for the TNMG staging system was 0.936 and 0.817 in the SEER and validation dataset, respectively; higher than those of the 8th AJCC system (vs. 0.843 and 0.779, respectively). DCA revealed that compared with the 8th AJCC TNM staging system, the TNMG staging system demonstrated superior net prognostic benefit in both the training and validation datasets. CONCLUSIONS: The proposed TNMG staging system could more accurately predict the 3- and 5-year OS rate of gNENs patients than the 8th AJCC TNM staging system.

Improved photoelectric performance of all-inorganic perovskite through different additives for green light-emitting diodes.

The exceptional optical and electronic properties of all-inorganic cesium lead bromide (CsPbBr3) perovskite make it an ideal new optoelectronic material, but low surface coverage limits its performance. The morphological characteristics of thin films have a great influence on the performance of perovskite light emitting diodes, especially at low coverage, and an inhomogeneous surface will lead to current leakage. To tackle this problem, the widespread adoption of composite layers including polymers poly(ethylene oxide) (PEO) and organic insulating poly(vinylpyrrolidone) (PVP) and all-inorganic perovskites is an effective way to increase the surface coverage and uniformity of perovskite films and improve the performance of perovskite light emitting devices. In our work, the perovskite thin films are investigated by using PEO and PVP dual additives, and the optimized CsPbBr3-PEO-PVP LED with maximum luminance, current efficiency, and external quantum efficiency of 2353 cd m-2 (at 7.2 V), 2.14 cd A-1 (at 6.5 V) and 0.85% (at 6.5 V) was obtained. This work indicates that the method of using additives is not only the key to enhancing the quality of perovskite thin film, but also the key to achieving a higher performance perovskite LED.