| Purposes:The 8th edition AJCC-TNM staging system of UICC-AJCC is the most mainstream and latest one at present.But it has always been challenged by other staging systems,such as the JGCA in Japan.The prognosis of GC patients with the same stage is still heterogeneous.In addition,according to the World Cancer Report 2021 of WHO,the number of new GC in China,Japan and South Korea accounts for 70%of the world,and China accounts for 43.9%of the world.On the contrary,GC is a rare cancer in western countries such as the United States.Therefore,China should provide important evidence for the improvement of diagnosis and treatment of GC.In this context,we aimed to explore the most critical clinical prognostic factors after D2 radical resection from a prospectively collected credible database,then to explore the most critical molecular mechanism and to construct a new prognostic risk model of GC,guided by the critical clinical prognostic factor.Methods:All Stage Ⅰ-Ⅲ GC patients’ medical records were retrospectively reviewed from a prospectively collected database in our country.A nested case-control study was included in this study.Chi-squared test and Fisher’s exact test were used for univariate analysis;Logistics regression analysis was used for multivariate analysis;The Cox proportional regression model was used for the univariate and multivariate analyses of DFS.The log-rank test was used for comparing the Kaplan Meier curves.A basic study was carried out to explore the molecular mechanism after a clinical study.To explore the most critical clinical prognostic factors after D2 radical resection,then to explore the most critical molecular mechanism in TCGA database using bioinformatics analysis,guided by the most critical clinical prognostic factors.And basic experiments were carried out to verify it.Finally,we aimed to construct a new prognostic risk model of GC based on the key genes obtained.Results:Part 1.Multivariate analysis confirmed that the LNM,lymphovascular or nerve invasion and operative type were the independent prognostic factors for DFS(P<0.05),among these,the LNM’s difference was the most significant prognostic factor(P<0.01).Part 2.Multivariate regression analyses revealed that the increases of group No.12a and 5 LN were significant(P<0.05)in group 2D,the increases of group No.12a,8a and 6 LN were significant(P<0.05)in group 2T,respectively.Among these,the increases of group No.12a LNs were the most significant(P<0.01).Part 3.Multivariate analysis confirmed that only the LNM,lymphovascular or nerve invasion and operative type were the significant difference between the two groups(P<0.05).And between the group 12a-and group 12a+,the tumor size,lauren type,lymphovascular or nerve invasion,and N stage had a significant difference.Between the group NO’ and group N1-3’,only the immunohistochemical VEGF and Ki-67 had a significant difference.Part 4.Bioinformatics analysis showed that 6 miRNAs were most closely related to the prognosis of GC.Among them,the expression level and fold change of mir-675 were both the most significant.The expression level of mir-675 was not closely related to age,clinical stage,depth of invasion and distant metastasis,but closely related to LNM.A PCR experiment confirmed that mir-675 was highly expressed in GC tissue.Then we screened seven key target genes of mir-675 and constructed a new and effective prognostic risk model of GC.Combined with the histological expression of the gene,another analysis was carried out,which showed that GPX3 may be the most optimal target gene of mir-675.A further Western Blot experiment confirmed that GPX3 was lowly expressed in GC tissue.Conclusions:1.Perigastric LNM was an independent poor prognostic factor of stage Ⅰ-Ⅲ GC.The prognostic importance of LNM in the 8th edition TNM staging system might need to be further improved.2.The metastasis of No.12a LN might be a very important poor prognostic factor of stage Ⅲ GC.3.The diameter of the tumor≥4 cm,low or undifferentiation,vascular or nerve invasion,stage T3-4 tumor,and the immunohistochemical VEGF and Ki-67 were good predictors of perigastric LNM in stage Ⅰ-Ⅲ GC.4.mir-675 and its most optimal target gene GPX3 might be most closely related to the LNM and prognosis of the GC patients.Further experiments confirmed that there was a significant difference between the two in cancerous and adjacent normal tissues.5.A new prognostic risk model of GC was constructed guided by the LNM. |