Establishment Of A Methylation-Based Diagnostic Model Of Lymph Node Metastasis In Early Gastric Cancer And Its Clinical Research

Background:Lymph node metastatic status is not only a decision-making factor for the treatment of early gastric cancer(EGC)but also the most important risk factor for recurrence and progression.Current clinical evaluation methods are insufficient to evaluate lymph node metastasis(LNM)in EGC,leading to patients experiencing unnecessary overtreatment or undertreatment.This study developed and validated a diagnostic model based on qPCR fully quantitative DNA methylation to assess individual risk of LNM in EGC.Methods:Whole-epigenome sequencing was performed in 55 fresh frozen samples of EGC to screen for specific methylation markers to LNM.Highly sensitive Methylight qPCR was used to analyze differential methylation markers in 432 paraffin tissue sections of EGC.The samples were randomly divided into Train(n=151):Test(n=151):Validation(n=130)according to balancing the age、gender and clinical variables after quantifying the methylation level data based on qPCR.Logistic regression was used to construct an EGC LNM diagnostic model with methylation signatures and verify the performance of the model.The clinical application value of the model was evaluated by combining the preoperative clinic-pathological data.Results:The selected 55 differential methylation markers were used to construct the model.A 3-gene methylation LNM diagnostic model consisting of GNAS,FCGBP,and CCDC166 was constructed in 302 EGC samples with an AUC of 0.87(95%CI:0.80-0.93).The AUC of the model was 0.88(95%CI:0.80-0.95)in the validation cohort(n=130).Subsequently,the 3-gene methylation diagnostic model for EGC LNM was significantly superior to CT(AUC:0.64,95%CI:0.52-0.76),CEA(AUC:0.51,95%CI:0.38-0.64),CA19-9(AUC:0.51,95%CI:0.38-0.65),CA72-4(AUC:0.50,95%CI:0.37-0.63)and the traditional model(AUC:0.74,95%CI:0.68-0.81).In addition,the mixed model with preoperative CT images had a slightly higher AUC of 0.88(95%CI:0.81-0.94)in the model data set and 0.91(95%CI:0.85-0.97)in the validation set for the identification of lymph node metastatic status in early gastric cancer than the 3-gene methylation diagnostic model.When the 3-gene model was combined with the traditional model,the AUC of the combined model was 0.87(95%CI:0.82-0.92)in the model data set and 0.86(95%CI:0.78-0.94)in the validation set,significantly superior to the diagnostic ability of the traditional model alone.There was no statistical difference in diagnostic performance between the 3-gene model and the 3-gene model.When 3-gene methylation was combined with the conventional model,the specificity and accuracy for the diagnosis of EGC lymph node metastatic status were 89.0%and 85.6%,respectively.In addition,the specificity of the 3-gene methylation model and the combined model for the preoperative diagnosis of ulcerative early gastric cancer was 4 times higher than that of the traditional model.The specificity of the diagnosis of early gastric cancer with signet ring,mucinous,and poorly differentiated tissue types was improved by more than 50%compared with the conventional diagnosis.Conclusion:We screened 55 differential methylation markers of EGC LNM and validated them at the tissue level,and constructed and validated a new set of diagnostic models for 3-gene methylation.This model is a robust diagnostic tool,with good differential performance and clinical potential,and is expected to be a powerful tool for evaluating EGC LNM status and assisting in therapeutic decision-making.