Zhongguo linchuang yanjiu (Sep 2023)
Predictive value of preoperative inflammatory markers in gastric cancer with lymphovascular invasion
Abstract
Objective To investigate the predictive value of preoperative hematological inflammatory markers in gastric cancer (GC) with lymphovascular invasion(LVI). Methods A retrospective study was conducted on 341 GC patients who undergoing surgical resection and confirmed by pathological examination in Jiangsu Cancer Hospital from October 2020 to July 2022. The patients were divided into a positive group (n=163) and a negative group (n=178) based on the presence or absence of vascular cancer thrombi in postoperative pathology. Before operation, the following clinical data were collected, including neutrophil to lymphocyte ratio (NLR), platelet-lymphocyte ratio (PLR), lymphocyte-monocyte ratio(LMR), fibrinogen-lymphocyte ratio (FLR) and common tumor markers. The receiver operating characteristic (ROC) curve was used to evaluate the predictive ability of NLR, PLR and FLR for LVI to find the optimal cut-off values. Univariate logistic regression was used to analyze the influencing factors of LVI in GC patients, and a logistic regression model for predicting LVI in GC patients was constructed. Results Compared with those in the patients with LVI negative, the levels of NLR, PLR, FLR and carbohydrate antigen (CA)19-9, the tumor infiltration depth and the rate of lymph node metastasis rate significantly increased in LVI positive patients(P＜0.05). ROC curve analysis showed that the best cutoff values of NLR, PLR and FLR were 1.68, 115.29 and 1.96 respectively. The logistic regression model of LVI in GC patients showed that Logit(P)=-2.117+0.890×X1+0.286×X2+0.917×X3(X1=PLR, X2=cT, X3=lymph node metastasis) had a good model fit(χ=58.405, P＜0.01), with the prediction accuracy of 67.26% and an AUC of 0.733 (95%CI: 0.680-0.786, P＜0.01), and this model analysis suggested that PLR(＞115.29, OR=2.436, 95%CI: 1.386-4.282, P=0.002), cT stage (OR=1.331, 95%CI: 1.031-1.749, P=0.030) and lymph node metastasis(OR=2.503, 95%CI: 1.507-4.156, P＜0.001) were the independent risk factors for LVI in patients with GC. Conclusion Preoperative monitoring of PLR level has a certain predictive value in evaluating the occurrence of LVI in GC patients, and can be used as a supplementary indicator to judge the formation of LVI.
Keywords
