The Clinicopathological Analysis Of Liver Metastasis In Gastric Cancer And The Establishments Of Predictive Models

Background and purposeDespite general declining incidence worldwide,gastric cancer(GC)continues to be one of the significant causes of cancer deaths globally,which induces enormous health problems.Although diagnosis and treatment of gastric cancer have hugely improved over the past 20 years,including the development of radical surgical resection,the application of molecularly targeted therapy drugs,and the rise of immunotherapy,most gastric cancer patients were asymptomatic at early stages and progressed to advanced stages once diagnosed.The liver is the most common site for hematogenous metastases of gastric cancer.The incidence of liver metastasis in advanced gastric cancer reaches 44%,and patients’ 5-year survival rate is approximately 10%.Some researchers reported that serum markers might not be helpful for early cancer.They are still beneficial in detecting recurrence and distant metastasis.Furthermore,it has been reported that in some patients,the levels of serum tumor markers increased within 2-3 months ahead of the imaging findings in gastric cancer patients with liver metastasis.However,there has been no convenient and practical method based on serum tumor markers and hematological indicators for clinicians to detect liver metastasis in gastric cancer patients.Whether those indicators play an essential role in detecting liver metastasis is not clear.This study aimed to identify risk factors for liver metastasis based on clinicopathological and laboratory results and constructed an applicable predictive and prognosis nomogram to guide treatment in GC patients with liver metastasis and promote public health development in our country.MethodsClinicopathological data of 1861 patients diagnosed with gastric cancer at the First Affiliated Hospital Zhengzhou University between September 2013 and October 2019 were collected.After screening inclusion and exclusion criteria,we eventually enrolled 1604 patients for the present study.Patients were divided into two groups depending on the presence or absence of liver metastasis.Use univariate and multivariate logistic regression to analyze the risk factors of liver metastasis in gastric cancer.According to patients’ follow-up in the liver metastasis group,prognostic factors were analyzed using Kaplan-Meier analysis,Log-rank test,and Cox regression models.Based on those statistically significant indictors,a predictive and prognosis model was established.Resultsthere were 395 gastric patients with liver metastasis.Based on the results of univariate binomial logistic regression analysis,age,smoking status,tumor location,Lauren classification,the level of AST,ALP,NLR,PLR,and AFP are related to liver metastasis in GC patients.All of the variables significantly associated with liver metastasis in the univariate logistic regression analysis were included in multivariate logistic regression analysis.It was confirmed that the elevated level of AST(OR=2.012,P=0.045),ALP(OR=2.550,P=0.005),NLR(OR=2.656,P=0.003),PLR(OR=2.013,P=0.040),and AFP(OR=2.927,P<0.001)were the independent predictors for liver metastasis in GC patients.Based on the multivariate logistic analysis results,a nomogram predicting the probability of liver metastasis in GC patients was developed.A calibration plot with Bootstraps resamples was performed to verify the nomogram’s accuracy.Moreover,the nomogram’s effectiveness was also evaluated by using the ROC curves.The sensitivity and specificity of it are 0.635.and 0.734,respectively.ROC analysis showed that the area under the receiver operating characteristic curve(AUC)was 0.737(95%CI:0.702-0.772),which is the same as the C Index,indicating that the nomogram is reliable.The median survival time of patients with gastric cancer was 12 months after the diagnosis of liver metastasis.According to the Kaplan-Meier analysis,alcohol consumption,the depth of gastric cancer invasion,the degree of differentiation,whether accepted radical surgery,the level of NLR,PLR,and CA125 were statistically different prognosis of GC patients with liver metastasis.Based on the multivariate Cox analysis results,a nomogram predicting liver metastasis’s 1-year survival rate in GC patients was developed.A calibration plot with Bootstrap resamples was performed to verify the nomogram’s accuracy.Moreover,the nomogram’s effectiveness was also evaluated by using the ROC curves.The sensitivity and specificity of it are 0.816 and 0.609,respectively.ROC analysis showed that the area under the receiver operating characteristic curve(AUC)was 0.786(95%CI:0.721-0.851),which is the same as the C Index,indicating that the nomogram is reliable.Conclusions1.AST≥ 40 u/l,ALP≥ 105 u/l,AFP≥ 10ng/ml,NLR≥ 2.42,PLR≥ 154.5 are independent risk factors of liver metastasis with gastric cancer.2.The deeper gastric cancer invasion,the worse degree of differentiation,without gastric cancer radical surgery,CA125≥ 35U/mL and NLR≥ 1.9,PLR≥ 156 were statistically different in the prognosis of gastric cancer patients with liver metastasis.3.Combined with those independent risk indicators,we constructed a predictive nomogram.Thus,it will relatively accurately predict the risk of liver metastasis in gastric cancer patients without a pathological diagnosis;establish a nomogram for predicting the 1-year survival rate of gastric cancer patients with liver metastasis.Thus,it can help clinicians make effective interventions to personalize treatment.