Prognostic Analysis And Related Prediction Model Of Endometrial Cancer Based On Inflammatory Factors And Tumor Markers

Objective:To investigate the relationship between clinicopathological characteristics and neutrophil(N),monocyte(M),lymphocyte(L),neutrophil/lymphocyte ratio(NLR),monocyte/lymphocyte ratio(MLR)and tumor indicators in patients with endometrial cancer.To construct a prognostic model with the factors affecting the prognosis of endometrial cancer to evaluate the prognosis of patients and guide treatment.Methods:276 patients with endometrial cancer who underwent surgical treatment in the First Affiliated Hospital of Soochow University from May 2013 to February 2020 were collected with complete clinicopathological data,and were followed up to collect survival information.The collected data were sorted out,and SPSS 26.0 and R was used for data analysis.Multiple interpolation analysis was performed for partial missing values,descriptive analysis was performed for clinicopathological data,Mann-Whitney U test was used for comparison between two groups that did not conform to normal distribution,and Kruskal-Wallis H test was used for comparison between multiple groups.Spearson correlation analysis of inter-group data lines that do not follow normal distribution.Kaplan-Meier univariate prognostic analysis and COX multivariate regression analysis were performed to screen out independent risk factors for endometrial cancer.A prognosis model was established and visualized by nomogram for each case of endometrial carcinoma to evaluate prognosis and predict survival.Results:1.Lymph node metastasis was positively correlated with NLR(P=0.035),MLR(P=0.031),CA125(P=0.002),CA199(P=0.002),HE4(P=0.028)and CEA(P=0.005).The correlation index was 0.145,0.149,0.211,0.215,0.152,0.195.2.FIGO staging was positively correlated with CA125(P=0.000011),CA199(P=0.009039)and HE4(P=0.000011),and the correlation index was 0.29,0.20 and 0.30.3.The infiltration depth of muscle layer had statistical significance with CA125(P=0.001),CA199(P=0.003)and HE4(P=0.003),and the correlation index was 0.22,0.20 and 0.19.4.Pathological types had statistical significance with NLR(P=0.027),MLR(P=0.008)and CA199(P=0.019).and the correlation index was 0.13,0.16 and-0.14.5.There was statistical significance between estrogen receptor(ER)and L(P=0.022),N(P=0.04)and HE4(P=0.049),and the correlation indexes were 0.14.0.12 and 0.12.respectively.6.Univariate prognostic analysis indicated that age(P=0.005),whether there was hypertension(P=0.05).FIGO stage(P=1.2028E-11),depth of muscle infiltration(P=0.00005),pathological classification(P=0.000362).histological grade(P=0.007),lymph node metastasis(P=0.12)),lymphatic vascular space infiltration(P=0.000064),extrautominal metastasis(P=1.9051E-8),ER(P=0.001),progesterone receptor(PR)(P=0.000142),P53(P=0.00119),CA125(P=0.017).HE4(P=0.031),CEA(P=0.008)was associated with prognosis.7.Multivariate COX risk model included hypertension,ER and CEA in the model(x2=70.74,P=3.2926E-9)functioned as an independent factor affecting the prognosis of endometrial cancer.8.The C-index value of internal verification in the column chart was 0.801.indicating the accuracy of model prediction.ROC curve predicted one-year survival rate(AUC=0.978),three-year survival rate(AUC=0.812)and five-year survival rate(AUC=0.817).Conclusions:1.NLR and MLR were positively correlated with lymph node metastasis and pathological type.CEA was positively correlated with lymph node metastasis.CA125,CA199 and HE4 were positively correlated with lymph node metastasis,FIGO stage and depth of myometrial invasion.CA199 was negatively correlated with pathological types.The levels of L,N and HE4 were positively correlated with ER.2.Age>55 years old,combined with hypertension,late FIGO stage,deep infiltration of muscle layer,pathological type Ⅱ,poorly differentiated histological grade,lymph node metastasis,lymphatic vasculature infiltration,extrauterine metastasis,ER negative,PR negative,P53 positive,CA125>40U/mL,HE4>88pmol/L,CEA>2ng/mL was associated with poor prognosis.3.COX regression model consisting of whether hypertension,ER and CEA can be used as a more accurate prognostic model,which can be visually displayed by the prognostic score of each patient and be used to evaluate the prognosis of patients in clinical work.