Clinical Study Of LMR-PLR Scoring System For Predicting Long-Term Survival In Patients With Pancreatic Ductal Adenocarcinoma After Surgery

ObjectiveTo explore the predictive value of preoperative LMR-PLR scoring system based on lymphocyte to monocyte ratio(LMR)combined with platelet to lymphocyte ratio(PLR)for long-term survival in pancreatic ductal adenocarcinoma after surgery.MethodA retrospective study was used.A total of 167 patients who underwent pancreaticoduodenectomy in our hospital from November 2012 to November 2020 and were pathologically diagnosed as pancreatic ductal adenocarcinoma after surgery were collected.General information includes:gender and age;Laboratory test indicators include:lymphocyte,monocyte,platelet,albumin,globulin,total bilirubin,carcinoembryonic antigen,carbohydrate antigen 125 and carbohydrate antigen 199levels.Pathological data included tumor size,tumor stage,tumor differentiation degree,surgical margin,presence of vascular and nerve invasion.Treatment options include postoperative chemotherapy.Follow-up was carried out in outpatient clinic,by telephone and by case review.Follow-up included clinical and laboratory examinations to assess the survival of patients after surgery.The follow-up period is to March 2022.Use X-tile software to determine the best cutoff values of LMR,PLR,AGR,TBIL,CEA,CA125 and CA199.LMR-PLR is divided into three groups according to the best cutoff value and defined.Logistic regression was used to analyze the correlation between LMR,PLR and clinicopathological characteristics.Use R4.2.1 as the drawing tool to draw the box-type scatter map of the correlation between LMR and PLR and clinical pathological characteristics.UseX~2test to analyze the relationship between LMR-PLR scoring system and clinical pathological characteristics of patients.Use R4.2.1 as the drawing tool to draw Kaplan-Meier survival curve,and use Log-rank Test to compare the survival conditions between groups.Univariate and multivariate analyses of Cox proportional risk regression models were performed using SPSS 26.0,and Hazard ratio(HR)and 95%Confidence interval(95%CI)were calculated.Independent risk factors affecting Overall survival(OS)were obtained.Based on the screened independent risk factors affecting the prognosis of pancreatic cancer patients,a Nomogram model to predict the 1-3-year postoperative survival rate of pancreatic cancer patients was drawn by using R4.2.1 software,and a web page diagram was drawn.The Concordance index(C-index)was used to evaluate the differentiation of the model.The calibration curve was used to evaluate the consistency of the model,and the calculation method was Bootstrap method with 1000 resampling times.the Receiver operating characteristic curve(ROC)of the prediction model was drawn and its Area under the curve(AUC)was calculated.The clinical benefit was evaluated by Decision curve analysis(DCA).Result1.Use X-tile software to calculate the best cutoff values of LMR,PLR,Albumin to globulin ratio(AGR),TBIL,CEA,CA125 and CA199 as 1.9、115.3、1.2、267.4μmol/L、2.0 ng/m L、32.0 U/m L、954.6 U/m L respectively.According to these optimal cutoff values,LMR-PLR is divided into three groups and defined:LMR<1.9 and PLR≥115.3,LMR-PLR score is 0;LMR≥1.9 or PLR<115.3,LMR-PLR score is 1 point;LMR≥1.9 and PLR<115.3,LMR-PLR score is 2 points.2.Logistic regression was used to analyze the correlation between LMR and PLR and clinicopathologic features.The results showed that age<65 years old(P=0.041),AGR≥1.2(P=0.029)and Vascular invasion(Negative,P=0.032)were closely related to low LMR;CEA<2.0(P=0.002),and stageⅡ、Ⅲ(P=0.028,0.022)were closely related to low PLR.Using R 4.2.1 software,a box-type scatter map of the correlation between LMR and PLR and clinicopathological characteristics was made.The results showed that male(P<0.0001),AGR<1.2(P=0.039),PLR≥115.3(P<0.0001),TBIL≥267.4μmol/L(P=0.00083),low differentiation(P=0.0039),vascular invasion(positive,P=0.044)were closely related to LMR;CEA<2.0 ng/m L(P=0.015),LMR≥1.9(P=0.00017),TBIL<267.4μmol/L(P=0.00018),and chemotherapy status(P=0.047)are closely related to PLR.X~2test was used to analyze the relationship between LMR-PLR scoring system and clinicopathologic features of patients.The results showed that LMR-PLR scoring system was related to CEA of pancreatic cancer patients(P<0.05).3.The median survival time of 167 enrolled patients was 18 months.The 1-year,2-year and 3-year survival rates were 72.9%,38.7%and 14.0%,respectively.4.The results of univariate COX regression analysis showed that age,LMR-PLR score,TBIL,CEA,CA125,CA199,tumor maximum diameter,T stage,N stage,TNM stage,surgical margin and chemotherapy status were all related to the prognosis of pancreatic cancer patients(P<0.05).The factors related to the prognosis of pancreatic cancer(P<0.05)were further analyzed by multivariate COX regression analysis.The results suggested that age,LMR-PLR score,TBIL,CA199 and TNM stage were independent risk factors for pancreatic cancer prognosis(P<0.05).5.Based on the above independent prognostic risk factors,the Nomogram prognostic model of the overall survival rate after radical surgery for pancreatic cancer was established with R4.2.1 software,and the webpage version of the nomogram is constructed,thus realizing the dynamic network presentation of the nomogram.Using R4.2.1 as a mapping tool,draw the ROC of each index of Nomogram prognosis model for 1,2 and 3 years.The results showed:(1)One year after operation:Nomogram AUC=0.745,Age AUC=0.581,LMR-PLR score AUC=0.662,TBIL AUC=0.568,CA199 AUC=0.564,TNM stage AUC=0.584;(2)Two years after operation:Nomogram AUC=0.747,Age AUC=0.578,LMR-PLR score AUC=0.617,TBIL AUC=0.543,CA199 AUC=0.558,TNM stage AUC=0.668;(3)Three years after operation:Nomogram AUC=0.815,Age AUC=0.600,LMR-PLR score AUC=0.706,TBIL AUC=0.524,CA199 AUC=0.565,TNM stage AUC=0.684.Based on the bootstrap resampling method,the prediction efficiency of Nomogram’s prognostic model was verified.The calculated consistency index(C-index)was 0.698(95%CI:0.647-0.749),which showed that the prognostic model established in this study had good prediction ability.The calibration curves of 1-year,2-year and 3-year survival rates fitted well,indicating that the observed survival rates were in good agreement with the predicted survival rates;DCA shows that the model has high clinical practicability within a certain critical threshold.Conclusion1.According to the Logistic regression results,age<65 years old,AGR≥1.2 and Vascular invasion(Negative)are closely related to low LMR;CEA<2.0 ng/m L and stageⅡ、Ⅲwere closely related to low PLR.According to the results of box-type scatter diagram,male,AGR<1.2,PLR≥115.3,TBIL≥267.4μmol/L,low differentiation,vascular invasion(positive)are closely related to LMR;CEA<2.0,LMR≥1.9,TBIL<267.4μmol/L,and chemotherapy status are closely related to PLR.UsingX~2test analysis LMR-PLR scoring system and clinical pathologic features of patients with relationships.The results showed that LMR-PLR scoring system was related to CEA of pancreatic cancer patients.2.The independent risk factors influencing the prognosis of pancreatic cancer are age,LMR-PLR score,TBIL,CA199 and TNM stage.3.The LMR-PLR scoring system composed of LMR and PLR may be a new prognostic score for patients with pancreatic ductal carcinoma after radical surgery.By including age,LMR-PLR score,TBIL,CA199 and TNM stage into Nomogram,a more accurate and reliable prediction model can be obtained.