Based On Naples Prognostic Score To Construct A Novel Nomogram To Predict Overall Survival Of Resected Pancreatic Cancer Patients

Objective:NPS score is an important prognostic evaluation index for tumor patients,and it has been demonstrated that NPS score is related to the prognosis of various tumors such as colon cancer,gastric cancer,breast cancer,and ovarian cancer.The purpose of this study was to investigate the relationship between NPS score and the prognosis of postoperative pancreatic cancer patients,and to construct a prognostic nomogram based on NPS score to predict the long-term survival of postoperative pancreatic cancer patients;and to compare the advantages and disadvantages of the line chart prognostic model with other prognostic scores for predicting long-term survival of patients.Methods:The clinical data of 214 patients who underwent radical resection of pancreatic cancer at the Second Affiliated Hospital of Kunming Medical University from January 2015 to December 2020 were retrospectively analyzed,and the patients’ blood and biochemical results within one week before surgery,postoperative pathological results:including gender,age,BMI,total bilirubin(TB),CA125,CA199,tumor size,peripheral nerve invasion,vascular cancer thrombus,pathological differentiation grade,lymph node metastasis,whether adjuvant therapy was performed,absolute neutrophil count,absolute lymphocyte count,absolute monocyte count,serum albumin,and total cholesterol.Patients were randomized into the training set(161)and the validation set(53)in a 7:3 ratio,and CONUT,PNI,SIS,and NPS scores were calculated and grouped according to preoperative neutrophil-to-lymphocyte ratio(NLR),lymphocyte-to-monocyte ratio(LMR),serum albumin,and total cholesterol to compare whether there was heterogeneity between the two data sets.Kaplan-Meier survival analysis was used to analyze the relationship between CONUT,PNI,SIS and NPS scores and patient survival time(OS);univariate and multivariate COX regression analyses were performed to screen independent prognostic factors in the training set,and nomogram prognostic models were constructed with independent prognostic factors at P<0.05 in the multivariate analysis.The area under the subject operating characteristic(ROC)curve(AUC)and C-index were used to compare the discriminative ability of each prognostic score with the nomogram model,and calibration curves and decision curves were used to verify the predictive ability of the nomogram model and validated in the validation set.Results:The univariate results showed that gender,lymph node metastasis,CA125,tumor size,nerve infiltration,vascular cancer embolism,chemotherapy,TNM stage,degree of tumor differentiation,NPS score,CONUT score,PNI score,and SIS score were prognostic factors for patients’ OS;the multifactorial results showed that CA125,chemotherapy,NPS score,degree of differentiation,and TNM stage were independent prognostic factors for patients’ Higher NPS score predicted worse prognosis of patients,and patients in NPS1 group or 2 group had worse OS in the prediction model compared with NPS0 group(NPS1 vs.NPS0:HR=2.277,95%CI:(1.106,4.685),p=0.025;NPS2 vs.NPS0:HR=4.232,95%CI:(2.031,8.819),P<0.001).In the training set,the AUC values for possible OS at 1,2 and 3 years for the columnar graph model were:0.841,0.796,0.781,C-index:0.734,respectively;the AUC values for possible OS at 1,2 and 3 years for the SIS score were:0.623,0.574,0.535,Cindex:0.55,respectively;the PNI score at 1,2 and 3 years for the AUC values were:0.579,0.548,0.546,C-index:0.562;AUC values for possible OS at 1,2,and 3 years for CONUT score were 0.656,0.589,0.521,Cindex:0.522;AUC values for possible OS at 1,2,and 3 years for the column line graph model in the validation set were 0.848,0.823,0.801,Cindex:0.796;AUC values for possible OS at 1,2,and 3 years for SIS score:0.753,0.642,0.641,C-index:0.621;AUC values for possible OS at 1,2,and 3 years for PNI score:0.721,0.636,0.592,C-index:0.602 The calibration curves showed acceptable agreement between the predicted OS and the actual OS of the nomogram model.the DCA curves indicated that the column line graph model has good clinical utility.Kaplan-Meier survival curves for the new nomogram prediction model scores demonstrated good risk stratification ability(p<0.001).Conclusion:NPS score is an important prognostic evaluation index for postoperative pancreatic cancer patients.The nomogram prognostic model based on NPS score is superior to other prognostic scores in predicting patients’ prognosis.The column line graph prognostic model with NPS score has significant ability to stratify patients’ prognosis and can improve patients’ prognosis by early intervention in high-risk patients.