Abstract
1.1Objective
According to relevant studies, the incidence of pancreatic tumors has increased in both Asian and Western countries [1]. Surgical resection is the cornerstone of treatment for this aggressive disease [2] [3]. According to relevant studies, postoperative pancreatic fistula (POPF) in pancreatic tumors is one of the most serious complications after surgery, which seriously affects the course of patients' treatment and their prognosis [2] [3]. In the present study, we aime to identify the risk factors associated with clinically relevant postoperative pancreatic fistula (CR-POPF) based on systemic inflammatory markers (SIB), and preoperative, intraoperative, and postoperative dimensions, and to establish a multidimensional columnar graphical model for predicting postoperative pancreatic fistula (POPF) in pancreatic tumors [4] [5].
1.2Methods
Ninety patients who underwent surgery for pancreatic tumors at the Department of Hepatobiliary Surgery of the First Affiliated Hospital of Dalian Medical University between November 21, 2022 and November 21, 2023 were retrospectively studied. This study has been approved by the Ethics Committee of the First Affiliated Hospital of Dalian Medical University, and the relevant guidelines and regulations have been strictly followed.The collected clinical data were first processed by feature engineering and analyzed with relevant statistics such as chi-square test according to the criteria described in the Supplementary file, and risk factors with P-Value ≤ 0.05were selected. Then, the data set was randomly divided into training set and validation set according to 7:3, and then batch one-way logistic regression analysis was performed and risk factors with P-Value ≤ 0.05 were selected, and then multifactorial logistic regression analysis was performed and the optimal multivariate regression model was constructed. Based on the results of multifactors logistic regression analysis, a multidimensional column-line diagram was developed . The Bootstrap method was used to validate the model in a relevant way, and the predictive performance was assessed using the subject work characteristics (ROC) curve, and the clinical application value of the model was verified using the decision curve analysis (DCA) and calibration (Calibration) curve.
1.3Results
The incidence of POPF in this study was 50.0% (45/90). Multivariate logistic regression analysis identified the following variables as independent risk factors for POPF: preoperative albumin level(ALB, OR:5.06,P=0.012), neutrophil to lymphocyte ratio (NLR, OR: 9.55,P < 0.001), and systemic immunoinflammatory markers (SII, OR: 0.006,P =4.58), the ratio of amylase concentration in drainage fluid to preoperative blood amylase concentration (DSAR, OR=26.73,P < 0.001). We create a multi-dimensional nomogram by combining the above risk factors. The multidimensional nomogram model shows better predictive value. AUC under receiver Operating characteristic (ROC) curve: training set AUC=0.889 > 0.50, validation set AUC=0.949 > 0.50. The results show that the multidimensional nomogram model has good predictive performance, and the DCA and Calibration analyses of training set and validation set also show that the multidimensional nomogram model has higher clinical net benefit and better fit.
1.4Conclusion
Based on SIB, the column diagram we constructed can scientifically, accurately and objectively predict the risk of POPF in patients after pancreatic tumor surgery, which can better assist clinicians to conduct scientific POPF risk assessment on patients about to undergo pancreatic tumor surgery, and timely conduct relevant clinical intervention, so as to better develop the mitigation strategy of pancreatic fistula and corresponding postoperative management. The quality of life and prognosis of patients after operation were improved.