ObjectiveCancer of the pancreas is a life-threatening condition and has a high distant metastasis (DM) rate of over 50% at diagnosis. Therefore, this study aimed to determine whether patterns of distant metastases correlated with prognosis in pancreatic ductal adenocarcinoma (PDAC) with metastatic spread, and build a novel nomogram capable of predicting the 6, 12, 18-month survival rate with high accuracy.MethodsWe analyzed data from the Surveillance, Epidemiology, and End Results (SEER) database for cases of PDAC with DM. Kaplan-Meier analysis, log-rank tests and Cox-regression proportional hazards model were used to assess the impact of site and number of DM on the cancer-specific survival (CSS) and over survival (OS). A total of 2709 patients with DM were randomly assigned to the training group and validation group in a 7:3 ratio. A nomogram was constructed by the dependent risk factors which were determined by multivariate Cox-regression analysis. An assessment of the discrimination and ability of the prediction model was made by measuring AUC, C-index, calibration curve and decision curve analysis (DCA). In addition, we collected 98 patients with distant metastases at the time of initial diagnosis from Ningbo University Affiliated LiHuili Hospital to verify the efficacy of the prediction model.ResultsThere was a highest incidence of liver metastases from pancreatic cancer (2387,74.36%), followed by lung (625,19.47%), bone (190,5.92%), and brain (8,0.25%). The prognosis of liver metastases differed from that of lung metastases, and the presence of multiple organ metastases was associated with poorer prognosis. According to univariate and multivariate Cox-regression analyses, seven factors (i.e., diagnosis age, tumor location, grade of tumor differentiation, T-stage, receipt of surgery, receipt of chemotherapy status, presence of multiple organ metastases) were included in our nomogram model. In internal and external validation, the ROC curves, C-index, calibration curves and DCA were calculated, which confirmed that this nomogram can precisely predict prognosis of PDAC with DM.ConclusionMetastatic PDAC patients with liver metastases tended to have a worse prognosis than those with lung metastases. The number of DM had significant effect on the overall survival rate of metastatic PDAC. This study had a high prediction accuracy, which was helpful clinicians to analyze the prognosis of PDAC with DM and implement individualized diagnosis and treatment.

IntroductionSurgery represents a primary therapeutic approach for borderline resectable and locally advanced pancreatic cancer (BR/LAPC). However, BR/LAPC lesions exhibit high heterogeneity and not all BR/LAPC patients who undergo surgery can derive bene cial outcomes. The present study aims to employ machine learning (ML) algorithms to identify those who would obtain bene ts from the primary tumor surgery. MethodsWe retrieved clinical data of patients with BR/LAPC from the Surveillance, Epidemiology, and End Results (SEER) database and classi ed them into surgery and non-surgery groups based on primary tumor surgery status. To eliminate confounding factors, propensity score matching (PSM) was employed. We hypothesized that patients who underwent surgery and had a longer median cancer-speci c survival (CSS) than those who did not undergo surgery would certainly bene t from surgical intervention. Clinical and pathological features were utilized to construct six ML models, and model effectiveness was compared through measures such as the area under curve (AUC), calibration plots, and decision curve analysis (DCA). We selected the best-performing algorithm (i.e., XGBoost) to predict postoperative bene ts. The SHapley Additive exPlanations (SHAP) approach was used to interpret the XGBoost model.Additionally, data from 53 Chinese patients prospectively collected was used for external validation of the model. ResultsAccording to the results of the 10-fold cross-validation in the training cohort, the XGBoost model yielded the best performance (AUC = 0.823, 95%CI 0.707-0.938). The internal (74.3% accuracy) and external (84.3% accuracy) validation demonstrated the generalizability of the model. The SHAP analysis provided explanations independent of the model, highlighting important factors related to postoperative survival bene ts in BR/LAPC, with age, chemotherapy, and radiation therapy being the top three important factors. ConclusionBy integrating of ML algorithms and clinical data, we have established a highly e cient model to facilitate clinical decision-making and assist clinicians in selecting the population that would bene t from surgery.

Introduction Surgery represents a primary therapeutic approach for borderline resectable and locally advanced pancreatic cancer (BR/LAPC). However, BR/LAPC lesions exhibit high heterogeneity and not all BR/LAPC patients who undergo surgery can derive beneficial outcomes. The present study aims to employ machine learning (ML) algorithms to identify those who would obtain benefits from the primary tumor surgery.Methods We retrieved clinical data of patients with BR/LAPC from the Surveillance, Epidemiology, and End Results (SEER) database and classified them into surgery and non-surgery groups based on primary tumor surgery status. To eliminate confounding factors, propensity score matching (PSM) was employed. We hypothesized that patients who underwent surgery and had a longer median cancer-specific survival (CSS) than those who did not undergo surgery would certainly benefit from surgical intervention. Clinical and pathological features were utilized to construct six ML models, and model effectiveness was compared through measures such as the area under curve (AUC), calibration plots, and decision curve analysis (DCA). We selected the best-performing algorithm (i.e., XGBoost) to predict postoperative benefits. The SHapley Additive exPlanations (SHAP) approach was used to interpret the XGBoost model. Additionally, data from 53 Chinese patients prospectively collected was used for external validation of the model.Results According to the results of the 10-fold cross-validation in the training cohort, the XGBoost model yielded the best performance (AUC = 0.823, 95%CI 0.707–0.938). The internal (74.3% accuracy) and external (84.3% accuracy) validation demonstrated the generalizability of the model. The SHAP analysis provided explanations independent of the model, highlighting important factors related to postoperative survival benefits in BR/LAPC, with age, chemotherapy, and radiation therapy being the top three important factors.Conclusion By integrating of ML algorithms and clinical data, we have established a highly efficient model to facilitate clinical decision-making and assist clinicians in selecting the population that would benefit from surgery.