Abstract:Objective To construct an early diagnosis model of acute pulmonary thromboembolism (PTE) based on chest pain center database. Methods According to the database of Chest Pain Center of Affiliated Hospital of Jining Medical University, the clinical data of patients who visited the emergency department from January to December 2020 were retrospectively collected, and the patients were divided into PTE group and non-PTE chest pain group according to whether diagnosed PTE. The clinically relevant indicators of the two groups were compared, and the study indicators with statistical differences between the two groups were included in the multivariate logistic regression analysis, and a nomogram model for the early diagnosis of PTE was established. The receiver operating characteristic (ROC) curve of the model was plotted to assess the predictive accuracy, and the model was tested for goodness of fit using the Hosmer-Lemeshow test. An additional 654 patients presenting with chest pain between January 2021 and March 2021 were collected to externally validate the model. Results A total of 2 738 patients were included for the construction of the model, of whom 117 (4.27%) had confirmed PTE. On multivariate analysis, a history of surgical trauma, ＞ 3 d of immobilization / bed rest on the lower extremities, with dyspnea, syncope, low pulse oxygen staturation (SpO₂) at admission, high D-dimer, right deviation of electrocardiographic axis, and complete right bundle branch block were independent factors for the diagnosis of PTE in patients with chest pain (P＜0.05) . ROC curve analysis showed that the area under the curve was 0.985 (95%CI: 0.969-0.999) in internal validation data and 0.924 (95%CI: 0.872-0.977) in external validation data, showing that the model had a good discrimination. The goodness of fit test was performed using Hosmer-Lemeshow, and validated internally (χ²=14.077, P=0.080) and external validation (χ²=615.690, P=0.986) both indicating a good fit of the model. Conclusion This study constructed a nomogram model for the diagnosis of PTE in patients with acute chest pain, and this model could effectively predict the risk probability of PTE in patients with acute chest pain.