Analysis of clinical and pathological characteristics of patients with gastric stromal tumors and construction and validation of a prognostic nomogram model

Background : Explore the clinical and pathological characteristics of patients with gastric stromal tumors and the factors influencing postoperative recurrence, and establish a nomogram model to predict the recurrence free survival (RFS) of patients with gastric stromal tumors. Methods: The data of patients with gastric stromal tumors admitted to the 900th hospital of Joint Logistics Support Force from August 2009 to December 2020 were analyzed retrospectively. To analyze the clinical and pathological characteristics of patients with gastric stromal tumors undergoing surgery. The Kaplan Meier method was used to draw the survival curves to analyze the total survival time of patients, and log-rank test was used to analyze the comparison between groups. Logistic regression model and Cox regression model were used for univariate and multivariate analysis. A nomogram prediction model for predicting RFS in patients with gastric stromal tumors was constructed and verified by calibration curve and consistency curve. Results : Among 184 patients with gastric stromal tumor, abdominal pain was the most common clinical symptom, followed by gastrointestinal bleeding. In patients with gastric stromal tumors, the most common location of tumors is the stomach body, followed by the stomach floor and antrum; The diameter of tumor is 2.1 ~ 5 cm and the number of mitosis is ≤5/50 HPF. The 5-year recurrence rate of patients who regularly took imatinib for 3 years after operation was significantly lower than that of patients who did not take imatinib (14.16% vs. 43.80%, P<0.05), while the 5-year RFS was higher than that of patients who did not take imatinib (73.30% vs. 55.10%, P<0.05). Multivariate Logistic regression analysis showed that the modified NIH criteria, tumor necrosis and oral imatinib treatment were independent influencing factors for postoperative recurrence of gastric stromal tumors (P<0.05). Multivariate Cox regression analysis showed that the modified NIH criteria and oral imatinib treatment were independent influencing factors for postoperative RFS of gastric stromal tumors (P<0.05). Kaplan-meier method was used to calculate DFS and draw the survival curve of the correlation between the modified NIH criteria and oral imatinib treatment with the prognosis of gastric stromal tumor patients. The results showed that patients with higher modified NIH criteria and those without oral imatinib treatment had shorter DFS and worse prognosis. The factors (age, gender, tumor diameter, mitotic index, tumor rupture, tumor necrosis, modified NIH criteria, gastrointestinal bleeding, oral imatinib treatment, and surgical method) that will affect patients' RFS were selected to construct a nomogram for predicting RFS, and the consistency index (C-index) was 0.828 and 0.881, and the external verification C-index was 0.837. The calibration curve indicates that the nomogram has relatively accurate prediction ability. Conclusions : The first clinical symptoms of patients with gastric stromal tumor are abdominal pain and gastrointestinal bleeding. Patients with higher risk of modified NIH criteria, tumor necrosis and no oral imatinib treatment are prone to relapse. The higher the risk of modified NIH criteria and the shorter the RFS of patients who have not received oral imatinib treatment, the worse the prognosis of patients. For patients with medium and high risk gastric stromal tumor, it is recommended to carry out imatinib adjuvant therapy for 3 years or more after operation, which can effectively improve the prognosis of patients. In addition, the nomogram prediction model based on the factors affecting patients' RFS can effectively predict the 3-and 5-year recurrence-free survival rate, which is conducive to individualized diagnosis and treatment of patients' prognosis in clinic.