Construction and application of a nomogram model for predicting postoperative cerebral edema in meningiomas based on radiomics and clinical features

Background There is a lack of unified standard and effective methods for the diagnosis and treatment of postoperative cerebral edema. Purpose To test the effectiveness of a predictive model in the diagnostic and treatment strategies for postoperative cerebral edema in patients with a meningioma. Material and Methods A prediction model was constructed based on the data of 300 patients with a meningioma. The predictive model was used to evaluate the diagnosis and treatment effectiveness among another 100 patients. The 100 patients were randomly divided into a control group (n = 50) and an intervention group (n = 50). The control group received conventional diagnosis and treatment, and the intervention group was evaluated, diagnosed, and treated under the instruction of the prediction model. Results The calibration curves, decision curves, and receiver operating characteristic curves showed that the model had good calibration and good utility performance. A significant and effective rate of cerebral edema treatment was higher in the intervention group compared to the control group. In addition, a shorter time to cerebral edema regression, shorter hospital stay, lower cost, and lower incidence of postoperative complications characterized the intervention group compared to the control group ( P <0.05). Conclusion The prediction model based on radiomics and clinical features has a high classification performance and clinical utility. The diagnostic and therapeutic decision under this model can improve the therapeutic effect and outcome of patients with postoperative cerebral edema and reduce the hospitalization time and cost.