The Application Value of CT Radiomics Combined with Clinical Parameters in Predicting the Duration of Invasive Mechanical Ventilation in Patients with Traumatic Brain Injury

Background: Traumatic brain injury (TBI) is a commonly devastating injury that is a leading cause of neurological disability and mortality worldwide. Prolonged invasive mechanical ventilation (IMV) can cause complications for TBI patients. We explored the application value to predict the duration of IMV by using a quantitative radiomics approach combined with clinical parameters. Methods A retrospective analysis of 71 TBI patients was conducted. Among them, 30 patients (42.3%) showed IMV duration exceeding 7 days. The intraparenchymal hemorrhage (IPH) on the initial CT was manually segmented for radiomics analysis. After feature extraction and selection, a total of 5 different machine learning algorithms were used to predict IMV duration exceeding 7 days and to derive radiomics scores (R-scores). Univariate logistic regression analyses were implemented to screen clinical parameters. We conducted a comparative study of IVM for more than 7 days between the clinical factor alone and fusion radiomics features. Results The individual prediction of the radiomics model for predicting IMV duration exceeding 7 days resulted in an AUC of 0.8626. Clinical parameters were used to build a comparative model. For predicting IMV duration exceeding 7 days, the AUC generated by combining variables such as Glasgow Coma Scale (GCS) at the emergency department, Injury Severity Score (ISS), albumin count and neutrophil count at admission was 0.8496. After incorporating the R-scores into the combined model, the AUC increased to 0.8959. Conclusions The findings suggest that by performing radiomics analysis on initial CT images and combining clinical parameters, it is possible to effectively predict the duration of IMV duration exceeding 7 days in TBI patients.