Radiomic features as a risk factor for early postoperative seizure in patients with meningioma.

PURPOSE

This study aim to identify the clinical risk factors of and to develop a radiomics-based predictive model for early postoperative seizure.

METHODS

We retrospectively assessed 322 operative patients with meningioma who met the inclusion criteria from January 2014 to December 2016 at The First Affiliated Hospital of Wenzhou Medical University. Univariate and multivariate analyses were performed to determine the predictive value of clinical variables. Magnetic resonance imaging (MRI) was performed to obtain the radiomic score (Rscore) for early postoperative seizure. Radiological features were evaluated using the AK software. The minimal redundancy (mRMR) and least absolute shrinkage and selection operator (LASSO) methods were used to assess for radiomic features, and the Rscore was obtained based on radiomic characteristics using a specific formula.

RESULTS

In total, 260 patients who met the inclusion criteria were finally enrolled in this study. Among them, 20 experienced early postoperative seizure. Logistic regression analysis showed that Rscore was associated with a significantly high risk of seizure (p<0.000). Receiver operating characteristic (ROC) curve analysis revealed that the area under the ROC curve of the Rscore was 0.92 (95% confidence interval: 0.853-0.987). The model had a high accuracy for predicting early postoperative seizure.

CONCLUSIONS

The Rscore was found to be associated with a high risk of early postoperative seizures. Thus, a higher Rscore (>-1.644) can identify high-risk patients requiring intensive care.