Dual-Energy CT-Based Thrombus Radiomics Can Predict Functional Outcome of Intravenous Thrombolysis in Acute Ischemic Stroke.

Abstract

To explore the predictive value of dual-energy CT-based thrombus radiomics for the functional outcome of intravenous thrombolysis in patients with acute ischemic stroke (AIS). One hundred four AIS patients who received intravenous thrombolysis were enrolled and classified into favorable and unfavorable outcome based on their modified Rankin Scale (mRS) scores at 90 days. All patients underwent a one-stop-shop CT scan upon admission, including NCCT, dual-energy CTA, and CTP. The thrombus radiological and radiomics models were developed using NCCT, CTA, and iodine overlay map (IOM) images. The clinical model was developed using clinical information and other radiological data. The best-performing radiomics model was selected for the further development of a clinical-radiomics nomogram. The performance of these models was evaluated using receiver operating characteristic (ROC) curves, clinical decision curves, calibration curves, and DeLong's test. The AUCs of the model_Thrombus built using the thrombus characteristics were lower than those of most radiomics models (training, 0.77; test, 0.75). The AUCs of the model_IOM were higher than those of model_CTA (training, 0.84 vs. 0.71; test, 0.78 vs. 0.66) and were comparable to model_NCCT (training, 0.84 vs. 0.82; test, 0.78 vs. 0.78). The model_NCCT+IOM demonstrated improved predictive performance compared to either single-sequence model alone (training, 0.92; test, 0.83). Systolic blood pressure and baseline NIHSS score were independent predictors of favorable outcome. Among all models, the nomogram has the highest predictive value (training, 0.96; test, 0.91). The thrombus radiomics model based on dual-energy CT can effectively predict functional outcome of intravenous thrombolysis in patients with AIS. The addition of clinical data to the model can improve predictive performance.