Radiomics Features in Contrast‐Enhanced and Nonenhanced Magnetic Resonance Imaging Images Are Associated With High Intracranial Aneurysmal Risk

BACKGROUND Aneurysm wall enhancement is a potential imaging biomarker for risk stratification of intracranial aneurysms (IAs). Variations in the texture magnetic resonance (MRI) signal could shed light on underlying pathobiology aneurysm wall. Radiomics can help quantify textural complexity MRI images, which lead to better understanding and IAs. Herein, we investigated use radiomics derived from nonenhanced contrast‐enhanced identify high‐risk IAs evaluated their performance different data sets. METHODS We obtained 126 centers extracted features each aneurysm. then built random forest model part 3‐T set based 5‐year population, hypertension, age, size aneurysm, earlier SAH another site (PHASES) score. tested this same set, 7‐T an external set. also performed multivariate analysis understand significance features. RESULTS found that 75 were significantly between high‐ low‐risk The had good when (accuracy, 90%; sensitivity, 86%; specificity, 92%); however, sets, it moderate 88%; 50%; 95% set; accuracy, 62%; 27%; 100% set). CONCLUSIONS show high accuracy identifying be used as tool quantifying enhancement.