CT Angiography-Based Radiomics for Classification of Intracranial Aneurysm Rupture

Abstract

Background: Intracranial aneurysm rupture is a devastating medical event with a high morbidity and mortality rate. Thus, timely detection and management are critical. The present study aimed to identify the aneurysm radiomics features associated with rupture and to build and evaluate a radiomics classification model of aneurysm rupture. Methods: Radiomics analysis was applied to CT angiography (CTA) images of 393 patients [152 (38.7%) with ruptured aneurysms]. Patients were divided at a ratio of 7:3 into retrospective training (n = 274) and prospective test (n = 119) cohorts. A total of 1,229 radiomics features were automatically calculated from each aneurysm. The feature number was systematically reduced, and the most important classifying features were selected. A logistic regression model was constructed using the selected features and evaluated on training and test cohorts. Radiomics score (Rad-score) was calculated for each patient and compared between ruptured and unruptured aneurysms. Results: Nine radiomics features were selected from the CTA images and used to build the logistic regression model. The radiomics model has shown good performance in the classification of the aneurysm rupture on training and test cohorts [area under the receiver operating characteristic curve: 0.92 [95% confidence interval CI: 0.89-0.95] and 0.86 [95% CI: 0.80-0.93], respectively, p < 0.001]. Rad-score showed statistically significant differences between ruptured and unruptured aneurysms (median, 2.50 vs. -1.60 and 2.35 vs. -1.01 on training and test cohorts, respectively, p < 0.001). Conclusion: The results indicated the potential of aneurysm radiomics features for automatic classification of aneurysm rupture on CTA images.

Keywords: aneurysm rupture; intracranial aneurysm; machine learning; radiomics; subarachnoid hemorrhage.

Figure 1

Flowchart of the patients' inclusion and exclusion process.

Figure 2

Illustration of the study workflow. Clinical and imaging data of the potential candidates were collected and assessed for enrollment eligibility. The aneurysms were segmented by two radiologists using 3D slicer software. Seven groups of radiomics features were extracted using Artificial Intelligence Kit (AK). The feature number was reduced using a step-wise process. The selected radiomics features were used to build the radiomics model. The model was evaluated using the area under receiver operating characteristic curve on training and test cohorts.

Figure 3

Receiver operating characteristic (ROC) curves of the radiomics model performance on training (A) and test cohorts (B). Radiomics score (Rad-score) of each patient on training (C) and test cohorts (D) show the association of high Rad-score with risk of aneurysm rupture.

Figure 4

Example cases of unruptured and ruptured aneurysm features. Case 1: sagittal reconstructed maximum intensity projection (MIP, A) and 3D volume-rendered (3D-VR) CT angiography (CTA, B) images of a 52-year-old-male with unruptured aneurysm. The aneurysm was small (3.5 mm in maximal diameter), regular, and located on the internal carotid artery. Radiomics score was −6.1 indicating a low risk of rupture. Case 2: MIP (C) and 3D-VR CTA (D) images of a 50-year-old-male with ruptured aneurysm. The aneurysm measured 9.4 mm in maximal diameter, was irregular (lobulated with multiple daughter sacs), and located on the anterior communicating artery. Radiomics score was 3.5 indicating a high risk of rupture.