| Autor: |
Elena S. Sagues Sese, Sebastian Sanchez, Sricharan Veeturi, Tatsat Patel, Vicent M. Tutino, Diego J. Ojeda, Jacob M. Miller, Arshaq Saleem, Andres Gudino, Bincheng Wang, Edgar A. Samaniego |
| Jazyk: |
angličtina |
| Rok vydání: |
2023 |
| Předmět: |
|
| Zdroj: |
Stroke: Vascular and Interventional Neurology, Vol 3, Iss S2 (2023) |
| Druh dokumentu: |
article |
| ISSN: |
2694-5746 |
| DOI: |
10.1161/SVIN.03.suppl_2.045 |
| Popis: |
Introduction High resolution vessel wall imaging (HR‐VWI) enables accurate visualization of intracranial atherosclerotic plaques. Radiomics can be utilized as an objective quantification method of plaque appearance and shape. We aimed to analyze the radiomics features (RFs) obtained from 7T‐HR‐VWI to differentiate between culprit and non‐culprit plaques in patients with intracranial atherosclerotic disease (ICAD). Methods Patients with ICAD as stroke etiology undergoing HR‐VWI were included in the study. Culprit plaques in the vascular territory of the stroke were identified. The degree of stenosis, area degree of stenosis and plaque burden were calculated. Three‐dimensional segmentation of the plaque was performed, and RFs were extracted. We then evaluated multiple machine learning models to predict and identify culprit plaques using significantly different RFs. The dataset was then randomly divided into training and testing datasets, and the trained model was evaluated on the independent testing dataset. Results The study included 33 patients with ICAD as the cause of stroke. Univariate analysis revealed 38 significantly different RFs between culprit and non‐culprit plaques in pre‐contrast MRI, 39 in post‐contrast MRI and 25 RFs that were different between pre and post contrast MRIs. Additionally, seven shape‐based RFs exhibited significant distinctions between the two plaque types. The random forest model achieved an accuracy rate of 81% (88% sensitivity and 75% specificity) in identifying culprit plaques in the independent testing dataset. This model successfully identified the culprit plaque in 7 out of 8 cases during the testing phase. Conclusion In this study symptomatic culprit plaques had a different signature RFs compared to other plaque within the same subject. A machine learning model built with RFs successfully identified the symptomatic atherosclerotic plaques in most cases. Radiomics is a promising tool for stratification of plaques in patients with ICAD. |
| Databáze: |
Directory of Open Access Journals |
| Externí odkaz: |
|
Plný text