Your browser doesn't support javascript.
loading
Identification of culprit plaques characteristics of intracranial atherosclerosis: a radiomic study / 国际脑血管病杂志
International Journal of Cerebrovascular Diseases ; (12): 252-259, 2019.
Article in Chinese | WPRIM | ID: wpr-751545
ABSTRACT
Objective To investigate the ability of quantitative radiomic method based on highresolution magnetic resonance imaging (HR-MRI) to distinguish between culprit plaques and non-culprit plaques of intracranial atherosclerosis.Methods Patients with middle cerebral artery and basilar artery stenosis underwent HR-MRI in Changhai Hospital Affiliated to the Naval Medical University from September 2013 to October 2016 were analyzed retrospectively.The minimum lumen area,plaque burden,severity of luminal stenosis,intraplaque hemorrhage (IPH),enhancement rate,and 109quantitative radiomic characteristics of the culprit and non-culprit plaques were measured.For clinical features and traditional plaque morphology,multivariate logistic regression models were used to determine independent risk factors for culprit plaque.A random forest-supervised machine learning method was used to determine the radiomic characteristics of distinguishing between symptomatic plaques and asymptomatic plaques.The receiver operating characteristic (ROC) curve was constructed,and the diagnostic efficacy was described by the area under the curve (AUC).Results During the study,158 subjects were enrolled,and they aged (59.42± 11.62) years.The plaques of 75 patients were located in middle cerebral artery,and the plaques of 83 patients were located in basilar artery.There were 111 symptomatic patients and 47 asymptomatic patients.Multivariate logistic regression analysis showed that smoking (odds ratio [OR] 2.724,95% confidence interval [CI] 1.200-6.183),IPH (OR 11.340,95% CI 1.441-89.221),and enhancement rate (OR 6.865,95% CI 1.052-44.802) were the independent risk factors for culprit plaques.The AUC of these three characteristics for predicting symptomatic plaques were 0.605,0.584,and 0.590,respectively.The combination of the three cloud improve the test efficacy for the intracranial atherosclerotic culprit plaques,AUC could reached 0.714.Radiomic analysis showed that 22 radiomic characteristics extracted from T-2 weighted imaging,T1 weighted imaging,and contrast-enhanced T1 weighted imaging were associated with the culprit plaques.Their AUCs were 0.801,0.835,and 0.846,respectively.After the combination of all morphological and radiomic characteristics,AUC could reach 0.976,the accuracy rate was 87.4%.However,the difference was not statistically significant compared to the combined AUC of all radiomic characteristics (0.953) (P=0.275).Conclusion Radiomic analysis could accurately distinguish between the culprit plaques and non-culprit plaques of intracranial atherosclerosis,and is superior to the traditional morphological methods.

Full text: Available Index: WPRIM (Western Pacific) Type of study: Diagnostic study / Prognostic study Language: Chinese Journal: International Journal of Cerebrovascular Diseases Year: 2019 Type: Article

Similar

MEDLINE

...
LILACS

LIS

Full text: Available Index: WPRIM (Western Pacific) Type of study: Diagnostic study / Prognostic study Language: Chinese Journal: International Journal of Cerebrovascular Diseases Year: 2019 Type: Article