Machine learning prediction model for the rupture status of middle cerebral artery aneurysm in patients with hypertension: a Chinese multicenter study.

Background: Hypertension is a common comorbidity in patients with unruptured intracranial aneurysms and is closely associated with the rupture of aneurysms. However, only a few studies have focused on the rupture risk of aneurysms comorbid with hypertension. This retrospective study aimed to construct prediction models for the rupture of middle cerebral artery (MCA) aneurysm associated with hypertension using machine learning (ML) algorithms, and the constructed models were externally validated with multicenter datasets. Methods: We included 322 MCA aneurysm patients comorbid with hypertension who were being treated in four hospitals. All participants underwent computed tomography angiography (CTA), and aneurysm morphological features were measured. Clinical characteristics included sex, age, smoking, and hypertension history. Based on the clinical and morphological characteristics, the training datasets (n=277) were used to fit the ML algorithms to construct prediction models, which were externally validated with the testing datasets (n=45). The prediction performances of the models were assessed by receiver operating characteristic (ROC) curves. Results: The areas under the ROC curve (AUCs) of the k-nearest-neighbor (KNN), neural network (NNet), support vector machine (SVM) and logistic regression (LR) models in the training datasets were 0.83 [95% confidence interval (CI): 0.78-0.88], 0.87 (95% CI: 0.82-0.92), 0.91 (95% CI: 0.88-0.95), and 0.83 (95% CI: 0.77-0.88), respectively, and in the testing datasets were 0.74 (95% CI: 0.59-0.89), 0.82 (95% CI: 0.69-0.94), 0.73 (95% CI: 0.58-0.88), and 0.76 (95% CI: 0.61-0.90), respectively. The aspect ratio (AR) was ranked as the most important variable in the ML models except for NNet. Further analysis showed that the AR had good diagnostic performance, with AUC values of 0.75 in the training datasets and 0.77 in the testing datasets. Conclusions: The ML models performed reasonably accurately in predicting MCA aneurysm rupture comorbid with hypertension. AR was demonstrated as the leading predictor for the rupture of MCA aneurysm with hypertension.

Predicting the rupture status of small middle cerebral artery aneurysms using random forest modeling.

Objective: Small intracranial aneurysms are increasingly being detected; however, a prediction model for their rupture is rare. Random forest modeling was used to predict the rupture status of small middle cerebral artery (MCA) aneurysms with morphological features. Methods: From January 2009 to June 2020, we retrospectively reviewed patients with small MCA aneurysms (<7 mm). The aneurysms were randomly split into training (70%) and internal validation (30%) cohorts. Additional independent datasets were used for the external validation of 78 small MCA aneurysms from another four hospitals. Aneurysm morphology was determined using computed tomography angiography (CTA). Prediction models were developed using the random forest and multivariate logistic regression. Results: A total of 426 consecutive patients with 454 small MCA aneurysms (<7 mm) were included. A multivariate logistic regression analysis showed that size ratio (SR), aspect ratio (AR), and daughter dome were associated with aneurysm rupture, whereas aneurysm angle and multiplicity were inversely associated with aneurysm rupture. The areas under the receiver operating characteristic (ROC) curves (AUCs) of random forest models using the five independent risk factors in the training, internal validation, and external validation cohorts were 0.922, 0.889, and 0.92, respectively. The random forest model outperformed the logistic regression model (p = 0.048). A nomogram was developed to assess the rupture of small MCA aneurysms. Conclusion: Random forest modeling is a good tool for evaluating the rupture status of small MCA aneurysms and may be considered for the management of small aneurysms.

Classifying Ruptured Middle Cerebral Artery Aneurysms With a Machine Learning Based, Radiomics-Morphological Model: A Multicentral Study.

OBJECTIVE: Radiomics and morphological features were associated with aneurysms rupture. However, the multicentral study of their predictive power for specific-located aneurysms rupture is rare. We aimed to determine robust radiomics features related to middle cerebral artery (MCA) aneurysms rupture and evaluate the additional value of combining morphological and radiomics features in the classification of ruptured MCA aneurysms. METHODS: A total of 632 patients with 668 MCA aneurysms (423 ruptured aneurysms) from five hospitals were included. Radiomics and morphological features of aneurysms were extracted on computed tomography angiography images. The model was developed using a training dataset (407 patients) and validated with the internal (152 patients) and external validation (73 patients) datasets. The support vector machine method was applied for model construction. Optimal radiomics, morphological, and clinical features were used to develop the radiomics model (R-model), morphological model (M-model), radiomics-morphological model (RM-model), clinical-morphological model (CM-model), and clinical-radiomics-morphological model (CRM-model), respectively. A comprehensive nomogram integrating clinical, morphological, and radiomics predictors was generated. RESULTS: We found seven radiomics features and four morphological predictors of MCA aneurysms rupture. The R-model obtained an area under the receiver operating curve (AUC) of 0.822 (95% CI, 0.776, 0.867), 0.817 (95% CI, 0.744, 0.890), and 0.691 (95% CI, 0.567, 0.816) in the training, temporal validation, and external validation datasets, respectively. The RM-model showed an AUC of 0.848 (95% CI, 0.810, 0.885), 0.865 (95% CI, 0.807, 0.924), and 0.721 (95% CI, 0.601, 0.841) in the three datasets. The CRM-model obtained an AUC of 0.856 (95% CI, 0.820, 0.892), 0.882 (95% CI, 0.828, 0.936), and 0.738 (95% CI, 0.618, 0.857) in the three datasets. The CRM-model and RM-model outperformed the CM-model and M-model in the internal datasets (p < 0.05), respectively. But these differences were not statistically significant in the external dataset. Decision curve analysis indicated that the CRM-model obtained the highest net benefit for most of the threshold probabilities. CONCLUSION: Robust radiomics features were determined related to MCA aneurysm rupture. The RM-model exhibited good ability in classifying ruptured MCA aneurysms. Integrating radiomics features into conventional models might provide additional value in ruptured MCA aneurysms classification.

Predictors of Poor-Grade Aneurysmal Subarachnoid Hemorrhage Caused by Anterior Communicating Artery Aneurysm.

BACKGROUND: Patients with poor-grade aneurysmal subarachnoid hemorrhage (aSAH) are considered to have a poor prognosis. However, the underlying reason for the association between the aneurysmal characteristics and poor-grade aSAH is still unclear. In the present study, we retrospectively evaluated the independent risk factors for patients with anterior communicating artery (ACoA) aneurysms with poor-grade aSAH. METHODS: From January 2009 to January 2016, 477 consecutive patients with ruptured ACoA aneurysms were included in the present study. Poor-grade aSAH was defined as a World Federation of Neurosurgical Society grade of IV or V, and good-grade aSAH was defined as a grade of I-III. Univariate and multivariable regression analyses were used to investigate the differences in aneurysm morphology and clinical characteristics between the 2 groups. RESULTS: On univariate analysis, older patients (P = 0.038), larger aneurysm size (P = 0.013), larger size ratio (P = 0.007), larger aspect ratio (P = 0.009), positive history of stroke (P = 0.001), and posterior projection aneurysms (P = 0.001) were associated with poor-grade aSAH. Multivariate analyses revealed that older patients (odds ratio [OR], 1.654; 95% confidence interval [CI], 1.004-2.728; P = 0.048), larger size ratio (OR, 1.280; 95% CI, 1.111-1.475; P = 0.001), positive history of stroke (OR, 6.051; 95% CI, 1.712-21.381; P = 0.005), and posterior projection aneurysms (OR, 2.718; 95% CI, 1.607-4.598; P < 0.001) were independently associated with poor-grade aSAH. CONCLUSIONS: Poor-grade aSAH was independently associated with older patients, a larger size ratio, a positive history of stroke, and posterior projection aneurysms in patients with a ruptured ACoA aneurysm. These parameters could contribute to screening for patients with the potential for poor-grade aSAH.

Computed Tomography Perfusion Deficits during the Baseline Period in Aneurysmal Subarachnoid Hemorrhage Are Predictive of Delayed Cerebral Ischemia.

BACKGROUND AND PURPOSE: Delayed cerebral ischemia (DCI) is a frequent and fearful complication following aneurysmal subarachnoid hemorrhage (aSAH). The aim of this study is to assess the diagnostic accuracy of computed tomography perfusion (CTP) during an admission baseline period for the prediction of DCI. METHODS: Fifty-four aSAH cases were screened by baseline CTP within 3 days after aSAH and were reexamined with CTP 7-17 days after aSAH. Relative cerebral blood volume, relative cerebral blood flow (CBF), and relative mean transit time were measured. DCI was confirmed by a combination of noncontrast CT, CTP reexamination, and clinical assessment of neurologic deficits. Quantitative baseline and reexamination CTP data for all patients were compared between DCI and without DCI groups using Student's t-tests. The quantitative baseline and reexamination CTP data of DCI patients were compared using paired Student's t-tests. The χ2 test was used to evaluate incidences of DCI between different baseline relative CBF levels. The optimal cutoff value for each parameter was established by receiver operating characteristic curve analysis. RESULTS: Of the patients included in this study, 33.3% (18 of 54) developed DCI. There was a significant difference in the incidence of DCI among different baseline relative CBF subsets (χ2 = 38.00, P < .05). A relative CBF of .84 had the highest specificity and sensitivity of predicting DCI. CONCLUSION: CTP parameters during the baseline period can be helpful for the early identification of aSAH patients who are at high risk for DCI.

[CT perfusion imaging evaluation on hemodynamic changes of acute spontaneous intracerebral hemorrhage surrounding tissues].

OBJECTIVE: To discuss the hemodynamic changes in patients with acute supratentorial spontaneous intracerebral hemorrhage (within 72 hours) by using 320-slice of low-dose volume CT perfusion imaging. METHODS: Twenty-six patients of The First Affiliated Hospital of Wenzhou Medical University during December 2012 to December 2013 with acute supratentorial SICH diagnosed by plain CT scanning and clinic were enrolled. With hematoma maximum level for reference, the hematoma volume, edema area and perfusion defect area were measured, and the perfusion parameters values of the marginal area and outer area of the intracerebral hematoma and contralateral mirror area were measured, including cerebral blood flow (CBF), cerebral blood volume (CBV), mean transit time (MTT) and time-to-peak (TTP), and rCBF, rCBV, rMTT and rTTP were calculated by ipsilateral/contralateral value. RESULTS: The CBF, CBV of the marginal area were lower than the contralateral mirror area (tCBF=-8.125, tCBV=-8.671, PCBF, CBV<0.01); the MTT of the marginal area was shorter than the contralateral mirror area (tMTT=-3.246, PMTT<0.05); the TTP of the marginal area was longer than the contralateral mirror area (tTTP=5.027, PTTP<0.01). The CBV of the outer area was lower than the contralateral mirror area (tCBV=-2.337, PCBV<0.05); the MTT of the outer area was shorter than the contralateral mirror area (tMTT=-2.421, PMTT<0.05); the TTP of the outer area was longer than the contralateral mirror area (tTTP=2.077, PTTP<0.05). There was a siginificant relationship between the volume of acute hematoma and rCBV, rMTT, rTTP of the marginal area (rrCBV=-0.412, PrCBV<0.05, rrMTT=-0.437, PrMTT<0.05, rrTTP=0.475, PrMTT<0.05). Perihematomal CBF perfusion defect area showed a positive linear relation with the volume of acute hematoma (r=0.440, P<0.05). There was a positive linear relationship between the maximum level edema area and the hematoma volume, perihematomal CBF perfusion defect area (r=0.400, r=0.81, P<0.05). CONCLUSIONS: 320-slice of low-dose and volume CT perfusion imaging can perfectly reflect the hemodynamic changes in brain tissuse after acute supratentorial SICH. Hypoperfusion was appeared in perihematomal area of acute supratentorial SICH. The perihematomal brain tissue may exists ischemic injury associated with the size of hematoma.The hematoma place holder effect, ischemic injury are the important cause of acute brain edema formation.