Comparison of Ruptured Intracranial Aneurysms Identification Using Different Machine Learning Algorithms and Radiomics.

Different machine learning algorithms have different characteristics and applicability. This study aims to predict ruptured intracranial aneurysms by radiomics models based on different machine learning algorithms and evaluate their differences in the same data condition. A total of 576 patients with intracranial aneurysms (192 ruptured and 384 unruptured intracranial aneurysms) from two institutions are included and randomly divided into training and validation cohorts in a ratio of 7:3. Of the 107 radiomics features extracted from computed tomography angiography images, seven features stood out. Then, radiomics features and 12 common machine learning algorithms, including the decision-making tree, support vector machine, logistic regression, Gaussian Naive Bayes, k-nearest neighbor, random forest, extreme gradient boosting, bagging classifier, AdaBoost, gradient boosting, light gradient boosting machine, and CatBoost were applied to construct models for predicting ruptured intracranial aneurysms, and the predictive performance of all models was compared. In the validation cohort, the area under curve (AUC) values of models based on AdaBoost, gradient boosting, and CatBoost for predicting ruptured intracranial aneurysms were 0.889, 0.883, and 0.864, respectively, with no significant differences among them. Of note, the performance of these models was significantly superior to that of the other nine models. The AUC of the AdaBoost model in the cross-validation was within the range of 0.842 to 0.918. Radiomics models based on the machine learning algorithms can be used to predict ruptured intracranial aneurysms, and the prediction efficacy differs among machine learning algorithms. The boosting algorithms might be superior in the application of radiomics combined with the machine learning algorithm to predict aneurysm ruptures.

Grey matter connectome abnormalities and age-related effects in antipsychotic-naive schizophrenia.

BACKGROUND: Convergent evidence is increasing to indicate progressive brain abnormalities in schizophrenia. Knowing the brain network features over the illness course in schizophrenia, independent of effects of antipsychotic medications, would extend our sight on this question. METHODS: We recruited 237 antipsychotic-naive patients with schizophrenia range from 16 to 73 years old, and 254 healthy controls. High-resolution T1 weighted images were obtained with a 3.0T MR scanner. Grey matter networks were constructed individually based on the similarities of regional grey matter measurements. Network metrics were compared between patient groups and healthy controls, and regression analyses with age were conducted to determine potential differential rate of age-related changes between them. FINDINGS: Nodal centrality abnormalities were observed in patients with untreated schizophrenia, particularly in the central executive, default mode and salience networks. Accelerated age-related declines and illness duration-related declines were observed in global assortativity, and in nodal metrics of left superior temporal pole in schizophrenia patients. Although no significant intergroup differences in age-related regression were observed, the pattern of network metric alternation of left thalamus indicated higher nodal properties in early course patients, which decreased in long-term ill patients. INTERPRETATIONS: Global and nodal alterations in the grey matter connectome related to age and duration of illness in antipsychotic-naive patients, indicating potentially progressive network organizations mainly involving temporal regions and thalamus in schizophrenia independent from medication effects. FUNDING: The National Natural Science Foundation of China, Sichuan Science and Technology Program, the Fundamental Research Funds for the Central Universities, Post-Doctor Research Project, West China Hospital, Sichuan University , the Science and Technology Project of the Health Planning Committee of Sichuan, Postdoctoral Interdisciplinary Research Project of Sichuan University and 1.3.5 Project for Disciplines of Excellence, West China Hospital, Sichuan University.

Morphological alterations of the corpus callosum in antipsychotic-naive first-episode schizophrenia before and 1-year after treatment.

OBJECTIVE: The corpus callosum (CC) is known to be altered in patients with schizophrenia. However, its morphologic characteristics are less well studied in treatment-naive first-episode schizophrenia patients, as is the effect of antipsychotic treatment on this structure. METHODS: T-1 weighted MRI scans were obtained from 160 antipsychotic-naïve first-episode schizophrenia patients (AN-FES) and 155 healthy controls (HCs) before treatment initiation. Among the patients, forty-four were available for follow-up studies after one year of antipsychotic treatment, and were divided into good-outcome (n = 31) and poor-outcome subgroups (n = 13) based on whether there was a 50% reduction in Positive and Negative Symptom Scale (PANSS) total scores from baseline. A computer algorithm was applied to automatically identify the mid-sagittal plane (MSP) and obtain morphological measurement parameters of the CC. RESULTS: Compared with HCs, AN-FES patients showed a significant reduction of thickness in the posterior midbody of the CC. This deficit was correlated with severity of negative symptoms. After one year of antipsychotic treatment, there was no significant change in CC morphological measurements in schizophrenia patients, nor was there a significant difference of CC morphological measurements between good-outcome and poor-outcome subgroups at baseline or at 1-year follow-up. CONCLUSION: Thickness of the posterior midbody of the CC is reduced in the early course of schizophrenia before treatment. This alteration was not affected by antipsychotic treatment and was unrelated to treatment outcome at 1-year.

Gray Matter Differences Between Premature Pubertal Girls With and Without the Reactivation of the Hypothalamic-Pituitary-Gonadal Axis.

The onset of puberty and related hormones exerts significant effects on brain morphometric and psychosocial development. The biological mechanisms underlying how the reactivation of the hypothalamic-pituitary-gonadal (HPG) axis and puberty-related hormonal maturation sculpts human brain architecture remain elusive. To address this question, 105 premature pubertal girls (age 8-11 years) without menstruation underwent brain structural scanning on a 3T MR system, and the luteinizing hormone releasing hormone (LHRH) stimulation test was used to identify the reactivation of the HPG axis. Among the 105 girls, 63 were positive for HPG axis reactivation (HPG+), while the others showed negative (HPG-). Cortical thickness was calculated and compared between the two groups after adjusting for age. The brain regions showing inter-group differences were then extracted and correlated with the peak value of serum hormone after the LHRH stimulation test in entire sample. Compared to HPG- girls, HPG+ girls showed reduced cortical thickness mainly in the the right precuneus, right inferior temporal gyrus, and right superior frontal gyrus, while increased cortical thickness primarily in the left superior parietal lobe and right inferior parietal lobe. Linear-regression analysis revealed negative correlations between the cortical thickness of the right inferior parietal lobe with the peak value of FSH and the right precuneus with LH and E. These findings provide evidence to support the notion that the reactivation of HPG axis and changes of hormones during the early phase of hormonal maturation exert influences on the development of gray matter.

Changes of Brain Structure in Patients With Metastatic Non-Small Cell Lung Cancer After Long-Term Target Therapy With EGFR-TKI.

PURPOSE: Epidermal growth factor receptor-tyrosine kinase inhibitor (EGFR-TKI) therapy is the routine treatment for patients with metastatic non-small cell lung cancer (NSCLC) harboring positive EGFR mutations. Patients who undergo such treatment have reported cognitive decline during follow-up. This study, therefore, aimed to evaluate brain structural changes in patients receiving EGFR-TKI to increase understanding of this potential symptom. METHOD: The medical records of 75 patients with metastatic NSCLC (without brain metastasis or other co-morbidities) who received EGFR-TKI therapy from 2010 to 2017 were reviewed. The modified Scheltens Visual Scale and voxel-based morphometry were used to evaluate changes in white matter lesions (WML) and gray matter volume (GMV), respectively. RESULTS: The WML scores were higher at the 12-month [8.65 ± 3.86; 95% confidence interval (CI), 1.60-2.35; p < 0.001] and 24-month follow-ups (10.11 ± 3.85; 95% CI, 2.98-3.87; p < 0.001) compared to baseline (6.68 ± 3.64). At the 24-month follow-up, the visual scores were also significantly higher in younger patients (3.89 ± 2.04) than in older patients (3.00 ± 1.78; p = 0.047) and higher in female patients (3.80 ± 2.04) than in male patients (2.73 ± 1.56; p = 0.023). Additionally, significant GMV loss was observed in sub-regions of the right occipital lobe (76.71 voxels; 95% CI, 40.740-112.69 voxels), left occipital lobe (93.48 voxels; 95% CI, 37.48-149.47 voxels), and left basal ganglia (37.57 voxels; 95% CI, 21.58-53.57 voxels) (all p < 0.005; cluster-level false discovery rate < 0.05). CONCLUSIONS: An increase in WMLs and loss of GMV were observed in patients with metastatic NSCLC undergoing long-term EGFR-TKI treatment. This might reflect an unknown side-effect of EGFR-TKI treatment. Further prospective studies are necessary to confirm our findings.