Medical image foundation models in assisting diagnosis of brain tumors: a pilot study.

OBJECTIVES: To build self-supervised foundation models for multicontrast MRI of the whole brain and evaluate their efficacy in assisting diagnosis of brain tumors. METHODS: In this retrospective study, foundation models were developed using 57,621 enhanced head MRI scans through self-supervised learning with a pretext task of cross-contrast context restoration with two different content dropout schemes. Downstream classifiers were constructed based on the pretrained foundation models and fine-tuned for brain tumor detection, discrimination, and molecular status prediction. Metrics including accuracy, sensitivity, specificity, and area under the ROC curve (AUC) were used to evaluate the performance. Convolutional neural networks trained exclusively on downstream task data were employed for comparative analysis. RESULTS: The pretrained foundation models demonstrated their ability to extract effective representations from multicontrast whole-brain volumes. The best classifiers, endowed with pretrained weights, showed remarkable performance with accuracies of 94.9, 92.3, and 80.4%, and corresponding AUC values of 0.981, 0.972, and 0.852 on independent test datasets in brain tumor detection, discrimination, and molecular status prediction, respectively. The classifiers with pretrained weights outperformed the convolutional classifiers trained from scratch by approximately 10% in terms of accuracy and AUC across all tasks. The saliency regions in the correctly predicted cases are mainly clustered around the tumors. Classifiers derived from the two dropout schemes differed significantly only in the detection of brain tumors. CONCLUSIONS: Foundation models obtained from self-supervised learning have demonstrated encouraging potential for scalability and interpretability in downstream brain tumor-related tasks and hold promise for extension to neurological diseases with diffusely distributed lesions. CLINICAL RELEVANCE STATEMENT: The application of our proposed method to the prediction of key molecular status in gliomas is expected to improve treatment planning and patient outcomes. Additionally, the foundation model we developed could serve as a cornerstone for advancing AI applications in the diagnosis of brain-related diseases.

Whole brain morphologic features improve the predictive accuracy of IDH status and VEGF expression levels in gliomas.

Glioma is a systemic disease that can induce micro and macro alternations of whole brain. Isocitrate dehydrogenase and vascular endothelial growth factor are proven prognostic markers and antiangiogenic therapy targets in glioma. The aim of this study was to determine the ability of whole brain morphologic features and radiomics to predict isocitrate dehydrogenase status and vascular endothelial growth factor expression levels. This study recruited 80 glioma patients with isocitrate dehydrogenase wildtype and high vascular endothelial growth factor expression levels, and 102 patients with isocitrate dehydrogenase mutation and low vascular endothelial growth factor expression levels. Virtual brain grafting, combined with Freesurfer, was used to compute morphologic features including cortical thickness, LGI, and subcortical volume in glioma patient. Radiomics features were extracted from multiregional tumor. Pycaret was used to construct the machine learning pipeline. Among the radiomics models, the whole tumor model achieved the best performance (accuracy 0.80, Area Under the Curve 0.86), while, after incorporating whole brain morphologic features, the model had a superior predictive performance (accuracy 0.82, Area Under the Curve 0.88). The features contributed most in predicting model including the right caudate volume, left middle temporal cortical thickness, first-order statistics, shape, and gray-level cooccurrence matrix. Pycaret, based on morphologic features, combined with radiomics, yielded highest accuracy in predicting isocitrate dehydrogenase mutation and vascular endothelial growth factor levels, indicating that morphologic abnormalities induced by glioma were associated with tumor biology.

Association between patent foramen ovale and migraine: evidence from a resting-state fMRI study.

A relationship between migraine without aura (MO) and patent foramen ovale (PFO) has been observed, but the neural basis underlying this relationship remains elusive. Utilizing independent component analysis via functional magnetic resonance imaging, we examined functional connectivity (FC) within and across networks in 146 patients with MO (75 patients with and 71 patients without PFO) and 70 healthy controls (35 patients each with and without PFO) to elucidate the individual effects of MO and PFO, as well as their interaction, on brain functional networks. The main effect of PFO manifested exclusively in the FC among the visual, auditory, default mode, dorsal attention and salience networks. Furthermore, the interaction effect between MO and PFO was discerned in brain clusters of the left frontoparietal network and lingual gyrus network, as well as the internetwork FC between the left frontoparietal network and the default mode network (DMN), the occipital pole and medial visual networks, and the dorsal attention and salience networks. Our findings suggest that the presence of a PFO shunt in patients with MO is accompanied by various FC changes within and across networks. These changes elucidate the intricate mechanisms linked to PFO-associated migraines and provide a basis for identifying novel noninvasive biomarkers.

Distinctive intrinsic functional connectivity alterations of anterior cingulate cortex subdivisions in major depressive disorder: A systematic review and meta-analysis.

Evidence of whether the intrinsic functional connectivity of anterior cingulate cortex (ACC) and its subregions is altered in major depressive disorder (MDD) remains inconclusive. A systematic review and meta-analysis were therefore performed on the whole-brain resting-state functional connectivity (rsFC) studies using the ACC and its subregions as seed regions in MDD, in order to draw more reliable conclusions. Forty-four ACC-based rsFC studies were included, comprising 25 subgenual ACC-based studies, 11 pregenual ACC-based studies, and 17 dorsal ACC-based studies. Specific alterations of rsFC were identified for each ACC subregion in patients with MDD, with altered rsFC of subgenual ACC in emotion-related brain regions, of pregenual ACC in sensorimotor-related regions, and of dorsal ACC in cognition-related regions. Furthermore, meta-regression analysis revealed a significant negative correlation between the pgACC-caudate hypoconnectivity and percentage of female patients in the study cohort. This meta-analysis provides robust evidence of altered intrinsic functional connectivity of the ACC subregions in MDD, which may hold relevance to understanding the origin of, and treating, the emotional, sensorimotor and cognitive dysfunctions that are often observed in these patients.

Evaluation of Key Molecular Markers in Adult Diffuse Gliomas Based on a Novel Combination of Diffusion and Perfusion MRI and MR Spectroscopy.

BACKGROUND: Preoperative identification of isocitrate dehydrogenase (IDH) mutation and 1p/19q codeletion status could help clinicians select the optimal therapy in patients with diffuse glioma. Although, the value of multimodal intersection was underutilized. PURPOSE: To evaluate the value of quantitative MRI biomarkers for the identification of IDH mutation and 1p/19q codeletion in adult patients with diffuse glioma. STUDY TYPE: Retrospective. POPULATION: Two hundred sixteen adult diffuse gliomas with known genetic test results, divided into training (N = 130), test (N = 43), and validation (N = 43) groups. SEQUENCE/FIELD STRENGTH: Diffusion/perfusion-weighted-imaging sequences and multivoxel MR spectroscopy (MRS), all 3.0 T using three different scanners. ASSESSMENT: The apparent diffusion coefficient (ADC) and cerebral blood volume (CBV) of the core tumor were calculated to identify IDH-mutant and 1p/19q-codeleted statuses and to determine cut-off values. ADC models were built based on the 30th percentile and lower, CBV models were built based on the 75th centile and higher (both in five centile steps). The optimal tumor region was defined and the metabolite concentrations of MRS voxels that overlapped with the ADC/CBV optimal region were calculated and added to the best-performing diagnostic models. STATISTICAL TESTS: DeLong's test, diagnostic test, and decision curve analysis were performed. A P value <0.05 was considered to be statistically significant. RESULTS: Almost all ADC models achieved good performance in identifying IDH mutation status, among which ADC_15th was the most valuable parameter (threshold = 1.186; Youden index = 0.734; AUC_train = 0.896). The differential power of CBV histogram metrics for predicting 1p/19q codeletion outperformed ADC histogram metrics, and the CBV_80th-related model performed best (threshold = 1.435; Youden index = 0.458; AUC_train = 0.724). The AUCs of ADC_15th and CBV_80th models in the validation set were 0.857 and 0.733. These models tended to improve after incorporation of N-acetylaspartate/total_creatine and glutamate-plus-glutamine/total_creatine, respectively. DATA CONCLUSION: The intersection of ADC-, CBV-based histogram and MRS provide a reliable paradigm for identifying the key molecular markers in adult diffuse gliomas. EVIDENCE LEVEL: 3 TECHNICAL EFFICACY: Stage 3.

Cortical myelin and thickness mapping provide insights into whole-brain tumor burden in diffuse midline glioma.

Systemic infiltration is a hallmark of diffuse midline glioma pathogenesis, which can trigger distant disturbances in cortical structure. However, the existence and effects of these changes have been underexamined. This study aimed to investigate whole-brain cortical myelin and thickness alternations induced by diffuse midline glioma. High-resolution T1- and T2-weighted images were acquired from 90 patients with diffuse midline glioma with H3 K27-altered and 64 patients with wild-type and 86 healthy controls. Cortical thickness and myelin content was calculated using Human Connectome Project pipeline. Significant differences in cortical thickness and myelin content were detected among groups. Short-term survival prediction model was constructed using automated machine learning. Compared with healthy controls, diffuse midline glioma with H3 K27-altered patients showed significantly reduced cortical myelin in bilateral precentral gyrus, postcentral gyrus, insular, parahippocampal gyrus, fusiform gyrus, and cingulate gyrus, whereas diffuse midline glioma with H3 K27 wild-type patients exhibited well-preserved myelin content. Furtherly, when comparing diffuse midline glioma with H3 K27-altered and diffuse midline glioma with H3 K27 wild-type, the decreased cortical thickness in parietal and occipital regions along with demyelination in medial orbitofrontal cortex was observed in diffuse midline glioma with H3 K27-altered. Notably, a combination of cortical features and tumor radiomics allowed short-term survival prediction with accuracy 0.80 and AUC 0.84. These findings may aid clinicians in tailoring therapeutic approaches based on cortical characteristics, potentially enhancing the efficacy of current and future treatment modalities.

Metabolic heterogeneity in different subtypes of malformations of cortical development causing epilepsy: a proton magnetic resonance spectroscopy study.

Background: The most common subtypes of malformations of cortical development (MCDs) are gray matter heterotopia (GMH), focal cortical dysplasia (FCD), and polymicrogyria (PMG). This study aimed to characterize the possible neurometabolic abnormalities and heterogeneity in different MCDs subtypes using proton magnetic resonance spectroscopy (1H-MRS). Methods: In this prospective cross-sectional study, we recruited 29 patients with MCDs and epilepsy, including ten with GMH, ten with FCD, and nine with PMG, as well as 25 age- and sex-matched healthy controls (HC) from the Epilepsy Center of West China Hospital of Sichuan University between August 2018 and November 2021. Inclusion criteria for the patients were based upon typical magnetic resonance imaging (MRI) findings of MCDs and full clinical assessment for epilepsy. Single-voxel point-resolved spectroscopy was used to acquire data from both the lesion and the normal-appearing contralateral side (NACS) in patients and from the frontal lobe in HC. Metabolite measures, including N-acetyl aspartate (NAA), myoinositol (Ins), choline (Cho), creatine (Cr), and glutamate + glutamine (Glx) concentrations, were quantitatively estimated with linear combination model (LCModel) software and corrected for the partial volume effect of cerebrospinal fluid (CSF). Results: The NAA concentration was lower and the Ins concentration was higher in the MCDs lesions than in the NACS and in HC (P=0.002-0.007), and the Cho and Cr concentrations were higher in MCDs lesions than in HC (P=0.001-0.016). Moreover, the Cho concentration was higher in NACS than in HC (P=0.015). In the GMH lesions, the only metabolic alteration was an NAA reduction (GMH_lesion vs. HC: P=0.001). In the FCD lesions, there were more metabolite abnormalities than in the other two subtypes, particularly a lower NAA and a higher Ins than in HC and NACS (P=0.012-0.042). In the PMG lesions, Cr (lesion vs. HC or NACS: P=0.017-0.021) and Glx (lesion vs. NACS: P=0.043) were increased, while NAA was normal. Correlation analysis revealed that the Cr concentration in MCDs lesions was positively correlated with seizure frequency (r=0.411; P=0.027). Conclusions: Based upon 1H-MRS, our study demonstrated that different MCDs subtypes exhibited variable metabolic features, which may be associated with distinct functional and cytoarchitectural properties.

Deep learning for malignancy risk estimation of incidental sub-centimeter pulmonary nodules on CT images.

OBJECTIVES: To establish deep learning models for malignancy risk estimation of sub-centimeter pulmonary nodules incidentally detected by chest CT and managed in clinical settings. MATERIALS AND METHODS: Four deep learning models were trained using CT images of sub-centimeter pulmonary nodules from West China Hospital, internally tested, and externally validated on three cohorts. The four models respectively learned 3D deep features from the baseline whole lung region, baseline image patch where the nodule located, baseline nodule box, and baseline plus follow-up nodule boxes. All regions of interest were automatically segmented except that the nodule boxes were additionally manually checked. The performance of models was compared with each other and that of three respiratory clinicians. RESULTS: There were 1822 nodules (981 malignant) in the training set, 806 (416 malignant) in the testing set, and 357 (253 malignant) totally in the external sets. The area under the curve (AUC) in the testing set was 0.754, 0.855, 0.928, and 0.942, respectively, for models derived from baseline whole lung, image patch, nodule box, and the baseline plus follow-up nodule boxes. When baseline models externally validated (follow-up images not available), the nodule-box model outperformed the other two with AUC being 0.808, 0.848, and 0.939 respectively in the three external datasets. The resident, junior, and senior clinicians achieved an accuracy of 67.0%, 82.5%, and 90.0%, respectively, in the testing set. The follow-up model performed comparably to the senior clinician. CONCLUSION: The deep learning algorithms solely mining nodule information can efficiently predict malignancy of incidental sub-centimeter pulmonary nodules. CLINICAL RELEVANCE STATEMENT: The established models may be valuable for supporting clinicians in routine clinical practice, potentially reducing the number of unnecessary examinations and also delays in diagnosis. KEY POINTS: • According to different regions of interest, four deep learning models were developed and compared to evaluate the malignancy of sub-centimeter pulmonary nodules by CT images. • The models derived from baseline nodule box or baseline plus follow-up nodule boxes demonstrated sufficient diagnostic accuracy (86.4% and 90.4% in the testing set), outperforming the respiratory resident (67.0%) and junior clinician (82.5%). • The proposed deep learning methods may aid clinicians in optimizing follow-up recommendations for sub-centimeter pulmonary nodules and may lead to fewer unnecessary diagnostic interventions.

A Lightweight Deep Learning Framework for Automatic MRI Data Sorting and Artifacts Detection.

The purpose of this study is to develop a lightweight and easily deployable deep learning system for fully automated content-based brain MRI sorting and artifacts detection. 22092 MRI volumes from 4076 patients between 2017 and 2021 were involved in this retrospective study. The dataset mainly contains 4 common contrast (T1-weighted (T1w), contrast-enhanced T1-weighted (T1c), T2-weighted (T2w), fluid-attenuated inversion recovery (FLAIR)) in three perspectives (axial, coronal, and sagittal), and magnetic resonance angiography (MRA), as well as three typical artifacts (motion, aliasing, and metal artifacts). In the proposed architecture, a pre-trained EfficientNetB0 with the fully connected layers removed was used as the feature extractor and a multilayer perceptron (MLP) module with four hidden layers was used as the classifier. Precision, recall, F1_Score, accuracy, the number of trainable parameters, and float-point of operations (FLOPs) were calculated to evaluate the performance of the proposed model. The proposed model was also compared with four other existing CNN-based models in terms of classification performance and model size. The overall precision, recall, F1_Score, and accuracy of the proposed model were 0.983, 0.926, 0.950, and 0.991, respectively. The performance of the proposed model was outperformed the other four CNN-based models. The number of trainable parameters and FLOPs were the smallest among the investigated models. Our proposed model can accurately sort head MRI scans and identify artifacts with minimum computational resources and can be used as a tool to support big medical imaging data research and facilitate large-scale database management.

Artificial Intelligence Applications in Glioma With 1p/19q Co-Deletion: A Systematic Review.

As an important genomic marker for oligodendrogliomas, early determination of 1p/19q co-deletion status is critical for guiding therapy and predicting prognosis in patients with glioma. The purpose of this study is to systematically review the literature concerning the magnetic resonance imaging (MRI) with artificial intelligence (AI) methods for predicting 1p/19q co-deletion status in glioma. PubMed, Scopus, Embase, and IEEE Xplore were searched in accordance with the Preferred Reporting Items for systematic reviews and meta-analyses guidelines. Methodological quality of studies was assessed according to the Quality Assessment of Diagnostic Accuracy Studies-2. Finally, 28 studies were included in the quantitative analysis. Diagnostic test accuracy reached an area under the ROC curve of 0.71-0.98 were reported in 24 studies. The remaining four studies with no available AUC provided an accuracy of 0.75-0. 89. The included studies varied widely in terms of imaging sequences, input features, and modeling methods. The current review highlighted that integrating MRI with AI technology is a potential tool for determination 1p/19q status pre-operatively and noninvasively, which can possibly help clinical decision-making. However, the reliability and feasibility of this approach still need to be further validated and improved in a real clinical setting. EVIDENCE LEVEL: 2. TECHNICAL EFFICACY: 2.

Effects of randomness in the development of machine learning models in neuroimaging studies of schizophrenia.

Numerous studies have used machine learning with neuroimaging data for identifying individuals with a schizophrenia diagnosis. However, inconsistent results have limited the ability of the psychiatric community to objectively judge and accept the value of this approach. One factor that has contributed to the inconsistency, but has long been ignored, is randomness in the practice of machine learning. This is manifest when executing the same machine learning pipeline multiple times on the same dataset but getting different results. In the current study, a dataset of anatomical MRI scans from 158 patients with first-episode medication-naïve schizophrenia and 166 matched controls was used to investigate the effect of randomness on classifier performance estimates under different algorithm complexity and data splitting ratios. The maximum discriminatory accuracy that could be reached was 62.6 % ± 4.7 % (43.5 %-79.3 %) obtained when using extra-trees classifiers without feature normalization. Regions contributing to discrimination were located at bilateral temporal lobes and right frontal lobe. The results show that randomness has a significant impact on the precision of model performance estimates, especially when the size of test set is small. Current neuroimaging feature engineering combined with machine learning still falls short of being able to make diagnoses in the clinical context, but has value in revealing patterns of regional brain alteration associated with the illness. The current results indicate that effects of randomness on model performance should be reported and considered in interpreting model utility and it is necessary to evaluate models on large test sets to obtain valid estimates of model performance.

Unsupervised anomaly detection in brain MRI: Learning abstract distribution from massive healthy brains.

PURPOSE: To develop a general unsupervised anomaly detection method based only on MR images of normal brains to automatically detect various brain abnormalities. MATERIALS AND METHODS: In this study, a novel method based on three-dimensional deep autoencoder network is proposed to automatically detect and segment various brain abnormalities without being trained on any abnormal samples. A total of 578 normal T2w MR volumes without obvious abnormalities were used for model training and validation. The proposed 3D autoencoder was evaluated on two different datasets (BraTs dataset and in-house dataset) containing T2w volumes from patients with glioblastoma, multiple sclerosis and cerebral infarction. Lesions detection and segmentation performance were reported as AUC, precision-recall curve, sensitivity, and Dice score. RESULTS: In anomaly detection, AUCs for three typical lesions were as follows: glioblastoma, 0.844; multiple sclerosis, 0.858; cerebral infarction, 0.807. In anomaly segmentation, the mean Dice for glioblastomas was 0.462. The proposed network also has the ability to generate an anomaly heatmap for visualization purpose. CONCLUSION: Our proposed method was able to automatically detect various brain anomalies such as glioblastoma, multiple sclerosis, and cerebral infarction. This work suggests that unsupervised anomaly detection is a powerful approach to detect arbitrary brain abnormalities without labeled samples. It has the potential to support diagnostic workflow in radiology as an automated tool for computer-aided image analysis.

An MRI Study Combining Virtual Brain Grafting and Surface-Based Morphometry Analysis to Investigate Contralateral Alterations in Cortical Morphology in Patients With Diffuse Low-Grade Glioma.

BACKGROUND: The human brain has ability to reorganize itself in response to glioma. However, the mechanism of cortical reorganization remains unclear. PURPOSE: To investigate alterations in cortical thickness and local gyration index (LGI) in patients with unilateral frontal lobe diffuse low-grade glioma (DLGG). STUDY TYPE: Retrospective. SUBJECTS: Ninety-nine patients with histopathologically proven DLGG invading the left frontal lobe (LF; N = 56) or the right frontal lobe (RF; N = 43), and healthy controls (HC; N = 53). FIELD STRENGTH/SEQUENCE: 3.0 T, 3D T1-weighted images and gadolinium enhanced T1-weighted images using magnetization-prepared rapid gradient echo sequence, T2-weighted images, and fluid-attenuated inversion recovery using turbo spin echo sequence. ASSESSMENT: In patients with DLGG, virtual brain grafting combined with Freesurfer was utilized to enable automated cortical thickness and LGI calculation. In HC, standard FreeSurfer pipeline was applied to calculate these measures. Radiomic features were extracted from glioma using Pyradiomic software. STATISTICAL TESTS: General linear model and Pearson's correlation analysis. A P value <0.05 was considered statistically significant. RESULTS: For LF patients, there was significantly increased cortical thickness in the rostral middle frontal gyrus, significantly reduced cortical thickness in the precentral gyrus and hypogyrification in the lingual and medial orbitofrontal (MOF) gyrus in contralateral hemisphere. For RF patients, there was significantly increased cortical thickness in the middle temporal, lateral occipital extending to isthmus cingulate gyrus, significantly reduced cortical thickness in the precentral gyrus and hypogyrification in the lingual gyrus in the contralateral hemisphere. A negative association between four textural features of DLGG and LGI in the right MOF gyrus of LF group was found (r = -0.609, -0.442, -0.545, and -0.417, respectively). DATA CONCLUSION: Cortical thickness compensation was shown in contralateral homotopic location and some distant contralateral regions. Additionally, there was decreased cortical thickness in the contralateral precentral gyrus and hypogyrification in contralateral lingual gyrus. EVIDENCE LEVEL: 4 TECHNICAL EFFICACY: Stage 2.

Morphological fingerprinting: Identifying patients with first-episode schizophrenia using auto-encoded morphological patterns.

Although a large number of case-control statistical and machine learning studies have been conducted to investigate structural brain changes in schizophrenia, how best to measure and characterize structural abnormalities for use in classification algorithms remains an open question. In the current study, a convolutional 3D autoencoder specifically designed for discretized volumes was constructed and trained with segmented brains from 477 healthy individuals. A cohort containing 158 first-episode schizophrenia patients and 166 matched controls was fed into the trained autoencoder to generate auto-encoded morphological patterns. A classifier discriminating schizophrenia patients from healthy controls was built using 80% of the samples in this cohort by automated machine learning and validated on the remaining 20% of the samples, and this classifier was further validated on another independent cohort containing 77 first-episode schizophrenia patients and 58 matched controls acquired at a different resolution. This specially designed autoencoder allowed a satisfactory recovery of the input. With the same feature dimension, the classifier trained with autoencoded features outperformed the classifier trained with conventional morphological features by about 10% points, achieving 73.44% accuracy and 0.8 AUC on the internal validation set and 71.85% accuracy and 0.77 AUC on the external validation set. The use of features automatically learned from the segmented brain can better identify schizophrenia patients from healthy controls, but there is still a need for further improvements to establish a clinical diagnostic marker. However, with a limited sample size, the method proposed in the current study shed insight into the application of deep learning in psychiatric disorders.

Abnormal effective connectivity in visual cortices underlies stereopsis defects in amblyopia.

The neural basis underlying stereopsis defects in patients with amblyopia remains unclear, which hinders the development of clinical therapy. This study aimed to investigate visual network abnormalities in patients with amblyopia and their associations with stereopsis function. Spectral dynamic causal modeling methods were employed for resting-state functional magnetic resonance imaging data to investigate the effective connectivity (EC) among 14 predefined regions of interest in the dorsal and ventral visual pathways. We adopted two independent datasets, including a cross-sectional and a longitudinal dataset. In the cross-sectional dataset, we compared group differences in EC between 31 patients with amblyopia (mean age: 26.39 years old) and 31 healthy controls (mean age: 25.71 years old) and investigated the association between EC and stereoacuity. In addition, we explored EC changes after perceptual learning in a novel longitudinal dataset including 9 patients with amblyopia (mean age: 15.78 years old). We found consistent evidence from the two datasets indicating that the aberrant EC from V2v to LO2 is crucial for the stereoscopic deficits in the patients with amblyopia: it was weaker in the patients than in the controls, showed a positive linear relationship with the stereoscopic function, and increased after perceptual learning in the patients. In addition, higher-level dorsal (V3d, V3A, and V3B) and ventral areas (LO1 and LO2) were important nodes in the network of abnormal ECs associated with stereoscopic deficits in the patients with amblyopia. Our research provides insights into the neural mechanism underlying stereopsis deficits in patients with amblyopia and provides candidate targets for focused stimulus interventions to enhance the efficacy of clinical treatment for the improvement of stereopsis deficiency.

Psychoradiological abnormalities in treatment-naive noncomorbid patients with posttraumatic stress disorder.

BACKGROUND: Neuroimaging studies in posttraumatic stress disorder (PTSD) have identified various alterations in white matter (WM) microstructural organization. However, it remains unclear whether these are localized to specific regions of fiber tracts, and what diagnostic value they might have. This study set out to explore the spatial profile of WM abnormalities along defined fiber tracts in PTSD. METHODS: Diffusion tensor images were obtained from 77 treatment-naive noncomorbid patients with PTSD and 76 demographically matched trauma-exposed non-PTSD (TENP) controls. Using automated fiber quantification, tract profiles of fractional anisotropy, axial diffusivity, mean diffusivity, and radial diffusivity were calculated to evaluate WM microstructural organization. Results were analyzed by pointwise comparisons, by correlation with symptom severity, and for diagnosis-by-sex interactions. Support vector machine analyses assessed the ability of tract profiles to discriminate PTSD from TENP. RESULTS: Compared to TENP, PTSD showed lower fractional anisotropy accompanied by higher radial diffusivity and mean diffusivity in the left uncinate fasciculus, and lower fractional anisotropy accompanied by higher radial diffusivity in the right anterior thalamic radiation. Tract profile alterations were correlated with symptom severity, suggesting a pathophysiological relevance. There were no significant differences in diagnosis-by-sex interaction. Tract profiles allowed individual classification of PTSD versus TENP with significant accuracy, of potential diagnostic utility. CONCLUSIONS: These findings add to the knowledge of the neuropathological basis of PTSD. WM alterations based on a tract-profile quantification approach are a potential biomarker for PTSD.

Multimodal MR imaging signatures to identify brain diffuse midline gliomas with H3 K27M mutation.

BACKGROUND: Conventional MR imaging has limited value in identifying H3 K27M mutations. We aimed to investigate the capacity of quantitative MR imaging variables in identifying the H3 K27M mutation status of diffuse midline glioma. MATERIALS AND METHODS: Twenty-three patients with H3 K27M mutation and thirty-two wild-type patients were recruited in this retrospective study, all of whom underwent multimodal MR imaging. Clinical data and quantitative MR imaging variables were explored by subgroup analysis stratified by age (juveniles and adults). Then, a logistic model for all patients was constructed to identify potential variables for predicting K27M mutation status. Besides, a retrospective validation set including 13 patients was recruited. The C-index and F1 score were used to evaluate the performance of the prediction model. RESULTS: It turned out that patients with H3 K27M mutation were younger in the adult subgroup. In the mutation group, some relative apparent diffusion coefficient (rADC) histogram parameters and myo-inositol/creatine plus phosphocreatine (Ins/tCr) ratio were lower than in the wild-type group of both juveniles and adults (p < 0.05). After nested cross-validation and LASSO algorithm, the age, Ins/tCr, and rADC_15th were selected as potential predictors for H3 K27M mutation in the model. The nomogram model showed good diagnostic power with a validated C-index of 0.884. In addition, the area under the curve (AUC) was 0.898 (0.976 in validation set) and the F1 score was 0.732. CONCLUSIONS: In conclusion, age, rADC_15th, and Ins/tCr values were helpful in identifying H3 K27M mutations in midline gliomas.

Linked brain connectivity patterns with psychopathological and cognitive phenotypes in drug-naïve first-episode schizophrenia.

Background: Schizophrenia is considered to be a disorder of dysconnectivity characterized by abnormal functional integration between distinct brain regions. Different brain connection abnormalities were found to be correlated with various clinical manifestations, but whether a common deficit in functional connectivity (FC) in relation to both clinical symptoms and cognitive impairments could present in first-episode patients who have never received any medication remains elusive. Objective: To find a core deficit in the brain connectome that is related to both psychopathological and cognitive manifestations. Methods: A total of 75 patients with first-episode schizophrenia and 51 healthy control participants underwent scanning of the brain and clinical ratings of behaviors. A principal component analysis was performed on the clinical ratings of symptom and cognition. Partial correlation analyses were conducted between the main psychopathological components and resting-state FC that were found abnormal in schizophrenia patients. Results: Using the principal component analysis, the first principal component (PC1) explained 37% of the total variance of seven clinical features. The ratings of GAF and BACS contributed negatively to PC1, while those of PANSS, HAMD, and HAMA contributed positively. The FCs positively correlated with PC1 mainly included connections related to the insula, precuneus gyrus, and some frontal brain regions. FCs negatively correlated with PC1 mainly included connections between the left middle cingulate cortex and superior and middle occipital regions. Conclusion: In conclusion, we found a linked pattern of FC associated with both psychopathological and cognitive manifestations in drug-naïve first-episode schizophrenia characterized as the dysconnection related to the frontal and visual cortex, which may represent a core deficit of brain FC in patients with schizophrenia.

Developing of risk models for small solid and subsolid pulmonary nodules based on clinical and quantitative radiomics features.

BACKGROUND: Accurate evaluation of pulmonary nodule malignancy is important for lung cancer management. This current study aimed to develop risk models for small solid and subsolid pulmonary nodules based on clinical and quantitative radiomics features. METHODS: This study enrolled 5-20 mm pulmonary nodules detected on thoracic high-resolution computed tomography (HRCT), which were all confirmed pathologically. There were 548 solid nodules (242 malignant vs. 306 benign) and 623 subsolid nodules (SSNs 519 malignant vs. 104 benign). Relevant clinical characteristics were recorded. The CT image prior to the initial treatment was chosen for manual segmentation of the targeted nodule using the ITK-SNAP software. Subsequently, the marked image was processed to quantitatively extract 1218 radiomics features using PyRadiomics. We performed five-fold cross-validation to select potential predictors from clinical and radiomics features using the LASSO method and to evaluate the performance of the established models. In total, four types of models were tried: random forest, XGBOOST, SVM, and logistic models. The established models were compared with the Mayo model. RESULTS: Lung cancer risk models were developed among four nodule groups: all nodules (410 benign vs. 761 malignant; 1:1.86), nodules ≤10 mm (185 benign vs. 224 malignant; 1:1.21), solid nodules (306 benign vs. 242 malignant; 1.26:1), and SSNs (104 benign vs. 104 malignant; 1:1 matched). Significant clinical and radiomics predictors were selected for each group. The accuracy, area under the ROC curve, sensitivity, and specificity of the best model on validation dataset was 0.86, 0.91, 0.93, 0.73 for all nodules (XGBOOST), 0.82, 0.90, 0.86, 0.76 for nodules ≤10 mm (XGBOOST), 0.80, 0.89, 0.78, 0.82 for solid nodules (XGBOOST) and 0.70, 0.73, 0.73, 0.67 for SSNs (Random Forest). Except for the SSN models, the established clinical-radiomics models were superior to the Mayo model. CONCLUSIONS: Predictive models based on both clinical and radiomics features can be used to assess the malignancy of small solid and subsolid pulmonary nodules, even for nodules that are 10 mm or smaller.

An optimized two-stage cascaded deep neural network for adrenal segmentation on CT images.

Segmentation of adrenal glands from CT images is a crucial step in the AI-assisted diagnosis of adrenal gland-related disease. However, highly intrasubject variability in shape and adhesive boundaries with surrounding tissues make accurate segmentation of the adrenal gland a challenging task. In the current study, we proposed a novel two-stage deep neural network for adrenal gland segmentation in an end-to-end fashion. In the first stage, a localization network that aims to determine the candidate volume of the target organ was used in the preprocessing step to reduce class imbalance and computational burden. Then, in the second stage, a Small-organNet model trained with a novel boundary attention focal loss was designed to refine the boundary of the organ within the screened volume. The experimental results show that our proposed cascaded framework outperforms the state-of-the-art deep learning method in segmenting the adrenal gland with respect to accuracy; it requires fewer trainable parameters and imposes a smaller demand on computational resources.