Dose-dependent early white matter alterations in patients with brain metastases after radiotherapy.

PURPOSE: Previous diffusion tensor imaging (DTI) studies have mainly focused on dose-dependent white matter (WM) alterations 1 month to 1 year after radiation therapy (RT) with a tract-average method. However, WM alterations immediately after RT are subtle, resulting in early WM alterations that cannot be detected by tract-average methods. Therefore, we performed a study with an along-tract method in patients with brain metastases to explore the early dose-response pattern of WM alterations after RT. METHODS: Sixteen patients with brain metastases underwent DTI before and 1-3 days after brain RT. DTI metrics, such as fractional anisotropy (FA), axial diffusivity (AD), radial diffusivity (RD) and mean diffusivity (MD), were calculated. Along-tract statistics were then used to resample WM fibre streamlines and generate a WM skeleton fibre tract. DTI metric alterations (post_RT-pre_RT DTI metrics) and the planned doses (max or mean doses) were mapped to 18 WM tracts. A linear fixed model was performed to analyse the main effect of dose on DTI metric alterations. RESULTS: AD alterations in the left hemispheric uncinated fasciculus (UNC_L) were associated with max doses, in which decreased AD alterations were associated with higher doses. CONCLUSION: Our findings may provide pathological insight into early dose-dependent WM alterations and may contribute to the development of max dose-constrained RT techniques to protect brain microstructure in the UNC_L.

The radiomic-clinical model using the SHAP method for assessing the treatment response of whole-brain radiotherapy: a multicentric study.

OBJECTIVE: To develop and validate a pretreatment magnetic resonance imaging (MRI)-based radiomic-clinical model to assess the treatment response of whole-brain radiotherapy (WBRT) by using SHapley Additive exPlanations (SHAP), which is derived from game theory, and can explain the output of different machine learning models. METHODS: We retrospectively enrolled 228 patients with brain metastases from two medical centers (184 in the training cohort and 44 in the validation cohort). Treatment responses of patients were categorized as a non-responding group vs. a responding group according to the Response Assessment in Neuro-Oncology Brain Metastases (RANO-BM) criteria. For each tumor, 960 features were extracted from the MRI sequence. The least absolute shrinkage and selection operator (LASSO) was used for feature selection. A support vector machine (SVM) model incorporating clinical factors and radiomic features wase used to construct the radiomic-clinical model. SHAP method explained the SVM model by prioritizing the importance of features, in terms of assessment contribution. RESULTS: Three radiomic features and three clinical factors were identified to build the model. Radiomic-clinical model yielded AUCs of 0.928 (95%CI 0.901-0.949) and 0.851 (95%CI 0.816-0.886) for assessing the treatment response in the training cohort and validation cohort, respectively. SHAP summary plot illustrated the feature's value affected the feature's impact attributed to model, and SHAP force plot showed the integration of features' impact attributed to individual response. CONCLUSION: The radiomic-clinical model with the SHAP method can be useful for assessing the treatment response of WBRT and may assist clinicians in directing personalized WBRT strategies in an understandable manner. KEY POINTS: • Radiomic-clinical model can be useful for assessing the treatment response of WBRT. • SHAP could explain and visualize radiomic-clinical machine learning model in a clinician-friendly way.

Modeling the association between functional connectivity and lateralization with the activity flow framework.

The human brain is localized and distributed. On the one hand, each cognitive function tends to involve one hemisphere more than the other, also known as the principle of lateralization. On the other hand, interactions among brain regions in the form of functional connectivity (FC) are indispensable for intact function. Recent years have seen growing interest in the association between lateralization and FC. However, FC metrics vary from spurious correlation to causal associations. If lateralization manifests local processing and causal network interactions, more causally valid FC metrics should predict lateralization index (LI) better than FC based on simple correlations. The present study directly investigates this hypothesis within the activity flow framework to compare the association between lateralization and four brain connectivity metrics: correlation-based FC, multiple-regression FC, partial-correlation FC, and combinedFC. We propose two modeling approaches: the one-step approach, which models the relationship between LI and FC directly, and the two-step approach, which predicts the brain activation and calculates the LI. Our results indicated that multiple-regression FC, partial-correlation FC, and combinedFC could significantly improve the model prediction compared to correlation-based FC, which was consistent in a spatial working memory task (typically right-lateralized) and a language task (typically left-lateralized). The one-step and two-step approach yielded similar conclusions. In addition, the finding was replicated in a clinical sample of schizophrenia (SZ), bipolar disorder (BP), and attention deficit hyperactivity disorder (ADHD). The present study suggests that the causal interactions among brain regions help shape the lateralization pattern.

TSP-GNN: a novel neuropsychiatric disorder classification framework based on task-specific prior knowledge and graph neural network.

Neuropsychiatric disorder (ND) is often accompanied by abnormal functional connectivity (FC) patterns in specific task contexts. The distinctive task-specific FC patterns can provide valuable features for ND classification models using deep learning. However, most previous studies rely solely on the whole-brain FC matrix without considering the prior knowledge of task-specific FC patterns. Insight by the decoding studies on brain-behavior relationship, we develop TSP-GNN, which extracts task-specific prior (TSP) connectome patterns and employs graph neural network (GNN) for disease classification. TSP-GNN was validated using publicly available datasets. Our results demonstrate that different ND types show distinct task-specific connectivity patterns. Compared with the whole-brain node characteristics, utilizing task-specific nodes enhances the accuracy of ND classification. TSP-GNN comprises the first attempt to incorporate prior task-specific connectome patterns and the power of deep learning. This study elucidates the association between brain dysfunction and specific cognitive processes, offering valuable insights into the cognitive mechanism of neuropsychiatric disease.