BACKGROUND:Growing evidence points an association between internet gaming disorder (IGD) and suicide risk. However, the interplay with adverse childhood experiences (ACEs) and comorbid negative emotional states remains unclear. This knowledge gap hinders effective risk identification and targeted intervention for young people, underscoring the urgent need for comprehensive research in this area. METHODS:We conducted a cross-sectional, multicenter study involving 13,509 participants aged 12-25 years from six provinces in China. Participants completed the Internet Gaming Disorder Scale-Short Form (IGDS9-SF), Adverse Childhood Experiences-International Questionnaire (ACE-IQ), Depression Anxiety Stress Scales-21 (DASS-21). Suicidal thoughts and suicidal behavior items were analyzed with Chi-square tests, ANOVA, logistic regression models and mediation analysis. RESULTS:Among 2141 participants who met the criteria for IGD, 32.8% reported experiencing suicidal thoughts while 6.4% disclosed having suicidal behavior. High exposure to ACEs (≥5 events) was significantly more prevalent among those with suicidality (50.74% for thoughts; 45.23% for behaviors) compared to those without (16.36%). Key risk factors associated with increased suicide risk included male gender (OR = 1.456), depression (OR = 1.091), emotional abuse (OR = 1.727), exclusion/bullying (OR = 1.565), and parental separation or divorce (OR = 1.581). LIMITATIONS:The retrospective, cross-sectional design cannot draw causality, and biases in self-report measurements cannot be ignored. CONCLUSION:This large-scale study demonstrates that male gender, depression and cumulative ACEs are significantly associated with suicide risk in young individuals with IGD in China. These findings highlight the necessity of early identification of ACEs, trauma-informed support, and prompt management of depression to mitigate suicidality in this population.

Clinical management of breast cancer lung metastasis is challenging because of the complexity of dynamic lesion assessment. Traditional methods based on RECIST1.1 rely on size measurement, and existing studies require image registration and are limited to lesion-level assessment. In this study, we proposed a patient-level spatiotemporal assessment framework without registration to comprehensively analyze multiple lesions evolvement based on longitudinal CT images. Our method considers metastatic lesions that vary in size and often overlap with complex structures such as blood vessels and bones, and avoids potential registration errors. Our method outperforms state-of-the-art methods on both the Peking Union Medical College Hospital breast cancer lung metastasis dataset and the publicly available dataset. The model also showed excellent performance in a multicenter validation across four medical centers. We established a patient-level metastatic breast cancer assessment framework, providing a practical solution for longitudinal treatment monitoring.

ABSTRACT Advanced pancreatic ductal adenocarcinoma (PDAC) often progresses rapidly during chemotherapy despite initial assessments of stable disease or partial response by Response Evaluation Criteria in Solid Tumors (RECIST 1.1), underscoring the limitations of the current methods for predicting short‐term progressive disease (PD). To address this, the study developed a spatiotemporal deep learning framework that integrates convolutional and long short‐term memory (LSTM) neural networks to dynamically predict PD at the next follow‐up visit using serial computed tomography (CT) scans and baseline clinical variables. The model was trained on a retrospective cohort of 243 patients (415 predicted events, defined as temporal sequences for the next follow‐up PD prediction) and evaluated across internal, external, and prospective cohorts. The model achieved area under the curve (AUC) values of 0.77, 0.76, and 0.74, respectively. Performance remained robust across chemotherapy regimens (AG or Gemcitabine‐based, FOLFIRINOX, and SOXIRI; AUC 0.68–0.79), PD subtypes (target lesion growth vs. new metastases; AUC 0.72 vs. 0.77), and baseline disease stages (locally advanced vs. metastatic; AUC 0.85 vs. 0.71). This framework enables the noninvasive, real‐time prediction of imminent PD in advanced PDAC, facilitating timely treatment modification. Its validated generalizability and reliance on routine clinical data underscore its potential for seamless integration into chemotherapy management.

Adolescent borderline personality disorder (aBPD) is linked to severe psychological problems and social dysfunction in the affected adolescents. However, the neural bases of this disorder are still poorly understood. The objective of this review is to summarize the current understanding of brain structural and functional alterations identified by neuroimaging methods and to propose future research directions. This systematic review identified widespread alterations of three main brain networks in aBPD, i.e., the Central Executive Network, the Default Mode Network, and the Affective Network. We also provide insights for future research using novel analytical techniques and artificial intelligence, to facilitate the identification of potential neuroimaging biomarkers for aBPD. The potential for targeting the dysfunctional networks via neurostimulation method is also discussed from a theoretical and hypothesis-generating perspective.

BACKGROUND:Adolescents with bipolar disorder (BD) experience rapid mood fluctuations and cognitive disturbance. However, the underlying neural mechanism for sad-happy emotional switching remains poorly understood in adolescents with BD. METHODS:We used an emotional Go/No-Go functional magnetic resonance imaging (fMRI) task paradigm to assess brain activity during the transition between sad and happy emotional states in 43 adolescents with BD and 18 age- and gender-matched healthy control (HC) participants. Neuroimaging data were analyzed using a general linear model with age and gender as covariates, and results were thresholded at p < 0.001 (voxel-wise, uncorrected) with cluster-level False Discovery Rate (FDR) correction at p < 0.05. Behavioral performance data and clinical scale data were compared, and the statistical significance was set at p < 0.05 (FDR-corrected). RESULTS:Compared to the HC participants, adolescents with BD showed hyperactivation in the default mode network (right precuneus, left medial superior frontal gyrus), and hypoactivation in the extended salience network (left insula, right middle temporal gyrus) and in the cerebellum (p < 0.05). Notably, default mode network hyperactivation was associated with higher omission rates and longer reaction times when performing the Go/No-Go fMRI task paradigm (r = 0.319, p = 0.049; r = 0.320, p = 0.044), while decreased insular and temporal activity was linked to impaired inhibitory control (r = -0.315, p = 0.049; r = -0.389, p = 0.011), and reduced cerebellar activity was linked with poorer executive function, particularly under sad conditions (r = 0.308, p = 0.044). CONCLUSIONS:Adolescents with BD showed aberrant brain activity patterns for regulating emotions, shifting attention, and maintaining cognitive control during emotional switching. These network alterations may serve as potential neuroimaging biomarkers for adolescents with BD, assisting individualized treatment strategies targeting emotion-cognition dysfunctions.

BACKGROUND:Current neuroimaging research on paranoid traits remains limited. Existing studies have largely relied on small samples, focused on categorical diagnoses or transient paranoid states, and examined either structural or functional measures in isolation. Crucially, the joint contribution of brain structure and intrinsic functional activity to paranoid personality traits (PPT) in the general population and their relationship with other psychological traits, remains unknown OBJECTIVES: The present study aimed to identify network-level neural markers of paranoid personality traits in a large sample by integrating gray matter morphology and resting-state brain activity, to test the hypothesis that networks associated with social and affective dysfunctions, predict PTT METHODS: We applied an unsupervised multimodal data-fusion approach (parallel ICA) to gray matter concentration and fractional ALFF in 197 healthy individuals. Paranoid personality traits were assessed dimensionally. In complementary analyses, we examined whether multimodal component loadings captured trait-related variance beyond demographic covariates RESULTS: Analyses identified a resting-state component encompassing the precuneus and angular gyrus, partially overlapping with the default mode network, significantly associated with paranoid personality traits. This functional component was positively correlated with a gray matter component including orbitofrontal and insular regions indicating a linked structural-functional pattern. CONCLUSIONS:By jointly modeling gray matter and resting-state activity, this study provides the first multimodal evidence of network-level markers underlying paranoid personality traits in the general population.

Objective: Differentiating between brain metastasis (BM) and glioblastoma (GBM) preoperatively is challenging due to their similar imaging features on conventional brain MRI. This study aimed to enhance diagnostic accuracy through a machine learning model based on MRI radiomics data. Methods: This retrospective study included 235 patients with confirmed solitary BM and 273 patients with GBM. Patients were randomly assigned to the training (n = 356) or the validation (n =152) cohort. Conventional brain MRI sequences including T1-weighted imaging (T1WI), contrast-enhanced_T1WI, and T2-weighted imaging (T2WI) were acquired. Brain tumors were delineated on all three sequences and segmented. Features were selected from demographic, clinical, and radiomic data. An integrated ensemble machine learning model, i.e., the elastic regression-SVM-SVM model (ERSS) and a multivariable logistic regression (LR) model combining demographic, clinical, and radiomic data were built for predictive modeling. Model efficiency was evaluated using discrimination, calibration, and decision curve analyses. Additionally, external validation was performed using an independent cohort consisting of 47 patients with GBM and 43 patients with isolated BM to assess the ERSS model generalizability. Results: The ERSS model demonstrated more optimal classification performance (AUC: 0.9548, 95% CI: 0.9337-0.9734 in training cohort; AUC: 0.9716, 95% CI: 0.9485-0.9895 in validation cohort) as compared to the LR model according to the receiver operating characteristic (ROC) curve and decision curve for the internal cohort. The external validation cohort had less optimal but still robust performance (AUC: 0.7174, 95% CI: 0.6172-0.8024). The ERSS model with integration of multiple classifiers, including elastic net, random forest and support vector machine, produced robust predictive performance and outperformed the LR method. Conclusion: The results suggested that the integrated machine learning model, i.e., the ERSS model, had the potential for efficient and accurate preoperative differentiation of BM from GBM, which may improve clinical decision-making and outcomes of patients with brain tumors.

Amyotrophic lateral sclerosis is a fatal neurodegenerative disease involving progressive degeneration of upper and lower motor neurons. Beyond well-established grey and white matter pathology, alterations in cortical gyrification have recently been observed, yet their clinical relevance and molecular underpinnings remain to be understood. Here, we investigated this premise by examining its microstructural and transcriptional basis in 60 patients with amyotrophic lateral sclerosis (median age = 55, range = 25–72 years) and 60 matched controls (median age = 56, range = 27–72 years) using structural and diffusion MRI. Patients exhibited a significant reduction in local gyrification index within bilateral precentral and postcentral gyri, left middle frontal gyrus and left superior parietal lobule. This was accompanied by reduced fractional anisotropy in the white matter tracts, primarily involving the corticospinal tract and corpus callosum. Higher local gyrification index and fractional anisotropy values were associated with better motor function as measured by the Amyotrophic Lateral Sclerosis Functional Rating Scale-Revised, and local gyrification index also showed positive associations with global cognitive status. A mediation analysis indicated that fractional anisotropy partially accounted for the relationship between local gyrification index and functional disability, suggesting that disrupted white matter pathways contribute to the clinical impact of gyrification changes. To explore underlying mechanisms, we integrated neuroimaging findings with transcriptomic data from the Allen Human Brain Atlas. Regions of reduced local gyrification index showed spatial convergence with cortical expression of amyotrophic lateral sclerosis-related genes such as TARDBP and C9orf72, enriched for biological processes related to protein aggregation, axon guidance and synaptic signalling. Together, these findings suggest that cortical gyrification abnormalities in amyotrophic lateral sclerosis are closely linked to white matter degeneration, functional impairment and genetic vulnerability, thereby offering an integrative window into the multiscale pathology of amyotrophic lateral sclerosis.

Objectives: To develop and validate a prediction model based on brain MRI features to predict disease-free survival (DFS) and overall survival (OS) for patients with intracranial extraventricular ependymoma (IEE). Methods: The study included 114 patients with pathology-proven IEE, of whom 80 were randomly assigned to a training group and 34 to a validation group. Preoperative brain MRI images were assessed with the Visually AcceSAble Rembrandt Images (VASARI) feature set. Clinical variables were assessed including age, gender, KPS, pathological grade of the tumor and blood test data such as eosinophil, blood urea nitrogen and serum creatinine. Multivariate Cox proportional hazards regression analysis was performed to select the independent prognostic factors for DFS and OS. Three prediction models were built with clinical variables, MRI-VASARI features, and combined clinical and MRI-VASARI data, respectively. The predictive power of survival models was assessed using c-index and calibration curve. Results: Clinical variables such as eosinophil, blood urea nitrogen and serum creatinine, and MRI-VASARI feature for definition of the non-enhancing margin (F13) were significantly correlated with the prognosis of DFS. Blood urea nitrogen, D-dimer, tumor location (F1), eloquent brain (F3), and T1/FLAIR ratio (F10) were independent predictors of OS. Based on these factors, prediction models were constructed. The concordance indices of the three survival models for OS were 0.732, 0.729, and 0.768, respectively. For DFS, the concordance indices were respectively 0.694, 0.576, and 0.714. Conclusion: Predictive modelling combining both clinical and MRI-VASARI features is robust and may assist in the assessment of prognosis in patients with IEE.

Background Neoadjuvant immunotherapy significantly improves the pathological complete response (pCR) rate in colorectal cancer (CRC). However, the lack of reliable tools to accurately identify responders remains a key barrier to its widespread clinical adoption. This study aimed to develop an interpretable radiomics model guided by immunophenotypes to predict response to preoperative immunotherapy in CRC, with the goal of enabling more precise and personalized treatment strategies.Methods First, we retrospectively collected 108 patients with CRC from the center who underwent preoperative CT and RNA sequencing. Immunophenotypes were characterized through unsupervised clustering of tumor-infiltrating immune cells at RNA level, with subsequent validation by CD3 and CD8 spatial distributions at immunohistochemical (IHC) level. Furthermore, patients from center II (n=19) and center III (n=22) receiving neoadjuvant immunotherapy were assigned to training and validation cohorts. Through a two-stage selection process, immunophenotype-associated radiomic features were initially identified, followed by immunotherapy response-related radiomics features that were further identified. Eventually, an interpretable immunotherapy response prediction model was developed by integrating a decision tree algorithm with SHapley Additive exPlanations (SHAP) analysis.Results Two immunophenotypes (immune-hot and immune-cold) were identified, compared with the immune-cold, the former exhibited more abundant RNA-based immune cell infiltration and higher densities of CD3 and CD8 T cells in both the core tumor and invasive margin areas. A decision tree model integrating three key radiomic features achieved an area under the curve of 0.904 (95% CI: 0.679 to 1.000) in an independent validation cohort. SHAP analysis identified higher large dependence emphasis and lower variance as potential predictors of pCR, which quantified the homogeneity of the tumors.Conclusion Two immunophenotypes were constructed and identified at both bulk RNA level and IHC level. An interpretable and accurate radiomics model was constructed to guide personalized immunotherapy strategies in clinical practice.

Photothermal therapy is a safe and effective tumour treatment strategy due to its excellent spatiotemporal controllability. However, interferon gamma in the tumour microenvironment is upregulated after photothermal therapy, which enhances the expression of programmed cell death ligand 1 (PD-L1) in tumour cells. This further promotes immunosuppression and tumour metastasis, resulting in a poor prognosis in cancer therapy. Traditional nanodrugs often face challenges in penetrating the dense extracellular matrix of solid tumours, whereas certain probiotics possess the ability to specifically colonise the core regions of tumours. In this research, we used Escherichia coli Nissle 1917 (ECN) as a chassis cell and self-assembly polydopamine (PDA) on the ECN surface. The black PDA@ECN (notes as PE) actively colonises at the tumour site and produces a photothermal effect under 808 nm laser irradiation to kill tumour cells. To overcome the high expression of PD-L1 induced after photothermal therapy, metformin (MET) was also encapsulated in PE to form PDA@MET@ECN (notes as PME). In vivo experiments demonstrated that PME effectively inhibited the PD-L1 expression and growth of CT26 tumour cells. Overall, PME reverses the immunosuppressive tumour microenvironment and enhances the effect of photothermal/immune therapy in tumour treatment.

BACKGROUND:Major depressive disorder (MDD) is a severe mental illness, and the Hamilton Depression Rating Scale (HAMD) is commonly used to quantify its severity. Our aim is to develop a predictive model for MDD symptoms using machine learning techniques based on effective connectivity (EC) from resting-state functional magnetic resonance imaging (rs-fMRI). NEW METHOD:We obtained large-scale rs-fMRI data and HAMD scores from the multi-site REST-meta-MDD dataset. Average time series were extracted using different atlases. Brain EC features were computed using Granger causality analysis based on symbolic path coefficients, and a machine learning model based on EC was constructed to predict HAMD scores. Finally, the most predictive features were identified and visualized. RESULTS:Experimental results indicate that different brain atlases significantly impact predictive performance, with the Dosenbach atlas performing best. EC-based models outperformed functional connectivity, achieving the best predictive accuracy (r = 0.81, p < 0.001, Root Mean Squared Error=3.55). Among various machine learning methods, support vector regression demonstrated superior performance. COMPARISON WITH EXISTING METHODS:Current phenotype score prediction primarily relies on FC, which cannot indicate the direction of information flow within brain networks. Our method is based on EC, which contains more comprehensive brain network information and has been validated on large-scale multi-site data. CONCLUSIONS:Brain network connectivity features effectively predict HAMD scores in MDD patients. The identified EC feature network may serve as a biomarker for predicting symptom severity. Our work may provide clinically significant insights for the early diagnosis of MDD, thereby facilitating the development of personalized diagnostic tools and therapeutic interventions.

BACKGROUND:Major depressive disorder (MDD) is a severe and common mental illness. The first-episode drugs-naive MDD (FEDN-MDD) patients, who have not undergone medication intervention, contribute to understanding the biological basis of MDD. Multimodal Magnetic Resonance Imaging can provide a comprehensive understanding of brain functional and structural abnormalities in MDD. However, most MDD studies use single-modal, small-scale MRI data. And several multimodal studies of MDD are limited to simple linear combinations of functional and structural features. METHODS:We screened a large sample of FEDN-MDD patients and healthy controlsmultimodal MRI data. Extracting the fractional amplitude of low-frequency fluctuations (fALFF) feature from functional magnetic resonance imaging and the gray matter volume (GMV) feature from structural magnetic resonance imaging. The mCCA-jICA method was used to integrate these two modal features to investigate the functional-structural co-variation abnormalities in MDD. To validate the stability of the extracted functional-structural covariant abnormalities features, we apply them to identify FEDN-MDD patients. RESULTS:The results show that compared to healthy controls, FEDN-MDD patients exhibit joint group-discriminative independent component and modality-specific group-discriminative independent component, suggesting functional-structural covariant abnormalities in MDD patients. Using lightGBM classifier, we achieve a classification accuracy of 99.84 %. LIMITATION:We use GMV and fALFF for multimodal fusion shows promise, but requires further validation with other datasets and exploration of additional multimodal features. CONCLUSIONS:This may indicate that multimodal fusion features can effectively explore information between different modalities and can accurately identify FEDN-MDD patients, suggesting their potential as multimodal brain imaging biomarkers for MDD.

Supplementary Fig. 4 A. Progression-free survival by PD-L1 status. B. Overall survival by PD-L1 status. Dot marks indicate censored data.

Introduction:Borderline personality disorder (BPD) is one of the most frequently diagnosed disorders in psychiatric settings. Beyond the categorical diagnosis, borderline personality traits (BPT) are common in the general population and vary along a continuum from mild to severe. While prior research has reported functional connectivity alterations in the default mode network (DMN), the salience network (SN), and the central-executive network (CEN) in patients with BPD, the impairment of these networks in subclinical BPT remain underexplored. To fill this gap, this study aims to investigate dynamic functional connectivity alterations associated with BPT in a subclinical population. We expect to find abnormal connectivity inside the DMN, the SN and in regions ascribed to mentalization processes associated with BPT. We also expect these networks to be associated with psychological symptoms experienced by borderline patients such as impulsivity and anger issues, as well as lack of self-control and neuroticism among others. Method:An unsupervised machine learning method known as Group-ICA, was applied to resting state fMRI images of 200 individuals to predict BPT from the temporal variability of independent macro networks. Results:Results indicated abnormal dynamic functional connectivity inside the SN including areas implicated in emotional reactivity and sensitivity, and in a network that partially overlaps with the DMN, including regions involved in social cognition and mind reading. Specifically, the higher the BPT, the higher the temporal variability inside the SN, and the lower the temporal variability in a network that includes DMN and mentalization regions. Notably, the BOLD variability of the SN correlated with neuroticism, anger problems, lack of self- control, and distorted inner dialogue, all symptoms displayed by individuals with borderline personality. Discussion:These findings indicate that abnormalities in resting state networks are visible in subclinical populations with varying degrees of borderline traits, with impaired DMN and SN. These insights may pave the way for designing interventions to prevent the development of the full disorder.

Supplementary Fig. 3 A. Progression-free survival by CD8+ TIL density. B. Overall survival by CD8+ TIL density. Dot marks indicate censored data.

OBJECTIVE:Previous research suggests that adolescents with BPD (aBPD) exhibit distinct neuroanatomical alterations, although methodological limitations such as low sample size, and the reliance on univariate massive statistical analyses, prevent conclusions. Moreover, the possibility to associate these abnormalities with clinical features has been only partially explored. This study aims to investigate structural brain differences in the largest sample of adolescents with BPD to date, using a combination of unsupervised and supervised machine learning approaches: Source-Based Morphometry (SBM) and a deep neural network whose architecture was optimized through genetic algorithms. We hypothesize that adolescents with BPD will exhibit increased gray matter volume in the default mode network (DMN) and cerebellum, strictly related with emotion dysregulation and borderline symptoms, alongside reduced gray matter in frontal control networks. METHODS:High-resolution T1-weighted structural MRI of 129 adolescents with BPD (aged 12-17) and 107 age-, gender-, and education-matched healthy controls (HCs) were analyzed using SBM to identify networks of covarying gray matter concentration (GMC). Borderline symptomatology, difficulties in emotion regulation, anxiety-related problems, and global functioning were assessed to characterize the meaning of neural findings. RESULTS:Compared to HCs, adolescents with BPD exhibited significantly increased GMC in regions overlapping with the posterior hub of the DMN, and the cerebellum, and reduced GMC in frontal control regions. Importantly, the GMC alterations inside the cerebellum and the DMN positively correlated with the difficulties in emotion regulation such as emotional clarity and emotion regulation difficulties, self-harm injuries, anxiety and depressive symptoms, and negatively correlated with global assessment functioning. The deep learning model confirmed these findings and provided a good generalization performance. CONCLUSIONS:Our findings suggest that gray matter alterations in regions ascribed to the default mode network, cerebellum, and frontal control regions play a crucial role in emotional regulation deficits and self-injurious behaviors in adolescent BPD. This study provides new insights into the neurobiological mechanisms of BPD in youth and offering potential biomarker and targets for treatments.

Supplementary Fig. 4 A. Progression-free survival by PD-L1 status. B. Overall survival by PD-L1 status. Dot marks indicate censored data.

Supplementary Fig. 1 Forest plot of subgroup analysis for progression-free survival (A) and overall survival (B).