Background and Objective Electroencephalography (EEG) provides a promising non-invasive approach for Alzheimer's disease (AD) detection. However, cross-subject EEG-based diagnosis remains challenging due to substantial inter-subject variability and the complex pathological characteristics of AD, which involve both dynamic temporal alterations and abnormal brain functional organization. Existing methods often focus on either temporal patterns or spatial connectivity, while the effective interaction between these complementary representations remains insufficiently explored. To address these limitations, we propose STFormer, a Transformer-based spatio-temporal fusion network for subject-independent AD-related EEG classification. Methods STFormer is designed to jointly model temporal dynamics, brain-region topology, and their interactions through these complementary components. The Multiscale Masked Prediction Transformer (MMPT) captures robust multi-scale temporal representations by incorporating auxiliary masked reconstruction. The Matrix Fusion Graph Convolution (MFGC) learns disease-related brain-network topology by integrating brain-region-aware feature aggregation and complementary connectivity matrices. Furthermore, the Gated Cross-Attention Fusion (GCAF) module performs adaptive bidirectional interaction between temporal and structural representations, enabling more effective spatio-temporal feature integration. Results Conducted on three public EEG datasets under subject-independent five-fold cross-validation and compared with eight representative EEG classification methods, STFormer achieves the best overall performance in both binary and three-class classification tasks. Specifically, STFormer improved accuracy and F1-score by 5.67 and 8.39 percentage points in binary classification, and by 12.93 and 15.66 percentage points in three-class classification, respectively. Conclusions The results demonstrate that jointly modeling multi-scale temporal dynamics, brain-region topology, and bidirectional spatio-temporal interactions can effectively improve the cross-subject generalization of EEG-based AD detection. STFormer provides a robust and computationally efficient framework for subject-independent EEG analysis and offers potential value for intelligent neurodegenerative disease assessment.
Objective Surgical treatment of benign bone tumors in the pelvic region is technically challenging. Traditional cortical fenestration methods rely heavily on the surgeon's experience, making it difficult to consistently control the position and extent of the bone window. This study reports the design of a 3D-printed fenestrated guide and evaluates its ability to control fenestration dimensions and its preliminary clinical outcomes in intralesional curettage of benign pelvic bone tumors. Methods This retrospective study included six patients with benign or locally aggressive pelvic bone tumors who underwent intralesional curettage assisted by an individualized 3D-printed fenestrated guide between January 2022 and August 2024. The guide was designed from preoperative CT data to conform to the local bony contours and assist intraoperative localization and cortical fenestration of the planned bone window. The primary outcome was the dimensional consistency between the intraoperative bone flap and the planned guide fenestration (bone-contacting side, projected onto the cortical bone surface). Secondary outcomes included perioperative parameters, MSTS-93 functional scores, complications, and radiographic follow-up. Results All surgeries were completed successfully with guide assistance. The achieved bone flap dimensions closely matched those of the planned guide fenestration: median long-axis deviation −1.65 mm (range, −2.8 to +2.4) and median short-axis deviation −0.90 mm (range, −1.8 to +1.5). All deviations fell within ±3 mm. The median operative time was 150 min (range, 130–170) and the median intraoperative blood loss was 200 mL (range, 100–300). At a median follow-up of 25 months (range, 20–34), all patients showed satisfactory bone window healing with no local recurrence. The median MSTS-93 score increased from 26 (87%) at 3 months to 28.5 (95%) at final follow-up (p=0.031). One patient developed a Clavien–Dindo grade IIIb infection that resolved after debridement and antibiotic therapy. Conclusion The 3D-printed fenestrated guide may assist intraoperative localization and controlled cortical fenestration in selected benign pelvic bone tumors.
Objectives Medical image synthesis using Generative Adversarial Networks (GANs) has gained increasing attention to address data scarcity, missing modalities, and image quality enhancement in clinical workflows. Despite encouraging technical performance, concerns persist regarding the clinical reliability, generalizability, and safety of GAN-based approaches. This systematic review critically examines experimental and observational studies applying GANs to medical image synthesis, with emphasis on methodological design, evaluation practices, clinical validation, and sources of bias. Material and Methods A PRISMA-guided search of PubMed, IEEE Xplore, Scopus, and Web of Science was conducted to identify studies published between January 2018 and December 2025. Studies were included if they focused on GAN-based medical image synthesis using human subject data, provided quantitative evaluation measures, and reported sufficient methodological detail. Data extraction covered study characteristics, GAN architecture, evaluation metrics, validation strategies, and clinical relevance. Methodological quality and risk of bias were assessed using adapted frameworks. Results Twenty studies met inclusion criteria, spanning multiple imaging modalities (MRI: 40%, CT: 35%, PET: 20%) and anatomical regions. Modality translation (40%) and super-resolution (20%) were the predominant applications, with conditional GAN architectures dominating (75% paired training). While improvements in image similarity metrics such as PSNR and SSIM were frequently reported, external validation (0–5%), multi-center evaluation (0–5%), and clinical reader studies (0–5%) were rare. Common methodological limitations included overreliance on pixel-level metrics, inadequate reporting of failure cases (85%), and dataset biases arising from single-center cohorts (80%). Conclusion Although GAN-based medical image synthesis demonstrates substantial technical promise, current evidence remains insufficient to support routine clinical deployment. Critical barriers include evaluation bias, limited generalizability, hallucination risks, and inadequate safety assessment. Future work should prioritize clinically meaningful endpoints, robust multi-center validation, standardized reporting, and explicit safety protocols. By concentrating on GAN-based methods with demonstrated experimental maturity, this work provides a clinically grounded evidence synthesis intended to inform both researchers and clinicians regarding the responsible use of image synthesis technologies in medical settings.
Introduction and Objectives This study evaluates the performance of machine learning (ML) algorithms in predicting hospital outcomes: readmission, length of stay (LOS), and mortality. Methods A systematic review was registered on the PROSPERO database (CRD42024505704) and conducted according to PRISMA guidelines. A detailed search was conducted across six databases to identify studies applying ML algorithms to forecast readmission, LOS, or mortality in general, heterogeneous hospital populations. Data collection extracted model types, input variables, and reported performance metrics. Risk of bias was assessed using the PROBAST tool. Results Fifty-two studies met the eligibility criteria. Random Forest, XGBoost, and Support Vector Machines demonstrated consistent discriminative capacity across independent cohorts. Reported area under the receiver operating characteristic curve (AUC) values reached up to 0.90 for mortality (Random Forest), >0.75 for LOS (XGBoost), and 0.79 for readmission (Random Forest). AUC was the most frequently reported metric (60–85%), whereas sensitivity, specificity, and calibration measures (e.g., Brier score) appeared infrequently. Age, comorbidity burden, and prior admissions were the predominant predictors. Socioeconomic variables were underrepresented. Conclusions Despite high discriminative performance, systemic methodological deficits prevent clinical adoption. The limited reporting of calibration measures restricts the assessment of the reliability of predicted risks, and the underrepresentation of socioeconomic determinants may compromise algorithmic equity. Future clinical implementation mandates standardized metric reporting, prospective validation, and the strict integration of social determinants of health.
In the current era of precision psychiatry, the diagnosis of psychiatric illnesses, particularly schizophrenia, remains challenging and complex task for clinician. The objective of this study is to bring an help tool in this diagnosis using biomarkers such as electroencephalograms (EEGs) and approaches based on machine learning (ML). In our work, we propose to quantify functional connectivity in schizophrenia using several methods (Phase Lag Index, Weighted Phase Lag Index, Squared Coherence and Shannon Entropy), to extract connectivity graphs from EEG signals and to compute topological features such as clustering coefficient, number of triangles, strength, and eigenvector centrality. These features will be used for classification using ML approaches. The main contribution of our approach is the comparison of several connectivity methods based on the computing time, the accuracy, recall, the classifier algorithm, and the train set size. The choice of the threshold for the connectivity methods is also discussed. The results obtained with around 17 classification model for each connectivity method show that Shannon entropy provides better results with an accuracy of around 98%, a sensitivity and specificity of 96%, and a computing time in the order of a few seconds. Our results show also that it is possible to diagnosis the schizophrenia using EEG as biomarker if the classification pipeline is well designed.
