Browse Articles
Discover research articles across all indexed journals
Towards an innovative modular cutter of leeks for improving retained stalk quality and prolonging tool life
Empyema Necessitans
Perceived determinants of clinical practice guideline implementation for stroke rehabilitation in LMICs a multinational REFORM survey
Glucocorticoids for Pneumonia in Africa — Old Therapy, New Context
Weakly supervised regression enables interpretable tumor detection in whole-slide histopathology without negative cases
Abstract Accurate tumor detection in digital pathology whole-slide images (WSIs) is crucial for cancer diagnosis and treatment planning. Multiple Instance Learning (MIL) has emerged as a widely used approach for weakly-supervised tumor detection with large-scale data without the need for manual annotations. However, traditional MIL methods often depend on classification tasks that require tumor-free cases as negative examples, which are challenging to obtain in real-world clinical workflows, especially for surgical resection specimens. We address this limitation by reformulating tumor detection as a regression task, estimating tumor percentages from WSIs, a clinically available target across multiple cancer types. In this paper, we provide an analysis of the proposed weakly-supervised regression framework by applying it to multiple organs, specimen types and clinical scenarios. We characterize the robustness of our framework to tumor percentage as a noisy regression target, and introduce a novel concept of “amplification technique” to improve tumor detection sensitivity when learning from small tumor regions. Finally, we provide interpretable insights into the model’s predictions by analyzing visual attention and logit maps. Our code is available at https://github.com/imi-bigpicture/tumor-percentage-mil-regression .
Orforglipron, an Oral GLP-1 Receptor Agonist, in Early Type 2 Diabetes
Serum MicroRNA signatures associated with hypertrophic callus formation in polytrauma patients with traumatic brain injury
Trial of Pegcetacoplan in C3 Glomerulopathy and Immune-Complex MPGN
The association of multigenerational family characteristics with loneliness and social isolation in older adults
Abstract Loneliness and social isolation are widespread among older adults, especially in aging societies undergoing demographic and social transitions. Despite growing global interest, few studies have examined how multigenerational family characteristics contribute to these outcomes in later life. This study explored the associations between loneliness and social isolation in older adults and the demographic, health, and socioeconomic features of three family generations. This community-based cross-sectional study used a two-stage stratified random sampling method. Data were collected through structured interviews using standardized questionnaires. A total of 185 older adults (first generation), 185 adult children (second generation), and 185 grandchildren (third generation) participated. Standardized tools, including the De Jong Gierveld Loneliness Scale and the Social Isolation Questionnaire by Chalabi and Amirkafi, were used to assess outcomes. Sociodemographic and health characteristics were recorded across all generations. Statistical analyses included descriptive statistics, chi-square, t-tests, ANOVA, and both univariate and multivariate linear regression models. The prevalence of loneliness and social isolation in the older adults sample was 68.6% and 29%, respectively. Financial sufficiency in older adults was significantly associated with lower loneliness levels (B = − 0.69, p = 0.039), and rural residence predicted lower social isolation (B = − 4.43, p = 0.033). Among second-generation variables, only adult children’s residential proximity to the older parent significantly predicted lower social isolation (B = 3.73, p = 0.020). Characteristics of grandchildren showed no significant association with either outcome. Financial adequacy specifically reduces loneliness, while rural residence and children’s proximity protect against social isolation. Other intergenerational factors showed limited impact. Findings advocate for targeted interventions ensuring economic security and family proximity, offering more effective strategies than broad multigenerational approaches for promoting elder well-being.
A Randomized Trial of Shunting for Idiopathic Normal-Pressure Hydrocephalus
Multi-indicator feature extraction and temperature prediction model for spontaneous coal combustion and oxidation
Updated Evidence for Covid-19, RSV, and Influenza Vaccines for 2025–2026
Alzheimer disease predicting from clinical and MRI data using DeepALZNET dual pathway framework
Abstract Alzheimer’s Disease (AD) is a significant neurological condition characterized by progressive cognitive deterioration, with prevalence rising exponentially with age. Currently, there is no effective cure, and the disease progression impacts patients’ quality of life, often leading to severe symptoms before death. The gradual development of AD symptoms, often mistaken for typical aging, frequently leads to delayed diagnosis. This underscores the critical need for precise, early diagnostic methodologies, as timely intervention plays a crucial role in managing progression. Accordingly, this paper introduces DeepALZNET, a dual-pathway computational framework designed to enhance AD prediction by offering two independent processing pathways: one for structured clinical data and another for unstructured brain MRI scans. The first pathway combines a 1D Convolutional Neural Network (CNN) with a Random Forest classifier to analyze clinical data, while the second employs a transfer-learning-based VGG19 architecture to detect subtle structural changes in MRI scans. Empirical validation on two publicly available datasets (2k clinical cases and 40k MRI images) demonstrated that both pathways achieved competitive accuracy ( $$>90\%$$ ), with further evaluation on ADNI and OASIS benchmarks confirming robustness. Unlike recent transformer-based or attention-driven methods, which often demand large multimodal datasets and high computational resources, DeepALZNET prioritizes practical applicability, interpretability, and adaptability, operating effectively with either clinical or imaging data alone. This design bridges the gap between benchmark-driven research and deployable real-world solutions, with potential for multimodal fusion or attention integration in future work.