Deep learning techniques have been increasingly used to provide more accurate and more accessible diagnosis of thorax diseases on chest radiographs. However, due to the lack of dense annotation of large-scale chest radiograph data, this computer-aided diagnosis task is intrinsically a weakly supervised learning problem and remains challenging. In this paper, we propose an attention regularized deep convolutional neural network called Thorax-Net to diagnose 14 thorax diseases using chest radiography. Thorax-Net consists of a classification branch and an attention branch. The classification branch serves as a uniform feature extraction-classification network to free users from the troublesome handcrafted feature extraction and classifier construction. The attention branch exploits the correlation between class labels and the locations of pathological abnormalities via analyzing the feature maps learned by the classification branch. Feeding a chest radiograph to the trained Thorax-Net, a diagnosis is obtained by averaging the outputs of two branches. The proposed Thorax-Net model has been evaluated against three state-of-the-art deep learning models using the patient-wise official split of the Chest X-ray 14 dataset [1] and against other five deep learning models using the image-wise random data split. Our results show that Thorax-Net achieves an average per-class AUC of 0.7876 and 0.896 in both experiments, respectively, which are higher than the AUC values obtained by other deep models when they were trained without using external data.

Focal segmental glomerulosclerosis (FSGS) is a chronic glomerular disease with poor clinical outcomes. Podocyte loss via apoptosis is one important mechanism underlying the pathogenesis of FSGS. Recently, Yes-associated-protein (YAP), a key downstream protein in the Hippo pathway, was identified as an activator for multiple gene transcriptional factors in the nucleus to control cell proliferation and apoptosis. To investigate the potential role of YAP in the progression of FSGS, we examined kidney samples from patients with minimal change disease or FSGS and found that increases in podocyte apoptosis is positively correlated with the cytoplasmic distribution of YAP in human FSGS. Utilizing an established mT/mG transgenic mouse model and primary cultured podocytes, we found that YAP was distributed uniformly in nucleus and cytoplasm in the podocytes of control animals. Adriamycin treatment induced gradual nuclear exclusion of YAP with enhanced phospho-YAP/YAP ratio, accompanied by the induction of podocyte apoptosis both in vivo and in vitro. Moreover, we used verteporfin to treat an Adriamycin-induced FSGS mouse model, and found YAP inhibition by verteporfin induced nuclear exclusion of YAP, thus increasing podocyte apoptosis and accelerating disease progression. Therefore, our findings suggest that YAP nuclear distribution and activation in podocytes is an important endogenous anti-apoptotic mechanism during the progression of FSGS.

Objectives: To determine the prevalence of possible sarcopenia and its association with other conditions in older adults in Bengbu, China. Design, setting and participants: A cross-sectional study of 1082 community-dwelling Chinese people aged at least 60 years from March to June 2022. Methods: Handgrip strength and information regarding associated conditions were collected. Possible sarcopenia was estimated based on handgrip strength with cut-off values (<28 kg in men; <18 kg in women) recommended by the Asia Working Group for Sarcopenia in 2019. Mann-Whitney U tests, χ ² tests and binary logistic regression analyses were used to explore relationships between possible sarcopenia and associated conditions. Results: Possible sarcopenia was more prevalent in men (52.79%, n=246, age 79.43±7.33 years among men with possible sarcopenia) than in women (44.48%, n=274, age 78.90±7.71 years among women with possible sarcopenia). In men, possible sarcopenia positively correlated with high age (OR 2.658, 95% CI 1.758 to 4.019), physical inactivity (OR 2.779, 95% CI 1.646 to 4.691) and diabetes (OR 4.269, 95% CI 2.397 to 7.602), and negatively with hypertension (OR 0.586, 95% CI 0.384 to 0.893). The risk of possible sarcopenia in men decreased by 12.6% for every 1 kg/m ² increase of body mass index (OR 0.874, 95% CI 0.817 to 0.935). In women, possible sarcopenia positively correlated with high age (OR 3.821, 95% CI 2.677 to 5.455), physical inactivity (OR 2.185, 95% CI 1.488 to 3.210) and arthritis (OR 2.076, 95% CI 1.411 to 3.056). Conclusion: Possible sarcopenia is prevalent in older adults and the factors affecting possible sarcopenia are different in men and women. Health education about these target factors can be considered as a potential measure to prevent possible sarcopenia.

Chest radiography is the most common radiology examination for thoracic disease diagnosis, such as pneumonia. A tremendous number of chest X-rays prompt data-driven deep learning models in constructing computer-aided diagnosis systems for thoracic diseases. However, in realistic radiology practice, a deep learning-based model often suffers from performance degradation when trained on data with noisy labels possibly caused by different types of annotation biases. To this end, we present a novel stochastic neural ensemble learning (SNEL) framework for robust thoracic disease diagnosis using chest X-rays. The core idea of our method is to learn from noisy labels by constructing model ensembles and designing noise-robust loss functions. Specifically, we propose a fast neural ensemble method that collects parameters simultaneously across model instances and along optimization trajectories. Moreover, we propose a loss function that both optimizes a robust measure and characterizes a diversity measure of ensembles. We evaluated our proposed SNEL method on three publicly available hospital-scale chest X-ray datasets. The experimental results indicate that our method outperforms competing methods and demonstrate the effectiveness and robustness of our method in learning from noisy labels. Our code is available at https://github.com/hywang01/SNEL.

Background: Post-stroke cognitive impairment (PSCI) is a prevalent complication among stroke survivors. It not only directly impacts patients' cognitive abilities but also hampers their capacity to regain independence in daily activities, consequently diminishing their quality of life. Among the various cognitive deficits following stroke, impaired attention is the most frequently observed, influencing not only daily functioning but also higher cognitive functions like working memory, executive functioning, and language.Emerging evidence indicates that Baduanjin, a traditional Chinese exercise, may have a positive impact on enhancing attention in older adults with mild cognitive impairment and stroke survivors. However, the precise mechanisms behind this effect remain unclear. In this study, we employed Baduanjin training as an intervention to address attention decline in post-stroke cognitive impairment patients and to delve into the potential mechanisms through which Baduanjin training may enhance attention in individuals with PSCI. Methods: In this prospective randomized controlled trial, we plan to recruit 72 participants diagnosed with post-stroke cognitive impairment (PSCI). These participants will be randomly assigned in a 1:1:1 ratio to one of three groups: Baduanjin training(left hemisphere stroke and right hemisphere stroke) and conventional treatment.The conventional treatment group will receive standard rehabilitation sessions. In addition to conventional treatment, participants in the octogenarian training groups will undergo octogenarian training sessions lasting 40 min, five times a week, over a total period of 12 weeks.The primary outcome measures will include the Montreal Cognitive Assessment (MoCA) scale and the Attentional Lateralization Index. These assessments will be conducted by a trained evaluator before the start of the intervention and at weeks 6 and 12 after the intervention begins.Secondary outcome measures will be assessed using the baseline Mandarin version of the Oxford Cognitive Screening (OCS-P) scale, the simplified Fugl-Meyer Motor Function Assessment (FMA) scale, the Pittsburgh Rehabilitation Participation (PRPS) scale, and the Activities of Daily Living (ADL) scale before and after the intervention, respectively. Discussion: This trial aims to examine the impact of Baduanjin training on attentional lateralization among patients with post-stroke cognitive impairment (PSCI). Functional brain imaging utilizing near-infrared spectroscopy will be employed to investigate how Baduanjin exercise influences the structural and functional connectivity of distinct brain regions or brain networks. Trial registration: Chictr.org.cn, ID: ChiCTR2300076533. Registered on 11 October 2023.