Technology and Health Care最新文献

Background: Postoperative hypoxemia is a severe complication in patients undergoing surgery for acute Type A aortic dissection (AAD), with significant impacts on recovery and clinical outcomes. Technological advancements in risk assessment models offer opportunities for early intervention and optimized care.

Objective: To develop and validate a technology-driven predictive model for hypoxemia based on clinical and intraoperative risk factors, enhancing postoperative management strategies.

Methods: A retrospective cohort of 242 patients was analyzed, including 77 with hypoxemia (PaO₂/FiO₂ ≤ 200 mmHg) and 165 without. Key clinical variables, intraoperative factors, and postoperative outcomes were examined. Spearman correlation analysis and receiver operating characteristic (ROC) curve analysis were conducted to identify and validate predictive markers.

Results: Prolonged time from symptom onset to surgery (>48 h), aortic cross-clamp time, and deep hypothermic circulatory arrest time (DHCA) emerged as the most significant predictors (all p < 0.001). DHCA time demonstrated the highest sensitivity (0.961) and area under the curve (AUC = 0.891). Additional significant predictors included intraoperative blood product use and prolonged mechanical ventilation, with cumulative predictive value for hypoxemia risk.

Conclusion: The integration of clinical variables into a technology-enhanced prediction model provides robust early warnings of postoperative hypoxemia risk. Implementing timely surgical interventions and refined intraoperative management can minimize adverse respiratory outcomes, improving recovery in AAD patients.

BackgroundRobot-assisted feeding systems aim to promote independence for individuals with motor impairments. Despite significant technological progress, widespread adoption remains limited due to challenges related to adaptability, safety, and cost.ObjectiveThis review investigates recent advancements in robot-assisted feeding, highlights key technical and usability challenges, and outlines future directions to improve system adaptability, autonomy, and cost-effectiveness.MethodsA systematic literature search was conducted for peer-reviewed articles published in the past decade. The analysis focuses on critical domains including hardware architecture, human-robot interaction (HRI) modalities, and control strategies.ResultsAdvances in artificial intelligence (AI) and HRI have enhanced system autonomy and user adaptability. Nevertheless, unresolved issues persist in handling diverse food types, achieving real-time responsiveness, and minimizing system costs. Emerging solutions-such as adaptive learning, Artificial Intelligence of Things (AIoT) integration, and modular design-offer promising pathways to overcome these barriers and support scalable deployment in real-world care settings.

Background: Gliomas are malignant brain tumors with poor prognosis, and aging is believed to play a role in their malignant transformation. However, the relationship between aging and glioma prognosis remains unclear.

Objective: This study aims to construct and validate a prognostic risk model based on aging-related differential expression genes (ARDEGs) to understand their role in glioma prognosis and tumorigenesis, with a particular focus on immune responses.

Methods: ARDEGs were identified between LGG and HGG through LASSO regression and Cox regression. A prognostic risk model was developed and validated. GSEA and KEGG pathway analyses were performed to explore tumorigenic mechanisms in high- and low-risk groups. The correlation between the model genes and immune cell infiltration, as well as immune checkpoint molecules, was also analyzed. The protein expression of NOG was evaluated in glioma cells using WB and IHC.

Results: Three aging-related genes-IGFBP2, AGTR1, and NOG-were identified, and a prognostic model was established. KEGG and GSEA analysis revealed that the high-risk group enriched pathways related to inflammation and immune responses, while the low-risk group showed enrichment in oxidative phosphorylation and metabolism pathways. IGFBP2 and AGTR1 expression correlated positively with immunosuppressive cells and immune checkpoint molecules, whereas NOG showed an opposite trend. NOG protein expression was reduced in glioma cells and lower in high-grade gliomas compared to low-grade gliomas.

Conclusions: The prognostic risk model based on aging-related genes shows strong predictive power for glioma prognosis, highlighting the potential role of immune-related pathways and NOG in tumor progression.

BackgroundCoronary artery disease (CAD) is a leading cause of ischemic heart disease, and accurate identification of coronary lesion-specific ischemia (CLSI) is crucial for treatment. Coronary computed tomography angiography (CCTA) provides detailed visualization of coronary lesions, but its multiparameter analysis for predicting ischemia remains underexplored.ObjectiveTo develop a nomogram prediction model for CLSI based on multiparameters derived from CCTA.MethodsA total of 160 patients with CAD were divided into non-ischemic and ischemic groups according to the target-vessel CT-fractional flow reserve (CT-FFR). The baseline data of the two groups were collected, and the quantitative parameters of CCTA were compared. The predictive value of these parameters for CLSI was analyzed by the receiver operator characteristic (ROC) curve, and independent risk factors were analyzed by logistic regression.ResultsThe ischemic group showed significant differences in maximum diameter stenosis (MDS), maximum area stenosis (MAS), minimum lumen area (MLA), plaque burden (PB), pericoronary fat attenuation index (FAI), and low-attenuation plaque compared to the non-ischemic group (P < 0.05). Logistic regression revealed that MAS, MLA, FAI, and PB were independent risk factors for CLSI. The area under the curve (AUC) for MAS, MLA, FAI, and PB were 0.783, 0.947, 0.804, and 0.935, respectively. The calibration curve of the nomogram showed a good fit to the actual values [0.995 (95%CI: 0.988-1.000)].ConclusionsThis study constructed a nomogram risk prediction model for CLSI based on MAS, MLA, FAI, and PB, which holds significant clinical value.

Background: Post-stroke cognitive impairment (PSCI) poses significant challenges to patient independence, yet technological interventions like high-frequency repetitive transcranial magnetic stimulation (rTMS) remain underexplored in clinical neurorehabilitation.

Objective: This study evaluates the integration of high-frequency rTMS into standard care, focusing on its technological efficacy in modulating neuroplasticity and serum biomarkers to enhance cognitive and functional recovery.

Methods: A retrospective analysis of 80 PSCI patients (2021-2023) compared outcomes between a conventional care group (n = 30) and an rTMS group (n = 50) receiving 20 Hz stimulation (YRD-CCY-I device) targeting the dorsolateral prefrontal cortex. Key metrics included Montreal Cognitive Assessment (MoCA), Barthel Index (BI), cortical silent period (CL), central motor conduction time (CMCT), and serum neurotrophic factors (BDNF, VEGF, IGF-1).

Results: Post-intervention, the rTMS group demonstrated superior MoCA scores (19.25 vs. 15.24, p = 0.001), BI (76.36 vs. 70.13, p = 0.001), and IADL (20.38 vs. 18.13, p = 0.001) compared to controls. Neurophysiological markers revealed prolonged CL (25.30 vs. 24.02 ms, p = 0.001) and shortened CMCT (12.05 vs. 12.98 ms, p = 0.001), alongside elevated BDNF (9.56 vs. 7.34 ng/mL), VEGF (156.48 vs. 110.54 pg/mL), and IGF-1 (153.74 vs. 112.90 ng/mL, p = 0.001). Overall efficacy was 94% for rTMS versus 73.33% for conventional care (p = 0.047).

Conclusion: High-frequency rTMS, as a targeted neurostimulation technology, enhances cognitive recovery and cortical adaptability in PSCI by modulating neuroplasticity and upregulating neurotrophic biomarkers. These findings underscore its potential as a scalable adjunct in technology-driven neurorehabilitation programs.

Background: The complexity and diversity of brain activity patterns make it difficult to accurately diagnose neurological disorders such epilepsy, Parkinson's disease, schizophrenia, stroke, and Alzheimer's disease. Integrated and effective analysis of multiple data sources is often beyond the scope of traditional diagnostic procedures. With the use of multi-modal data, recent developments in neural network approaches present encouraging opportunities for raising diagnostic accuracy.

Objectives: A novel approach has been proposed toward the integration of different Nondestructive Evaluation data with EEG signals for improving the diagnosis of neurological disorders such as stroke, epilepsy, Parkinson's disease, and schizophrenia, by leveraging advanced neural network techniques in order to improve the identification and correlation of shared latent features across heterogeneous NDE datasets.

Methods: We determined the 2D scalogram images using a specific encoder-decoder neural network after transforming the EEG signals using wavelet signal processing. Several NDE data types can be easily integrated for thorough analysis due to this network's ability to extract and correlate important aspects from each form of data. Aiming to uncover common patterns indicating of neurological disorders, the technique was evaluated on datasets containing EEG signals and corresponding NDE data.

Results: Our method demonstrated a significant improvement in diagnostic accuracy and efficiency. The encoder-decoder network effectively identified shared latent features across the heterogeneous NDE datasets, leading to more precise and reliable diagnoses. The fusion of multi-modality NDE data with EEG signals provided a robust framework for the automatic identification of multiple neurological disorders.

Conclusions: This innovative approach represents a substantial advancement in the field of neurological disorder diagnosis. By integrating diverse NDE data with EEG signals through advanced neural network techniques, we have developed a method that enhances the accuracy and efficiency of diagnosing multiple neurological conditions. This fusion of multi-modality data has the potential to revolutionize current diagnostic practices in neurology, paving the way for more precise and automated identification of neurological disorders.

Objective: To detect the changes in the thickness of the Retinal Nerve Fiber Layer (RNFL) in patients with High Myopia (HM) complicated with glaucoma through Optical Coherence Tomography (OCT).

Methods: 80 patients (160 eyes) with HM complicated with glaucoma treated from March 2018 to March 2020 were enrolled as the experimental group, and 60 healthy volunteers (120 eyes) undergoing physical examination in the same period were selected as the control group. OCT measured their RNFL thicknesses.

Results: Compared with that in the control group, the nasal, supratemporal, subnasal, supranasal, and infratemporal RNFL thickness and overall mean RNFL thickness in the experimental group was significantly decreased, while the temporal RNFL thickness was significantly increased in the experimental group (P < 0.05). According to the diopter, patients in the experimental group were assigned into group A (n = 25, 50 eyes, diopter range: ≥ -6.00 D and ≤ -8.00 D), group B (n = 30, 60 eyes, diopter range: > -8.00 D and ≤ -10.00 D) and group C (n = 25, 50 eyes, diopter range: > -10.00 D). The nasal, supratemporal, subnasal, supranasal, and infratemporal RNFL thickness and overall mean RNFL thickness in group A were significantly greater than those in groups B and C (P < 0.05). Spearman correlation analysis revealed that the absolute value of diopter was negatively correlated with the nasal, supratemporal, subnasal, supranasal, and infratemporal RNFL thickness and overall mean RNFL thickness (P < 0.05), and positively correlated with the thickness of temporal RNFL (P < 0.05).

Conclusion: In patients with HM complicated with glaucoma, RNFL is thinner in all quadrants except for temporal RNFL.

Ovarian cancer is a highly prevalent cancer among women; However, it remains difficult to find effective pharmacological solutions to treat this deadly disease. However, early detection can significantly increase life expectancy. To address this issue, a predictive model for early diagnosis of ovarian cancer was developed by applying statistical techniques and machine learning models to clinical data from 349 patients. A hybrid evolutionary deep learning model was proposed by integrating genetic and histopathological imaging modalities within a multimodal fusion framework. Machine learning pipelines have been built using feature selection and dilution approaches to identify the most relevant genes for disease classification. A comparison was performed between the UNeT and transformer models for semantic segmentation, leading to the development of an optimized fuzzy C-means clustering algorithm (FCM-NPOA-PM-UI) for the classification of gynecological abdominopelvic tumors. Performing better than individual classifiers and other machine learning methods, the suggested ensemble model achieved an average accuracy of 98.96%, precision of 97.44%, and F1 score of 98.7%. With average Dice scores of 0.98 and 0.97 for positive tumors and 0.99 and 0.98 for malignant tumors, the Transformer model performed better in segmentation than the UNeT model. Additionally, we observed a 92.8% increase in accuracy when combining five machine learning models with biomarker data: random forest, logistic regression, SVM, decision tree, and CNN. These results demonstrate that the hybrid model significantly improves the accuracy and efficiency of ovarian cancer detection and classification, offering superior performance compared to traditional methods and individual classifiers.

BackgroundMedical waste should be collected, classified, and transported to the treatment plant within 48 h. If it is not disposed of in time, it will cause cross-infection, increasing the risk of disease transmission and environmental pollution. How to reasonably plan transportation routes to ensure that the medical waste can be transported to the treatment plant in time is very important.ObjectiveThere are usually two modes of transportation, the fastest speed and shortest path, how to reasonably plan the transportation scheme so that medical waste can be transported to the treatment plant for disposal in the specified time is the main purpose of this article.MethodsThe multi-agent modeling method is adopted. AnyLogic simulation software is used to model the transportation routes of 118 Grade III hospitals and 2 treatment plants in Beijing under the two transportation modes of fastest speed and shortest path.ResultsBased on the traffic index in Beijing, the speed range of 20 km/h-32 km/h is set up and divided into 4 parts and 24 levels with 0.5 km/h as the unit, and the 24 levels of medical waste transportation data set is formed. The key speed nodes of 21 km/h, 24 km/h and 29.5 km/h are identified.ConclusionsThe medical waste transportation model and transport data set formed in this paper have enriched the theory and data basis of medical waste transportation management. The key speed nodes of transportation model selection have important practical significance for the transportation management decision of medical waste in big cities.

Background: Despite wearable sensors' ability to provide continuous physiologic monitoring, migraine remains challenging to predict due to unpredictability of onset and a variety of triggers. Developing an accurate prediction model requires reducing signal variability by using effective filtering techniques.

Objective: The main objective of this study is to evaluate machine learning models for predicting migraines and analyze the effect of different filtering techniques and classifiers on prediction performance.

Methods: A feature set based on ANOVA analysis of four key physiological signals was used. After the pre-processing, filtering methods, including median, Butterworth, and Savitzky-Golay filter, were applied. Five classification models, Extreme Gradient Boosting, Histogram-Based Gradient Boosting, Random Forest, Support Vector Machine, and K-Nearest Neighbors, were evaluated.

Results: The highest predictive performance was achieved using the Savitzky-Golay filter. The Random Forest model demonstrated the best accuracy (0.858) and precision (0.815), and an F1-score of 0.677, indicating the potential of investigated signals for migraine prediction. Furthermore, the Histogram-Based Gradient Boosting model achieved the highest recall using the Savitzky-Golay filter (0.719), demonstrating its effectiveness in identifying true positive cases of migraines.

Conclusion: The results indicate significant potential for healthcare applications for early migraine prediction and treatment using wearable technology. The study highlights the importance of selecting appropriate features and filtering methods to improve the accuracy and reliability of the predictions.