Plasma proteomics-based biomarkers for predicting response to mesenchymal stem cell therapy in severe COVID-19 DOI Creative Commons
Tiantian Li, Weiqi Yao, Haibo Dong

et al.

Stem Cell Research & Therapy, Journal Year: 2023, Volume and Issue: 14(1)

Published: Dec. 10, 2023

Abstract Background The objective of this study was to identify potential biomarkers for predicting response MSC therapy by pre-MSC treatment plasma proteomic profile in severe COVID-19 order optimize choice. Methods A total 58 patients selected from our previous RCT cohort were enrolled study. responders ( n = 35) defined as whose resolution lung consolidation ≥ 51.99% (the median value consolidation) 28 days post-MSC treatment, while non-responders 23) < 51.99%. Plasma before detected using data-independent acquisition (DIA) proteomics. Multivariate logistic regression analysis used that might distinguish between and therapy. Results In total, 1101 proteins identified plasma. Compared with the non-responders, had three upregulated (CSPG2, CTRB1, OSCAR) 10 downregulated (ANXA1, AGRG6, CAPG, DDX55, KV133, LEG10, OXSR1, PICAL, PTGDS, S100A8) treatment. Using model, lower levels ANXA1 higher CTRB1 predictors therapy, AUC ROC at 0.910 (95% CI 0.818–1.000) training set. validation set, 0.767 0.459–1.000). Conclusions responsiveness appears depend on baseline level ANXA1. Clinicians should take these factors into consideration when making decision initiate COVID-19.

Language: Английский

Issues and Limitations on the Road to Fair and Inclusive AI Solutions for Biomedical Challenges DOI Creative Commons
Oliver Faust, Massimo Salvi, Prabal Datta Barua

et al.

Sensors, Journal Year: 2025, Volume and Issue: 25(1), P. 205 - 205

Published: Jan. 2, 2025

Objective: In this paper, we explore the correlation between performance reporting and development of inclusive AI solutions for biomedical problems. Our study examines critical aspects bias noise in context medical decision support, aiming to provide actionable solutions. Contributions: A key contribution our work is recognition that measurement processes introduce arising from human data interpretation selection. We concept “noise-bias cascade” explain their interconnected nature. While current models handle well, remains a significant obstacle achieving practical these models. analysis spans entire lifecycle, collection model deployment. Recommendations: To effectively mitigate bias, assert need implement additional measures such as rigorous design; appropriate statistical analysis; transparent reporting; diverse research representation. Furthermore, strongly recommend integration uncertainty during deployment ensure utmost fairness inclusivity. These comprehensive recommendations aim minimize both noise, thereby improving future support systems.

Language: Английский

Citations

2
Potential diagnostic application of a novel deep learning- based approach for COVID-19 DOI Creative Commons
Alireza Sadeghi, Mahdieh Sadeghi, Ali Sharifpour

et al.

Scientific Reports, Journal Year: 2024, Volume and Issue: 14(1)

Published: Jan. 2, 2024

Abstract COVID-19 is a highly communicable respiratory illness caused by the novel coronavirus SARS-CoV-2, which has had significant impact on global public health and economy. Detecting patients during pandemic with limited medical facilities can be challenging, resulting in errors further complications. Therefore, this study aims to develop deep learning models facilitate automated diagnosis of from CT scan records patients. The also introduced COVID-MAH-CT, new dataset that contains 4442 images 133 patients, as well 3D volumes. We proposed evaluated six different transfer for slide-level analysis are responsible detecting multi-slice spiral CT. Additionally, multi-head attention squeeze excitation residual (MASERes) neural network, model was developed patient-level analysis, analyzes all slides given patient whole accurately diagnose COVID-19. codes available at https://github.com/alrzsdgh/COVID . were able detect an accuracy more than 99%, while MASERes volumes 100%. These achievements demonstrate useful automatically both patients’ records, applied real-world utilization, particularly diagnosing cases areas facilities.

Language: Английский

Citations

10
Medical Imaging-based Artificial Intelligence in Pneumonia: A Narrative Review DOI
Yanping Yang, Wenyu Xing, Yiwen Liu

et al.

Neurocomputing, Journal Year: 2025, Volume and Issue: unknown, P. 129731 - 129731

Published: Feb. 1, 2025

Language: Английский

Citations

1
Enhancing COVID-19 detection using CT-scan image analysis and disease classification: the DI-QL approach DOI
Meshal Alharbi, Sultan Ahmad

Health and Technology, Journal Year: 2025, Volume and Issue: unknown

Published: March 6, 2025

Language: Английский

Citations

1
Segmentation of liver and liver lesions using deep learning DOI

Maryam Fallahpoor,

Dan Nguyen, Ehsan Montahaei

et al.

Physical and Engineering Sciences in Medicine, Journal Year: 2024, Volume and Issue: 47(2), P. 611 - 619

Published: Feb. 21, 2024

Language: Английский

Citations

5
Exploration of deep learning models for real-time monitoring of state and performance of anaerobic digestion with online sensors DOI

Ru Jia,

Young‐Chae Song,

Dong-Mei Piao

et al.

Bioresource Technology, Journal Year: 2022, Volume and Issue: 363, P. 127908 - 127908

Published: Sept. 7, 2022

Language: Английский

Citations

21
Improving COVID-19 CT classification of CNNs by learning parameter-efficient representation DOI Creative Commons
Yujia Xu, Hak‐Keung Lam, Guangyu Jia

et al.

Computers in Biology and Medicine, Journal Year: 2022, Volume and Issue: 152, P. 106417 - 106417

Published: Dec. 15, 2022

COVID-19 pandemic continues to spread rapidly over the world and causes a tremendous crisis in global human health economy. Its early detection diagnosis are crucial for controlling further spread. Many deep learning-based methods have been proposed assist clinicians automatic based on computed tomography imaging. However, challenges still remain, including low data diversity existing datasets, unsatisfied resulting from insufficient accuracy sensitivity of learning models. To enhance diversity, we design augmentation techniques incremental levels apply them largest open-access benchmark dataset, COVIDx CT-2A. Meanwhile, similarity regularization (SR) derived contrastive is this study enable CNNs learn more parameter-efficient representations, thus improving CNNs. The results seven commonly used demonstrate that CNN performance can be improved stably through applying designed SR techniques. In particular, DenseNet121 with achieves an average test 99.44% three trials three-category classification, normal, non-COVID-19 pneumonia, pneumonia. And achieved precision, sensitivity, specificity pneumonia category 98.40%, 99.59%, 99.50%, respectively. These statistics suggest our method has surpassed state-of-the-art CT-2A dataset.

Language: Английский

Citations

20
A Systematic Review on Deep Structured Learning for COVID-19 Screening Using Chest CT from 2020 to 2022 DOI Open Access
K. C. Santosh, Debasmita GhoshRoy, Suprim Nakarmi

et al.

Healthcare, Journal Year: 2023, Volume and Issue: 11(17), P. 2388 - 2388

Published: Aug. 24, 2023

The emergence of the COVID-19 pandemic in Wuhan 2019 led to discovery a novel coronavirus. World Health Organization (WHO) designated it as global on 11 March 2020 due its rapid and widespread transmission. Its impact has had profound implications, particularly realm public health. Extensive scientific endeavors have been directed towards devising effective treatment strategies vaccines. Within healthcare medical imaging domain, application artificial intelligence (AI) brought significant advantages. This study delves into peer-reviewed research articles spanning years 2022, focusing AI-driven methodologies for analysis screening through chest CT scan data. We assess efficacy deep learning algorithms facilitating decision making processes. Our exploration encompasses various facets, including data collection, systematic contributions, emerging techniques, encountered challenges. However, comparison outcomes between 2022 proves intricate shifts dataset magnitudes over time. initiatives aimed at developing AI-powered tools detection, localization, segmentation cases are primarily centered educational training contexts. deliberate their merits constraints, context necessitating cross-population train/test models. encompassed review 231 publications, bolstered by meta-analysis employing search keywords (COVID-19 OR Coronavirus) AND (deep imaging) both PubMed Central Repository Web Science platforms.

Language: Английский

Citations

12
MSSFN: A multi-scale sequence fusion network for ct-based diagnosis of pulmonary complications DOI

Hongfu Zeng,

Xinyu Li, Haipeng Xu

et al.

Neurocomputing, Journal Year: 2025, Volume and Issue: unknown, P. 129878 - 129878

Published: March 1, 2025

Language: Английский

Citations

0
Advancing coal and gangue classification: A novel approach using 3D-CT data and deep learning DOI

Yinyu Ye,

Liang Dong, Chenyang Zhou

et al.

Measurement, Journal Year: 2025, Volume and Issue: unknown, P. 117118 - 117118

Published: March 1, 2025

Language: Английский

Citations

0