Review
Cardiac & Cardiovascular Systems
Evangelos K. Oikonomou, Rohan Khera
Summary: Artificial intelligence and machine learning have the potential to revolutionize healthcare, particularly in the management of diabetes and its cardiovascular complications. This review provides an overview of the various data-driven methods and their application in personalized care for diabetes patients at increased cardiovascular risk. The article discusses the role of artificial intelligence in diagnosis, prognostication, phenotyping, and treatment, as well as the challenges and ethical considerations that arise. It also emphasizes the need for regulatory standards to ensure the effectiveness and safety of medical artificial intelligence products in transforming cardiovascular care and outcomes in diabetes.
CARDIOVASCULAR DIABETOLOGY
(2023)
Article
Computer Science, Artificial Intelligence
Xiyue Wang, Sen Yang, Jun Zhang, Minghui Wang, Jing Zhang, Wei Yang, Junzhou Huang, Xiao Han
Summary: In this study, a novel self-supervised learning strategy called SRCL is proposed to generate more positive pairs by comparing relevance between instances, increasing the diversity of positives and resulting in more informative representations. The CTransPath backbone, pretrained on massively unlabeled histopathological images, is used to learn universal feature representations for tasks in the histopathology image domain. The results show that the SRCL-pretrained CTransPath achieves state-of-the-art performance in various downstream tasks and exhibits more robustness and transferability than other SSL methods and ImageNet pretraining.
MEDICAL IMAGE ANALYSIS
(2022)
Review
Engineering, Biomedical
Bjoern M. Eskoficr, Jochen Klucken
Summary: Artificial intelligence (AI) and machine learning (ML) methods are widely used in medicine and healthcare. However, there is a lack of comprehensive research on the use of these methods in longitudinal analysis and prediction of an individual patient's health status within a personalized disease pathway. This review aims to fill that gap by providing an overview of the AI and ML methods employed in this field and discussing the strengths, limitations, and future research directions.
ANNUAL REVIEW OF BIOMEDICAL ENGINEERING
(2023)
Article
Computer Science, Theory & Methods
Guan-Nan Dong, Chi-Man Pun, Zheng Zhang
Summary: The study introduces a novel deep collaborative multi-modal learning (DCML) to integrate underlying facial properties information adaptively for effective unsupervised kinship verification. By constructing an adaptive feature fusion mechanism and using self-supervised learning, the proposed method outperforms state-of-the-art kinship verification methods.
IEEE TRANSACTIONS ON INFORMATION FORENSICS AND SECURITY
(2021)
Review
Biochemistry & Molecular Biology
Jaroslaw Polanski
Summary: The availability of computers has opened new possibilities in drug design, and neural networks have played an important role. However, with the recent success of deep learning, there has been a resurgence in the use of neural networks in deep chemistry. Self-organizing maps have been found to be highly efficient in molecular representation. While deep learning has shown efficiency in other areas, its application in deep chemistry still faces challenges.
INTERNATIONAL JOURNAL OF MOLECULAR SCIENCES
(2022)
Article
Computer Science, Artificial Intelligence
Franz A. Van-Horenbeke, Angelika Peer
Summary: Unsupervised feature learning refers to learning useful feature extraction functions from unlabeled data. We propose a new neural network, Neocortex-Inspired Locally Recurrent Neural Network, which combines ideas from the structure and function of the neocortex to machine learning and neural networks. Our system mimics the connection patterns in the neocortex to generalize the success of convolutional neural networks to non-image data. By comparing accuracies of classifiers trained using our learned features, we have shown that our system outperforms other shallow feature learning systems in terms of accuracy and speed of learning.
COGNITIVE COMPUTATION
(2023)
Review
Neurosciences
Katharina Schultebraucks, Bernard P. Chang
Summary: Personalized medicine offers tailored screening and management, while biomarker-enriched clinical trials in cancer research have shown increased efficiency. Providers in acute stress situations may need to make rapid decisions based on abbreviated assessments. Trauma survivors admitted to the Emergency Department have a high risk of developing posttraumatic stress psychopathologies, yet there is still a lack of accurate prognostic models and cost-effective prevention methods.
EXPERIMENTAL NEUROLOGY
(2021)
Review
Clinical Neurology
Anna K. Bonkhoff, Christian Grefkes
Summary: This review focuses on the application of artificial intelligence in stroke outcome research and discusses how a large amount of patient data can be used for individualized predictions. Different types of data, including demographic, clinical, electrophysiological, and various imaging modalities, can be utilized for predicting the outcome of stroke patients. The review also highlights the methodological issues of novel machine learning approaches and the potential of artificial intelligence in improving stroke outcomes.
Review
Cardiac & Cardiovascular Systems
Diego Sadler, Tochukwu Okwuosa, A. J. Teske, Avirup Guha, Patrick Collier, Rohit Moudgil, Abdullah Sarkar, Sherry-Ann Brown
Summary: The rapid growth of cardio-oncology programs and networks has been hindered by the limited availability of large clinical trials and challenges in providing equitable access to care. The emergence of personalized medicine, AI, and machine learning in cardio-oncology presents an opportunity to improve treatment outcomes and eliminate healthcare disparities.
FRONTIERS IN CARDIOVASCULAR MEDICINE
(2022)
Review
Biotechnology & Applied Microbiology
Yihao Liu, Minghua Wu
Summary: Deep learning has been successfully applied to various tasks in different fields, including disease diagnosis in medicine. By extracting multilevel features from medical data, deep learning helps doctors automatically assess diseases and monitor patients' physical health.
BIOENGINEERING & TRANSLATIONAL MEDICINE
(2023)
Review
Chemistry, Medicinal
Julien Guiot, Akshayaa Vaidyanathan, Louis Deprez, Fadila Zerka, Denis Danthine, Anne-Noelle Frix, Philippe Lambin, Fabio Bottari, Nathan Tsoutzidis, Benjamin Miraglio, Sean Walsh, Wim Vos, Roland Hustinx, Marta Ferreira, Pierre Lovinfosse, Ralph T. H. Leijenaar
Summary: Radiomics is a method for quantitatively analyzing medical images to create diagnostic, prognostic, and/or predictive models. It utilizes sophisticated image analysis tools and statistical methods to extract hidden information in medical images, but caution is needed to avoid overenthusiastic claims and scientific pollution.
MEDICINAL RESEARCH REVIEWS
(2022)
Review
Biochemistry & Molecular Biology
Elettra Barberis, Shahzaib Khoso, Antonio Sica, Marco Falasca, Alessandra Gennari, Francesco Dondero, Antreas Afantitis, Marcello Manfredi
Summary: This review discusses the application of recent technological innovations in mass spectrometry to metabolomics analysis, with a focus on the use of artificial intelligence (AI) strategies. The article also explores the challenges and limitations of implementing metabolomics-AI systems, as well as recent tools and studies in disease classification and biomarker identification.
INTERNATIONAL JOURNAL OF MOLECULAR SCIENCES
(2022)
Review
Health Care Sciences & Services
Maged N. Kamel Boulos, Peng Zhang
Summary: A digital twin is a virtual model of a physical entity, increasingly used in various industry sectors. In the fields of medicine and public health, digital twin technology can drive radical transformation towards precision medicine and personalized treatments. Digital twins enable learning, discovering new knowledge, and hypothesis generation, while also playing a key role in formulating highly personalized treatments and interventions.
JOURNAL OF PERSONALIZED MEDICINE
(2021)
Review
Medical Laboratory Technology
Daniel S. Herman, Daniel D. Rhoads, Wade L. Schulz, Thomas J. S. Durant
Summary: AI and ML technologies have significantly impacted laboratory medicine, but their current implementation is still in the preliminary stages. To facilitate the use of reliable and advanced ML-based technologies, further best practices need to be established, and information systems and communication infrastructure must be improved.
CLINICAL CHEMISTRY
(2021)
Article
Computer Science, Information Systems
Arvin Tashakori, Wenwen Zhang, Z. Jane Wang, Peyman Servati
Summary: Recent advances in wearable devices and IoT have resulted in a massive growth of sensor data in edge devices. However, labeling this massive data for classification tasks is challenging due to personal attributes and edge heterogeneity. In addition, concerns over data privacy and communication costs prevent centralized data accumulation and training. To address these issues, we propose SemiPFL, which supports edge users with limited labeled data sets and a large amount of unlabeled data. The framework utilizes collaboration between edge users and server training to generate personalized autoencoders for each user.
IEEE INTERNET OF THINGS JOURNAL
(2023)