Article
Medicine, General & Internal
Abin Abraham, Brian Le, Idit Kosti, Peter Straub, Digna R. Velez-Edwards, Lea K. Davis, J. M. Newton, Louis J. Muglia, Antonis Rokas, Cosmin A. Bejan, Marina Sirota, John A. Capra
Summary: Machine learning models based on billing codes from electronic health records can accurately predict singleton preterm birth risk and outperform models trained on known risk factors. These models also stratify deliveries into interpretable groups and predict preterm birth subtypes, mode of delivery, and recurrent preterm birth. This study suggests that machine learning has great potential to improve medical care during pregnancy.
Article
Mathematics
Nicholas Christakis, Dimitris Drikakis
Summary: This paper presents the development of a novel algorithm called RUN-ICON for unsupervised learning. It aims to improve the reliability and confidence of unsupervised clustering by leveraging the K-means++ method and introducing novel metrics. The algorithm has notable characteristics such as robustness, high-quality clustering, automation, and flexibility, and extensive testing has demonstrated its capability to determine the optimal number of clusters under different scenarios. It will soon undergo rigorous testing in real-world scenarios to further prove its effectiveness.
Article
Computer Science, Information Systems
Sarvesh Soni, Surabhi Datta, Kirk Roberts
Summary: This article proposes a system called quEHRy, which retrieves precise and interpretable answers to natural language questions from structured data in electronic health records (EHRs). The performance of quEHRy is evaluated on 2 clinical question answering (QA) datasets, and the results show high precision and accuracy. However, errors in medical concept extraction affect the downstream generation of correct logical structures, indicating the need for QA-specific clinical concept normalizers.
JOURNAL OF THE AMERICAN MEDICAL INFORMATICS ASSOCIATION
(2023)
Review
Cardiac & Cardiovascular Systems
Pishoy Gouda, Justin Ezekowitz
Summary: The use of electronic medical records has provided new opportunities for cardiovascular research, such as improving quality of care and examining genetic and pharmacological associations. However, further assessment of true clinical utility is needed before fully embracing these opportunities.
JOURNAL OF CARDIOVASCULAR TRANSLATIONAL RESEARCH
(2023)
Review
Medical Informatics
Xinyu Yang, Dongmei Mu, Hao Peng, Hua Li, Ying Wang, Ping Wang, Yue Wang, Siqi Han
Summary: This study reviewed the application and limitations of artificial intelligence based on electronic health records in cancer care. The results demonstrated the good performance of artificial intelligence in cancer emergencies, prognostic estimates, diagnosis and prediction, tumor stage detection, cancer case detection, and treatment pattern recognition. Improvement in AI performance and the development of new methods and electronic health record data sharing are needed, along with increased support from cancer specialists.
JMIR MEDICAL INFORMATICS
(2022)
Article
Cardiac & Cardiovascular Systems
Emily Kogan, Eva-Maria Didden, Eileen Lee, Anderson Nnewihe, Dimitri Stamatiadis, Samson Mataraso, Deborah Quinn, Daniel Rosenberg, Christel Chehoud, Charles Bridges
Summary: This study developed a machine learning model based on a US-based electronic health record database to identify patients with pulmonary hypertension (PH). The model used diagnostic, treatment, and procedure codes to identify PH and control patients, and achieved an AUROC of 0.92. The model showed good performance in subgroups of patients with different types of PH.
INTERNATIONAL JOURNAL OF CARDIOLOGY
(2023)
Article
Health Care Sciences & Services
Philipp Roechner, Franz Rothlauf
Summary: This study investigates two unsupervised anomaly detection approaches, pattern-based (FindFPOF) and compression-based (autoencoder), to identify implausible electronic health records in cancer registries. The results show that both methods are effective in detecting implausible electronic health records.
BMC MEDICAL RESEARCH METHODOLOGY
(2023)
Review
Nutrition & Dietetics
Stefania Russo, Stefano Bonassi
Summary: This article describes the applications of machine learning in nutritional epidemiology and provides guidelines to avoid common pitfalls. Machine learning has the potential to improve modeling of nonlinear associations and confounding in nutritional data, but its cautious application is necessary to ensure the scientific quality of the results.
Article
Mathematics
Nicholas Christakis, Dimitris Drikakis
Summary: This paper discusses the use of unsupervised learning to classify particle-like dispersion and its relevance to various applications such as virus transmission and atmospheric pollution. The RUN-ICON algorithm of unsupervised learning is applied to classify particle spread with higher confidence and lower uncertainty compared to other algorithms, even in the presence of noise. The combination of unsupervised learning and the RUN-ICON algorithm provides a tool for studying particle dynamics and their impact on air quality, health, and climate.
Article
Computer Science, Information Systems
Chathurika S. Wickramasinghe, Kasun Amarasinghe, Daniel L. Marino, Craig Rieger, Milos Manic
Summary: Cyber-Physical Systems (CPSs) are crucial in modern infrastructure, but issues of reliability, performance, and security persist. While predictive Machine Learning (ML) models offer opportunities for CPSs, their black-box nature poses challenges for safety-critical systems. To maximize the use of ML in CPSs, explainable unsupervised ML models are necessary.
Article
Computer Science, Information Systems
Peyman Shokrollahi, Juan M. Zambrano Chaves, Jonathan P. H. Lam, Avishkar Sharma, Debashish Pal, Naeim Bahrami, Akshay S. Chaudhari, Andreas M. Loening
Summary: The objective of this study is to develop an algorithm to predict correct radiology titles using machine learning techniques. The proposed system can guide physicians in selecting appropriate radiology titles and improve imaging utility and diagnostic accuracy.
Review
Health Care Sciences & Services
Sahar Borna, Michael J. Maniaci, Clifton R. Haider, Karla C. Maita, Ricardo A. Torres-Guzman, Francisco R. Avila, Julianne J. Lunde, Jordan D. Coffey, Bart M. Demaerschalk, Antonio J. Forte
Summary: This article discusses the utilization of AI models in Health Information Exchange (HIE) and evaluates their predictive capabilities and limitations. The study found that machine learning models, particularly in oncology and cardiac failures, have shown high accuracy in predicting clinical outcomes. However, variations in sensitivity metrics pose challenges.
Article
Orthopedics
Bardia Khosravi, Alexander D. Weston, Fred Nugen, John P. Mickley, Hilal Maradit Kremers, Cody C. Wyles, Rickey E. Carter, Michael J. Taunton
Summary: Electronic health records have made it easier to extract and analyze large amounts of data, but traditional statistical methods struggle with the complexities of high-dimensional data. As a result, researchers are turning to machine learning methods to uncover hidden patterns in patient data.
JOURNAL OF ARTHROPLASTY
(2023)
Review
Orthopedics
Christina M. Eckhardt, Sophia J. Madjarova, Riley J. Williams, Mattheu Ollivier, Jon Karlsson, Ayoosh Pareek, Benedict U. Nwachukwu
Summary: Unsupervised machine learning methods are crucial tools for analyzing high-dimensional data, identifying latent patterns and hidden structures, and supporting health sciences research.
KNEE SURGERY SPORTS TRAUMATOLOGY ARTHROSCOPY
(2023)
Review
Computer Science, Information Systems
Amir Kamel Rahimi, Oliver J. Canfell, Wilkin Chan, Benjamin Sly, Jason D. Pole, Clair Sullivan, Sally Shrapnel
Summary: This study systematically reviewed the literature on the development and validation of machine learning (ML) models using electronic medical records (EMR) for improving the care of hospitalized adult patients with diabetes. The study found a limited number of ML models developed for inpatient management of diabetes, but none of these models have been implemented in real hospital settings.
INTERNATIONAL JOURNAL OF MEDICAL INFORMATICS
(2022)
Article
Computer Science, Artificial Intelligence
Chen Zhao, Joyce H. Keyak, Jinshan Tang, Tadashi S. Kaneko, Sundeep Khosla, Shreyasee Amin, Elizabeth J. Atkinson, Lan-Juan Zhao, Michael J. Serou, Chaoyang Zhang, Hui Shen, Hong-Wen Deng, Weihua Zhou
Summary: In this study, we aimed to develop a deep-learning-based method for automatic proximal femur segmentation in quantitative computed tomography (QCT) images. We proposed a spatial transformation V-Net (ST-V-Net) that incorporates a shape prior into the segmentation network to improve model performance. Experimental results showed excellent performance of the proposed ST-V-Net for automatic proximal femur segmentation in QCT images.
COMPLEX & INTELLIGENT SYSTEMS
(2023)
Article
Psychology, Clinical
Euijung Ryu, Gregory D. Jenkins, Yanshan Wang, Mark Olfson, Ardesheer Talati, Lauren Lepow, Brandon J. Coombes, Alexander W. Charney, Benjamin S. Glicksberg, J. John Mann, Myrna M. Weissman, Priya Wickramaratne, Jyotishman Pathak, Joanna M. Biernacka
Summary: This study aimed to identify the most relevant social determinants of health (SDoH) related to major depressive disorder (MDD) in older adults. The results showed that the perceived level of social activity was the most influential SDoH variable, with a lower level of social activity associated with a higher risk of MDD.
PSYCHOLOGICAL MEDICINE
(2023)
Article
Cell Biology
Abhishek Chandra, Anthony B. Lagnado, Joshua N. Farr, Madison Doolittle, Tamara Tchkonia, James L. Kirkland, Nathan K. LeBrasseur, Paul D. Robbins, Laura J. Niedernhofer, Yuji Ikeno, Joao F. Passos, David G. Monroe, Robert J. Pignolo, Sundeep Khosla
Summary: This study demonstrates that cellular senescence-driven radiation-induced osteoporosis is primarily mediated by p21(Cip1) rather than p16(Ink4a), based on the clearance of senescent cells using genetic models. This approach may be used to investigate the contributions of these pathways in other senescence-associated conditions, including aging across tissues.
Editorial Material
Endocrinology & Metabolism
Sundeep Khosla, Nicole C. Wright, Ann L. Elderkin, Douglas P. Kiel
LANCET DIABETES & ENDOCRINOLOGY
(2023)
Article
Geriatrics & Gerontology
Sara E. Espinoza, Sundeep Khosla, Joseph A. Baur, Rafael de Cabo, Nicolas Musi
Summary: The geroscience hypothesis suggests that targeting key hallmarks of aging can improve healthspan and prevent age-related diseases. Several pharmacological interventions, including senolytics, NAD(+) boosters, and metformin, are being studied for their potential benefits. Preclinical studies show that senolytic drugs improve healthspan in rodents, while increasing NAD(+) through supplementation appears to extend healthspan in model organisms. Metformin, on the other hand, has pleiotropic effects and is being examined for its potential to improve healthspan and prevent frailty in clinical trials. However, further research is needed to determine their efficacy, safety, target populations, and long-term outcomes.
JOURNALS OF GERONTOLOGY SERIES A-BIOLOGICAL SCIENCES AND MEDICAL SCIENCES
(2023)
Article
Endocrinology & Metabolism
Japneet Kaur, Dominik Saul, Madison L. Doolittle, Joshua N. Farr, Sundeep Khosla, David G. Monroe
Summary: Aging is associated with an accumulation of senescent cells in various tissues, including bones. This study found that a specific miRNA, miR-19a-3p, decreases with age in mouse and human bones. Furthermore, inducing senescence in mouse bone marrow stromal cells also reduced the levels of miR-19a-3p. The study suggests that miR-19a-3p could be a potential therapeutic target for age-related bone loss.
Article
Endocrinology & Metabolism
Stefan Bartenschlager, Alexander Cavallaro, Tobias Pogarell, Oliver Chaudry, Michael Uder, Sundeep Khosla, Georg Schett, Klaus Engelke
Summary: Opportunistic screening is a promising technique for identifying individuals at high risk for osteoporotic fracture using CT scans. This study compared the performance of four existing phantomless calibration methods and found that precalibrated phantomless calibration methods performed well.
JOURNAL OF BONE AND MINERAL RESEARCH
(2023)
Article
Multidisciplinary Sciences
Madison L. Doolittle, Dominik Saul, Japneet Kaur, Jennifer L. Rowsey, Stephanie J. Vos, Kevin D. Pavelko, Joshua N. Farr, David G. Monroe, Sundeep Khosla
Summary: The article provides a detailed characterization of senescent skeletal cells in vivo, identifying a population of senescent cells associated with age and increased in late osteoblasts/osteocytes and CD24(high) osteolineage cells. The authors also establish CD24 as a marker for skeletal cells cleared by senolytics.
NATURE COMMUNICATIONS
(2023)
Article
Endocrinology & Metabolism
Madison L. Doolittle, Sundeep Khosla, Dominik Saul
Summary: The regulation of bone mineral density (BMD) is influenced by genetics and age. Genome-wide association studies (GWAS) have identified many genes associated with BMD, but their specific mechanisms in different cell types and during aging are still unclear. By analyzing age-related transcriptomics and single-cell RNA-sequencing (scRNA-seq) datasets, this study investigated the cell-specific expression of GWAS candidate genes and identified enrichment in various cells related to bone metabolism. The findings provide potential therapeutic targets for osteoporosis treatment.
Article
Endocrinology & Metabolism
Madison L. L. Doolittle, Brittany A. A. Eckhardt, Stephanie J. J. Vos, Sarah Grain, Jennifer L. L. Rowsey, Ming Ruan, Dominik Saul, Joshua N. N. Farr, Megan M. M. Weivoda, Sundeep Khosla, David G. G. Monroe
Summary: Estrogen plays a crucial role in regulating bone mass, primarily through its action on ERalpha. Recent studies have shown that estrogen action in osteocytes is more important than in osteoclasts, and the loss of ERalpha in specific cell types results in decreased bone volume and reduced bone formation rate.
Article
Endocrinology & Metabolism
Pamela Rufus-Membere, Kara L. Holloway-Kew, Adolfo Diez-Perez, Natasha M. Appelman-Dijkstra, Mary L. Bouxsein, Erik F. Eriksen, Joshua N. Farr, Sundeep Khosla, Mark A. Kotowicz, Xavier Nogues, Mishaela Rubin, Julie A. Pasco
Summary: Impact microindentation (IMI) is a novel technique for assessing bone material strength index (BMSi) in vivo. The aim of this study was to define the reference intervals for men and women by evaluating healthy adults from multiple countries. BMSi values ranged from 48 to 101, with mean values of 84.4 +/- 6.9 for men and 79.0 +/- 9.1 for women.
CALCIFIED TISSUE INTERNATIONAL
(2023)
Article
Computer Science, Interdisciplinary Applications
Sonish Sivarajkumar, Yufei Huang, Yanshan Wang
Summary: This study proposes a novel method to pre-train fair and unbiased patient representations from EHR data using a weighted loss function. The experimental results show that this method outperforms the baseline models in fairness metrics and achieves comparable predictive performance. The study also reveals that the method captures more information from clinical features.
JOURNAL OF BIOMEDICAL INFORMATICS
(2023)
Review
Health Care Sciences & Services
David Oniani, Jordan Hilsman, Yifan Peng, Ronald K. Poropatich, Jeremy C. Pamplin, Gary L. Legault, Yanshan Wang
Summary: This article discusses the core similarities between the military and medical service and the ethical concerns posed by the application of generative AI in healthcare. It proposes a set of ethical principles called GREAT PLEA from the military perspective and introduces a framework for applying these principles. By addressing ethical dilemmas and challenges, the aim is to integrate generative AI into healthcare practice.
NPJ DIGITAL MEDICINE
(2023)
Article
Endocrinology & Metabolism
Brandon M. Wagner, Jerid W. Robinson, Timothy C. R. Prickett, Eric A. Espiner, Sundeep Khosla, Dana Gaddy, Larry J. Suva, Lincoln R. Potter
Summary: C-type natriuretic peptide (CNP) activates guanylyl cyclase-B (GC-B) to catalyze the synthesis of cGMP, which affects bone length and mass. GC-B mutant mice with increased CNP-dependent GC-B activity show increased bone length, mass, and strength. However, the mechanism by which GC-B increases bone mass remains unknown.
CALCIFIED TISSUE INTERNATIONAL
(2022)
Meeting Abstract
Endocrinology & Metabolism
Japneet Kaur, Jennifer L. Rowsey, Stephanie J. Vos, Joshua N. Farr, Sundeep Khosla, David G. Monroe
JOURNAL OF BONE AND MINERAL RESEARCH
(2022)