Search

Abstract Since the first studies on bowhead whale singing behaviour, song variations have been consistently reported. However, there has been little discussion regarding variability in bowhead whale singing display and its ecological significance. Unlike the better studied humpback whales, bowhead whales do not appear to share songs at population level, but several studies have reported song sharing within clusters of animals. Over the winter season 2013–2014, in an unstudied wintering ground off Northeast Greenland, 13 song groups sharing similar hierarchical structure and units were identified. Unit types were assessed through multidimensional maps, showing well separated clusters corresponding to manually labelled units, and revealing the presence of unit subtypes. Units presented contrasting levels of variability over their acoustic parameters, suggesting that bowhead whales keep consistency in some units while using a continuum in values of frequency, duration and modulation parameters for other unit types. Those findings emphasise the need to account for variability in song analysis to better understand the behavioural ecology of this endangered species. Additionally, shifting from song toward units or phrase-based analysis, as it has been suggested for humpback whales, offers the opportunity to identify and track similarities in songs over temporal and geographical scales relevant to population monitoring.

Enteroviruses reorganize cellular endomembranes into replication organelles (ROs) for genome replication. Although enterovirus replication depends on phosphatidylinositol 4-kinase type IIIβ (PI4KB), its role, and that of its product, phosphatidylinositol 4-phosphate (PI4P), is only partially understood. Exploiting a mutant coxsackievirus resistant to PI4KB inhibition, we show that PI4KB activity has distinct functions both in proteolytic processing of the viral polyprotein and in RO biogenesis. The escape mutation rectifies a proteolytic processing defect imposed by PI4KB inhibition, pointing to a possible escape mechanism. Remarkably, under PI4KB inhibition, the mutant virus could replicate its genome in the absence of ROs, using instead the Golgi apparatus. This impaired RO biogenesis provided an opportunity to investigate the proposed role of ROs in shielding enteroviral RNA from cellular sensors. Neither accelerated sensing of viral RNA nor enhanced innate immune responses was observed. Together, our findings challenge the notion that ROs are indispensable for enterovirus genome replication and immune evasion.

ABSTRACT Enterovirus genome replication occurs at virus-induced structures derived from cellular membranes and lipids. However, the origin of these replication organelles (ROs) remains uncertain. Ultrastructural evidence of the membrane donor is lacking, suggesting that the sites of its transition into ROs are rare or fleeting. To overcome this challenge, we combined live-cell imaging and serial block-face scanning electron microscopy of whole cells to capture emerging enterovirus ROs. The first foci of fluorescently labeled viral protein correlated with ROs connected to the endoplasmic reticulum (ER) and preceded the appearance of ROs stemming from the trans-Golgi network. Whole-cell data sets further revealed striking contact regions between ROs and lipid droplets that may represent a route for lipid shuttling to facilitate RO proliferation and genome replication. Our data provide direct evidence that enteroviruses use ER and then Golgi membranes to initiate RO formation, demonstrating the remarkable flexibility with which enteroviruses usurp cellular organelles. IMPORTANCE Enteroviruses are causative agents of a range of human diseases. The replication of these viruses within cells relies on specialized membranous structures termed replication organelles (ROs) that form during infection but whose origin remains elusive. To capture the emergence of enterovirus ROs, we use correlative light and serial block-face scanning electron microscopy, a powerful method to pinpoint rare events in their whole-cell ultrastructural context. RO biogenesis was found to occur first at ER and then at Golgi membranes. Extensive contacts were found between early ROs and lipid droplets (LDs), which likely serve to provide LD-derived lipids required for replication. Together, these data establish the dual origin of enterovirus ROs and the chronology of their biogenesis at different supporting cellular membranes.

ABSTRACT The enterovirus genus of the picornavirus family includes a large number of important human pathogens such as poliovirus, coxsackievirus, enterovirus A71, and rhinoviruses. Like all other positive-strand RNA viruses, genome replication of enteroviruses occurs on rearranged membranous structures called replication organelles (ROs). Phosphatidylinositol 4-kinase IIIβ (PI4KB) is required by all enteroviruses for RO formation. The enteroviral 3A protein recruits PI4KB to ROs, but the exact mechanism remains elusive. Here, we investigated the role of acyl-coenzyme A binding domain containing 3 (ACBD3) in PI4KB recruitment upon enterovirus replication using ACBD3 knockout (ACBD3KO) cells. ACBD3 knockout impaired replication of representative viruses from four enterovirus species and two rhinovirus species. PI4KB recruitment was not observed in the absence of ACBD3. The lack of ACBD3 also affected the localization of individually expressed 3A, causing 3A to localize to the endoplasmic reticulum instead of the Golgi. Reconstitution of wild-type (wt) ACBD3 restored PI4KB recruitment and 3A localization, while an ACBD3 mutant that cannot bind to PI4KB restored 3A localization, but not virus replication. Consistently, reconstitution of a PI4KB mutant that cannot bind ACBD3 failed to restore virus replication in PI4KBKO cells. Finally, by reconstituting ACBD3 mutants lacking specific domains in ACBD3KO cells, we show that acyl-coenzyme A binding (ACB) and charged-amino-acid region (CAR) domains are dispensable for 3A-mediated PI4KB recruitment and efficient enterovirus replication. Altogether, our data provide new insight into the central role of ACBD3 in recruiting PI4KB by enterovirus 3A and reveal the minimal domains of ACBD3 involved in recruiting PI4KB and supporting enterovirus replication. IMPORTANCE Similar to all other positive-strand RNA viruses, enteroviruses reorganize host cellular membranes for efficient genome replication. A host lipid kinase, PI4KB, plays an important role in this membrane rearrangement. The exact mechanism of how enteroviruses recruit PI4KB was unclear. Here, we revealed a role of a Golgi-residing protein, ACBD3, as a mediator of PI4KB recruitment upon enterovirus replication. ACBD3 is responsible for proper localization of enteroviral 3A proteins in host cells, which is important for 3A to recruit PI4KB. By testing ACBD3 and PI4KB mutants that abrogate the ACBD3-PI4KB interaction, we showed that this interaction is crucial for enterovirus replication. The importance of specific domains of ACBD3 was evaluated for the first time, and the domains that are essential for enterovirus replication were identified. Our findings open up a possibility for targeting ACBD3 or its interaction with enteroviruses as a novel strategy for the development of broad-spectrum antienteroviral drugs.

ABSTRACT Picornaviruses induce dramatic rearrangements of endomembranes in the cells that they infect to produce dedicated platforms for viral replication. These structures, termed replication organelles (ROs), have been well characterized for the Enterovirus genus of the Picornaviridae. However, it is unknown whether the diverse RO morphologies associated with enterovirus infection are conserved among other picornaviruses. Here, we use serial electron tomography at different stages of infection to assess the three-dimensional architecture of ROs induced by encephalomyocarditis virus (EMCV), a member of the Cardiovirus genus of the family of picornaviruses that is distantly related. Ultrastructural analyses revealed connections between early single-membrane EMCV ROs and the endoplasmic reticulum (ER), establishing the ER as a likely donor organelle for their formation. These early single-membrane ROs appear to transform into double-membrane vesicles (DMVs) as infection progresses. Both single- and double-membrane structures were found to support viral RNA synthesis, and progeny viruses accumulated in close proximity, suggesting a spatial association between RNA synthesis and virus assembly. Further, we explored the role of phosphatidylinositol 4-phosphate (PI4P), a critical host factor for both enterovirus and cardiovirus replication that has been recently found to expedite enterovirus RO formation rather than being strictly required. By exploiting an EMCV escape mutant, we found that low-PI4P conditions could also be overcome for the formation of cardiovirus ROs. Collectively, our data show that despite differences in the membrane source, there are striking similarities in the biogenesis, morphology, and transformation of cardiovirus and enterovirus ROs, which may well extend to other picornaviruses. IMPORTANCE Like all positive-sense RNA viruses, picornaviruses induce the rearrangement of host cell membranes to form unique structures, or replication organelles (ROs), that support viral RNA synthesis. Here, we investigate the architecture and biogenesis of cardiovirus ROs and compare them with those induced by enteroviruses, members of the well-characterized picornavirus genus Enterovirus. The origins and dynamic morphologies of cardiovirus ROs are revealed using electron tomography, which points to the endoplasmic reticulum as the donor organelle usurped to produce single-membrane tubules and vesicles that transform into double-membrane vesicles. We show that PI4P, a critical lipid for cardiovirus and enterovirus replication, is not strictly required for the formation of cardiovirus ROs, as functional ROs with typical morphologies are formed under phosphatidylinositol 4-kinase type III alpha (PI4KA) inhibition in cells infected with an escape mutant. Our data show that the transformation from single-membrane structures to double-membrane vesicles is a conserved feature of cardiovirus and enterovirus infections that likely extends to other picornavirus genera.

Itraconazole (ITZ) is a well-known antifungal agent that also has anticancer activity. In this study, we identify ITZ as a broad-spectrum inhibitor of enteroviruses (e.g., poliovirus, coxsackievirus, enterovirus-71, rhinovirus). We demonstrate that ITZ inhibits viral RNA replication by targeting oxysterol-binding protein (OSBP) and OSBP-related protein 4 (ORP4). Consistently, OSW-1, a specific OSBP/ORP4 antagonist, also inhibits enterovirus replication. Knockdown of OSBP inhibits virus replication, whereas overexpression of OSBP or ORP4 counteracts the antiviral effects of ITZ and OSW-1. ITZ binds OSBP and inhibits its function, i.e., shuttling of cholesterol and phosphatidylinositol-4-phosphate between membranes, thereby likely perturbing the virus-induced membrane alterations essential for viral replication organelle formation. ITZ also inhibits hepatitis C virus replication, which also relies on OSBP. Together, these data implicate OSBP/ORP4 as molecular targets of ITZ and point to an essential role of OSBP/ORP4-mediated lipid exchange in virus replication that can be targeted by antiviral drugs.

ABSTRACT Like all other positive-strand RNA viruses, enteroviruses generate new organelles (replication organelles [ROs]) with a unique protein and lipid composition on which they multiply their viral genome. Suitable tools for live-cell imaging of enterovirus ROs are currently unavailable, as recombinant enteroviruses that carry genes that encode RO-anchored viral proteins tagged with fluorescent reporters have not been reported thus far. To overcome this limitation, we used a split green fluorescent protein (split-GFP) system, comprising a large fragment [strands 1 to 10; GFP(S1-10)] and a small fragment [strand 11; GFP(S11)] of only 16 residues. The GFP(S11) (GFP with S11 fragment) fragment was inserted into the 3A protein of the enterovirus coxsackievirus B3 (CVB3), while the large fragment was supplied by transient or stable expression in cells. The introduction of GFP(S11) did not affect the known functions of 3A when expressed in isolation. Using correlative light electron microscopy (CLEM), we showed that GFP fluorescence was detected at ROs, whose morphologies are essentially identical to those previously observed for wild-type CVB3, indicating that GFP(S11)-tagged 3A proteins assemble with GFP(S1-10) to form GFP for illumination of bona fide ROs. It is well established that enterovirus infection leads to Golgi disintegration. Through live-cell imaging of infected cells expressing an mCherry-tagged Golgi marker, we monitored RO development and revealed the dynamics of Golgi disassembly in real time. Having demonstrated the suitability of this virus for imaging ROs, we constructed a CVB3 encoding GFP(S1-10) and GFP(S11)-tagged 3A to bypass the need to express GFP(S1-10) prior to infection. These tools will have multiple applications in future studies on the origin, location, and function of enterovirus ROs. IMPORTANCE Enteroviruses induce the formation of membranous structures (replication organelles [ROs]) with a unique protein and lipid composition specialized for genome replication. Electron microscopy has revealed the morphology of enterovirus ROs, and immunofluorescence studies have been conducted to investigate their origin and formation. Yet, immunofluorescence analysis of fixed cells results in a rather static view of RO formation, and the results may be compromised by immunolabeling artifacts. While live-cell imaging of ROs would be preferred, enteroviruses encoding a membrane-anchored viral protein fused to a large fluorescent reporter have thus far not been described. Here, we tackled this constraint by introducing a small tag from a split-GFP system into an RO-resident enterovirus protein. This new tool bridges a methodological gap by circumventing the need for immunolabeling fixed cells and allows the study of the dynamics and formation of enterovirus ROs in living cells.

ABSTRACT Positive-strand RNA [(+)RNA] viruses are true masters of reprogramming host lipid trafficking and synthesis to support virus genome replication. Via their membrane-associated 3A protein, picornaviruses of the genus Enterovirus (e.g., poliovirus, coxsackievirus, and rhinovirus) subvert Golgi complex-localized phosphatidylinositol 4-kinase IIIβ (PI4KB) to generate “replication organelles” (ROs) enriched in phosphatidylinositol 4-phosphate (PI4P). PI4P lipids serve to accumulate oxysterol-binding protein (OSBP), which subsequently transfers cholesterol to the ROs in a PI4P-dependent manner. Single-point mutations in 3A render enteroviruses resistant to both PI4KB and OSBP inhibition, indicating coupled dependency on these host factors. Recently, we showed that encephalomyocarditis virus (EMCV), a picornavirus that belongs to the Cardiovirus genus, also builds PI4P/cholesterol-enriched ROs. Like the hepatitis C virus (HCV) of the Flaviviridae family, it does so by hijacking the endoplasmic reticulum (ER)-localized phosphatidylinositol 4-kinase IIIα (PI4KA). Here we provide genetic evidence for the critical involvement of EMCV protein 3A in this process. Using a genetic screening approach, we selected EMCV mutants with single amino acid substitutions in 3A, which rescued RNA virus replication upon small interfering RNA (siRNA) knockdown or pharmacological inhibition of PI4KA. In the presence of PI4KA inhibitors, the mutants no longer induced PI4P, OSBP, or cholesterol accumulation at ROs, which aggregated into large cytoplasmic clusters. In contrast to the enterovirus escape mutants, we observed little if any cross-resistance of EMCV mutants to OSBP inhibitors, indicating an uncoupled level of dependency of their RNA replication on PI4KA and OSBP activities. This report may contribute to a better understanding of the roles of PI4KA and OSBP in membrane modifications induced by (+)RNA viruses. IMPORTANCE Positive-strand RNA viruses modulate lipid homeostasis to generate unique, membranous “replication organelles” (ROs) where viral genome replication takes place. Hepatitis C virus, encephalomyocarditis virus (EMCV), and enteroviruses have convergently evolved to hijack host phosphatidylinositol 4-kinases (PI4Ks), which produce PI4P lipids, to recruit oxysterol-binding protein (OSBP), a PI4P-binding protein that shuttles cholesterol to ROs. Consistent with the proposed coupling between PI4K and OSBP, enterovirus mutants resistant to PI4KB inhibitors are also resistant to OSBP inhibitors. Here, we show that EMCV can replicate without accumulating PI4P/cholesterol at ROs, by acquiring point mutations in nonstructural protein 3A. Remarkably, the mutations conferred resistance to PI4K but not OSBP inhibitors, thereby uncoupling the levels of dependency of EMCV RNA replication on PI4K and OSBP. This work may contribute to a deeper understanding of the roles of PI4K/PI4P and OSBP/cholesterol in membrane modifications induced by positive-strand RNA viruses.

Cardioviruses, including encephalomyocarditis virus (EMCV) and the human Saffold virus, are small non-enveloped viruses belonging to the Picornaviridae, a large family of positive-sense RNA [(+)RNA] viruses. All (+)RNA viruses remodel intracellular membranes into unique structures for viral genome replication. Accumulating evidence suggests that picornaviruses from different genera use different strategies to generate viral replication organelles (ROs). For instance, enteroviruses (e.g. poliovirus, coxsackievirus, rhinovirus) rely on the Golgi-localized phosphatidylinositol 4-kinase III beta (PI4KB), while cardioviruses replicate independently of the kinase. By which mechanisms cardioviruses develop their ROs is currently unknown. Here we show that cardioviruses manipulate another PI4K, namely the ER-localized phosphatidylinositol 4-kinase III alpha (PI4KA), to generate PI4P-enriched ROs. By siRNA-mediated knockdown and pharmacological inhibition, we demonstrate that PI4KA is an essential host factor for EMCV genome replication. We reveal that the EMCV nonstructural protein 3A interacts with and is responsible for PI4KA recruitment to viral ROs. The ensuing phosphatidylinositol 4-phosphate (PI4P) proved important for the recruitment of oxysterol-binding protein (OSBP), which delivers cholesterol to EMCV ROs in a PI4P-dependent manner. PI4P lipids and cholesterol are shown to be required for the global organization of the ROs and for viral genome replication. Consistently, inhibition of OSBP expression or function efficiently blocked EMCV RNA replication. In conclusion, we describe for the first time a cellular pathway involved in the biogenesis of cardiovirus ROs. Remarkably, the same pathway was reported to promote formation of the replication sites of hepatitis C virus, a member of the Flaviviridae family, but not other picornaviruses or flaviviruses. Thus, our results highlight the convergent recruitment by distantly related (+)RNA viruses of a host lipid-modifying pathway underlying formation of viral replication sites.

ABSTRACT All positive-strand RNA viruses induce membrane structures in their host cells which are thought to serve as suitable microenvironments for viral RNA synthesis. The structures induced by enteroviruses, which are members of the family Picornaviridae, have so far been described as either single- or double-membrane vesicles (DMVs). Aside from the number of delimiting membranes, their exact architecture has also remained elusive due to the limitations of conventional electron microscopy. In this study, we used electron tomography (ET) to solve the three-dimensional (3-D) ultrastructure of these compartments. At different time points postinfection, coxsackievirus B3-infected cells were high-pressure frozen and freeze-substituted for ET analysis. The tomograms showed that during the exponential phase of viral RNA synthesis, closed smooth single-membrane tubules constituted the predominant virus-induced membrane structure, with a minor proportion of DMVs that were either closed or connected to the cytosol in a vase-like configuration. As infection progressed, the DMV number steadily increased, while the tubular single-membrane structures gradually disappeared. Late in infection, complex multilamellar structures, previously unreported, became apparent in the cytoplasm. Serial tomography disclosed that their basic unit is a DMV, which is enwrapped by one or multiple cisternae. ET also revealed striking intermediate structures that strongly support the conversion of single-membrane tubules into double-membrane and multilamellar structures by a process of membrane apposition, enwrapping, and fusion. Collectively, our work unravels the sequential appearance of distinct enterovirus-induced replication structures, elucidates their detailed 3-D architecture, and provides the basis for a model for their transformation during the course of infection. IMPORTANCE Positive-strand RNA viruses hijack specific intracellular membranes and remodel them into special structures that support viral RNA synthesis. The ultrastructural characterization of these “replication structures” is key to understanding their precise role. Here, we resolved the three-dimensional architecture of enterovirus-induced membranous compartments and their transformation in time by applying electron tomography to cells infected with coxsackievirus B3 (CVB3). Our results show that closed single-membrane tubules are the predominant initial virus-induced structure, whereas double-membrane vesicles (DMVs) become increasingly abundant at the expense of these tubules as infection progresses. Additionally, more complex multilamellar structures appear late in infection. Based on compelling intermediate structures in our tomograms, we propose a model for transformation from the tubules to DMVs and multilamellar structures via enwrapping events. Our work provides an in-depth analysis of the development of an unsuspected variety of distinct replication structures during the course of CVB3 infection.

Cells infected with dengue virus release a high proportion of immature prM-containing virions. In accordance, substantial levels of prM antibodies are found in sera of infected humans. Furthermore, it has been recently described that the rates of prM antibody responses are significantly higher in patients with secondary infection compared to those with primary infection. This suggests that immature dengue virus may play a role in disease pathogenesis. Interestingly, however, numerous functional studies have revealed that immature particles lack the ability to infect cells. In this report, we show that fully immature dengue particles become highly infectious upon interaction with prM antibodies. We demonstrate that prM antibodies facilitate efficient binding and cell entry of immature particles into Fc-receptor-expressing cells. In addition, enzymatic activity of furin is critical to render the internalized immature virus infectious. Together, these data suggest that during a secondary infection or primary infection of infants born to dengue-immune mothers, immature particles have the potential to be highly infectious and hence may contribute to the development of severe disease.

Dengue virus (DENV) is an enveloped RNA virus that causes the most common arthropod-borne infection worldwide. The mechanism by which DENV infects the host cell remains unclear. In this work, we used live-cell imaging and single-virus tracking to investigate the cell entry, endocytic trafficking, and fusion behavior of DENV. Simultaneous tracking of DENV particles and various endocytic markers revealed that DENV enters cells exclusively via clathrin-mediated endocytosis. The virus particles move along the cell surface in a diffusive manner before being captured by a pre-existing clathrin-coated pit. Upon clathrin-mediated entry, DENV particles are transported to Rab5-positive endosomes, which subsequently mature into late endosomes through acquisition of Rab7 and loss of Rab5. Fusion of the viral membrane with the endosomal membrane was primarily detected in late endosomal compartments.

Wearable electroencephalography devices emerge as a cost-effective and ergonomic alternative to gold-standard polysomnography, paving the way for better health monitoring and sleep disorder screening. Machine learning allows to automate sleep stage classification, but trust and reliability issues have hampered its adoption in clinical applications. Estimating uncertainty is a crucial factor in enhancing reliability by identifying regions of heightened and diminished confidence. In this study, we used an uncertainty-centred machine learning pipeline, U-PASS, to automate sleep staging in a challenging real-world dataset of single-channel electroencephalography and accelerometry collected with a wearable device from an elderly population. We were able to effectively limit the uncertainty of our machine learning model and to reliably inform clinical experts of which predictions were uncertain to improve the machine learning model's reliability. This increased the five-stage sleep-scoring accuracy of a state-of-the-art machine learning model from 63.9% to 71.2% on our dataset. Remarkably, the machine learning approach outperformed the human expert in interpreting these wearable data. Manual review by sleep specialists, without specific training for sleep staging on wearable electroencephalography, proved ineffective. The clinical utility of this automated remote monitoring system was also demonstrated, establishing a strong correlation between the predicted sleep parameters and the reference polysomnography parameters, and reproducing known correlations with the apnea-hypopnea index. In essence, this work presents a promising avenue to revolutionize remote patient care through the power of machine learning by the use of an automated data-processing pipeline enhanced with uncertainty estimation., (© 2024 European Sleep Research Society.)

Purpose: This brief report aims to summarize and discuss the methodologies of eXplainable Artificial Intelligence (XAI) and their potential applications in surgery., Methods: We briefly introduce explainability methods, including global and individual explanatory features, methods for imaging data and time series, as well as similarity classification, and unraveled rules and laws., Results: Given the increasing interest in artificial intelligence within the surgical field, we emphasize the critical importance of transparency and interpretability in the outputs of applied models., Conclusion: Transparency and interpretability are essential for the effective integration of AI models into clinical practice., Competing Interests: Declarations. Competing interests: The authors declare no competing interests., (© 2025. The Author(s).)

Objectives: To facilitate the stratification of patients with osteoarthritis (OA) for new treatment development and clinical trial recruitment, we created an automated machine learning (autoML) tool predicting the rapid progression of knee OA over a 2-year period., Methods: We developed autoML models integrating clinical, biochemical, X-ray and MRI data. Using two data sets within the OA Initiative-the Foundation for the National Institutes of Health OA Biomarker Consortium for training and hold-out validation, and the Pivotal Osteoarthritis Initiative MRI Analyses study for external validation-we employed two distinct definitions of clinical outcomes: Multiclass (categorising OA progression into pain and/or radiographic) and binary. Key predictors of progression were identified through advanced interpretability techniques, and subgroup analyses were conducted by age, sex and ethnicity with a focus on early-stage disease., Results: Although the most reliable models incorporated all available features, simpler models including only clinical variables achieved robust external validation performance, with area under the precision-recall curve (AUC-PRC) 0.727 (95% CI: 0.726 to 0.728) for multiclass predictions; and AUC-PRC 0.764 (95% CI: 0.762 to 0.766) for binary predictions. Multiclass models performed best in patients with early-stage OA (AUC-PRC 0.724-0.806) whereas binary models were more reliable in patients younger than 60 (AUC-PRC 0.617-0.693). Patient-reported outcomes and MRI features emerged as key predictors of progression, though subgroup differences were noted. Finally, we developed web-based applications to visualise our personalised predictions., Conclusions: Our novel tool's transparency and reliability in predicting rapid knee OA progression distinguish it from conventional 'black-box' methods and are more likely to facilitate its acceptance by clinicians and patients, enabling effective implementation in clinical practice., (Copyright © 2024 The Author(s). Published by Elsevier B.V. All rights reserved.)

Rapid advancements in medical AI necessitate targeted educational initiatives for clinicians to ensure AI tools are safe and used effectively to improve patient outcomes. To support decision-making among stakeholders in medical education, we propose three tiers of medical AI expertise and outline the challenges for medical education at different educational stages. Additionally, we offer recommendations and examples, encouraging stakeholders to adapt and shape curricula for their specific healthcare setting using this framework., Competing Interests: TS and TO have nothing to disclose. RDS is an advisor to PIPRA AG, Ceribell Inc, and B-Secur Ltd and has received an NIH/NIA research grant. PHM is a member of the Boards of Scottish Mortgage Investment Trust and the International Biotechnology Trust, and the Scientific Advisory Board of Mission Therapeutics. MvdS leads the Cambridge Centre for AI in Medicine; this centre has received research funding from AstraZeneca and GlaxoSmithKline., (© 2024 The Author(s).)

Aims: Machine learning (ML), a branch of artificial intelligence that uses algorithms to learn from data and make predictions, offers a pathway towards more personalized and tailored surgical treatments. This approach is particularly relevant to prevalent joint diseases such as osteoarthritis (OA). In contrast to end-stage disease, where joint arthroplasty provides excellent results, early stages of OA currently lack effective therapies to halt or reverse progression. Accurate prediction of OA progression is crucial if timely interventions are to be developed, to enhance patient care and optimize the design of clinical trials., Methods: A systematic review was conducted in accordance with PRISMA guidelines. We searched MEDLINE and Embase on 5 May 2024 for studies utilizing ML to predict OA progression. Titles and abstracts were independently screened, followed by full-text reviews for studies that met the eligibility criteria. Key information was extracted and synthesized for analysis, including types of data (such as clinical, radiological, or biochemical), definitions of OA progression, ML algorithms, validation methods, and outcome measures., Results: Out of 1,160 studies initially identified, 39 were included. Most studies (85%) were published between 2020 and 2024, with 82% using publicly available datasets, primarily the Osteoarthritis Initiative. ML methods were predominantly supervised, with significant variability in the definitions of OA progression: most studies focused on structural changes (59%), while fewer addressed pain progression or both. Deep learning was used in 44% of studies, while automated ML was used in 5%. There was a lack of standardization in evaluation metrics and limited external validation. Interpretability was explored in 54% of studies, primarily using SHapley Additive exPlanations., Conclusion: Our systematic review demonstrates the feasibility of ML models in predicting OA progression, but also uncovers critical limitations that currently restrict their clinical applicability. Future priorities should include diversifying data sources, standardizing outcome measures, enforcing rigorous validation, and integrating more sophisticated algorithms. This paradigm shift from predictive modelling to actionable clinical tools has the potential to transform patient care and disease management in orthopaedic practice., Competing Interests: S. Castagno is supported by the Louis and Valerie Freedman Studentship in Medical Sciences from Trinity College Cambridge, the Orthopaedic Research UK (ORUK) / Versus Arthritis: AI in MSK Research Fellowship (G124606), the Addenbrooke’s Charitable Trust (ACT) Research Advisory Committee grant (G123290), and NIHR Academic Clinical Fellowship in Trauma and Orthopaedics ((ACF-2021-14-003)). B. Gompels is supported by the Geoffrey Fisk Studentship from Darwin College Cambridge. A. McCaskie and M. Birch are supported by the NIHR Cambridge Biomedical Research Centre (BRC) (NIHR203312) and receive funding from Versus Arthritis (grant 21156) and UKRMP (grant MR/R015635/1). M. Birch is also a member of the editorial board of The Bone & Joint Journal. M. van der Schaar reports funding from AstraZeneca and GSK, related to this study., (© 2024 Castagno et al.)

Synthetic data promise privacy-preserving data sharing for healthcare research and development. Compared with other privacy-enhancing approaches-such as federated learning-analyses performed on synthetic data can be applied downstream without modification, such that synthetic data can act in place of real data for a wide range of use cases. However, the role that synthetic data might play in all aspects of clinical model development remains unknown. In this work, we used state-of-the-art generators explicitly designed for privacy preservation to create a synthetic version of ever-smokers in the UK Biobank before building prognostic models for lung cancer under several data release assumptions. We demonstrate that synthetic data can be effectively used throughout the medical prognostic modeling pipeline even without eventual access to the real data. Furthermore, we show the implications of different data release approaches on how synthetic biobank data could be deployed within the healthcare system., (© 2024. The Author(s).)

Background: Classification of perioperative risk is important for patient care, resource allocation, and guiding shared decision-making. Using discriminative features from the electronic health record (EHR), machine-learning algorithms can create digital phenotypes among heterogenous populations, representing distinct patient subpopulations grouped by shared characteristics, from which we can personalize care, anticipate clinical care trajectories, and explore therapies. We hypothesized that digital phenotypes in preoperative settings are associated with postoperative adverse events including in-hospital and 30-day mortality, 30-day surgical redo, intensive care unit (ICU) admission, and hospital length of stay (LOS)., Methods: We identified all laminectomies, colectomies, and thoracic surgeries performed over a 9-year period from a large hospital system. Seventy-seven readily extractable preoperative features were first selected from clinical consensus, including demographics, medical history, and lab results. Three surgery-specific datasets were built and split into derivation and validation cohorts using chronological occurrence. Consensus k -means clustering was performed independently on each derivation cohort, from which phenotypes' characteristics were explored. Cluster assignments were used to train a random forest model to assign patient phenotypes in validation cohorts. We reconducted descriptive analyses on validation cohorts to confirm the similarity of patient characteristics with derivation cohorts, and quantified the association of each phenotype with postoperative adverse events by using the area under receiver operating characteristic curve (AUROC). We compared our approach to American Society of Anesthesiologists (ASA) alone and investigated a combination of our phenotypes with the ASA score., Results: A total of 7251 patients met inclusion criteria, of which 2770 were held out in a validation dataset based on chronological occurrence. Using segmentation metrics and clinical consensus, 3 distinct phenotypes were created for each surgery. The main features used for segmentation included urgency of the procedure, preoperative LOS, age, and comorbidities. The most relevant characteristics varied for each of the 3 surgeries. Low-risk phenotype alpha was the most common (2039 of 2770, 74%), while high-risk phenotype gamma was the rarest (302 of 2770, 11%). Adverse outcomes progressively increased from phenotypes alpha to gamma, including 30-day mortality (0.3%, 2.1%, and 6.0%, respectively), in-hospital mortality (0.2%, 2.3%, and 7.3%), and prolonged hospital LOS (3.4%, 22.1%, and 25.8%). When combined with the ASA score, digital phenotypes achieved higher AUROC than the ASA score alone (hospital mortality: 0.91 vs 0.84; prolonged hospitalization: 0.80 vs 0.71)., Conclusions: For 3 frequently performed surgeries, we identified 3 digital phenotypes. The typical profiles of each phenotype were described and could be used to anticipate adverse postoperative events., Competing Interests: Conflicts of Interest: See Disclosures at the end of the article., (Copyright © 2023 International Anesthesia Research Society.)

Despite significant technical advances in machine learning (ML) over the past several years, the tangible impact of this technology in healthcare has been limited. This is due not only to the particular complexities of healthcare, but also due to structural issues in the machine learning for healthcare (MLHC) community which broadly reward technical novelty over tangible, equitable impact. We structure our work as a healthcare-focused echo of the 2012 paper "Machine Learning that Matters", which highlighted such structural issues in the ML community at large, and offered a series of clearly defined "Impact Challenges" to which the field should orient itself. Drawing on the expertise of a diverse and international group of authors, we engage in a narrative review and examine issues in the research background environment, training processes, evaluation metrics, and deployment protocols which act to limit the real-world applicability of MLHC. Broadly, we seek to distinguish between machine learning ON healthcare data and machine learning FOR healthcare-the former of which sees healthcare as merely a source of interesting technical challenges, and the latter of which regards ML as a tool in service of meeting tangible clinical needs. We offer specific recommendations for a series of stakeholders in the field, from ML researchers and clinicians, to the institutions in which they work, and the governments which regulate their data access., Competing Interests: L.A.C. is the Editor-in Chief of PLOS Digital Health., (Copyright: © 2024 Balagopalan et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.)

The use of data from randomized clinical trials to justify treatment decisions for real-world patients is the current state of the art. It relies on the assumption that average treatment effects from the trial can be extrapolated to patients with personal and/or disease characteristics different from those treated in the trial. Yet, because of heterogeneity of treatment effects between patients and between the trial population and real-world patients, this assumption may not be correct for many patients. Using machine learning to estimate the expected conditional average treatment effect (CATE) in individual patients from observational data offers the potential for more accurate estimation of the expected treatment effects in each patient based on their observed characteristics. In this review, we discuss some of the challenges and opportunities for machine learning to estimate CATE, including ensuring identification assumptions are met, managing covariate shift, and learning without access to the true label of interest. We also discuss the potential applications as well as future work and collaborations needed to further improve identification and utilization of CATE estimates to increase patient benefit., (© 2023 The Authors. Clinical Pharmacology & Therapeutics published by Wiley Periodicals LLC on behalf of American Society for Clinical Pharmacology and Therapeutics.)

This article deals with the discovery of causal relations from a combination of observational data and qualitative assumptions about the nature of causality in the presence of unmeasured confounding. We focus on applications where unobserved variables are known to have a widespread effect on many of the observed ones, which makes the problem particularly difficult for constraint-based methods, because most pairs of variables are conditionally dependent given any other subset, rendering the causal effect unidentifiable. In this article, we show that under the principle of independent mechanisms, unobserved confounding in this setting leaves a statistical footprint in the observed data distribution that allows for disentangling spurious and causal effects. Using this insight, we demonstrate that a sparse linear Gaussian directed acyclic graph (DAG) among observed variables may be recovered approximately and propose a simple adjusted score-based causal discovery algorithm that may be implemented with general-purpose solvers and scales to high-dimensional problems. We find, in addition, that despite the conditions we pose to guarantee causal recovery, performance in practice is robust to large deviations in model assumptions, and extensions to nonlinear structural models are possible.

Causal machine learning (ML) offers flexible, data-driven methods for predicting treatment outcomes including efficacy and toxicity, thereby supporting the assessment and safety of drugs. A key benefit of causal ML is that it allows for estimating individualized treatment effects, so that clinical decision-making can be personalized to individual patient profiles. Causal ML can be used in combination with both clinical trial data and real-world data, such as clinical registries and electronic health records, but caution is needed to avoid biased or incorrect predictions. In this Perspective, we discuss the benefits of causal ML (relative to traditional statistical or ML approaches) and outline the key components and steps. Finally, we provide recommendations for the reliable use of causal ML and effective translation into the clinic., (© 2024. Springer Nature America, Inc.)

With the increasing prevalence of machine learning in critical fields like healthcare, ensuring the safety and reliability of these systems is crucial. Estimating uncertainty plays a vital role in enhancing reliability by identifying areas of high and low confidence and reducing the risk of errors. This study introduces U-PASS, a specialized human-centered machine learning pipeline tailored for clinical applications, which effectively communicates uncertainty to clinical experts and collaborates with them to improve predictions. U-PASS incorporates uncertainty estimation at every stage of the process, including data acquisition, training, and model deployment. Training is divided into a supervised pre-training step and a semi-supervised recording-wise finetuning step. We apply U-PASS to the challenging task of sleep staging and demonstrate that it systematically improves performance at every stage. By optimizing the training dataset, actively seeking feedback from domain experts for informative samples, and deferring the most uncertain samples to experts, U-PASS achieves an impressive expert-level accuracy of 85% on a challenging clinical dataset of elderly sleep apnea patients. This represents a significant improvement over the starting point at 75% accuracy. The largest improvement gain is due to the deferral of uncertain epochs to a sleep expert. U-PASS presents a promising AI approach to incorporating uncertainty estimation in machine learning pipelines, improving their reliability and unlocking their potential in clinical settings., Competing Interests: Declaration of competing interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper., (Copyright © 2024 The Author(s). Published by Elsevier Ltd.. All rights reserved.)

Machine learning (ML)-based risk prediction models hold the potential to support the health-care setting in several ways; however, use of such models is scarce. We aimed to review health-care professional (HCP) and patient perceptions of ML risk prediction models in published literature, to inform future risk prediction model development. Following database and citation searches, we identified 41 articles suitable for inclusion. Article quality varied with qualitative studies performing strongest. Overall, perceptions of ML risk prediction models were positive. HCPs and patients considered that models have the potential to add benefit in the health-care setting. However, reservations remain; for example, concerns regarding data quality for model development and fears of unintended consequences following ML model use. We identified that public views regarding these models might be more negative than HCPs and that concerns (eg, extra demands on workload) were not always borne out in practice. Conclusions are tempered by the low number of patient and public studies, the absence of participant ethnic diversity, and variation in article quality. We identified gaps in knowledge (particularly views from under-represented groups) and optimum methods for model explanation and alerts, which require future research., Competing Interests: Declaration of interests RG is funded by the National Institute for Health and Care Research (NIHR) for this study (NIHR302604). TC is a co-founder of and owns stock in Mortimer Health and is supported by the Wellcome Trust via a Wellcome Clinical PhD Training Fellowship (REF 22890/Z/21/Z). SMJ is supported by Cancer Research UK (CRUK; EDDCPGM\100002) and Medical Research Council (MRC; MR/W025051/1) programme grants; received fees for advisory board membership in the last three years from Bard1 Lifescience; received grant income from Owlstone and GRAIL; is an unpaid member of a GRAIL advisory board; has received lecture fees for academic meetings from Cheisi and AstraZeneca; and receives support from the CRUK Lung Cancer Centre and the CRUK City of London Centre, the Rosetrees Trust, the Roy Castle Lung Cancer foundation, the Longfonds BREATH Consortia, MRC UKRMP2 Consortia, the Garfield Weston Trust, and University College London (UCL) Hospitals Charitable Foundation where this work was partly undertaken and who also received a proportion of funding from the Department of Health's NIHR Biomedical Research Centre's funding scheme. SMJ's wife works for AstraZeneca. MvdS is a Director at the Cambridge Centre for AI in Medicine, which receives funding from AstraZeneca and GSK. JS is funded by NIHR Applied Research Collaboration North Thames, NIHR (policy research programme, programme grant, health service and delivery grant, and capacity building funding), and the UK Prevention Research Partnership. NN is supported by an MRC Clinical Academic Research Partnership (MR/T02481X/1); received a grant from the CRUK early diagnosis grant for the Real-time cancer analytics (REACT) study (EICEDAAP\100012); reports honoraria for non-promotional educational talks or advisory boards from Amgen, AstraZeneca, Boehringer ingelheim, Bristol Myers Squibb, EQRx, Fujifilm, Guardant Health, Intuitive, Janssen, Lilly, MSD, Olympus, OncLive, PeerVoice, Pfizer, Roche, and Takeda outside of the current work; holds positions as the Director of UK Lung cancer coalition, member of the Steering Committee of British Thoracic Oncology Group, and Clinical Director of National Lung Cancer Audit; and reports stock and stock options in OneWelbeck, The Physicians’ Clinic (HCA), and Pharmacierge. AJ declares no competing interests. The views expressed in this publication are those of the authors and not necessarily those of the NIHR, National Health Service, or the UK Department of Health and Social Care., (Copyright © 2024 The Author(s). Published by Elsevier Ltd. This is an Open Access article under the CC BY 4.0 license. Published by Elsevier Ltd.. All rights reserved.)

The sea anemone Calliactis parasitica, which is found in the East Atlantic (Portugal to Senegal) and the Mediterranean Sea, forms a symbiotic relationship with the red hermit crab, Dardanus calidus, in which the anemone provides protection from predators such as the octopus while it gains mobility, and possibly food scraps, from the hermit crab. Acoustic pollution is recognised by the scientific community as a growing threat to ocean inhabitants. Recent findings on marine invertebrates showed that exposure to artificial sound had direct behavioural, physiological and ultrastructural consequences. In this study we assess the impact of artificial sound (received level 157 ± 5 dB re 1 μPa 2 with peak levels up to 175 dB re 1 μPa 2 ) on the red hermit crab and its symbiotic sea anemone. Scanning electron microscopy analyses revealed lesions in the statocyst of the red hermit crab and in the tentacle sensory epithelia of its anemone when exposed to low-intensity, low-frequency sounds. These ultrastructural changes under situations of acoustic stress in symbiotic partners belonging to different phyla is a new issue that may limit their survival capacity, and a new challenge in assessing the effects of acoustic disturbance in the oceanic ecosystem. Despite the lesions found in the red hermit crab, its righting reflex time was not as strongly affected showing only an increase in the range of righting times. Given that low-frequency sound levels in the ocean are increasing and that reliable bioacoustic data on invertebrates is very scarce, in light of the results of the present study, we argue that anthropogenic sound effects on invertebrates species may have direct consequences in the entire ecosystem., Competing Interests: Declaration of competing interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper., (Copyright © 2023 The Authors. Published by Elsevier B.V. All rights reserved.)

Precision medicine requires individualized modeling of disease and drug dynamics, with machine learning-based computational techniques gaining increasing popularity. The complexity of either field, however, makes current pharmacological problems opaque to machine learning practitioners, and state-of-the-art machine learning methods inaccessible to pharmacometricians. To help bridge the two worlds, we provide an introduction to current problems and techniques in pharmacometrics that ranges from pharmacokinetic and pharmacodynamic modeling to pharmacometric simulations, model-informed precision dosing, and systems pharmacology, and review some of the machine learning approaches to address them. We hope this would facilitate collaboration between experts, with complementary strengths of principled pharmacometric modeling and flexibility of machine learning leading to synergistic effects in pharmacological applications., (© 2023. The Author(s), under exclusive licence to Springer Nature Switzerland AG.)

Background: Risk-based screening for lung cancer is currently being considered in several countries; however, the optimal approach to determine eligibility remains unclear. Ensemble machine learning could support the development of highly parsimonious prediction models that maintain the performance of more complex models while maximising simplicity and generalisability, supporting the widespread adoption of personalised screening. In this work, we aimed to develop and validate ensemble machine learning models to determine eligibility for risk-based lung cancer screening., Methods and Findings: For model development, we used data from 216,714 ever-smokers recruited between 2006 and 2010 to the UK Biobank prospective cohort and 26,616 high-risk ever-smokers recruited between 2002 and 2004 to the control arm of the US National Lung Screening (NLST) randomised controlled trial. The NLST trial randomised high-risk smokers from 33 US centres with at least a 30 pack-year smoking history and fewer than 15 quit-years to annual CT or chest radiography screening for lung cancer. We externally validated our models among 49,593 participants in the chest radiography arm and all 80,659 ever-smoking participants in the US Prostate, Lung, Colorectal and Ovarian (PLCO) Screening Trial. The PLCO trial, recruiting from 1993 to 2001, analysed the impact of chest radiography or no chest radiography for lung cancer screening. We primarily validated in the PLCO chest radiography arm such that we could benchmark against comparator models developed within the PLCO control arm. Models were developed to predict the risk of 2 outcomes within 5 years from baseline: diagnosis of lung cancer and death from lung cancer. We assessed model discrimination (area under the receiver operating curve, AUC), calibration (calibration curves and expected/observed ratio), overall performance (Brier scores), and net benefit with decision curve analysis. Models predicting lung cancer death (UCL-D) and incidence (UCL-I) using 3 variables-age, smoking duration, and pack-years-achieved or exceeded parity in discrimination, overall performance, and net benefit with comparators currently in use, despite requiring only one-quarter of the predictors. In external validation in the PLCO trial, UCL-D had an AUC of 0.803 (95% CI: 0.783, 0.824) and was well calibrated with an expected/observed (E/O) ratio of 1.05 (95% CI: 0.95, 1.19). UCL-I had an AUC of 0.787 (95% CI: 0.771, 0.802), an E/O ratio of 1.0 (95% CI: 0.92, 1.07). The sensitivity of UCL-D was 85.5% and UCL-I was 83.9%, at 5-year risk thresholds of 0.68% and 1.17%, respectively, 7.9% and 6.2% higher than the USPSTF-2021 criteria at the same specificity. The main limitation of this study is that the models have not been validated outside of UK and US cohorts., Conclusions: We present parsimonious ensemble machine learning models to predict the risk of lung cancer in ever-smokers, demonstrating a novel approach that could simplify the implementation of risk-based lung cancer screening in multiple settings., Competing Interests: I have read the journal’s policy and the authors of this manuscript have the following competing interests: NN reports honoraria for non-promotional educational talks, conference support or advisory boards from Amgen, Astra Zeneca, Boehringer Ingelheim, Bristol Myers Squibb, Guardant Health, Janssen, Lilly, Merck Sharp & Dohme, Olympus, OncLive, PeerVoice, Pfizer, and Takeda. SMJ has received fees for advisory board membership in the last three years from Astra-Zeneca, Bard1 Lifescience, and Johnson and Johnson. He has received grant income from Owlstone and GRAIL Inc. He has received assistance with travel to an academic meeting from Cheisi. TC and SMJ are founders of, and own stock in, Mortimer Health., (Copyright: © 2023 Callender et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.)

In high-stakes applications of data-driven decision-making such as healthcare, it is of paramount importance to learn a policy that maximizes the reward while avoiding potentially dangerous actions when there is uncertainty. There are two main challenges usually associated with this problem. First, learning through online exploration is not possible due to the critical nature of such applications. Therefore, we need to resort to observational datasets with no counterfactuals. Second, such datasets are usually imperfect, additionally cursed with missing values in the attributes of features. In this article, we consider the problem of constructing personalized policies using logged data when there are missing values in the attributes of features in both training and test data. The goal is to recommend an action (treatment) when ~ X , a degraded version of X with missing values, is observed. We consider three strategies for dealing with missingness. In particular, we introduce the conservative strategy where the policy is designed to safely handle the uncertainty due to missingness. In order to implement this strategy, we need to estimate posterior distribution p(X| ~ X) and use a variational autoencoder to achieve this. In particular, our method is based on partial variational autoencoders (PVAEs) that are designed to capture the underlying structure of features with missing values.

Using passive acoustic monitoring (PAM) and convolutional neural networks (CNN), we monitored the movements of the two endangered Amazon River dolphin species, the boto (Inia geoffrensis) and the tucuxi (Sotalia fluviatilis) from main rivers to floodplain habitats (várzea) in the Mamirauá Reserve (Amazonas, Brazil). We detected dolphin presence in four main areas based on the classification of their echolocation clicks. Using the same method, we automatically detected boat passages to estimate a possible interaction between boat and dolphin presence. Performance of the CNN classifier was high with an average precision of 0.95 and 0.92 for echolocation clicks and boats, respectively. Peaks of acoustic activity were detected synchronously at the river entrance and channel, corresponding to dolphins seasonally entering the várzea. Additionally, the river dolphins were regularly detected inside the flooded forest, suggesting a wide dispersion of their populations inside this large area, traditionally understudied and particularly important for boto females and calves. Boats overlapped with dolphin presence 9% of the time. PAM and recent advances in classification methods bring a new insight of the river dolphins' use of várzea habitats, which will contribute to conservation strategies of these species., (© 2023. The Author(s).)

Background Empiric antimicrobial therapy with azithromycin is highly used in patients admitted to the hospital with COVID-19, despite prior research suggesting that azithromycin may be associated with increased risk of cardiovascular events. Methods and Results This study was conducted using data from the ISACS-COVID-19 (International Survey of Acute Coronavirus Syndromes-COVID-19) registry. Patients with a confirmed diagnosis of SARS-CoV-2 infection were eligible for inclusion. The study included 793 patients exposed to azithromycin within 24 hours from hospital admission and 2141 patients who received only standard care. The primary exposure was cardiovascular disease (CVD). Main outcome measures were 30-day mortality and acute heart failure (AHF). Among 2934 patients, 1066 (36.4%) had preexisting CVD. A total of 617 (21.0%) died, and 253 (8.6%) had AHF. Azithromycin therapy was consistently associated with an increased risk of AHF in patients with preexisting CVD (risk ratio [RR], 1.48 [95% CI, 1.06-2.06]). Receiving azithromycin versus standard care was not significantly associated with death (RR, 0.94 [95% CI, 0.69-1.28]). By contrast, we found significantly reduced odds of death (RR, 0.57 [95% CI, 0.42-0.79]) and no significant increase in AHF (RR, 1.23 [95% CI, 0.75-2.04]) in patients without prior CVD. The relative risks of death from the 2 subgroups were significantly different from each other ( P interaction =0.01). Statistically significant association was observed between AHF and death (odds ratio, 2.28 [95% CI, 1.34-3.90]). Conclusions These findings suggest that azithromycin use in patients with COVID-19 and prior history of CVD is significantly associated with an increased risk of AHF and all-cause 30-day mortality. Registration URL: https://www.clinicaltrials.gov; Unique identifier: NCT05188612.