RESUMO
Artificial intelligence (AI) has been developed for echocardiography1-3, although it has not yet been tested with blinding and randomization. Here we designed a blinded, randomized non-inferiority clinical trial (ClinicalTrials.gov ID: NCT05140642; no outside funding) of AI versus sonographer initial assessment of left ventricular ejection fraction (LVEF) to evaluate the impact of AI in the interpretation workflow. The primary end point was the change in the LVEF between initial AI or sonographer assessment and final cardiologist assessment, evaluated by the proportion of studies with substantial change (more than 5% change). From 3,769 echocardiographic studies screened, 274 studies were excluded owing to poor image quality. The proportion of studies substantially changed was 16.8% in the AI group and 27.2% in the sonographer group (difference of -10.4%, 95% confidence interval: -13.2% to -7.7%, P < 0.001 for non-inferiority, P < 0.001 for superiority). The mean absolute difference between final cardiologist assessment and independent previous cardiologist assessment was 6.29% in the AI group and 7.23% in the sonographer group (difference of -0.96%, 95% confidence interval: -1.34% to -0.54%, P < 0.001 for superiority). The AI-guided workflow saved time for both sonographers and cardiologists, and cardiologists were not able to distinguish between the initial assessments by AI versus the sonographer (blinding index of 0.088). For patients undergoing echocardiographic quantification of cardiac function, initial assessment of LVEF by AI was non-inferior to assessment by sonographers.
Assuntos
Inteligência Artificial , Cardiologistas , Ecocardiografia , Testes de Função Cardíaca , Humanos , Inteligência Artificial/normas , Ecocardiografia/métodos , Ecocardiografia/normas , Volume Sistólico , Função Ventricular Esquerda , Método Simples-Cego , Fluxo de Trabalho , Reprodutibilidade dos Testes , Testes de Função Cardíaca/métodos , Testes de Função Cardíaca/normasRESUMO
Polynyas, areas of open water embedded within sea ice, are a key component of ocean-atmosphere interactions that act as hotspots of sea-ice production, bottom-water formation, and primary productivity. The specific drivers of polynya dynamics remain, however, elusive and coupled climate models struggle to replicate Antarctic polynya activity. Here, we leverage a 44-y time series of Antarctic sea ice to elucidate long-term trends. We identify Antarctic-wide linear increases and a hitherto undescribed cyclical pattern of polynya activity across the Ross Sea region that potentially arises from interactions between the Amundsen Sea Low and Southern Annular Mode. While their specific drivers remain unknown, identifying these emerging patterns augments our capacity to understand the processes that influence sea ice. As we enter a potentially new age of Antarctic sea ice, this advance in understanding will, in turn, lead to more accurate predictions of environmental change, and its implications for Antarctic ecosystems.
RESUMO
BACKGROUND: Diagnosis of mitral regurgitation (MR) requires careful evaluation by echocardiography with Doppler imaging. This study presents the development and validation of a fully automated deep learning pipeline for identifying apical 4-chamber view videos with color Doppler echocardiography and detecting clinically significant (moderate or severe) MR from transthoracic echocardiograms. METHODS: A total of 58 614 transthoracic echocardiograms (2 587 538 videos) from Cedars-Sinai Medical Center were used to develop and test an automated pipeline to identify apical 4-chamber view videos with color Doppler across the mitral valve and then assess MR severity. The model was tested internally on a test set of 1800 studies (80 833 videos) from Cedars-Sinai Medical Center and externally evaluated in a geographically distinct cohort of 915 studies (46 890 videos) from Stanford Healthcare. RESULTS: In the held-out Cedars-Sinai Medical Center test set, the view classifier demonstrated an area under the curve (AUC) of 0.998 (0.998-0.999) and correctly identified 3452 of 3539 echocardiography videos as having color Doppler information across the mitral valve (sensitivity of 0.975 [0.968-0.982] and specificity of 0.999 [0.999-0.999] compared with manually curated videos). In the external test cohort from Stanford Healthcare, the view classifier correctly identified 1051 of 1055 manually curated videos with color Doppler information across the mitral valve (sensitivity of 0.996 [0.990-1.000] and specificity of 0.999 [0.999-0.999]). In the Cedars-Sinai Medical Center test cohort, MR moderate or greater in severity was detected with an AUC of 0.916 (0.899-0.932) and severe MR was detected with an AUC of 0.934 (0.913-0.953). In the Stanford Healthcare test cohort, the model detected MR moderate or greater in severity with an AUC of 0.951 (0.924-0.973) and severe MR with an AUC of 0.969 (0.946-0.987). CONCLUSIONS: In this study, a novel automated pipeline for identifying clinically significant MR from full transthoracic echocardiography studies demonstrated excellent performance across large numbers of studies and across multiple institutions. Such an approach has the potential for automated screening and surveillance of MR.
RESUMO
BACKGROUND AND AIMS: Contrasting patterns of host and microbiome biogeography can provide insight into the drivers of microbial community assembly. Distance-decay relationships are a classic biogeographical pattern shaped by interactions between selective and non-selective processes. Joint biogeography of microbiomes and their hosts is of increasing interest owing to the potential for microbiome-facilitated adaptation. METHODS: In this study, we examine the coupled biogeography of the model macroalga Durvillaea and its microbiome using a combination of genotyping by sequencing (host) and 16S rRNA amplicon sequencing (microbiome). Alongside these approaches, we use environmental data to characterize the relationship between the microbiome, the host, and the environment. KEY RESULTS: We show that although the host and microbiome exhibit shared biogeographical structure, these arise from different processes, with host biogeography showing classic signs of geographical distance decay, but with the microbiome showing environmental distance decay. Examination of microbial subcommunities, defined by abundance, revealed that the abundance of microbes is linked to environmental selection. As microbes become less common, the dominant ecological processes shift away from selective processes and towards neutral processes. Contrary to expectations, we found that ecological drift does not promote structuring of the microbiome. CONCLUSIONS: Our results suggest that although host macroalgae exhibit a relatively 'typical' biogeographical pattern of declining similarity with increasing geographical distance, the microbiome is more variable and is shaped primarily by environmental conditions. Our findings suggest that the Baas Becking hypothesis of 'everything is everywhere, the environment selects' might be a useful hypothesis to understand the biogeography of macroalgal microbiomes. As environmental conditions change in response to anthropogenic influences, the processes structuring the microbiome of macroalgae might shift, whereas those governing the host biogeography are less likely to change. As a result, increasingly decoupled host-microbe biogeography might be observed in response to such human influences.
Assuntos
Microbiota , Humanos , RNA Ribossômico 16S/genética , GeografiaRESUMO
The structure of tubular transport networks is thought to underlie much of biological regularity, from individuals to ecosystems. A core assumption of transport network models is either area-preserving or area-increasing branching, such that the summed cross-sectional area of all child branches is equal to or greater than the cross-sectional area of their respective parent branch. For insects, the most diverse group of animals, the assumption of area-preserving branching of tracheae is, however, based on measurements of a single individual and an assumption of gas exchange by diffusion. Here we show that ants exhibit neither area-preserving nor area-increasing branching in their abdominal tracheal systems. We find for 20 species of ants that the sum of child tracheal cross-sectional areas is typically less than that of the parent branch (area-decreasing). The radius, rather than the area, of the parent branch is conserved across the sum of child branches. Interpretation of the tracheal system as one optimized for the release of carbon dioxide, while readily catering to oxygen demand, explains the branching pattern. Our results, together with widespread demonstration that gas exchange in insects includes, and is often dominated by, convection, indicate that for generality, network transport models must include consideration of systems with different architectures.
Assuntos
Formigas/fisiologia , Transporte Biológico/fisiologia , Biologia Computacional/métodos , Modelos Biológicos , Traqueia/fisiologia , Animais , Dióxido de Carbono/metabolismo , Oxigênio/metabolismoRESUMO
Soil systems are being increasingly exposed to the interactive effects of biological invasions and climate change, with rising temperatures expected to benefit alien over indigenous species. We assessed this expectation for an important soil-dwelling group, the springtails, by determining whether alien species show broader thermal tolerance limits and greater tolerance to climate warming than their indigenous counterparts. We found that, from the tropics to the sub-Antarctic, alien species have the broadest thermal tolerances and greatest tolerance to environmental warming. Both groups of species show little phenotypic plasticity or potential for evolutionary change in tolerance to high temperature. These trait differences between alien and indigenous species suggest that biological invasions will exacerbate the impacts of climate change on soil systems, with profound implications for terrestrial ecosystem functioning.
Assuntos
Aclimatação/fisiologia , Ecossistema , Aquecimento Global , Espécies Introduzidas , Modelos Biológicos , Fenômenos Fisiológicos Vegetais , Plantas , AustráliaRESUMO
Thermal traits are frequently used to explain variation in species distributions, abundance, and sensitivity to climate change. Due to their utility and ease of measurement, critical thermal limits in particular have proliferated across the ecophysiological literature. Critical limit assays can, however, have deleterious or even lethal effects on individuals and there is growing recognition that intermediate metrics of performance can provide a further, nuanced understanding of how species interact with their environments. Meanwhile, the scarcity of data describing sub-critical or voluntary limits, which have been proposed as alternatives to critical limits and can be collected under less extreme conditions, reduces their value in comparative analyses and broad-scale syntheses. To overcome these limitations and determine if sub-critical limits are viable proxies for upper and lower critical thermal limits we measured and compared the critical and sub-critical thermal limits of 2023 ants representing 51 species. Sub-critical limits in isolation were a satisfactory linear predictor for both individual and species critical limits and when species identity was also considered there were substantial gains in variance explained. These gains indicate that a species-specific conversion factor can further improve estimates of critical traits using sub-critical proxies. Sub-critical limits can, therefore, be integrated into broader syntheses of critical limits and confidently used to calculate common ecological metrics, such as warming tolerance, so long as uncertainty in estimates is explicitly acknowledged. Although lower thermal traits exhibited more variation than their upper counterparts, the stronger phylogenetic signal of lower thermal traits indicates that appropriate conversions for lower thermal traits can be inferred from congenerics or other closely related taxa.
Assuntos
Aclimatação , Formigas/fisiologia , Animais , Formigas/genética , Filogenia , Especificidade da Espécie , TemperaturaRESUMO
Thermal performance traits are regularly used to make forecasts of the responses of ectotherms to anthropogenic environmental change, but such forecasts do not always differentiate between fundamental and realised thermal niches. Here we determine the relative extents to which variation in the fundamental and realised thermal niches accounts for current variation in species abundance and occupancy and assess the effects of niche-choice on future-climate response estimations. We investigated microclimate and macroclimate temperatures alongside abundance, occupancy, critical thermal limits and foraging activity of 52 ant species (accounting for >95% individuals collected) from a regional assemblage from across the Western Cape Province, South Africa, between 2003 and 2014. Capability of a species to occupy sites experiencing the most extreme temperatures, coupled with breadth of realised niche, explained most deviance in occupancy (up to 75%), while foraging temperature range and body mass explained up to 50.5% of observed variation in mean species abundance. When realised niches are used to forecast responses to climate change, large positive and negative effects among species are predicted under future conditions, in contrast to the forecasts of minimal impacts on all species that are indicated by fundamental niche predictions.
Assuntos
Mudança Climática , Ecossistema , Animais , Temperatura Alta , África do Sul , TemperaturaRESUMO
For over a century, the hypothesis of temperature compensation, the maintenance of similar biological rates in species from different thermal environments, has remained controversial. An alternative idea, that fitness is greater at higher temperatures (the thermodynamic effect), has gained increasing traction. This alternative hypothesis is also being used to understand large-scale biodiversity responses to environmental change. Yet evidence in favour of each of these contrasting hypotheses continues to emerge. In consequence, the fundamental nature of organismal thermal responses and its implications remain unresolved. Here, we investigate these ideas explicitly using a global dataset of 619 observations of four categories of organismal performance, spanning 14 phyla and 403 species. In agreement with both hypotheses, we show a positive relationship between the temperature of maximal performance rate (Topt) and environmental temperature (Tenv) for developmental rate and locomotion speed, but not growth or photosynthesis rate. Next, we demonstrate that relationships between Tenv and the maximal performance rate (Umax) are rarely significant and positive, as expected if a thermodynamic effect predominates. By contrast, a positive relationship between Topt and Umax is always present, but markedly weaker than theoretically predicted. These outcomes demonstrate that while some form of thermodynamic effect exists, ample scope is present for biochemical and physiological adaptation to thermal environments in the form of temperature compensation.
Assuntos
Adaptação Fisiológica , Temperatura , Termodinâmica , Adaptação Fisiológica/genética , Adaptação Fisiológica/fisiologia , Animais , Crescimento/fisiologia , Locomoção/fisiologia , Fotossíntese/fisiologia , Filogenia , Fenômenos Fisiológicos Vegetais , PlantasRESUMO
Advancements in medical imaging and artificial intelligence (AI) have revolutionized the field of cardiac diagnostics, providing accurate and efficient tools for assessing cardiac function. AI diagnostics claims to improve upon the human-to-human variation that is known to be significant. However, when put in practice, for cardiac ultrasound, AI models are being run on images acquired by human sonographers whose quality and consistency may vary. With more variation than other medical imaging modalities, variation in image acquisition may lead to out-of-distribution (OOD) data and unpredictable performance of the AI tools. Recent advances in ultrasound technology has allowed the acquisition of both 3D as well as 2D data, however 3D has more limited temporal and spatial resolution and is still not routinely acquired. Because the training datasets used when developing AI algorithms are mostly developed using 2D images, it is difficult to determine the impact of human variation on the performance of AI tools in the real world. The objective of this project is to leverage 3D echos to simulate realistic human variation of image acquisition and better understand the OOD performance of a previously validated AI model. In doing so, we develop tools for interpreting 3D echo data and quantifiably recreating common variation in image acquisition between sonographers. We also developed a technique for finding good standard 2D views in 3D echo volumes. We found the performance of the AI model we evaluated to be as expected when the view is good, but variations in acquisition position degraded AI model performance. Performance on far from ideal views was poor, but still better than random, suggesting that there is some information being used that permeates the whole volume, not just a quality view. Additionally, we found that variations in foreshortening didn't result in the same errors that a human would make.
Assuntos
Inteligência Artificial , Biologia Computacional , Humanos , AlgoritmosRESUMO
Climate change is altering environmental conditions, with microclimates providing small-scale refuges within otherwise challenging environments. Durvillaea (southern bull kelp; rimurapa) is a genus of large intertidal fucoid algae, and some species harbour diverse invertebrate communities in their holdfasts. We hypothesised that animal-excavated Durvillaea holdfasts provide a thermal refuge for epibiont species, and tested this hypothesis using the exemplar species D. poha. Using a southern Aotearoa New Zealand population as a case-study, we found extreme temperatures outside the holdfast were 4.4 °C higher in summer and 6.9 °C lower in winter than inside the holdfast. A microclimate model of the holdfasts was built and used to forecast microclimates under 2100 conditions. Temperatures are predicted to increase by 2-3 °C, which may exceed the tolerances of D. poha. However, if D. poha or a similar congeneric persists, temperatures inside holdfasts will remain less extreme than the external environment. The thermal tolerances of two Durvillaea-associated invertebrates, the trochid gastropod Cantharidus antipodum and the amphipod Parawaldeckia kidderi, were also assessed; C. antipodum, but not P. kidderi, displayed metabolic depression at temperatures above and below those inside holdfasts, suggesting that they would be vulnerable outside the holdfast and with future warming. Microclimates, such as those within D. poha holdfasts or holdfasts of similar species, will therefore be important refuges for the survival of species both at the northern (retreating edge) and southern (expanding edge) limits of their distributions.
Assuntos
Mudança Climática , Invertebrados , Kelp , Microclima , Animais , Kelp/fisiologia , Nova Zelândia , Invertebrados/fisiologia , Temperatura , Anfípodes/fisiologiaRESUMO
Host-associated microbial communities are shaped by host migratory movements. These movements can have contrasting impacts on microbiota, and understanding such patterns can provide insight into the ecological processes that contribute to community diversity. Furthermore, long-distance movements to new environments are anticipated to occur with increasing frequency due to host distribution shifts resulting from climate change. Understanding how hosts transport their microbiota with them could be of importance when examining biological invasions. Although microbial community shifts are well-documented, the underlying mechanisms that lead to the restructuring of these communities remain relatively unexplored. Using literature and ecological simulations, we develop a framework to elucidate the major factors that lead to community change. We group host movements into two types-regular (repeated/cyclical migratory movements, as found in many birds and mammals) and irregular (stochastic/infrequent movements that do not occur on a cyclical basis, as found in many insects and plants). Ecological simulations and prior research suggest that movement type and frequency, alongside environmental exposure (e.g. internal/external microbiota) are key considerations for understanding movement-associated community changes. From our framework, we derive a series of testable hypotheses, and suggest means to test them, to facilitate future research into host movement and microbial community dynamics.
Assuntos
Microbiota , Animais , Migração Animal , Biodiversidade , Aves/microbiologia , Mudança Climática , Interações entre Hospedeiro e Microrganismos , Mamíferos/microbiologiaRESUMO
Background: Echocardiography is the most common modality for assessing cardiac structure and function. While cardiac magnetic resonance (CMR) imaging is less accessible, CMR can provide unique tissue characterization including late gadolinium enhancement (LGE), T1 and T2 mapping, and extracellular volume (ECV) which are associated with tissue fibrosis, infiltration, and inflammation. While deep learning has been shown to uncover findings not recognized by clinicians, it is unknown whether CMR-based tissue characteristics can be derived from echocardiography videos using deep learning. We hypothesized that deep learning applied to echocardiography could predict CMR-based measurements. Methods: In a retrospective single-center study, adult patients with CMRs and echocardiography studies within 30 days were included. A video-based convolutional neural network was trained on echocardiography videos to predict CMR-derived labels including wall motion abnormality (WMA) presence, LGE presence, and abnormal T1, T2 or ECV across echocardiography views. The model performance was evaluated in a held-out test dataset not used for training. Results: The study population included 1,453 adult patients (mean age 56±18 years, 42% female) with 2,556 paired echocardiography studies occurring on average 2 days after CMR (interquartile range 2 days prior to 6 days after). The model had high predictive capability for presence of WMA (AUC 0.873 [95%CI 0.816-0.922]), however, the model was unable to reliably detect the presence of LGE (AUC 0.699 [0.613-0.780]), native T1 (AUC 0.614 [0.500-0.715]), T2 0.553 [0.420-0.692], or ECV 0.564 [0.455-0.691]). Conclusions: Deep learning applied to echocardiography accurately identified CMR-based WMA, but was unable to predict tissue characteristics, suggesting that signal for these tissue characteristics may not be present within ultrasound videos, and that the use of CMR for tissue characterization remains essential within cardiology. Clinical Perspective: Tissue characterization of the heart muscle is useful for clinical diagnosis and prognosis by identifying myocardial fibrosis, inflammation, and infiltration, and can be measured using cardiac MRI. While echocardiography is highly accessible and provides excellent functional information, its ability to provide tissue characterization information is limited at this time. Our study using a deep learning approach to predict cardiac MRI-based tissue characteristics from echocardiography showed limited ability to do so, suggesting that alternative approaches, including non-deep learning methods should be considered in future research.
RESUMO
Passive samplers are enabling the scaling of environmental DNA (eDNA) biomonitoring in our oceans, by circumventing the time-consuming process of water filtration. Designing a novel passive sampler that does not require extensive sample handling time and can be connected to ocean-going vessels without impeding normal underway activities has potential to rapidly upscale global biomonitoring efforts onboard the world's oceanic fleet. Here, we demonstrate the utility of an artificial sponge sampler connected to the continuous pump underway seawater system as a means to enable oceanic biomonitoring. We compared the performance of this passive sampling protocol with standard water filtration at six locations during a research voyage from New Zealand to Antarctica in early 2023. Eukaryote metabarcoding of the mitochondrial COI gene revealed no significant difference in phylogenetic α-diversity between sampling methods and both methods delineated a progressive reduction in number of Zero-Radius Operational Taxonomic Units (ZOTUs) with increased latitudes. While both sampling methods revealed comparable trends in geographical community compositions, distinct clusters were identified for passive samplers and water filtration at each location. Additionally, greater variability between replicates was observed for passive samplers, resulting in an increased estimated level of replication needed to recover 90 % of the biodiversity. Furthermore, traditional water filtration failed to detect three phyla observed by passive samplers and extrapolation analysis estimated passive samplers recover a larger number of ZOTUs compared to water filtration for all six locations. Our results demonstrate the potential of this passive eDNA sampler protocol and highlight areas where this emerging technology could be improved, thereby enabling large-scale offshore marine eDNA biomonitoring by leveraging the world's oceanic fleet without interfering with onboard activities.
Assuntos
Monitoramento Biológico , DNA Ambiental , Monitoramento Ambiental , Água do Mar , Monitoramento Ambiental/métodos , Monitoramento Ambiental/instrumentação , Monitoramento Biológico/métodos , DNA Ambiental/análise , Nova Zelândia , Biodiversidade , Oceanos e MaresRESUMO
Marine sponges have recently emerged as efficient natural environmental DNA (eDNA) samplers. The ability of sponges to accumulate eDNA provides an exciting opportunity to reconstruct contemporary communities and ecosystems with high temporal and spatial precision. However, the use of historical eDNA, trapped within the vast number of specimens stored in scientific collections, opens up the opportunity to begin to reconstruct the communities and ecosystems of the past. Here, we define the term 'heDNA' to denote the historical environmental DNA that can be obtained from the recent past with high spatial and temporal accuracy. Using a variety of Antarctic sponge specimens stored in an extensive marine invertebrate collection, we were able to recover information on Antarctic fish biodiversity from specimens up to 20 years old. We successfully recovered 64 fish heDNA signals from 27 sponge specimens. Alpha diversity measures did not differ among preservation methods, but sponges stored frozen had a significantly different fish community composition compared to those stored dry or in ethanol. Our results show that we were consistently and reliably able to extract the heDNA trapped within marine sponge specimens, thereby enabling the reconstruction and investigation of communities and ecosystems of the recent past with a spatial and temporal resolution previously unattainable. Future research into heDNA extraction from other preservation methods, as well as the impact of specimen age and collection method, will strengthen and expand the opportunities for this novel resource to access new knowledge on ecological change during the last century.
RESUMO
Atrial fibrillation (AF) often escapes detection, given its frequent paroxysmal and asymptomatic presentation. Deep learning of transthoracic echocardiograms (TTEs), which have structural information, could help identify occult AF. We created a two-stage deep learning algorithm using a video-based convolutional neural network model that (1) distinguished whether TTEs were in sinus rhythm or AF and then (2) predicted which of the TTEs in sinus rhythm were in patients who had experienced AF within 90 days. Our model, trained on 111,319 TTE videos, distinguished TTEs in AF from those in sinus rhythm with high accuracy in a held-out test cohort (AUC 0.96 (0.95-0.96), AUPRC 0.91 (0.90-0.92)). Among TTEs in sinus rhythm, the model predicted the presence of concurrent paroxysmal AF (AUC 0.74 (0.71-0.77), AUPRC 0.19 (0.16-0.23)). Model discrimination remained similar in an external cohort of 10,203 TTEs (AUC of 0.69 (0.67-0.70), AUPRC 0.34 (0.31-0.36)). Performance held across patients who were women (AUC 0.76 (0.72-0.81)), older than 65 years (0.73 (0.69-0.76)), or had a CHA2DS2VASc ≥2 (0.73 (0.79-0.77)). The model performed better than using clinical risk factors (AUC 0.64 (0.62-0.67)), TTE measurements (0.64 (0.62-0.67)), left atrial size (0.63 (0.62-0.64)), or CHA2DS2VASc (0.61 (0.60-0.62)). An ensemble model in a cohort subset combining the TTE model with an electrocardiogram (ECGs) deep learning model performed better than using the ECG model alone (AUC 0.81 vs. 0.79, p = 0.01). Deep learning using TTEs can predict patients with active or occult AF and could be used for opportunistic AF screening that could lead to earlier treatment.
RESUMO
BACKGROUND: Preoperative risk assessments used in clinical practice are insufficient in their ability to identify risk for postoperative mortality. Deep-learning analysis of electrocardiography can identify hidden risk markers that can help to prognosticate postoperative mortality. We aimed to develop a prognostic model that accurately predicts postoperative mortality in patients undergoing medical procedures and who had received preoperative electrocardiographic diagnostic testing. METHODS: In a derivation cohort of preoperative patients with available electrocardiograms (ECGs) from Cedars-Sinai Medical Center (Los Angeles, CA, USA) between Jan 1, 2015 and Dec 31, 2019, a deep-learning algorithm was developed to leverage waveform signals to discriminate postoperative mortality. We randomly split patients (8:1:1) into subsets for training, internal validation, and final algorithm test analyses. Model performance was assessed using area under the receiver operating characteristic curve (AUC) values in the hold-out test dataset and in two external hospital cohorts and compared with the established Revised Cardiac Risk Index (RCRI) score. The primary outcome was post-procedural mortality across three health-care systems. FINDINGS: 45 969 patients had a complete ECG waveform image available for at least one 12-lead ECG performed within the 30 days before the procedure date (59 975 inpatient procedures and 112 794 ECGs): 36 839 patients in the training dataset, 4549 in the internal validation dataset, and 4581 in the internal test dataset. In the held-out internal test cohort, the algorithm discriminates mortality with an AUC value of 0·83 (95% CI 0·79-0·87), surpassing the discrimination of the RCRI score with an AUC of 0·67 (0·61-0·72). The algorithm similarly discriminated risk for mortality in two independent US health-care systems, with AUCs of 0·79 (0·75-0·83) and 0·75 (0·74-0·76), respectively. Patients determined to be high risk by the deep-learning model had an unadjusted odds ratio (OR) of 8·83 (5·57-13·20) for postoperative mortality compared with an unadjusted OR of 2·08 (0·77-3·50) for postoperative mortality for RCRI scores of more than 2. The deep-learning algorithm performed similarly for patients undergoing cardiac surgery (AUC 0·85 [0·77-0·92]), non-cardiac surgery (AUC 0·83 [0·79-0·88]), and catheterisation or endoscopy suite procedures (AUC 0·76 [0·72-0·81]). INTERPRETATION: A deep-learning algorithm interpreting preoperative ECGs can improve discrimination of postoperative mortality. The deep-learning algorithm worked equally well for risk stratification of cardiac surgeries, non-cardiac surgeries, and catheterisation laboratory procedures, and was validated in three independent health-care systems. This algorithm can provide additional information to clinicians making the decision to perform medical procedures and stratify the risk of future complications. FUNDING: National Heart, Lung, and Blood Institute.
Assuntos
Aprendizado Profundo , Humanos , Medição de Risco/métodos , Algoritmos , Prognóstico , EletrocardiografiaRESUMO
Urban heat islands, where temperatures are elevated relative to non-urban surrounds, are near-ubiquitous in cities globally. Yet, the magnitude and form of urban heat islands in the tropics, where heat has a large morbidity and mortality burden, is not well understood, especially for those of urban informal settlements. We used 29 years of Landsat satellite-derived surface temperature, corroborated by in situ temperature measurements, to provide a detailed spatial and temporal assessment of urban heat islands in Makassar, Indonesia, a city that is representative of rapidly growing urban settlements across the tropics. Our analysis identified surface urban heat islands of up to 9.2 °C in long-urbanised parts of the city and 6.3 °C in informal settlements, the seasonal patterns of which were driven by change in non-urban areas rather than in urban areas themselves. In recently urbanised areas, the majority of urban heat island increase occurred before land became 50% urbanised, whereas the established heat island in long-urbanised areas remained stable in response to urban expansion. Green and blue space protected some informal settlements from the worst urban heat islands observed across the city and maintenance of such space will be essential to mitigate the growing heat burden from urban expansion and anthropogenic climate change. Settlements further than 4 km from the coast and with Normalised Difference Vegetation Index (NDVI) less than 0.2 had higher surface temperatures, with modelled effects of more than 5 °C. Surface temperature measurements were representative of in situ heat exposure, measured in a subset of 12 informal settlements, where mean indoor temperature had the strongest relationship with surface temperature (R2 = 0.413, P = 0.001). We advocate for green space to be prioritised in urban planning, redevelopment and informal settlement upgrading programs, with consideration of the unique environmental and socioeconomic context of tropical cities.
Assuntos
Monitoramento Ambiental , Temperatura Alta , Cidades , Temperatura , Planejamento de CidadesRESUMO
Importance: Early detection of atrial fibrillation (AF) may help prevent adverse cardiovascular events such as stroke. Deep learning applied to electrocardiograms (ECGs) has been successfully used for early identification of several cardiovascular diseases. Objective: To determine whether deep learning models applied to outpatient ECGs in sinus rhythm can predict AF in a large and diverse patient population. Design, Setting, and Participants: This prognostic study was performed on ECGs acquired from January 1, 1987, to December 31, 2022, at 6 US Veterans Affairs (VA) hospital networks and 1 large non-VA academic medical center. Participants included all outpatients with 12-lead ECGs in sinus rhythm. Main Outcomes and Measures: A convolutional neural network using 12-lead ECGs from 2 US VA hospital networks was trained to predict the presence of AF within 31 days of sinus rhythm ECGs. The model was tested on ECGs held out from training at the 2 VA networks as well as 4 additional VA networks and 1 large non-VA academic medical center. Results: A total of 907 858 ECGs from patients across 6 VA sites were included in the analysis. These patients had a mean (SD) age of 62.4 (13.5) years, 6.4% were female, and 93.6% were male, with a mean (SD) CHA2DS2-VASc (congestive heart failure, hypertension, age, diabetes mellitus, prior stroke or transient ischemic attack or thromboembolism, vascular disease, age, sex category) score of 1.9 (1.6). A total of 0.2% were American Indian or Alaska Native, 2.7% were Asian, 10.7% were Black, 4.6% were Latinx, 0.7% were Native Hawaiian or Other Pacific Islander, 62.4% were White, 0.4% were of other race or ethnicity (which is not broken down into subcategories in the VA data set), and 18.4% were of unknown race or ethnicity. At the non-VA academic medical center (72 483 ECGs), the mean (SD) age was 59.5 (15.4) years and 52.5% were female, with a mean (SD) CHA2DS2-VASc score of 1.6 (1.4). A total of 0.1% were American Indian or Alaska Native, 7.9% were Asian, 9.4% were Black, 2.9% were Latinx, 0.03% were Native Hawaiian or Other Pacific Islander, 74.8% were White, 0.1% were of other race or ethnicity, and 4.7% were of unknown race or ethnicity. A deep learning model predicted the presence of AF within 31 days of a sinus rhythm ECG on held-out test ECGs at VA sites with an area under the receiver operating characteristic curve (AUROC) of 0.86 (95% CI, 0.85-0.86), accuracy of 0.78 (95% CI, 0.77-0.78), and F1 score of 0.30 (95% CI, 0.30-0.31). At the non-VA site, AUROC was 0.93 (95% CI, 0.93-0.94); accuracy, 0.87 (95% CI, 0.86-0.88); and F1 score, 0.46 (95% CI, 0.44-0.48). The model was well calibrated, with a Brier score of 0.02 across all sites. Among individuals deemed high risk by deep learning, the number needed to screen to detect a positive case of AF was 2.47 individuals for a testing sensitivity of 25% and 11.48 for 75%. Model performance was similar in patients who were Black, female, or younger than 65 years or who had CHA2DS2-VASc scores of 2 or greater. Conclusions and Relevance: Deep learning of outpatient sinus rhythm ECGs predicted AF within 31 days in populations with diverse demographics and comorbidities. Similar models could be used in future AF screening efforts to reduce adverse complications associated with this disease.
Assuntos
Fibrilação Atrial , Aprendizado Profundo , Acidente Vascular Cerebral , Veteranos , Humanos , Masculino , Feminino , Pessoa de Meia-Idade , Fibrilação Atrial/complicações , Fibrilação Atrial/diagnóstico , Fibrilação Atrial/epidemiologia , Acidente Vascular Cerebral/epidemiologia , EletrocardiografiaRESUMO
BACKGROUND: Coronary artery calcification (CAC), often assessed by computed tomography (CT), is a powerful marker of coronary artery disease that can guide preventive therapies. Computed tomographies, however, are not always accessible or serially obtainable. It remains unclear whether other widespread tests such as transthoracic echocardiograms (TTEs) can be used to predict CAC. METHODS: Using a data set of 2,881 TTE videos paired with coronary calcium CTs, we trained a video-based artificial intelligence convolutional neural network to predict CAC scores from parasternal long-axis views. We evaluated the model's ability to classify patients from a held-out sample as well as an external site sample into zero CAC and high CAC (CAC ≥ 400 Agatston units) groups by receiver operating characteristic and precision-recall curves. We also investigated whether such classifications prognosticated significant differences in 1-year mortality rates by the log-rank test of Kaplan-Meier curves. RESULTS: Transthoracic echocardiogram artificial intelligence models had high discriminatory abilities in predicting zero CAC (receiver operating characteristic area under the curve [AUC] = 0.81 [95% CI, 0.74-0.88], F1 score = 0.95) and high CAC (AUC = 0.74 [0.68-0.8], F1 score = 0.74). This performance was confirmed in an external test data set of 92 TTEs (AUC = 0.75 [0.65-0.85], F1 score = 0.77; and AUC = 0.85 [0.76-0.93], F1 score = 0.59, respectively). Risk stratification by TTE-predicted CAC performed similarly to CT CAC scores in prognosticating significant differences in 1-year survival in high-CAC patients (CT CAC ≥ 400 vs CT CAC < 400, P = .03; TTE-predicted CAC ≥ 400 vs TTE-predicted CAC < 400, P = .02). CONCLUSIONS: A video-based deep learning model successfully used TTE videos to predict zero CAC and high CAC with high accuracy. Transthoracic echocardiography-predicted CAC prognosticated differences in 1-year survival similar to CT CAC. Deep learning of TTEs holds promise for future adjunctive coronary artery disease risk stratification to guide preventive therapies.