Validation of a Deep Learning Algorithm for Continuous, Real-Time Detection of Atrial Fibrillation Using a Wrist-Worn Device in an Ambulatory Environment.

BACKGROUND: Wearable devices may be useful for identification, quantification and characterization, and management of atrial fibrillation (AF). To date, consumer wrist-worn devices for AF detection using photoplethysmography-based algorithms perform only periodic checks when the user is stationary and are US Food and Drug Administration cleared for prediagnostic uses without intended use for clinical decision-making. There is an unmet need for medical-grade diagnostic wrist-worn devices that provide long-term, continuous AF monitoring. METHODS AND RESULTS: We evaluated the performance of a wrist-worn device with lead-I ECG and continuous photoplethysmography (Verily Study Watch) and photoplethysmography-based convolutional neural network for AF detection and burden estimation in a prospective multicenter study that enrolled 117 patients with paroxysmal AF. A 14-day continuous ECG monitor (Zio XT) served as the reference device to evaluate algorithm sensitivity and specificity for detection of AF in 15-minute intervals. A total of 91 857 intervals were contributed by 111 subjects with evaluable reference and test data (18.3 h/d median watch wear time). The watch was 96.1% sensitive (95% CI, 92.7%-98.0%) and 98.1% specific (95% CI, 97.2%-99.1%) for interval-level AF detection. Photoplethysmography-derived AF burden estimation was highly correlated with the reference device burden (R2=0.986) with a mean difference of 0.8% (95% limits of agreement, -6.6% to 8.2%). CONCLUSIONS: Continuous monitoring using a photoplethysmography-based convolutional neural network incorporated in a wrist-worn device has clinical-grade performance for AF detection and burden estimation. These findings suggest that monitoring can be performed with wrist-worn wearables for diagnosis and clinical management of AF. REGISTRATION INFORMATION: URL: https://www.clinicaltrials.gov; Unique identifier: NCT04546763.

Interobserver Agreement and Correlation of an Automated Algorithm for B-Line Identification and Quantification With Expert Sonologist Review in a Handheld Ultrasound Device.

OBJECTIVES: B-lines are ultrasound artifacts that can be used to detect a variety of pathologic lung conditions. Computer-aided methods to detect and quantify B-lines may standardize quantification and improve diagnosis by novice users. We sought to test the performance of an automated algorithm for the detection and quantification of B-lines in a handheld ultrasound device (HHUD). METHODS: Ultrasound images were prospectively collected on adult emergency department patients with dyspnea. Images from the first 124 patients were used for algorithm development. Clips from 80 unique subjects for testing were randomly selected in a predefined proportion of B-lines (0 B-lines, 1-2 B-lines, 3 or more B-lines) and blindly reviewed by five experts using both a manual and reviewer-adjusted process. Intraclass correlation coefficient (ICC) and weighted kappa were used to measure agreement, while an a priori threshold of an ICC (3,k) of 0.75 and precision of 0.3 were used to define adequate performance. RESULTS: ICC between the algorithm and manual count was 0.84 (95% confidence interval [CI] 0.75-0.90), with a precision of 0.15. ICC between the reviewer-adjusted count and the algorithm count was 0.94 (95% CI 0.90-0.96), and the ICC between the manual and reviewer-adjusted counts was 0.94 (95% CI 0.90-0.96). Weighted kappa was 0.72 (95% CI 0.49-0.95), 0.88 (95% CI 0.74-1), and 0.85 (95% CI 0.89-0.96), respectively. CONCLUSIONS: This study demonstrates a high correlation between point-of-care ultrasound experts and an automated algorithm to identify and quantify B-lines using an HHUD. Future research may incorporate this HHUD in clinical studies in multiple settings and users of varying experience levels.

A Quantitative Angiographic Comparison of Restenotic Tissue Following Placement of Drug-Eluting Stents and Bare Metal Stents in Symptomatic Patients With Femoropopliteal Disease.

PURPOSE: To evaluate the patterns of in-stent restenosis (ISR) within femoropopliteal bare metal stents (BMS) and drug-eluting stents (DES) as determined by quantitative angiographic analysis. METHODS: Utilizing results from independent core laboratory angiographic imaging analysis, quantitative assessment of the restenotic tissue burden was evaluated in 33 patients with symptomatic femoropopliteal ISR, including 20 lesions in 19 patients (mean age 71.5±8.1 years; 11 men) treated with DES and 14 lesions in 14 patients (mean age 70.6±9.2 years; 9 men) treated with BMS. RESULTS: The average time to target lesion revascularization was similar (8.7 months) for the DES and BMS groups. The DES group had significantly less recurrent disease burden (17.1%) compared with the BMS group (27.8%, p=0.03), representing a 39% relative reduction. CONCLUSION: Reduced restenotic tissue after endovascular intervention is associated with improved hemodynamics and fewer clinical symptoms and may explain the reduced need for reintervention in restenotic lesions initially treated with DES as compared with BMS. Further study of treatment failure modes may lead to improved device selection criteria to treat patients with peripheral artery disease.

Mechanism for maturation-related reorganization of flavivirus glycoproteins.

Flaviviruses, such as dengue, West Nile, and yellow fever viruses, assemble as fusion-incompetent particles and subsequently undergo a large reorganization of their glycoprotein envelope resulting in formation of mature infectious virions. Here we used a combination of three-dimensional cryo-electron tomography and two-dimensional image analysis to study pleomorphic maturation intermediates of dengue virus 2. Icosahedral symmetries of immature and mature regions within one particle were mismatched relative to each other. Furthermore, the orientation of the two regions relative to each other differed among particles. Therefore, there cannot be a specific pathway determining the maturation of all particles. Instead, the region with mature structure expands when glycoproteins on its boundary acquire suitable orientation and conformation to allow them to become a stable part of the mature region. This type of maturation is possible because the envelope glycoproteins are anchored to the phospholipid bilayer that is a part of flavivirus virions and are thus restricted to movement on the two-dimensional surface of the particle. Therefore, compounds that limit movement of the glycoproteins within the virus membrane might be used as inhibitors of flavivirus maturation.

Rubella virus capsid protein structure and its role in virus assembly and infection.

Rubella virus (RV) is a leading cause of birth defects due to infectious agents. When contracted during pregnancy, RV infection leads to severe damage in fetuses. Despite its medical importance, compared with the related alphaviruses, very little is known about the structure of RV. The RV capsid protein is an essential structural component of virions as well as a key factor in virus-host interactions. Here we describe three crystal structures of the structural domain of the RV capsid protein. The polypeptide fold of the RV capsid protomer has not been observed previously. Combining the atomic structure of the RV capsid protein with the cryoelectron tomograms of RV particles established a low-resolution structure of the virion. Mutational studies based on this structure confirmed the role of amino acid residues in the capsid that function in the assembly of infectious virions.

Sample preparation induced artifacts in cryo-electron tomographs.

We investigated the effects of sample preparation and of the exposure to an electron beam on particles in cryo-electron tomographs. Various virus particles with icosahedral symmetry were examined, allowing a comparison of symmetrically related components that should be identical in structure but might be affected differently by these imaging artifacts. Comparison of tomographic reconstructions with previously determined structures established by an independent method showed that neither freezing nor electron beam exposure produced a significant amount of shrinkage along the z axis (thickness). However, we observed damage to regions of the particles located close to the surface of the vitreous ice.

Structure and assembly of a paramyxovirus matrix protein.

Many pleomorphic, lipid-enveloped viruses encode matrix proteins that direct their assembly and budding, but the mechanism of this process is unclear. We have combined X-ray crystallography and cryoelectron tomography to show that the matrix protein of Newcastle disease virus, a paramyxovirus and relative of measles virus, forms dimers that assemble into pseudotetrameric arrays that generate the membrane curvature necessary for virus budding. We show that the glycoproteins are anchored in the gaps between the matrix proteins and that the helical nucleocapsids are associated in register with the matrix arrays. About 90% of virions lack matrix arrays, suggesting that, in agreement with previous biological observations, the matrix protein needs to dissociate from the viral membrane during maturation, as is required for fusion and release of the nucleocapsid into the host's cytoplasm. Structure and sequence conservation imply that other paramyxovirus matrix proteins function similarly.

Cryo-electron tomography of rubella virus.

Rubella virus is the only member of the Rubivirus genus within the Togaviridae family and is the causative agent of the childhood disease known as rubella or German measles. Here, we report the use of cryo-electron tomography to examine the three-dimensional structure of rubella virions and compare their structure to that of Ross River virus, a togavirus belonging the genus Alphavirus. The ectodomains of the rubella virus glycoproteins, E1 and E2, are shown to be organized into extended rows of density, separated by 9 nm on the viral surface. We also show that the rubella virus nucleocapsid structure often forms a roughly spherical shell which lacks high density at its center. While many rubella virions are approximately spherical and have dimensions similar to that of the icosahedral Ross River virus, the present results indicate that rubella exhibits a large degree of pleomorphy. In addition, we used rotation function calculations and other analyses to show that approximately spherical rubella virions lack the icosahedral organization which characterizes Ross River and other alphaviruses. The present results indicate that the assembly mechanism of rubella virus, which has previously been shown to differ from that of the alphavirus assembly pathway, leads to an organization of the rubella virus structural proteins that is different from that of alphaviruses.

Maturation of flaviviruses starts from one or more icosahedrally independent nucleation centres.

Flaviviruses assemble as fusion-incompetent immature particles and subsequently undergo conformational change leading to release of infectious virions. Flavivirus infections also produce combined 'mosaic' particles. Here, using cryo-electron tomography, we report that mosaic particles of dengue virus type 2 had glycoproteins organized into two regions of mature and immature structure. Furthermore, particles of a maturation-deficient mutant had their glycoproteins organized into two regions of immature structure with mismatching icosahedral symmetries. It is therefore apparent that the maturation-related reorganization of the flavivirus glycoproteins is not synchronized across the whole virion, but is initiated from one or more nucleation centres. Similar deviation from icosahedral symmetry might be relevant to the asymmetrical mode of genome packaging and cell entry of other viruses.

Future prospects.

Cryo-electron microscopy (cryo-EM) in combination with single-particle analysis has begun to complement crystallography in the study of large macromolecules at near-atomic resolution. Furthermore, advances in cryo-electron tomography have made possible the study of macromolecules within their cellular environment. Single-particle and tomographic studies will become even more useful when technologies for improving the signal-to-noise ratio such as direct electron detectors and phase plates become widely available. Automated image acquisition has significantly reduced the time and effort required to determine the structures of macromolecular assemblies. As a result, the number of structures determined by cryo-EM is growing exponentially. However, there is an urgent need for improved criteria for validating both the reconstruction process and the atomic models derived from cryo-EM data. Another major challenge will be mitigating the effects of anisotropy caused by the missing wedge and the excessively low signal-to-noise ratio for tomographic data. Parallels between the development of macromolecular crystallography and cryo-EM have been used to tentatively predict the future of cryo-EM.

Structural studies of Hantaan virus.

Hantaan virus is the prototypic member of the Hantavirus genus within the family Bunyaviridae and is a causative agent of the potentially fatal hemorrhagic fever with renal syndrome. The Bunyaviridae are a family of negative-sense RNA viruses with three-part segmented genomes. Virions are enveloped and decorated with spikes derived from a pair of glycoproteins (Gn and Gc). Here, we present cryo-electron tomography and single-particle cryo-electron microscopy studies of Hantaan virus virions. We have determined the structure of the tetrameric Gn-Gc spike complex to a resolution of 2.5 nm and show that spikes are ordered in lattices on the virion surface. Large cytoplasmic extensions associated with each Gn-Gc spike also form a lattice on the inner surface of the viral membrane. Rod-shaped ribonucleoprotein complexes are arranged into nearly parallel pairs and triplets within virions. Our results differ from the T=12 icosahedral organization found for some bunyaviruses. However, a comparison of our results with the previous tomographic studies of the nonpathogenic Tula hantavirus indicates a common structural organization for hantaviruses.

An icosahedral algal virus has a complex unique vertex decorated by a spike.

Paramecium bursaria Chlorella virus-1 is an icosahedrally shaped, 1,900-A-diameter virus that infects unicellular eukaryotic green algae. A 5-fold symmetric, 3D reconstruction using cryoelectron microscopy images has now shown that the quasiicosahedral virus has a unique vertex, with a pocket on the inside and a spike structure on the outside of the capsid. The pocket might contain enzymes for use in the initial stages of infection. The unique vertex consists of virally coded proteins, some of which have been identified. Comparison of shape, size, and location of the spike with similar features in bacteriophages T4 and P22 suggests that the spike might be a cell-puncturing device. Similar asymmetric features may have been missed in previous analyses of many other viruses that had been assumed to be perfectly icosahedral.

Structural comparison of different antibodies interacting with parvovirus capsids.

The structures of canine parvovirus (CPV) and feline parvovirus (FPV) complexed with antibody fragments from eight different neutralizing monoclonal antibodies were determined by cryo-electron microscopy (cryoEM) reconstruction to resolutions varying from 8.5 to 18 A. The crystal structure of one of the Fab molecules and the sequence of the variable domain for each of the Fab molecules have been determined. The structures of Fab fragments not determined crystallographically were predicted by homology modeling according to the amino acid sequence. Fitting of the Fab and virus structures into the cryoEM densities identified the footprints of each antibody on the viral surface. As anticipated from earlier analyses, the Fab binding sites are directed to two epitopes, A and B. The A site is on an exposed part of the surface near an icosahedral threefold axis, whereas the B site is about equidistant from the surrounding five-, three-, and twofold axes. One antibody directed to the A site binds CPV but not FPV. Two of the antibodies directed to the B site neutralize the virus as Fab fragments. The differences in antibody properties have been linked to the amino acids within the antibody footprints, the position of the binding site relative to the icosahedral symmetry elements, and the orientation of the Fab structure relative to the surface of the virus. Most of the exposed surface area was antigenic, although each of the antibodies had a common area of overlap that coincided with the positions of the previously mapped escape mutations.

The capsid proteins of a large, icosahedral dsDNA virus.

Chilo iridescent virus (CIV) is a large (approximately 1850 A diameter) insect virus with an icosahedral, T=147 capsid, a double-stranded DNA (dsDNA) genome, and an internal lipid membrane. The structure of CIV was determined to 13 A resolution by means of cryoelectron microscopy (cryoEM) and three-dimensional image reconstruction. A homology model of P50, the CIV major capsid protein (MCP), was built based on its amino acid sequence and the structure of the homologous Paramecium bursaria chlorella virus 1 Vp54 MCP. This model was fitted into the cryoEM density for each of the 25 trimeric CIV capsomers per icosahedral asymmetric unit. A difference map, in which the fitted CIV MCP capsomers were subtracted from the CIV cryoEM reconstruction, showed that there are at least three different types of minor capsid proteins associated with the capsomers outside the lipid membrane. "Finger" proteins are situated at many, but not all, of the spaces between three adjacent capsomers within each trisymmetron, and "zip" proteins are situated between sets of three adjacent capsomers at the boundary between neighboring trisymmetrons and pentasymmetrons. Based on the results of segmentation and density correlations, there are at least eight finger proteins and three dimeric and two monomeric zip proteins in one asymmetric unit of the CIV capsid. These minor proteins appear to stabilize the virus by acting as intercapsomer cross-links. One transmembrane "anchor" protein per icosahedral asymmetric unit, which extends from beneath one of the capsomers in the pentasymmetron to the internal leaflet of the lipid membrane, may provide additional stabilization for the capsid. These results are consistent with the observations for other large, icosahedral dsDNA viruses that also utilize minor capsid proteins for stabilization and for determining their assembly.

Binding of a neutralizing antibody to dengue virus alters the arrangement of surface glycoproteins.

The monoclonal antibody 1A1D-2 has been shown to strongly neutralize dengue virus serotypes 1, 2 and 3, primarily by inhibiting attachment to host cells. A crystal structure of its antigen binding fragment (Fab) complexed with domain III of the viral envelope glycoprotein, E, showed that the epitope would be partially occluded in the known structure of the mature dengue virus. Nevertheless, antibody could bind to the virus at 37 degrees C, suggesting that the virus is in dynamic motion making hidden epitopes briefly available. A cryo-electron microscope image reconstruction of the virus:Fab complex showed large changes in the organization of the E protein that exposed the epitopes on two of the three E molecules in each of the 60 icosahedral asymmetric units of the virus. The changes in the structure of the viral surface are presumably responsible for inhibiting attachment to cells.

Cryo-EM study of the Pseudomonas bacteriophage phiKZ.

The phiKZ virus is one of the largest known bacteriophages. It infects Pseudomonas aeruginosa, which is frequently pathogenic in humans, and, therefore, has potential for phage therapy. The phiKZ virion consists of an approximately 1450 A diameter icosahedral head and an approximately 2000 A long contractile tail. The structure of the phiKZ tail has been determined using cryo-electron microscopy. The phiKZ tail is much longer than that of bacteriophage T4. However, the helical parameters of their contractile sheaths, surrounding their tail tubes, are comparable. Although there is no recognizable sequence similarity between the phiKZ and T4 tail sheath proteins, they are similar in size and shape, suggesting that they evolved from a common ancestor. The phiKZ baseplate is significantly larger than that of T4 and has a flatter shape. Nevertheless, phiKZ, similar to T4, has a cell-puncturing device in the middle of its baseplate.

Cryo-electron microscopy study of bacteriophage T4 displaying anthrax toxin proteins.

The bacteriophage T4 capsid contains two accessory surface proteins, the small outer capsid protein (Soc, 870 copies) and the highly antigenic outer capsid protein (Hoc, 155 copies). As these are dispensable for capsid formation, they can be used for displaying proteins and macromolecular complexes on the T4 capsid surface. Anthrax toxin components were attached to the T4 capsid as a fusion protein of the N-terminal domain of the anthrax lethal factor (LFn) with Soc. The LFn-Soc fusion protein was complexed in vitro with Hoc(-)Soc(-)T4 phage. Subsequently, cleaved anthrax protective antigen heptamers (PA63)(7) were attached to the exposed LFn domains. A cryo-electron microscopy study of the decorated T4 particles shows the complex of PA63 heptamers with LFn-Soc on the phage surface. Although the cryo-electron microscopy reconstruction is unable to differentiate on its own between different proposed models of the anthrax toxin, the density is consistent with a model that had predicted the orientation and position of three LFn molecules bound to one PA63 heptamer.

The structures of bacteriophages K1E and K1-5 explain processive degradation of polysaccharide capsules and evolution of new host specificities.

External polysaccharides of many pathogenic bacteria form capsules protecting the bacteria from the animal immune system and phage infection. However, some bacteriophages can digest these capsules using glycosidases displayed on the phage particle. We have utilized cryo-electron microscopy to determine the structures of phages K1E and K1-5 and thereby establish the mechanism by which these phages attain and switch their host specificity. Using a specific glycosidase, both phages penetrate the capsule and infect the neuroinvasive human pathogen Escherichia coli K1. In addition to the K1-specific glycosidase, each K1-5 particle carries a second enzyme that allows it to infect E. coli K5, whose capsule is chemically different from that of K1. The enzymes are organized into a multiprotein complex attached via an adapter protein to the virus portal vertex, through which the DNA is ejected during infection. The structure of the complex suggests a mechanism for the apparent processivity of degradation that occurs as the phage drills through the polysaccharide capsule. The enzymes recognize the adapter protein by a conserved N-terminal sequence, providing a mechanism for phages to acquire different enzymes and thus to evolve new host specificities.

Asymmetric binding of transferrin receptor to parvovirus capsids.

Although many viruses are icosahedral when they initially bind to one or more receptor molecules on the cell surface, such an interaction is asymmetric, probably causing a breakdown in the symmetry and conformation of the original infecting virion in preparation for membrane penetration and release of the viral genome. Cryoelectron microscopy and biochemical analyses show that transferrin receptor, the cellular receptor for canine parvovirus, can bind to only one or a few of the 60 icosahedrally equivalent sites on the virion, indicating that either canine parvovirus has inherent asymmetry or binding of receptor induces asymmetry. The asymmetry of receptor binding to canine parvovirus is reminiscent of the special portal in tailed bacteriophages and some large, icosahedral viruses. Asymmetric interactions of icosahedral viruses with their hosts might be a more common phenomenon than previously thought and may have been obscured by averaging in previous crystallographic and electron microscopic structure determinations.

Molecular dissection of ø29 scaffolding protein function in an in vitro assembly system.

An in vitro assembly system was developed to study prolate capsid assembly of phage ø29 biochemically, and to identify regions of scaffolding protein required for its functions. The crowding agent polyethylene glycol can induce bacteriophage ø29 monomeric capsid protein and dimeric scaffolding protein to co-assemble to form particles which have the same geometry as either prolate T=3 Q=5 procapsids formed in vivo or previously observed isometric particles. The formation of particles is a scaffolding-dependent reaction. The balance between the fidelity and efficiency of assembly is controlled by the concentration of crowding agent and temperature. The assembly process is salt sensitive, suggesting that the interactions between the scaffolding and coat proteins are electrostatic. Three N-terminal ø29 scaffolding protein deletion mutants, Delta 1-9, Delta 1-15 and Delta 1-22, abolish the assembly activity. Circular dichroism spectra indicate that these N-terminal deletions are accompanied by a loss of helicity. The inability of these proteins to dimerize suggests that the N-terminal region of the scaffolding protein contributes to the dimer interface and maintains the structural integrity of the dimeric protein. Two C-terminal scaffolding protein deletion mutants, Delta 79-97 and Delta 62-97, also fail to promote assembly. However, the secondary structure and the dimerization ability of these mutants are unchanged relative to wild-type, which suggests that the C terminus is the likely site of interaction with the capsid protein.