Automated Spontaneous Echo Contrast Detection Using a Multisequence Attention Convolutional Neural Network.

OBJECTIVE: Spontaneous echo contrast (SEC) is a vascular ultrasound finding associated with increased thromboembolism risk. However, identification requires expert determination and clinician time to report. We developed a deep learning model that can automatically identify SEC. Our model can be applied retrospectively without deviating from routine clinical practice. The retrospective nature of our model means future works could scan archival data to opportunistically correlate SEC findings with documented clinical outcomes. METHODS: We curated a data set of 801 archival acquisitions along the femoral vein from 201 patients. We used a multisequence convolutional neural network (CNN) with ResNetv2 backbone and visualized keyframe importance using soft attention. We evaluated SEC prediction performance using an 80/20 train/test split. We report receiver operating characteristic area under the curve (ROC-AUC), along with the Youden threshold-associated sensitivity, specificity, F1 score, true negative, false negative, false positive and true positive. RESULTS: Using soft attention, we can identify SEC with an AUC of 0.74, sensitivity of 0.73 and specificity of 0.68. Without soft attention, our model achieves an AUC of 0.69, sensitivity of 0.71 and specificity of 0.60. Additionally, we provide attention visualizations and note that our model assigns higher attention score to ultrasound frames containing more vessel lumen. CONCLUSION: Our multisequence CNN model can identify the presence of SEC from ultrasound keyframes with an AUC of 0.74, which could enable screening applications and enable more SEC data discovery. The model does not require the expert intervention or additional clinician reporting time that are currently significant barriers to SEC adoption. Model and processed data sets are publicly available at https://github.com/Ouwen/automatic-spontaneous-echo-contrast.

A Hybrid Deep Learning Approach to Identify Preventable Childhood Hearing Loss.

OBJECTIVE: Childhood hearing loss has well-known, lifelong consequences. Infection-related hearing loss disproportionately affects underserved communities yet can be prevented with early identification and treatment. This study evaluates the utility of machine learning in automating tympanogram classifications of the middle ear to facilitate layperson-guided tympanometry in resource-constrained communities. DESIGN: Diagnostic performance of a hybrid deep learning model for classifying narrow-band tympanometry tracings was evaluated. Using 10-fold cross-validation, a machine learning model was trained and evaluated on 4810 pairs of tympanometry tracings acquired by an audiologist and layperson. The model was trained to classify tracings into types A (normal), B (effusion or perforation), and C (retraction), with the audiologist interpretation serving as reference standard. Tympanometry data were collected from 1635 children from October 10, 2017, to March 28, 2019, from two previous cluster-randomized hearing screening trials (NCT03309553, NCT03662256). Participants were school-aged children from an underserved population in rural Alaska with a high prevalence of infection-related hearing loss. Two-level classification performance statistics were calculated by treating type A as pass and types B and C as refer. RESULTS: For layperson-acquired data, the machine-learning model achieved a sensitivity of 95.2% (93.3, 97.1), specificity of 92.3% (91.5, 93.1), and area under curve of 0.968 (0.955, 0.978). The model's sensitivity was greater than that of the tympanometer's built-in classifier [79.2% (75.5, 82.8)] and a decision tree based on clinically recommended normative values [56.9% (52.4, 61.3)]. For audiologist-acquired data, the model achieved a higher AUC of 0.987 (0.980, 0.993), had an equivalent sensitivity of 95.2 (93.3, 97.1), and a higher specificity of 97.7 (97.3, 98.2). CONCLUSIONS: Machine learning can detect middle ear disease with comparable performance to an audiologist using tympanograms acquired either by an audiologist or a layperson. Automated classification enables the use of layperson-guided tympanometry in hearing screening programs in rural and underserved communities, where early detection of treatable pathology in children is crucial to prevent the lifelong adverse effects of childhood hearing loss.

Phosphorylation barcodes direct biased chemokine signaling at CXCR3.

G protein-coupled receptor (GPCR)-biased agonism, selective activation of certain signaling pathways relative to others, is thought to be directed by differential GPCR phosphorylation "barcodes." At chemokine receptors, endogenous chemokines can act as "biased agonists", which may contribute to the limited success when pharmacologically targeting these receptors. Here, mass spectrometry-based global phosphoproteomics revealed that CXCR3 chemokines generate different phosphorylation barcodes associated with differential transducer activation. Chemokine stimulation resulted in distinct changes throughout the kinome in global phosphoproteomics studies. Mutation of CXCR3 phosphosites altered ß-arrestin 2 conformation in cellular assays and was consistent with conformational changes observed in molecular dynamics simulations. T cells expressing phosphorylation-deficient CXCR3 mutants resulted in agonist- and receptor-specific chemotactic profiles. Our results demonstrate that CXCR3 chemokines are non-redundant and act as biased agonists through differential encoding of phosphorylation barcodes, leading to distinct physiological processes.

Phosphorylation barcodes direct biased chemokine signaling at CXCR3.

G protein-coupled receptor (GPCR) biased agonism, the activation of some signaling pathways over others, is thought to largely be due to differential receptor phosphorylation, or "phosphorylation barcodes." At chemokine receptors, ligands act as "biased agonists" with complex signaling profiles, which contributes to the limited success in pharmacologically targeting these receptors. Here, mass spectrometry-based global phosphoproteomics revealed that CXCR3 chemokines generate different phosphorylation barcodes associated with differential transducer activation. Chemokine stimulation resulted in distinct changes throughout the kinome in global phosphoproteomic studies. Mutation of CXCR3 phosphosites altered ß-arrestin conformation in cellular assays and was confirmed by molecular dynamics simulations. T cells expressing phosphorylation-deficient CXCR3 mutants resulted in agonist- and receptor-specific chemotactic profiles. Our results demonstrate that CXCR3 chemokines are non-redundant and act as biased agonists through differential encoding of phosphorylation barcodes and lead to distinct physiological processes.

MimickNet, Mimicking Clinical Image Post- Processing Under Black-Box Constraints.

Image post-processing is used in clinical-grade ultrasound scanners to improve image quality (e.g., reduce speckle noise and enhance contrast). These post-processing techniques vary across manufacturers and are generally kept proprietary, which presents a challenge for researchers looking to match current clinical-grade workflows. We introduce a deep learning framework, MimickNet, that transforms conventional delay-and-summed (DAS) beams into the approximate Dynamic Tissue Contrast Enhanced (DTCE™) post-processed images found on Siemens clinical-grade scanners. Training MimickNet only requires post-processed image samples from a scanner of interest without the need for explicit pairing to DAS data. This flexibility allows MimickNet to hypothetically approximate any manufacturer's post-processing without access to the pre-processed data. MimickNet post-processing achieves a 0.940 ± 0.018 structural similarity index measurement (SSIM) compared to clinical-grade post-processing on a 400 cine-loop test set, 0.937 ± 0.025 SSIM on a prospectively acquired dataset, and 0.928 ± 0.003 SSIM on an out-of-distribution cardiac cine-loop after gain adjustment. To our knowledge, this is the first work to establish deep learning models that closely approximate ultrasound post-processing found in current medical practice. MimickNet serves as a clinical post-processing baseline for future works in ultrasound image formation to compare against. Additionally, it can be used as a pretrained model for fine-tuning towards different post-processing techniques. To this end, we have made the MimickNet software, phantom data, and permitted in vivo data open-source at https://github.com/ouwen/MimickNet.

Staphylococcus aureus Biofilms Induce Macrophage Dysfunction Through Leukocidin AB and Alpha-Toxin.

UNLABELLED: The macrophage response to planktonic Staphylococcus aureus involves the induction of proinflammatory microbicidal activity. However, S. aureus biofilms can interfere with these responses in part by polarizing macrophages toward an anti-inflammatory profibrotic phenotype. Here we demonstrate that conditioned medium from mature S. aureus biofilms inhibited macrophage phagocytosis and induced cytotoxicity, suggesting the involvement of a secreted factor(s). Iterative testing found the active factor(s) to be proteinaceous and partially agr-dependent. Quantitative mass spectrometry identified alpha-toxin (Hla) and leukocidin AB (LukAB) as critical molecules secreted by S. aureus biofilms that inhibit murine macrophage phagocytosis and promote cytotoxicity. A role for Hla and LukAB was confirmed by using hla and lukAB mutants, and synergy between the two toxins was demonstrated with a lukAB hla double mutant and verified by complementation. Independent confirmation of the effects of Hla and LukAB on macrophage dysfunction was demonstrated by using an isogenic strain in which Hla was constitutively expressed, an Hla antibody to block toxin activity, and purified LukAB peptide. The importance of Hla and LukAB during S. aureus biofilm formation in vivo was assessed by using a murine orthopedic implant biofilm infection model in which the lukAB hla double mutant displayed significantly lower bacterial burdens and more macrophage infiltrates than each single mutant. Collectively, these findings reveal a critical synergistic role for Hla and LukAB in promoting macrophage dysfunction and facilitating S. aureus biofilm development in vivo. IMPORTANCE: Staphylococcus aureus has a propensity to form multicellular communities known as biofilms. While growing in a biofilm, S. aureus displays increased tolerance to nutrient deprivation, antibiotic insult, and even host immune challenge. Previous studies have shown that S. aureus biofilms thwart host immunity in part by preventing macrophage phagocytosis. It remained unclear whether this was influenced solely by the considerable size of biofilms or whether molecules were also actively secreted to circumvent macrophage-mediated phagocytosis. This is the first report to demonstrate that S. aureus biofilms inhibit macrophage phagocytosis and induce macrophage death through the combined action of leukocidin AB and alpha-toxin. Loss of leukocidin AB and alpha-toxin expression resulted in enhanced S. aureus biofilm clearance in a mouse model of orthopedic implant infection, suggesting that these toxins could be targeted therapeutically to facilitate biofilm clearance in humans.

Binding of amelogenin to MMP-9 and their co-expression in developing mouse teeth.

Amelogenin is the most abundant matrix protein in enamel. Proper amelogenin processing by proteinases is necessary for its biological functions during amelogenesis. Matrix metalloproteinase 9 (MMP-9) is responsible for the turnover of matrix components. The relationship between MMP-9 and amelogenin during tooth development remains unknown. We tested the hypothesis that MMP-9 binds to amelogenin and they are co-expressed in ameloblasts during amelogenesis. We evaluated the distribution of both proteins in the mouse teeth using immunohistochemistry and confocal microscopy. At postnatal day 2, the spatial distribution of amelogenin and MMP-9 was co-localized in preameloblasts, secretory ameloblasts, enamel matrix and odontoblasts. At the late stages of mouse tooth development, expression patterns of amelogenin and MMP-9 were similar to that seen in postnatal day 2. Their co-expression was further confirmed by RT-PCR, Western blot and enzymatic zymography analyses in enamel organ epithelial and odontoblast-like cells. Immunoprecipitation assay revealed that MMP-9 binds to amelogenin. The MMP-9 cleavage sites in amelogenin proteins across species were found using bio-informative software program. Analyses of these data suggest that MMP-9 may be involved in controlling amelogenin processing and enamel formation.