Automated Detection of Aortic Stenosis Using Machine Learning.

BACKGROUND: Aortic stenosis (AS) is a degenerative valve condition that is underdiagnosed and undertreated. Detection of AS using limited two-dimensional echocardiography could enable screening and improve appropriate referral and treatment of this condition. The aim of this study was to develop methods for automated detection of AS from limited imaging data sets. METHODS: Convolutional neural networks were trained, validated, and tested using limited two-dimensional transthoracic echocardiographic data sets. Networks were developed to accomplish two sequential tasks: (1) view identification and (2) study-level grade of AS. Balanced accuracy and area under the receiver operator curve (AUROC) were the performance metrics used. RESULTS: Annotated images from 577 patients were included. Neural networks were trained on data from 338 patients (average n = 10,253 labeled images), validated on 119 patients (average n = 3,505 labeled images), and performance was assessed on a test set of 120 patients (average n = 3,511 labeled images). Fully automated screening for AS was achieved with an AUROC of 0.96. Networks can distinguish no significant (no, mild, mild to moderate) AS from significant (moderate or severe) AS with an AUROC of 0.86 and between early (mild or mild to moderate AS) and significant (moderate or severe) AS with an AUROC of 0.75. External validation of these networks in a cohort of 8,502 outpatient transthoracic echocardiograms showed that screening for AS can be achieved using parasternal long-axis imaging only with an AUROC of 0.91. CONCLUSION: Fully automated detection of AS using limited two-dimensional data sets is achievable using modern neural networks. These methods lay the groundwork for a novel method for screening for AS.

Preference for opioids in emergency department patients with acute musculoskeletal pain.

BACKGROUND: Public awareness of the opioid epidemic is increasing nationally, emphasizing the need to develop methods to reduce opioid use. We determined patient preference for analgesics before and after a brief educational intervention informing them of the risks and benefits of opioids versus non-steroidal anti-inflammatory drugs (NSAID's). We hypothesized 50% of patients would prefer opioids pre-intervention and that this would be reduced by the intervention by at least 15%. METHODS: Study Design-Before and after study. Setting-Suburban ED with annual census of 110,000. Patients-English-speaking adult ED patients with acute musculoskeletal pain. Interventions-An anonymous survey was administered by an investigator not involved in the patient's clinical care prior to physician evaluation, before and after a video describing the risks and benefits of opioids versus NSAID's. Patients were asked if they desired analgesics. Data Analysis-Descriptive statistics were used to summarize the data. Univariate analysis and logistic regression were used to predict patient demographics and pain characteristics associated with desire for analgesics. RESULTS: Of all 94 patients, 48 (51% [95% CI 41-62%]) desired an analgesic pre-intervention. Of these 48 patients, 10 (11% [5-19%]) specifically preferred an opioid. Of the 10 patients who preferred an opioid pre-intervention, one had no preference for analgesic post-intervention. CONCLUSIONS: Many adult ED patients with acute musculoskeletal pain do not desire any analgesics and few specifically prefer opioids. This knowledge can prove helpful to ED physicians across the country in discussing pain management options with patients as we attempt to combat the opioid epidemic.

Structure-Activity Relationships of Cbx7 Inhibitors, Including Selectivity Studies against Other Cbx Proteins.

The five human polycomb (Pc) paralog proteins, chromobox homolog (Cbx) 2/4/6/7/8, are a family of chromodomain containing methyllysine reader proteins that are canonical readers of trimethyllysine 27 on histone 3 (H3K27me3). The aberrant expression of the Cbx7 gene is implicated in several cancers including prostate, gastric, thyroid, pancreas, and colon cancer. Previous reports on antagonizing the molecular recognition of Cbx7-H3K27me3 with chemical inhibitors showed an impact on prostate cancer cell lines. We report here on the design, synthesis, and structure-activity relationships of a series of potent peptidomimetic antagonists that were optimized on a trimethyllysine-containing scaffold to target Cbx7. The ligands were characterized using fluorescence polarization (FP) for their binding efficiency and selectivity against the Pc paralog Cbx proteins. The most selective ligand 9, as indicated by the FP data analysis, was further characterized using the isothermal titration calorimetry (ITC). Compound 9 exhibits a 220 nM potency for Cbx7 and exhibits 3.3, 1.8, 7.3 times selective for Cbx7 over Cbx2/4/8 and 28-fold selective over the HP1 family member Cbx1. Our research provides several potent and partially selective inhibitors for Cbx2/4/7 that do not contain trimethyllysine. Our models and binding data suggest that the aromatic cages of Cbx7/Cbx4 can accommodate larger alkyl groups such as diisobutyl substitution on the lysine nitrogen.