Background
COVID-19 is caused by the SARS-CoV-2 virus and has strikingly heterogeneous clinical manifestations, with most individuals contracting mild disease but a substantial minority experiencing ...fulminant cardiopulmonary symptoms or death. The clinical covariates and the laboratory tests performed on a patient provide robust statistics to guide clinical treatment. Deep learning approaches on a data set of this nature enable patient stratification and provide methods to guide clinical treatment.
Objective
Here, we report on the development and prospective validation of a state-of-the-art machine learning model to provide mortality prediction shortly after confirmation of SARS-CoV-2 infection in the Mayo Clinic patient population.
Methods
We retrospectively constructed one of the largest reported and most geographically diverse laboratory information system and electronic health record of COVID-19 data sets in the published literature, which included 11,807 patients residing in 41 states of the United States of America and treated at medical sites across 5 states in 3 time zones. Traditional machine learning models were evaluated independently as well as in a stacked learner approach by using AutoGluon, and various recurrent neural network architectures were considered. The traditional machine learning models were implemented using the AutoGluon-Tabular framework, whereas the recurrent neural networks utilized the TensorFlow Keras framework. We trained these models to operate solely using routine laboratory measurements and clinical covariates available within 72 hours of a patient’s first positive COVID-19 nucleic acid test result.
Results
The GRU-D recurrent neural network achieved peak cross-validation performance with 0.938 (SE 0.004) as the area under the receiver operating characteristic (AUROC) curve. This model retained strong performance by reducing the follow-up time to 12 hours (0.916 SE 0.005 AUROC), and the leave-one-out feature importance analysis indicated that the most independently valuable features were age, Charlson comorbidity index, minimum oxygen saturation, fibrinogen level, and serum iron level. In the prospective testing cohort, this model provided an AUROC of 0.901 and a statistically significant difference in survival (P<.001, hazard ratio for those predicted to survive, 95% CI 0.043-0.106).
Conclusions
Our deep learning approach using GRU-D provides an alert system to flag mortality for COVID-19–positive patients by using clinical covariates and laboratory values within a 72-hour window after the first positive nucleic acid test result.
The expansion of the recommend uniform screening panel to include more than 50 primary and secondary target conditions has resulted in a substantial increase of false positive results. As an ...alternative to subjective manipulation of cutoff values and overutilization of molecular testing, here we describe the performance outcome of an algorithm for disorders of methionine, cobalamin, and propionate metabolism that includes: (1) first tier screening inclusive of the broadest available spectrum of markers measured by tandem mass spectrometry; (2) integration of all results into a score of likelihood of disease for each target condition calculated by post-analytical interpretive tools created byCollaborative Laboratory Integrated Reports (CLIR), a multivariate pattern recognition software; and (3) further evaluation of abnormal scores by a second tier test measuring homocysteine, methylmalonic acid, and methylcitric acid. This approach can consistently reduce false positive rates to a <0.01% level, which is the threshold of precision newborn screening. We postulate that broader adoption of this algorithm could lead to substantial savings in health care expenditures. More importantly, it could prevent the stress and anxiety experienced by many families when faced with an abnormal newborn screening result that is later resolved as a false positive outcome.
Newborn screening for congenital hypothyroidism remains challenging decades after broad implementation worldwide. Testing protocols are not uniform in terms of targets (TSH and/or T4) and protocols ...(parallel vs. sequential testing; one or two specimen collection times), and specificity (with or without collection of a second specimen) is overall poor. The purpose of this retrospective study is to investigate the potential impact of multivariate pattern recognition software (CLIR) to improve the post-analytical interpretation of screening results. Seven programs contributed reference data (N = 1,970,536) and two sets of true (TP, N = 1369 combined) and false (FP, N = 15,201) positive cases for validation and verification purposes, respectively. Data were adjusted for age at collection, birth weight, and location using polynomial regression models of the fifth degree to create three-dimensional regression surfaces. Customized Single Condition Tools and Dual Scatter Plots were created using CLIR to optimize the differential diagnosis between TP and FP cases in the validation set. Verification testing correctly identified 446/454 (98%) of the TP cases, and could have prevented 1931/5447 (35%) of the FP cases, with variable impact among locations (range 4% to 50%). CLIR tools either as made here or preferably standardized to the recommended uniform screening panel could improve performance of newborn screening for congenital hypothyroidism.
To discover lipidomic alterations during pregnancy in mothers who subsequently delivered small for gestational age (SGA) neonates and identify predictive lipid markers that can help recognize and ...manage these mothers, we carried out untargeted lipidomics on maternal serum samples collected between 24-28 weeks of gestation. We used a nested case-control study design and serum from mothers who delivered SGA and appropriate for gestational age babies. We applied untargeted lipidomics using mass spectrometry to characterize lipids and discover changes associated with SGA births during pregnancy. Multivariate pattern recognition software Collaborative Laboratory Integrated Reports (CLIR) was used for the post-analytical recognition of range differences in lipid ratios that could differentiate between SGA and control mothers and their integration for complete separation between the two groups. Here, we report changes in lipids from serum collected during pregnancy in mothers who delivered SGA neonates. In contrast to normal pregnancies where lysophosphatidic acid increased over the course of the pregnancy owing to increased activity of lysophospholipase D, we observed a decrease (32%;
P
= 0.05) of 20:4-lysophosphatidic acid in SGA mothers, which could potentially compromise fetal growth and development. Integration of lipid ratios in an interpretive tool (CLIR) could completely separate SGA mothers from controls demonstrating the power of untargeted lipidomic analyses for identifying novel predictive biomarkers. Additional studies are required for further assessment of the lipid biomarkers identified in this report.
Lysophosphatidic acid was found to be decreased in mothers who subsequently delivered small for gestational age neonates during 24-28 weeks of pregnancy while it was increased in mothers who delivered control neonates.
•The MEA and BNST displayed greater microglial density than other sites examined.•Late aging was associated with greater numbers of microglia in MEA in males and females.•Aged females (but not males) ...displayed a greater number of microglia in the BNST.•Microglia in aged rats were larger in BNST, MEA, and CA3, regardless of sex.
Aging is associated with a substantial decline in the expression of social behavior as well as increased neuroinflammation. Since immune activation and subsequent increased expression of cytokines can suppress social behavior in young rodents, we examined age and sex differences in microglia within brain regions critical to social behavior regulation (PVN, BNST, and MEA) as well as in the hippocampus. Adult (3-month) and aged (18-month) male and female F344 (N = 26, n = 5–8/group) rats were perfused and Iba-1 immunopositive microglia were assessed using unbiased stereology and optical density. For stereology, microglia were classified based on the following criteria: (1) thin ramified processes, (2) thick long processes, (3) stout processes, or (4) round/ameboid shape. Among the structures examined, the highest density of microglia was evident in the BNST and MEA. Aged rats of both sexes displayed increased total number of microglia number exclusively in the MEA. Sex differences also emerged, whereby aged females (but not males) displayed greater total number of microglia in the BNST relative to their young adult counterparts. When morphological features of microglia were assessed, aged rats exhibited increased soma size in the BNST, MEA, and CA3. Together, these findings provide a comprehensive characterization of microglia number and morphology under ambient conditions in CNS sites critical for the normal expression of social processes. To the extent that microglia morphology is predictive of reactivity and subsequent cytokine release, these data suggest that the expression of social behavior in late aging may be adversely influenced by heightened inflammation.
Hepatitis C virus (HCV) genotype (gt) 3 is highly prevalent globally, with non‐gt3a subtypes common in Southeast Asia. Resistance‐associated substitutions (RASs) have been shown to play a role in ...treatment failure. However, the role of RASs in gt3 is not well understood. We report the prevalence of RASs in a cohort of direct‐acting antiviral treatment‐naive, gt3‐infected patients, including those with rarer subtypes, and evaluate the effect of these RASs on direct‐acting antivirals in vitro. Baseline samples from 496 gt3 patients enrolled in the BOSON clinical trial were analyzed by next‐generation sequencing after probe‐based enrichment for HCV. Whole viral genomes were analyzed for the presence of RASs to approved direct‐acting antivirals. The resistance phenotype of RASs in combination with daclatasvir, velpatasvir, pibrentasvir, elbasvir, and sofosbuvir was measured using the S52 ΔN gt3a replicon model. The nonstructural protein 5A A30K and Y93H substitutions were the most common at 8.9% (n = 44) and 12.3% (n = 61), respectively, and showed a 10‐fold and 11‐fold increase in 50% effect concentration for daclatasvir compared to the unmodified replicon. Paired RASs (A30K + L31M and A30K + Y93H) were identified in 18 patients (9 of each pair); these combinations were shown to be highly resistant to daclatasvir, velpatasvir, elbasvir, and pibrentasvir. The A30K + L31M combination was found in all gt3b and gt3g samples. Conclusion: Our study reveals high frequencies of RASs to nonstructural protein 5A inhibitors in gt3 HCV; the paired A30K + L31M substitutions occur in all patients with gt3b and gt3g virus, and in vitro analysis suggests that these subtypes may be inherently resistant to all approved nonstructural protein 5A inhibitors for gt3 HCV. (Hepatology 2018).
This article is based on an exploratory study that examines factors which predict children's performance on the numeracy component of the Australian National Assessment Program—Literacy and Numeracy ...(NAPLAN). Utilizing an ecological theoretical model, this study examines child, home and school variables which may enable or constrain NAPLAN numeracy performance. Data are presented from a nationally-representative sample of 2450 children participating in the Longitudinal Study of Australian Children (LSAC). Twenty-four children, home and school variables are tested as predictors of performance on the Year 3 NAPLAN numeracy assessment. The analysis includes an investigation of bivariate relationships between the outcome variable and each of the predictor variables. Following this a series of linear regression models are used to analyse the relation between child, home and school-related variables and NAPLAN numeracy performance. The results support the ecological model and point to the importance of a supportive home-school relationship on children's numeracy performance.
•Aged males did not exhibit increased basal inflammation in hippocampal tissue.•Extracellular release of cytokines and chemokines did not vary by age.•Release of CCL3 and CCL20 was higher in males, ...regardless of age.•Chemokine release was rapidly induced upon probe insertion and suppressed by BzATP.
Aging is associated with an enhanced neuroinflammatory response to acute immune challenge, often termed “inflammaging.” However, there are conflicting reports about whether baseline levels of inflammatory markers are elevated under ambient conditions in the aging brain, or whether such changes are observed predominantly in response to acute challenge. The present studies utilized two distinct approaches to assess inflammatory markers in young and aging Fischer 344 rats. Experiment 1 examined total tissue content of inflammatory markers from hippocampus of adult (3 month), middle-aged (12 month), and aging (18 month) male Fischer (F) 344 rats using multiplex analysis (23-plex). Though trends emerged for several cytokines, no significant differences in basal tissue content were observed across the 3 ages examined. Experiment 2 measured extracellular concentrations of inflammatory factors in the hippocampus from adult (3 month) and aging (18 month) males and females using large-molecule in vivo microdialysis. Although few significant aging-related changes were observed, robust sex differences were observed in extracellular concentrations of CCL3, CCL20, and IL-1α. Experiment 2 also evaluated the involvement of the P2X7 purinergic receptor in neuroinflammation using reverse dialysis of the selective agonist BzATP. BzATP produced an increase in IL-1α and IL-1β release and rapidly suppressed the release of CXCL1, CCL2, CCL3, CCL20, and IL-6. Other noteworthy sex by aging trends were observed in CCL3, IL-1β, and IL-6. Together, these findings provide important new insight into late-aging and sex differences in neuroinflammation, and their regulation by the P2X7 receptor.
Affordable next-generation sequencing (NGS) technologies for hepatitis C virus (HCV) may potentially identify both viral genotype and resistance genetic motifs in the era of directly acting antiviral ...(DAA) therapies. This study compared the ability of high-throughput NGS methods to generate full-length, deep, HCV sequence data sets and evaluated their utility for diagnostics and clinical assessment. NGS methods using (i) unselected HCV RNA (metagenomics), (ii) preenrichment of HCV RNA by probe capture, and (iii) HCV preamplification by PCR implemented in four United Kingdom centers were compared. Metrics of sequence coverage and depth, quasispecies diversity, and detection of DAA resistance-associated variants (RAVs), mixed HCV genotypes, and other coinfections were compared using a panel of samples with different viral loads, genotypes, and mixed HCV genotypes/subtypes geno(sub)types. Each NGS method generated near-complete genome sequences from more than 90% of samples. Enrichment methods and PCR preamplification generated greater sequence depth and were more effective for samples with low viral loads. All NGS methodologies accurately identified mixed HCV genotype infections. Consensus sequences generated by different NGS methods were generally concordant, and majority RAVs were consistently detected. However, methods differed in their ability to detect minor populations of RAVs. Metagenomic methods identified human pegivirus coinfections. NGS provided a rapid, inexpensive method for generating whole HCV genomes to define infecting genotypes, RAVs, comprehensive viral strain analysis, and quasispecies diversity. Enrichment methods are particularly suited for high-throughput analysis while providing the genotype and information on potential DAA resistance.
The purpose of this prospective multi-center cross-sectional study was to identify key biopsychosocial factors that impact quality of life (QOL) of youth with congenital heart disease (CHD). ...Patient-parent pairs were recruited at a regular hospital follow-up visit. Patient- and parent-proxy-reported QOL were assessed using the Pediatric Cardiac Quality of Life Inventory (PCQLI). Wallander’s and Varni’s disability-stress coping model guided factor selection, which included disease factors, educational impairment, psychosocial stress, child psychological and parent/family factors. Measures utilized for these factors included the Pediatric Inventory for Parents, Self-Perception Profile for Children/Adolescents, Child Behavior Checklist, Revised Children’s Manifest Anxiety Scale, Child PTSD Symptom Scale, State-Trait Anxiety Inventory, and Posttraumatic Diagnostic Scale. Ordinary least squares regression was applied to test the theoretical model, with backwards stepwise elimination process. The models accounted for a substantial amount of variance in QOL (Patient-reported PCQLI
R
2
= 0.58,
p
< 0.001; Parent-proxy-reported PCQLI
R
2
= 0.60,
p
< 0.001). For patient-reported QOL, disease factors, educational impairment, poor self-esteem, anxiety, patient posttraumatic stress, and parent posttraumatic stress were associated with lower QOL. For parent-proxy-report QOL, disease factors, educational impairment, greater parental medical stress, poorer child self-esteem, more child internalizing problems, and parent posttraumatic stress were associated with lower QOL. The results highlight that biopsychosocial factors account for over half the variance in QOL in CHD survivors. Assessing and treating psychological issues in the child and the parent may have a significant positive impact on QOL.