Machine Learning of Plasma Proteomics Classifies Diagnosis of Interstitial Lung Disease.

RATIONALE: Distinguishing connective tissue disease associated interstitial lung disease (CTD-ILD) from idiopathic pulmonary fibrosis (IPF) can be clinically challenging. OBJECTIVES: Identify proteins that separate and classify CTD-ILD from IPF patients. METHODS: Four registries with 1247 IPF and 352 CTD-ILD patients were included in analyses. Plasma samples were subjected to high-throughput proteomics assays. Protein features were prioritized using Recursive Feature Elimination (RFE) to construct a proteomic classifier. Multiple machine learning models, including Support Vector Machine, LASSO regression, Random Forest (RF), and imbalanced-RF, were trained and tested in independent cohorts. The validated models were used to classify each case iteratively in external datasets. MEASUREMENT AND MAIN RESULTS: A classifier with 37 proteins (PC37) was enriched in biological process of bronchiole development and smooth muscle proliferation, and immune responses. Four machine learning models used PC37 with sex and age score to generate continuous classification values. Receiver-operating-characteristic curve analyses of these scores demonstrated consistent Area-Under-Curve 0.85-0.90 in test cohort, and 0.94-0.96 in the single-sample dataset. Binary classification demonstrated 78.6%-80.4% sensitivity and 76%-84.4% specificity in test cohort, 93.5%-96.1% sensitivity and 69.5%-77.6% specificity in single-sample classification dataset. Composite analysis of all machine learning models confirmed 78.2% (194/248) accuracy in test cohort and 82.9% (208/251) in single-sample classification dataset. CONCLUSIONS: Multiple machine learning models trained with large cohort proteomic datasets consistently distinguished CTD-ILD from IPF. Identified proteins involved in immune pathways. We further developed a novel approach for single sample classification, which could facilitate honing the differential diagnosis of ILD in challenging cases and improve clinical decision-making.

Associations of Plasma Omega-3 Fatty Acids With Progression and Survival in Pulmonary Fibrosis.

BACKGROUND: Preclinical experiments suggest protective effects of omega-3 fatty acids and their metabolites in lung injury and fibrosis. Whether higher intake of omega-3 fatty acids is associated with disease progression and survival in humans with pulmonary fibrosis is unknown. RESEARCH QUESTION: What are the associations of plasma omega-3 fatty acid levels (a validated marker of omega-3 nutritional intake) with disease progression and transplant-free survival in pulmonary fibrosis? STUDY DESIGN AND METHODS: Omega-3 fatty acid levels were measured from plasma samples of patients with clinically diagnosed pulmonary fibrosis from the Pulmonary Fibrosis Foundation Patient Registry (n = 150), University of Virginia (n = 58), and University of Chicago (n = 101) cohorts. The N-3 index (docosahexaenoic acid + eicosapentaenoic acid) was the primary exposure variable of interest. Linear-mixed effects models with random intercept and slope were used to examine associations of plasma omega-3 fatty acid levels with changes in FVC and diffusing capacity for carbon monoxide over a period of 12 months. Cox proportional hazards models were used to examine transplant-free survival. Stratified analyses by telomere length were performed in the University of Chicago cohort. RESULTS: Most of the cohort were patients with idiopathic pulmonary fibrosis (88%) and male patients (74%). One-unit increment in log-transformed N-3 index plasma level was associated with a change in diffusing capacity for carbon monoxide of 1.43 mL/min/mm Hg per 12 months (95% CI, 0.46-2.41) and a hazard ratio for transplant-free survival of 0.44 (95% CI, 0.24-0.83). Cardiovascular disease history, smoking, and antifibrotic usage did not significantly modify associations. Omega-3 fatty acid levels were not significantly associated with changes in FVC. Higher eicosapentaenoic acid plasma levels were associated with longer transplant-free survival among University of Chicago participants with shorter telomere length (P value for interaction = .02). INTERPRETATION: Further research is needed to investigate underlying biological mechanisms and whether omega-3 fatty acids are a potential disease-modifying therapy.

Sirt2 promotes white matter oligodendrogenesis during development and in models of neonatal hypoxia.

Delayed oligodendrocyte (OL) maturation caused by hypoxia (Hx)-induced neonatal brain injury results in hypomyelination and leads to neurological disabilities. Previously, we characterized Sirt1 as a crucial regulator of OL progenitor cell (OPC) proliferation in response to Hx. We now identify Sirt2 as a critical promoter of OL differentiation during both normal white matter development and in a mouse model of Hx. Importantly, we find that Hx reduces Sirt2 expression in mature OLs and that Sirt2 overexpression in OPCs restores mature OL populations. Reduced numbers of Sirt2+ OLs were also observed in the white matter of preterm human infants. We show that Sirt2 interacts with p27Kip1/FoxO1, p21Cip1/Cdk4, and Cdk5 pathways, and that these interactions are altered by Hx. Furthermore, Hx induces nuclear translocation of Sirt2 in OPCs where it binds several genomic targets. Overall, these results indicate that a balance of Sirt1 and Sirt2 activity is required for developmental oligodendrogenesis, and that these proteins represent potential targets for promoting repair following white matter injury.