Daily Respiratory Research Analysis
Three impactful studies advance respiratory health: (1) a multicenter machine-learning model (AutoCOPD) using quantitative inspiratory CT reliably identifies COPD across diverse cohorts; (2) a US claims-based matched cohort shows substantially elevated short- and long-term mortality after lower respiratory tract illness, even for ambulatory cases; and (3) a pediatric sepsis study integrates early endothelial biomarkers with clinical data to predict persistent acute respiratory dysfunction by day
Summary
Three impactful studies advance respiratory health: (1) a multicenter machine-learning model (AutoCOPD) using quantitative inspiratory CT reliably identifies COPD across diverse cohorts; (2) a US claims-based matched cohort shows substantially elevated short- and long-term mortality after lower respiratory tract illness, even for ambulatory cases; and (3) a pediatric sepsis study integrates early endothelial biomarkers with clinical data to predict persistent acute respiratory dysfunction by day 3.
Research Themes
- AI-enabled respiratory diagnostics using quantitative imaging
- Short- and long-term outcomes after lower respiratory tract infections
- Biomarker-driven risk stratification in pediatric critical care
Selected Articles
1. Derivation and Validation of a Clinical and Endothelial Biomarker Risk Model to Predict Persistent Pediatric Sepsis-Associated Acute Respiratory Dysfunction.
In prospectively enrolled children with septic shock, machine-learning models that incorporate day-1 endothelial biomarkers and clinical variables predicted persistent sepsis-associated acute respiratory dysfunction on day 3. Children with day-3 dysfunction had higher mortality, longer ventilation, and longer PICU stays; CART models validated in holdout and independent cohorts identified early predictors.
Impact: Provides a validated, early risk stratification tool for persistent respiratory dysfunction in pediatric sepsis by combining endothelial biology with clinical data, enabling targeted interventions and trial enrichment.
Clinical Implications: Supports early identification of high-risk children with sepsis for closer monitoring, timely ventilatory strategies, and potential biomarker-guided therapies; facilitates enrichment strategies in interventional trials targeting pediatric respiratory failure.
Key Findings
- Day-1 endothelial biomarkers plus clinical variables predicted day-3 persistent sepsis-associated acute respiratory dysfunction using TreeNet and CART.
- Children with day-3 respiratory dysfunction had higher mortality, longer mechanical ventilation, and longer PICU length of stay.
- Models validated in both holdout and an independent test cohort, demonstrating reproducibility.
Methodological Strengths
- Prospective enrollment with multicenter derivation and independent test cohort
- Holdout validation and external testing of ML models (TreeNet/CART)
Limitations
- Single-center external test cohort may limit generalizability
- No prospective implementation trial to assess impact on outcomes
Future Directions: Prospective, multi-center impact studies to test biomarker-guided care pathways; assess integration with ventilatory strategies and anti-endothelial therapies; calibration across varied PICU settings.
BACKGROUND: Sepsis-associated ARDS results in high morbidity and mortality in children. However, heterogeneity among patients makes identifying those at risk of persistent acute respiratory dysfunction challenging. Endothelial dysfunction is a key feature of ARDS pathophysiologic characteristics, contributing to lung injury in sepsis. Incorporating endothelial biomarkers into risk models may enhance prediction of those with persistent acute respiratory dysfunction. RESEARCH QUESTION: Can clinical variables and endothelial biomarkers measured early in the course of sepsis predict risk of persistent acute respiratory dysfunction among critically ill children? STUDY DESIGN AND METHODS: This was a multicenter derivation and single center test cohort study of prospectively enrolled children with sepsis. The derivation cohort was split into training and holdout validation sets. We trained TreeNet (Minitab, LLC) and classification and regression tree (CART) models using clinical and endothelial biomarkers measured on day 1 of septic shock to predict risk of sepsis-associated acute respiratory dysfunction (SA ARD) on day 3. The performance of the CART model was tested in the holdout validation data set and in the independent test cohort. RESULTS: In the derivation (n = 625) and test (n
2. AutoCOPD-A novel and practical machine learning model for COPD detection using whole-lung inspiratory quantitative CT measurements: a retrospective, multicenter study.
Using only ten quantitative features from inspiratory whole-lung CT, AutoCOPD achieved AUCs ≈0.86–0.92 across internal and multiple external validations, including robustness on low-dose CT (NLST). Decision curve analysis supports clinical utility across risk thresholds, suggesting feasibility for early detection of mild or asymptomatic COPD.
Impact: Delivers a practical, validated ML tool for COPD detection using widely available inspiratory CT, potentially reducing underdiagnosis and enabling earlier intervention across varied clinical settings.
Clinical Implications: AutoCOPD could be integrated into CT workflows (including lung cancer screening LDCT) to flag at-risk patients for spirometry and management, improving early COPD identification and care pathways.
Key Findings
- AutoCOPD achieved internal AUC 0.860 and external AUCs 0.903–0.915 across three hospital cohorts; NLST LDCT validation AUC 0.881.
- Only ten inspiratory QCT features were required, enabling a parsimonious and practical model.
- Decision curve analysis demonstrated net clinical benefit across COPD risk thresholds (0.12–0.66).
Methodological Strengths
- Large multicenter derivation with three external validations plus NLST cohort
- Transparent model with limited features and decision curve analysis for clinical utility
Limitations
- Retrospective design limits causal inference and may introduce selection bias
- Clinical impact on outcomes not yet tested in prospective implementation studies
Future Directions: Prospective implementation trials assessing care pathways, spirometry uptake, and outcomes; calibration across scanners/vendors and diverse populations; integration into screening programs.
BACKGROUND: The rate of diagnosis for chronic obstructive pulmonary disease (COPD) is low worldwide. Quantitative computed tomography (QCT) parameters add value to quantify alterations in airway and lung parenchyma for COPD. This study aimed to assess the performance of QCT features in COPD detection using a whole-lung inspiratory CT model. METHODS: This multicenter retrospective study was performed on 4106 participants. The derivation cohort containing 1950 participants who enrolled in Guangzhou communities from August 2017 to December 2019, was separated for training and internal validation cohorts, and three external validation cohorts containing 1703 participants were recruited from the public hospitals (Cohort 1: the First Affiliated Hospital of Guangzhou Medical University; Cohort 2: Xiangyang central hospital; Cohort 3: the Second Affiliated Hospital of Xi'an Jiaotong University) in China between April 2017 and May 2024. Questionnaire information, CT reports, and QCT features derived from inspiratory CT were extracted for model development. A novel multimodal framework using eXtreme gradient boosting and hybrid feature selection was established for COPD detection. National Lung Screening Trial (NLST) cohort (n = 453) was applied to validate the multiracial extrapolation and robustness on low-dose CT scans. FINDINGS: The QCT model (referred to as AutoCOPD) with ten features achieved the highest AUC of 0·860 (95% CI: 0·823-0·898) in the internal validation cohort, and showed excellent discrimination when externally validated [Cohort 1: AUC = 0·915 (95% CI: 0·898-0·931); Cohort 2: AUC = 0·903 (95% CI: 0·864-0·943); Cohort 3: AUC = 0·914 (95% CI: 0·882-0·947); NLST: AUC = 0·881 (95% CI: 0·846-0·915)]. Decision curve analysis demonstrated that AutoCOPD was valuable across a range of COPD risk thresholds between 0·12 and 0·66 compared with intervention in all patients with COPD or no intervention. INTERPRETATION: Heterogeneous COPD can be well identified using AutoCOPD (https://lwj-lab.shinyapps.io/autocopd/) constructed by a subset of only ten QCT features. It may be generalizable across clinical settings and serve as a feasible tool for early detecting patients with mild or asymptomatic COPD to reduce delayed diagnosis in routine practice. FUNDING: The National Natural Science Foundation of China, Guangzhou Laboratory, Natural Science Foundation of Guangdong Province, Guangzhou Municipal Science and Technology grant, State Key Laboratory of Respiratory Disease.
3. Short-term and Long-term Mortality Following Hospitalized and Ambulatory Lower Respiratory Tract Illnesses Among US Adults.
In >2.46 million US adults with LRTI, 30-day and 360-day all-cause mortality were markedly higher than matched controls for both hospitalized and ambulatory cases, with risk escalating with age and comorbidity. Findings underscore the need for prevention (e.g., vaccination) and post-episode risk management.
Impact: Quantifies substantial short- and long-term mortality burden after LRTI across care settings, informing vaccination policies, resource planning, and follow-up strategies for high-risk patients.
Clinical Implications: Prioritize prevention (influenza, pneumococcal, RSV vaccines), early detection of deterioration, and targeted post-LRTI follow-up for older and comorbid adults to reduce excess mortality.
Key Findings
- Hospitalized LRTI: 30-day mortality 5.8% and 360-day 18.3% (7.5× and 2.6× vs matched controls).
- Ambulatory LRTI: 30-day mortality 1.2% and 360-day 3.6% (6.5× and 2.1× vs controls).
- Mortality risks increased with age and were higher in adults with chronic or immunocompromising conditions.
Methodological Strengths
- Very large matched-cohort across inpatient and ambulatory settings with multiple time horizons
- Robust subgroup analyses by age and comorbidity
Limitations
- Claims-based retrospective design with potential misclassification and residual confounding
- Cause-specific attribution of LRTI etiology not available
Future Directions: Evaluate interventions to mitigate post-LRTI mortality (e.g., vaccination coverage, post-acute follow-up protocols) and validate in other health systems; explore pathogen-specific effects.
BACKGROUND: Lower respiratory tract illness (LRTI) is a significant cause of morbidity among adults, particularly older adults and adults with underlying medical conditions. Evidence on short- and long-term risks of mortality among adults requiring hospitalization or ambulatory care for LRTI, overall and within subgroups, is currently lacking. METHODS: A retrospective observational matched-cohort design and Optum's de-identified Clinformatics Data Mart Database (2012-2019) were used. The study population included adults who were hospitalized or received ambulatory care for LRTI and matched (1:1) comparison patients. All-cause mortality was ascertained during the 30-, 60-, 90-, 180-, and 360-day periods following the beginning of the LRTI episode. Risks of mortality were estimated for all LRTI patients and comparison patients as well as within age/comorbidity-specific subgroups. RESULTS: Among LRTI-hospitalized patients (n = 60.2K), 30-day mortality risk was 5.8% and 360-day risk was 18.3%, 7.5 and 2.6 times higher than corresponding values for comparison patients. Among LRTI-ambulatory patients (n = 2.4M), 30-day mortality risk was 1.2% and 360-day risk was 3.6%, 6.5 and 2.1 times higher than comparison patients. Among both LRTI-hospitalized and LRTI-ambulatory patients, mortality risk increased with increasing age and was higher for adults with chronic or immunocompromising conditions (vs without medical conditions). CONCLUSIONS: Short- and long-term mortality were higher among patients who were hospitalized or received ambulatory care for LRTI vs matched comparison patients, and risks increased markedly with increasing age and worsening comorbidity profile. Strategies for preventing LRTI, especially among persons at elevated risk, may reduce premature deaths and yield important public health benefits.