Daily Respiratory Research Analysis
Today’s top respiratory research advances feature AI-enabled prognostication after terminal extubation to optimize donation after circulatory death, an externally validated neonatal model that improves early prediction of bronchopulmonary dysplasia using dynamic postnatal factors, and prospective evidence that high-quality home spirometry can reliably track lung function in cystic fibrosis. These studies collectively highlight precision prediction and remote monitoring as maturing pillars of res
Summary
Today’s top respiratory research advances feature AI-enabled prognostication after terminal extubation to optimize donation after circulatory death, an externally validated neonatal model that improves early prediction of bronchopulmonary dysplasia using dynamic postnatal factors, and prospective evidence that high-quality home spirometry can reliably track lung function in cystic fibrosis. These studies collectively highlight precision prediction and remote monitoring as maturing pillars of respiratory care.
Research Themes
- AI-driven risk prediction and clinical decision support in respiratory care
- Remote monitoring and telehealth for chronic lung disease
- Early-life risk stratification and outcome forecasting in neonatal respiratory disorders
Selected Articles
1. Deep learning unlocks the true potential of organ donation after circulatory death with accurate prediction of time-to-death.
An ODE-RNN model trained on 3,238 ICU patients and externally validated on 1,908 patients accurately classified whether death would occur within 30 and 60 minutes after terminal extubation. Key predictors included heart rate, respiratory rate, MAP, SpO2, and GCS, and performance surpassed existing clinical scores, suggesting potential to streamline DCD logistics and improve transplant outcomes.
Impact: This study provides externally validated AI that addresses a key operational barrier in DCD by predicting time-to-death after extubation. The approach could reduce ischemic injury risk, optimize resource allocation, and increase usable organ yield.
Clinical Implications: Incorporation of the model into ICU workflows could enable real-time prognostication to guide DCD attempts, coordination with procurement teams, and mitigation of prolonged warm ischemia.
Key Findings
- ODE-RNN trained on 3,238 patients and externally validated on 1,908 across six hospitals.
- Accurately predicted whether death would occur within 30 and 60 minutes after terminal extubation with good calibration.
- Heart rate, respiratory rate, MAP, SpO2, and GCS were the most important predictors.
- Outperformed existing clinical scores for this task, indicating clinical utility.
Methodological Strengths
- Large, multi-center external validation enhances generalizability.
- Model tailored for irregularly sampled ICU time-series (ODE-RNN) with calibration assessment.
Limitations
- Retrospective observational design; prospective interventional evaluation absent.
- Exact performance metrics (numerical accuracies) not detailed in abstract.
Future Directions: Prospective, multi-center implementation studies assessing decision-impact, ethics, and workflow integration; evaluation across diverse ICUs and organ systems with cost-effectiveness and equity analyses.
Increasing the number of organ donations after circulatory death (DCD) has been identified as one of the most important ways of addressing the ongoing organ shortage. While recent technological advances in organ transplantation have increased their success rate, a substantial challenge in increasing the number of DCD donations resides in the uncertainty regarding the timing of cardiac death after terminal extubation, impacting the risk of prolonged ischemic organ injury, and negatively affecting post-transplant outcomes. In this study, we trained and externally validated an ODE-RNN model, which combines recurrent neural network with neural ordinary equations and excels in processing irregularly-sampled time series data. The model is designed to predict time-to-death following terminal extubation in the intensive care unit (ICU) using the history of clinical observations. Our model was trained on a cohort of 3,238 patients from Yale New Haven Hospital, and validated on an external cohort of 1,908 patients from six hospitals across Connecticut. The model achieved accuracies of [Formula: see text] and [Formula: see text] for predicting whether death would occur in the first 30 and 60 minutes, respectively, with a calibration error of [Formula: see text]. Heart rate, respiratory rate, mean arterial blood pressure (MAP), oxygen saturation (SpO2), and Glasgow Coma Scale (GCS) scores were identified as the most important predictors. Surpassing existing clinical scores, our model sets the stage for reduced organ acquisition costs and improved post-transplant outcomes.
2. Development and external validation of a machine learning model to predict bronchopulmonary dysplasia using dynamic factors.
In preterm infants <32 weeks, integrating early respiratory support, FiO2, and blood gas data within 7 days with perinatal factors significantly improved BPD prediction. The integrated model achieved AUROC 0.841 (development), 0.912 (internal validation vs 0.805 static, p<0.0001), and 0.814 (external validation).
Impact: Provides a validated, practically obtainable early-life prediction tool that could inform individualized respiratory care and trial enrollment for BPD prevention.
Clinical Implications: Early risk stratification may guide oxygen/ventilation strategies, nutrition, and targeted prophylaxis, and support timely counseling and follow-up in neonatal care units.
Key Findings
- Integrated model using static perinatal and dynamic postnatal factors reached AUROC 0.841 in development.
- Internal validation: AUROC 0.912 vs 0.805 for static model (p<0.0001).
- External validation at another center maintained performance (AUROC 0.814).
- Early respiratory support metrics and blood gas analyses within 7 days were influential predictors.
Methodological Strengths
- Internal and external validation across distinct cohorts.
- Incorporation of time-sensitive clinical variables likely to capture evolving disease biology.
Limitations
- Retrospective design from two centers; potential center-specific practices limit generalizability.
- Short early-life window only (first 7 days) may miss later dynamic signals.
Future Directions: Prospective, multi-center impact studies integrating the model into neonatal workflows; evaluate decision-impact, safety, and whether tailored interventions reduce BPD incidence or severity.
We hypothesized that incorporating postnatal dynamic factors would enhance the prediction accuracy of bronchopulmonary dysplasia in preterm infants. This retrospective cohort study included neonates born before 32 weeks of gestation at Seoul National University Hospital between 2013 and 2022. The primary outcome was moderate or severe bronchopulmonary dysplasia. We assessed both static perinatal risk factors and dynamic factors, such as respiratory support type, inspired oxygen concentration, and blood gas analysis results within the first 7 days. The model was developed using data from 546 infants born between 2013 and 2021, with internal validation on 75 infants born in 2022. External validation was based on 105 infants recruited at the Boramae Medical Center. The integrated prediction model, combining static and dynamic factors, showed superior predictive performance, with an area under the receiver operating characteristic curve (AUROC) of 0.841 in the development set, outperforming the static perinatal factor model. Internal validation confirmed the robustness of the integrated model (AUROC: 0.912 vs. 0.805, p < 0.0001). The performance was maintained in the external validation (AUROC: 0.814). Incorporating early respiratory support and blood gas analysis into predictive models substantially improved the accuracy of bronchopulmonary dysplasia prediction in preterm infants.
3. Remote monitoring of cystic fibrosis lung disease in children and young adults.
Across seven Swedish CF centers, 110 participants were followed for a median 12 months; 779 usable home spirometry sessions were recorded, and 50% were graded high-quality (ATS/ERS A–C) from 84% of eligible participants. When adjusted for clinical stability and antibiotic use, home spirometry provided lung function levels and trends comparable to hospital spirometry.
Impact: Demonstrates feasibility and quality of remote spirometry in children and young adults with CF, supporting telehealth-enabled personalized monitoring and timely treatment adjustments.
Clinical Implications: Clinics can integrate standardized home spirometry workflows to complement or reduce in-clinic testing, while using antibiotic logs and stability status to interpret trajectories and trigger interventions.
Key Findings
- Prospective multi-center cohort with 110 participants followed for a median of 12 months.
- A total of 779 usable home spirometry sessions; 50% graded as high-quality per ATS/ERS, from 84% of eligible participants ≥5 years.
- Home spirometry trends, adjusted for clinical stability and antibiotic use, were comparable to hospital spirometry.
Methodological Strengths
- Prospective, multi-center design with standardized quality grading (ATS/ERS).
- Integration of clinical stability and antibiotic use to contextualize lung function trends.
Limitations
- Abstract does not report exact numeric agreement metrics versus hospital spirometry.
- Potential selection/adherence bias in home measurements.
Future Directions: Define thresholds for action based on home spirometry, evaluate cost-effectiveness and equity, and automate quality control and alerts within telehealth platforms.
AIM: Cystic fibrosis (CF) care increasingly demands flexible and personalised approaches, particularly with the growing role of telemedicine in disease management. This study aimed to evaluate the use of home-based spirometry and antibiotic monitoring for assessing lung function trends and treatment patterns in individuals with CF. METHOD: Individuals aged 0-25 years from seven Swedish CF centres participated in 12 months of routine CF care, digitally recording antibiotic usage and performing home spirometry (aged ≥5 years). Home spirometry sessions were graded according to ATS/ERS criteria, with A-C representing high-quality sessions. Longitudinal FEV RESULTS: Of 126 invited participants, 110 were enrolled and followed for a median (range) duration of 12 months (9-17). A total of 779 usable home spirometry sessions were conducted, with 388 sessions (50 %) from 80 out of 95 (84 %) participants aged ≥5 years graded as high-quality. Mean (95 % CI) FEV CONCLUSION: High-quality home spirometry measurements, adjusted for clinical stability and antibiotic usage, may provide lung function levels and trends closely comparable to hospital spirometry.