Daily Respiratory Research Analysis
Three impactful respiratory studies emerged today: a deep learning approach that derives parametric response mapping and fSAD from a single inspiratory CT, a multicenter ICU analysis showing two-thirds of switches from controlled to assisted ventilation fail with poor outcomes, and the first global description of post‑pandemic seasonality for hMPV and RSV. Together they advance diagnostics, ventilator management strategy, and viral surveillance for immunization planning.
Summary
Three impactful respiratory studies emerged today: a deep learning approach that derives parametric response mapping and fSAD from a single inspiratory CT, a multicenter ICU analysis showing two-thirds of switches from controlled to assisted ventilation fail with poor outcomes, and the first global description of post‑pandemic seasonality for hMPV and RSV. Together they advance diagnostics, ventilator management strategy, and viral surveillance for immunization planning.
Research Themes
- AI-enabled respiratory imaging and small airways disease detection
- Critical care ventilation strategy and weaning risks
- Global seasonality and circulation of respiratory viruses post-pandemic
Selected Articles
1. Single Inspiratory Chest CT-based Generative Deep Learning Models to Evaluate Functional Small Airways Disease.
This study develops predictive and generative deep learning models that infer PRM and fSAD using only a single inspiratory CT, also generating synthetic expiratory images. Trained on 308 individuals and validated on internal and external sets, voxelwise metrics showed strong performance, suggesting reduced need for dual-phase acquisitions.
Impact: By replacing paired inspiratory–expiratory CT PRM with single-inspiratory CT modeling, this approach can lower radiation exposure, simplify workflows, and scale fSAD diagnostics.
Clinical Implications: Potential to reduce dual-phase CT requirements for small airways assessment, enabling broader PRM use in COPD/asthma clinics and trials with lower radiation and logistical burden.
Key Findings
- Deep learning models predicted PRM-derived fSAD from a single inspiratory CT with strong voxelwise performance (e.g., AUC and sensitivity reported).
- Generative modeling produced expiratory-equivalent CT images with high structural similarity indices.
- Models trained on 308 subjects and validated across multiple internal and one external test set, supporting generalizability.
Methodological Strengths
- Use of voxelwise metrics and external validation to assess generalizability.
- Generative modeling to synthesize expiratory images from inspiratory CT.
Limitations
- Retrospective design; full sample sizes for all test sets not detailed in abstract.
- Clinical thresholds and impact on decision-making were not prospectively assessed.
Future Directions: Prospective clinical validation to quantify diagnostic yield, impact on management, and radiation savings; harmonization across scanners and reconstruction kernels.
Purpose To develop a deep learning model that uses a single inspiratory chest CT scan to perform parametric response mapping (PRM) and predict functional small airways disease (fSAD). Materials and Methods In this retrospective study, predictive and generative deep learning models for PRM using inspiratory chest CT were developed using a model development dataset with fivefold cross-validation, with PRM derived from paired respiratory CT as the reference standard. Voxelwise metrics, including sensitivity, area under the receiver operating characteristic curve (AUC), and structural similarity index measure, were used to evaluate model performance in predicting PRM and generating expiratory CT images. The best-performing model was tested on three internal test sets and an external test set. Results The model development dataset of 308 individuals (median age, 67 years [IQR: 62-70 years]; 113 female) was divided into the training set (
2. Post-Pandemic Dynamics of the Global Circulation of Human Metapneumovirus and Respiratory Syncytial Virus.
Across 26 countries (2022–2024), RSV consistently peaked before hMPV; global circulation suggested opposite directional patterns between RSV and hMPV, with rapid multi-country seeding windows. These insights enable data borrowing across geographies and inform timing for future immunization or prophylaxis.
Impact: It is the first global seasonality description for RSV and hMPV post‑pandemic, directly supporting surveillance, modeling, and vaccine program design.
Clinical Implications: Helps optimize timing of maternal/older-adult RSV prevention and future hMPV immunization strategies; supports regional planning where local seasonality data are sparse.
Key Findings
- In most countries, RSV peaks occur systematically before hMPV peaks.
- Directional global circulation patterns differ between RSV (clockwise) and hMPV (counterclockwise on Mercator projection).
- Rapid, near-synchronous multi-country seeding windows were observed with year-to-year variation.
- Global surveillance gaps were evident—only 26 countries had suitable public data.
Methodological Strengths
- Multi-country weekly surveillance synthesis with standardized activity and seeding definitions.
- Comparative analysis of RSV vs hMPV timing and circulation patterns.
Limitations
- Publicly available surveillance data limited to 26 countries, risking geographic bias.
- Laboratory testing practices and positivity thresholds may vary across settings.
Future Directions: Expand standardized global surveillance, integrate genomic data to resolve lineage-specific dynamics, and link seasonality to immunization policy optimization.
BACKGROUND: Understanding the seasonality of human metapneumovirus (hMPV) and respiratory syncytial virus (RSV) is important for public health planning. It can support rationale for using another country data to model immunization strategies where seasonality data are lacking. While some studies have investigated (sub)-national seasonality drivers, this is the first to describe global seasonality for RSV and hMPV. METHODS: We included 26 countries with consistent reporting and >10 detections at the peak, after searching international databases and local reports. Weekly surveillance data from January 2022 to June 2024 were included. Viral activity was defined by comparing the 4-week moving average of the positivity rate to its annual average. "Seeding" events were the first 2 consecutive weeks with a statistically significant increase in detections. RESULTS: Most countries showed seasonal patterns of RSV and hMPV, except for some tropical countries. The RSV peak occurred systematically before the hMPV peak. On a Mercator projection, hMPV appeared to circulate in a counterclockwise manner, opposite to RSV. Although global information was incomplete, the first seeding events occurred in a short time in multiple countries with year-to-year variations. CONCLUSIONS: We have provided critical information on the circulation of hMPV and RSV. We only found 26 countries reporting suitable surveillance data in publicly accessible reports, which likely reflects true gaps in surveillance.
3. Switching from controlled to assisted mechanical ventilation: a multi-center retrospective study (SWITCH).
Across three centers (n=6715 switch attempts), 67% of first attempts to transition from controlled to assisted ventilation failed and were linked to worse outcomes, despite similar baseline severity. LASSO-based prediction performed poorly, emphasizing prevention and timing rather than prediction.
Impact: This large, granular, multicenter analysis reframes weaning strategy: failures are common, prognostically important, and hard to predict, supporting protocolized prevention and prospective trials.
Clinical Implications: ICUs should minimize failed switches by structured readiness assessments, cautious stepwise trials, and early detection of intolerance; prioritize prospective protocols to define safe initiation of spontaneous breathing.
Key Findings
- Among 6715 switch attempts, 67% of first attempts failed, returning to controlled ventilation within 72 hours.
- Failed switches were associated with worse clinical outcomes despite similar baseline severity to successful attempts.
- Pre-switch and early post-switch features were similar, and LASSO logistic models struggled to predict failure.
- First switches were attempted at similar normalized PaCO2 levels regardless of outcome (per abstract).
Methodological Strengths
- Large, multicenter dataset with high temporal resolution of ventilatory and physiologic variables.
- Comparative analyses at multiple time anchors (ICU admission, immediately before, and 3 h after switch) and use of LASSO modeling.
Limitations
- Retrospective design limits causal inference; potential unmeasured confounding.
- Incomplete visibility into clinician decision rules and heterogeneity of practices across centers.
Future Directions: Prospective trials to define readiness criteria, stepwise protocols, and physiologic targets minimizing failure; integration of continuous waveform analytics and patient-ventilator synchrony metrics.
BACKGROUND: Switching from controlled to assisted ventilation is crucial in the trajectory of intensive care unit (ICU) stay, but no guidelines exist. We described current practices, analyzed patient characteristics associated with switch success or failure, and explored the feasibility to predict switch failure. METHODS: In this retrospective study, we obtained highly granular longitudinal ICU data sets from three medical centers, covering demographics, severity scores, vital signs, ventilation, and laboratory parameters. The primary endpoint was switch success, considering a switch attempt to be successful if a patient did not return to controlled ventilation for the next 72 h while alive, and to be failed otherwise. We compared the characteristics of patients with successful vs. failed first switch attempts at ICU admission, immediately before, and 3 h after the attempt. We trained LASSO logistic regression models to predict switch failure. RESULTS: In 4524/6715 (67%) patients attempting a switch, the first attempt failed. The first switch attempt, regardless of success or failure, was generally made at normalized PaCO CONCLUSIONS: Approximately two-thirds of attempts to switch patients to assisted ventilation fail, which are associated with significantly worse clinical outcomes, despite similar baseline disease severity. Contrary to our hypotheses, patients with successful and failed attempts showed similar characteristics, making switch failure difficult to predict. These findings underscore the importance of preventing switch failures and, given the retrospective nature of this study, highlight the need for prospective studies to better understand the reasons for switch failure and when spontaneous breathing can be safely initiated.