Daily Respiratory Research Analysis
Real-world evidence shows that nirsevimab substantially reduces RSV-related healthcare utilization in infants, while RSV vaccines for older adults are effective and rare serious safety signals were infrequent. A deep learning CT framework (PVDNet) accurately distinguishes pulmonary artery sarcoma from pulmonary thromboembolism, potentially preventing misdiagnosis. In Malawi, a prospective cohort revealed that acute breathlessness is highly multifactorial with high one-year mortality, supporting
Summary
Real-world evidence shows that nirsevimab substantially reduces RSV-related healthcare utilization in infants, while RSV vaccines for older adults are effective and rare serious safety signals were infrequent. A deep learning CT framework (PVDNet) accurately distinguishes pulmonary artery sarcoma from pulmonary thromboembolism, potentially preventing misdiagnosis. In Malawi, a prospective cohort revealed that acute breathlessness is highly multifactorial with high one-year mortality, supporting integrated diagnostics (e.g., BNP/NT-proBNP, CRP) and care pathways.
Research Themes
- RSV prevention effectiveness and safety in real-world settings
- AI-enabled diagnostic differentiation of pulmonary vascular diseases
- Integrated management of acute breathlessness in resource-limited settings
Selected Articles
1. Real-world effectiveness and safety of nirsevimab, RSV maternal vaccine and RSV vaccines for older adults: a living systematic review and meta-analysis.
Across 50 studies including ~7.6 million individuals, nirsevimab reduced RSV-related ED visits and hospitalizations by ~81% and ICU admissions by ~76%, with no severe safety signals. RSV vaccines for older adults reduced RSV-related hospital admissions by ~80%; serious adverse events (e.g., Guillain–Barré syndrome) were rare. Effectiveness data for maternal RSV vaccination remain limited.
Impact: This living synthesis provides timely, large-scale real-world confirmation of RSV prophylaxis effectiveness and safety, directly informing immunization policies and implementation decisions.
Clinical Implications: Supports widespread use of nirsevimab for infant RSV prevention and continued deployment of older adult RSV vaccines, with ongoing pharmacovigilance. Highlights evidence gaps for maternal RSV vaccination to guide targeted studies.
Key Findings
- Nirsevimab effectiveness: 80.7% against ED visits (95% CI 75.7–85.7), 80.7% against hospitalizations (95% CI 76.1–85.2), and 75.6% against ICU admissions (95% CI 63.3–87.9).
- Older adult RSV vaccines reduced RSV-related hospitalizations by 79.6% (95% CI 73.8–85.3).
- Safety: no severe adverse events for nirsevimab; RSV vaccines in older adults had <10 Guillain–Barré syndrome cases per million doses.
- No real-world effectiveness data available for RSV maternal vaccine; safety evidence limited.
Methodological Strengths
- Living systematic review with regular updates and broad coverage (~7.6 million participants).
- Random-effects meta-analyses and stratified assessments (age, outcome types) increase robustness.
Limitations
- Heterogeneity across observational designs and healthcare systems; residual confounding possible.
- Limited effectiveness and safety data for maternal RSV vaccination; potential publication bias.
Future Directions: Expand real-world effectiveness studies for maternal RSV vaccines and longer-term safety surveillance for all RSV immunizations; head-to-head comparisons and cost-effectiveness analyses to guide programmatic choices.
BACKGROUND: The long-acting monoclonal antibody nirsevimab and respiratory syncytial virus (RSV) vaccines became available for prevention of severe RSV-associated disease in 2023. While clinical trials showed good efficacy and safety, their restrictive inclusion criteria, small sample sizes and short follow-up limit generalisability. We aimed to summarise real-world evidence on the effectiveness and safety of nirsevimab, RSV maternal vaccine and RSV vaccines for older adults. METHODS: A living systematic review and meta-analysis, with 5 monthly updated searches in three databases was performed. Eligible studies were published from 1 December 2022 to 10 March 2025. Meta-analyses for the effectiveness of nirsevimab and RSV vaccines were carried out using random-effects model. Safety data were summarised narratively. RESULTS: A total of 50 publications, covering approximately 7.6 million people, were included. Nirsevimab showed 80.7% effectiveness (95% CI: 75.7% to 85.7%; seven studies) against RSV-related emergency department visits, 80.7% (95% CI: 76.1% to 85.2%; 17 studies) against hospital admissions and 75.6% (95% CI: 63.3% to 87.9%; eight studies) against intensive care unit admissions. The effectiveness of RSV vaccines for older adults against RSV-related hospital admissions was 79.6% (95% CI: 73.8% to 85.3; three studies). No effectiveness data were available for RSV maternal vaccine. No severe adverse events were reported for nirsevimab, while RSV vaccines in older adults had fewer than 10 Guillain-Barré syndrome cases per million doses. No severe adverse events were reported for RSV maternal vaccine, although evidence was limited. CONCLUSIONS: Our review demonstrated high effectiveness of nirsevimab in reducing RSV-related healthcare utilisation in infants and a favourable safety profile. More evidence is needed for evaluating RSV vaccines in pregnant people and older adults. PROSPERO REGISTRATION NUMBER: CRD42025643585.
2. Acute breathlessness as a cause of hospitalisation in Malawi: a prospective, patient-centred study to evaluate causes and outcomes.
In a multicenter Malawian cohort, 44% of acutely admitted adults had breathlessness and experienced a 1-year mortality of 51% versus 26% without breathlessness (adjusted HR 1.8). Heart failure, anemia, pneumonia, and tuberculosis were prevalent with substantial mortality, and 63% had multiple concurrent conditions. BNP/NT-proBNP (AUC 0.89/0.88) and CRP (AUC 0.77) showed strong diagnostic performance for heart failure and pneumonia, respectively.
Impact: This study reframes acute breathlessness as a multifactorial syndrome in LMIC settings, quantifies high mortality, and validates practical biomarkers, guiding integrated diagnostic and treatment pathways.
Clinical Implications: Implement integrated care bundles for breathlessness that include early cardiac (BNP/NT-proBNP), infection (CRP/PCT, TB testing), and anemia evaluation, coupled with context-adapted therapeutics and follow-up to reduce mortality.
Key Findings
- Breathlessness present in 44% (334/751); 1-year mortality 51% vs 26% without breathlessness (adjusted HR 1.8, 95% CI 1.4–2.3).
- High prevalence and mortality: heart failure 35% prevalence with 69% 1-year mortality; anemia 40% with 57% mortality; pneumonia 41% with 53% mortality; tuberculosis 29% with 47% mortality.
- Multimorbidity common: 63% had multiple concurrent conditions.
- Diagnostic performance: BNP AUC 0.89 and NT-proBNP AUC 0.88 for heart failure; CRP AUC 0.77 and PCT AUC 0.69 for pneumonia.
Methodological Strengths
- Prospective, multicenter design with 1-year follow-up and enhanced diagnostic screening.
- Use of standardized biomarker evaluation (BNP/NT-proBNP, CRP/PCT) with AUC reporting.
Limitations
- Single-country LMIC setting may limit generalizability to other regions.
- Some diagnostic tests may be constrained by resource availability; potential misclassification.
Future Directions: Randomized or stepped-wedge evaluations of integrated breathlessness care bundles in LMICs; implementation research on biomarker-guided triage and treatment; scalable diagnostic pathways.
INTRODUCTION: Breathlessness is a common cause of hospital admission globally and is associated with high mortality, particularly in low-income countries. In sub-Saharan Africa, there is a paucity of data on breathlessness, with existing data focused on individual diseases. There is a need for patient-centred approaches to understand interactions between multiple conditions to address population needs and inform health system responses. This multicentre prospective study in Malawi aimed to characterise the aetiologies, outcomes and biomarker accuracy for breathless patients. METHODS: Adults (aged ≥18 years) admitted to medical wards were consecutively recruited within 24 hours of hospital presentation and followed up for 1 year. Participants with breathlessness (defined as a composite of patient-reported shortness of breath; tachypnoea (respiratory rate ≥25/min); hypoxaemia (SpO RESULTS: Of 751 participants, 44% (n=334) had breathlessness, and 316 underwent enhanced diagnostic screening. One-year mortality was higher in breathless patients (51% (157/307)) than those without (26% (100/385)); adjusted HR 1.8 (95% CI 1.4 to 2.3). We identified high prevalence and mortality of heart failure (35% (112/316) prevalence; 69% (75/109) 1-year mortality), anaemia (40% (126/316); 57% (70/122)), pneumonia (41% (131/316); 53% (68/129)) and tuberculosis (29% (91/316); 47% (41/87)). Most participants (63% (199/316)) had multiple conditions. Diagnostic accuracy (area under the curve) for heart failure was 0.89 (brain natriuretic peptide) and 0.88 (N-terminal pro-B-type natriuretic peptide); for pneumonia, CRP was 0.77 and PCT was 0.69. DISCUSSION: Breathlessness-related hospital admissions in Malawi are common, multifactorial and associated with poor survival. This study demonstrates that co-existing conditions are common, highlighting the limitation of single-disease-focused health system responses. Integrated care pathways with context-sensitive diagnostic and treatment approaches are urgently needed to improve survival.
3. Developing a deep learning-based imaging diagnostic framework, PVDNet, for differentiating pulmonary artery sarcoma and pulmonary thromboembolism: a multi-center observational study.
Using 952 CTPA cases from 15 hospitals, PVDNet distinguished PAS from PTE with AUC 0.972 internally and 0.973 in external validation, matching a senior subspecialist radiologist (p=0.308). Agreement with the senior reader was the highest (kappa 0.651). Performance for APE vs CPE classification was lower (AUC ~0.90) and warrants refinement.
Impact: Accurate differentiation between PAS and PTE is clinically critical yet challenging; this externally validated DL model approaches expert performance and could reduce misdiagnosis and delays in definitive therapy.
Clinical Implications: Integrating PVDNet into CTPA workflows could flag suspected PAS for urgent multidisciplinary review, expedite surgical referral or biopsy, and avoid inappropriate anticoagulation when malignancy is likely.
Key Findings
- Internal test AUCs: PAS 0.972 (95% CI 0.945–0.994), APE 0.902 (95% CI 0.855–0.944), CPE 0.900 (95% CI 0.852–0.946).
- External validation: PAS vs PTE AUC 0.973, comparable to senior subspecialist radiologist (0.943; p=0.308); highest agreement with SRPV (kappa 0.651, p<0.001).
- Fine-grained classification enabled differentiation across PAS, APE, and CPE; APE vs CPE performance requires further optimization.
Methodological Strengths
- Large, multicenter dataset with external validation across 12 centers.
- Direct comparison with radiologists of varying expertise; reporting of AUCs and inter-rater agreement.
Limitations
- Retrospective imaging analysis; clinical impact not yet tested in prospective workflows.
- Generalizability to different scanners/protocols and non-Chinese populations needs further study; APE vs CPE classification suboptimal.
Future Directions: Prospective clinical utility trials, domain adaptation across scanners/sites, calibration with clinical data, and active learning to improve APE vs CPE discrimination.
BACKGROUND: Differentiating pulmonary artery sarcoma (PAS) from pulmonary thromboembolism (PTE) based on CT pulmonary angiography (CTPA) is a big challenge, necessitating the incorporation of other methods, such as deep learning (DL). This study aimed to develop and validate a DL-based model, PVDNet, for differentiating PAS and PTE on CTPA. METHODS: This study retrospectively analyzed CTPA image datasets from the prospective CHinese pulmOnary embolism multimodality Imaging artifiCial intelligencE (CHOICE) study to develop and validate a DL model for differentiating PAS from PTE. CTPA image datasets of 952 patients (470 acute PTE [APE], 363 chronic PTE [CPE], and 119 PAS) from 15 hospitals were included. The training set comprised CTPA images from 590 patients, and the internal test set comprised those from 186 patients, all obtained from the same three centers. CTPA images of 176 patients from 12 centers were used for external validation. A DL framework, PVDNet, was employed to perform fine-grained classification. Meanwhile, CTPA images in the external validation set were independently assessed by four radiologists with different levels of expertise. The main outcome measures were area under the curve (AUC) and the consistency test. FINDINGS: In the internal test set, PVDNet achieved an AUC of 0.972, 0.902, and 0.900 for PAS (95% CI: [0.945, 0.994]), APE (95% CI: [0.855, 0.944]), and CPE (95% CI: [0.852, 0.946]), respectively. Furthermore, PVDNet model demonstrated effective differentiation between PAS and PTE, showing comparable AUC values to a senior radiologist specialized in pulmonary vascular diseases (SRPV) in the external validation set (0.973 vs. 0.943, p = 0.308). The model achieved moderate agreement with SRPV (kappa = 0.651, p < 0.001), which was the highest among four readers. INTERPRETATION: PVDNet model could differentiate PAS from PTE, with performance approaching the proficiency level of a senior radiologist specializing in pulmonary vascular diseases. PVDNet's performance in distinguishing APE from CPE requires further optimization. FUNDING: National Natural Science Foundation of China (No. 82272081), Medical and Health Science and Technology Innovation Project of Chinese Academy of Medical Science (No. 2021-I2M-1-049), and the National Key Research and Development Program of China (No. 2023YFC2507200).