Daily Respiratory Research Analysis
Analyzed 50 papers and selected 3 impactful papers.
Summary
Three high-impact respiratory studies stood out today: a multicenter pathomics model that predicts which lung squamous cell carcinoma patients benefit from first-line chemo-immunotherapy, a UK-wide ICU study revealing that many patients with known or predicted difficult airways are not clearly identified or planned for, and a real-world metagenomics diagnostic study showing markedly higher sensitivity than standard methods for lower respiratory tract infections.
Research Themes
- Precision oncology and treatment selection
- Airway safety and critical care quality
- Metagenomic diagnostics for respiratory infections
Selected Articles
1. A pathomics model for predicting response to chemo-immunotherapy in lung squamous cell carcinoma: A multicenter study.
Using whole-slide pathology images, the authors built a pathomics score that predicts a T cell–inflamed GEP and identifies lung squamous cell carcinoma patients who derive superior PFS and OS from first-line chemo-immunotherapy compared with chemotherapy. The treatment–PS interaction was significant in a prospective multicenter trial and replicated in two external cohorts, with high-PS tumors corresponding to an immune-hot microenvironment.
Impact: This study operationalizes precision immuno-oncology from routine pathology, enabling treatment selection without expensive molecular assays. It shows robust multicenter validation and clinically meaningful survival interaction.
Clinical Implications: Pathomics-based stratification could guide first-line CIT use in LUSC by identifying patients with immune-hot biology likely to benefit, potentially reducing overtreatment and optimizing resource use when molecular profiling is unavailable.
Key Findings
- Pathomics score predicted T cell–inflamed GEP with AUC 0.80 (training) and 0.71 (validation) in TCGA-LUSC.
- Significant PS×treatment interactions in AK105-302 for PFS (p=0.011) and OS (p<0.001), with high-PS patients benefiting from CIT (PFS HR 0.31; OS HR 0.30 versus chemotherapy).
- Findings replicated in two independent cohorts (n=82 and n=50), and high PS associated with an immune-hot tumor microenvironment.
Methodological Strengths
- Prospective multicenter validation with formal treatment–biomarker interaction testing.
- External replication across independent clinical cohorts and linkage to immune microenvironment biology.
Limitations
- Development used retrospective TCGA data; real-world implementation and workflow integration need prospective utility studies.
- Follow-up duration and thresholds for clinical deployment not detailed in the abstract; generalizability beyond included cohorts remains to be shown.
Future Directions: Prospective, biomarker-stratified trials to test treatment selection using the pathomics score, head-to-head comparisons with PD-L1 and GEP assays, and integration into digital pathology pipelines.
BACKGROUND: The identification of lung squamous cell carcinoma (LUSC) patients who may benefit from first-line chemo-immunotherapy (CIT) remains a challenge. This study aimed to develop a pathomics model to predict the T cell-inflamed gene-expression profile (GEP) status and validate its utility in identifying patients who derive survival benefit from CIT. METHODS: The pathomics model was developed using whole-slide images and RNA-sequencing data from The Cancer Genome Atlas (TCGA) LUSC cohort (n = 334) to predict the GEP status and generate a pathomics score (PS). The predictive value of PS was validated in a prospective, multicenter trial (AK105-302, n = 267) by assessing its interaction with treatment (CIT vs. chemotherapy) for progression-free survival (PFS) and overall survival (OS). Two additional independent cohorts (n = 82 and n = 50) were used for external validation. RESULTS: The pathomics model accurately predicted GEP status, achieving area under the curve (AUC) values of 0.80 and 0.71 in the TCGA training and validation sets, respectively. In the AK105-302 cohort, significant interactions were identified between PS and treatment modalities for both PFS (interaction p = 0.011) and OS (interaction p < 0.001). Patients with high PS who received CIT exhibited significantly prolonged PFS (hazard ratio [HR]: 0.31, 95 % confidence interval [CI]: 0.21-0.48, p < 0.001) and OS (HR: 0.30, 95 % CI: 0.18-0.50, p < 0.001) compared to high PS patients receiving chemotherapy. However, this survival benefit was not observed in the low-PS patients. These findings were corroborated in two independent clinical cohorts. Furthermore, biological assessments revealed a significant association between high PS and an immune-hot tumor microenvironment. CONCLUSION: We developed a GEP-based pathomics model that provides a practical, cost-effective strategy to identify patients most likely to derive superior survival benefit from first-line CIT over chemotherapy alone.
2. Identification of difficult airways in critical care units (ID-ACCT): a multicentre, survey-based prospective observational study.
In 106 UK ICUs, 16.1% of patients had a known difficult airway, yet only 28.7% were clearly identifiable as at risk and 30.7% had a documented airway plan. Among those with predicted difficulty, identification (13.6%) and planning (19.6%) were even lower, indicating substantial safety and systems gaps.
Impact: This large, prospective multicenter study quantifies a critical gap in airway risk identification and planning in ICUs, directly informing quality improvement and patient safety initiatives.
Clinical Implications: Units should implement systematic difficult airway screening, standardized documentation, and proactive airway plans for at-risk patients, potentially via checklists, EHR flags, and multidisciplinary airway rounds.
Key Findings
- Known difficult airways were present in 16.1% (n=244/1519) of ICU patients with intubation details.
- Only 28.7% of known difficult airways were clearly identifiable to investigators and 30.7% had airway management plans.
- Among 922 patients with predicted difficulty, only 13.6% were clearly identifiable and 19.6% had plans, highlighting system-level gaps.
Methodological Strengths
- Prospective multicenter design across 106 ICUs with standardized bedside assessments.
- Clear operational definitions for known and predicted difficult airways.
Limitations
- Observational design cannot assess causality or patient outcomes from identification gaps.
- Investigator-led assessments may vary; generalizability outside the UK requires validation.
Future Directions: Test standardized ICU airway risk flags and care bundles in implementation trials to reduce complications; link identification and planning to patient-centered outcomes.
BACKGROUND: Airway-related complications are common in ICUs and cause serious patient harm. We determined the prevalence of known and predicted difficult airways in ICU patients in the UK, assessed whether these patients are identifiable as at risk, and evaluated the presence of airway management plans through local investigator-led bedside assessment. METHODS: We conducted a prospective, multicentre, observational study in 106 UK ICUs. Adult patients without treatment escalation plans precluding artificial airway insertion were included. Patients with known difficult airways (defined as difficulty in face-mask ventilation, laryngoscopy, or intubation) or predicted difficult airways (defined as BMI ≥30 kg m RESULTS: Amongst 1965 patients, 872 (44.4%) had an artificial airway in situ (tracheal tube n=639 [73.3%], tracheostomy tube n=233 [26.7%]). Among 1519 patients with intubation details, 244 (16.1%) had a known difficult airway (difficult face-mask ventilation n=32 [2.1%], laryngoscopy n=37 [2.4%], intubation n=182 [12.0%]). In this population, 70 (28.7%) were clearly identifiable to investigators, and 75 (30.7%) had airway management plans. Excluding patients with a known difficult airway, 922 patients had predicted difficult airway management, of whom 125 (13.6%) were clearly identifiable and 181 (19.6%) had airway management plans. CONCLUSIONS: Known and predicted difficult airways are common in UK ICU patients, but these at-risk patients are often not clearly identifiable to healthcare professionals.
3. Clinical metagenomics for pathogen detection in lower respiratory infections: a diagnostic study.
In a real-world retrospective cohort of 186 patients with respiratory samples, mNGS achieved an 81.2% positivity rate and higher sensitivity than culture and conventional testing for LRTIs, particularly excelling in mixed infection detection. Specificity was lower, underscoring the need for clinical adjudication to distinguish colonization from infection.
Impact: The study provides pragmatic evidence that clinical metagenomics can substantially enhance pathogen detection in LRTIs, especially for mixed infections, informing antimicrobial stewardship and early targeted therapy.
Clinical Implications: mNGS can complement culture and conventional assays to rapidly identify pathogens—including viruses, mycobacteria, and mixed infections—supporting earlier, targeted treatment, while requiring careful interpretation to avoid overtreatment due to lower specificity.
Key Findings
- Overall mNGS positivity was 81.2% (151/186); detection rates in LRTIs were bacteria 84.6%, fungi 89.0%, viruses 100%, M. tuberculosis 88.9%, and NTM 100%.
- mNGS detected mixed infections far more often than CMTs (91.3% vs 43.5%, P<0.01).
- Against a composite standard, mNGS sensitivity was higher than culture (89.0% vs 32.4%) and CMTs (89.0% vs 57.2%), but specificity was lower (46.3% vs 87.8% and 82.9%).
Methodological Strengths
- Head-to-head comparison with culture and conventional testing on the same samples using a composite diagnostic standard.
- Subgroup analyses by ventilation and immunosuppression status; comprehensive pathogen spectrum assessment.
Limitations
- Retrospective single-center design with potential selection and information biases.
- Lower specificity raises risk of detecting colonizers; clinical correlation is essential.
Future Directions: Prospective multicenter diagnostic accuracy and impact studies integrating mNGS into antimicrobial stewardship, with standardized clinical adjudication algorithms to address specificity.
BACKGROUND: Metagenomic next-generation sequencing (mNGS) enables comprehensive detection of all potential pathogens in a sample. However, its diagnostic performance in various clinical settings requires further validation through real-world data. METHODS: This retrospective study included 186 patients from the First Affiliated Hospital of Zhejiang University School of Medicine who underwent mNGS testing on respiratory samples. We compared mNGS with traditional culture method using the same samples and also assessed its performance against conventional microbiological testing combinations (CMTs) for the same patients. Additionally, we analyzed the diagnostic performance of mNGS in different disease states (immunosuppressive status and mechanical ventilation). RESULTS: The positivity rate of mNGS was 81.2 % (151/186). In lower respiratory tract infections (LRTIs), mNGS successfully detected 84.6 % (137/162) of bacteria, 89.0 % (65/73) of fungi, 100.0 % (72/72) of viruses, 88.9 % (16/18) of Mycobacterium tuberculosis, and 100.0 % (9/9) of non-tuberculous mycobacteria. Mixed infections were the most common infection type in LRTIs in this study (69/145, 47.6 %). The detection rate of mixed infections by mNGS was significantly higher than that of CMTs (91.3 % vs 43.5 %, P < 0.01). In the comparison based on a composite LRTI diagnostic standard, mNGS showed significantly higher sensitivity than paired culture (89.0 % vs 32.4 %, P < 0.01) and CMTs (89.0 % vs 57.2 %, P < 0.01), but lower specificity (46.3 % vs 87.8 % and 46.3 % vs 82.9 %, P < 0.01). In non-mechanically ventilated patients, mNGS maintained high sensitivity (87.0 % vs 96.7 %, P = 0.19), while culture and CMTs showed significant sensitivity decline (P < 0.01). CONCLUSION: mNGS demonstrates superior diagnostic performance for LRTIs compared to CMTs.