Pierluigi Ciet

Background: Although Elexacaftor–tezacaftor–ivacaftor (ETI) improves lung function and reduces exacerbations in cystic fibrosis (CF), airway inflammation and bronchial wall thickening do not completely resolve. Characterizing regional heterogeneity in bronchial wall thickness before and after ETI may provide insight into residual disease. We investigated whether bronchial wall thickening during ETI reflects ongoing inflammation, mucus burden, and infection at the segmental level. Methods: In this prospective observational cohort study, paired pre-ETI and 1.5-year post-ETI chest CT scans and bronchoalveolar lavage (BAL) samples were analyzed from five preselected lung segments in nine adults with CF. An AI-based pipeline generated automated segmental bronchus–artery metrics, including bronchial wall area (B _wa /B _oa ), wall thickness (B _wt /A), inner diameter (B _in /A), and outer diameter (B _out /A) ratios, as well as mucus plug number and volume. These metrics were correlated with segment-matched BAL markers of inflammation, mucus burden, and Pseudomonas aeruginosa infection. Results: Responses to ETI were heterogeneous both between and within subjects. Across 176 segments, 91.3% ( sd 19.1) showed reductions in B _wa /B _oa , 86.7% ( sd 18.4) reductions in B _wt /A, and 77.1% ( sd 20.9) increases in B _in /A, whereas B _out /A remained stable. However, differences in the magnitude of these changes were observed at the segmental level within subjects. Pre-ETI, bronchial wall thickness was most strongly associated with neutrophil elastase. Post-ETI, this association weakened but persisted, while sialic acid emerged as an additional predictor of residual wall thickness. P. aeruginosa burden was also associated with wall thickness post-ETI but explained less of the variance than inflammatory and mucus markers. Conclusions: ETI produces significant but variable improvements at the segmental level in bronchial dimensions. Persistent segmental heterogeneity suggests that local inflammation, mucus retention, and infection continue to shape residual structural lung disease. Automated segmental CT analysis may enable more individualized monitoring and management in the era of CFTR modulation.

Objective: Accurate imaging is essential for assessing structural lung abnormalities in children with bronchopulmonary dysplasia (BPD). Chest MRI offers a radiation-free alternative to CT and enables silent scanning. This study aimed to develop and evaluate a chest MRI protocol for BPD using a standard 1.5-Tesla (T) MRI scanner with a dedicated neonatal chest coil. Materials and methods: In this prospective pilot study, infants underwent feed-and-swaddle chest MRI at ~40 weeks postmenstrual age, and chest CT at six months corrected age. The MRI protocol included free-breathing axial T2-weighted (T2-W) fast spin echo (PROPELLER) and axial proton density-weighted (PD-W) gradient zero echo time (ZTE) sequences. Scans were assessed for image quality, and quantified for normal, hypo-/hyperintense lung tissue and bronchopathy. MRI and CT scores were correlated using Pearson’s or Spearman’s coefficients, based on data distribution. Results: Eight infants participated (seven preterm < 28 weeks’ gestation with severe BPD; one term neonate with asphyxia). T2-w PROPELLER provided superior soft tissue contrast and fewer artefacts than ZTE. ZTE enabled silent scanning and better visualisation of hypointense structures. Quantitative scores were comparable between PROPELLER and ZTE sequences. Significant correlations were found between MRI and CT scores for normal and hyperintense lung tissue ( p < 0.05; r = 0.84–0.96), but not for hypointense regions and bronchopathy. Conclusion: We developed a feasible and safe chest MRI protocol for imaging severe BPD-related lung abnormalities in neonates using a standard 1.5-T system. While technically promising, MRI is not yet clinically equivalent to CT. Further validation is needed to define its potential role in BPD assessment. Key Points: Question Can a standard 1.5 T chest MRI protocol performed without anaesthesia reliably image lung abnormalities in neonates with BPD? Findings T2-W PROPELLER sequence provided superior image quality and fewer artefacts compared to the ZTE, especially in depicting fissures and bronchi . Clinical relevance Our BPD-MRI protocol enables safe, radiation-free lung imaging in neonates with severe BPD without requiring anaesthesia. It offers a technically feasible alternative to CT and may support early structural assessment before discharge in routine clinical care .

Introduction: Infants with severe bronchopulmonary dysplasia (BPD) are at increased risk of hypoxemia and often fail hypoxic challenge testing (HCT). The predictive value of clinical characteristics and structural lung abnormalities on HCT outcomes is unclear. Methods: This prospective cohort study included preterm‐born infants (≤ 32 weeks of gestation) with severe BPD enrolled in a standardized follow‐up program. Perinatal and neonatal data were extracted from medical records. At 6 months corrected gestational age, structural lung abnormalities were quantified on chest CT using PRAGMA‐BPD scoring. HCT was performed using a sealed body plethysmograph and failure was defined as inability to maintain oxygen saturation ≥ 85% for 20 min. Univariate and multivariable logistic regression analyses identified predictors of HCT failure. Results: Among 156 infants, 28.8% failed HCT. Of the perinatal and neonatal factors, patent ductus arteriosus (odds ratio 2.73, 95% confidence interval 1.29–6.14), pulmonary hypertension (2.57, 0.99–6.62), and longer duration of supplemental oxygen therapy (odds ratio per interquartile range increase [OR _IQR ] 2.40, 1.55–3.71) were associated with failure. For structural lung abnormalities, higher composite PRAGMA‐BPD scores (OR _IQR 2.18, 1.36–3.48), higher hyper‐attenuation (OR _IQR 1.75, 1.17–2.63), and hypo‐attenuation (OR _IQR 1.45, 1.08–1.98) scores predicted failure. In multivariable analysis, only longer duration of supplemental oxygen therapy (OR _IQR 2.04, 1.22–3.42) and higher composite PRAGMA‐BPD scores (OR _IQR 1.95, 1.15–3.31) remained independent predictors. Conclusion: Duration of supplemental oxygen therapy and structural lung abnormalities independently predict HCT failure, highlighting the clinical relevance of both structural and functional impairments in severe BPD.

Artificial intelligence (AI) has potential to revolutionize radiology, yet current solutions and guidelines are predominantly focused on adult populations, often overlooking the specific requirements of children. This is important because children differ significantly from adults in terms of physiology, developmental stages, and clinical needs, necessitating tailored approaches for the safe and effective integration of AI tools. This multi-society position statement systematically addresses four critical pillars of AI adoption: (1) regulation and purchasing, (2) implementation and integration, (3) interpretation and post-market surveillance, and (4) education. We propose pediatric-specific safety ratings, inclusion of datasets from diverse pediatric populations, quantifiable transparency metrics, and explainability of models to mitigate biases and ensure AI systems are appropriate for use in children. Risk assessment, dataset diversity, transparency, and cybersecurity are important steps in regulation and purchasing. For successful implementation, a phased strategy is recommended, involving early pilot testing, stakeholder engagement, and comprehensive post-market surveillance with continuous monitoring of defined performance benchmarks. Clear protocols for managing discrepancies and adverse incident reporting are essential to maintain trust and safety. Moreover, we emphasize the need for foundational AI literacy courses for all healthcare professionals which include pediatric safety considerations, alongside specialized training for those directly involved in pediatric imaging. Public and patient engagement is crucial to foster understanding and acceptance of AI in pediatric radiology. Ultimately, we advocate for a child-centered framework for AI integration, ensuring that the distinct needs of children are prioritized and that their safety, accuracy, and overall well-being are safeguarded.