Randomized trial identifies distinct cardiopulmonary phenotypes in very preterm children, suggesting tailored treatments.
Rationale Very preterm birth is linked to cardiopulmonary complications throughout childhood, yet individual trajectories beyond five years remain poorly understood. This study used xenon MRI to characterize lung phenotypes in 6-year-old children born very preterm and examined associations with neonatal factors, spirometry, and respiratory quality of life (St. George Respiratory Questionnaire, SGRQ). Methods In a population-based cohort of 150 children (121 very preterm, 29 term controls), participants underwent same-day xenon MRI, spirometry, and SGRQ at age six. MRI assessed ventilation, alveolar dimensions, and gas exchange (membrane uptake and RBC transfer). Metrics included mean values, coefficient of variation, and lung-volume percent of abnormal lung function. Low-quality data were excluded. Phenotyping involved: (1) imputing missing MRI metrics via ensemble regression (predictors: demographics, SGRQ), (2) normalizing to term control medians/IQRs, (3) applying PCA to each image type, and (4) clustering using a density-based algorithm with standardized Euclidean distance. Final phenotype evaluation used non-imputed data and included term controls. Results Very preterm subjects were classified in infancy as 48% without BPD, 27% with grade 1 BPD, 17% with grade 2 BPD, and 7% with grade 3 BPD. Three primary clusters/phenotypes were found using xenon MRI. Cluster 0 (n = 33) primarily had xenon metrics consistent with healthy term controls (meanventilation≈0.997, meanairspace≈0.0247, meanmembrane≈0.0102, meanRBC≈0.0036). Cluster 1 (n = 84) had the most impaired ventilation and alveolar simplification - despite the alveolar simplification, “normal” membrane uptake was observed suggestive of possible inflammation, thickened interstitium, or proliferative tissue consistent with a partial-restrictive phenotype seen in other studies. Cluster 2 (n = 33) had moderately impaired ventilation with alveolar simplification (increased airspace size, reduced membrane-uptake, reduced RBC-transfer). Cluster 0 (n = 33) included the most term controls, had a longer gestation (34.7 weeks, P < 0.001), a higher birth weight (2320g, P < 0.001), and better lung function (FEF Z=-0.4, P = 0.04) than the other two clusters. Cluster 2 included very preterm patients and had a shorter gestation (27.6 vs 28.8 weeks), lower birth weight (1047 vs 1231 g), and worse SGRQ scores (14.1 vs 9.7) compared to Cluster 1. Conclusions Xenon MRI identified three distinct phenotypes: 17% of very preterm children were “normal” (Cluster 0), 71% showed moderate obstruction with possible interstitial pathology (Cluster 1), and 22% exhibited alveolar simplification with mild obstruction but normal interstitium (Cluster 2). These findings highlight xenon MRI’s potential for phenotype-specific, personalized care in preterm populations. This abstract is funded by: NIH
No takes yet. Share an insight, caveat, or question.
Willmering et al. (2026) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: