Volv Global model flags early ARDS risk in pneumonia

Volv Global says a machine learning model can identify community-acquired pneumonia patients at risk of acute respiratory distress syndrome up to five days before diagnosis. The retrospective work, built on U.S. claims data and validated on ICU data and expert chart review, will be presented at ERS Congress 2026 in Barcelona.

Why it matters: - Acute respiratory distress syndrome is a life-threatening complication with a narrow window for intervention. - Earlier risk detection could help clinicians act sooner and could support patient enrichment in ARDS trials. - Volv Global says the approach may also improve decisions in other diseases that are difficult to detect early.

What happened: - Volv Global developed a machine learning model for CSL Behring’s research question on community-acquired pneumonia and ARDS. - The results will be presented at ERS Congress 2026 in Barcelona. - The model flagged patients likely to progress from community-acquired pneumonia to ARDS up to five days before an ARDS diagnosis code appeared. - The retrospective analysis used de-identified U.S. claims data from the Komodo Health database.

The details: - The training dataset covered 341,697 patient records from 2016 to 2023. - The model reached ROC-AUC 0.89 on U.S. claims data. - Independent validation on the MIMIC-IV ICU dataset produced ROC-AUC 0.88. - Three independent specialists in the U.S., U.K. and Germany reviewed flagged phenotypes and found high concordance with ARDS pathophysiology. - The model’s top 20 high-risk cases matched expert review at 95% agreement. - Volv Global said the model is built through inFlow, its prognostic modeling and outcome prediction solution. - The system learns disease-specific biomarkers from population-scale real-world data. - CSL Behring helped shape the clinical questions and how the results could inform trial design and patient care. - Academic collaborators in the U.S., U.K. and Germany contributed clinical expertise. - Volv Global said the model supports clinical decision-making and research but does not diagnose ARDS or replace clinician judgment.

Between the lines: - The work is retrospective, so the results still need prospective validation before clinical use can be established. - The strong performance on both claims data and ICU data suggests the model may generalize beyond the original dataset. - Independent chart review adds credibility because the output aligned with specialist assessment, not just statistical metrics. - Volv Global is positioning the project as a template for similar pharma collaborations in diseases with short treatment windows.

What's next: - Full results will appear in the ERS Congress 2026 abstract, “Early prediction of ARDS in community-acquired pneumonia patients using machine learning.” - Prospective validation would be the next step before broader clinical adoption. - If confirmed, the model could help clinical teams identify high-risk patients sooner and help trial sponsors enrich study populations more precisely.

The bottom line: - Volv Global says its AI model can spot ARDS risk in pneumonia patients days earlier than diagnosis, but the evidence remains retrospective and needs prospective testing before routine clinical use.