On 18 May, at the American Thoracic Society (ATS) 2026 annual meeting, a cohort analysis of the Study of Mechanisms of Action of Omalizumab in Severe Asthma (SoMOSA) (ISRCTN15124178) was presented.
SoMOSA is an open-label, real-world study undertaken at 18 hospitals within the UK National Health Service (NHS) system, in which severe, uncontrolled asthma patients were administered Novartis’s Xolair (omalizumab) at doses between 75mg and 600mg based on body weight for 52 weeks, while remaining on their standard pre-study treatments throughout. The goal of this analysis was to establish and determine the accuracy of a metabolite-based pre-treatment clinical biomarker for omalizumab response.
Go deeper with GlobalData
Access deeper industry intelligence
Experience unmatched clarity with a single platform that combines unique data, AI, and human expertise.
With the steady rise of approved therapies to treat severe asthma, particularly an increase in biological treatments, that bring added interest with novel mechanisms of actions (MOAs) and potentially elevated therapeutic efficacies, the issue of how to best select the correct drug for each patient to ensure proper symptom control and sustained disease remission has become an obstacle. The severe asthma biologics market is increasingly crowded, with agents including omalizumab, mepolizumab, benralizumab, dupilumab, and tezepelumab each targeting distinct immunological pathways, making patient stratification a critical and unresolved issue.
SoMOSA and similarly designed studies provide the unique opportunity to further develop and evaluate patient-based precision treatment. According to key opinion leaders (KOLs) previously interviewed by GlobalData, there is a high unmet need for efficient diagnostic and biomarker tests to better determine the appropriate medicine per patient, not simply with asthma, but throughout immunological diseases.
For the purpose of this analysis, longitudinal plasma samples from 168 participants underwent global and targeted metabolomic profiling. Treatment response was set at a 50% or greater reduction in oral corticosteroid (OCS) use, disease exacerbations, and Asthma Control Test (ACT) improvement at 52 weeks. Logistic, elastic net, and mixed-effects regression models assessed the relationship between metabolites and these outcomes. The primary findings were validated in an independent cohort from the Mass General Brigham Biobank (MGBB).
Investigators’ evaluations of findings indicated that no one metabolite or ratio correlated significantly with reduction in disease exacerbation in response to omalizumab. However, multiple Week-52 metabolite ratios, particularly sphingolipid-to-steroid and ceramide-to-steroid measures, were significantly associated with OCS reduction (p=0.041 to 0.05, Meff corrected) and ACT improvement (fdr p=0.043 to 0.05).
Models built from a subset of Week-52 metabolite ratios identified via regularised feature selection achieved high accuracy in classifying treatment response (median AUC=0.86). Surrogate measures of these ratios using baseline metabolites predicted treatment response with nearly the same accuracy (median AUC=0.85), a finding of considerable practical significance since a pre-treatment biomarker is far more clinically actionable than one measured at the end of a 52-week course. Strong associations between these metabolite ratios and treatment response were replicated in the MGBB cohort (p=0.046 to 0.05).
These findings presented at ATS 2026 represent an early but meaningful step towards metabolomics-guided precision medicine in severe asthma. The ability to predict omalizumab response before treatment initiation could reduce the trial-and-error prescribing cycles that KOLs have identified as a key unmet need.
However, important questions remain: the cost-effectiveness of metabolomic profiling at scale within routine clinical practice has not been established, and whether this biomarker framework can be extended to other approved biologics in severe asthma remains to be determined. Further validation in larger, prospective cohorts will be required before this approach could be considered for clinical implementation.
