DrDominic Marshall

Dr Marshall’s research focuses on leveraging multimodal data – including high-frequency physiological data, clinical variables and chest imaging – to improve the diagnosis, phenotyping and management of acute lung injury and ARDS in the intensive care unit.

Key themes include:

Trajectory phenotypes in acute hypoxic respiratory failure
Using machine learning to identify distinct clinical trajectories and treatment responses in patients with severe respiratory failure, with the aim of supporting prognostic and predictive enrichment in clinical trials.

Integrating clinical and radiological data
Developing methods to combine longitudinal clinical data with chest radiographs and radiology reports, including work on generative and explainable models for chest X-ray analysis and radiology report evaluation. 

AI and decision support in mechanical ventilation
Studying how machine learning and clinical decision support systems can assist with invasive mechanical ventilation, personalising ventilator strategies and improving bedside decision-making in ICU. 

Global and population-level respiratory epidemiology
Analysing large international datasets (e.g. Global Burden of Disease, WHO mortality data) to describe trends in prevalence, mortality and disease burden for respiratory conditions such as COPD, interstitial lung disease, pulmonary hypertension, pneumonia and other major non-communicable diseases.