Study Design and Planning
Defining clinical research questions, study objectives, patient eligibility criteria, outcome measures and analysis plans.
Clinical data · Imaging · AI
Research
My research integrates clinical data, medical imaging, artificial intelligence and quantitative biomarkers to investigate neurological disease, occupational lung disease and cancer. Working with clinical and research teams, I connect clinically meaningful questions with study design, data collection, analysis and validation to generate evidence for diagnosis, disease characterisation and clinical decision-making.
Research Design and Delivery
My experience spans the design and delivery of research involving clinical data, medical imaging and artificial intelligence. Working with clinical and research teams, I connect study planning, ethics and hospital coordination with data collection, reference standards, analytical methods, validation and scientific reporting.
Ethics and Study Coordination
Preparing ethics documentation, supporting research governance, coordinating patient recruitment and working with hospital and research teams.
Clinical Data and Reference Standards
Collecting and linking clinical records and images from hospital systems (eMR/PACS), de-identifying and preparing data, and establishing reference standards with clinical experts.
Analysis and Model Development
Statistical analysis, image preprocessing, feature engineering, quantitative biomarker development and machine learning using clinical data and medical images.
Validation and Clinical Interpretation
Evaluating predictive and segmentation performance, testing generalisability across cohorts, investigating errors and interpreting findings with clinical collaborators.
Reporting and Reproducibility
Preparing research reports, manuscripts and conference presentations, supported by documented methods, reproducible analysis workflows and clear reporting of limitations.
Research Areas and Selected Projects
Neuroimaging & neurological disease
Epilepsy: lesion detection and segmentation
How can MRI and machine learning help detect and delineate subtle brain abnormalities associated with focal epilepsy?
I develop unsupervised lesion-detection methods for 7T MRI and multimodal segmentation methods using T1-weighted and FLAIR images, with a focus on focal cortical dysplasia, small lesions and model failures.
Predicting seizure recurrence after a first seizure
Can clinical information and MRI-derived brain measurements help estimate the likelihood of another seizure?
I develop imaging features and machine-learning methods, integrate clinical information, and evaluate predictive performance. This includes NeuroMorphix, which captures brain asymmetry from structural MRI.
Alzheimer’s disease: structural MRI biomarkers
How can structural MRI characterise brain changes across normal cognition, mild cognitive impairment and Alzheimer’s disease?
I analyse cortical and subcortical measurements and networks describing similarities between brain regions, combining machine learning with harmonisation across imaging sites to investigate potential biomarkers.
Friedreich’s ataxia: diffusion MRI and cerebellar pathways
How can diffusion MRI quantify changes in the cerebellar pathways affected by Friedreich’s ataxia?
I develop automated methods to segment the cerebellar peduncles and extract measurements of white matter microstructure, evaluating segmentation accuracy and consistency across imaging sites.
Occupational lung disease & thoracic imaging
Chest X-ray and CT assessment in dust-exposed workers
How closely does chest X-ray assessment agree with high-resolution CT (HRCT), and how well does it identify abnormalities in dust-exposed workers?
At I-MED Radiology Network, I lead research delivery and analysis in occupational lung imaging. ExactDust examines agreement and diagnostic performance between chest radiography and HRCT, alongside related work on quantitative assessment of structural lung abnormalities.
Linking quantitative CT with lung function
How do quantitative CT measurements relate to spirometry and other measures of lung function in coal mine dust lung disease?
In ACARP-supported research, I evaluate CT-derived measurements against spirometry and other available pulmonary function measures to investigate relationships between lung structure and function.
Cancer imaging
PET–CT tumour segmentation across cancer types
How well do PET–CT tumour segmentation models generalise across cancer types and patient cohorts?
I evaluate nnU-Net across oesophageal and lung cancer cohorts and public datasets, comparing training strategies and analysing segmentation errors.