MultiMeDIA (Multimodal Medical Data Integration & Analysis) Lab
Our human body is a team of interconnected organs working in perfect harmony in healthy body, or disharmony during pathology. All the current medical imaging modalities can only provide a piece of the anatomical puzzle, either through 2D projections (e.g. x-ray), or 2D sparse planes (e.g. MRI), or without accounting for motions (e.g. CT, MRI). It gets further complicated due to the presence of breathing, heart beating, and patient movements. Beyond the anatomy, there are functional characteristics including physiology, electrophysiology, electromechanics, etc., which are also related to simple demographic to complex blood, kidney, protein, and genomics biomarker data. So clearly some Assembly is required to understand and analyse this Multiverse of Madness.
That’s why we are in the Multimodal Medical Data Integration & Analysis Lab, in short MultiMeDIA Lab, working to develop automated, AI-assisted tools for real-time clinical assistance both for the early diagnosis as well as during the interventions. We are both working on the development of optimal patient care pathways in the high-income countries, as well as providing early diagnosis and intervention support during the scarcity of resources such as in the low- and middle-income countries. Our aim is to develop generalisable approaches that will work optimally anywhere in the world.
Modelling and Analysis of 3D Vasculatures
X-ray angiography is the most commonly used imaging modality for the detection of coronary stenoses. However, the high inter- and intra-observer variability in interpreting the geometry of 3D vascular structure based on multiple 2D image projections is a limitation in the accurate determination of lesion severity. In order to address these issues, a novel point-cloud based approach is developed for 3D coronary arterial tree reconstruction. Comparison of the reconstructed 3D lumen surface with optical coherence tomography (OCT) measurements shows no statistically significant difference in the luminal cross-sections.
3D Heart Mesh Reconstruction
Cardiac magnetic resonance (CMR) is increasingly used for non-invasive evaluation of the myocardium, providing accurate assessments of ventricular function, perfusion, oedema, and scar. From the CMR images acquired using a standard clinical protocol, an accurate high-resolution 3D representation of the myocardium is generated using patient-specific 3D surface mesh reconstruction, integrating slice alignments and segmentation refinements for a spatially consistent slice arrangement, including deep learning methodology, both of which optimise consistency between long and short axis contours.