The cooperation between clinicians and engineers is a key condition to delivering healthcare engineering innovations. Dr Carole Sudre (UCL, London) and Dr Beatriz Gómez-Ansón (Hospital de Sant Pau, Barcelona) teamed up to develop more efficient medical imaging assessment.


From raw data to novel diagnosis techniques 

Big data and machine learning are often seen as having great potential for innovation in the medical sector. But what’s the path from building databases to revolutionising medical diagnosis?

Dr Sudre uses big sets of data and develops machine learning algorithms as new tools to be used in the field of medical imaging. The outputs of these novel assessment methods are then compared at scale to the neuroradiological visual assessment performed by Dr Gómez-Ansón. Once the reliability of the tools has been established, these can then be used to bring new insights and deliver more rapid medical evaluation.

Take this research project as an example: 4500 cases of lesions have been used by Dr Sudre to build a visualisation tool that sequentially presents a summary of lesion measurements in an anatomically-driven manner. This interface facilitated the assessment process done by Dr Gómez-Ansón, who told us “By having the data displayed in a user-friendly way, I was able to classify >4500 lesions in a week’s time.” Tools like this one have the potential to simplify clinicians’ work to detect and categorise large amount of cases having different pathologies.


Two sides of each healthcare engineering story

 To provide such healthcare engineering innovations requires close coupling between software engineering and clinical validation. “One of the benefits of this close collaboration is being able to enlarge the perspectives on both sides and understand what different engineering and clinical aspects should be factored in”, affirms Dr Sudre.

 A further example in this regard is the bullseye biomarker diagram that Dr Sudre produced. Dr Gómez-Ansón applied this to a small, prospective MRI study of patients with Cushing’s disease carried out at the Hospital Sant Pau and compared the use of this regional display to the more conventional radiological approach of reading MRI images. The bullseye diagrams proved to be more time efficient and provided more quantitative information in a single glimpse.

 Projects like these ones showcase ways in which machine learning and engineering-clinician interdisciplinary efforts impact the future of medical diagnosis. It is this kind of synergy that drives innovation within the field of healthcare engineering.


This research work is being developed under the guidance of Prof Rolf Jager and Dr M. Jorge Cardoso.


IMG 3358RE