A proprietary, computational platform developed through machine learning to generate powerful digital tools for disease screening, prevention and intervention through:
Development of digital biomarkers to assess efficacy of drugs on the multiple parameters of biological ageing
Assessment of patient health in a clinical setting
Enhancement of drug discovery processes
Proprietary predictive models are differentiated through the integration of large phenotypic and genomic data sets providing a unique insight into the status of biological clocks and ageing. Models have been developed for application in metabolic, pulmonary and neurodegenerative diseases.
Case Study 1 – blood based biological age
Blood-based assessments are perhaps some of the most widespread and accessible approaches for biological age estimation, relying on relatively easy-to-get clinically measurable quantities. At Oxcitas we have developed an array of ML-based methodologies (with varying degrees of complexity) that consider an individual’s general demographic characteristics together with different combinations of anthropometric measurements and blood biomarkers to predict biological age and relevant associated risk metrics. These predictions are in turn used as input of an inverse optimisation problem to derive optimal biomarkers modifications that would allow the end user of our technology to reach a desired goal (e.g., a specific reduction in biological age). Finally, we add further value to our predictions by providing actionable solutions as sets of highly personalised lifestyle suggestions. These valuable developments are accessible and can be used as effective monitoring tools as well as sources of tangible and actionable recommendations for improvement of patient health and quality of life.
Case Study 2 – MRI based brain age
MRI brain age models developed at Oxcitas gather state-of-the-art scientific research in the field and harness such knowledge for the development of a set of usable and useful digital tools. The mathematical models behind these tools are able to accurately predict brain age and the risk of developing certain classes of neurodegenerative diseases and conditions like dementia. We employ both ML and DL techniques to derive our brain age estimations (based on brain MRI scans and/or MRI-derived features) and to accurately predict the risk of having or developing within a time frame ageing diseases associated with abnormal brain age scores. Apart from being a valuable support in clinical assessment and for patient stratification, our developments are complemented with a set of personalised recommendations with the aim to promote changes in the lifestyle of the individual that can help reduce their risk factors and have a significant impact on their future.