With cardiovascular disease and mild cognitive impairment increasingly recognised as interconnected conditions, the DORIAN GRAY project aims to bridge the gap between cardiology and neurology through an innovative, artificial intelligence-driven approach. Dr Ivana Galinovic discusses how this groundbreaking initiative leverages digital twin models, real-world data and avatar-based exergaming to revolutionise risk assessment, prevention and intervention.
In recent decades, there has been a notable rise in the frequency of non-communicable diseases and their effect on public health, with cardiovascular disease (CVD) being a prime representative of this category.
Simultaneously, partly due to increased human longevity, mild cognitive impairment (MCI) is becoming more prevalent in the population, and these conditions often co-occur. It is estimated that one-third of patients with CVD experience some degree of cognitive impairment.
In response to this, the 2019 World Health Organization guidelines on Risk Reduction of Cognitive Decline and Dementia advocate for personalised, proactive risk reduction initiatives to combat MCI through improved risk assessment and individualised prevention. This is precisely what the DORIAN GRAY project hopes to achieve.
The exact interaction between cardiovascular risk factors and brain pathologies remains unclear, prompting the project’s primary objective: to integrate various laboratory, genetic and imaging biomarkers with clinical variables to develop an artificial intelligence (AI)-powered digital twin model for personalised risk assessment and stratification.
We aim to provide a more individualised approach to preventing MCI through avatar-based coaching exergaming, which will function as both a cognitive enhancement tool via the exergame component and a lifestyle intervention tool through the coaching component.
Why is the link between CVD and mild cognitive impairment so significant?
The two diseases are certainly linked through the damage that CVD inflicts on the central nervous system. Clinical and subclinical cerebral infarcts adversely impact brain function and connectivity. Arteriosclerosis leads to arterial stiffness, which increases arterial pulse pressure on the cerebral vasculature. Together with low cardiac output – caused by, for example, atrial fibrillation or heart failure – this results in chronic cerebral hypoperfusion, contributing to neuroinflammation and white matter injury.
Much of the relationship between CVD and MCI still needs to be explored and is often overlooked in clinical practice. This is partly due to the lack of effective treatment for MCI. In addition, the organisation of conventional medicine into specific disciplines supports a reductionist approach, sometimes making the connections between organs and systems less apparent.
Tackling the interaction of CVD and MCI requires an interdisciplinary effort, which is what DORIAN GRAY will provide.
Leveraging avatars, exergaming and digital coaching via DORIAN GRAY
The AI-based tool being developed as part of the DORIAN GRAY project is a digital twin and will function as a stratification tool for MCI risk assessment. While it aims to facilitate cognitive screening for the high-risk subgroup of CVD patients and individuals with cerebrovascular risk factors, it is also designed to be adaptable for the general population.
Our proposed intervention – exergaming – is defined as ‘physical exercise interactively combined with cognitive stimulation in a gaming environment’. Existing evidence suggests that such activities are associated with broad benefits in cognition, emotional and social functioning, as well as overall health. However, several psychological barriers may hinder its implementation, particularly for individuals experiencing cognitive decline, such as diminished initiative and interest.
By leveraging an avatar approach to exergaming, we believe we can address some of the psycho-sociological barriers to motivation and create a personalised and adaptive physical–cognitive intervention, enhancing user experience and compliance through quests in a ‘juicy’ role-playing game (RPG)-like environment – meaning one that feels alive and responds to everything the user does.
The digital coaching system will be integrated with an avatar-based exergame within an application called MIRROR. The coaching system will provide a behavioural intervention through motivational and educational messages based on behavioural change techniques.
The integrated multidomain exergame will utilise pervasive technology, including smartwatches, smartphones and tablets, to deliver personalised and dynamically adaptive combined physical training and cognitive stimulation programmes. These will take the form of structured ‘gamified’ training sessions, realised as goal-oriented mini-quests with specific cognitive-physical objectives. They will be virtually situated within an RPG-like environment to enhance engagement and adherence to the intervention.
We anticipate that our intervention will demonstrate high levels of acceptability and will be capable of improving or at least maintaining cognitive functioning over a follow-up period of 12 months.
The MIRROR app will collect qualitative and quantitative real-world data on heart rate, blood pressure, and blood oxygen levels during interactions with the exergame.
This information will be provided as summary metrics for each exergaming session within the MIRROR app for both the study physicians and participants. It will be used to monitor the participants’ safety and progress, as well as to adjust the overall exergaming intensity for each individual.
What are the anticipated benefits of DORIAN GRAY for CVD and MCI?
DORIAN GRAY will investigate the interconnected physiological pathways linking CVD and MCI, providing insights into their shared mechanisms and aiming to develop a neurobiology-based hypothesis relating MCI to CVD.
This will allow us to create a reliable AI tool for risk stratification of MCI across various subgroups, which healthcare systems can utilise as a screening method for primary and secondary MCI prevention.
By developing an interactive and immersive healthcare intervention like the MIRROR app, DORIAN GRAY aims to boost patient engagement and positively influence social dynamics within healthcare environments, empowering individuals to take control of their health.
A clinical study planned as part of the DORIAN GRAY project will evaluate the efficacy of this intervention in a high-risk population of patients with heart failure and consider efficacy and implementation. A second clinical study will consider individuals with MCI and cardiovascular risk factors and the socioeconomic impact of the multidomain intervention. The studies will identify insights for adherence and assessing its cost-effectiveness in the context of secondary prevention.
Both the MIRROR app and the exergame, if proven effective, will provide a solution that can be adapted into a commercial product for broader use or utilised as a medical device to treat patients with varying degrees of MCI, ideally leading to reduced healthcare expenses.
In this way, DORIAN GRAY fully endorses the 5-P Medicine vision adopted by the EU, which recommends digital transformation towards the goals of predictive, preventive, personalised, participatory and psycho-cognitive healthcare as a target for 2030.
The challenges of integrating and analysing an extensive dataset
With data expected from over 300,000 individuals across various countries, the main challenge we are currently facing is obtaining detailed information and access to these diverse cohorts in a timely manner. Once this is achieved, the primary challenge will be addressing data heterogeneity.
Neuroimaging features and electrocardiogram variables will be extracted from scratch using a consistent postprocessing pipeline for all datasets to ensure maximum harmonisation of parameters. Clinical variables will be recoded across all datasets, and the variability in the platforms used to measure metabolomics and proteomics will be addressed by evaluating all individual metabolites and protein analytes.
Despite these efforts, we anticipate missing data, which we plan to address by employing multivariate data imputation methods, such as k-nearest neighbours – a machine learning technique for handling missing values – along with supervised learning approaches like regression trees and artificial neural networks.
However, training our digital twin model with the derived dataset may still encounter the challenge of insufficient data. This can be addressed through data augmentation and transfer learning techniques to maximise the use of the available data. Alternatively, simpler models could also be developed using a reduced set of variables or information.
Unifying clinical practice and raising awareness
Beyond merely raising awareness of the connection between CVD and MCI, our project aims to identify and provide new, reliable biomarkers for assessing cognitive decline in this demographic. This could promote the establishment of a cohesive clinical pathway across cardiology, neurology and psychiatry for the diagnosis and eventual prevention of MCI.
Companies might utilise the MIRROR app to create various versions tailored for use as a commercial tool or as a medical device for patients. Integrating the coaching tool and the exergame into systems for the primary and secondary prevention of MCI would positively affect cognitive function while promoting healthy ageing throughout the UK and Europe. It would also enhance confidence in digital health interventions and hopefully foster greater public engagement.
We estimate potential savings of between €200bn and €400bn across Europe and the UK over a 27-year period by preventing the progression of up to two million individuals to dementia.
By promoting interdisciplinary collaboration and personalised digital health solutions, DORIAN GRAY has the potential to transform patient care, impact healthcare policies and usher in a new era of dementia prevention and healthy ageing throughout the UK and Europe.
Author
Ivana Galinovic MD PhD
Radiologist, researcher and head of CSB neuroradiology, Center for Stroke Research Berlin (CSB), Charité – Universitaetsmedizin Berlin, Germany
