Here, Professor Nirmalie Wiratunga of Robert Gordon University discusses how the iCARE Health Hub uses remote monitoring and clinical insight to improve follow-up care for adolescent and young adult cancer survivors via an interface that prioritises ethical, personalised and clinician-informed artificial intelligence.
Adolescents and young adults (AYA) who survive cancer often face a silent burden: the ‘late effects’ of their treatment. This can include risks to their cardiovascular and pulmonary health. Despite the known dangers, traditional follow-up care involves infrequent check-ups and lacks sufficient support for lifestyle modification. As a result, many risks go undetected.
The iCARE Health Hub device, launched by the EU-funded MAYA project, seeks to change this by providing continuous, home-based monitoring tailored to young cancer survivors. Drawing on AI reasoning and established clinical guidelines, this smart mirror is equipped with an AI-powered conversational agent and tracks metrics such as heart rate, blood pressure and stress levels.
Its intelligence is drawn from two key modules: iHEART for cardiotoxicity risk tracking and iLIFE for lifestyle modification recommendations. These core components are supported by the iCARE Onto knowledge framework and together provide real-time advice, early warnings and personalised coaching to help users manage modifiable risks such as hypertension, obesity and stress.
The goals of the platform are ambitious: to achieve a 30% reduction in major cardiac events, a 15% reduction in heart-related hospital visits, and a marked improvement in quality of life (70%) and health literacy for at least 20% of participants. Data from the platform will also contribute to the broader UNCAN.eu initiative, helping to improve overall understanding of cancer survivorship care across Europe.
Learning from experience and clinicians
Unlike rule-based systems, the iCARE Health Hub uses case-based reasoning. It analyses a user’s live data, and provides recommendations informed by the most relevant real-world successes from similar survivors and is therefore grounded in both real-world outcomes and medical best practices.
Clinicians are integral to the platform’s design. Through ‘Living Labs’ set up across Europe, including one led by Robert Gordon University in Aberdeen, they co-develop and test the smart mirror system alongside survivors, researchers and industry partners. Their input is embedded into the Hub’s knowledge base and training protocols, ensuring that the advice remains practical, evidence-based and current.
Feedback from pilot studies directly informs ongoing updates to the Hub’s knowledge base and training materials, ensuring that its guidance remains practical and aligned with current medical standards. Clinical involvement helps guarantee that the Hub provides care that is both clinically robust and genuinely responsive to the complex needs of AYA cancer survivors.
All health measurements are securely stored according to FHIR/HL7 standards – the global framework for exchanging healthcare information electronically – enabling seamless integration with hospital dashboards where permitted. This keeps the system connected to mainstream healthcare workflows, allowing doctors to track patient trends without relying on siloed digital tools.
Ethical and transparent AI to inform clinical practice in cancer
For AI to be effective in healthcare, particularly with younger populations, it must not only be accurate but also understandable, unbiased and fair. That is a challenge, given that most cardiovascular datasets come from older adults, which may skew predictions for younger users.
Bias in AI remains a key concern, as unequal access to devices and skewed training data can subtly influence recommendations. Equally important is the need for plain-language clarity. Complex medical guidelines must be translated into everyday language that is clear and trustworthy. Above all, accuracy is critical, and it is essential that all recommendations align precisely with clinical guidelines and reflect the highest standards of medical reliability.
To address this, the iCARE team is building protective ‘guard-rails’ into the system. This includes training models on AYA-specific data, running continual bias checks and ensuring all outputs are presented in plain language appropriate to a diverse 15-39-year-old demographic.
In the planned user trials, plain-language consent will be obtained from every participant, and they will be clearly told what the AI can and cannot do, which data it uses and how it generates advice. All participants retain the right to withdraw at any time. Their clinicians remain a reassuring presence throughout, reinforcing that while the AI offers personalised guidance, it does not deliver diagnoses.
The MAYA project spans four years, with major development milestones and clinical trials scheduled through to the final year. Predictive models are expected to reach over 80% accuracy by Year 3 and 90% by Year 4, with full-scale piloting and data analysis running through the final phase.
Alongside product development, the consortium will publish a White Paper to disseminate its best practices across disciplines through academic journals, conferences, multidisciplinary health AI events and policy networks to engage stakeholders and foster knowledge sharing.
Policy recommendations will be generated and shared with Members of the European Parliament and other stakeholders to inform and improve healthcare policy across the EU.
Multidisciplinary collaboration as a core strength in cancer care
The MAYA project requires a large-scale, multidisciplinary effort. This diversity of expertise is a significant advantage in developing comprehensive AI-driven healthcare solutions for AYA cancer survivors. A total of 16 organisations from across Europe bring together technical, clinical, ethical and patient-centred expertise.
The University of Ioannina in Greece leads the project and brings extensive experience in coordinating complex health-related projects, while technical coordination is handled by Spain’s University of Castilla-La Mancha. Key partners also include Robert Gordon University, which is leading on AI, reasoning and digital solutions to form the iCARE Health Hub, and the Medical University of Gdańsk, which provides clinical leadership in oncology. Other partners share expertise in digital health solution development, quality and risk management, and impact and dissemination to ensure the project’s visibility.
This collaboration ensures that the solution is not developed in a vacuum, and it reflects the real-world needs of AYA cancer survivors and the professionals who care for them. The project’s scale and structure allow it to move beyond isolated innovation and toward a widely implementable, ethical and scalable model for survivorship care.
Looking towards a preventive future with AI
AI technology is well placed to move preventive care from occasional check-ups to continuous, data-driven recommendations. As algorithms learn to combine wearables, imaging, genomics and lifestyle data, they will spot subtle risk patterns long before symptoms appear. In turn, this can trigger personalised behaviour changes or medication nudges and give clinicians objective evidence to intervene early.
The result could be fewer hospital admissions, slower disease progression and better quality of life for millions living with, or at risk of, chronic conditions.
The iCARE Health Hub represents a shift from reactive to proactive cancer care. By catching warning signs early, issuing personalised interventions and empowering patients to take control, AI-driven tools like this could significantly improve long-term outcomes for many patients.
The key to success lies in thoughtful design. Retraining models for specific populations, ensuring transparent logic, embedding expert oversight and clearly communicating limitations are all necessities.
Regulatory frameworks such as the EU AI Act are a start, but local governance, opt-out mechanisms and continual monitoring will be just as important.
Conclusion
The iCARE Health Hub promises a step-change in care for AYA cancer survivors. Continuous, home-based monitoring and AI-driven coaching aim to spot heart-risk early, tailor lifestyle advice and provide round-the-clock emotional support by helping users control modifiable factors – all by giving clear, personalised guidance the moment physiological readings drift outside safe ranges.
Projects like iCARE signal a growing movement to treat survivorship not as an afterthought, but as a vital phase of cancer care. With the right tools, AYA survivors can live not just longer, but better, supported by a system that anticipates risk, adapts in real time and listens to both clinicians and patients every step of the way.
Author
Nirmalie Wiratunga PhD
Professor in intelligent systems and associate dean for research, Robert Gordon University School of Computing, UK