Towards precision medicine: platform for transdiagnostic stratification of brain dysfunction

Expected Impact:

The action under this topic is expected to achieve the following impacts:

Data platform for precision medicine: A comprehensive, sustainable data-driven health platform linking behaviours and symptoms to quantitative biological markers. This will deliver much-needed refinements to existing diagnostic frameworks and treatment paradigms, providing a clear step towards personalised healthcare for CNS-driven symptoms. This platform could be a model for other disease areas where there is need for more biology-driven precision medicine.

Advancing mechanistic understanding: Clarification of the biological basis of CNS transdiagnostic symptoms expediting the identification of novel and more effective precision therapies across the relevant disorders, boosting the competitiveness of European industry and beyond. Advanced mechanistic understanding will also galvanise innovation in diagnostics.

Patient outcomes and stigma: A significant improvement in care quality (more integrated and personalised care) and in health outcomes for people within the RM&I driven relevant disorders. This includes healthcare innovations arising from improved understanding of the relationship between psychiatric and physical health and their underlying biology with far-reaching implications for patient health beyond psychiatry and neurology. This will also lead to reduction of stigma and provide opportunities for early intervention.

Efficiency in the healthcare system: Precision treatments reduce avoidable waste in healthcare resources, leading to overall cost reduction, higher productivity, and a positive economic impact on the European health care budget. Overall, a transformative shift towards a more integrated and personalised approach to healthcare will benefit patients across Europe and beyond.

The action will also support the EU political priority to boost European competitiveness and contribute to a number of European policies/initiatives, which include the European Health Data Space Regulation (EHDS)¹, the EU Artificial Intelligence Act² and the European Commission’s Communication on a comprehensive approach on mental health³ and the Healthier Together- EU Non-Communicable Diseases initiative⁴.

¹_{https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=OJ:L_202500327}

²_{https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=CELEX:32024R1689}

³_{https://commission.europa.eu/strategy-and-policy/priorities-2019-2024/promoting-our-european-way-life/european-health-union/comprehensive-approach-mental-health_en}

⁴_{https://health.ec.europa.eu/non-communicable-diseases/healthier-together-eu-non-communicable-diseases-initiative_en}

Expected Outcome:

The action generated by this topic is expected to contribute to all the following outcomes:

1. A sustainable and collaborative large, multimodal data platform that can identify novel transdiagnostic candidate markers and endpoints for the symptom domains of reward/motivation (including anhedonia) and impulsivity (RM&I) [1],[2] in neuropsychiatric, neurodegenerative, and physical health disorders. Relevant disorders include Alzheimer’s disease (AD), major depressive disorder (MDD) and obesity (priority areas). Other relevant disorders/diseases include but are not limited to substance use and associated disorders, schizophrenia, bipolar disorder, borderline personality disorder and Parkinson’s disease. For a disorder/disease to be relevant, there must be evidence to show that reward/motivation and/or impulsivity are clinically significant symptom domains;

2. Novel transdiagnostic candidate markers and endpoints are identified and progressed towards validation. Learnings are applied in drug discovery to increase probability of success (PoS);

3. A clear roadmap to achieve full validation of candidate markers and endpoints by regulatory and health technology assessment (HTA) bodies. Clinical best-practice guidelines are developed, and recommendations are made to the current diagnostic classifications¹ to expedite the adoption of precision medicine;

4. A greater understanding of the biological foundations of RM&I symptom domains and their role in AD, MDD, obesity and other relevant disorders, enabling the generation of novel therapeutic approaches by industry;

5. Closer alignment between psychiatry, neurology, and physical health disciplines to enable dialogue between healthcare professionals (HCPs) and other medical specialists to optimise outcomes, particularly for individuals with complex healthcare needs and comorbidities.

¹_{Diagnostic and Statistical Manual of Mental Disorders (DSM) and International Classification of Diseases (ICD)}

Objective:

Objectives of the topic

Applicants are expected to address all four main objectives of the topic in their proposal:

1. Adapt and extend an existing federated data platform that is sustainable, enabling collaborative curation, access and analysis of clinical datasets and samples (as mentioned above). Consolidate existing multimodal datasets and samples from cohorts with relevant disorders¹ into the adapted platform;

2. Collect additional new clinical datasets and samples to address gaps and integrate these into the adapted platform;

3. Test hypotheses for candidate markers and endpoints within a defined context (e.g. patient selection, diagnosis, or treatment monitoring) in a transdiagnostic patient population presenting symptom domains of RM&I, including those with AD, MDD, and obesity;

4. Establish a collaborative platform to bring together people with LE, HCPs, regulators, HTA bodies and payers to achieve consensus on the value of candidate markers and endpoints, how to operationalise them into new diagnostic and treatment frameworks and achieve readiness in the healthcare system.

Activities under objective 1:

The success of this topic hinges on access to a large of amount of high-quality, multimodal data and biological samples collated from applicants and other partners (including industry). The applicants must list the datasets and samples that they will bring and confirm that they will be made accessible to the whole public-private partnership (PPP) from the start of the action.

1.1 Collate existing multimodal, longitudinal and transdiagnostic datasets at an individual level, including relevant parameters outlined under 1.2. These datasets can come from public or private databases, observational studies, clinical trials, real-world evidence (RWE) studies, biobanks, electronic health records, registries, and/or other digital health technologies and platforms. In their short proposal applicants must include a strategy to utilise relevant data from the European Platform for Neurodegenerative Diseases (EPND) catalogue² as much as possible as well as other relevant datasets available from previous projects (including pre-clinical data).

1.2 Relevant multimodal datasets ideally include as many as possible from the following: neurophysiology data (e.g. electroencephalography (EEG), magnetoencephalography (MEG)), brain imaging data (e.g. functional magnetic resonance imaging (fMRI), MRI), qualitative subjective assessments, behavioural data, real-world data, medical claims and billing data, routine clinical data (from medical and psychological assessments including data on metabolic status), physiological/activity monitoring data (polysomnography, actigraphy, digital data from wearables, etc.), speech/language data, patient reported outcome data (e.g. questionnaires), molecular biodata (e.g. “-omics”), and potentially data gained via therapeutic protocols (drugs, neuromodulation (deep brain stimulation, transcranial magnetic stimulation, transcranial functional ultrasound, etc.)). Biological samples (e.g. blood, urine, stools, cerebrospinal fluid) from biobanks should be leveraged. Datasets should be from individuals with relevant disorders as well as healthy controls.

1.3 Propose a strategy to integrate and connect the datasets from different sources.

1.4 Outline an approach to inclusive and equitable practices, including data representation, including but not limited to gender, ethnicity, and age (e.g. paediatric and adolescent populations).

1.5 Adapt and extend a federated data platform by building on existing infrastructures proven effective in PPPs, including the AD Workbench³ and the EPND hub⁴ (made available via the pre-identified industry consortium). The adapted platform should leverage available resources (including standard operating procedures) from EPND [4]. The adapted platform must be scalable and adaptable to curate high-quality, multimodal, retrospective, prospective and longitudinal data as mentioned under 1.1 and 1.2. It must enable data/sample discovery, access, and support AI analysis, while ensuring interoperability with other global data platforms.

1.6 Ensure high data quality by verifying the robustness of methodologies before integration into the adapted platform. This could be achieved by establishing a Data Quality Assessment Committee.

1.7 Implement fair and transparent governance for data- and sample-sharing including model interpretability, data provenance, and traceability of AI decision-making processes. Applicants must explain how they will develop a consensus on data sharing principles, complying with legal and ethical standards (e.g. General Data Protection Regulation (GDPR) and intellectual property rights (IPR)) and ensuring robust protection of data volunteers' rights. For example, leveraging the Data Sharing Playbook⁵ and setting up a Data Access Review Committee.

1.8 Review the collated dataset to identify key gaps and develop a strategy to guide new data collection under Objective 2.

1.9 Identify potential algorithms for activities under 3.3 in Objective 3.

1.10 Ensure platform sustainability by creating a strong value proposition and user ecosystem beyond the consortium, with a clear strategy for a long-term, AI-powered platform accessible to the broader research community, including the healthcare industry. Relevant activities need to be in place from the start of the action.

Activities under objective 2

2.1 Collect new prospective multimodal and ideally longitudinal data from transdiagnostic cohorts, focussing on individuals affected by RM&I abnormalities in the relevant disorders, and ideally also collect biological samples. The datasets should close data gaps identified under 1.8 and be integrated into the adapted platform, meeting the same criteria described in objective 1.

2.2 Continue to recognise and fill data gaps to expand and maintain the adapted data platform, keeping it current with technological and scientific advancements. Whenever appropriate, utilise AI/ML, such as synthetic data generation, image analysis, natural language processing etc., to enhance the dataset.

2.3 Continue identification of potential algorithms for activities under 3.3 in Objective 3.

Activities under objective 3

3.1 The short proposal should propose an initial pilot clinical case study designed to test a scientifically robust and data-supported hypothesis on candidate markers and/or endpoints in RM&I symptom domains during the project’s initial year. It must include transdiagnostic populations from AD, MDD, and obesity. The case study must include as a minimum neurophysiological data (e.g. EEG or MEG) and brain imaging data (e.g. MRI, fMRI) from each subject. In addition, datasets should include as many parameters as possible from the list described in 1.2. The precise scope of the initial clinical case study will be developed by the full consortium during the preparation of the full proposal. (Additional case studies are described under 3.4).

3.2 In the first 6 months, prepare a systematic literature review (white paper) of the available potential markers in RM&I symptom domains in relevant disorders to support hypothesis generation and subsequent testing. This should be kept up to date throughout the action.

3.3 Apply suitable statistical methods, advanced computational analytics (including, whenever appropriate, AI/ML as part of the statistical/analytical toolbox), modelling, and simulation across the multimodal data in the adapted platform to cluster biologically similar subjects across disorders/diseases, stratified independently of their conventional diagnostic classification. This should enable to identify and confirm clinically significant, quantitative candidate markers for RM&I symptom domains in relevant disorders, incorporating hypothesis-driven and data-driven approaches. It should also establish the foundation for a new transdiagnostic framework based on phenotypes/biotypes to enable detection of factors for susceptibility, risk stratification, diagnostic precision, disease monitoring, treatment response prediction, and overall patient outcomes. In addition, it should elucidate the biological underpinnings of the relationship between psychiatric and physical health (e.g. for obesity, understanding the interplay between metabolic disturbances, mental health and eating behaviours).

3.4 Test putative transdiagnostic markers and endpoint hypotheses derived from 3.2 and 3.3 through additional non-sequential pilot clinical case studies, incorporating insights from stakeholder consultations as mentioned in objective 4. These case studies must test the same transdiagnostic marker/endpoints in separate pre-defined patient populations in two or more of the relevant disorders to strengthen the transdiagnostic approach. As a preference, the three priority disorders should be included in at least one study each as a lead indication. For instance, one study with AD, one with MDD and one with obesity as the lead indication, each including at least one additional relevant disorder. Each study must include neurophysiological and brain imaging data and include as many other parameters as possible from the list outlined under 1.2. Studies must be powered sufficiently to allow analyses both within and across the included disorders. All results must be integrated into the adapted platform. The studies should enhance the platform's ability to accelerate hypothesis testing of new candidate markers and endpoints within a defined context of use (e.g. patient selection, diagnosis, or treatment monitoring) in representative patient populations. These studies must not involve the development of new in vitro diagnostic tools or digital sensors. The resulting evidence from pilot case studies (including the initial pilot clinical study under 3.1) should:

• be verifiable and applicable for patient stratification and/or monitoring in future clinical trials;
• demonstrate clinical utility to foster new patient pathways and clinical guidelines;
• contribute to bridging the gap between health care needs and capacity.

Activities under objective 4

4.1 Create an efficient collaborative platform to support seamless communication and collaboration among key stakeholders in the field of the relevant disorders. This includes innovators, researchers, clinicians, people with LE, carers, patient advocates, HCPs, regulators, scientific societies, HTA bodies, payers, and policy makers to collectively define and implement a new framework for the diagnosis and treatment of these disorders.

4.2 Form advisory/working groups comprising different stakeholders to support activities under objectives 1, 2 and 3, and co-create solutions. Ensure active and meaningful participation of people with LE, carers, and advocacy organisations throughout the activities and governance.

4.3 Engage with regulators (via experts with relevant expertise), e.g. EMA and/or national competent authorities, proactively initiating early consultations as appropriate. This should set the basis for continuation towards full validation of markers and endpoints beyond the action. Applicants are expected to consider the potential regulatory impact of the results and as relevant, develop a regulatory strategy and interaction plan early on to define a strategic approach to evidence collection and analysis where feasible (including case studies under objective 3) for generating appropriate evidence, as well as engaging with regulators in a timely manner (e.g. national competent authorities, EMA Innovation Task Force, qualification advice). Similarly, appropriately engage with HTA bodies and payers on the value of new transdiagnostic framework, candidate markers and endpoints when used to support claims of effectiveness of new therapies, paving the way for future reimbursement.

4.4 Craft evidence-based clinical guidelines through consultations with stakeholders, including people with LE, regulators, HTA bodies, payers, and medical organisations. Achieve consensus on best practices for implementing the new transdiagnostic framework. Develop recommendations and provide proposals for updates to the classification of disorders⁶.

4.5 Design and implement a comprehensive training programme for HCPs to adopt the new transdiagnostic framework. Create educational materials and implement trainings for people with LE, families and carers in multiple languages, ensuring readiness across the healthcare system for the paradigm shift in healthcare delivery throughout Europe and helping to reduce stigma.

Applicants are expected to leverage and build on the learnings and outputs from previous and ongoing relevant PPPs⁷ and other relevant global, European and national initiatives. They should consider synergies with the future European Genomic Data Infrastructure (GDI)⁸, part of the European 1+Million Genomes Initiative⁹, the future European Partnership for Brain Health¹⁰ and other relevant upcoming projects¹¹.

_{¹ Relevant disorders: Alzheimer’s disease, major depressive disorder and obesity are priority areas for this topic. Other relevant disorders/diseases include but are not limited to substance use disorders, schizophrenia, bipolar disorder, borderline personality disorder and Parkinson’s disease. For a disorder/disease to be relevant, there must be evidence to show that reward/motivation and/or impulsivity are clinically significant symptom domains.}

²_{https://discover.epnd.org/catalogue/studies}

³_{https://www.alzheimersdata.org/ad-workbench}

⁴_{https://epnd.org/news-and-resources}

⁵_{Data Sharing Playbook: https://www.ihi.europa.eu/sites/default/files/uploads/Documents/ProjectResources/IMI_IHI_DataSharingPlayBook_2024.pdf}

⁶ _{see footnote 1}

_{7 EPND, PRISM/ PRISM2, RADAR-CNS, RADAR-AD, MOBILISE-D, IDEA-FAST, EU-PEARL, SOFIA, READI, AIMS-2-TRIALS, EHDEN, PROMINENT, PREDICTOM, AD-RIDDLE, among others.}

⁸ _{Synergies with the future European Genomic Data Infrastructure (GDI) are encouraged. https://gdi.onemilliongenomes.eu/}

⁹_{https://gdi.onemilliongenomes.eu/ ; https://digital-strategy.ec.europa.eu/en/policies/1-million-genomes}

¹⁰_{https://www.brainhealth-partnership.eu/about/}

¹¹_{IHI Call 11 T1: Understanding how infections foster and induce non-communicable diseases as well as relevant projects funded under IHI Call 9.}

Scope:

Current diagnosis and patient stratification in health disorders with Central Nervous System (CNS)-driven symptoms are based on DSM-5 / ICD-11 codes, which are not aligned with underlying biological processes and mechanisms. Subsequent suboptimal disease classification and patient stratification is a key reason for the low PoS of clinical development and the historical lack of new and more efficacious treatments. This topic aims to address these challenges by adopting a holistic, transdiagnostic approach [3] focused on the common underlying biology of RM&I symptom domains across the relevant disorders listed in the first expected outcome.

The topic seeks to build on an existing federated data platform to consolidate, curate, link and analyse robust, multimodal datasets from relevant patient populations. Thus, activities related to building a new platform or a biorepository from scratch are out of scope. The data platform must enable data/sample discovery, access, and support advanced computational analysis including artificial intelligence (AI)/ machine learning (ML) technologies while ensuring interoperability with other global data platforms to illuminate the biological basis of the RM&I symptom domains and identify related candidate markers and endpoints. The hypotheses will be prospectively tested in clinical case studies focusing on but not limited to AD, MDD, and obesity. Post-project, the platform will be available as open access for ongoing research and validation.

The topic also priorities collaboration with relevant stakeholders, including people with lived experience (LE), carers, HCPs, providers, regulators, HTA bodies and payers, to prepare the healthcare system for this transformative shift. People with LE can provide unique insights and expertise that comes as the result of first-hand experience of health challenges. Integrating LE expertise improves research by bringing an understanding beyond academic and clinical knowledge. The perspectives of people with LE across the relevant symptom domains must be represented within the consortium and applied wherever appropriate.

Applicants must outline their approach to inclusive and equitable practices throughout the initiative, possibly through a risk register and appropriate mitigations. Example areas for consideration include data representation, bias mitigation, stakeholder engagement, ethics, and feedback mechanisms.