AI-based clinical software for fully automated Cardiac Magnetic Resonance reporting

Projektinformationen

AI4CMR

ID Finanzhilfevereinbarung: 190149120

DOI

10.3030/190149120

Projekt abgeschlossen

EK-Unterschriftsdatum 16 August 2022

Startdatum 1 September 2022

Enddatum 28 Februar 2025

Finanziert unter

• The European Innovation Council (EIC)

Gesamtkosten

€ 2 505 488,57

EU-Beitrag

€ 1 753 842,00

Koordiniert durch

AI4MEDIMAGING MEDICAL SOLUTIONS SA
Portugal

Periodic Reporting for period 2 - AI4CMR (AI-based clinical software for fully automated Cardiac Magnetic Resonance reporting)

Berichtszeitraum: 2023-09-01 bis 2025-02-28

Zusammenfassung vom Kontext und den Gesamtzielen des Projekts

Cardiovascular disease (CVD) is the leading global cause of death, responsible for 18 million lives annually (32% of all deaths). CVD imposes a significant economic burden, costing €54 billion in healthcare and €45 billion in productivity losses in the EU in 2017. Factors like population growth, ageing, and risk factors such as obesity exacerbate this challenge. Addressing CVD has become a global public health priority, but existing diagnostic methods exhibit accuracy and scalability limitations, resulting in around 7 million misdiagnosed deaths yearly.
AI4MedImaging is pioneering AI4CMR.Plus a project revolutionizing CVD diagnosis. It centres on advanced clinical software automating heart function and structure assessment via Cardiac Magnetic Resonance (CMR) imaging. CMR is the gold standard but faces operational barriers, e.g. long queues and costs, stemming from complex analysis of over 300 contracting heart images. The scarcity of CMR experts (0.06-3 per 1 million Europeans) leads to errors (85% inter-/intra-observer variability) and unmet diagnosis demand (~22 citizens per 1 million Europeans).
AI4CMR.Plus seeks to democratize CMR diagnosis through automated image analysis, unlocking CMR's full potential. AI4MedImaging's MVP software, AI4CMR, was introduced in the EU market. With EIC support, AI4CMR.Plus development accelerates, targeting a 2025 market launch. This growth enhances company stature, fortifies EU digital sovereignty, and transforms the lives of CVD-threatened individuals.
The project's vision is to democratize CMR through automated AI, making advanced cardiac exams accessible and scalable. This innovation's impact extends beyond healthcare, affecting the economy, society, and humanity. Addressing limitations of existing CVD diagnostics, AI4CMR.Plus shapes cardiovascular disease management, contributing to a healthier society on a global scale.

Arbeit, die ab Beginn des Projekts bis zum Ende des durch den Bericht erfassten Berichtszeitraums geleistet wurde, und die wichtigsten bis dahin erzielten Ergebnisse

AI4CMR.Plus is an upgraded version of our MVP AI4CMR, recently launched to the EU market (Q1 2022). AI4CMR was the result of 3 years of AI4MedImaging collaboration with the Clinical Academic Centre of Braga and School of Medicine of Minho University. Devised to automatically post-process CINE sequences and to be fully agnostic, interoperable, and cloud-based, AI4CMR is well positioned relative to competing CMR software. However, the full value proposition and market potential of our AI tool will only be materialized by evolving AI4CMR to AI4CMR.Plus. We have Obtained a patent for 'Artificial Intelligence Cardiac Motion Classification' in the EU and another in the US, and have achieved CE Mark and FDA approval for AI4CMR.
So far, in this first year we have: Improved CINE interpretation pipeline – developed technology now allows multi-class and segmental approaches to model training; - Matured LGE analysis and interpretation feature – new features enable automatic segmentation, quantification, and classification of fibrosis, for different LGE-CMR acquisition protocols. We have also started the development of technological protocols and frameworks for interoperability and the improvement of the cloud infrastructure, resulting in improved scalability, availability, and security. Regarding marketing achievements we have established agreements with strategic partners and engaged potential users to run early-stage trials of AI4CMR.Plus. We have also obtained SFDA approval (KSA) and started the preparation of the application for the Indian CDSCO. Change requests are being prepared for CE-mark and FDA approval of the newly developed features.

Fortschritte, die über den aktuellen Stand der Technik hinausgehen und voraussichtliche potenzielle Auswirkungen (einschließlich der bis dato erzielten sozioökonomischen Auswirkungen und weiter gefassten gesellschaftlichen Auswirkungen des Projekts)

AI4CMR.Plus aims to streamline cardiac magnetic resonance (CMR) exams using deep learning technology. This project seeks to reduce CMR exam duration, enhance interpretation consistency, and seamlessly integrate with radiology workflows. Key partnerships with industry leaders are expected to expand accessibility.
Results and Potential Impacts:
1. Efficiency: AI4CMR.Plus will significantly shorten CMR exams, benefiting healthcare providers and patients.
2. Consistency: Deep learning algorithms will standardize CMR image interpretation, reducing variability and improving diagnosis reliability.
3. Workflow Enhancement: Integration into the physician's workflow will boost efficiency and care quality.
4. Accessibility: Partnerships with industry leaders will promote widespread adoption, potentially enabling earlier detection of cardiac issues.
Key Needs for Further Uptake and Success:
1. R&D: Continual refinement is crucial to adapt to evolving medical practices.
2. Interoperability: Ensuring compatibility with diverse healthcare systems is essential.
3. IP Protection: Robust intellectual property protection is necessary.
4. Regulatory Approval: Compliance with healthcare regulations is vital for global access.
5. Funding and Market Access: Securing resources and market access is critical for scalability.
6. Global Expansion: Adapting to local regulations and practices is essential for international growth.
7. Advocacy: Promoting favourable regulatory policies and standards can facilitate adoption.