Periodic Reporting for period 2 - AIDPATH (Artificial Intelligence-driven, Decentralized Production for Advanced Therapies in the Hospital)
Berichtszeitraum: 2022-07-01 bis 2023-12-31
AIDPATH aims to transform the use of CAR T-cell therapy in EU hospitals and simplify complex manufacturing workflows. The project will develop an innovative AI-driven platform within Smart Manufacturing Hospitals (SMH) to improve efficiency and accessibility. The AIDPATH platform seamlessly integrates the modular manufacturing of CAR-T cells through end-to-end automation and leverages the scalability of the Aglaris Facer bioreactor. The project will use experimentally derived process models to manage patient variability, utilizing data from automated QC devices and clinical trials. Early optimization through metabolic monitoring improves bioprocess outcomes. Advanced AI technology integrates patient-specific data and biomarkers and implements flexible manufacturing schemes to ensure optimal CAR-T cell suitability. AI controls also optimize scheduling, reduce costs and improve patient access. Supported by a robust data architecture, the AI solutions enable continuous learning and optimization by consolidating data from multiple sources. This comprehensive approach ensures thorough patient monitoring and optimized CAR-T production. AIDPATH is pioneering the use of AI to revolutionize CAR-T therapy and promote accessibility, cost-effectiveness and clinical success in a sophisticated and credible way.
In the second reporting phase of the AIDPATH project, significant strides were made across various work packages. The consortium defined manufacturing SOPs, production sequence diagrams, and material flow mappings for the automated CAR-T cell platform. Despite an eleven-month delay in PLC component delivery, the automated system's design was finalized, and hardware build-up commenced, with ongoing integration of the PLC cabinet and LHU programming. CE certification and GMP documentation progress, including risk analysis and operator's manual, are subcontracted.
Ethics meetings, though delayed, remain essential for addressing concerns. Collaboration between different partners focused on GMP-compliant documentation, qualification, validation, and risk analysis. Experimental work, guided by Design of Experiments, generated high-throughput datasets for CAR-T manufacturing, enhancing the understanding of crucial process parameters. We could also presenting concepts and methodologies for AI modules and actively contributing to Project Board and Executive Group discussions.
The provision of SOPs for production, the optimization of CAR-T production through small-scale experiments and the preparation for the introduction of the AIDPATH platform were also covered. The Bioreactor FACER prototype can be customized now, refining single-use cartridges, hardware, and software. Variants of T-cell culture media are being developed to operate the bioreactor and perform biological experiments for process optimization. Databases are also provided, which are crucial for CAR-T therapy and AI training.
Successfully developed soft sensors for adaptive control and monitoring in AI use case 2, which will be integrated into the newly developed AI-COPE interface. The LogiqSuite structure for managing medical data was further developed and prepared for patient planning and AI use case 5. Furthermore, surveys and workshops were conducted to capture the trust and acceptance of the stakeholder groups. This contributed to the social guideline and the development of the business model.
Overall, AIDPATH's second phase demonstrated progress in hardware build-up, AI module development, regulatory approaches, and stakeholder engagement. Ongoing collaboration, despite delays, sets the stage for the platform's final implementation, emphasizing its potential impact on personalized CAR-T cell therapy.
Ethics meetings, though delayed, remain essential for addressing concerns. Collaboration between different partners focused on GMP-compliant documentation, qualification, validation, and risk analysis. Experimental work, guided by Design of Experiments, generated high-throughput datasets for CAR-T manufacturing, enhancing the understanding of crucial process parameters. We could also presenting concepts and methodologies for AI modules and actively contributing to Project Board and Executive Group discussions.
The provision of SOPs for production, the optimization of CAR-T production through small-scale experiments and the preparation for the introduction of the AIDPATH platform were also covered. The Bioreactor FACER prototype can be customized now, refining single-use cartridges, hardware, and software. Variants of T-cell culture media are being developed to operate the bioreactor and perform biological experiments for process optimization. Databases are also provided, which are crucial for CAR-T therapy and AI training.
Successfully developed soft sensors for adaptive control and monitoring in AI use case 2, which will be integrated into the newly developed AI-COPE interface. The LogiqSuite structure for managing medical data was further developed and prepared for patient planning and AI use case 5. Furthermore, surveys and workshops were conducted to capture the trust and acceptance of the stakeholder groups. This contributed to the social guideline and the development of the business model.
Overall, AIDPATH's second phase demonstrated progress in hardware build-up, AI module development, regulatory approaches, and stakeholder engagement. Ongoing collaboration, despite delays, sets the stage for the platform's final implementation, emphasizing its potential impact on personalized CAR-T cell therapy.
The ambition of AIDPATH is to take CAR-T manufacturing to a new level by demonstrating that decentralized production and AI-optimized processes lead to higher quality and availability of CAR-T cell therapeutics.
The Project will deliver important innovations:
Build-up of an automated manufacturing plant for CAR-T cell therapy including centralized AI-integrated control software COPE
Better understanding and synergies concerning the intelligent and automated manufacturing of advanced therapies in hospitals.
Development of a real-time digital twin platform to enable real-time prediction for biomanufacture of advanced therapeutics.
New techniques for scheduling in uncertain domains combined with hard time-constraints; novel reinforcement learning based solutions for personalized therapies in stochastic, dynamic environments.
early dialogue with regulatory authorities on adaptive manufacturing and advanced PATs control.
Modify the FACER equipment to achieve full automation and seamless integration with the AIDPATH platform. Allow adaptive manufacturing in CAR T production processes.
Development of a novel ADCF T cell medium that requires no addition of serum for growth of T cells such as CAR-T cells. The medium is safe and suited to be subsequently upgraded to a GMP-compliant quality grade
Results from metabolomics studies of samples during manufacturing process in order to improve manufacturing process and AI training of manufacturing platform
Development of AI manufacturing platform and informatic system thanks to provided databases.
Current monitoring of bioreactor cell perfusion process uses PID controller based on Glucose sensor. Some issues exist with PID controllers and Glucose sensor, such as distinguishing transient disturbances which do not require reassignment of set-points. The adaptive control and monitoring (soft sensors) developed will use a diversity of sensors (Glucose, Lactate, O2, CO2, pH, ..) and different advanced statistical and AI techniques to provide a more complete vision of the behavior of the bioreactor in relation to the set-points.
Identification of factors, concerns, and challenges related to trust in and acceptance of artificial intelligence (AI) applications in medicine
developing a societal guide for implementing Smart Manufacturing Hospitals
co-developing a business model plan and providing insight into the financial analysis, external environment, and a risk analysis for hospitals interested to transform their current business model into a future oriented business model by means of the AIDPATH platform
early cost-effectiveness analysis of the AIDPATH system.
Medical data management system to integrate care and lab data, with syncing to analytics, facilitating
collaboration of local and network data, and real-time analysis of real-world data.
Preparing data framework for AI development in personalized CAR-T cell therapy.
Impact:
AIDPATH aims to advance CAR T-cell therapy in Europe by improving access, reducing costs and implementing smart control solutions. This will be achieved through regulatory adjustments for optimized CAR-T products, faster and more robust manufacturing and the commercialization of the developed ADCF medium. The goals include accelerating the development of CAR-T therapy, making it more affordable and providing human operators with more comprehensive information for better decision-making. In addition, the project contributes to building trust in AI technology, societal aspects and scalability, with a focus on influencing patient outcomes and facilitating precision medicine through effective data management and continuous learning loops.
The Project will deliver important innovations:
Build-up of an automated manufacturing plant for CAR-T cell therapy including centralized AI-integrated control software COPE
Better understanding and synergies concerning the intelligent and automated manufacturing of advanced therapies in hospitals.
Development of a real-time digital twin platform to enable real-time prediction for biomanufacture of advanced therapeutics.
New techniques for scheduling in uncertain domains combined with hard time-constraints; novel reinforcement learning based solutions for personalized therapies in stochastic, dynamic environments.
early dialogue with regulatory authorities on adaptive manufacturing and advanced PATs control.
Modify the FACER equipment to achieve full automation and seamless integration with the AIDPATH platform. Allow adaptive manufacturing in CAR T production processes.
Development of a novel ADCF T cell medium that requires no addition of serum for growth of T cells such as CAR-T cells. The medium is safe and suited to be subsequently upgraded to a GMP-compliant quality grade
Results from metabolomics studies of samples during manufacturing process in order to improve manufacturing process and AI training of manufacturing platform
Development of AI manufacturing platform and informatic system thanks to provided databases.
Current monitoring of bioreactor cell perfusion process uses PID controller based on Glucose sensor. Some issues exist with PID controllers and Glucose sensor, such as distinguishing transient disturbances which do not require reassignment of set-points. The adaptive control and monitoring (soft sensors) developed will use a diversity of sensors (Glucose, Lactate, O2, CO2, pH, ..) and different advanced statistical and AI techniques to provide a more complete vision of the behavior of the bioreactor in relation to the set-points.
Identification of factors, concerns, and challenges related to trust in and acceptance of artificial intelligence (AI) applications in medicine
developing a societal guide for implementing Smart Manufacturing Hospitals
co-developing a business model plan and providing insight into the financial analysis, external environment, and a risk analysis for hospitals interested to transform their current business model into a future oriented business model by means of the AIDPATH platform
early cost-effectiveness analysis of the AIDPATH system.
Medical data management system to integrate care and lab data, with syncing to analytics, facilitating
collaboration of local and network data, and real-time analysis of real-world data.
Preparing data framework for AI development in personalized CAR-T cell therapy.
Impact:
AIDPATH aims to advance CAR T-cell therapy in Europe by improving access, reducing costs and implementing smart control solutions. This will be achieved through regulatory adjustments for optimized CAR-T products, faster and more robust manufacturing and the commercialization of the developed ADCF medium. The goals include accelerating the development of CAR-T therapy, making it more affordable and providing human operators with more comprehensive information for better decision-making. In addition, the project contributes to building trust in AI technology, societal aspects and scalability, with a focus on influencing patient outcomes and facilitating precision medicine through effective data management and continuous learning loops.