Periodic Reporting for period 2 - METAflow (Development and validation of the first digital, clinical grade flow cytometry platform for enhanced diagnosis accuracy and sustainability in immunology and hemato oncology)
Reporting period: 2023-10-01 to 2024-06-30
Flow cytometry (FC) is a technology massively used in R&D, supporting many IVD tests, it mobilizes 3 components: Reagents, Instrument and Software for data analysis. Thanks to recent progresses in hardware, measurement of tens of parameters on millions of cells creates complex, multidimensional data sets. However, FC still faces 2 main bottlenecks:
- biologists analyze data manually, creating a technological gap between hardware capabilities and human skills, limiting FC development for clinical applications.
- most recent hardware enables the measurement of many features of the cell function, but the key role of cell metabolism in several diseases stays unchecked due to lack of available biomarkers.
Our METAflow solution is a dual track platform combining unique technology and biology with:
- a ground-breaking cloud-based software with an intuitive interface and generating automated, multiplexed and reliable data analyses.
- RBDs, proprietary biomarkers able to characterize cell metabolism by binding to nutrient transporters.
We are positioning our METAflow platform to serve 3 markets with a SaaS business model:
- A RUO version;
- a 21 CFR Part11/820 compliant version and
- a Clinical grade version (CE-IVD and FDA-approved).
METAflow is an AI-based software in the cloud with an intuitive interface, enabling automated analysis of FC data with speed and reliability. It outperforms manual analysis by enabling the analysis of complex multidimensional data for a broader use from discovery to diagnostics.
Our objective is for METAflow to become the new standard in FC, replacing cumbersome and error-prone manual analyses. We want to make instruments smarter by providing users an augmented intelligence solution allowing them to get quick, objective and reproducible results.
Our developments will turn METAflow into the 1st digital, clinical-grade FC platform able to support an expanding menu of applications in oncology, immunology, infectious diseases and many more.
- biologists analyze data manually, creating a technological gap between hardware capabilities and human skills, limiting FC development for clinical applications.
- most recent hardware enables the measurement of many features of the cell function, but the key role of cell metabolism in several diseases stays unchecked due to lack of available biomarkers.
Our METAflow solution is a dual track platform combining unique technology and biology with:
- a ground-breaking cloud-based software with an intuitive interface and generating automated, multiplexed and reliable data analyses.
- RBDs, proprietary biomarkers able to characterize cell metabolism by binding to nutrient transporters.
We are positioning our METAflow platform to serve 3 markets with a SaaS business model:
- A RUO version;
- a 21 CFR Part11/820 compliant version and
- a Clinical grade version (CE-IVD and FDA-approved).
METAflow is an AI-based software in the cloud with an intuitive interface, enabling automated analysis of FC data with speed and reliability. It outperforms manual analysis by enabling the analysis of complex multidimensional data for a broader use from discovery to diagnostics.
Our objective is for METAflow to become the new standard in FC, replacing cumbersome and error-prone manual analyses. We want to make instruments smarter by providing users an augmented intelligence solution allowing them to get quick, objective and reproducible results.
Our developments will turn METAflow into the 1st digital, clinical-grade FC platform able to support an expanding menu of applications in oncology, immunology, infectious diseases and many more.
The project allowed us to develop key features to make METAflow a robust AI-powered data analysis software, fulfilling regulatory expectations of 21 CFR Part 11/820 and CE-IVD. We also worked on the platform architecture to go at scale, and be integrated in clinical environments.
We developed the AI technology for the robust automated analysis and multi-parametric visualization of FC data. The automated analysis is done in 2 phases: 1st an unsupervised algorithm suite is applied to clean and preprocess the raw data for the clustering of cell populations; then supervised machine learning techniques identify and label cells of interest.
Integrated to the core unsupervised algorithm, additional algorithms have been developed to upgrade the performance of METAflow on clustering performance and clustering quality assessment. The output of these optimizations is a higher robustness in precision and reproducibility. In addition to the extra unsupervised algorithms enhancing the performance of the software, we have developed a series of AI-based supervised algorithms to enable the automated recognition of cell populations, to automatically determine the proportions of cell populations, and enable inter-patients FC data comparison in specific diseases and finally identify rare cells populations.
In conclusion, the “all-terrain” capability of the unsupervised layer has been improved, providing a better robustness, and a 1st series of machine learning tools have been integrated in METAflow, allowing the platform to demonstrate high quality results on challenging applications.
We also worked to ensure the scalability and the clinical-grade compliance of METAflow SaaS platform. More specifically, we integrated the AI technologies and visualization tools developed to the cloud platform, ensuring its scalability and performance as the workload increases. We have developed new security features to ensure compliance with relevant regulations, as well as preparing for the integration of the platform to Laboratory information management systems.
Regarding the platform’s infrastructure, our architecture has evolved to allow METAflow to work in different cloud environments; offering great flexibility to customers’ requirements.
FDA requirements have been taken into account to comply with 21 CFR part 11 regulation for FC data in a cloud-based software system, and the solution is ready to apply for a CE-mark.
We developed the AI technology for the robust automated analysis and multi-parametric visualization of FC data. The automated analysis is done in 2 phases: 1st an unsupervised algorithm suite is applied to clean and preprocess the raw data for the clustering of cell populations; then supervised machine learning techniques identify and label cells of interest.
Integrated to the core unsupervised algorithm, additional algorithms have been developed to upgrade the performance of METAflow on clustering performance and clustering quality assessment. The output of these optimizations is a higher robustness in precision and reproducibility. In addition to the extra unsupervised algorithms enhancing the performance of the software, we have developed a series of AI-based supervised algorithms to enable the automated recognition of cell populations, to automatically determine the proportions of cell populations, and enable inter-patients FC data comparison in specific diseases and finally identify rare cells populations.
In conclusion, the “all-terrain” capability of the unsupervised layer has been improved, providing a better robustness, and a 1st series of machine learning tools have been integrated in METAflow, allowing the platform to demonstrate high quality results on challenging applications.
We also worked to ensure the scalability and the clinical-grade compliance of METAflow SaaS platform. More specifically, we integrated the AI technologies and visualization tools developed to the cloud platform, ensuring its scalability and performance as the workload increases. We have developed new security features to ensure compliance with relevant regulations, as well as preparing for the integration of the platform to Laboratory information management systems.
Regarding the platform’s infrastructure, our architecture has evolved to allow METAflow to work in different cloud environments; offering great flexibility to customers’ requirements.
FDA requirements have been taken into account to comply with 21 CFR part 11 regulation for FC data in a cloud-based software system, and the solution is ready to apply for a CE-mark.
METAflow will help improve results quality of clinical FC and will directly impact patient care. It will also help testing labs maintain sustainability of this kind of technology, by improving analysis cost-efficiency. Its robustness provides unmatched accuracy and reliability with a quick turnaround time.
METAflow will reduce by 80% the time needed for analyzing a sample batch and could generate a 40% cost saving. To solve the problem of expertise shortage, METAflow will provide a standard and homogeneous analysis without the need of special coding skills. Enhanced performance will improve the scalability of flow-based activities. Additionally, METAflow will drive the development of innovative assays with a positive impact on multiple disorders in the long run. Our solution will have a broad societal impact considering that the end beneficiaries, the patients, will get better healthcare services in terms of diagnosis, enabling better treatment. Our solution will also ensure that innovation spreads across territories to benefit all citizens, wherever they are; aiming equal access to high quality medicine. Our company is fully engaged in this objective of access to a better medicine, as demonstrated by the reimbursement in France of our first blood test METAglut1, which is impacting children life by saving years of diagnostic wandering.
Our solution will also have an impact on the broader R&D and academic ecosystem due to the knowledge creation and exchange of expertise, thanks to METAflow’s accessibility for the training of young biologists in academic centers.
METAflow will reduce by 80% the time needed for analyzing a sample batch and could generate a 40% cost saving. To solve the problem of expertise shortage, METAflow will provide a standard and homogeneous analysis without the need of special coding skills. Enhanced performance will improve the scalability of flow-based activities. Additionally, METAflow will drive the development of innovative assays with a positive impact on multiple disorders in the long run. Our solution will have a broad societal impact considering that the end beneficiaries, the patients, will get better healthcare services in terms of diagnosis, enabling better treatment. Our solution will also ensure that innovation spreads across territories to benefit all citizens, wherever they are; aiming equal access to high quality medicine. Our company is fully engaged in this objective of access to a better medicine, as demonstrated by the reimbursement in France of our first blood test METAglut1, which is impacting children life by saving years of diagnostic wandering.
Our solution will also have an impact on the broader R&D and academic ecosystem due to the knowledge creation and exchange of expertise, thanks to METAflow’s accessibility for the training of young biologists in academic centers.