Periodic Reporting for period 1 - NEUROSENSE (NEUROendocrine SENSor for Sudden Unexpected Death in Epilepsy (SUDEP) prediction and prevention)
Periodo di rendicontazione: 2022-06-01 al 2023-05-31
Imagine the helplessness of being told your child, recently diagnosed with epilepsy, is at a high risk of dying suddenly, without warning and without a known cause, often in their sleep. The doctors inform you this phenomenon threatening your child's life is called Sudden Unexpected Death in Epilepsy (SUDEP), but that there is very little you can do to prevent it. Imagine the disruption to your life, as you spend hours riddled with constant worry and care-worn sleepless nights. And despite taking extensive measures to monitor your child and their epilepsy, they tragically still succumb to a silent death that you were powerless to predict or prevent. For many families living with epilepsy, this is their reality. 65 million people with epilepsy are at risk of SUDEP. However, SUDEP cannot be predicted nor prevented. The NEUROSENSE project aims to develop the first SUDEP Medical Device (SMD) prototype supported by artificial intelligence (AI) capable of anticipating life-threatening seizures and trigger an automatic emergency drug administration to prevent SUDEP.
Achieving such disruptive technology requires a close unconventional collaboration model between actors in different scientific domains. The NEUROSENSE Consortium has a deep interdisciplinary nature bridging different domains of science and technology: animal modelling and neurobiology (Centre National de la Recherche Scientifique – CNRS), clinical neurology (Filadelfia Epilepsy Hospital – Dianalund and Centre Hospitalier Universitaire Vaudois - CHUV), analytical mass spectrometry (Instituto de Investigação e Inovação em Saúde - i3S), mathematical modelling, e-health software development, product design in health sector, medical device implementation, GPDR/HIPAA knowledge (Kinetikos Health), biotechnology, product design, marketing and IP (Biostrike), biosensors and medical electronics (Karolinska Institutet). This synergy of expertise will make it possible to present a breakthrough solution for a major societal challenge helping the EU to attain leadership in the specific domain of SUDEP. Based on the new SMD technology, we will build a diverse portfolio of future projects that will result in long-term benefits for people with epilepsy, their families, caregivers, society and the economy.
Achieving such disruptive technology requires a close unconventional collaboration model between actors in different scientific domains. The NEUROSENSE Consortium has a deep interdisciplinary nature bridging different domains of science and technology: animal modelling and neurobiology (Centre National de la Recherche Scientifique – CNRS), clinical neurology (Filadelfia Epilepsy Hospital – Dianalund and Centre Hospitalier Universitaire Vaudois - CHUV), analytical mass spectrometry (Instituto de Investigação e Inovação em Saúde - i3S), mathematical modelling, e-health software development, product design in health sector, medical device implementation, GPDR/HIPAA knowledge (Kinetikos Health), biotechnology, product design, marketing and IP (Biostrike), biosensors and medical electronics (Karolinska Institutet). This synergy of expertise will make it possible to present a breakthrough solution for a major societal challenge helping the EU to attain leadership in the specific domain of SUDEP. Based on the new SMD technology, we will build a diverse portfolio of future projects that will result in long-term benefits for people with epilepsy, their families, caregivers, society and the economy.
During the first year, the technical and scientific efforts of NEUROSENSE were focused on the optimization of methodology, algorithm development and exploratory experimentation. This strategy resulted in important breakthroughs that laid solid foundations for educated confirmatory studies and pioneering work that will be performed in the subsequent years of the project. Main technical and scientific achievements include:
1- Set up equipment and optimized protocols for in vivo microdialysis of interstitial fluid (ISF) in SUDEP mice models.
2- Developed a reliable, robust and sensitive mass spectrometry method for the analytical quantification of potential SUDEP biomarkers in microdialysis samples of ISF.
3- Produced solid microneedles and sensing electrodes. These are now under optimization and will subsequently combined with novel biorecognition elements (research in frame of NEUROSENSE) to later incorporate in what will be the SUDEP Predictive Smart Biosensor (SPSB) Prototype.
4- Develop a video-based seizure detection model that can continuously analyse video data captured in animal experimental settings. By leveraging machine learning techniques, we have created a classification model capable of detecting seizure events in “real-time”. Upon seizure detection, the model triggers a predefined intervention (microdialysis operation), that was previously defined by an expert. The first version of the seizure detection algorithm and the algorithm itself can be found at https://zenodo.org/record/7969738.
5- A Statistical Analysis Plan was created. This document provides a technical and detailed explanation of the statistical analysis of the preclinical and clinical data, envisioning the development of the SUDEP risk algorithm. The goal of the SAP is to ensure the reliability and transparency of the NEUROSENSE results.
6- Developed the first (temporary) release of a Clinical Database –a cloud-based platform for storage of pseudonymized raw clinical data and raw biosignals from patients accepting to participate in the NEUROSENSE clinical study, which aims to corroborate in the clinical setting, the SUDEP biomarker identified in the preclinical studies. The NEUROSENSE Clinical Database employs various security measures, including data transmission encryption via HTTPS and cloud traffic encryption. The data storage servers are isolated within a Virtual Private Cloud with no direct external access, and all accesses are recorded in audit logs. Data is stored on two servers, with user identification metadata on one server and the remaining data, along with a matching key, on the second server. The application server holds a separate key that, together with the matching key and both databases, allows for data matching. This design ensures that even if there is a breach and access to the content occurs, the data from the two databases cannot be correlated without the second key stored in the application server. Access to the Clinical Database is via a web browser at http://platform.neurosense.pt/. Users must log in with a password and undergo multiple-factor authentication using an application token. Passwords undergo stringent verification to prevent the use of common or repeated passwords, employing standard protection measures. Additional protection includes safeguards against "Brute Force" login attempts. Different levels of data access are managed within the application, and all account and patient data access is logged to meet legal requirements. Both databases have daily backups retained for 7 days before automatic deletion. Continuous integration/continuous delivery methods were employed to develop and deploy the solution.
7- Protocol for clinical study finalized and submitted for Ethics Committee approval. We wrote a comprehensive document that, besides detailing the study protocol, includes a brief version for layman (non-experts), patient information - for adults, for parents and (separately) for teenagers, and consent forms. Since the study involves children too, there was the need to obtain an assessment from an independent expert paediatrician, not involved in the study and without any conflict-of-interest for the study. Additionally, we created a clinical sample processing manual that details all the time-points, processing instructions and matrices to be used throughout clinical sample collection and analytes quantification. Infographics were also created for the staff in the Epilepsy Monitoring Unit and the Laboratory, to ensure a smooth start of the clinical phase.
1- Set up equipment and optimized protocols for in vivo microdialysis of interstitial fluid (ISF) in SUDEP mice models.
2- Developed a reliable, robust and sensitive mass spectrometry method for the analytical quantification of potential SUDEP biomarkers in microdialysis samples of ISF.
3- Produced solid microneedles and sensing electrodes. These are now under optimization and will subsequently combined with novel biorecognition elements (research in frame of NEUROSENSE) to later incorporate in what will be the SUDEP Predictive Smart Biosensor (SPSB) Prototype.
4- Develop a video-based seizure detection model that can continuously analyse video data captured in animal experimental settings. By leveraging machine learning techniques, we have created a classification model capable of detecting seizure events in “real-time”. Upon seizure detection, the model triggers a predefined intervention (microdialysis operation), that was previously defined by an expert. The first version of the seizure detection algorithm and the algorithm itself can be found at https://zenodo.org/record/7969738.
5- A Statistical Analysis Plan was created. This document provides a technical and detailed explanation of the statistical analysis of the preclinical and clinical data, envisioning the development of the SUDEP risk algorithm. The goal of the SAP is to ensure the reliability and transparency of the NEUROSENSE results.
6- Developed the first (temporary) release of a Clinical Database –a cloud-based platform for storage of pseudonymized raw clinical data and raw biosignals from patients accepting to participate in the NEUROSENSE clinical study, which aims to corroborate in the clinical setting, the SUDEP biomarker identified in the preclinical studies. The NEUROSENSE Clinical Database employs various security measures, including data transmission encryption via HTTPS and cloud traffic encryption. The data storage servers are isolated within a Virtual Private Cloud with no direct external access, and all accesses are recorded in audit logs. Data is stored on two servers, with user identification metadata on one server and the remaining data, along with a matching key, on the second server. The application server holds a separate key that, together with the matching key and both databases, allows for data matching. This design ensures that even if there is a breach and access to the content occurs, the data from the two databases cannot be correlated without the second key stored in the application server. Access to the Clinical Database is via a web browser at http://platform.neurosense.pt/. Users must log in with a password and undergo multiple-factor authentication using an application token. Passwords undergo stringent verification to prevent the use of common or repeated passwords, employing standard protection measures. Additional protection includes safeguards against "Brute Force" login attempts. Different levels of data access are managed within the application, and all account and patient data access is logged to meet legal requirements. Both databases have daily backups retained for 7 days before automatic deletion. Continuous integration/continuous delivery methods were employed to develop and deploy the solution.
7- Protocol for clinical study finalized and submitted for Ethics Committee approval. We wrote a comprehensive document that, besides detailing the study protocol, includes a brief version for layman (non-experts), patient information - for adults, for parents and (separately) for teenagers, and consent forms. Since the study involves children too, there was the need to obtain an assessment from an independent expert paediatrician, not involved in the study and without any conflict-of-interest for the study. Additionally, we created a clinical sample processing manual that details all the time-points, processing instructions and matrices to be used throughout clinical sample collection and analytes quantification. Infographics were also created for the staff in the Epilepsy Monitoring Unit and the Laboratory, to ensure a smooth start of the clinical phase.