Periodic Reporting for period 1 - ONCOVALUE (Implementing value-based oncology care at European cancer hospitals: An AI-based framework for assessing real-life effectiveness of novel cancer therapies in real-time)
Reporting period: 2022-12-01 to 2024-05-31
ONCOVALUE will unlock the full potential of real world data (RWD) collected in European cancer hospitals and institutes to ease the decision-making of regulators on cost-effectiveness of novel cancer therapies. To achieve this, we build up data collection and processing capabilities of leading European cancer hospitals to create high-quality data sources with clinical outcomes, quality of life, and adverse events data.
With the use powerful artificial intelligence (AI) technologies, we will transform unstructured data originating from medical notes and medical images into structured data to enable analytics and real world evidence (RWE). This RWE will be directly available for hospital decision-making, and for regulatory and health technology assessment (HTA) bodies to adopt optimized data-driven methodologies for the effective assessment of medicinal products and digital health innovations. For that, we will provide an end-to-end infrastructure for RWD reporting and address the legal constraints in the cancer hospitals to ensure secure and legal access to RWD. Furthermore, ONCOVALUE will ensure the implementation of the developed guidelines and methodologies, by providing trainings for the collection and management of high-quality RWD in European cancer centres and for the analysis of this data by HTA and regulatory bodies.
By opening the door to widespread regulatory and HTA integration of RWD, ONCOVALUE will lead to safer, more efficient, and affordable therapies, technologies, and digital solutions for cancer care. As such, ONCOVALUE is positioned to contribute to increased cost-effectiveness and subsequent sustainability of cancer care, as well as improved wellbeing of patients.
With the use powerful artificial intelligence (AI) technologies, we will transform unstructured data originating from medical notes and medical images into structured data to enable analytics and real world evidence (RWE). This RWE will be directly available for hospital decision-making, and for regulatory and health technology assessment (HTA) bodies to adopt optimized data-driven methodologies for the effective assessment of medicinal products and digital health innovations. For that, we will provide an end-to-end infrastructure for RWD reporting and address the legal constraints in the cancer hospitals to ensure secure and legal access to RWD. Furthermore, ONCOVALUE will ensure the implementation of the developed guidelines and methodologies, by providing trainings for the collection and management of high-quality RWD in European cancer centres and for the analysis of this data by HTA and regulatory bodies.
By opening the door to widespread regulatory and HTA integration of RWD, ONCOVALUE will lead to safer, more efficient, and affordable therapies, technologies, and digital solutions for cancer care. As such, ONCOVALUE is positioned to contribute to increased cost-effectiveness and subsequent sustainability of cancer care, as well as improved wellbeing of patients.
During project implementation, partners have investigated what data is available to each clinical partner and established a common framework for selecting clinical use cases. The feasibility of the use cases has been examined at each center, and plans have been made on how to validate the developed solutions in the pilot studies.
The goal of ONCOVALUE is to increase the availability of structured data that can be used for creating Real-World Evidence (RWE). This is done with two tracks; the first track is to build structured data entry into modern electronic medical record (EMR) systems so that the data created during standard clinical routines is in structured form already in the recording phase. The second track is to develop AI tools that can extract structured information from unstructured data such as radiological images or free-text medical notes.
The cornerstone of any HTA process is access to relevant clinical data, harmonised and curated for this specific purpose. One of the most important obstacles to the efficient use of RWD has been the lack of a standardized data model, which makes it difficult to analyse data from different data sources. The ONCOVALUE project is using Observational Medical Outcomes Partnership (OMOP) common data model (CDM) and shall contribute to the processes of OMOP harmonization in each hospital, so that data is truly comparable.
The goal of ONCOVALUE is to increase the availability of structured data that can be used for creating Real-World Evidence (RWE). This is done with two tracks; the first track is to build structured data entry into modern electronic medical record (EMR) systems so that the data created during standard clinical routines is in structured form already in the recording phase. The second track is to develop AI tools that can extract structured information from unstructured data such as radiological images or free-text medical notes.
The cornerstone of any HTA process is access to relevant clinical data, harmonised and curated for this specific purpose. One of the most important obstacles to the efficient use of RWD has been the lack of a standardized data model, which makes it difficult to analyse data from different data sources. The ONCOVALUE project is using Observational Medical Outcomes Partnership (OMOP) common data model (CDM) and shall contribute to the processes of OMOP harmonization in each hospital, so that data is truly comparable.
For collecting and processing RWD, ONCOVALUE introduces automated and standardized methods as part of standard clinical routines. The project is first focusing on developing practices through set use cases, developing and testing a structured real-time data collection pathway for breast cancer and non-small cell lung cancer (NSCLC), with plans to widen horizons in the near future to additional cancer types.
First versions of guidelines and standard operating procedures (SOPs) have been developed for the collection, processing and basic analytics of structured data. The documents can be used by other cancer centres that aim at building fully structured data collection practices into their clinical routines. In structured EMRs, data elements that are available in a structured format have been evaluated and novel documentation templates implemented to increase the availability of structured data. The templates will help both the clinicians to document relevant data consistently and ensure that all necessary RWD is documented for later utilization.
The project is developing a “future-proof” HTA-framework, which will provide guidance for using RWD-based HTAs, utilizing both structured and unstructured data. To accomplish this, the HTA recommendations related to the acceptance of RWE have been comprehensively explored.
Developing and validating AI models requires annotated image and text data. For medical images, annotated imaging data has been prepared using a tool developed within the project to streamline annotation. This data is used for the development of AI algorithms for automatic disease progression quantification. For clinical report data, the capabilities of the Clinical Data Collector (CDC)-tool to automatically identify and extract structured data from unstructured medical notes have been explored. Next, text data will also be annotated to further develop natural language processing (NLP) models for data structuring. For this purpose, a text annotation training has been created, simultaneously launching the Knowledge Hub which will serve as an integral centre of excellence, hosting the trainings developed in the project to share the gained knowledge from the project.
The solutions developed within the ONCOVALUE project will be validated in cancer hospitals. This will be executed by performing pilot studies, which will validate the guidelines and SOPs developed for the collection and processing of structured RWD, the hybrid RWD-based HTA-framework and the AI-tools developed for the automatic extraction of structured information from unstructured data. To execute this, two pilot studies have been designed, one in the breast cancer setting and one in lung cancer, and a validation process handbook for both studies created.
ONCOVALUE is aiming to develop a RWD source for federated analysis by HTA and regulatory bodies across Europe, containing harmonized and curated data. Eventually, we aim to display data from the participating hospitals into a federated dashboard. There are many advantages in federated analysis: it enables bigger sample size, more variation in patients and also variation across the clinical centers and countries that are needed for value-based HTA decisions.
First versions of guidelines and standard operating procedures (SOPs) have been developed for the collection, processing and basic analytics of structured data. The documents can be used by other cancer centres that aim at building fully structured data collection practices into their clinical routines. In structured EMRs, data elements that are available in a structured format have been evaluated and novel documentation templates implemented to increase the availability of structured data. The templates will help both the clinicians to document relevant data consistently and ensure that all necessary RWD is documented for later utilization.
The project is developing a “future-proof” HTA-framework, which will provide guidance for using RWD-based HTAs, utilizing both structured and unstructured data. To accomplish this, the HTA recommendations related to the acceptance of RWE have been comprehensively explored.
Developing and validating AI models requires annotated image and text data. For medical images, annotated imaging data has been prepared using a tool developed within the project to streamline annotation. This data is used for the development of AI algorithms for automatic disease progression quantification. For clinical report data, the capabilities of the Clinical Data Collector (CDC)-tool to automatically identify and extract structured data from unstructured medical notes have been explored. Next, text data will also be annotated to further develop natural language processing (NLP) models for data structuring. For this purpose, a text annotation training has been created, simultaneously launching the Knowledge Hub which will serve as an integral centre of excellence, hosting the trainings developed in the project to share the gained knowledge from the project.
The solutions developed within the ONCOVALUE project will be validated in cancer hospitals. This will be executed by performing pilot studies, which will validate the guidelines and SOPs developed for the collection and processing of structured RWD, the hybrid RWD-based HTA-framework and the AI-tools developed for the automatic extraction of structured information from unstructured data. To execute this, two pilot studies have been designed, one in the breast cancer setting and one in lung cancer, and a validation process handbook for both studies created.
ONCOVALUE is aiming to develop a RWD source for federated analysis by HTA and regulatory bodies across Europe, containing harmonized and curated data. Eventually, we aim to display data from the participating hospitals into a federated dashboard. There are many advantages in federated analysis: it enables bigger sample size, more variation in patients and also variation across the clinical centers and countries that are needed for value-based HTA decisions.