The relevance of open data for medical research: The Case of FISABIO Foundation

In the digital age, technological advancements have transformed the field of medical research. One of the factors contributing to technological development in this area is data, particularly open data. The openness and availability of information obtained from health research provide multiple benefits to the scientific community. Open data in the healthcare sector promotes collaboration among researchers, accelerates the validation process of study results, and ultimately helps save lives.

The significance of this type of data is also evident in the prioritized intention to establish the European Health Data Space (EHDS), the first common EU data space emerging from the European Data Strategy and one of the priorities of the Commission for the 2019-2025 period. As proposed by the European Commission, the EHDS will contribute to promoting better sharing and access to different types of health data, not only to support healthcare delivery but also for health research and policymaking.

However, the handling of this type of data must be appropriate due to the sensitive information it contains. Personal data related to health is considered a special category by the Spanish Data Protection Agency (AEPD), and a personal data breach, especially in the healthcare sector, has a high personal and social impact.

To avoid these risks, medical data can be anonymized, ensuring compliance with regulations and fundamental rights, thereby protecting patient privacy. The Basic Anonymization Guide developed by the AEPD based on the Personal Data Protection Commission Singapore (PDPC) defines key concepts of an anonymization process, including terms, methodological principles, types of risks, and existing techniques.

Once this process is carried out, medical data can contribute to research on diseases, resulting in improvements in treatment effectiveness and the development of medical assistance technologies. Additionally, open data in the healthcare sector enables scientists to share information, results, and findings quickly and accessibly, thus fostering collaboration and study replicability.

In this regard, various institutions share their anonymized data to contribute to health research and scientific development. One of them is the FISABIO Foundation (Foundation for the Promotion of Health and Biomedical Research of the Valencian Community), which has become a reference in the field of medicine thanks to its commitment to open data sharing. As part of this institution, located in the Valencian Community, there is the FISABIO-CIPF Biomedical Imaging Unit, which is dedicated, among other tasks, to the study and development of advanced medical imaging techniques to improve disease diagnosis and treatment.

This research group has developed different projects on medical image analysis. The outcome of all their work is published under open-source licenses: from the results of their research to the data repositories they use to train artificial intelligence and machine learning models.

To protect sensitive patient data, they have also developed their own techniques for anonymizing and pseudonymizing images and medical reports using a Natural Language Processing (NLP) model, whereby anonymized data can be replaced by synthetic values. Following their technique, facial information from brain MRIs can be erased using open-source deep learning software.

BIMCV: Medical Imaging Bank of the Valencian Community

One of the major milestones of the Regional Ministry of Universal Health and Public Health, through the Foundation and the San Juan de Alicante Hospital, is the creation and maintenance of the Medical Imaging Bank of the Valencian Community, BIMCV (Medical Imaging Databank of the Valencia Region in English), a repository of knowledge aimed at achieving "technological advances in medical imaging and providing technological coverage services to support R&D projects," as explained on their website.

BIMCV is hosted on XNAT, a platform that contains open-source images for image-based research and is accessible by prior registration and/or on-demand. Currently, the Medical Imaging Bank of the Valencian Community includes open data from research conducted in various healthcare centers in the region, housing data from over 90,000 subjects collected in more than 150,000 sessions.

New Dataset of Radiological Images

Recently, the FISABIO-CIPF Biomedical Imaging Unit and the Prince Felipe Research Center (FISABIO-CIPF) released in open access the third and final iteration of data from the BIMCV-COVID-19 project. They released image data of chest radiographs taken from patients with and without COVID-19, as well as the models they had trained for the detection of different chest X-ray pathologies, thanks to the support of the Regional Ministry of Innovation, the Regional Ministry of Health and the European Union REACT-EU Funds. All of this was made available "for use by companies in the sector or simply for research purposes," explains María de la Iglesia, director of the unit. "We believe that reproducibility is of great relevance and importance in the healthcare sector," she adds. The datasets and the results of their research can be accessed here.

The findings are mapped using the standard terminology of the Unified Medical Language System (UMLS), as proposed by the results of Dr. Aurelia Bustos' doctoral thesis, an oncologist and computer engineer. They are stored in high resolution with anatomical labels in a Medical Image Data Structure (MIDS) format. Among the stored information are patient demographic data, projection type, and imaging study acquisition parameters, among others, all anonymized.

The contribution that such open data projects make to society not only benefits researchers and healthcare professionals but also enables the development of solutions that can have a significant impact on improving healthcare. One of these solutions can be generative AI, which provides interesting results that healthcare professionals can consider in personalized diagnosis and propose more effective treatment, prioritizing their own judgment.

On the other hand, the digitization of healthcare systems is already a reality, including 3D printing, digital twins applied to medicine, telemedicine consultations, or portable medical devices. In this context, the collaboration and sharing of medical data, provided their protection is ensured, contribute to promoting research and innovation in the sector. In other words, open data initiatives for medical research stimulate technological advancements in healthcare.

Therefore, the FISABIO Foundation, together with the Prince Felipe Research Center, where the platform hosting BIMCV is located, stands out as an exemplary case in promoting the openness and sharing of data in the field of medicine. As the digital age progresses, it is crucial to continue promoting data openness and encouraging its responsible use in medical research, for the benefit of society.