One of the world’s largest public, machine-readable databases for X-ray images of COVID-19 patients is stored securely in the Open Telekom Cloud, in compliance with data protection regulations. Medizinische Hochschule Hannover (MHH) can thereby provide physicians, researchers, companies, and health institutions with data reliably to enable further research into the coronavirus.
PCR, antibody, or rapid test: Since the outbreak of the coronavirus pandemic, several test procedures have become established for detecting the virus SARS-CV-2 in the human body. However, such tests are sometimes have limitations in terms of the accuracy of analysis and the associated diagnosis process. Chest X-ray examinations and computed tomography (CT) provide more precise information, even if tests for a coronavirus infection come back negative. Physicians at Medizinische Hochschule Hannover (MHH) also rely primarily on X-ray images in severe cases to be able to better assess the course of the disease in the event of a coronavirus infection, to detect possible complications at an early stage, and to act more quickly. While it is true that coronavirus cannot always be directly and unambiguously proven using the two- or three-dimensional images, typical changes in the lungs that indicate coronavirus can be identified. This means that every image may provide life-saving information.
To enable physicians, researchers, companies, and health institutions worldwide to benefit from the COVID-19 data gathered and the findings gained as a result, the Medizinische Hochschule Hannover (MHH) team under Dr. Hinrich Winther, head of the Machine Learning working group, Prof. Jens Vogel-Claussen, Chief Physician, and Prof. Frank Wacker, Director of the Institute for Diagnostic and Interventional Radiology at MHH, decided in April 2020 to make X-ray images and additional information accessible worldwide for selected physicians, researchers, companies, and health institutes, while complying with data protection. All patient data is very sensitive and should therefore be accessible in a secure and anonymized way on the internet. The result was one of the world’s largest public databases with machine-readable, structured COVID-19 patient data. It serves as the basis for intelligent analyses right through to machine learning evaluations, with the potential to provide important information for researching and combatting the disease. And for avoiding new coronavirus infections in future.
As MHH supplements the X-ray images of the lungs with age data as well as patients’ blood and cell analyses, huge volumes of data are created. However, the development platform GitHub previously used by the team only enabled the storage of metadata. To be able to digitally process and store all X-ray images, considerably higher capacities were required. In terms of data security, the medical school also had considerable reservations about storing the data on a U.S. cloud platform. The data protection pledge lacking among large U.S. cloud providers is provided by the Open Telekom Cloud. The Open Telekom Cloud impressed MHH in particular because it was designed in accordance with the strict and proven security and data protection standards of Deutsche Telekom and because it is operated exclusively by T-Systems in Europe’s most modern high-security datacenter. MHH now uses the Open Telekom Cloud Object Storage Service (OBS) as a secure storage location in Germany.
In its search for a suitable storage location, MHH quickly found the right solution with T-Systems: the Object Storage Service (OBS) from the Open Telekom Cloud. The provider’s cloud service enables all large data volumes to be saved in a fail-safe manner with minimal operational effort. The standard OBS solution includes automated backup, meaning that all data is automatically replicated and saved with high availability. The jointly concluded service level agreement provides for availability of up to 99.95 percent, which equates to an annual outage time of a maximum of just four hours in the event of 24/7 operation.
In practice, physicians, researchers, companies, and healthcare institutions access the metadata of the images and patient data via the internet in GitHub, and these are then loaded via a direct link from the Open Telekom Cloud. In order to work exclusively with data that does not allow any conclusions as to patient identity in line with the General Data Protection Regulation (GDPR), MHH anonymizes personal data retrospectively at 30-day intervals and only shares value ranges instead of exact figures.
At present, around 100 users use the database per month, with that figure increasing. On average, they work with an average data volume of 200 gigabytes. For comparison: one X-ray image generally equates to 18 megabytes, and one CT image is almost one gigabyte. Should considerably more cloud users access the health data in future, the installation on the Open Telekom Cloud can be expanded with additional cloud resources as required at very short notice or might also be supplemented in future with application areas such as artificial intelligence. For MHH and all parties involved, yet another reason to have opted for the Open Telekom Cloud.
We support your digital transformation with industry-specific consultation, first-class cloud services, digital solutions, and strong systemic security. With us, advanced industry expertise meets a perfectly integrated solution from a single source. Let's power higher performance – together!