Date: 27th Jul 2022
The use of AI algorithms in Azure to analyze diagnostic test images in oncology translates into improved quality of life for the patient, achieving a reduction of up to 80% in interpretation time by the physician and detecting users who will best respond to a particular therapy.
Quibim has positioned itself as one of the primary references in the medical imaging field, managing today more than 10 million images in the Azure cloud, which meets the most demanding security requirements in the medical field.
Among the company’s projects is a model that, through the analysis of chest CT scans, determines whether lung cancer patients will respond to new immunotherapy treatments or to know, from a prostate MRI, whether prostate cancer patients will have a relapse in the following years.
Madrid, July 27, 2022 – In the field of health, getting an accurate diagnosis on time is a priority. Artificial Intelligence allows a huge qualitative leap through machine learning mechanisms, which learn and improve with the analysis of large amounts of information, to help medical staff to offer better diagnoses and treatments. The Spanish company Quibim (acronym for Quantitative Imaging Biomarkers In Medicine), which designs and creates pioneering tools that extract information from medical images, uses this technology to accelerate diagnosis and identify possible diseases early, increasing the chances of survival and avoiding more invasive and harmful therapeutic techniques for the patient.
Relying on Microsoft Azure cloud services, Quibim has centralized the image data needed for the creation of its AI models, creating a repository of more than 10 million anonymized medical images for some of the most relevant innovative projects and biobanks worldwide in pediatric cancer (Primage), prostate cancer (ProCancer-I) or COVID19 (Imaging COVID-19 AI), among others.
Quibim, headquartered in Valencia, with offices and subsidiaries in Madrid, Barcelona, Cambridge (UK), and New York (USA), was founded with the ambition of making medical imaging one of the catalysts of precision healthcare. These tools provide objective results from image analysis and reduce the physician’s interpretation time by up to 80%.
In recent years, both the number of relevant Quibim customers and new treatments in the fields of oncology, rheumatology, and neurology have been increasing, and the need to predict the clinical evolution of patients from medical images has arisen.
In this scenario, the company sought to create AI tools that would analyze diagnostic images under the new concept of ‘medical imaging panels’. Especially in cases of lung and prostate cancer, to identify these patients early to be treated at an earlier stage of the disease. For example, with a model that, through the analysis of chest CT scans, determines whether the patient will respond to the new immunotherapy treatments or to know, from a prostate MRI, whether he/she will develop a relapse and metastasis in the coming years.
Quibim has a global business strategy, which requires a modular, flexible, and fully scalable cloud infrastructure, capable of serving a large volume of customers located around the world. For this reason, the company uses a multitenant architecture, which optimizes resources and improves the security and reliability of both the environment and the customers.
Unlike other healthcare centers and services, which have on-premises developments and deployments integrated within their private infrastructure, Quibim has opted for a Cloud-Smart approach that allows balancing the adoption of the cloud with the circumstances and objectives of the organization, having a fully automated infrastructure ready to take on any challenge. And for this, they have relied on Microsoft Azure cloud services for their compliance in terms of data security and privacy.
“Working with a secure, private, and traceable tool is one of the biggest challenges in this sector. The fact of having to be certified as a medical device to be used in clinical routines, or the requirements of healthcare centers and data processing regulations, implies compliance with specific regulations. For us, Microsoft has been the best choice thanks to the large investment they have made over the years in these fields”, explains Bas Hulsken, Chief Technology Officer at Quibim.
As a cornerstone of its security strategy, the company has consolidated its management, identity, and local access systems within Microsoft Azure Active Directory. Application Gateway has also been implemented to improve the security of access to the platforms. The orchestration and management of algorithms responsible for the analysis of the different parts of the body (breast, brain, prostate, colon…) are carried out in Azure Kubernetes Services, while the storage of medical images is managed by Quibim. -more than 10 million- is deployed in Azure Storage Accounts. Finally, Azure Webapps makes it easy to deploy web applications.
Microsoft services help Quibim to have a flexible, secure, and powerful infrastructure, enjoying tools and services tailored to their needs to keep operational key aspects such as labeling, storage, and management of clinical data, as well as scalability in cost and time.
Quibim’s next great challenge is to create the world’s largest biobank of medical images, generating a universal and centralized repository of cases classified according to different variables, such as pathology, anatomical region, or the technique used to obtain the image.
Over the next five years, Quibim aims to introduce significant improvements in the care of thousands of patients – especially children – through virtual biopsy diagnostics: “We hope to be able to monitor patients closely, with a high degree of accuracy and without the need for a highly invasive procedure, avoiding hospital surgeries and the hassle of testing. This is the case of children, who for ethical reasons cannot undergo a standard biopsy, or in certain lung cancers, whose location makes diagnosis complicated,” says Ángel Alberich-Bayarri, CEO and co-founder of Quibim.
In the long term, the company sees the future in digital human twins: “Our current healthcare model is reactive. Patients are treated when they are sick. We have projects where our goal is to try to anticipate symptoms using a digital twin, to monitor the inside of the human body when we are healthy, and to be able to have information and a model about the age of each organ, preventing possible diseases” concludes Ángel Alberich-Bayarri.