Datrix, an international ecosystem of B2B vertical software companies, leverages the empowering value of Artificial Intelligence as a data activator, with a strong focus on innovation through research. Thanks to its internal R&D, Datrix has launched collaborations with over 20 international research centers and universities, carrying out projects funded by European and Italian funds for a total of over 30 million euros since 2016.
This activity has led to the development of concrete and high-impact technological solutions, with a particular focus on the fields of Cybersecurity and Healthcare.
A distinctive aspect of these innovations is that they are not only born in the laboratory, but are developed in collaboration with Italian excellences, such as the Politecnico di Milano or the Vittore Buzzi Hospital in Milan, representing a real contribution to the scientific and healthcare community, but also to the Italian social fabric
We had the opportunity to talk about some of these technologies with Renzo Vanna, Senior Researcher at the CNR at the Institute of Photonics and Nanotechnologies of the Politecnico di Milano, who collaborated with us within the framework of some projects funded by the European Union that saw Datrix as a technological enabler. Renzo has been a unit leader in three European projects focused on biomedical and diagnostic applications, and he illustrated the path that led to the creation of an innovative tool for oncological medical research.
Raman Technology and Artificial Intelligence: a new standard in biomedical research thanks to the CRIMSON project
CRIMSON is a project funded by the European Union that aims to revolutionize oncology through an advanced microscope based on the Raman effect with a significant impact not only on the daily work of researchers, but – potentially – on the quality of life of cancer patients.
“Traditionally, the study of biological samples begins with laborious and invasive processes, such as hematoxylin-eosin staining, a technique used for over a century. However, this method, although effective, is based on stains that provide indirect information on tissue chemistry, requiring the expert eye of clinicians and pathologists to subjectively interpret the structures visible under the microscope” Renzo explains to us.
CRIMSON revolutionizes this approach by using precisely the Raman effect, which allows for the direct analysis of the chemical and molecular composition of tissues, without altering the sample. Specifically: “by illuminating the sample with light, a spectrum is obtained that reveals specific details of the molecules present, without destroying the tissue. This process allows for biomolecular information to be obtained with sub-micrometric, extremely detailed resolution. For researchers, this means being able to examine tissues and cells with very high levels of precision and specificity. Furthermore, the use of artificial intelligence to analyze the obtained spectra allows for a significant acceleration of the diagnostic process, quickly identifying the molecular composition of the samples and, last but not least, producing virtual images understandable to clinicians and pathologists”.
For patients, especially oncology patients, this AI-based technology could represent a significant improvement in diagnosis and treatment. By being less subjective and more precise, CRIMSON could enable faster and more accurate diagnoses, contributing to improved care and, consequently, the prospects and quality of life of patients.
The impact of CRIMSON on leukemia research and patient health: a practical example
The CRIMSON technology also has a significant impact on leukemia research. Leukemia, with its over 40 subtypes, represents a complex challenge for diagnosis and treatment. Currently, subtype classification often requires microscopic analysis of blood or bone marrow, a process that can be invasive and relies heavily on the pathologist’s experience.
Renzo clarifies:
“Using Raman technology, CRIMSON can allow for a biochemical molecular classification in an objective and automatable way. Raman, in fact, allows for detailed information on the chemical composition of cells to be obtained, overcoming the limits of traditional visual analysis.
In the published study with the technological support of Datrix, the ability of Raman technology to identify leukemia cells and differentiate them into various subtypes was evaluated, comparing it with the diagnostic accuracy of a pathologist. The results were promising, showing that the Raman technique can distinguish the different subtypes of leukemia with an accuracy in some cases comparable to that of traditional methods, with the added value of offering an automatable, objective and rapid assessment, a relevant approach in contexts where the presence of expert pathologists is limited or absent”.
The implementation of CRIMSON in clinical practice will require further development and validation, but the potential benefits are enormous. The ability to provide rapid and precise diagnoses can significantly improve the management of leukemia, leading to timely and targeted treatments.
Contrary to the fear that artificial intelligence might replace human workers, CRIMSON demonstrates how technology that harnesses AI potential in accelerating data can expand expert skills, speeding up diagnostic processes and offering more precise decision support tools.
Ultimately, the goal is to provide pathologists with advanced tools that can integrate with their experience and thus improve the accuracy of diagnoses.
RamApp: ease of use for doctors and researchers to bridge the technological gap
Exactly within the CRIMSON project, RamApp was born, a free web tool co-developed by Datrix, which uses AI technologies to simplify the analysis of hyperspectral images, offering valuable support to doctors and researchers who are not IT experts, facilitating their interpretation capacity and bridging the technological gap.
Vanna recalls: “During the COVID-19 period, not being able to access the laboratories for many weeks, we took the opportunity to critically examine how we analyzed our data. In this context, we developed more intuitive and faster methods to study a significant amount of data. We therefore considered it useful not to keep these advancements within our group, making these advantages accessible to a wider audience. RamApp moves in this direction, allowing for the uploading and analysis of hyperspectral images directly online, offering users an intuitive interface to explore and interpret complex data without the need for advanced programming skills. To date, it can be used by undergraduate and doctoral students in just a few days, with a pool of about 250-300 users globally, including researchers and doctors.”
This tool represents a novelty in the way data is analyzed and shared in the scientific community. Being an online and free application, it facilitates collaboration and access to powerful analysis tools without economic or technical barriers. Furthermore, it is continuously updated with new features based on user feedback.
The use of RamApp by researchers from prestigious institutions like MIT testifies to its international relevance and solid reputation in the global scientific community. The platform has been used to produce results published by international journals such as Nature Communications and Science Advance, recognitions that attest to its innovative value and ability to improve research in the biomedical field.
Renzo finally concludes: “ease of use is the heart of RamApp, but the collaborative aspect is equally important: the results of the analyses can be immediately shared with colleagues in other parts of the world, facilitating direct and multidisciplinary comparison. This distributed approach allows for the aggregation of multiple opinions on the same sample, making RamApp a fundamental tool for comparative research and knowledge sharing. RamApp’s cloud technology has been optimized to ensure efficiency and calculation speed, with an intuitive interface that allows for remote computational load management, without impacting the resources of individual devices.”
