By akademiotoelektronik, 08/04/2022

Imageens and I&D AP-HP: structuring and enhancing medical imaging data

Founded in 2017, the French start-up is basically a spin-off of the Biomedical Imaging Laboratory (INSERM/Sorbonne University). Imageens, in collaboration with AP-HP (France's leading hospital group), is developing the Label software aimed at exploiting medical imaging data stored in healthcare warehouses. A solution that could be a game-changer in the future of medical research and diagnostics.

Designed and offered by

35 million is the number of medical imaging data stored in the AP-HP servers and cloud within the EDS AP-HP project managed by the Innovation & Data (I&D) department. Today, many diagnoses or routine treatments go through a medical imaging device such as an MRI, a CT scanner, or an ultrasound. The images emanating from these examinations are anonymized and placed in health data warehouses (EDS) for activities of public interest such as scientific research, innovation, or the management of hospital activity. These data thus represent a gold mine of information for AP-HP researchers.

A need for visibility necessary for health actors

The usability of this data is therefore essential because it allows us to enter into a virtuous clinical evolution. Despite this potential, finding the right images within EDS is like finding a needle in a haystack. The images present in the EDS come from care and are difficult for researchers to sort, which limits their valuation. It is in this perspective that the Label platform, co-developed by IMAGEENS and AP-HP as part of an innovative public-private partnership, takes on its full importance.

Revolutionary artificial intelligence technology

In medicine, AI can streamline medical imaging research. Like a binder, Label's algorithms place tags, which can be likened to labels, on the images, one by one, according to their characteristics. Data can be categorized by body part imaged, type of imaging exam, or even quality. Researchers no longer spend time pruning a mass of images, they now study between data already sorted and annotated by algorithms - an essential shift from "big data" to "smart data". The potential of this technology and its alignment with the priorities of the national health system, has enabled Imageens to emerge as a French leader in the field. Imageens was notably invited to the National Assembly during the Digital and health for all event: what actions for the public health of tomorrow? co-organized by AI for Health, so that Robert Baldwin, President of Imageens, can share his vision on health data management.

Imageens: structure and enhance medical imaging data Imageens

Ethics in the use of data

By collaborating with the AP-HP and its health data warehouse, Imageens respects the standards established by the CNIL. All data processed by Label's artificial intelligence is anonymized. Personal information is not attributable to an individual, to ensure patient privacy. The Imageens technology is grafted directly to the hospital device. To guarantee its ethics in the processing of personal data, the start-up is in constant collaboration with the world of radiology which monitors the algorithms, allowing them to be in permanent progress.

Medical artificial intelligence: the future of innovative solutions

Organizing and using medical data using AI is tomorrow's challenge. In the near future, it is more than likely that healthcare professionals will use an artificial intelligence tool to make a diagnosis. The objective of Imageens and AP-HP I&D is simple: to provide a bank of structured medical data that can be used by researchers and AP-HP partners. Thus, Imageens would become one of the leading innovative solutions for the enhancement of care data via AI.