Your tissues have
so much to tell
Next Generation Tumor Board Meetings.
Our mission is to optimize the management of cancer patients. Thanks to AI and digital pathology, we offer support for Tumor Board meetings by integrating diagnosis, molecular biology and therapeutic recommendations directly on the image of the digitized biopsy.
We are defining a new approach
Deep Learning is able to identify morphological abnormalities that are currently unknown to pathology classifications. By combining it with medical expertise, we propose new tumoral
biomarkers that refine the diagnosis and predict the patient’s response to the therapeutic arsenal available to doctors.
We make Precision Oncology accessible.
Tissue images are a data gold mine. Pathologists analyze tissue architecture, cells and nuclei abnormalities, their pecial relationships at small and large scales. The computing power of the neural networks we design allows us to decode visually imperceptible information.
Our algorithms are designed to support the oncologist, surgeon, radiotherapist and pathologist during their Tumor Board meetings. They include a diagnostic aid for the pathologist and a virtual molecular biology analysis on the image (NGS, transcriptomics, epigenetics, etc.) to help the oncologist choose the right complementary examinations to carry out. Based on this information, our databases and the literature, we statistically estimate which treatment options have the best chance of working out for a specific patient.
CE-marked and patent-protected automated biopsy and smear image readers. Our medical device software improves the reliability of screening and diagnosis by providing a fast, automated dual reading. They then show the most at-risk areas to the pathologist, who will decide on the diagnosis.
These tools make the cancer diagnosis more reliable by acting as quality controls to support physicians.
This programme has been awarded a French Tech Emergence grant and is supported by the Bourgogne-Franche-Comté region.
We are developing a virtual molecular biology solution that predicts the presence of molecular tumor alterations (DNA sequencing, transcriptomics, epigenetics, etc.) based on the analysis of morphological anomalies. The aim is to optimize access to molecular biology platforms to urgently orient patients who can respond to innovative targeted therapies and help reducing not relevant testing. We are working on all solid tumors, for more information we invite you to consult our publications.
This project is a 2022 winner of the Bpifrance i-Lab national innovation competition.
The routine use of neural networks is still a challenge. Indeed, the performance of the algorithms tends to drop once used in real life due to the variability of the data to be taken into account (type of
scanner, staining machine, technician, etc.). Moreover, neural networks are often very confident in their predictions, even when they are wrong. We propose UmmonQC, a set of tools to monitor the performance of AI algorithms, from their training phase to their daily use. Indeed, the trust that a doctor, a patient or a representative of a healthcare organization places in the results generated by an AI solution must be earned at every use.
This project has received a grant from the Fonds Régional d’Aide à l’Innovation of the Bourgogne-Franche-Comté region.
If you are interested in learning more about our products or are looking for expertise like ours for your own solutions, we would be happy to talk to you.
Ce qu’apportent nos solutions
gained to initiate therapy.
optimisation of molecular biology platforms.
+1 billion $
of annual savings for health systems.
UMMON HealthTechIn a few figures...
Year of creation
of average age
Quality certification :
A Deep Learning solution for triaging patient with cancer according to their predicted mutational status using histopathological images
The presence of mutation in cancer can be associated with a response to a targeted cancer therapy. Therefore it has become an important information while it
Abstract—In computer vision, data shift has proven to be a major barrier for safe and robust deep learning applications. In medical applications, histopathological images are often