MammoScreen® is anArtificial Intelligence (AI) solution that assists radiologistsin reading 2D&3D screening mammograms
Our Technology
The true value of our technology isn’t measured by how powerful it is, but by how much it empowers you.Winner of the DREAM Mammography Challenge, to date, the largest competition ever organized in the field of AI for breast cancer screening, we use the latest Artificial Intelligence (AI) techniques to analyze screening mammograms.
Based on cases interpreted by expert radiologists and using state of the art deep learning techniques, our neural networks learn to detect and characterize suspicious findings.
MammoScreen® builds on the latest machine learning techniques to leverage information contained at lesion, breast and mammogram level.
MammoScreen® received its first FDA clearance for use on Full Field Digital Mammography (FFDM) in May 2020, and its second for Digital Breast Tomosynthesis (DBT) in November 2021, respectively named MammoScreen® 2D and MammoScreen® 3D.
Explainability
Our AI is not a black box. Our algorithm always links the characterization of malignancy at the mammogram level to the specific finding(s). For this, we combine a top-down approach that gives us an overall assessment for the mammogram with a bottom-up approach that analyzes specific portions of it.
Reliability
Our AI software carefully combines different families of machine learning algorithms to perform a wide range of tasks: detection of findings, characterization of their level of suspicion and pairing findings from multiple views. The algorithms are tested on independent sources never seen or used by the algorithms during the development process. This avoids the risk of biased performances and is more representative of real-life conditions.
Resilience
Reducing the variability of predictions and increasing AI robustness is a key success factor. Our solution exploits state-of-the-art preprocessing and machine learning techniques to reduce possible biases between the data used for training and the data used in production.
System Integration
Installation
MammoScreen® is easy to deploy, allowing a fully-remote installation.
A virtual local server is installed on site, mammograms are automatically pushed to this server and then sent to our cloud where the AI algorithm processes the images. Then the AI-reports are sent back and stored inside your network. All the data transiting to and from our cloud are de-identified and deleted after processing.
Flexibility
MammoScreen® can easily integrate into common radiology workflows. In most cases, it can be synchronized to your existing viewer worklist or your reporting software.
Reading Environment Neutral
MammoScreen® can be deployed in any environment, from a single general practice to a large and fragmented health system. It can be used with any PACS and RIS.
Focus on Security
We use a HIPAA compliant cloud and as a double security layer, we systematically de-identify the DICOM images before sending them out to the cloud and delete them from the cloud right after analysis.