Nantes-based medtech adds two new features to its Breast-Slim View software suite. Sylvie Davila, its founder, foresees an imminent take-off of artificial intelligence in mammography. It intends to raise 4 to 5 M€ to accompany it.
Nantes-based medtech Hera-mi adds two features that are almost unique in the world to its Breast-Slim View software, a breast cancer screening support solution based on artificial intelligence. For the record, the original innovation of Hera-mi, founded in 2017, is negativation. This consists, thanks to AI, of automatically identifying and masking all the physiologically normal elements of the breast (vessels, connective tissues, mammary gland, muscles, etc.) to only make visible the areas potentially at risk of malignancy. This reduces the volume of information and facilitates the interpretation of the radiologist.
AI detects non-compliant images
Mammography Technical Evaluation (MTE), the new functionality unveiled, makes it possible to evaluate in real time, via an algorithm, the quality of images in mammography and tomosynthesis in order to identify without delay technically insufficient images. The AI automatically analyzes several essential criteria such as breast compression, positioning (pectoral and nipple, for example) and the general quality of the image on criteria such as artifacts, folds…
Avoid anxiety-provoking reminders
The technology has several advantages, knowing that to date 10 to 15% of the shots must be redone. This represents a waste of time for the patient, not to mention the anxiety of being called back for a second appointment. In addition, the X-ray dose is also limited. What’s more, MTE presents a personalized tracking and improvement axes interface that directs radio manipulators towards rapid correction for an optimal image. In France, the national organized screening program for breast cancer provides for a second reading outside the practice. Here again MTE must contribute to reducing the recall rate.
Mammograms, ultrasounds and MRIs
A third technological stratum called Breast Insight Linker should be certified in 2024. It allows correlated interpretation of mammograms, ultrasounds and MRIs in order to perform biopsies precisely. “Today it is the eye of the radiologist who does this work”, notes Sylvie Davila, CEO of Hera-mi, specifying that these three sources are difficult to associate because, depending on the techniques, the images are captured on patients in a standing or lying position.
Hera-mi, which employs 27 people, including 15 in R&D, does not communicate the amount of its turnover but reports about thirty equipped mammography centers, the only French market having nearly 2,500 mammographers knowing that the AI technologies are still absent. “2023 will see the real boom in AI,” predicts the leader, who intends to raise €4 to 5 million in the first half to support this boom.