L'immagine mostra un'illustrazione relativa al progetto SABREEN


Smart Assistant for BREast screENing

SABREEN is the solution designed by HTLab (Healthcare Technology Lab), patented with the obligation of humanitarian use, developed in collaboration with the staff of vEyes, the Computer Science Department of the University of Verona, the Department of Electrical, Electronic and Computer Engineering of the University of Catania and the CNIT (National Inter-University Consortium for Telecommunications), in order to allow young women to perform breast ultrasound screening independently, using a linear probe connected to an Android smartphone and to detect the presence of suspicious masses through an artificial intelligence system, so as to send an alert to a senology center assigned by SABREEN, for the necessary clinical evaluations and the subsequent convocation to visit the patient.

Breast ultrasound, in fact, is an optimal diagnostic tool for testing a young breast (generally under 40 years of age, or in any case in fertile age, which can extend statistically up to 48 ÷ 50 years), which has a glandular component very rich, difficult to explore radiologically with a mammogram. After the age of 50, SABREEN still does not lose efficacy, as it can still be used for breast screening. However, it becomes advisable to place side by side a mammogram (which SABREEN will suggest to the patients to carry out, each time they perform the test, if the age is compatible with what indicated above).

L'immagine mostra uno schema esemplificativo della rete neurale di SABREEN

As is known, an ultrasound examination requires manual skills and is an examination highly dependent on the operator. To simplify the execution of the test and make it clinically reliable, the probe is equipped with appropriate sensors that can guide the woman in the movement of the probe, discarding the scans made with incorrect maneuvers. During the scans, no ultrasound images will be shown on the display, nor will any clinical evaluation of the results of the scans be provided. The smartphone merely confirms that the requested scans have been performed correctly.

The application on the smartphone also remembers when it is time to perform ultrasound screening, taking into account the time interval since the last check-up and that the patient falls between the seventh and fourteenth day of the menstrual cycle. The neural network at the edges also builds a 3D model of the breast, using two-dimensional ultrasound images produced with the probe, simplifying the analysis of the suspect masses to the doctor.

Categories: Project