Research group at imec & Vrije Universiteit Brussel


Safe and patient-centered breast cancer screening by next-generation dynamic thermal imaging and Artificial Intelligence

Project Description

Breast cancer is the most diagnosed cancer worldwide. Well targeted screening is a proven way to prevent deaths by cancers by early diagnoses. However, current techniques come with different limitations, like being uncomfortable and painful for women, the need for body contact, radiation of mammography, limitation in screening options for breasts with dense breast tissue. ThermoBreast aims to drastically improve the prevention, diagnosis and monitoring of breast cancer while reducing the burden on women and their families, healthcare professionals and others who are directly or indirectly affected by breast cancer.  The project consists of multidisciplinary research teams – 16 partners including three SMEs and one patient association from 10 countries with complementary expertise in breast cancer diagnostics, vessel and tissue analysis, AI modelling and computer vision, health economics, social sciences and clinical science know-how.   

How does the technology work? 

ThermoBreast proposes a new solution for accurate, harmless and non-contact screening, equally applicable for all age groups and breast densities and capable of detecting pre-cancerous states. This risk-free screening technology, recently patented by the project coordinator ThermoMind LTD, can detect vascular anomalies and asymmetry, caused by cancerous growth. It combines innovative screening through multiple sensitive infrared sensors with advanced AI analysis of temporal dynamic thermal patterns. Through its patient-centred integrated diagnostics approach, this project converges intelligent computer vision, blood vessel extraction and tissue analysis with advanced information technology to deliver a medical class 1 device that will be validated in an international multicentre clinical study

Role of SMIT

To enable this technology to attain its potential in future clinical practice, it is important to keep evaluating and designing with end users iteratively throughout the innovation process. A human centered design approach is therefore needed, looking at the different actors, coming in contact with this service. Learning about local and cultural differences and translating them into requirements for the design of the technology and the design of the process when implementing.  SMIT will focus on close collaboration cancer patient association to research explainability and trustworthiness of this innovative screening approach, enhancing the quality of the user experience during screening and follow up of both patients and clinicians. We will study and develop methods to assess the ethical aspect (privacy, autonomy, transparency, bias, …) of this AI driven solution and their social impact.