SCARtool · Scattered radiation reduction tool to improve computer-aided diagnosis performance in digital breast tomosynthesis
Horizon 2020 — Marie Skłodowska-Curie Actions
- Duration
- 2015-09-01 → 2017-08-31
- EU contribution
- €158,122
- Participants
- 1
- Scheme
- MSCA-IF-EF-ST
Lines connect the coordinator with its partners.
Results in brief
Scattered radiation reduction tool to improve computer-aided diagnosis performance in digital breast tomosynthesis
Breast cancer is the most common form of cancer in the female population and represents the principal cause of cancer mortality in women worldwide. Around the world, every 19 seconds a woman is diagnosed with breast cancer and also, a woman dies from breast cancer every 74 seconds. Thus, further research, based on early detection, is necessary to reduce its high impact on the population. Breast cancer screening programs (BSPs) is the current approach used to find early cancer signs in an asymptomatic target population. Screening population is invited to have regular mammograms (MG). However, digital breast tomosynthesis (DBT) has shown potential to be the future imaging technique for BSPs. DBT setup is similar to MG, however, DBT images are generated from several low-dose image projections at different angles, producing pseudo-3D images of the breast. The SCARtool project investigates algorithms to improve image quality in DBT. The algorithms developed within the project have the aim to reduce scattered radiation in the DBT images, a type of radiation which might affect lesion detectability. Therefore, SCARtool's main goal is to improve early detection of breast cancer signs via DBT imaging.
Data: CORDIS, © European Union
Project objective
Breast cancer is the most common cancer diagnosed in women worldwide. This has motivated the establishment of breast screening (BSPs). X-ray mammography is the current gold standard technique in BSPs for early breast cancer detection. However, this technique suffers from performance limitations due to tissue superposition which can either mimic or obscure malignant pathology. For this reason digital breast tomosynthesis (DBT) is currently being investigated.DBT consists of a series of projected images of the breast with potential to overcome the aforementioned limitations. However, DBT shows high levels of scattered radiation within the image receptor which can reduce significantly the visibility of breast lesions and the lesion detection performance of computer-aided diagnosis (CAD) systems. CAD systems are frequently used in BSPs to provide useful information to radiologist regarding breast lesion location and characterisation, thus CAD’s performance is critical to achieve acceptable breast cancer diagnosis rates. The aim of this fellowship is to develop a novel tool to reduce scattered radiation directly from clinically acquired DBT projections. These scatter-reduced images will be incorporated, for the first time, into a new DBT CAD system to significantly improve the detection of subtle lesions as microcalcification clusters (MCC). The scatter reduction will be achieved after adapting an existing convolution-based algorithm for synthetic DTB images to work with clinically acquired DBT images. The CAD experience of ViCOROB and the DBT experts’ support from UDIAT guarantee the success of this project as well as the correct transfer of knowledge from the lab to the clinical site. The proposed tool will enable radiologists to detect more accurately breast lesions and therefore reduce the mortality in breast cancer patients. It is also planned to commercialise the output CAD tool to clinical sites through existing medical imaging companies or via a spin-off.
Original text from CORDIS.
Participants
- UNIVERSITAT DE GIRONA · GironaCoordinatorSpain
Links
Data: CORDIS, © European Union