Development of image reconstruction algorithms for photoacoustic breast imaging
Project SACAMIR is a cooperation between the industrial and academic partners Achmea, the Friesland zorgverzekeraar, the University Twente and PA Imaging R&D, with the goal to develop algorithms for the application in medical imaging and in particular photoacoustic mammography. Photoacoustic mammography applies photoacoustic technology with the aim of early breastcancer diagnosis in women of all ages. This technology is noninvasive, painless and without the need for exposure to ionizing radiation (X-rays). Photoacoustics, in combination with existing imaging-technologies has the potential to reduce the number of “false positive” and “false negative” outcomes of breast cancer diagnoses significantly.
13.000 women per year are diagnosed with breast cancer in the Netherlands. It is the most common form of cancer among women and yearly, 3.200 women die as a result of it. Currently, several imaging technologies exist to detect breast cancer (X-ray mammography, US-imaging and MRI). The value of these techniques is not in question, but room for improvement is present.
Project SACAMIR entails both practical aspects of photoacoustic measurements as well as theoretical aspects of modelling physics, with the aim to develop image reconstruction algorithms that are both of high quality as well as practical in the sense that they are fast to reconstruct an image. The project has run in parallel with a clinical study on volunteers and patient-volunteers. This offered the unique possibility to test algorithms and measurement protocols in an early stage. In the case of patient-volunteers, the obtained images are being compared with clinical images of conventional diagnostic imaging modalities.
Project SACAMIR has resulted in very significantly improved image reconstruction algorithms, thereby proving the potential of photoacoustic mammography as a welcome addition to existing imaging modalities. Furthermore, the results of the research into the use of “artificial intelligence” in developing learning algorithms, offer a view into future improvements.
The goals of project SACAMIR were challenging; on the one hand, develop algorithms and measurement protocols that deliver the best possible images in terms of resolution and imaging depth and on the other hand, to ensure that these algorithms are fast enough for use on advanced, but cost pricewise, manageable workstations. These deliverables have been achieved.
