Medical ultrasound is a safe imaging modality with relatively inexpensive and easily portable equipment. As opposed to other modalities that require patient transportation to a dedicated examination room (e.g., CT and MR), ultrasound can be used bedside. Ultrasound imaging is therefore widely accessible and used for many medical applications. Downsides include that the ultrasound field of view is limited and therefore provides a reduced overview of the patient’s anatomy, image quality is sometimes poor, and both acquisition and interpretation of images are operator dependent and require a high degree of expertise.
For several clinical applications, CT or MR images from a prior examination is already available at the time of ultrasound imaging. Being able to combine these images with information from ultrasound in real-time during the ultrasound exam could be of huge benefit, both for diagnostic purposes and for guidance of minimally-invasive therapy.
Combining information from several image sources is referred to as image-fusion. A key to image-fusion is to determine the spatial relationship between images defined in different coordinate systems and with different resolution. Methods devised to find this relationship are known as registration methods, and has been a long-term research focus in SINTEF Health Research. This functionality is available on several high-end commercial ultrasound scanners, but current solutions require manual landmark selection for registration, and are therefore quite labour intensive. We have therefore developed an automatic, real-time method that has the potential to increase accuracy, reduce time and labour, and open for increased use of image fusion by more operators and for new applications.
In this project, the aim is to investigate the technological and commercial potential of a software comprising this method. Objectives include to examine application areas, and to implement and validate the technology in a lab setup.