• Add to new collection
    Rights: The University of Waikato
    Published 21 July 2007 Referencing Hub media

    Dr Eli Van Houten and Dr Ashton Peters at the University of Canterbury talk about how they’ve gone about developing a new technique to screen for breast cancer.

    There are a number of challenges involved in developing new medical imaging techniques. The process usually starts with proving you can do it in theory, using pen and paper or computer simulations, and then it moves to phantom studies. In this case, the phantoms are ‘breasts’ made of silicon. Once they can accurately work out the internal structure of the phantom breasts, which have fake tumours in them, they can then move to imaging real breasts, with the associated challenges of working with real tissue and real people.


    When you are developing a new medical imaging technique, the standard process is to start with a theoretical development and prove that you can actually do it somehow in theory, either on a piece of paper by writing out some equations or in our case we prove the theory to ourselves using computer simulations.

    The next step after that is usually to move to phantom studies where you take phantoms, in our case we make them out of silicon gel and we used some earlier ones that were made out of actual gelatin just like you'd eat for dessert at night, and we’ve construct phantoms where you can actually control what it is you are looking for and the reason for this is that that way you know when you are getting the image right. So we have built phantoms where we know what the elastic property distribution in the phantom is, so when we get our resulting images back we can say, yeah, okay we have captured that elastic property image accurately, or no we are missing something, or there is something here that shouldn't be here. So a phantom experiment can only tell you so much.

    A real breast or a real biological tissue in general is very complex. To describe the material properties of real biological tissue is actually quite complicated because it has visco-elastic properties, it has anisotropic properties, there is a certain amount of material non-linearity, there is all these things that make describing real tissue very difficult, and it’s actually almost impossible to mimic those perfectly in gelatins. So the final step in developing any medical imaging modality is to move on and start imaging actual tissue because you will only really experience the full gamut of mechanical properties when you are working with real tissue.

    So we have a major step still to take in this research which is how do we take the results that we got from silicon phantoms and how we do then apply those same techniques to real breasts. We’d have a silicon phantom that might be the same material all the way through. In a real breast we have areas of different stiffness naturally within healthy breasts.

    So it’s going to be a big challenge for our project to move from the phantom data that we currently have which is working quite well to applying this system to real people, and there’s a number of challenges, some of which we’ve identified, like how do we actually get the data. It’s much more difficult and challenging than just getting it from a small phantom that we can play with you actually have to get real people and then there’s the other aspect of the inverse problem, so how do we go from motion to stiffness – that becomes more difficult when we know that the internal stiffness of a real breast isn’t as simple as the phantom models that we’ve been using so far.

    Patrick J. Lynch and Dr C.Carle Jaffe