A simple technique for looking for breast cancer involves feeling the breasts for an unusual lump which could be a tumour, but this is like feeling for a small marble in a wrapped present. In hospitals, mammography is used to screen for tumours. An X-ray of the breast is taken and specialists look for changes in the density of the tissue. Mammograms can be uncomfortable, because the breast is pressed firmly between two plates to get a good X-ray, and the results are not always conclusive.

Dr Eli Van Houten and his team at the University of Canterbury are researching different non-invasive methods to detect and diagnose breast cancer that are more accurate and less painful.

Digital image-based elasto-tomography (DIET)

Other researchers have shown that cancerous tissue is stiffer (less elastic) than healthy tissue because it has a denser cell structure, so Eli's team are testing to see if they can detect the stiffer cancer tissue by vibrating breasts very rapidly and looking at the change in motion of reference points on the breast surface.

Eli’s team use cylinders of silicon, which they call ‘phantoms’, as models to test how well surface movement predicts the elasticity of underlying tissues. Some phantoms have stiffer objects inserted in them, to act as tumours. The team puts dots on the outside of the phantoms as reference points to measure for movement.

They vibrate the phantom and use standard digital cameras that are wired up to computers to take photos.

Normal digital cameras cannot take pictures quickly enough, so strobe light flashes at the same rate as the silicon model is vibrating. This means that, even though the phantom is moving really fast, the pictures appear still.

About twenty photographs are put together into a slow motion video – the dots on the surface actually move in an elliptical path.

The video information is compared to simulated computer models. An image is produced with different colours showing varying stiffness in the phantom. The images of a normal phantom and one with a ‘tumour’ can then be compared and differences can be investigated.

Once the method is refined, Eli and his team will start imaging the surface of breasts with the cameras and look at their motion.

When Eli and his team change from using phantoms to using people, they will have some new challenges. One of the first is trying to model the boundary of the breast tissue and the muscle and ribs in the chest wall. (This is even harder to do in men.) The team also need their computer models to allow for the fact that breasts come in many shapes and sizes, but still be able to indicate possible cancer tissue.

This technique can only be used to detect the presence of tumours. Other techniques are needed to diagnose what kind of tumours are there.

Magnetic resonance elastography (MRE)

Where DIET looks at the breast surface, this method looks at movement of tissue inside the breast and is more accurate, rather than inferring from the movement on the breast surface. However, it is expensive – instead of using digital camera imaging, it uses magnetic resonance imaging (MRI) and the equipment needed to do this is far from cheap. One picture costs approximately $4,500, and several are needed to get an accurate diagnosis.

Eli's team wants to provide a more accurate diagnosis of the type and location of the cancer within the breast. They started by building a special container that shakes the silicon gel phantoms in a very regular and precise way. The motion data from the MRI is turned into stiffness data – unhealthy tissue is about 10 percent stiffer than healthy tissue. The tissue’s varying stiffness can be calculated and made into an image.

Eli’s team is now using this technique on patients with suspected tumours.

Healthy and unhealthy tissue can be very difficult to detect using current medical imaging techniques such as mammography and ultrasound. In an MRI, the different tissue shows up much more clearly. Eli's team is taking this one step further by looking at the movement of the tissue in the MRI – with very promising results.

    Published 23 July 2007