Early warning

The detection of skin cancer could be further improved with a new technology that aims to help dermatologists find subtle changes in the shape and size of moles.

The development — from researchers at Photometrix Imaging, a spin-out from Glamorgan University's Medical Imaging Laboratories — could also be used by the cosmetics and computer gaming sectors.

The technology, called Measurement of Area and Volume Instrument System (MAVIS) II, captures and measures 3D images of dermatological anomalies, such as malignant melanomas, using a high-resolution camera, a stereo adapter and a computer.

Peter Plassmann, chairman of Photometrix, said that moles become a cause for concern when they develop a 3D structure and no longer lay flat on the skin.

'With our technology we can spot that moment in time and say "now it's time to do something about it",' he said.

Currently, non-invasive assessments are made by doctors who observe the anomaly directly or look at digital images of it. Subtle changes, which can indicate a melanoma is malignant, can be missed unless invasive testing is carried out.

The MAVIS technique, which is completely non-contact, produces 3D images using specialised software based on a technique known as stereophotogrammetry.

The technique operates on the principle that human eyes are placed slightly apart, and when a person's eyes observe the same object, each will produce subtly different images because the object is shown at slightly different lateral positions.

MAVIS uses a normal digital camera fitted with a dual lens for taking stereo photographs. Two tube-shaped LED projectors on the base of the camera produce two beams of light, forming a single spot of light when the camera is held at the right angle and distance from the subject.

The picture is taken and the photo file is loaded into a computer via a memory card. The lateral displacement is a measure of the distance between the object and the 'eyes' of the camera. The software program processes the data to identify common points on the image taken from both the left and the right. Within minutes, an on-screen 3D map of the object is built up.

The technology was originally designed to monitor the kind of chronic wound that either heals very slowly or becomes deep, resembling a crater in the skin.

Plassmann and his team realised that doctors needed a highly accurate monitoring system so they could determine whether the treatment of such wounds was producing any effect.

Existing techniques involve measuring either the area or volume of wounds. The current method of measuring the area is to place a transparent acetate sheet on to the wound and trace its perimeter. The tracing is then placed on to graph paper and the number of squares is counted. This method of determining the area of a wound can lead to inaccuracies.

Similarly, there are two main methods for measuring the volume of a wound. The first is to fill it with a saline solution. The volume dispensed from the syringe equals the volume of the wound. The main source of error is that the wound absorbs the saline. The second method is to fill the wound with an alginate or silicone-based paste and weigh the amount of material used.

The main problem with these measurements is they require contact with the wound, which may be painful to the patient and incur the risk of infection, a situation the non-contact MAVIS system is able to avoid.

The National Physical Laboratory conducted tests on the accuracy of MAVIS and found it achieves a measurement precision of five per cent or better, which is higher than existing techniques.

The MAVIS developers plan to incorporate wireless image transfer in the future.

In addition to skin cancer detection and wound monitoring, Photometrix is also targeting its 3D measurement technology at the cosmetics and plastic surgery sector. The company is in talks with manufacturers and service providers in the sector to develop a device that could, for example, determine the effectiveness of products for wrinkle reduction.

The technology could also be used for full-body scanning with computer games like Second Life where the production of a 3D model of a player's virtual 'real-life' appearance, or 'avatar,' is currently a time-consuming and expensive manual task.

'Our main problem is there are virtually hundreds of applications out there and we just need strong partners to help us to exploit that,' said Plassmann.