A hand-held probe could cut the number of patients who develop infections following operations.
The probe detects the colour of blood cells in the wound
The device can predict, as little as 12 hours after surgery, whether a wound is likely to become infected.
The results could help doctors and nurses act quickly to prevent infections in the most vulnerable patients.
Details of the University Hospital of North Durham breakthrough feature in the British Journal of Surgery.
Surgical wound infections, including those caused by MRSA, are a significant cause of problems for patients trying to recover from major operations.
Contamination of the wound with bacteria is one obvious cause for this, and there are strict hygiene procedures in place to try to minimise the risk of this in operating theatres and on wards.
However, some patients have a far higher risk of developing an infection.
This is because not enough oxygen-rich blood is reaching their wound.
This not only slows down healing, offering more time for an infection to take hold, but the lack of oxygen also hampers the body's immune system as it tackles harmful bacteria.
The new approach works on a simple principle - blood cells carrying oxygen are bright red, while blood cells which have no oxygen are purple in colour.
The Durham team used a handheld device which bounces infra-red light into the skin around the wound.
The signal that reflects back is different depending on the colour of blood cells in the wound.
The study looked at 59 patients recovering from abdominal surgery, who were scanned at 12, 24 and 48 hours after their operation, then examined a week later to check for signs of infection.
In all, 17 patients developed an infection in their wound, and scans from these patients had suggested significantly lower levels of oxygen in the tissue surrounding the wound.
Lead researcher Dr David Harrison said it could make a real difference to patients.
He said: "The beauty of this device is that there is no need to even remove the transparent film that is placed across the wound after surgery - it's completely non-invasive.
"To be able to identify those patients most at risk of infection at just 12 hours after surgery gives you the opportunity to actually do something about it."
Dr Andrew Berrington, a consultant microbiologist, said: "It sounds very promising, and is certainly a potentially useful development in infection control.
"But it needs to be prospectively evaluated in more patients, and in more clinical situations, with a robust definition of what constitutes an infected wound.
"The main questions are does it really predict infection with a high degree of accuracy, and if so does this translate into a genuine clinical benefit?"
The technique is now to be the subject of a much larger clinical trial in the US after a major biotechnology firm found out about the Durham team's results.