Combining different types of imaging techniques gives the clearest picture of breast tumours, researchers claim.
The technique picks up cancerous tumours
Magnetic resonance imaging (MRI) scanners can be programmed to take images highlighting one type of tissue, such as fatty tissues or cysts.
Researchers said combining three techniques meant they could identify cancerous and benign tumours.
The Johns Hopkins University team said the technique may also help detect how quickly tumours may spread.
They used MRI scanners programmed to take images in different ways.
A technique labelled T1-weighted imaging is best at examining fatty tissues, T2-weighted imaging shows fluids, like those found inside cysts and 3D MRI were combined to obtain as complete a picture as possible.
Dyes were also injected into patients before the MRI scanning which would concentrate in the tumour to enhance the images.
In the study, 36 patients were scanned, half of whom had breast cancer and half who had been diagnosed with benign tumours.
The researchers assessed the scans without knowing which images came from which patient.
The combined, or multiparametric MRI technique, was able to identify and characterise breast tumours, or lesions, in all 36 patients, correctly identifying which were benign and which were malignant.
The images were even more well defined when the dye was added.
Dr Michael Jacobs, the radiologist who led the research, said each technique had its advantages.
But he added: "When all these techniques are combined into one data set, you can achieve an approach that shows the characteristics of a lesion not normally available using just one imaging technique."
He said larger studies were needed to see if the approach could be used to show how likely tumours were to spread.
MRI expert Professor Martin Leach, based at the Institute of Cancer Research said the research was a step forward, but stressed the findings needed to be reproduced in much larger studies.
"While a number of these methods are of limited value by themselves - for example T1 weighted and T2 weighted, they show that they add to the value of contrast enhanced imaging.
"On the other hand the researchers do not appear to be extracting as much information as possible from contrast enhanced imaging and they ignore information in the tumour shape all together.
"And although they obtain respectable sensitivity - how good they are at detecting abnormality - and specificity - how good they are at knowing whether it is tumour or benign - they do not do as well as some of the earlier single centre studies. "
Professor Leach said a number of image processing approaches were being tested, but there were often problem replicating results in other centres because of differences in equipment.
The research is published in the journal Radiology.