They likened their discovery to a device featured on the television show "Star Trek" that, when passed over the body, revealed the molecular secrets within.
"You don't have to invade the body in any way. We can actually obtain this information in a non-invasive manner," said Dr. Howard Chang of Stanford University School of Medicine, whose work appears in Monday's issue of the journal Nature Biotechnology.
Dr. Chang and colleagues at the University of California, San Diego, compared images from radiology scans such as CT scans commonly used to track cancer with lab tools called DNA microarrays -- gene chips -- that screen thousands of genes at a time.
What they found was a way to translate the data from the images into a computer model that could predict what was going on with the genetic material within the tumours.
"The main finding was a correspondence between radiographic images and genetic activity," Dr. Chang said in a telephone interview.
"You can think of it like a dictionary that can translate between two different languages," he added.
The result was a sophisticated computer model that predicts what is going on within a patient's cells.
The researchers said the technique may help eliminate the need for a biopsy, a procedure in which a needle is injected into a tumour to determine what type of cancer a patient has.
"This is a technology that potentially makes it simpler to very cheaply understand the genetic activity of a person's disease," Dr. Chang said.
But first it needs to be tested on various types of cancers and diseases to see whether the translation tool is universal.
"Maybe we got really lucky. Only time will tell," he said.