A news report about a dog that could smell cancerous lesions triggered a Penn researcher to consider the far-reaching possibilities of smell technologies and their use in detecting and diagnosing all kinds of medical maladies.

Using technology developed by British company AromaScan, C. William Hanson III, M.D., associate professor of anesthesia and chief of anesthesia/critical-care medicine at Penn's Medical Center, set out to apply an "electronic nose" to sniff out disease.

The notion is not necessarily a new one, but one that has not been explored due to the perceived subjectivity and lack of applied technology.

"It goes back thousands of years," Hanson said of the diagnostic powers of smell. "We've kind of gotten away from it because we didn't have the tools to objectively do anything with it."

The electronic nose, however, could change all that. Already, Hanson has applied the nose (which he said basically looks like a personal computer), to diagnose lung infections faster, sooner and cheaper than current, invasive techniques.

The breath of diabetes patients is a unique factor in diagnosing the disease. Other infections could be next, Hanson said, particularly in liver transplants, where the patient's breath smells different if the liver is failing.

While aroma analysis is not a widespead area of research, Hanson said it is on the verge of an explosion of interest in research medicine.

A study Hanson conducted with Heather Steinberger, R.N., a research nurse coordinator in the Department of Anesthesia, analyzed breath samples of already-diagnosed pneumonia patients and non-pneumonia patients.

The resulting computer-generated patterns provided clear distinctions between the patients who were infected and those who were not. Hanson published those findings in the September supplement of Anesthesiology, and presented them in October at the annual meeting of the American Society of Anesthesiologists in San Diego.

The electronic device offers major advantages over conventional tools, Hanson said, and could eventually extend far beyond detecting pulmonary diseases.

"The potential is huge," Hanson said. "Rather than waiting two or three days for the results of a bacterial culture or relying on chest X-rays that may be inaccurate, doctors can almost instantaneously evaluate their patients for infection."

These concepts of aroma analysis are not used as widely as they once were, Hanson said, because physicians felt the methods lacked objectivity. And, naturally, Hanson does not foresee the electronic nose replacing other standard tests for infection or pneumonia.

Still, it's hard to ignore concepts based on the facts that dogs can identify when their owners are about to have a seizure, or sniff out malignancies, simply based on smell.

Already, Hanson said, several different computer patterns have been discovered associated with certain bacteria and fungi, and more should follow.

At this point, spotting cancer or liver failure solely on the basis of smell is "clearly fantasy stuff," Hanson said. "But, it's also clearly not out of the bounds of possibility."