Dr. C. William Hanson, III, a University of Pittsburgh School of Medicine lung expert, said the electronic nose is so sensitive that it can identify specific types of bacteria such as strep or staph "by making a fingerprint of the odor."
In his pilot study, the nose was as accurate at diagnosing pneumonia as a traditional test.
Hanson presented his findings at CHEST 2002, the annual meeting of the American College of Chest Physicians.
Hanson used the nose, called Cyranose and is made by Cyrano Sciences Inc. of Pasadena, Calif., to test air samples drawn from the breathing tubes of patients on ventilator support.
Ventilated patients often develop at type of pneumonia called ventilator acquired pneumonia, or VAP. Usually physicians use X-rays or sputum cultures to diagnose the disease but those methods can take hours or days.
Hanson and colleagues screened 415 hospitalized patients who had VAP, identified by a measurement tool called a clinical pneumonia score. Cyranose was tested in 20 patients who had pneumonia scores of 50 or higher, which indicated that they had VAP, and 20 patients whose scores were 20 or lower, meaning that they did not have VAP.
When Hanson compared Cyranose's diagnoses with the pneumonia scores, he found that the electronic nose had correctly identified the pneumonia patients.
Cyranose uses 32 carbon-black sensors imbedded in a polymer grid. As the odor passes over these sensors it produces unique patterns that are displayed in two-dimensional maps or dot patterns on a computer screen. The computer then uses pattern recognition algorithms to identify the odor.
In an interview with United Press International, Hanson said that the electronic nose technology has the advantage of speed. While waiting for conventional diagnoses, a patient's condition could worsen or the patient might receive unnecessary treatments, he said.
"The electronic nose takes six airway samples, a process that takes seconds, and then the samples are analyzed and the 'nose' is cleared for reuse," he said.
The whole process takes 40 minutes.
Dr. Michael Niederman of Mineola, New York, who chaired Tuesday's meeting, questioned the need for diagnosis by nose.
"Why are we doing this when we have a perfectly acceptable way to diagnose pneumonia?" Niederman asked.
He noted that Hanson had failed to produce data to show the nose's diagnostic calls met scientific standards for specificity and selectivity.
Hanson agreed that the nose technology is not ready for general use, but maintained it is accurate, fast and inexpensive. He said the nose costs "about $8,000, but after that the cost for using it is extremely minimal."
He said he and his colleague, Dr. Erica Thaler, already have conducted lab tests in which the nose correctly identified bacteria grown in test tubes.
Thaler said they plan to test the nose on patients with sinusitis "because these patients frequently are given antibiotics when they don't have proven bacterial infections. Using this technology, we would quickly confirm infection."