Biosensor 'Sniffs' Pneumonia and Offers Fast Diagnosis

November 06, 2002

Peggy Peck

Nov. 6, 2002 (San Diego) — A novel biosensor that is about the size of portable CD player can identify ventilator-acquired pneumonia (VAP) with a degree of accuracy that matches traditional diagnostic techniques, but "is faster and noninvasive," according to lead investigator C. William Hanson, III, MD, professor of anesthesia and surgery at the University of Pittsburgh in Pennsylvania.

Dr. Hanson demonstrated the efficacy of "Cyranose" (Cyrano Sciences, Pasadena, California) in a study of mechanically ventilated patients with suspected VAP. He presented the results here at CHEST 2002, the 68th annual meeting of the American College of Chest Physicians.

The biosensor, or "electronic nose," samples exhaled gas in the expiratory limb of the ventilator or the circuit of critically ill patients, he said. A total of 415 mechanically ventilated patients were screened for the presence of VAP using a clinical pneumonia score (CPIS). Twenty patients were enrolled with CPIS scores > 50, and they were compared with 20 patients with no evidence of pneumonia and with scores between 0 and 25.

The "nose" contains an array of sensors consisting of carbon-black/polymer composites with variable sensitivity and specificity for a wide variety of volatile chemicals. When the sample passes over this polymer grid, "a fingerprint of the odor is sent to the computer, which then uses pattern recognition algorithms to identify the odor," he said. Currently the nose, which was donated for the study, is used in the food industry to verify ingredients and quality, Dr. Hanson said.

Each patient was sampled six times. "Taking each breath sample requires just seconds but the entire process takes about 40 minutes," he told Medscape.

Dr. Hanson said there was good correlation between the actual CPIS score and the CPIS score predicted by the data from the sensor array. Furthermore, the cross-validated prediction model suggests that the commercial biosensor would be reasonably successful in predicting the presence of VAP in an unknown subject. Improved predictive abilities might be achieved by optimizing the sensor array for this task.

But Michael Niederman, MD, chair of the Department of Medicine at Winthrop University Hospital in Mineola, New York, and chair of the session, was not convinced that the diagnosis by a nose is needed.

"Why are we doing this when we have a perfectly acceptable way to diagnose pneumonia?" Niederman asked. He said, too, that Dr. Hanson had failed to produce data that demonstrated adequate specificity and selectivity for the nose.

Dr. Hanson agreed that the nose technology is not ready for general use but he maintained that 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."

Dr. Hanson and coauthor Erica Thaler, MD, associate professor of otorhinolaryngology at the University of Pittsburgh, said that the nose is also capable of discriminating among plate-grown bacteria, which suggests that it has potential for use in other infections. Dr. Thaler said they plan to test the apparatus in sinusitis patients. In that case, she said, the nose would analyze odor taken by swab from the sinuses. Dr. Hanson said they also plan a study of the efficacy of the nose for identifying community-acquired pneumonia.

CHEST 2002: Abstract S137. Presented Nov. 5, 2002.

Reviewed by Gary D. Vogin, MD

Peggy Peck is a freelance writer for Medscape.

Comments

3090D553-9492-4563-8681-AD288FA52ACE

processing....