Chinese physicians have long used the sense of smell to identify specific diseases and infections. In the Western world, physicians have relied on laboratory testing. Over the last two decades, scientists and medical professionals have been working to combine technology and the sense of smell. The result of their efforts is the artificial nose.

In 1984, a team of scientists led by Dr. Joseph Stetter, who has a PhD in physical chemistry, produced the first working artificial nose at Argonne National Laboratory in Argonne, , a suburb of Chicago. Called the E-Nose, the device was totally self-contained and consisted of a sensor array, sampling system, and computerized pattern-recognition "smart" software. Battery operated and weighing 15 pounds, the device could identify up to 30 different compounds or chemical smells. The E-Nose was made possible by advances in computer technology and progress in chemical analysis.

Since its inception, the artificial nose has been used in a wide range of applications, and is slowly replacing the professional human nose and its canine companion in several areas, such as sniffing out explosives, narcotics, and foods, and detecting spoilage in products such as coffee and cosmetics. Today's artificial noses can reliably recognize and identify thousands of smells. More importantly, they can do so quickly and cost effectively. However, only recently has the medical community been open to using the artificial nose as a diagnostic tool.

Leading the way is Dr. C. William Hanson, an anesthesiologist and intensivist who runs the surgical ICU and the critical care section of the department of anesthesia of the University of Pennsylvania (UP) (part of the University of Pennsylvania Health System, in Philadelphia. In 1997, Hanson collected breath samples in plastic bags from the ventilators of 19 intubated intensive care patients, nine of whom were already diagnosed with pneumonia. He fed the samples into an aroma-analysis device. The device correctly analyzed the exhaled gas, distinguishing the non-infected from infected patients.

Hanson notes the device has many advantages over traditional diagnostic methods. "Traditional methods require the acquisition of blood samples for white count, sputum samples for bacterial culture, and clinical evidence to make the diagnosis of pneumonia. We hope that the E-Nose will be as accurate as traditional diagnostic measures for pneumonia, and permit earlier diagnosis." And, he says this is only the beginning: "One can imagine the electronic nose being used in the diagnosis of any number of lung and systemic processes." He is now working with Dr. Erica Thaler (an otolaryngologist who specializes in rhinology) on applying the artificial nose to the diagnosis of sinusitis.

Meanwhile, at the Cranfield Postgraduate Medical School, in Cranfield, Bedfordshire, , Dr. Selly Saini, an optical physicist, and Jan Leiferkus, a PhD student in medical diagnostics are working to refine the Diag-Nose, a tool they developed in 1999 to identify the presence of pathogens in urine. Saini says that although the use of smell has been largely forgotten with the shift to laboratory analysis, creating machines to extend the senses makes perfect sense. With the development of better sniffing equipment, this ancient method has modern-day applications. Saini saw urinary tract infections (UTIs) as the perfect testing ground for the Diag-Nose. "They're prevalent in the general population, being second only in incidence to the common cold for conditions that are presented to the physician."

With the device, physicians were able to cut down the diagnosis time for UTIs from two days to a few hours, and significantly lessen the patients' discomfort and the spread of the infection. Furthermore, the Diag-Nose was able to differentiate between various microorganisms and allow doctors to tailor treatment to the specialized needs of each patient. Saini sees a great future for the Diag-Nose. "Early work in 1999 showed that (Diag-Nose) analysis of urine samples could detect the two most common causes of UTIs, E. coli and proteus mirabilis, in about four to six hours with 80% accuracy. Since then, we have refined our methods and are able to achieve detection in less than an hour, with 98% accuracy." Saini continues to devise new forms of diagnostic technology, including a diabetic glucose-testing strip already on the market. He is currently focusing on clinical trials of odor-sniffing strips based on the Diag-Nose.

In 1998, Dr. Peter Lykos, a chemistry professor at The Illinois Institute of Technology (IIT), Chicago, and a group of students from five different disciplines (chemistry, physics, electrical engineering, environmental engineering, and chemical engineering) worked on a research project to make an electronic nose from scratch. They designed the sensor housing, electronics, computer interface, and odor pumps, and used it to test samples of coffee for spoilage and to test samples of other substances for industrial contaminants, such as benzene.

Then, in 1999, Christopher Morong, a chemistry/physics student, and Asha Joseph, an electrical/computer engineering student, along with chemistry faculty at ITT and lab technicians at the Provident Hospital of Cook County, Chicago, worked together to refine Lykos's E-Nose to test blood samples. Morong explains that the idea of working with blood sprang from a discovery microbiologist Dr. Pravin Patel made while testing for pathogens at Provident Hospital. "Patel and his colleagues noticed that many times they could identify the bacteria just by the smell of the sample before the more rigorous tests were performed."

Lykos is enthusiastic about the results. "The traditional methods of diagnosis involve first allowing the bacteria to grow, and then be identified, incurring a delay of 48 hours," he explains. "With the E-Nose, we can come up with a diagnosis in just 24 hours. Because the nose can focus on a particular pathogen, we can reduce the number of possibilities significantly, so the pathologist can proceed in a more educated way in determining what would be the most appropriate antibiotic for that particular patient and that particular pathogen."

Efforts to optimize the diagnostic applications for the E-Nose are in the works; these include developing the ability to test accurately for tuberculosis. Meanwhile, Stetter is moving ahead, developing a machine that emulates another human sense: taste. His new ChemArray chip, which he describes as an electronic nose or tongue, shows promise for detecting individual pathogens and toxins in biological fluids such as anthrax antibodies in blood by "sniffing" or "tasting" them.