Electronic nose’ could detect TB by sniffing patient’s breath

New device undergoing sensitivity training

No matter how keen your sense of smell, it’s a good bet you can’t detect TB on a patient’s breath. Someday in the not-too-distant future, an "electronic nose" being developed at the Illinois Institute of Technology in Chicago may be able to do just that.

The electronic nose spends its days sniffing mostly harmless varieties of bacteria in the laboratory of Joseph Stetter, PhD, laboratory chief at the chemical, biological, and physical sciences department at the university. The nose isn’t much to look at, adds Stetter’s colleague, senior research associate Bill Penrose, PhD. In fact, it’s "kind of ugly," Penrose says, explaining that it consists basically of a bench-size collection of parts that includes a computer hooked to a box filled with electronic sensors.

Looks aren’t everything, though. Working with TB isolates from the state TB lab in Chicago, Penrose has put the new device to the test. Of 24 lab samples, it has identified the 18 that were positive for M. tuberculosis and the six that were negative.

When the experiment was repeated a second time, the nose got all the answers right again.

A sensitive nose knows more

Like many a parent who envisions a big future for his brainchild, Penrose still isn’t content. He wants the device to be more sensitive. "The key is not being able to distinguish one kind of bacteria from another, because we already know it can do that. What we want is for it to be able to distinguish bacteria at the very low levels you’d find in a medical situation," he says. "A breath test would be the brass ring. We’ll really have accomplished something if you could just breathe into a machine, and we could tell you if you have TB or not." Even if the nose never meets that expectation, it could still dramatically shorten time to diagnosis by conventional methods, Penrose adds.

Just as with human noses, the sensors that Stetter and Penrose’s electronic nose uses produce a characteristic set of responses. If graphed, the responses make up what could be called olfactory fingerprints, each one unique.

Technology is rather simple

The concept of a breath test for TB isn’t as far-fetched as it might sound. Many health care providers have whiffed the distinctive odor of ketones on the breath of a person with diabetes. Using the same technology, another group of researchers has managed to sniff out pneumonia, Penrose says. All the patients were intubated.

If Penrose succeeds in increasing his device’s sensitivity sufficiently, the implications for TB diagnostics will no doubt be wide-reaching. Plus, because the technology involved isn’t especially complicated, manufacturing such a device shouldn’t prove to be difficult or costly. "The undergraduate student across the hall just made one that works," he says. "That was basically working from spare parts. It’s not as if you need a micro-fabrication unit to make one of these things."

If a cheap, smart nose could be built, such a device would have special value in the Third World. "If we could pull that off, it would be a double brass ring," he says.

The nose works in much the same way as its flesh-and-blood counterpart. "We used to joke about it being an electronic nose,’ but one day somebody pointed out that’s exactly what it is — it works basically the same way as the human olfactory system," says Penrose.

People have about 100 types of receptors in their olfactory epithelium yet can distinguish thousands of odors. That’s because each sensor produces its own characteristic reaction to an odor. Likewise, a given chemical sensor doesn’t respond in the same way to one substance but reacts in an idiosyncratic fashion. "So if you put two, three, or four sensors together and expose them to a solvent vapor, each one will respond in its own way," he says.

Electronic noses already are being used by the food and beverage industries to sort out complex bouquets of odorous components. A few years ago, scientists in Penrose’s lab taught their device to distinguish sour or moldy batches of grain from good grain, a feat that had eluded government scientists using high-powered chemical methods.

So when will TB docs be able to order their own TB breathalyzers? "It’s reasonable to think we’ll be in pre-clinical within a year," Penrose says. The device has attracted a fair amount of outside interest, he admits, though not anything he can talk about yet.

When asked the other obvious question — what does TB smell like? — Penrose chuckles. "I don’t know," he admits. "We did some tests a few years back with E. coli and other coliform bacteria. As any microbiologist well knows, you can tell them quite easily apart."

The same probably goes for TB in high enough concentrations, since researchers working with petri plates and deep cultures have been able to discern many kinds of bacteria from one another, both at the genus and species level.