Sniffing Out Cancer

Makarand writes "Researchers at the Univeristy of Rome are developing an electronic nose that can sniff out cancer by sampling people's breath. The instrument uses sensors that respond to the presence of chemical compounds in the patient's breath. For example, lung cancer patients exhale alkanes and benzene derivatives which the electronic nose will try to detect. The sensors are quartz crystal sensors coated with a substance that binds to a range of organic chemicals. If certain molecules in the breath bind to this surface coating they change the natural vibration frequency of the crystal."

Comments

Science put to good use? Wierd.

[OwnerOfWhinyCat]

Cheap tests are what HMOs love. If this test could be run for $5 a patient, they could add it to twice-yearly checkup for people over 55 and catch tumors when they are small and more cheaply removed. I kinda shrug when they invent new million dollar procedures for helping with a disease as they won't be in widespread use for many years, but cheap accurate tests like these could be saving thousands of lives a year, in just a few years. This seems very cool.

It seems likely it comes at a cost though. The accuracy of chemical detection they are talking about would make for some damnably accurate breath and air analysis tools. I certainly hope we resolve our most recent bout of prohibition in the states before Breathalyzers that can detect days old residue in the lungs are on the hip of every officer in the state.

A good nurse can tell by the smell

[Mondorescue]

My grandmother did this for 20+ years. As head matron of a major metropolitan hospital, she learned that patients suffering from certain ailments exuded certain odors through their pores and often their breath too. My girlfriend, a nurse, told me something similar. The skin is one of the body's organs for expelling toxins, so it's no surprise that we can tell what toxins are in a person's body, for example, by sniffing them.

Re:A good nurse can tell by the smell

[Daniel Dvorkin]

Yes, quite a few serious illnesses produce distinctive odors. The problem is, of course, that this is pretty useless as a diagnostic tool (with the human sense of smell) because by the time you can smell it, the patient's doomed. Something with the sensitivity to detect the odors in time to be diagnostically useful would be pretty cool, and it sounds like the researchers are on track.

Re:A good nurse can tell by the smell

[gad_zuki!]

Actually, they've trained dogs to sniff out the distinctive cancer "stink." Heres the first link from google.

Imagine walking into a hospital to visit a sick friend and having a german pointer point at you while everyone is the waiting room gasps.

Or maybe not...

[Spamalamadingdong]

Because the United States becomes more and more dystopian every day, let's look at another possibility. At your pre-employment screening, you have a mandatory breathalyzer test to see if you've been drinking. Unbeknownst to you, your breath is also analyzed for the presence of indications of cancer. If you have any, the company decides that you might very well be too expensive and drive up their insurance costs, so you don't get hired.

Pre-employment "drug" tests have been used to screen women for pregnancy, so I have no doubt that a cancer-detecting breathalyzer will be used to screen for other expensive conditions (or at least certify them as "pre-existing" and thus not covered by the company).

Dogs ...

[Anonymous Coward]

Dogs have been able to do this for a long time. Several (under controlled, restricted, scientific circumstances) tests show that they (the dogs) can be taught this - a lot easier and cheaper than trying to build an electronic nose.

But ... "Hey - it can be done"

Probably similar to the computer wine taster

[YllabianBitPipe]

I remember a couple of years ago seeing this documentary with a computer wine taster. It would sniff the vapors coming off of a glass of wine and identify the vintage. There are professional winetasters (humans) that do the same thing for a living ... needless to say they had one of these winetasters on the show and the computer was more accurate than him. I'm glad they found a practical application of this idea, since, although the technology was cool in this wine example it seemed rather pointless ... except to piss off the human winetaster.

Why am I sceptical? Let me enumerate

[sam_handelman]

1) The New Scientist article which is the source of this story isn't nearly as upbeat about it.

Quoth:

But Richard Sullivan, head of clinical programmes for the charity Cancer Research UK, is sceptical. "Smell is very important for detecting disease and this is an interesting twist," he says. "But this study is much too small to mean anything."

Sullivan adds that even an extremely sensitive nose could only ever detect tumours on the surface of the lungs, so it could never replace the blood tests or scans needed to alert doctors to the onset of secondary tumours.

2) Biosensors and Bioelectronics is not a very disciplined journal, AFAIK (those in the field please correct me if I've been misinformed); you find a lot of good work in second tier journals, don't get me wrong, but you also find a lot of crap.

3) My dad does measurements of breath alkanes; ethane is produced by oxidized fatty acids, so it is a marker for patients with high tissue free radicals (what some people call "oxidative stress" even though there is no reason to think it is harmful, in and of itself.) They are highly variable - diabetics, for example, exhale a lot of them.

4) "e-nose"? Anyone who'd use that name has to be a sheister.

Re:I can even do this.

[Rolo Tomasi]

You mean benzpyrene. 3,4-benzpyrene looks like a base, chemically, and gets built into the DNA of the cells it comes into contact with. Basically, it's like writing random data on your hard drive. Each cigarette smoked causes about 20,000 mutations, but most of the time they are corrected by error correction mechanisms in the DNA replication process.

Re:Electronic nose could sniff out lung cancer

[Beryllium Sphere(tm)]

>The idea that dogs can distinguish between cancerous and non-cancerous tissue first surfaced in the United States in 1989, when a woman said that her border collie-doberman cross had spotted a skin tumour on her leg.

There may have been more than one such case. I read about a woman whose dog became obsessed with a mole on her back. One day when she was outside sunbathing her normally loving dog bit the area where the mole was. That got her to the doctor, who treated the bite and sent the mole to the pathologist. Melanoma.

My question is, how did the dog know that the abnormal smell was (A) important, (B) a problem?

Old story resurfacing...

[pruneau]

I did my Ph.D. in just that field: gas detection by means of arrays of sensor.

Ok, it was 9 years ago, but the subject of it changed from possibilities of sensors array processing to trying to cope with the fact that those sensors were exhibiting horrible sensibility drift over time.

Because the main problem with those sensors is that they are using a chemical compounds that binds the gas molecule. To simplify, the weight of the bound gas molecules increases the load of the crystal, thus affecting its frequency response.

The main problem of such a system is that the binding tends to have permanent effects, thus altering durably the sensor response over time, up until it becomes unusable or exhibit too different a behavior for its signal to be processed efficiently.

What usually happens is that a misinformed journalist just happens to hear about that "famous new electronic nose"...

But up to now, such noses failed to find any industrial applications, just because of the sensivity drift. I clearly remember reading some literature from that Di Natale guy 9 years ago, making the same bold claims.

If someone from the italian team reads /., I will be greatly interested by there take on the drift matter.

Re:Only detects it, doesn't cure it

[rc5-ray]

I'd have to agree that easier detection would almost certainly increase screening rates. According to the American Gastroenterology Association, only 30% of people who should be screened for colon cancer actually get the screening.

As someone with two second-degree relatives with colon cancer, this is something that's personally very important. I'd rather blow on the cancer detector than get medieval with the colonoscope.