Follow Slashdot blog updates by subscribing to our blog RSS feed

 


Forgot your password?
Close
typodupeerror
×
Medicine Biotech

A Breathalyzer For Cancer 123

Posted by timothy from the your-breath-smells-like-carcinoma dept.
Tiger4 writes "Cancer researchers in the UK have come up with a way to sniff for lung cancer on the breath. 'From the results, the researchers identified 42 "volatile organic compounds" (VOCs) present in the breath of 83% of cancer patients but fewer than 83% of healthy volunteers. Four of the most reliable were used to develop a nine-sensor array made from tiny gold particles coated with reactive chemicals sensitive to the compounds.' Other sources have picked up the story as well. Obviously, this would be a big breakthrough for rapid screening, and early detection significantly improves outcomes."
This discussion has been archived. No new comments can be posted.

A Breathalyzer For Cancer More

A Breathalyzer For Cancer

Comments Filter:

  • Oscar (cat) (Score:3, Interesting)

    by juancnuno ( 946732 ) * <juancnuno@gmail.com> on Monday August 31, 2009 @01:46AM (#29257573)

    There's a cat that seems to be able to tell when someone is about to die. [wikipedia.org]

  • Sensitivity and specificity? (Score:5, Interesting)

    by yali ( 209015 ) on Monday August 31, 2009 @01:48AM (#29257591)

    83% of cancer patients but fewer than 83% of healthy volunteers

    Let me introduce you to my friend Reverend Bayes [hmc.edu].

    • Re:Sensitivity and specificity? (Score:4, Insightful)

      by graft ( 556969 ) on Monday August 31, 2009 @02:18AM (#29257717) Homepage
      Yeah... that was my first reaction. With a false-positive rate that high, this is useless as a diagnostic tool.

      • Re:Sensitivity and specificity? (Score:4, Insightful)

        by DeadDecoy ( 877617 ) on Monday August 31, 2009 @02:28AM (#29257769)
        It depends on how quick and cheap this test is. If it's really cheap, then it would be useful in validating more accurate (and inherently more expensive) tests that would be used for initial detection or risk assessment. As a hospital manager would you rather run 83 cheap tests and 17 expensive ones to confirm or 100 expensive tests; again it depends on the cost. On the other hand, I'd be more worried about the false negatives. Not being diagnosed with cancer, when you have it, and getting early treatment is much worse than getting an extra test, at least from the patient's perspective.

        • Re: (Score:1)

          by Anonymous Coward

          The way I understand it, a negative test doesn't mean you are free of cancer. A positive test doesn't mean you have cancer. With such low accuracy the test is useless.

          What do you do with the result? You cannot skip further tests if it says you don't have cancer and you cannot skip them if it says you do. So where's the point?

          Maybe some bullshit MBA will think it will save 83% of the expensive tests. Let's just let all those false negatives die because it's cheaper.

        • Re:Sensitivity and specificity? (Score:5, Insightful)

          by shentino ( 1139071 ) <shentino@gmail.com> on Monday August 31, 2009 @06:34AM (#29258695)

          I would run whichever test had the highest profit margin.

          • Re: (Score:1)

            by jd2112 ( 1535857 )

            I would run whichever test had the highest profit margin.

            Except that the patient's insurance policy will only pay for the cheap one.

        • Re:Sensitivity and specificity? (Score:5, Informative)

          by Bakkster ( 1529253 ) <Bakkster.man@NOspam.gmail.com> on Monday August 31, 2009 @10:05AM (#29259999)

          That's not how it works. Assuming 83% is the accuracy for both positives and negatives, and given a group of 100 people, 10 of whom have lung cancer, we would expect on average:
          2 patients with cancer will not be detected.
          17 cancer-free patients will be told they may have cancer.

          If our patients have no increased risk for cancer, then our group of 17 false-positives are suddenly scared into receiving further (possibly expensive) further testing. Our two patients whose cancer was not detected ignore futher testing until it is too late.

          Basically, this technique is mostly useless for general screening. Of course, widespread screening generally is a bad idea, due to false positives, so the best bet is always to stick with those with higher risks only. I could see this method gaining use in developing nations where an expensive test is not possible, as the false-negative rate is still smaller than the current rate of undiagnosed cancer. However, dealing with the false-positives is still a bitch.

          • However, dealing with the false-positives is still a bitch.

            That really depends on why you're running the test in the first place. Yes, general screening would probably be a bad idea as the accuracy is too low, but, if the patient was going to have a cancer test regardless of the situation, the cheap test could serve as a decision aid: you tested positive on the cancer-breath test which warrants more testing or you tested negative so you don't have to worry about it. Without the breath-test, these 17 false positives may have taken the normal cancer test anyways beca

            • That really depends on why you're running the test in the first place. Yes, general screening would probably be a bad idea as the accuracy is too low, but, if the patient was going to have a cancer test regardless of the situation, the cheap test could serve as a decision aid: you tested positive on the cancer-breath test which warrants more testing or you tested negative so you don't have to worry about it. Without the breath-test, these 17 false positives may have taken the normal cancer test anyways because they come from the subset of people who get cancer screening and not the general population.

              There are 4 results, only one of which is beneficial, if you were already planning to take the test.
              1) The test correctly diagnoses cancer free, at a fraction of the cost (good)
              2) The test correctly diagnoses cancer, in which case it's just an additional test
              3) False-positive, leading to stress until the actual test can be taken (bad)
              4) False-negative, leading to a loss of early detection, and possibly death (also bad)

              Basically, its use is limited significantly due to those false results, assuming it's

        • No, you don't understand. This test will end up costing people a lot MORE.

          You have signs of cancer. You run this test as part of a routine screening. If you already thought you might have cancer and the tests says negative, you still run the more expensive test because 17% of the time it is wrong.

          You have no signs of cancer. You run this test and it says you are positive. SO you run the real test.

          Net net, this will always always cost society more, a lot more. The false positives and false negative

      • Re: (Score:3, Insightful)

        by Romancer ( 19668 )

        I think that the article is trying to say that "fewer than 83 percent" means that the remaining percentage is the "fewer", AKA the 17 percent.
        I'm hopeful that it's just a case of poor wording, as it wouldn't be an effective test or really news worthy other than a study if it got 83% false positives.
        Making the actual testing device seems to give credit to the idea that it is a mis-statement of the results.

      • I dunno, breathalyzers with a hit rate lower than that get you arrested for DUI...

    • Well... I didn't understand any of the key parts of that to be honest. I was wondering though about level of detection. How much cancer do these cancer patients have? If they are, for example, on death's door of lung cancer, they might be exhaling large amounts of a chemical produced by the tumors. For this to be of real use in preventative medicine, you'd want to detect the cancer before there were obvious signs, before the cancer spreads, when it's just a few cells.

      It will be amazing if this thing can

    • 83% of cancer patients but fewer than 83% of healthy volunteers

      Let me introduce you to my friend Reverend Bayes [hmc.edu].

      Worst. Gun. Introduction. Evar.

  • Driving with cancer is a crime? (Score:3, Funny)

    by kenh ( 9056 ) on Monday August 31, 2009 @01:50AM (#29257595) Homepage Journal
    I didn't know that it was illegal to drive while suffering with cancer...

  • False positives? (Score:2, Insightful)

    by feedayeen ( 1322473 )
    "present in the breath of 83% of cancer patients but fewer than 83% of healthy volunteers" So, 83% of people with cancer have this chemical, and 82% without cancer can also have it. That test will still leave 17% false negatives and then it would give false positives to nearly everyone else.
    • I don't think "fewer than 83%" indicates that it is something around 82% but rather that they didn't have an actual number to give without doing more research. It could be 3%, it could be 50%. It's probably not 82%.
      • I think the point is that there *must* have been some difference, or there'd be no point in reporting the statistic. Water is found in 100% of cancer patient's breath samples, for instance, but that doesn't really help us. My guess is that the original article includes a typo - probably they just copied in the wrong statistic.

        Neither of the other articles includes a statistic, so I'd just assume there's a (more-than-just-statistically-)significant difference for now.

      • Re: (Score:2)

        by Trepidity ( 597 )

        True, but it would have to be a lot less than 83% to be useful. Even if only 50% of people without cancer have it, add in the fact that the vast majority of people do not have cancer, and you have a test where almost everyone is a false positive. You'd really need closer to the 3% to have a test that can actually discriminate well.

        • True, but it would have to be a lot less than 83% to be useful.

          Either that or the false negatives will have to be a lot less than 17%. I wouldn't care about false positives so much if false negatives were close to 0 - at least it would eliminate the need to do more expensive tests on those who get a negative result. If you always have to test further regardless of the result then what's the point?

    • Re: (Score:2)

      by evanbd ( 210358 )
      Hence the part about "four of the most reliable." Trust me, medical professionals are aware of the base rate problem and the difference between type I and type II errors.

  • Hrm.. (Score:5, Funny)

    by derfy ( 172944 ) on Monday August 31, 2009 @01:57AM (#29257633) Homepage Journal

    So, next time you get pulled over:

    "Sir, were you aware of how fast your cancer was progressing?"

  • Early detection doesn't always improve outcomes (Score:5, Insightful)

    by RonBurk ( 543988 ) on Monday August 31, 2009 @02:06AM (#29257673) Homepage Journal

    Non-oncs generally don't understand that a whole lot of cancer is "clinically irrelevant". That is, it would never go on to kill you. Thus, as early detection gets better in most areas, you detect a greater percentage of cancer that was never going to hurt the patient. However, once you see the cancer, you are duty-bound to slash/burn/poison (Susan Love's famous chapters) to cure it. Statistically speaking, you know you are actually harming some patients, but it is a dilemma -- you hurt all the patients in order to serve a greater good for some percentage of them. A good example is the growing backlash against general PSA screening. Even just a biopsy for prostate cancer can't be 100% risk-free, but the treatment is really risky, assuming you're not enthusiastic about being impotent and/or incontinent for the rest of your life.

    So don't get too excited about increased early detection of cancer. Currently, it is usually a double-edged sword that brings suffering to some percentage of patients who would have avoided it before the new test existed. An exciting development would be a detection test for distinguishing cancer that's just sitting there from cancer that's on the move and likely to kill.

    • Re:Early detection doesn't always improve outcomes (Score:5, Interesting)

      by RotateLeftByte ( 797477 ) on Monday August 31, 2009 @02:28AM (#29257767)

      I wish my Cancer had been detected earlier (Hairy Cell Leukaemia). Certainly before my immune system had been virtually trashed. Then the Chemo killed it completely but that is the nature of Chemo.

      There are many cancers that creep up on you slowly and almost unrecognised until they hit a critical mass. Any early detection of this type of cancer would be most welcomed by the people who have the misfortune to suffer from them.

      Just my $0.02 worth.

      • There are many cancers that creep up on you slowly and almost unrecognised until they hit a critical mass

        But the immune system may get rid of a lot of those asymptomatic cancers. In your case, early detection might have helped, but in the case of many other people, the treatment may just get rid of a cancer that the immune system would have gotten rid of anyway.

    • Re: (Score:3, Interesting)

      by F34nor ( 321515 )

      Great point, but like your point, you point totally misses the point. Detecting cancer in Phase 1 means you will almost certainly live. Detecting in Phase 3 means you're dead. False positives in Phase 1 are not a big worry as you point points out.Grow a pair of nuts and raise, live without fear, and ed-u-ma-cate yo sef about reality.

      ***Like most overly critical (and totally correct) slashdot.org points this one illustrates that "perfect is the enemy of the good".

      • It depends on how you value various things. Detecting a lot of people who would not have died anyway in Phase 1 will raise your overall survival rates (those people will still live), but may make a whole lot of people's lives shittier. Is removing an X% change you'd have died worth a Y% chance that you unnecessarily made your life shitty for years? Depends on X, Y, and your preferences.

      • Many of the treatments for cancer cause cancer, so while you may survive that stage one detection (which you might have survived anyhow), the treatment can kill ya.

        So yeah, false positives in any stage are a big deal, unless you think that cojones means that you volunteer to get cancer.

    • Re: (Score:3, Informative)

      by techno-vampire ( 666512 )
      Even just a biopsy for prostate cancer can't be 100% risk-free,

      ...and is considerably more unpleasant than they tell you it will be. They also tend not to tell you how messy it can get while you're healing afterwords. When I was finished, I told the doctor that IMO, before he performed another prostate biopsy, he should undergo it himself to find out just how bad the "slight sting" was, and he told me that he'd refuse the procedure.

      Incidentally, one thing I learned from the experience is that although

    • A good example is the growing backlash against general PSA screening.

      I hate PSA screenings. The "This is your brain on drugs." one was just lame.

    • Re: (Score:3, Informative)

      by vandelais ( 164490 )

      How did the parent get modded as insightful? Let's try this again. Comparing almost always fatal lung cancer (from TFA) affecting a variety of age groups to slow-moving often 'clinically irrelevant' prostate cancer (not from TFA) whose onset affects primarily the elderly is both Troll and Offtopic.
      -
      Primary lung cancers themselves most commonly metastasize to the adrenal glands, liver, brain, and bone. Secondary lung cancers can be indicative of other cancers whose prognosis is also dependent upon the ear

  • Dogs can do this for $2000 in training, yellow jackets for even less. It is hubris to think techmology is better than a billion years of dice and death.

    • Let me re-phrase that...

      "This is a story about a way to patent and monetize something that can be done on the cheap already."

      • Re: (Score:2)

        by TheLink ( 130905 )
        Dogs are likely to get bored and want to do other stuff after a while. Also the results might not be as consistent from one dog to another.

        The device in the article sounds really crap though - consistently crap is not very useful.

        I'd be more interested when it is far more accurate. Right now it's not news at all.

    • Re: (Score:1)

      by Anonymous Coward
      Oh, okay. Everybody, we can all go home. Apparently there is no use in finding out what compounds are indicative of lung cancer, because dogs can already detect it for us. Also, we should halt research into blindness, because we have guide dogs to help out with that too. While I'm thinking about it, let's scrap all of our research into generating energy, the Sun has been doing it better for billions of years.

      What is wrong with these statements? The truth is more useful to us when we know what it is. Way t

      • Actually, ignoring the Sun as a source of energy really is stupid.

        Big oil is laughing all the way to the bank every time we burn gas instead of burn sun.

        The other stuff you mentioned, maybe.

        But you need to see the light.

    • yeah, and horses used to transport people everywhere. If everyone in those olden days thought "lol, a car. they're way to expensive to buy" we wouldnt have decent cars.

      • Re: (Score:2)

        by F34nor ( 321515 )

        If horses were cheaper many people would still use them. Horses are now freakishly expensive.

        Back to MY point, this is an incredibly expensive and difficult way to detect cancer compared with training a dog. Why would someone do this? 1. People are dumb. 2. People who are smart try and find ways to take dumb peoples money. 3. People are scared of cancer and think that technology will save them.

        REALLY think about what a $2000 dollar cancer detector that loves kids and is powered by dog food and tennis balls

        • Sounds like you have the beginnings of a business plan. While you may be a bit too optimistic on cost savings, why not give it a try? Compete with this techno-gadget in the open marketplace and win! If you don't do it you'll be kicking yourself when someone else does and makes a billion.

  • Bad test. (Score:3, Insightful)

    by rew ( 6140 ) <r.e.wolff@BitWizard.nl> on Monday August 31, 2009 @03:26AM (#29258041) Homepage

    Just recentely an article was published about the effects of a "non-perfect" screening test on a large group of people. (They did a case study for a specific test, which gives the same results as the theoretical example below)

    If you have a test that is 99% accurate and 1 in 1000 people have the illness, for every million people, you'll find 990 out of a thousand who DO have the disease. And you send home 10 people who do have the disease. Too bad, 990 got the proper treatment. Right? No! Of the 999000 who don't have the disease, you'll find 9990 people with positive results. Are they getting treated? Are there risks to further tests? (for example X-rays create a risk for cancer later on.... )

    And finding the 990 people WITH the disease from the 10980 people who tested positive is going to be a hassle. And costly.

    Now a test with a 99% accuracy is pretty good compared to this test. As I read the slashdot intro, there is a 17% false negative rate, and up to 82% false positve rate....

    • Re:Bad test. (Score:4, Insightful)

      by speedtux ( 1307149 ) on Monday August 31, 2009 @03:48AM (#29258131)

      "Just recently?" That's been a standard example in introductory statistics classes for many decades.

      • However, very recently the BBC and the like have been publishing these sorts of things for the general audience, with clever visualsiations. It's often hard to grasp how exactly a 1% false positive rate can completely undo the benefits of a 1% false negative rate.

      • "Just recently?" That's been a standard example in introductory statistics classes for many decades.

        I think the he meant in geologic time.

    • Re: (Score:3, Insightful)

      by Aladrin ( 926209 )

      There's no such thing as a 'bad test' that has a high rate of accuracy, there are only people who are using the test wrong.

      With 86% accuracy, you can't rely on this test... But if it comes up positive, then you know you had better check them with a better, more expensive, less fallible test.

      If this test were cheap enough to use it on every patient that came in for anything, it would be a -huge- boon to society. If it cost 10x what a checkup costs, it'd be utterly useless.

      • Re: (Score:3, Insightful)

        by rew ( 6140 )

        In my example, in the population 99.9% of the people don't have the disease. In those that tested positive, 90% still don't have the disease. That's with a 99% accurate test.

        It's very difficult to effectively use a test that has an error rate of on the order of 17%. Or to be more accurate: A false negative rate of about 17%, and a false positive rate of "less than 83""....

        But yes, you're right. It only becomes bad when people use the test wrong.

        The research I don't remember the reference to however showed t

  • Soshulizm.
  • Pardon the sarcasm. But I could use exactly the same methodology to try to determine the difference between drunks and non-drunks, and come up with a result of Spearmint.

  • Very poor reporting (Score:3, Insightful)

    by wvmarle ( 1070040 ) on Monday August 31, 2009 @04:44AM (#29258337)

    Very poor reporting here, and not juts from slashdot alone this time.

    Two of the links talk about research done in the UK researchers, the third link about that it is done in Israel. One article mentions 62 volunteers, another 94. While it seems they are talking about the same research project - none is referring to the other though.

    And of course that 83% number. That means 17% false negatives for cancer patients? How is the false positive rate then for non-patients? In other words: how useful is this test in real life? If healthy people still get a say 70% positive rate it's quite useless I'd say.

    Interesting and all that a breath analyses could indicate the presence of cancer but the articles as linked do not make me believe it really exists in a working sense or that it is actually useful at this time. The articles linked at least are too light on detail, and even state facts differently. Unless there are two research groups working on the same thing together (one in UK, the other in Israel), but then why don't the articles refer to this fact?

  • Less than 83%? (Score:3, Insightful)

    by Arancaytar ( 966377 ) <arancaytar.ilyaran@gmail.com> on Monday August 31, 2009 @05:00AM (#29258401) Homepage

    Hooray, they are only 17% likely to get a false negative on a sick person, and something less than 83% (82%? 50%?) likely to get a false positive on someone healthy.

    Combining that with the statistical problem of detecting a rare disease, that's not really useful as the number of true positives vanishes against that of false positives. (Even assuming their results aren't down to a correlation like smoking.)

    • Re: (Score:2)

      by Brianwa ( 692565 )
      The 83% figure was just used to roughly sort through the VOCs that they tested for in a lab. It will probably have very little impact on the accuracy of the final device.
  • Wait till your insurance agent carries a C-Breathalyzer. "Please breath, Sir, for a personal cover scheme". There goes Ol' Mutual.

  • Don't you think Cancer sufferers have enough to deal with that they don't need to be stopped at random while driving their cars and subjected to additional tests? They must be allowed to continue to drive or how will they get to their chemotherapy appointments? Is there any evidence that Cancer patients are the cause of accidents? I'm not sure I like this. :-)

  • this is just a easy way for you to have your health care taken away from you also your job as well as your employer does not want to have to cover the costs of cancer. We need a health plan that blocks that crap be for we start useing stuff like this.

  • Suppose 1 out of 10,000 people in the USA have HIV. Suppose an HIV test has a 99.9% chance of detecting HIV in infected people, and a 99.9% chance of testing negative in people without HIV.

    Then, suppose we test 10,000 people.

    Among the infected, we will get an average of .999 people who test positive. Among the non-infected, we will get .001 * 9999 = 9.99 people who test positive.

    Thus, among people who tested positive, only .999/9.99 = about 10% will actually be infected.

  • There is an easier way to detect lung cancer via sniffing someone's breath. Ask them how long and how much they smoke, sniff their breath. Cigarette breath? Cancer!

  • Comment removed based on user account deletion
  • What are all these people claiming "Early detection of cancer will help" basing this statement on?

    If everyone who had cancer detected early NEVER DIED, this might make some sort of sense. Since they WILL DIE, we have to change the way to tell if we've helped from "Didn't die" to "Didn't die in the next X years." Of course early detection helps THAT situation.

    Let's take the case of someone getting cancer at age 50 and dying of it at age 65. Without detection, they'd wouldn't know til near the end, sa
    • Early detection means that the "five year survival rate" goes up. This way glowing news stories can be printed bragging about how far we've come in conquering cancer.

      Just a statistical game.

  • Sir you may or may not have cancer; this test was inconclusive, another test is needed. Why go through the trouble, why not straight test for cancer? Come back when results are 100% accurate.

Slashdot Top Deals

I have hardly ever known a mathematician who was capable of reasoning. -- Plato

Close