Evidently, a machine has been developed that found 73 percent of the patients with lung cancer, and it correctly cleared patients of lung cancer 72 percent of the time.
My question comes from the following quote, also in the article: "So far, the most accurate detector of cancer has been dogs. In 2006, researchers found dogs could be trained to smell cancer on the breath of patients with 99 percent accuracy."
Then what the blazes are they investing that kind of time and effort into a very expensive machine?
Dogs are cheap!
The Stupid Science: Just because one can do something with fancy machinery doesn't mean they should.
If you can accomplish something at 1/10th the price using a dog, then use the dog!
I see this all the time in computers, whereby organizations computerize processes which were clean and efficient when conducted by paper and wind up costing them 10 times as much to do via computer without a single added benefit.
Neither computers nor medical testing equipment holds any inherent benefit over other approaches. The only justification for using higher technology is if the long-term cost/benefit ratio outweighs other, often known methods.
In the late 1800s, doctors would smell the hands of their patients to determine if they were chronic alcoholics. There's a musky oder characteristic to chronic alcoholics which isn't present in those who don't drink heavily or to excess.
Many other signs, now long lost to "modern" doctors who over-rely on labwork, were quick and inexpensive tell-tales concerning disease, signs from fingernail growth and development to the whites of one's eyes to skin tone and color to on-hands examinations. Even the smell of one's breath can tell a properly trained doctor loads of information, from diabetes (fruity-smelling ketones on one's breath) to several other common diseases.
And if the odor is too weak to be noticed by a doctor, then by all means, use a dog, not a $5000 machine test with 27 times less accuracy!