 | |  | |  | |  |
Robots Reading Radiograms
As of 2005, about 40 million mammograms were done each year with about 200,000 cases of breast
cancer involved, either detected or not detected. This means that one-half of one percent of the
mammograms would show breast cancer if the radiologist were to detect 100%. In the 1990s, a study
was done which showed that radiologists were not detecting 22% of the cancers which were present.
They would see 78% positive correctly but miss 22%. In addition, therer are a number of cases where
the physician might see possible but it was not present. These are called false-positives. A young
and less experienced physician might see possible cancer in 10% of the X-rays which he reads, which
is a high rate given that it is only present one-half percent of the time. To address the 22%
undetected rate, the US government mandated that at least two radiologists read each x-ray so that
fewer cases of breast cancer would go undetected. According the book "The End of Medicine", in
1999,
Corporations which manufacture computerized neural networks which can read mammograms better
than humans can:
iCAD Mammography: http://www.icadmed.com/html/mammography.asp
Next
As of 2005, about 40 million mammograms were done each year with about 200,000 cases of breast
cancer involved, either detected or not detected. This means that one-half of one percent of the
mammograms would show breast cancer if the radiologist were to detect 100%. In the 1990s, a study
was done which showed that radiologists were not detecting 22% of the cancers which were present.
They would see 78% positive correctly but miss 22%. In addition, therer are a number of cases where
the physician might see possible but it was not present. These are called false-positives. A young
and less experienced physician might see possible cancer in 10% of the X-rays which he reads, which
is a high rate given that it is only present one-half percent of the time. To address the 22%
undetected rate, the US government mandated that at least two radiologists read each x-ray so that
fewer cases of breast cancer would go undetected. According the book "The End of Medicine", in
1999,
Corporations which manufacture computerized neural networks which can read mammograms better
than humans can:
iCAD Mammography: http://www.icadmed.com/html/mammography.asp
Next
The Concebot Medical Index