Nearly 18 months behind Europe, the US FDA has approved a software system that automatically
searches for suspicious areas on CT scans, and apparently does better than
radiologists. Though currently intended as a tool to aid the radiologist, HFD
predicts a future version of the system will render the radiologist redundant.

Pattern-recognition algorithms
developed for the �Star Wars� missile defense system are being applied to the
early and reliable detection of breast cancers.

A US study has found that the genetic mutations responsible for the commonest
form of Alzheimer’s appear to be housed in mitochondrial DNA, rather than in the
chromosomes.

Reference: Unknown (2004). “Brain Cells:
Alzheimer’s Clues.” Reuters via Wired News, July 5.

FDA Approves New Cancer Imager

The US Food and Drug Administration has approved a software system,
ImageChecker, that automatically searches for suspicious areas in CT
scans, acting as �a second pair of eyes for the radiologist,� who �first
conducts a standard review,” according to the company that makes it. In a
clinical study, 15 radiologists independently reviewed 90 images from lung CT
scans with and without the system. They identified more nodules with the system
than without it. The system was approved in Europe in March 2003.

Reference: Unknown (2004). “FDA
Allows New Computer Detection of Lung Cancer.” Reuters via Yahoo News, July
12.

Star Wars Medicine

A group of military doctors and engineers is adapting technology developed
for the discredited but still active $100 billion US “Star Wars� program to
search for enemy warheads in space, to search instead for breast tumors �years�
ahead of current methods, writes Bradley Graham in the Washington Post.
Encouraging early results have yet to be subjected to peer review and �years� of
further testing remain. Other non-military spin-offs of Star Wars research
includes cameras that aid eye exams, facial imaging sensors to identify people,
and chemical detectors for environmental contaminants.

It turns out that the pattern recognition algorithms developed to pick enemy
warheads out of fields of decoys and background clutter can be adapted to
identify patterns in digital mammograms that may signal the presence of breast
cancer. If successful, the technique will reduce the large number of false
positives (resulting in unnecessary surgeries) produced by today�s mammography,
as well as be more accurate and reliable than a radiologist.

The team first had to work out how to digitize mammograms. It has so far
digitized nearly 100, to which the algorithms are now being applied.

Reference: Graham, Bradley (2004). “Using
Defense Know-How for Health Care: Technology to Hunt Warheads May One Day Find
Tumors.” Washington Post, July 2.