Board index » health » software to interpret an MRI Image

2005-02-14 09:18:17 AM
Is there software that would look at an image from an mri film and report
results? I have recently has some MRI's taken of a hip I've been
experiencing pain in. I've scanned all of the MRI films into a jpg format
and I am waiting for the report from the radiologist.
I think it would be interesting to see if such an application exists and who
the results would compare with the radiologist has to say.
Thanks for any help
Mike
-

jlkjl;jkj wrote:
the maker of the software. Diagnosis has to be done by physicians.
-

There are several companies working on such software. It's called CAD
(Computer Assisted Diagnosis/Detection) or CARDS (Computer Assisted
Radiologic Detection Systems). Unfortunately, the systems are very far from
being able to make diagnoses. In certain kinds of imaging studies, such as
breast mammography or chest imaging, they are able to use pattern
recognition combined with rudimentary anatomical abilities to detect unusual
or uncommon lesions. Generally such lesions are "flagged" for review by
radioligists. There is a lot of debate over whether or not these devices are
useful, if they create more problems than they solve, and if they will ever
be used routinely.
An MRI of the hip is more complex than mammography and chest imaging because
of the anatomy, and I doubt anyone is working on that kind of software.
You may be able to find normal images on the internet, and then compare them
with your own MRI. But you will probably end up relying on the radiologist's
report.
"jlkjl;jkj" <lkjlkj@aol.com>wrote in message

med radiology

I don't know of any program. Given current state of the art, this is
probably not going to happen too soon.
There is software that assists in reading Mammographs (looks for calcs
(easy) and masses or
areas of spiculation). It's sometimes calls your attention to
something. Software also exists
to check for pulmonary nodules. These are two applications where the
anatomy is simple and you are
looking for things that are easily defined (calcs and nodules - certain
density for example).
Looking at a hip MRI is much harder (yes, I'm a radiologist).
Software for the hip is probably not a high priority. It's a relatively
low volume procedure. Mammo and chest CT
(and Pap smears in pathology) are high volume procedure.
In response to Cindy's comments; the Mammo and Pap smear software I've
mentioned is FDA approved and used as an tool by the radiologist. The
limits for other applications are not legal, they are technical.
-

richard.epstein@alum.mit.edu wrote:
other MDs, aren't they? Or, since other MDs don't have a skill to read
diagnostic films, they are not able to use the software?
-

I accept the change from "radiologist" to "MD". I admit to a bias,
being a radiologist.
But, many radiographs, mammograms, MRIs, etc are read by
non-radiologists.
That's an endless (usually heated) discussion, for another time.
Obviously, I mispoke, in that Pap smears are not read by radiologists.
The mammography software doesn't "read" the study. Usually each
Mammogram will have
2-4 "findings" from the software. The MD has to decide which ones are
worth evaluating further
and which can be dismissed. Usually this is done by comparing to older
studies and looking
for changes. I assume some future version of the software will be able
to compare to older studies.
The version I use does not and I don't know if any other products
currently do. Also, usually there
are findings worth evaluating further that the software does not find.
Will the software improve
enough to "read" Mammography by itself? Probably.
The only legal requirements I know of are that to read Mammography you
have to meet MQSA
(Mammography Quality Standards Act) standards, which require CME and a
minimum number
of studies read to qualify and to maintain qualification. However, any
MD can meet the
requirements, it's not specific to radiologists. Interestingly, there
are no specific standards that
you must meet to read other imaging studies. However, hospitals have
their own requirements and
insurance companies/health plans might have their own requirements.
-

Cindy <cindyduels@att.net>wrote:
Its because the R&D costs are too "high".
For a few pennies these films can be emailed to India etc
with diagnosis back the same day.
No self-respecting computer (let alone the software engineer!) can
compete with a bowl of rice /day..
ray
-