Data cross check programs.....are they possible and would they work?

I have an idea this should be submitted to the philosophy and metaphysics forum science section.....but here goes...
First off im no technophile...being thrity years a logger,.....but heres my idea....We have all heard of the NSA meta data collection....how they process this data may be of benefit to the world a whole.....
The algorythms are transferrable to any set parameters for examination....the method iexists...
We have a great deal of stored data.......the volume is tremendous....
Imagine getting an MRI....now the machine, takes pictures of micro slice- like cross sections of your body.....
The top to bottom of your physiology is recorded in detail.....
If it is a spinal problem, the relevant sections get examined....and treatment or diagnosis is made.....this is specified in the MRI order itself usually....
My question here is what happens to all the other data that's recorded during the MRI?

My idea is to write a computer program, which would examine the data in its entirety, and offer diagnosis of anything out of the normal it finds.....besides the severely limited specifications of the MRI order .......(for example...but could apply to other things as well)

Nothing new perhaps (my info stream is pretty narrow...)but certainly a wide open field for early diagnostics and cheaper medical costs....as the data is currently unused except for interpreting on a narrow bandwidth...

Such a code once written, could act as a sort of "check some".? for fringe of interest data which could reveal serious problems for the patient that were heretofore unknown or suspected....

The MRI (or other data) to be examined would be simply fed into a computer and the programming would scan for abnormalities not specified in the patients treatment chart....AUTOMATICALLY....
No high priced humsn expertise required unless something is detected....

The process is entirely convertible to other advances in tech as well as other fields....as a standard operating procedure....sort of AI second guessing program that examines the accrued data OUTSIDE the parameters of implied needs of the moment ie...
The program may detect a tumour while also examining your spinal trauma....?? just a thought...could this be done by A. Intelligence as a sort of beneficial by product of your back trauma treatment?

Perhaps my brains foggy this morning (only two cups of brew yet....) but is there room for discussion of this type of endeavour alongside the regular medical practices?.Or other forms of research even?
It could be enormously cost efficient....as well as live saving if standardised with certain technology....
if im way behind the times here then cut my legs off and call me Shorty....

a reply to: stirling
You should probably ask the mods to move this to the medical forum since it's about MRI, a medical technology:

www.abovetopsecret.com...

I only know limited things about MRI and I could be completely wrong about this, but here's my guess.

The existing computer programs seem to create a computer-generated 3-dimensional image of the volume of the body which was scanned. Apparently different tissues/densities can be coded differently to highlight things like tumors, for example.

Then if say a tumor is found, the physician uses the 3-D imagery (or multiple thin 2D slices of the 3D imagery) to plan the most effective surgical procedure:
-Where to make the incision and how large,
-How to remove and extract the tumor, etc.

Now how much the existing software assists the physician beyond creating the 3D imagery, to identify the tumor tissue as being a tumor, I don't know, and this is why I think you'd have a better chance of getting someone who knows more about medicine to comment on this in the medical forum, if that's what you're talking about, which seems to be the case?

Nope this thread is NOT about MRIs that's just one example of what the program could or would cross examine......this is applicable to all kinds of different forms of inquiry....sort of a meta data sort that occurs when the tech is in use for any purpose.....
Its got to be programmed into the diagnostics or whatever to check sum the data for other possibilities that are not being focussed upon.....the time for humans to scan for abnormalities in the data would im[possible to calculate...but a program set up to do so could be very beneficial and cost effective,,,,even produce off shoots to the research or diagnosis in question...
The idea is that humans cannot correlate the immense data stream that a sufficiently capable computer can.....
We are close to such a breakthrough of that I am pretty sure.....

a reply to: stirling

It is possible that such a thing could be made. Once made, I'd hope it could be used for any type of medical scan, as the basic premise would be the same. Look for something abnormal.

Here's the trick: weeding out false abnormalities. It's tricky with simple things like text recognition, since there are so many different fonts, scripts, languages, etc... but just imagine if all of the text was squishy, and surround by other squishy things, and could potentially be relocated just by leaning the wrong way. I'm sure I'm not the only parent who has seen an ultrasound and just taken the tech at their word that there's a kid in there somewhere.

So, difficulty of identifying structures and abnormalities would make this a really complex system. Misidentification can lead to a couple of issues: "We're showing 3 dozen points for review, so it will take us a week to get you a clean bill of health at a huge cost." Or "Everything's great!" (When it isn't. )

Honestly, I'd be a bit surprised if there isn't something like this started out there... but we'd be wise not to trust it until it's been workdd on for a long time.

a reply to: stirling

So something like the NSA's metadata analysis program?

New system will run the Hadoop file system on commodity servers and include search, discovery, and analysis tools.

This paper has more detail:

Cost/Benefit Analysis of NSA's 215 Metadata Collection Program

Yes definitely that's mentioned above....the NSA can do it for their purposes, why not Medicine or other instances where programs could be written to re examine the collected data which may have been taken for another reason?
We could put that big billion dollar boondoggle to work out in Utah where they have the computing power and probably the data already stored....ha ha...