The following content has been provided by the University of Erlangen-Nürnberg.

So I guess I don't need to go through the concept of x-ray CT here, right?

That's fine.

Then we'll skip this beautiful slide.

I just want to emphasize modern x-ray CT systems are going from single slice to multi-slice

and of course to flat panel large area detectors.

And there are a couple of different applications where they're using large area flat panel

detectors.

I wanted to emphasize the second one.

Probably most of you have heard about CT in the interventional sweep flat panel attached

to a CRM.

One of the other places where it's being used is in radiation therapy systems for verification

of patient positioning and of treatment plan, except that it can't yet be used for verifying

the treatment plan because the CT numbers that you get from a large area detector aren't

accurate enough.

So one of the driving forces right now is to get as accurate CT numbers as you can get

with a clinical CT scanner but using a large area flat panel detector.

And there are problems to doing that.

So conceptually though, conventional CT and cone beam CT or large area flat panel CT are

conceptually the same.

So in terms of the data processing pipeline, again, most of the same data acquisition,

data correction, reconstruction, post reconstruction, correction and rendering.

And I'm going to touch on two topics today.

One of them is data correction, looking for and correcting for errors in the projection

data itself.

And then I'm going to show you some early work on a new reconstruction technique that

we're working on.

So the CRM system goes through a pie plus fan angle acquisition compared to a full scan

but most importantly has a large area detector compared to the CT detector.

What that means is that we get a lot of X-ray scatter that you don't normally see in a clinical

single slice CT scanner.

It depends on the amount of X-ray scatter, it depends on geometric factors like the air

graph, the focal spot to detector distance and depends a little bit but not very strongly

on the energy spectrum.

And what we're seeing right now is that cone angles on clinical CT scanners are getting

larger.

For example, 64 slice scanners have a cone angle of 4 degrees, 256 up to 13 degrees.

But if you're using a larger area flat panel then you can get 12 degrees for a 3 by 40

flat panel or 15 degrees for a 40 by 40 flat panel which is the GE implementation.

So basically scattered primary ratios are on the rise.

What does that mean for image quality?

Well it turns out that as a rough rule of thumb the scattered primary ratio is linearly

dependent certainly for lower smaller volumes on the total irradiated volume and I really

like this representation.

This is in the absence of collimation which is not the way you normally acquire data on

a clinical CT scanner at least.

So where are we now?

64 slice scanner is down here for the total irradiated volume.

This is in a 16 centimeter diameter head object.

So again if it depends on the volume and if you're imaging a head or a body you're going