The possibility of using redcode/cineform for cancer treatment

[Vladimir Eugene]

I'm a radiation therapist/filmmaker. I have been seeing the imaging advancements in therapy. From x-ray cassets to digital x-rays that are read instantly. Saving the patient 5-10 minutes while someone runs the old x-ray cassets to the darkroom, thru the processor, run it back and see if there needs to be any adjustments. If the patient is in pain- this seems forever. Additionally we are able to adjust contrast, and brightness, and zoom in for better detail. Having a foot in both worlds it seems things are trickling down from the film world to the treatment world.

As a means of quality assurance every patient is imaged once a week, and adjustments are made if necessary. The imager has the capability of capturing images during treatment. Images are only recorded roughly 1 image per second in B&W. Available for review later.

Looking forward. Wouldn't it be nice if every treatment was captured real time, and we were able to view every treatment field real time on a monitor to adjust treatments on the fly. This increase in accuracy would undoubtedly show tremendous benefit. I possed the question to both the head research doctor for implementing new strategies for the hospital, and to our Varian representative(the treatment equipment we use) Both were very interested.

If this turns out to be possible everyone would win. More licesing outlets of a product already created, better treatment on the patient side.

[Rob Lohman]

I know very little (nothing?) about that world, but here are two thoughts:

1) the higher the resolution the more difficult it is to get "real-time". This depends on your definition of real-time (how many frames per second) and the processing power available.

2) I seem to remember reading somewhere that medical image recording & processing needs to guarantee a certain quality level or implementation correctness etc. to make sure no (or certain type of?) rounding errors occur and such.

[Graeme Nattress]

JPEG2000 is heavily used in medical imaging, to store x-ray, MRI etc. It can deal with very large files, and show the image at multiple resolutions.

Graeme

[Vladimir Eugene]

I know very little (nothing?) about that world, but here are two thoughts:

1) the higher the resolution the more difficult it is to get "real-time". This depends on your definition of real-time (how many frames per second) and the processing power available.

2) I seem to remember reading somewhere that medical image recording & processing needs to guarantee a certain quality level or implementation correctness etc. to make sure no (or certain type of?) rounding errors occur and such.

Yes and no, there is a huge difference between diagnostic images and reference images. The later being what we use on a weekly basis for eval. The quality is significanly lower than diagnostic images and are primarily used for lining up bone structures, aortic biforcation, ...etc. That's were I felt visually lossless would fit in. What I had in mind was the 100 to one compression image you showed a while back. visually reasonable would be acceptable if a realtime alignment purposes 15-24 frames per second could be achieved.

[Chuck T.]

JPEG2000 is heavily used in medical imaging, to store x-ray, MRI etc. It can deal with very large files, and show the image at multiple resolutions.

Graeme

"DICOM" is what most vendors(GE, Phillips, Siemens ets...) use in Healthcare - http://medical.nema.org/

[Rob Lohman]

Vladimir: that sounds good, a higher compression level will also help with decode times.

[Vladimir Eugene]

Vladimir: that sounds good, a higher compression level will also help with decode times.

great, we'll continue this discussion in a couple weeks after the team returns from Asia. I look forward to the possibilities.

Vladimir