[Dcmlib] dicom image orientation

[Greg Book]

I have some questions about the image orientation of dicom images. The 
part I'm having the most trouble with is the direction cosines and how 
it affects the encoding from the HFS, HFP, etc.
According to the dicom documentation, the patient orientation determines 
how the data is encoded: positive X direction points to the patient's 
outstretched left arm, positive Y points from antierior to posterior 
direction, and positive Z points from feet to head. Do the cosines 
change this encoding?
For example, if the patient is head-first supine, then axial slices are 
acquired, and if the slices are not rotated... How do the directions 
change? Is X still pointing to the patient's left arm?
Now... if the slice box were rotated around the patient's feet/head 
direction by 180º, how would that change the cosines?

My ultimate goal is to have a 3D cube on my image viewer that shows how 
the current view relates to the patient's R/L, A/P, and H/F sides.
I'm not sure if knowing the orientation description (Heart Coronal, etc) 
will be helpful. I'm interested in how the cosines can be transformed 
into a cube with the 6 patient directions on it.

-Greg Book