# How to determine the patient orientation of a NM TOMO frame? - DICOM

This is a discussion on How to determine the patient orientation of a NM TOMO frame? - DICOM ; Hi, I am struggling very hard with the patient orientation of a NM TOMO frame. In the detector information sequence you can find the Patient Orientation attribute. According to the description of this attribute, it is valid for the first ...

# Thread: How to determine the patient orientation of a NM TOMO frame?

1. ## How to determine the patient orientation of a NM TOMO frame?

Hi,

I am struggling very hard with the patient orientation of a NM TOMO
frame.
In the detector information sequence you can find the Patient
Orientation attribute.
According to the description of this attribute, it is valid for the
first row and column pixel.
One can read from this that it is valid for the first frame.

Now, I think you can calculate the patient orientation for the other
frames by
determining the difference in rotation for the 1st frame and the frame
of interest
from the rotation information sequence and the angular view vector.

Is this correct?
Is the start angle given in the rotation information sequence of no
interest for this?

But then what if there are multiple detectors defined?
(I know this is not IHE, but not prohibited by DICOM).
Then I cannot calculate the difference in rotation,
because I do not know the start angle of the other detector.
It should not be given in the detector information sequence according
to DICOM for TOMO images.
Or should in these cases always 2 Rotation information items be
defined containing their own start angle??

Can someone provide a recipe for the correct way of determining
patient orientation for TOMO frames?

thanks,
Tony.

2. ## Re: How to determine the patient orientation of a NM TOMO frame?

On Apr 5, 4:34 pm, "theTone" wrote:
> I am struggling very hard with the patient orientation of a
> NM TOMO frame.

[...]

In principle, patient orientation should come from Patient
Orientation and aught to be independent of frame. If you want to
know the orientation of the images relative to the patient, they
can be found in Image Orientation Patient (0020,0037), which
provides direction cosines for the orientation vector. I believe
all rotation angles and camera positions should already have been
accounted for in calculating the direction cosines.

Would you be willing to share an anonymized study so I can look
at it?

Mike

3. ## Re: How to determine the patient orientation of a NM TOMO frame?

Hi Mike,

Thanks for the attempt, but I think you missed the point of where I
have trouble.
I know how to calculate orientation from basic Image Patient
Orientation attribute only.
Trouble is that Nuclear Medicine data is MultiFrame.
For NM an Image Patient Orientation attribute is given per detector.

For starters lets assume only one detector.
So, for a 360 degrees rotational scan, this orientation can only be
valid for one frame.
Question is:
- for which frame? The first? Or the frame with an angular view of 0
degrees?

Then more questions come up when multiple detectors are defined.
Lets say two detectors for example.
Then two Image Patient Orientations are given.
Now, if there is only one Rotation Information item valid for both
detectors,
there is only one start angle defined. For which detector is it valid?
Then what is the start angle of the other detector?
Or what is the 'first frame'of the other detector?

I do not have a specific DICOM scan.
I need to know the general recipe for determining orientation...

Thanks,
Tony.