[Dcmlib] Orientation / Position

[Jean-Pierre ROUX]

At 14:43 -0400 30/09/05, Mathieu Malaterre wrote:
>Mathieu Malaterre wrote:

PatientPosition (0x0010,0x5100) tells us the way the patient was 
introduced in the imager
HFS : Head First Supine
FFS : Feet First Supine
HFP : Head First Prone
FFP : Feet First Prone

This is simple to imagine what it means.
( HFP and FFP are not very common values, since the position must be 
pretty unconfortable for the Patient -the patient is lying on his 
belly; but, if he has handcuffs there is no other way ...-
--> Theralys guys, I think your algorithms dont'y deal with that case ...)


ImageOrientationPatient (0x0020,0x0037) gives 6 cosines (2 for each plane)

Patient Orientation (as found in the optional 0x0020,0x0020, or computed by
std::string Orientation::GetOrientation ( File *f ), according to 
David Clunnie's algorithm) tells us about the direction of X and Y 
axes.

The values can be
A/P anterior/posterior
L/R left/right
H/F head/feet

It sounds like "values within a 'Patient referential".

Imagine the patient, in "HFS" position.
Sagital images are requested.
All the cosines will be -1, 0, or +1;
"Patient Orientation" (deduced) will be "A/F".
Positive X axis is oriented 'towards patient's nose
Positive Y axis  is oriented 'towards patient's feet
OK ?

Imagine now that patient has a stiffneck and his head is *turned* 30 
degrees towards the left.
Sagital images are requested.
One of the cosines will be almost 0.5

Deduced "Patient Orientation" will be "AL\F"
(main X axis orientation is towards patient's nose, and a little bit 
towards the left)
but the image is *perfectly* sagital !

Imagine the patient's stiffneck causes head to be *bended* 30 degrees 
towards the left AND *turned* left.
Sagital images are requested...

You'll probabely have 3 letters for X axis and  Y axis, and the image 
remains *perfectly* sagital !

The values are given within the 'imager referential', better than 
within the 'Patient Referential' ...


Well ...
We have to think a lot more, to be able to supply user friendly tools 
to allow people to do *easyly* something clever.

(I think I will ask Medical doctor to put me in the MRI and make 
images in all the positions/orientations/incidences, straightneck and 
stiffneck, etyc.
We'll have a tremendous data set !)

JP



 


>>>--> Sorry : actually they use PatientPosition (0x0010,0x5100)
>>>assumed to be equal to "HF" or "FF".
>>>
>>>In *all* the images I have, it "HFS " or "FFS "  (S for supine 
>>>-the guy is lying on his back- , You can also have P, for Prone).
>>>
>>>Moreover, on MR, you can have *any* value -1< <+1 for the cosines  :
>>>
>>>ImageOrientationPatient (0x0020,0x0037)= [ 
>>>-0.002378\0.999985\0.004852\0.030516\0.004923\-0.999522]
>>>
>>>an other image shows me :
>>>
>>>ImageOrientationPatient (0x0020,0x0037)= 
>>>[0.569486\0.822001\0.000000\-0.000000\0.000000\-1.000000 ]
>>>
>>>JP
>>>
>>>
>>>>and the suppose the only values for cosines inside "Image 
>>>>Orientation (Patient)" (0020,0037) are -1, 0, +1 !
>>>>
>>>>Have a look  at :
>>>>http://www.dclunie.com/medical-image-faq/html/part2.html#DICOMOrientation
>>>>
>>>>It *computes* the "Patient Orientation" from "Image Orientation 
>>>>(Patient)" ...
>>>>
>>>>gdcm uses David Clunie's code in the method :
>>>>std::string Orientation::GetOrientation ( File *f )
>>
>>
>>I have to say I am clueless. How exactly does this thing works. For 
>>instance, SIEMENS_Sonata-16-MONO2-Value_Multiplicity.dcm
>>
>>I get:
>>V 0018|5100 [CS]                       [Patient Position] [HFS ]
>>
>>
>>but gdcm returns:
>>Orientation [LHP\PF]
>>
>>What does LHP stand for ? How about PF ?

   Jean-Pierre ROUX
   CREATIS - CNRS UMR 5515, INSERM U 630
   Laboratoire de Radiologie Experimentale
   Hopital Cardiologique
   28 Avenue du Doyen LEPINE
   B.P. Lyon-Montchat
   69394 Lyon Cedex 03
 
   Tel      : (+33) 04 72 35 74 12
   Fax      : (+33) 04 72 68 49 16
   URL      : http://www.creatis.univ-lyon1.fr
   e-mail   : jpr at creatis.univ-lyon1.fr