Selection
of a stacked image set
Stacked Image Set
Definition
A
stacked image set consists of a series of images taken sequentially without
patient motion along an axis of the body.
All the images must be coplanar, but need not be the same size. A stacked image set may consist of all
coronal, sagittal, or transverse images, but not a mixture. It is necessary that each image be specified
in relationship to some common coordinate system. The Dicom standard requires this for all images in the same
series. Such images are here considered
stackable to form a three dimensional model of the patient. Normally you would only pull the images from
a single series. However, the
capability is provided to pull images together from anywhere, as might be the
case if a body technique changed to a head technique when the head is
encountered. However, the head images
must be specified to the same coordinate system as the body images in this
example and so should be part of the same series. The program has no way of knowing the relationship between
different series, and the user must not combine images outside of a series
unless it is known apriori that all the images meet the above conditions. CT transverse scans may be taken with the
gantry rotated, as long as all images are parallel to each other. The program will not allow non-coplanar
images to be together in a stacked image set.
Nor may two images occupy the same plane within a stacked image set.
Coordinate System
Each
image file must consist of patient orientation information and the coordinates
specifying the plane of the image relative to a common patient coordinate
system. Dicom 3 specifies that positive
Z axis pointing towards the patient’s head, with the Y axis pointing down (for
a supine patient) in the posterior direction, X axis to the patient’s left. Be aware that this program uses the IEC
coordinate system with the positive Z axis pointing anterior, Y axis toward the
patient’s head, and X axis to the patient’s left. Dicom 2 and 1 specified images relative to the imaging equipment,
not the patient. In that instance we
must rely on designations within the file as to feet first or head first,
supine or prone.
New
To
create a stacked image set select New under Stacked Image Sets on the main tool
bar. A name unique to the patient must
be provided by the user. A patient may
have any number of stacked image sets.
To select images for a stacked image set, the program will provide a
file selection box starting at the directory specified in the file
NewImagesDirectory.loc in the program resource directory.
Choosing
a directory on the left of the popup will show the files in that directory on
the right. Navigation is also possible
with the filter function at the top of the popup. Note on the bottom that the files may be copied or moved into the
patient directory. Moving means that
the file in the images directory will be erased after it is successfully copied
into the patient’s directory under the new stacked image set name. The entire file is copied or moved
unchanged.
|
Select Images Files for a Stacked Image Set Popup |
Normally
you should configure your Dicom input so that a patient’s series is stored in a
separate directory under the new images directory.
After
you have selected all the images for a stacked image set, the images will all
be displayed on a screen.
Computed Scout Views
For
stacked image sets with more than one image, a coronal, sagittal, and
transverse scout view is computed. The
plane for these scout views is through the center of the volume of the image
set. Each scout view is computed by
considering parallel (non-diverging) rays through the image set and averaging
the pixel value of the pixels that the ray goes through. Pixels with a value less than a threshold
specified in the program resource file ComputedScoutImages are not
averaged. This is to reduce the dynamic
range of the resulting image.
Upon
display of a new image set, the computed scout views may be useful in locating
errors due to patient motion during the imaging process, as discontinuities may
result and be discernable in the scout views.
Orientation Labels
For
each plane displayed, the edges of the image will be labeled among anterior,
posterior, superior, inferior, right, and left. To see the labels you may have to enlarge the frame or an image
to full screen by hitting the button in the upper right hand corner of the
frame. The text for these labels comes
from the program resource file PatientOrientationLabels (under the language
subdirectory). For oblique planes the
labels may be combined for the dominate directions, such as Ant-Sup.
It
is absolutely essential that the orientation labels on the two dimensional
images be verified. Although one is
unlikely to confuse anterior and posterior, left and right is another matter,
particularly in the head and pelvis.
Even in the thorax, there are rare individuals whose organs are reversed
left to right from the norm.
We strongly
recommend that a protocol be established with the imaging center to place a
marker consistently on one side of the body to eliminate any possibility of
confusion between left and right.
If patient
orientation is critical in any way, you must have some means of verifying the
correctness of the orientation shown on the image display.
As
a corrective measure you may flip the stacked image set across any of the three
orthogonal planes in a x,y,z coordinate system. However, be extremely cautious if you do so, as it is unlikely
that the display would be wrong. An
error would have to been made during the imaging process in specifying the
patient orientation to the imaging system, such as a patient placed feet first
instead of head first but specified as being head first. Or an error occurred in writing out the
Dicom files.
Do not flip an
image set unless you know precisely where and what caused the orientation
labels to be wrong and that you know precisely how to flip the image set to
correct the error. It would not do to
create a mirror image of the patient, or create an error when there was none,
or create a different error.
Be
sure the labels are correct before you do any further processing.
As
an aid, routine rlDicomDump is provided to dump the contents of a Dicom image
file, with decoding of some pertinent fields.
Type rlDicomDump followed by the file name to read. rlDicomDump may be found in the subdirectory
tools.dir.
Edit
You
may edit an image set at any time.
However, we recommend that you complete the image set before any further
processing is done. Otherwise
confusion may result, for example, if an outline is made on an image but then
the image is deleted from the image set. Besides flipping an image set as
described above, you can delete an image from an image set if there is some
reason to do so. Normally you would
want the most complete model for the patient, but an image file may have been
corrupted or mistakenly included in the image set.
Edit
Stacked Image Set Toolbar shown with the Flip Image Set Popup
You
may also add images to the image set.
But, as noted above, these images must be stackable, taken of the
patient during the same imaging process without patient motion. This function is provided as a means to
correct a situation where all the images were accidentally not included in the
original selection of the image set.
Selection of a stacked image set
The
program allows for displaying more than one image set. Consequently, through out the program there
are instances where an image set must be specified. For example, in reformatting an image the program has to know
which image set the image is to be reformatted from. Be careful of situations where an image set is selected and then
a new tool bar is pushed. You will
operating on the selected image set.
Stacked Image Set Options
Controls
are found under the Stacked Image Sets pulldown on the main menu and are
described throughout this manual.
Under
Options is the Options toolbar.
|
Stacked Image Set Options Toolbar |
The
options selected here are covered under the appropriate chapters in this manual
for the particular topic that is selected for each choice presented in the
toolbar.