Stereotactic Frame... 1

Stereotactic Frame File. 1

Stereotactic Frame Test 4

Procedure. 4

Results. 5

Select Frame. 5

Locate Frame. 6

Verification of Solution. 7

Display Control 8

Getting Stereotactic Coordinates of Points. 8

Delete Stereotactic Frame Association. 9


Stereotactic Frame

The Stereotactic Toolbar

Select the toolbar for the stereotactic frame under Options under Stacked Image Sets on the main toolbar.

Stereotactic Frame File

You must write a file for the stereotactic frame that you are going to use.Example files are under the directory frames.d in the program resource directory.The format of the file is documented in the example files.See the ASCII file standard in the Program Resources Files Chapter.The comments do not have to be present in the file but help to document what is in the file.An example stereotactic frame file follows here:


/* file format version: */ 1

/* name of the system: */ <*example frame file*>

/* Number of rods = */ 9


/* Following below will be the diameter of each rod, and its end points.The rods will be numbered by the below order.


†††††††††††††††††††† All in Centimeters

Diameterx†††††† y††††† z††††††† x†††††† y†††††† z */


0.952514.0†††† 0.0††† 9.25††† 14.0†††† 0.0††† -7.65

0.635††† 7.36†† 11.51†† 9.25††† 13.64††† 0.62†† -7.65

0.635††† 7.0††† 12.1249.25†††† 7.0††† 12.124-7.65

0.635†† -7.0††† 12.1249.25††† -7.0††† 12.124-7.65

0.635-13.64††† 0.62†† 9.25††† -7.36†† 11.51†† -7.65

0.635-14.0†††† 0.0††† 9.25†† -14.0†††† 0.0††† -7.65

0.635†† -7.0†† -12.1249.25††† -7.0†† -12.124-7.65

0.635††† 6.29-12.1249.25††† -6.29-12.124-7.65

0.635††† 7.0†† -12.1249.25†††† 7.0†† -12.124-7.65


/* Next follows for the typical use of the frame the conversion matrix FROM the coordinate system of the frame TO patient IEC coordinates.This matix will typically be a rotation matrix only.The last row of:†† 0001††† is NOT entered.This matrix is used for the purpose of displaying patient IEC axes on a 3d view of the frame along with the frame's coordinate system axes.Locating the rods on the images will establish the actual transformation between the patient's coordinates and the frame's coordinates.


Here we have a translation of (1.0, 1.0, 1.0) so that the two axes are not drawn on top of each other initially.



†††† -1.0†††† 0.0†††† 0.0†††† 1.0

††††† 0.0†††† 0.0†††† 1.0†††† 1.0

††††† 0.0†††† 1.0†††† 0.0†††† 1.0


/*color name for the system of rods: */†† <*seashell*>

/*color name for a picked rod: */†††††††† <*red*>


/* The following are material properties for the rods: */

/* ambient */††† 0.5

/* diffuse */††† 1.0

/* specular */†† 0.0

/* emission */†† 0.0

/* shininess */0.0

/* transparency (0 clear, 1 opaque */ 0.75


The above is an example file and must not be used without confirmation and testing as described below.


In the file you must specify the coordinates of the end points of each rod and the diameter of each rod.To do so you must determine each end of each rod in the coordinate system of the stereotactic frame.What the program does is solve for a rigid transformation between the patient coordinate system and this stereotactic frame coordinate system (see under Image Fusion for the definition of a rigid transformation).Points located in the patient system can then be reported in stereotactic coordinates.


The file you write completely determines what will be the stereotactic coordinate system when you enter the rod coordinates.You must test the result.


In the file you will also provide the first three lines of the transformation matrix that will rotate and translate the stereotactic space to patient IEC space.This matrix should consist of a 3x3 rotation matrix with the last column representing a translation.The last row is the vector (0,0,0,1)which provides for translation in matrix multiplication with vectors of the form (x,y,z,1) and is not written in the file.The IEC system has the positive Y axis pointing toward the patientís head, the positive Z axis pointing anterior, and the positive X axis pointing to the patientís left.This transformation matrix must be provided so that the program can draw in both the patient IEC axes and the stereotactic frame axes on the same 3d view of the stereotactic rods.This drawing will help you to turn the 3d view of the stereotactic frame so that it is positioned similarly to its presentation on the 2d images you will be locating the rods on.The IEC axes will be drawn in the color used for axes (defined in file FrameAxesColor in the program resource directory).The stereotactic axes will be drawn in the unpicked color of the rods defined, in the stereotactic frame file.


The rods in the file will be numbered in the order in which they are listed in the file.A CT or MRI scan will represent a plane that intersects the rods.Where the image plane intersects a rod, a cross sectional image of the rod will appear in the image.You will have to locate on the image which rod is rod one in the file, and then which rod on the image is rod 2 in the file, and so on for all the rods.To do this you will pick each rod on a 3d view of the rods, and then locate the corresponding rod on the image.It may be important that you can orient the 3d view of the rods to correspond to the orientation as seen from the images.


Be aware of a rod frame design that is symmetrical.If there is more than one orientation that matches the intersection with images, you could mistakenly match the rods up with the wrong intersection images.To do so would result in an incorrect solution.


Consider for example the Leksell frame made by Elekta Instruments AB (Sweden).The CT rod frame consists of side pieces with two parallel lines and a connecting diagonal line.There is also an anterior plate again with two parallel lines connected with a diagonal line.However, the anterior plane can move relative to the side pieces and so moves relative to the coordinate system of the frame.For our purposes this presents a problem since we have to know the coordinates of the end points of the lines that show up on the CT scans.If the anterior plane can move, than the coordinates can change and may be wrong.Therefore you must either not use the anterior plane or be sure that it is in a particular position for which you have determined the coordinates (you might consider pinning or fixing the plate to one position if you are going to seriously use this frame with this software).But without the anterior piece, the side pieces are symmetrical; you can flip the system upside down (head to feet) and the rods can still be matched up to align with the CT scans.But you would be matching the wrong rod with the image of a rod.You would be determining stereotactic coordinates incorrectly, flipping points from anterior to posterior and head to toe in this case.


The program solves for the orientation of the rod frame mathematically without any assumptions about the rod frame being parallel to or oriented in any particular way with the imaging system.This removes the dependency of aligning the system in any particular way with the imaging system and provides for a generic approach that will work with any frame design.Further, misalignments during imaging are no longer a concern.


But frame systems do generally attach to the patient in a particular orientation.The transformation matrix between the stereotactic system and the patient IEC system provides a visual clue for matching up the rods correctly.In our Leksell example above, the Leksell system has the positive Z axis pointing towards the patients feet, the positive Y axis pointing anterior, and the positive X axis pointing towards the patientís left.The rotation and translation matrix from the Leksell stereotactic coordinate system to the IEC system is thus:

















†† 1.0


Note that in the file the last line of the matrix is assumed and not entered, since it will always be (0,0,0,1).Multiplying this matrix on the right by the column vector (x, y, z, 1.0) will give the result (x -10, -z +10, y -10, 1.0), where we see that the x coordinate remains the same, the new y coordinate is negative the z, and the new z coordinate corresponds to the old y.Note that we moved the origin of the stereotactic system to the center of the frame in IEC coordinates.This is just for convenience and appearance sake on the 3d view.


After solving for the transformation between stereotactic coordinates and patient IEC coordinates, the patient IEC coordinate axes are redrawn in the stereotactic space on the 3d view of the rods in the position determined by the solution.This serves as another visual clue to the correctness of the solution.


As noted above, you are completely responsible for testing the program with your frame file.Particularly, if you are going to locate points in stereotactic coordinates you must do a test with a phantom to test the ability to locate points.


To write a file for your frame, you could simply copy one of the example files and use it as a template for making the new one.

Stereotactic Frame Test


Choose a phantom of some sort that you can put in the frame and on which you can locate points that will show up on the scans.For CT you can use small lead markers.Place the markers on or inside the phantom.But you will have to measure the coordinates of the these points in stereotactic coordinates.You can do this either mechanically, or use a film method if the system provides a method for determining coordinates from orthogonal films.Also consider carefully how you determined the coordinates of the rods.A good test would be to use some method provided by the stereotactic system for determining the coordinates of the points to eliminate the possibility of performing the same error in determining the end points of the rods and the coordinates of the test points.


Scan the phantom.You should use a small slice spacing.Make note of the fact that the resolution perpendicular to the scans will depend upon the slice spacing and is generally larger than the resolution in the plane of the scans.Locate the stereotactic frame and solve for its position.Then use the points (under the stereotactic toolbar) popup to determine the stereotactic coordinates of the points visible on the images.


If the scans are perpendicular to an axis in the stereotactic system, you may want to compute the distance between the two determinations of each point both in the plane of the scan and the three dimensional distance.Let (stX,stY,stZ) be the coordinates of a point determined from direct observation of the phantom or measured with the stereotactic apparatus.Let (pX, pY, pZ) be the coordinates determined with this imaging system from the scans.Let us assume that transverse scans were taken and that they are parallel to the x,y plane of the stereotactic system.Then for each point compute the in plane distance:


In plane distance†† =†† sqrt( (stX-pX)*(stX-pX) + (stY-pY)*(stY-pY)


And in three space for each point compute:


Three space distance†† =†† sqrt( (stX-pX)*(stX-pX) + (stY-pY)*(stY-pY) + (stZ-pZ)*(stZ-pZ)


The results should be of the order of the larger of the measurement accuracy for determining the coordinates of the test points and the pixel size and slice spacing.The accuracy in plane is limited by the pixel size of the images.The accuracy in three space is limited by the the larger of either the pixel size or the spacing between the scans.In general, you should be within one millimeter for most of your test points.


You are responsible for establishing what the performance accuracy is for your equipment using the above test procedure.

Select Frame

Before locating a stereotactic frame you must first hit the New button on the stereotactic frame toolbar to choose one.You will be presented a list of all the files that are in the program resource directory frames.d.Choose the correct frame.As there obviously can only be one frame bolted to the patientís head (or one body frame in place), there is only one frame possible per stacked image set.You must select the one that was used.Most likely, you have only one frame to choose from.

Locate Frame

Locate Stereotactic Frame Popup Tool

Hit the Locate button on the stereotactic frame toolbar to solve for the location of the frame relative to the patient.The program solves for a transformation of coordinates between the patient space, defined by the stacked image set,and the stereotactic space.The transformation is rigid, with only rotation and translation allowed.


A popup will appear that will include a 3d view of the rods.The first time you should check this carefully against the actual rod cage.Carefully inspect the 3d presentation against the actual object.They should agree.Note also the direction of the stereotactic coordinate axes (shown in the same unpicked color) and the patient IEC axes (shown in the default color for axes and labeled IEC).The axes should be oriented correctly relative to the rods.


You can select specific rods by clicking the left mouse on the 3d view of the rods.The rod will change color.You can unselect the rod by clicking again on a selected rod.You can select all the rods or unselect all the rods with the SelectAll and UnSelectAll buttons.


For the selected rods, click the left mouse on the corresponding image of each rod in the 2d image from the stacked image set.Locate the rods in the same order that you selected them on the 3d view of the rods.If you had hit SelectAll, the order is the numerical order of the rods.As you locate each rod on the image, that rodís color will change back.The select and unselect colors are specified in the specific stereotactic frame file that you are responsible for.


You should locate all the rods in at least the first, middle, and last image in the series of stacked images.After locating the rods on the images, hit Compute Transform to generate a solution.A rigid transformation between the stereotactic frame and the coordinate system of the stacked image set is solved for.The images are not solved for individually.The relationship among the images is known.If there has been patient movement during the imaging process some images may not align well with the solution.This will be evident below when you project the frame onto all the images of the stacked image set.


From the solution, the points that were located on the images will be transformed to the space of the stereotactic frame and drawn in the 3d view.Make the rods more transparent to increase the visibility of these points.The points should all land within their respective rods.


Note also that the average distance of the points from their respective rods is computed, along with the maximum distance seen.This should give you some indication of how well things are going.If you made an error and associated a point with the wrong rod, you should see a larger average distance and the maximum error seen should be large.


After computing a transform you can add more rod location points and again compute the solution again.


To remove specific points from the list used for computing a solution, simply click the left mouse on the point on the 2d display.Note that the corresponding rod on the 3d display will not revert back to a picked status.You can delete all location points with the Delete All Points button and start over again.

Verification of Solution

You should then verify the correctness of your solution by examining where the rod points you located on the 2d images fall within the 3d view of the frame on the locate popup.Then use Display Control to view where the rods intersect all the images of the stacked image set.In both cases the rods and their images should correspond exactly.If you have made any kind of error, delete the solution and start over again.The biggest source of error may be due to associating the images of a rod with the wrong rod.Or patient movement may have occurred during the image scanning.

Display Control

Stereotactic Display Control

Here you can control what is shown about the frame with the stacked image set.†† The top toggle button controls whether or not the location points are drawn.If you turn these points off, and then add more points, the new added points will be drawn.To turn them off first turn all points back on and then all off again to catch the additional new points.


The second toggle button will allow the frame to be drawn where it intersects an image plane.Use this feature to further check the correctness of the found transformation between stereotactic frame coordinates and all images of the image set.


The last button will control the drawing of the frame in 3d views of the image set.The transparency scale is for those views.This does not effect the 3d view of the frame in the location popup.At this time we don't provide separate control over individual viewing windows.


Use these tools described here and above to check the accuracy of the stereotactic frame solution.

Getting Stereotactic Coordinates of Points

Get Stereotactic Coordinates for a Point Popup Tool

Use the Points button on the stereotactic tool bar to get the stereotactic coordinates of a specific point that you locate with the mouse on an image that is in the image set or is fused to the image set.You have located the stereotactic frame in a stacked image set.Here when you locate any point in that stacked image set with the mouse the stereotactic coordinates of that point will be displayed.


The stereotactic coordinates is the coordinate system of the stereotactic frame defined by the corresponding file in directory frames.d in the program resources directory.The coordinate system is defined by the coordinates of the end points of the respective rods used for localization of the frame.


You may also locate this point in any other stacked image set that is fused to the stacked image set the frame was found in.Be careful that you understand which frame coordinates you are locating a point in if that other fused stacked image set also has a frame.You are finding the frame coordinates for a point in the frame that belongs to the stacked image set that was current when you selected the stereotactic frame option under "Options..."

Delete Stereotactic Frame Association

Lastly, use Delete to remove the stereotactic frame from the image set.You should remove the frame if any error occurred concerning locating it.You do not want the association lingering around if it is not right.When the program is restarted and the image set is again selected, that association will remain unless deleted here.