Method of generating a radiation image of an elongate body
[DESCRIPTION]
FIELD OF THE INVENTION
The present invention relates to a method of generating a
radiation image of an elongate body by taking plural partial X-ray
images on a digital radiography detector using a multiple shot
exposure technique .
BACKGROUND OF THE INVENTION
In radiography, sometimes an image of a region with long length
is taken, such as the entire spine or the legs. In Computed
Radiography (CR) , images are taken with Imaging Plates (IP) which
partially overlap each other and a long length image is created by
combining the partial images. Accurate alignment and measurement can
be obtained by superimposing an object of known geometry covering
the region to be imaged and correcting and aligning the partial
images to reconstruct the object of known geometry (see EP0919858,
EP0866342) . This technique does not suffer from patient movement
since all images are acquired in a single X-ray exposure.
In recent years, Digital Radiography (DR) has become a valuable
alternative for CR. The flat panel detectors (FPD) used in DR are
more costly than the IP's for CR, so an alternative to the one-shot
long length imaging technique of CR is needed. This is achieved by
taking plural partial images by moving the position of the FPD while
turning the X-ray tube or moving the X-ray tube parallel to the FPD
and pasting the partial images to obtain a composed long length
image. During this FDP and X-ray tube movement, the patient may
move, hereby introducing artifacts which need to be compensated in
the image composition.
Long length images are mostly taken to perform length and angle
measurements on the subject across an area larger than a single FPD.
It is therefore important to create an image where the
alignment of the partial images of the subject and the calibration
is accurate. If the patient is capable to stand still perfectly,
this can be accomplished with known techniques such as described in
the co-pending European application EP11157111.3 filed on March 7 ,
2011 entitled 'Radiographic imaging method and apparatus' .
Unfortunately, most patients requiring a long length examination
suffer from a condition which makes it difficult to stand still
during the time of the image acquisition. Therefore, the partial
images are positioned with a certain amount of overlap such that a
user is capable of positioning the partial images to form a
composite image based on the image content in the area of the
overlap .
In order to reduce the amount of X-ray exposure, images with
different sizes are needed. This can be accomplished by manipulation
of the collimator which is adjusted to block the x-rays based on the
size of the exposure, hereby reducing exposure to the subject.
European patent application 1 484 016 A l relates to the
acquisition of a composite image with a digital detector. Positions
of individual component images of the composite image are
calculated. First the structures that tend to move are identified
and then non-uniform collimation angles may be used for the
different component images to avoid placing these structures in the
regions of overlap. This method requires identification or
estimation of the position of these structures.
It is an object of the present invention to optimize the
calculation of the sizes of the partial images so as to reduce the
above-mentioned problems occurring in imaging elongate bodies and
originating from patient movement during multiple shot exposure and
image recording.
SUMMARY OF THE INVENTION
The above-mentioned aspects are realised by a method having the
specific features set out in claim 1 . Specific features for
preferred embodiments of the invention are set out in the dependent
claims .
The method of the present invention comprises the steps of
generating a sequence of partially overlapping partial radiation
images of said elongate body by multiple shot irradiation and read
out of a direct radiography detector and pasting the partial images
to form a composite image.
According to the present invention the size and position of the
partial images is determined on the basis of the calculated number
of partial images so that the partial image representing a part of
the elongate body which is most susceptible to movement during said
multiple shot irradiation covers a larger area than the area covered
by partial images representing parts of the elongate body less
susceptible to movement.
Preferably the length of the partial image which is most
susceptible of movement during the multi-shot exposure covers
substantially the entire radiation sensitive area of the detector if
this configuration is allowed by the physical contraints of the
imaging system which is used.
In a preferred embodiment, the length of the partial images is
calculated so that the part of the patient that is most susceptible
to movement during the multiple- shot exposure is recorded in one
shot on a single detector area. The width of the partial images is
commonly constant. The setting of the partial image's length is
achieved by appropriate setting of the x-ray source and radiation
shutter or collimator.
For images of an elongate body such as full spine or full leg
images where the patient is standing in upright position, it has
been observed that the amount of movement is larger for positions
higher from the ground. There are two main reasons for this
observation. First, movement is introduced by the patient's
breathing which is inherently positioned at a high position.
Secondly, a patient is more stable and steady at positions near the
ground. This is illustrated in figure 1 . With this observation, it
has been decided that it is best to take the highest images as large
as possible (given the detector's dimensions) to minimize the
corrections needed to compensate the patient movement.
When a full spine image is taken of a patient in a horizontal
position, the body part that is most susceptible of movement may be
the upper body part. However in case of a lying patient for whom a
full leg image is taken, this may be the lower part of the legs.
In up-right position, physical constraints may imply that the
lower image is an image which covers the largest part of the
detector. In that case the lengths of the other partial images are
calculated so that the partial image corresponding with the body
part that is most susceptible to movement during the multi-shot
exposure covers substantially the entire radiation sensitive part of
the radiation detector.
According to an embodiment of the present invention the total
length of the elongate body to be imaged is first determined.
Next, on the basis of the determined length and (a) predefined
amount (s) of overlap between said partial images, the number of
partial images required to image the elongate body is calculated.
Then, on the basis of the calculated number of partial images,
the size and position of the partial images is calculated so that
the partial image which comprises the body part that is most
susceptible of movement during the multi-shot exposure is imaged on
a larger area (preferably substantially the entire area) of the x -
ray sensitive detector area than the area covered by the other
partial images. The size and position of partial images may be
expressed by settings of x-ray source and collimator or x-ray
shutter .
The size (length) of a partial image can be input by the user
or can be calculated by a processor. Alternatively it can be input
by a device capable of indicating the size on the patient.
The partial images generated by multiple shot exposure are read
out and pasted to form an elongate image.
Further advantages and embodiments of the present invention
will become apparent from the following description and drawings.
BRIEF DESCRIPTION OF THE DRAWINGS
Figure 1 is an illustration of the fact that a patient un up
right position is more stable closer to the ground. The displacement
at the height of the legs (d) is much smaller than the displacement
at the chest (D) ,
Figure 2 is an illustration of a ceiling suspended and a floor
mounted system supporting tube rotation,
Figure 3 is an illustration of the positions of the four
partial images for a given input area with length L . The left figure
uses the complete detector for the highest three images. The right
figure is an illustration of a setup where only the 2 highest images
use the complete detector,
Figure 4 is an illustration of a setup where the bottom image
needs to be bigger than the second lowest image because otherwise
the detector will collide with the floor,
Figure 5 is an illustration of a U-Arm which is positioned in
such a way to keep the X-ray source stable for three positions,
Figure 6 is the description of the geometry and its variables
used in the calculations to compute the positions of a U-Arm in a
three image setup,
Figure 7 illustrates the function of yi in function of with
given df = l,d m = 0.8,5 = 0.43, v = 0.07 ,
Figure 8 is a flowchart describing the steps involved to select
the number of images and geometry configuration for a system capable
of performing long length imaging.
DETAILED DESCRIPTION OF THE INVENTION
X-ray systems capable of performing long length imaging are
available in different configurations, each controlled differently
to obtain optimal results for long length imaging. Common parts of
the configurations are: an X-ray generation unit including an X-ray
source that generates x-rays ; a collimator unit which is adjustable
and reduces the area on which X-rays are projected; an X-ray imaging
unit capable of collecting images based on the generated X-rays.
Most modern systems include controllers to control the X-ray
generation unit and X-ray imaging unit. Automatic systems, needed to
perform automatic long length imaging, also include position
mechanisms and controllers for the positions of the X-ray generation
and X-ray imaging unit. In systems such as a C-Arm or U-Arm, some
mechanisms and controllers are combined.
According to the present invention it has been decided to take
the partial image representing that part of the elongate body which
is most susceptible of movement during the multiple shot exposure as
large as possible (given the detector dimensions) to minimize the
corrections needed to compensate for the occasional patient
movement. This has the added advantage that for a regular up-right
full spine image starting from the atlas, for most patients the area
containing the lungs and the heart is imaged in a single shot. This
further reduces the artifacts introduced by the patient movement and
avoids double exposure of the heart .
Most prior art systems divide the area which needs imaging into
equal parts. This reduces the complexity and the number of
computations involved to position the X-ray generation and X-ray
imaging unit correctly.
In the next sections, formulas for a ceiling suspended system
and a U-Arm will be derived to take the largest images at the
highest positions when the patient is in an up-right position.
It will be clear that similar formulae may be derived for
calculating partial image dimensions (expressed by means of settings
for x-ray source, collimator and detector) when the body part that
is most susceptible of movement is in another position and the
dimensions of that partial image is as large as possible given the
detector's dimensions.
Ceiling suspended or floor mounted system
For a ceiling suspended (or floor mounted) system where the X -
ray source can rotate independently from the X-ray imaging unit,
depicted in figure 2 , the computations involved are given below.
Given a top position T and a bottom position B , the total
length L for imaging is obtained by
L-T - B .
Illustrations which clarify the variables and geometry are
given in figure 3 and figure 4 . The number of partial images N in
which this area needs to be divided is
N = sma ,
where S is the maximum height the detector can image and ov is
the desired amount of overlap.
This allows us to calculate the top positions t , and bottom
positions b for each partial image where
t = T
b . =t,+S,\/i:i N-l
tM = b , - ov, i \< iN '
=
From each pair of positions tj,b , the position of the center
height and collimator size for each partial image is easily computed
with following equations:
MMoosstt mmooddaalliittiieess ssuuppppoorrtt tthhee ccoommbbiinnaattiioonn ooff tthhee vvaalluueess aass
ppaarrtt ooff tthhee iinnppuutt ttoo ppoossiittiioonn tthhee ddeetteeccttoorr aanndd aaddjjuusstt tthhee
ccoolllliimmaattoorr .
WWhheenn llooookkiinngg aatt tthhee lleefftt ppaarrtt ooff ffiigguurree 33,, wwee sseeee tthhaatt ssuucchh a
ccoommppuuttaattiioonn sscchheemmee ccaann lleeaadd ttoo vveerryy ssmmaallll ppaarrttiiaall iimmaaggeess aatt tthhee
bboottttoomm ooff tthhee aarreeaa.. TThhiiss ccaann eeaassiillyy bbee pprreevveenntteedd bbyy cchhaannggiinngg tthhee
ccoommppuuttaattiioonn sscchheemmee iiff a ccoonnddiittiioonn ssuucchh aass ee..gg..
i s true. A new scheme which can b e used i n such conditions i s
V .= s + - R + ov
Figure 3 illustrates the differences between the new
computation scheme (right side) and the old computation scheme (left
side) .
Because long length imaging i s often used t o acquire images
from the legs, the computation schemes described above can position
the imaging unit t o a location which i s unreachable. This i s
illustrated i n figure 4 . Because the imaging unit has fixed
dimensions, the previously proposed computation scheme can not b e
used because the imaging unit would b e positioned below the ground.
Suppose e i s the lowest point which can b e imaged b y the detector a t
its lowest position, w e can redefine a variable ' for which w e
apply the computation schemes above but now with B ' a s bottom
position and a resulting number o f N'+l positions:
T-B'-ov
N'= smallest integer >
S-ov
U-Arm configuration
The computations involved to correctly position the detector
part of a U-Arm while keeping the X-ray source stable are given
below. An illustration of a typical U-Arm configuration is given in
figure 5 . A schematic representation of an U-Arm for a sequence with
3 partial images is given in figure 6 , in which the line [syl-sy2]
represents the positions where the point of rotation of the U-arm
can only move in the vertical way. If we use the complete detector
for the first partial image
t]- = S ,
and put the source of the X-rays at position y = 0 , we can
describe the geometry with following equations:
d + dm cos /?, = s
dm sin/?, = sy
- - = d f+d m ) i fi
d f = x, + d f cos / ,
where d s the distance between the detector and point of
rotation of the U-Arm and d m is the distance from the point of
rotation of the U-Arm to the X-ray source for the top partial image.
Solving these equations for T gives
= sec ?, (d f + d cos/?, s + 2 sin/?, (d f + dm ))
2(df + dm ) "
Similar to the derivation above, we can deduce the position of
t in function of d ,d m , by solving
cos/?, =s
sin/?, = Sy
=x + cos ?,
for t
Similar to the derivation of a computation scheme for the
ceiling suspended system, we can describe 2 situations. In the first
situation, we assume the second partial image is taken using a
complete detector. In the second situation, the area between
(t -ov/2,b 2+ov/2) will be twice as large as the area between
(b +ov/2,B) .
The first situation can be explored as follows:
Solving both equations for t in /?,when ov,S,dm,d is given, is
possible but computational extensive. Since the equation for t2in /?,
behaves nice for optimization (see figure 7 ) , it is also possible to
determine the value for /?, numerically.
In the second situation, the value for t should be (again
under the assumption that our X-ray source is positioned at =0and
B is a negative value)
ov- B
Again both an analytical solution as a numerical solution are
possible for /?, , where the numerical is preferred because of the
lengthy computations involved in the analytical solution.
The formulas above generate all the necessary information for
the positions of the first 2 images, to find the position of the
last partial image we first determine the position of y in function
O f ,
y = s 
s = d + d m - sx
s = d f + d m - sx
m - X = d m - x
s = d + d m cos ?
This leads to
Equation 1
y = tan d f cos + d m cos ) .
To find another function of y in function of ,, first
determine functions for /3and b3 . Given
s - x s3 cos 
s3 = d f + d m - sx
s = d + d m - sx
s = d + d m cos ?,
S 2 2
ov
t —
2
we find
Given
s - x = s cos 
s3 = d + d m - sx2
s = d f + d - sx
s = d f + d m cos ?,
d ~ sx i ={d - sx2 )cosfi
B b .
s2 x 3
the solution for b , is
B d cos + mcos )
3 = d +dm cos /
If we substitute the functions for t3 and 3 in
ov
2(ί - ) =ί - +
we get
Equation 2
2i mcos (5 - ov + t )+ (- ov+ cos 2( 5 -OV + 2t ))
3 + mcos/?, )
Again, an analytical solution for 2 exists but requires a lot of
computational effort. Numerical solutions for the quadratic
difference between
Equation 1 and Equation 2 with all parameters fixed except , will
converge rapidly to the solution of 2 . It is clear that similar
computations can be used for a setup with 2 or more images or for
situations where the detector is positioned near the ground.
In order to determine the number of partial images needed to
cover a given area, no single formula exists. A solution is to use
an algorithm which selects the suitable geometry. Such an algorithm
is depicted in figure 8 .
An example order of geometries for a U-Arm is:
Geometry for 1 partial image
Geometry for 2 partial images: top partial image covers 2/3 of
area, bottom partial images covers 1/3
Geometry for 2 partial images: top partial image uses complete
detector
Geometry for 3 partial images: top partial image uses complete
detector, middle partial image covers 2/3 of area minus top
image, bottom partial image covers 1/3
Geometry for 3 partial images: top and middle partial images
use complete detector
[CLAIMS]
1 . Method of generating a radiation image of an elongate body by
generating a sequence of partially overlapping partial radiation
images of said elongate body by multiple shot irradiation and
read out of a direct radiography detector and by pasting said
partial images, comprising the steps of
determining the total length of said elongate body to be
imaged,
calculating on the basis of said determined length and (a)
predefined amount (s) of overlap between said partial images,
the number of partial images required to image said elongate
body,
determining on the basis of the calculated number of partial
images, the size and position of said partial images, whereby
the partial image representing a part of the elongate body
which is most susceptible to movement during said multiple shot
irradiation covers a larger area than the area covered by
partial images representing parts of the elongate body less
susceptible to movement,
pasting said partial images to form said image of said
elongate body.
2 . Method according to claim 1 wherein said image is a full spine
image of a patient in up-right position and said partial image
representing a part of the elongate body which is most
susceptible to movement is the top partial image.
3 . Method according to claim 1 wherein said image is a full leg
image and said partial image representing a part of the elongate
body which is most susceptible to movement is a partial image
representing the bottom leg part.
4 . Method according to claim 1 wherein the minimum number N of
partial images required to image said elongate body is determined
the formula
L - overlap^
= smallest integer wherein L is the total length of the
elongate body, overlap is a predefined amount of overlap and S is
the size of the digital radiography detector.