QCT as a Base of Computer Aided Diagnosis of Osteoporotical Changes

The osteoporosis is a systemic disease of bone which leads to progressive decrease of the bone mass and the changes of bone structure. As the years went by these changes became so serious that it can cause the disorders of functioning of bone in organism (bones become weak and more subject on fracture). At people suffering from this disease more often occurred the fractures e.g. in reach of hip joint and vertebral of spine. One of characteristic feature of osteoporosis is asymptomatic progress. The first signs – difficulties during move and appearing pains in the spine and in the hip joint appear when it is the big loss of bone mass and it is the large risk of fractures. Unfortunately it is a serious phase and the fractures are common – after fracture one should stabilization of places of fracture. Differences between correct tissue and tissue with osteoporosis are presented in Fig.1 [1].


Introduction
The osteoporosis is a systemic disease of bone which leads to progressive decrease of the bone mass and the changes of bone structure.As the years went by these changes became so serious that it can cause the disorders of functioning of bone in organism (bones become weak and more subject on fracture).At people suffering from this disease more often occurred the fractures e.g. in reach of hip joint and vertebral of spine.One of characteristic feature of osteoporosis is asymptomatic progress.The first signs -difficulties during move and appearing pains in the spine and in the hip joint appear when it is the big loss of bone mass and it is the large risk of fractures.Unfortunately it is a serious phase and the fractures are common -after fracture one should stabilization of places of fracture.Differences between correct tissue and tissue with osteoporosis are presented in Fig. 1 [1].From mechanical point of view the fracture of bone occurs in two cases: • the correct structure of bone but the loads are so big that cause the stresses larger than stress limit, • the disorders of bone structure caused decrease of strength properties of bone when normal activity of organism can results stresses larger than stress limit.

www.intechopen.com
Computed Tomography -Clinical Applications 152 The attention is concentrated on the second situation (when only the physiology loading of organism can cause accidental fracture of bone) because it takes place in osteoporosis.Very important factor is diagnosis of osteoporosis.It is a serious problem because this disease progress without symptoms.First signs appear when it is the big loss of bone mass (about 30%) and the risk of fracture is high.Comparison of radiological images of hip joint for healthy state and state with osteoporosis is shown in Fig. 2 [11].Usually patients come to doctors when have difficulties during move and pains in spine and/or in hip joint.Unfortunately until then osteoporosis is recognized and treatment is begun.The treatment of osteoporosis usually consists of providing analgesic and stabilization of places of fractures.It would be better to prevent this disease because lack of movement is causes of weakness of bones.Knowledge of physical properties of bone tissue is helpful in diagnosing of the diseases of the bone system (especially that properties change during progress of disease).
To diagnose the osteoporosis the following preventive examinations are realized: a. Radiology Absorptiometry (RA), b.Roentgen Absorptiometry: -Single X-ray Absorptiometry (SXA), -Dual-energy X-ray absorptiometry (DXA), c.Quantitative Ultrasonography (QUS), d.Quantitative Computed Tomography (QCT).These methods work on the base of Lambert-Beer's low which describes phenomenon of weakness of the radiation during crossing by through object.In the course of examination the part of radiation is absorbed or distracted.The intensity of radiation depends on thickness as well as the content of minerals in the bone.In result of iterative reading of photos for individual pixels, after transformation, the measurement density of whole object is obtained.In estimate the progress of osteoporosis the largest meaning have DXA and QCT.These are densitometry methods which risky of osteoporosis estimate on the basis of distributions of density of bone tissue.In literature there are many publications about using both of these methods in aid of diagnosis osteoporosis.Good source of knowledge about computed tomography is paper [3].Cierniak R. characterized this method in overall way.He presented the algorithms and the principles of creating images.He described also applying measuring techniques as well as CT scanners.Computed tomography is inverse problem -on the base of gathered data the model is created (on the basis of information about absorbed radiation in individual slice of scanning the images are generated).Adams J. E. in [2] described in details QCT.Authoress go into genesis of these method, its development, clinical applying, in particular diagnosis of osteoporosis.She presented different variety of QCT and described density phantoms Adams J. E. compare QCT with other methods and gave example when QCT is more useful: in case when we can obtain more information about researched bone (because images from QCT after conversion are good sources of information).Sawada K. et al. in papers [16] and [17] pay attention that if we want to find physical properties of bones same places are more useful (e.g.spine or hip joint) and same places are less useful (e.g.wrist or limbs).Engelke K. et al. in paper [4] presented the official state the International Densitometry Society about applying of quantitative computed tomography and peripheral quantitative computed tomography in diagnosis of osteoporosis at adults person.Authors pay attention that although the dual energy x-ray absorptiometry is often used in recognition of osteoporotical changes, in some cases better solution is applying the QCT.Engelke K. et al. emphasize that important matter is calibration of density phantom (the more accurate calibration the better results from tomography).Authors put many details about methods of performing QCT examinations, technical parameters of tomographs as well as gave advice to interpreting of results from tomography.On the base of the papers: Adams J. E. [2], Engelke K. et al. [4] and Nayak S., Roberts M.S. & Greenspan S.L. [14] we can say that when we compare both of these methods: DXA and QCT, more precise data of bone density is from QCT (from DXA we obtain "surface density" -in g/cm 2 , from QCT "volume density" -in g/cm 3 ).Besides, in DXA is more difficult to distinction the kind of bone at the densitometry images.Second reason is the fact that data from QCT is more useful to create numerical model.In our case better solution is to use quantitative computed tomography -that's why we applied this method.

Description of QCT
In this method the tomograms from CT are used to analyze the mineral density of bone.Through using the composition of projection images from different directions one can get cross-sectional and solid images in all researched structures (Fig. 3) [19].Tomograms consist of individual voxels.Each voxel is characterized by coordinates x, y, z and color in gray scale.On the base of amount of radiation (absorbed in different places) one can determined density in these places with exact to one voxel.QCT differs than standard approach occurrence of density phantom which is X-rayed together with patients.The phantom is composed of regions representing specimens of bone density in Hounsfield Units.The density phantom is presented in Fig. 4.

Computed Tomography -Clinical Applications 154
Here, phantom is composed of six specimens.On the base of these standard density the calibration curve is drown.Formula of this curve enables to determinate the density for each voxel of researched bone.The calibration curves for four series were presented on Fig. 5.

Quality of images from QCT
Very important problem is quality of images from radiology examinations.On the base of obtained information, it is deducted about patient's state, disease's progress and it is possible to plan the treatment and check correctness of therapy.In case of QCT, when dates are often converting, quality of tomograms has special meaning.There are a few important features of images, which decide about useful in medical diagnostic: contrast, sharpness, resolution, noise-to-signal ratio, artifacts and distortion of signals.
Occurrence hums and disturbances is one of the characteristic features of signals.In dependence on method there is different kind of hums.Increase of hums level lead to lowering of visibility and the decrease of contrast.In a consequence photographs gives little information and are less useful.Hums are observed on the every stage of transformation data and it would be the best if minimization of disturbances started from the very beginning of creation of images and lasted during whole registration process of the data.Hums and disturbances can influence on incorrect interpretation of photographs.Influence of hums can be limited by using the low-pass filters and locally enlarging the contrast.Unfortunately during reducing of disturbances it can lose the part of information which will influence on level of medical content.The artifacts are others problems during conversion of photographs from radiology.These are unwanted feature of images appearing in medical modalities.The artifacts do not reflect the properties of researches structures but they are the result of accidental factors.In  Sometimes it does not influence significantly on perception of diagnostic information, however in some cases it can limit accuracy of interpretation or even mislead (the artifacts pretend features of living structures).Many different factors can cause formation of artifacts e.g.: methods of data processing, algorithms of creation of images, movements of patients or shift of X-ray lamp.During QCT examinations frequent artifacts are appearing defects on boundaries of areas with different density -in images are local decrease of sharpness and contrast, the rise of shadows or the exaggerations.In a consequence it can cause to incorrect estimation of material properties (Fig. 12) [19].

Numerical model of human pelvic bone
After performing tomography examination next step is to convert obtained images to get effort state of bone.Data from QCT was used to create the numerical model of human pelvic bone.First, the geometry of the model should be prepared.The geometry from tomography or from measuring coordinate machine can be applied.After completing prepared model on boundary conditions (it is presented in Fig. 14), the calculations on the base of FEM are performed.Important problem is to determine the changeability of material properties in numerical model.It was achieved in two approach -they became describe later.

Numerical simulations of osteoporotical changes
In view of difficulties and limited access to images from tomography the numerical models containing osteoporotical changes were created also in different methods.Second approach consist of division whole model to subregions -groups.Cortical bone was divided into five subregions, in each subregions elastic modulus is in the some range of value.In Fig. 18    (5) where: E -elastic modulus, ρ -radiological density of bone tissue.
It is possible to change these formulas to another and computation distribution of material properties according to different relationships.
The exemplary distributions of stresses, strains and displacements were presented in Fig. 21, 22 and 23.The maximum reduced stresses (Fig. 21.) and strains (Fig. 22.) appear in the head of femur and in the joint between sacral and pelvic bone.Some results of numerical simulations are put into Table 1 and Table 2.The maximum value of displacement and reduced stress in dependence on loss of bone mass (in percent, in upper part of ilium ala) are shown.Next, in Table 2 the same dependence is shown when the loss of bone mass appear in pubic symphysis and sacral joint.Additionally, the value of maximum strain is prescribed.information about stresses only is not unambiguous (it is not enough information about loss of bone mass) however when we also consider displacements we can determinate how loss of bone mass is in each points.

Data base
After QCT examinations the tomograms of researches structure are obtained.At the first, it is necessary to determinate the radiological density.To transformation of that quantity one should rescale it to HU.The next step it is delimitation the density of bone tissue (on the base of density in HU).By using relationship between density of bone tissue and material properties it is possible to calculate material parameters of bone tissue in each voxel.From the other hand, on the base of QCT data it is possible to create the geometry of numerical model.When the model is prepared the material properties are inserted and the strength calculation (according to FEM) is performed.On the ground of obtained results (distributions of stresses, strains and displacements) one can get to know about effort in researched bone.
When the large amount of QCT examinations will be realized (the set of data will be converted and strength calculations will be performed) it will be possible to create the base of data.This base will be aid detection of osteoporotical changes in human pelvic bone.This data base one can compare to table -matrix which consists of CT photos: column symbolize images for individual patients, row vector places where images were done.The base one have to divide into two main parts: QCT images and numerical simulations.General structure of data base is shown in Fig. 25.The data base makes up folders with suitable data and group with the correct order.The fragment of the base is presented in Fig. 26.

Procedure to aid of diagnosis osteoporosis
After building the numerical model the next step of work is to develop procedure to aid of recognizing osteoporotical changes in pelvic bone.The general principle is as follow: after radiology examination it takes place searching to find the most similar images (searching CT photo from data base).When these images are found the whole model with strength parameters is assigned.Next the results (from strength calculations) are analyzed.As a consequence, particular images from QCT are subordinated to effort of bone.In case of need it is possible to return to searching of base and analyzing the larger number of data.The simplified block diagram of the program is presented in Fig. 27.In procedure of searching on the basis of QCT, one can first read image -which will be analyzed, next to read images from data base.When sorting is ready, one should to insert number of image, which is the most similar, to the textbox of the program -the information about dangerous state will appear in the window of the program.This selected views of procedure box are presented in Fig. 28.
In procedure of searching on the basis of loss of bone mass one should indicate the proportional decreasing of bone mass (from the combo box) in each subregions of cortical bone in numerical model of pelvic bone.When all list are ready in the window of the program will appear information about dangerous state and potential dangerous of osteoporosis.This procedure was presented in Fig. 29.

Future plans
• Extension of data base.The structure of data base enables its easy extension.The more data collected in the base it means the bigger possibilities in aid of diagnosis.

•
Widening this procedure to another element of bone system.Here, the example of pelvic bone was presented, but one can also applied this procedure to another bone and joint e.g. for the spine or metatarsus bone.To perform this one can dispose data necessary to prepare the numerical models and create proper data base.

Fig. 8 .
Fig. 8. Procedure to converting images from QCT to create numerical model

Fig. 9 .
Fig. 9. Tomography image and cross-section in numerical model Exemplary part of matrix with values of Young modulus is shown in Fig. 10.

Fig. 10 .
Fig. 10.Fragment of table with values of elastic modulus obtained from tomography [GPa] Fig. 11 the comparison between normal image and image with artifact is shown [4].

Fig. 12 .
Fig. 12.The artifact caused by metal filling [19] The model consists of three main parts: -pelvic bone (compact and trabecular tissue), -endoprosthesis of hip joint (cement layer and artificial acetabulum), -femur head (metal or ceramic).The components and whole model of human pelvic bone are presented in Fig. 13.

Fig. 13 .
Fig. 13.Component of model of pelvic bone Characteristic feature of the model is subordinating individual material parameters to each finite element -this represents non-homogenous structure of bone and changes in pelvic bone during osteoporosis (model is closer to real conditions).

Fig. 14 .
Fig. 14.Meshing and boundary conditions of the model

Fig. 20 .
Fig. 20.Determining of elastic modulus in individual finite element

Fig. 26 .
Fig. 26.The part of data base

Fig. 27 .
Fig. 27.The block schema of the programIn the program there are available two searching procedures: on the base of tomography images and decreasing of bone mass.In procedure of searching on the basis of QCT, one can first read image -which will be analyzed, next to read images from data base.When sorting is ready, one should to insert number of image, which is the most similar, to the textbox of the program -the information about dangerous state will appear in the window of the program.This selected views of procedure box are presented in Fig.28.In procedure of searching on the basis of loss of bone mass one should indicate the proportional decreasing of bone mass (from the combo box) in each subregions of cortical bone in numerical model of pelvic bone.When all list are ready in the window of the program will appear information about dangerous state and potential dangerous of osteoporosis.This procedure was presented in Fig.29.

Fig. 28 .
Fig. 28.The algorithm of the searching on the base QCT images

Table 1 .
Displacements and stresses for loss of bone mass in upper part of ilium ala

Table 2 .
Quantities for loss of mass in pubic symphysis and joint of pelvic and sacral bone

•
Applied procedure facilitates interpretation of data from QCT and it helps diagnosis what enable earlier detection of osteoporosis and enlarges chance of the treatment.•Presentedproceduredeliversadditional information about effort state in analyzed bone.•Informationfrom QCT can be helpful for researching progress of osteoporosis in individual clinical cases (because easier one can find the differences between earlier and later images).