- Email: [email protected]
Assessment of mineral density and atomic content of fracture callus by quantitative computerized tomography
J Orthop Sci (2000) 5:248–255
Assessment of mineral density and atomic content of fracture callus by quantitative computerized tomography Feza Korkusuz1, Serhat Akin2, Ozan Akkus¸ 3, and Petek Korkusuz4 . Medical Center, Middle East Technical University, Inönü Bulvarı, 06531 Ankara, Turkey Petroleum and Natural Gas Engineering Department, Middle East Technical University, Faculty of Engineering, 06531 Ankara, Turkey 3 Department of Mechanical and Aerospace Engineering, Orthopedic Biomechanics Laboratories, Case Western Reserve University, 10900 Euclid Avenue, Cleveland, OH 44106, USA 4 Department of Histology and Embryology, Hacettepe University, School of Medicine, Sıhhiye, 06100 Ankara, Turkey 1 2
Abstract: The mineral density and atomic numbers of elements in the periosteal callus and the cortex area of a healing fracture were measured by quantitative computerized tomography (QCT) to obtain accurate information on the mineralization process in rabbit tibia. The mineral density of the periosteal callus was highest on day 15 and decreased gradually throughout the experiment. This was initially detected by QCT, but not with conventional radiography. An apparent decrease in cortical bone density on days 28 and 42 after fracture was observed. Atomic numbers of elements in the cortex remained stable, indicating a possible homeostatic mechanism of mineral preservation at the fracture callus and involved cortical area. QCT may predict density alterations in the fracture callus more accurately than conventional X-ray. Further studies are essential to predict a relationship between mineral density, atomic numbers, and mechanical stability. Key words: bone healing, fracture callus, mineral density, atomic number, quantitative computerized tomography
Introduction Accurate assessment of fracture healing of long bones will obviously improve the outcome of surgical and nonsurgical treatment. Early cessation of immobilization or internal fixation is one of the most common reasons for malunion and non-union. Joint stiffness, on the other hand, is another major problem, caused by excessive immobilization of the adjacent joints in nonsurgical treatment. Precise knowledge of callus mineralization and organization and its impact on mechanical stability is essential to prevent these complications in patients.
Offprint requests to: F. Korkusuz Received for publication on March 8, 1999; accepted on Oct. 5, 1999
Various biological and mechanical factors play important roles in the mineralization process of the fracture callus.4,5,17,25 This process is still poorly understood, in spite of extensive ultrastructural and histochemical investigations. Adequate concentrations of calcium and phosphate ions, the presence of a calcifiable matrix, a nucleating agent, and control of regulators is required for mineralization. Callus mineralization and its relation to mechanical stability is under extensive investigation.14 Although various studies have attempted to quantify parameters of the fracture site,11,13,15,20,24,26 only a few1,8,10,22 were able to demonstrate a positive relation between the mineral content of the callus and the mechanical properties of the fractured bone. The aim of this study was to quantify the mineralization of the callus by quantitative computerized tomography (QCT) and to correlate the results with those of conventional radiography and histology. We also assessed atomic numbers of elements in the fracture callus that may be used to quantify tissue integrity. We investigated the relation between atomic number, mineral density, and histological changes.
Materials and methods Animals and surgical protocol A total number of 24 three-month-old local albino male rabbits, weighing 1500 6 150 g, were anesthetized with intramuscular ketamine hydrochloride and xylazine hydrochloride (100 mg cm23: 23.32 mg cm23) injection. The back limbs of the rabbits were shaved and cleansed with Betadine solution (Kim-Pa, Istanbal, Turkey). A mid-diaphyseal transverse tibial osteotomy at the level of the fibular intersection simulated an open fracture. The periosteum was well preserved and reattached after the osteotomy. Stability was provided by intra-
F. Korkusuz et al.: Fracture callus assessment by QCT
249
medullary K-wire fixation. The tip of the Kirschner wire was specially designed for this procedure; the proximal tip was a triangular pyramidal shape to allow easy passage through the proximal joint line with minimal damage to the cartilage. The distal tip was a conical shape to prevent a secondary fracture in the distal tibial component. A torque-controlled electric powered drill was used to minimize thermal damage and secondary fractures. The non-fractured left tibia served as the control. The animals were killed on days 15, 28, 42, and 90 by hypovolemia. The fractured and control tibiae were removed and preserved at 220°C until the day of assessment. Six bones were assessed for each time interval. The bones were thawed to room temperature and transferred into petri dishes containing saline for QCT. Scanning of each bone lasted approximately 7 to 8 min. The bones were then transferred to Falcon tubes (Becton, NJ, USA) containing 10% formalin and fixed for histological evaluation. Conventional X-ray pictures were obtained at the time of transfer into the fixative. The Ethics Committee of Gazi University Medical School approved all animal procedures. Procedures were in full agreement with Turkish law 6343/2, Veterinary Medicine Deontology Regulation 6.7.26, and with the Helsinki Declaration of animal rights. Density and atomic number measurements Fractured and control bones were assessed with a Philips Tomoscan 60/TX third generation scanner (Philips, NY, USA). The X-ray source of the computerized tomography (CT) permitted the emission of pulse X-rays. Opposite this X-ray source was an x-ray detector in a xenon ionization chamber to measure the distribution of the X-ray intensity. The detector has 576 channels in total, 4 of which at each end are used as reference channels. Scan time changed between 1.9 and 9.0 s, with which 80- to 130-kV energy could be utilized for up to 700-mA tube current. Absorption coefficient distribution was obtained for each pixel in a matrix of 320 3 320 or 512 3 512. The CT number, Ni, was then calculated from the corresponding linear attenuation coefficients, using the following equation:3
Ni =
1000(m i - m w ) mw
Ze =
1 3.8
(Â f Z ) i
3.8 i
(3)
where fi is the fraction of electrons on the ith atomic number species. In order to obtain the density and atomic number, eight to ten images of each bone were obtained at 5-mm intervals in the proximal-to-distal direction. Scans were conducted in 4.5 s, in a field of view of 160 mm with a slice thickness of 2 mm, at 100 and 130 kV. The density and atomic values for each pixel were obtained by solving equation (2). Mineral density and atomic numbers of the fracture callus were randomly measured in at least ten different locations, and the average was defined as the callus density (Fig. 1). Measurements on days 42 and 90 were made in only six locations, as the size of the callus decreased gradually. The cortical density of each bone was measured separately (Fig. 2). The mineral density of the cortex was measured by drawing four to six lines on each transverse section. Cortical mineral density and atomic numbers were measured on each pixel of each of these lines. The location of the line was chosen in such a way that the effect of beam hardening was minimal. Beam
(1)
The linear attenuation coefficient, µ, depends on both electron density (bulk density), r, and the atomic number, Z, in the form: m = r(a + bZ 3.8 E 3.2 )
where a is a nearly energy independent coefficient and b is a constant. The first term in equation (2) represents Compton scattering, which is predominant at X-ray energies above 100 kV at which medical CT scanners normally operate. The second term accounts for photoelectric absorption. One image representing bulk density and another image proportional to the atomic number could be obtained by scanning at high and low X-ray energies (ie, 130 and 100 kV) and by solving equation (2) on a pixel-by-pixel basis. It should be noted that when a mixture of atomic species is present, photoelectric absorption is proportional to the effective atomic number, Ze, which is given by the following equation:
(2)
Fig. 1. Density of the periosteal callus was repeatedly measured in ten different locations in the quantitative computerized tomography sections (schematic drawing of measurements on day 15)
250
F. Korkusuz et al.: Fracture callus assessment by QCT
hardening is due to the polychromatic nature of the X-ray beams passing through the object. As the beam passes through the material, the lower energy photons are absorbed preferentially. As the remaining beams become more monochromatic at higher energy, the beam becomes harder. This causes the overall attenuation to change with distance into the object. Calculations and image analysis were performed on a SUN Sparc-10 workstation (SUN Microsystems, Palo Alto, CA, USA).
distance of the X-ray source to the bones was 100 cm and the tube current was 3.2 mA per s; 42-Kwp energy was used to obtain the images. Films were developed using an Agfa Crurix 60 automatic developing machine. Each bone was exposed in the anterior-posterior and the lateral directions. Histological analysis All specimens were decalcified in 10% formic acid solution for 1 to 2 weeks; the periosteum was stripped off and evaluated separately. Specimens were dehydrated in a graded series of ethanol and embedded in paraffin and then 5-µm thin longitudinal sections were obtained and stained with hematoxylin and eosin (H&E). Photomicrographs were obtained with an Olympus BH7 light microscope (Olympus, Tokyo, Japan). Two-tailed heteroscedastic t-test was performed for statistical analysis.
Radiographic evaluation Conventional X-rays were obtained on Agfa Crurix films (Agfa, Mortsel, Belgium) with a Siemens Multix C X-ray machine (Siemens, Erlangen, Germany). The
Results Density and atomic number measurements Cortical density measurements in the region of interest of the bone sample on days 15, 28, 42, and 90 are presented in Table 1. The density of the fractured tibiae decreased significantly on days 28, 42, and 90 compared with that on day 15 (P , 0.001, P , 0.001, and P , 0.01, respectively). The cortical density of the non-fractured left tibiae (control values in Table 1) remained constant and was always slightly higher than that of the fractured side. Cortical density changed significantly in the fractured bone on days 28 and 42 when compared with values in the non-fractured control side (P , 0.001 and P , 0.001, respectively) and recovered to normal on day
Fig. 2. Cortical bone mineral density measurement. The mineral density of the fractured tibiae and non-fractured control side were measured on each pixel on the lines defined in the cortex
Table 1. Cortical density on days 15, 28, 42, and 90 Days
15
Sample Mean density, g/cm3 Standard deviation, g/cm3 Minimum density, g/cm3 Maximum density, g/cm3
28
42
E
C
E
C
E
C
E
C
2.57 0.17 1.16 1.75
2.56 0.16 1.15 1.57
2.00 0.11 1.30 1.76
2.25 0.15 1.09 1.48
2.07 0.11 1.04 1.42
2.34 0.13 1.11 1.66
2.31 0.20 1.03 1.45
2.41 0.14 1.10 1.54
E, Experiment; C, control
Days 15 28 42 90
90
15
28
42
90
1 1.71E-27 5.53E-18 0.01
1 5.59E-12 1.01E-09
1 0.02
1
F. Korkusuz et al.: Fracture callus assessment by QCT
251
90 (P 5 0.39). Periosteal callus formation was most apparent on day 15. The mean density of the callus on day 15 was significantly higher than that on days 28, 42, and 90 (Table 2). The effective atomic numbers of the elements in the cortical and callus regions are shown in Tables 3 and 4. The effective atomic numbers of the elements in the cortex were nearly constant throughout the experiment when the fractured bones were compared with the non-fractured side. The effective atomic numbers of the elements in the experimental and control sides on days 15, 28, 42, and 90 were in the margins of experimental error, excluding day 42 (P 5 0.09, P 5 0.50, P 5 0.03, and P 5 0.41, respectively).
28 and 42. Radiology showed total union on day 90 (Fig. 3).
Histological findings Highly vascularized fibrous connective tissue was observed at the fracture site on day 15. Osteoblasts occurred in the neighborhood of compact bone. These osteoblasts were synthesizing the new bone matrix. Early-stage callus calcification was observed (Fig. 4). On day 28, newly formed bone trabeculae showed a higher amount of vascularized fibrous connective tissue. Osteoblasts and osteoclasts were active at the fracture site. They were located at the surfaces of the cortical bone. The activity of osteoblasts was morphologically defined by their cuboidal-to-columnar shape and basophilic cytoplasm. As their synthesizing activity declined, they flattened, and their cytoplasmic baso-
Radiological findings Radiological findings were in accordance with the histological findings and showed normal fracture healing. Periosteal callus was most apparent on days
Table 2. Callus density on days 15, 28, 42, and 90 Days
15
Sample Mean density, g/cm3 Standard deviation, g/cm3 Minimum density, g/cm3 Maximum density, g/cm3
28
42
90
E
C
E
C
E
C
E
C
2.57 0.17 2.31 2.97
2.33 0.10 2.07 3.62
2.46 0.16 2.12 2.85
2.26 0.14 1.09 2.56
2.07 0.11 1.34 2.94
2.34 0.13 1.60 3.19
2.25 0.20 1.59 4.63
2.37 0.15 1.72 2.95
E, Experiment; C, control
Days 15 28 42 90
15
28
42
90
1 0.378 0.281 0.935
1 0.181 0.419
1 0.331
1
Table 3. Cortical atomic number on days 15, 28, 42, and 90 Days
15
Sample
E
28 C
E
42 C
E
Mean atomic number 15.58 14.51 14 14.45 13.80 Standard deviation 1.59 1.61 1.3 1.2 1.10 Minimum atomic number 11.79 9.87 6.74 12.10 10.88 Maximum atomic number 20.19 14.14 15.80 17.00 15.81 E, Experiment; C, control
Days 15 28 42 90
15 1 0.061 1.04E-05 3.91E-05
28 1 0.014 0.032
42
90
1 0.696
1
90 C
E
C
14.47 0.99 12.47 18.72
13.91 1.19 12.41 16.79
14.13 0.87 12.66 16.84
252
F. Korkusuz et al.: Fracture callus assessment by QCT
Table 4. Callus atomic numbers on days 15, 28, 42, and 90 Days
15
Sample
E
Mean atomic number Standard deviation Minimum atomic number Maximum atomic number
28 C
E
42 C
E
15.58 12.81 14.41 14.73 13.80 1.59 0.07 0.16 0.14 1.10 11.79 1.87 6.74 7.07 10.80 24.09 27.36 20.80 18.34 18.56
90 C
14.47 0.99 12.47 20.65
E
C
13.60 13.90 1.19 0.71 10.42 12.64 17.54 16.84
E, Experiment; C, control
Days 15 28 42 90
15
28
42
90
1 0.001 0.091 0.114
1 0.055 0.031
1 0.864
1
gation. Mineralization of the callus is one of the major and crucial steps in this event. The mineralization process in a healing fracture occurs either at the cortical ends (primary mineralization) or in the fracture hematoma (secondary mineralization). Mechanical stability improves as mineralization proceeds. Advanced methods, such as ultrasound,23,26 vibration analysis,7 and improved mechanical tests,7,11,19,24 have been used to assess the mechanical stability of the fracture site more accurately. Nevertheless, none of the above-mentioned methods have proved to be effective in the establishment of a relation between callus mineralization and mechanical stability.2,6,8,19,20,22,24 Conventional radiography, which is the most common tool in clinical practice, is particularly unreliable for the quantitative assessment of bone fracture healing. An b alternative method for the evaluation of callus a mineralization may be dual X-ray absorptiometry Fig. 3a,b. Radiological findings on a day 28 and b day 90. (DEXA); however, its sensitivity and specificity is still Periosteal callus was most apparent on day 28. Total union under investigation.16,19 occurred on day 90. Arrows show the fracture lines The aim of this study was to establish a relation between callus mineralization and quantification of fracture healing in the long bones of rabbits. QCT data philia declined. Ossification at the fracture site was for the fracture site were obtained for this reason, and apparent by the propagation of the trabeculae on day the results were compared with those of conventional 42 (Fig. 5). Histologically, ossification was almost radiography and histology. Further, the atomic numbers completed on day 90. The periosteal findings for healing of elements (ie, calcium and phosphate) present in the were in accordance with those for the bone. Periosteal fracture callus were measured to allow more accurate loose connective tissue thickened and became denser understanding of the mineralization process. Both the and more cellular, starting from day 15. The cells of the cortical ends of the fractured bone and the mineral cambium layer of the periosteum transformed into content of the periosteal callus were assessed by QCT. bone-forming cells, namely, osteoblasts (Figs. 6 and 7). Callus mineralization was most apparent in the early stages of fracture healing, and was detected only by Discussion QCT, but not with conventional radiography. The size and density of the periosteal callus decreased with time Bone fracture healing is known to be affected by and this callus was almost undetectable at the end of the various factors which are still under extensive investiexperiment. The mean density of the periosteal callus
F. Korkusuz et al.: Fracture callus assessment by QCT
253
Fig. 4. A Fracture site at lower magnification. B Higher magnification of area marked with an asterix in A, showing the junction of cortical (C) and young spongy (S) bone at the fracture site on day 15. A H&E, 34; B H&E, 310
Fig. 5. A New bone formation (short arrow) is apparent with premature Haversian systems (arrowheads) on day 42. TB, Trabecular bone. B Active osteoblasts are observed on higher
magnification at the location indicated by the arrow in A. A H&E, 34; B H&E, 310
was significantly higher on day 15 than on days 28, 42, and 90. The size of the periosteal callus had decreased significantly on day 42 and was almost undetectable on day 90. The cortical density of the involved area, on the other hand, showed a significant decrease on days 28 and 42. This may be explained by the osteoclastic resorption of the cortex on those days. Calcium and
phosphate removed from the cortex most probably accumulated in the fracture hematoma and this accelerated the early calcification of the periosteal callus. Compared with the QCT findings, conventional X-ray findings, however, were unable to detect this early cortical resorption and mineral accumulation in the callus. X-rays showed apparent callus formation on
254
Fig. 6. A thin layer of bone-forming cells (arrows) lying on loose connective tissue (LCT) is visible on day 15. H&E, 310
Fig. 7. The periosteum overlying the fracture site is thickened and highly cellular, and rather dense connective tissue (asterisks) is observed on day 28. H&E, 310
day 28, which was contrary to the QCT findings. The possible effects of intramedullary fixation (ie, suppression of intramedullary healing and mineralization) in this experiment should be kept in mind when these results are interpreted. The fracture line was still present on day 90 on three-dimensional images of QCT (data not presented), while conventional X-rays showed almost normal healing. These findings indicate that findings on conventional X-ray should always be assessed with caution, and only estimates of the results of the mineralization process should be made. Periosteal callus mineralization, as quantified by QCT, represents the early stage of fracture healing and its mechanism is still under investigation. Although two hypotheses have been put forward to explain callus
F. Korkusuz et al.: Fracture callus assessment by QCT
mineralization,9 the constant mineral components in the fracture area, as defined by atomic number measurements in the present study, supports the hypothesis of Glimcher.14 There is a milieu of minerals, mainly calcium and phosphates, in the fracture area, most probably removed from the cortical ends of the fractured bone, and mineralizaton occurs very early in the hematoma by the initiation of release of vesicular calcium from the cytoplasm of osteoblasts. Atomic numbers of elements in the cortex and the callus may also largely depend on the inorganic matrix of bone found in that area. Collagen orientation and its effect on hydroxyapatite accumulation are under investigation. One advantage of QCT is its effectivity in regard to elemental atomic number measurements. Although a descriptive element cannot be defined as the sum of the elements found in that area, further studies may determine the relation of atomic numbers and tissue integrity more precisely. The variance in density measurements on days 28, 42, and 90 should be interpreted by further experiments. It can be speculated that the homeostatic mechanism of the bone is active in fracture healing, and the elements found in the fracture area are possibly removed from the cortex by osteoclasts. Then they are recycled in the mineralization process of the callus, by an as yet unknown mechanism. Variables other than callus mineralization, ie collagen orientation, may create a more complicated situation for the quantitative assessment of fracture healing.4,5,14,17,21,25 Histologically, osteoblasts and osteoclasts were active during the period of fracture healing. Their morphology and basophilic staining during the early phase of healing defined the activity of the osteoblasts. The existence of early-ossifying highly vascularized connective tissue close to the compact bone on days 15 and 28 is in accordance with the QCT findings. The atomic numbers and the histological findings of bone fracture healing were in accordance. The average atomic number depends on the tissue type and integrity of the callus. Recent QCT literature, mainly based on osteoporosis studies, defines its effectiveness and sensitivity in density measurements.10 As predicted by Gasser,12 this method is also sensitive in small animals; therefore, the results of that study can be compared with ours. QCT can assess the changes in bone mass and architecture. Although Ferretti10 defines a formula to assess the mechanical quality and spatial distribution of the hard tissue from QCT data, we made an attempt to quantify the periosteal callus and cortical mineral density by QCT and compared these findings with those of conventional radiography and histology. It can be concluded that QCT is an effective tool for the assessment of callus mineralization. Highresolution, three-dimensional imaging and a volumetric
F. Korkusuz et al.: Fracture callus assessment by QCT
approach are the advantages of QCT.18 However, as callus mineralization occurs at an early stage of fracture healing, and as other variables, such as collagen orientation, may play a role in stability, callus mineralization should not be used as the definitive tool of quantitative evaluation. Although a correlation between bone mineral density and vibration transmission could not be established in a recent study,1 biomechanical methods seem to have the potential for use in the monitoring of fracture healing. QCT, on the other hand, is superior to conventional radiography, and, being non-invasive, it may be used in patients for whom immobilization for fracture healing may have to be terminated. Acknowledgments. Quantitative computerized tomographic analysis was performed at the Petroleum Research Laboratory (PAL) of the Department of Petroleum and Natural Gas Engineering, Middle East Technical University. Valuable contributions during the evaluation of histological sections were obtained from Professor Dr. Nur Çakar, Department of Histology and Embryology, Hacettepe University Medical School. Technical assistance for the X-ray pictures was received from Ahmet Yürekli and Haydar Yag˘cı, Medical Center, Middle East Technical University, Ankara, Turkey.
References 1. Akkus¸ O, Korkusuz F, Akın S, et al. Relation between mechanical stiffness and vibration transmission for fracture callus: an experimental study on rabbit tibia. Proc Instn Mech Engrs 1998;212:327–36. 2. Aro HT, Wippermann BW, Hodgson SF, et al. Prediction of properties of fracture callus by measurement of mineral density using micro-bone densitometry. J Bone Joint Surg Am 1989;71: 1020–30. 3. Brooks RA, Di Chiro G. Principles of computer assisted tomography (CAT) in radiographic and radioisotopic imaging. Phys Med Biol 1976;21:689–732. 4. Buckwalter JA, Glimcher MJ, Cooper RR, et al. Bone biology (part I): structure, blood supply, cells, matrix, and mineralization. J Bone Joint Surg Am 1995;77:1256–75. 5. Buckwalter JA, Glimcher MJ, Cooper RR, et al. Bone biology (part II): formation, form, modeling, remodeling, and regulation of cell function. J Bone Joint Surg Am 1995;77:1276–89. 6. Chakkalakal DA, Lippiello L, Wilson R, et al. Mineral and matrix contributions to rigidity in fracture healing. J Biomech 1990;23: 425–34. 7. Cunningham JL, Kenwright J, Kershaw CJ. Biomechanical measurement of fracture healing. J Med Eng Technol 1990;14:92– 101.
255 8. Dimarogonas AD, Abbasi-Jahromi SH, Avioli LV. Material damping for monitoring of density and strength of bones. Calcif Tissue Int 1993;52:244–7. 9. Dunham J. Metabolism of the fracture callus. In: Hall BK, editor. Bone (vol. 5). Boca Raton, Florida: CRC Press; 1992. p. 1–31. 10. Ferretti JL. Noninvasive assessment of bone architecture and biomechanical properties in animals and humans employing pQCT technology. J Jpn Soc Bone Morphom 1997;7:115–25. 11. Foux A, Uhthoff HK, Black RC. Healing of plated femoral osteotomies in dogs. A mechanical study using a new test method. Acta Orthop Scand 1993;64:345–53. 12. Gasser JA. Quantitative assessment of bone mass and geometry by pQCT in rats: in vivo and site specificity of changes at different skeletal sites. J Jpn Soc Bone Morphom 1997;7:107–14. 13. Genant HK, Engelke K, Fuerst T, et al. Noninvasive assessment of bone mineral and structure: state of the art. J Bone Min Res 1996;11:707–30. 14. Glimcher MJ. The nature of the mineral component of bone and the mechanism of calcification. In: Griffin PP, editor. Instructional course lectures 36. Park Ridge: American Academy of Orthopaedic Surgeons, 1987. p. 49–68. 15. Kapıcıog˘lu MIS, Tüzünalp Ö, Arslan A, et al. Sonospectrographic assessment of fracture healing. Proceedings of the 1994 National Biomedical Engineering Days. Istanbul: Bog˘aziçi University Press, 1994. p. 175–7. . 16. Korkusuz F, Karamete K, Irfanog˘lu B, et al. Do porous calcium hydroxyapatite ceramics cause porosis in bone? A bone densitometry and biomechanical study on cortical bones of rabbits. Biomaterials 1995;16:537–43. 17. Lane JM, Werntz JR. Biology of fracture healing. In: Lane JM, editor. Fracture healing. New York: Churchill Livingstone; 1987. p. 49–60. 18. Lang T, Augat P, Majumdar S, et al. Noninvasive assessment of bone density and structure using computed tomography and magnetic resonance. Bone 1998;22:149S–53S. 19. Markel MD, Wikenheiser MA, Morin RL, et al. The determination of bone fracture properties by dual-energy X-ray absorptiometry and single-photon absorptiometry: a comparative study. Calcif Tissue Int 1991;48:392–9. 20. Markel MD, Chao EYS. Noninvasive monitoring techniques for quantitative description of callus mineral content and mechanical properties. Clin Orthop 1993;293:37–45. 21. Mundy GR. Bone remodeling and its disorders. London: Martin Dunitz; 1995. 22. Powell ES, Lawford PV, Duckworth T, et al. Is callus calcium content an indicator of the mechanical strength of healing fractures? An experimental study in rat metatarsals. J Biomed Eng 1989;11:277–81. 23. Ricciardi L, Perissinotto A, Dabala M. Mechanical monitoring of fracture healing using ultrasound imaging. Clin Orthop 1993;293: 71–6. 24. Richardson JB, Cunningham JL, Goodship AE, et al. Measuring stiffness can define healing of tibial fractures. J Bone Joint Surg Br 1994;76:389–94. 25. Schenk RK, Huinziker EB. Histological and ultrastructural features of fracture healing. In: Brighton CT, Friedlaender GE, Lane JM, editors. Bone formation and repair. Park Ridge: American Academy of Orthopaedic Surgeons; 1994. p. 117–46. 26. Yeni YN, Günel U, Korkusuz F, et al. Ultrasonic properties of human bones in senile osteoporosis. Acta Orthop Traumatol Turc 1995;29:294–8.
Assessment of mineral density and atomic content of fracture callus by quantitative computerized tomography Feza Korkusuz1, Serhat Akin2, Ozan Akkus¸ 3, and Petek Korkusuz4 . Medical Center, Middle East Technical University, Inönü Bulvarı, 06531 Ankara, Turkey Petroleum and Natural Gas Engineering Department, Middle East Technical University, Faculty of Engineering, 06531 Ankara, Turkey 3 Department of Mechanical and Aerospace Engineering, Orthopedic Biomechanics Laboratories, Case Western Reserve University, 10900 Euclid Avenue, Cleveland, OH 44106, USA 4 Department of Histology and Embryology, Hacettepe University, School of Medicine, Sıhhiye, 06100 Ankara, Turkey 1 2
Abstract: The mineral density and atomic numbers of elements in the periosteal callus and the cortex area of a healing fracture were measured by quantitative computerized tomography (QCT) to obtain accurate information on the mineralization process in rabbit tibia. The mineral density of the periosteal callus was highest on day 15 and decreased gradually throughout the experiment. This was initially detected by QCT, but not with conventional radiography. An apparent decrease in cortical bone density on days 28 and 42 after fracture was observed. Atomic numbers of elements in the cortex remained stable, indicating a possible homeostatic mechanism of mineral preservation at the fracture callus and involved cortical area. QCT may predict density alterations in the fracture callus more accurately than conventional X-ray. Further studies are essential to predict a relationship between mineral density, atomic numbers, and mechanical stability. Key words: bone healing, fracture callus, mineral density, atomic number, quantitative computerized tomography
Introduction Accurate assessment of fracture healing of long bones will obviously improve the outcome of surgical and nonsurgical treatment. Early cessation of immobilization or internal fixation is one of the most common reasons for malunion and non-union. Joint stiffness, on the other hand, is another major problem, caused by excessive immobilization of the adjacent joints in nonsurgical treatment. Precise knowledge of callus mineralization and organization and its impact on mechanical stability is essential to prevent these complications in patients.
Offprint requests to: F. Korkusuz Received for publication on March 8, 1999; accepted on Oct. 5, 1999
Various biological and mechanical factors play important roles in the mineralization process of the fracture callus.4,5,17,25 This process is still poorly understood, in spite of extensive ultrastructural and histochemical investigations. Adequate concentrations of calcium and phosphate ions, the presence of a calcifiable matrix, a nucleating agent, and control of regulators is required for mineralization. Callus mineralization and its relation to mechanical stability is under extensive investigation.14 Although various studies have attempted to quantify parameters of the fracture site,11,13,15,20,24,26 only a few1,8,10,22 were able to demonstrate a positive relation between the mineral content of the callus and the mechanical properties of the fractured bone. The aim of this study was to quantify the mineralization of the callus by quantitative computerized tomography (QCT) and to correlate the results with those of conventional radiography and histology. We also assessed atomic numbers of elements in the fracture callus that may be used to quantify tissue integrity. We investigated the relation between atomic number, mineral density, and histological changes.
Materials and methods Animals and surgical protocol A total number of 24 three-month-old local albino male rabbits, weighing 1500 6 150 g, were anesthetized with intramuscular ketamine hydrochloride and xylazine hydrochloride (100 mg cm23: 23.32 mg cm23) injection. The back limbs of the rabbits were shaved and cleansed with Betadine solution (Kim-Pa, Istanbal, Turkey). A mid-diaphyseal transverse tibial osteotomy at the level of the fibular intersection simulated an open fracture. The periosteum was well preserved and reattached after the osteotomy. Stability was provided by intra-
F. Korkusuz et al.: Fracture callus assessment by QCT
249
medullary K-wire fixation. The tip of the Kirschner wire was specially designed for this procedure; the proximal tip was a triangular pyramidal shape to allow easy passage through the proximal joint line with minimal damage to the cartilage. The distal tip was a conical shape to prevent a secondary fracture in the distal tibial component. A torque-controlled electric powered drill was used to minimize thermal damage and secondary fractures. The non-fractured left tibia served as the control. The animals were killed on days 15, 28, 42, and 90 by hypovolemia. The fractured and control tibiae were removed and preserved at 220°C until the day of assessment. Six bones were assessed for each time interval. The bones were thawed to room temperature and transferred into petri dishes containing saline for QCT. Scanning of each bone lasted approximately 7 to 8 min. The bones were then transferred to Falcon tubes (Becton, NJ, USA) containing 10% formalin and fixed for histological evaluation. Conventional X-ray pictures were obtained at the time of transfer into the fixative. The Ethics Committee of Gazi University Medical School approved all animal procedures. Procedures were in full agreement with Turkish law 6343/2, Veterinary Medicine Deontology Regulation 6.7.26, and with the Helsinki Declaration of animal rights. Density and atomic number measurements Fractured and control bones were assessed with a Philips Tomoscan 60/TX third generation scanner (Philips, NY, USA). The X-ray source of the computerized tomography (CT) permitted the emission of pulse X-rays. Opposite this X-ray source was an x-ray detector in a xenon ionization chamber to measure the distribution of the X-ray intensity. The detector has 576 channels in total, 4 of which at each end are used as reference channels. Scan time changed between 1.9 and 9.0 s, with which 80- to 130-kV energy could be utilized for up to 700-mA tube current. Absorption coefficient distribution was obtained for each pixel in a matrix of 320 3 320 or 512 3 512. The CT number, Ni, was then calculated from the corresponding linear attenuation coefficients, using the following equation:3
Ni =
1000(m i - m w ) mw
Ze =
1 3.8
(Â f Z ) i
3.8 i
(3)
where fi is the fraction of electrons on the ith atomic number species. In order to obtain the density and atomic number, eight to ten images of each bone were obtained at 5-mm intervals in the proximal-to-distal direction. Scans were conducted in 4.5 s, in a field of view of 160 mm with a slice thickness of 2 mm, at 100 and 130 kV. The density and atomic values for each pixel were obtained by solving equation (2). Mineral density and atomic numbers of the fracture callus were randomly measured in at least ten different locations, and the average was defined as the callus density (Fig. 1). Measurements on days 42 and 90 were made in only six locations, as the size of the callus decreased gradually. The cortical density of each bone was measured separately (Fig. 2). The mineral density of the cortex was measured by drawing four to six lines on each transverse section. Cortical mineral density and atomic numbers were measured on each pixel of each of these lines. The location of the line was chosen in such a way that the effect of beam hardening was minimal. Beam
(1)
The linear attenuation coefficient, µ, depends on both electron density (bulk density), r, and the atomic number, Z, in the form: m = r(a + bZ 3.8 E 3.2 )
where a is a nearly energy independent coefficient and b is a constant. The first term in equation (2) represents Compton scattering, which is predominant at X-ray energies above 100 kV at which medical CT scanners normally operate. The second term accounts for photoelectric absorption. One image representing bulk density and another image proportional to the atomic number could be obtained by scanning at high and low X-ray energies (ie, 130 and 100 kV) and by solving equation (2) on a pixel-by-pixel basis. It should be noted that when a mixture of atomic species is present, photoelectric absorption is proportional to the effective atomic number, Ze, which is given by the following equation:
(2)
Fig. 1. Density of the periosteal callus was repeatedly measured in ten different locations in the quantitative computerized tomography sections (schematic drawing of measurements on day 15)
250
F. Korkusuz et al.: Fracture callus assessment by QCT
hardening is due to the polychromatic nature of the X-ray beams passing through the object. As the beam passes through the material, the lower energy photons are absorbed preferentially. As the remaining beams become more monochromatic at higher energy, the beam becomes harder. This causes the overall attenuation to change with distance into the object. Calculations and image analysis were performed on a SUN Sparc-10 workstation (SUN Microsystems, Palo Alto, CA, USA).
distance of the X-ray source to the bones was 100 cm and the tube current was 3.2 mA per s; 42-Kwp energy was used to obtain the images. Films were developed using an Agfa Crurix 60 automatic developing machine. Each bone was exposed in the anterior-posterior and the lateral directions. Histological analysis All specimens were decalcified in 10% formic acid solution for 1 to 2 weeks; the periosteum was stripped off and evaluated separately. Specimens were dehydrated in a graded series of ethanol and embedded in paraffin and then 5-µm thin longitudinal sections were obtained and stained with hematoxylin and eosin (H&E). Photomicrographs were obtained with an Olympus BH7 light microscope (Olympus, Tokyo, Japan). Two-tailed heteroscedastic t-test was performed for statistical analysis.
Radiographic evaluation Conventional X-rays were obtained on Agfa Crurix films (Agfa, Mortsel, Belgium) with a Siemens Multix C X-ray machine (Siemens, Erlangen, Germany). The
Results Density and atomic number measurements Cortical density measurements in the region of interest of the bone sample on days 15, 28, 42, and 90 are presented in Table 1. The density of the fractured tibiae decreased significantly on days 28, 42, and 90 compared with that on day 15 (P , 0.001, P , 0.001, and P , 0.01, respectively). The cortical density of the non-fractured left tibiae (control values in Table 1) remained constant and was always slightly higher than that of the fractured side. Cortical density changed significantly in the fractured bone on days 28 and 42 when compared with values in the non-fractured control side (P , 0.001 and P , 0.001, respectively) and recovered to normal on day
Fig. 2. Cortical bone mineral density measurement. The mineral density of the fractured tibiae and non-fractured control side were measured on each pixel on the lines defined in the cortex
Table 1. Cortical density on days 15, 28, 42, and 90 Days
15
Sample Mean density, g/cm3 Standard deviation, g/cm3 Minimum density, g/cm3 Maximum density, g/cm3
28
42
E
C
E
C
E
C
E
C
2.57 0.17 1.16 1.75
2.56 0.16 1.15 1.57
2.00 0.11 1.30 1.76
2.25 0.15 1.09 1.48
2.07 0.11 1.04 1.42
2.34 0.13 1.11 1.66
2.31 0.20 1.03 1.45
2.41 0.14 1.10 1.54
E, Experiment; C, control
Days 15 28 42 90
90
15
28
42
90
1 1.71E-27 5.53E-18 0.01
1 5.59E-12 1.01E-09
1 0.02
1
F. Korkusuz et al.: Fracture callus assessment by QCT
251
90 (P 5 0.39). Periosteal callus formation was most apparent on day 15. The mean density of the callus on day 15 was significantly higher than that on days 28, 42, and 90 (Table 2). The effective atomic numbers of the elements in the cortical and callus regions are shown in Tables 3 and 4. The effective atomic numbers of the elements in the cortex were nearly constant throughout the experiment when the fractured bones were compared with the non-fractured side. The effective atomic numbers of the elements in the experimental and control sides on days 15, 28, 42, and 90 were in the margins of experimental error, excluding day 42 (P 5 0.09, P 5 0.50, P 5 0.03, and P 5 0.41, respectively).
28 and 42. Radiology showed total union on day 90 (Fig. 3).
Histological findings Highly vascularized fibrous connective tissue was observed at the fracture site on day 15. Osteoblasts occurred in the neighborhood of compact bone. These osteoblasts were synthesizing the new bone matrix. Early-stage callus calcification was observed (Fig. 4). On day 28, newly formed bone trabeculae showed a higher amount of vascularized fibrous connective tissue. Osteoblasts and osteoclasts were active at the fracture site. They were located at the surfaces of the cortical bone. The activity of osteoblasts was morphologically defined by their cuboidal-to-columnar shape and basophilic cytoplasm. As their synthesizing activity declined, they flattened, and their cytoplasmic baso-
Radiological findings Radiological findings were in accordance with the histological findings and showed normal fracture healing. Periosteal callus was most apparent on days
Table 2. Callus density on days 15, 28, 42, and 90 Days
15
Sample Mean density, g/cm3 Standard deviation, g/cm3 Minimum density, g/cm3 Maximum density, g/cm3
28
42
90
E
C
E
C
E
C
E
C
2.57 0.17 2.31 2.97
2.33 0.10 2.07 3.62
2.46 0.16 2.12 2.85
2.26 0.14 1.09 2.56
2.07 0.11 1.34 2.94
2.34 0.13 1.60 3.19
2.25 0.20 1.59 4.63
2.37 0.15 1.72 2.95
E, Experiment; C, control
Days 15 28 42 90
15
28
42
90
1 0.378 0.281 0.935
1 0.181 0.419
1 0.331
1
Table 3. Cortical atomic number on days 15, 28, 42, and 90 Days
15
Sample
E
28 C
E
42 C
E
Mean atomic number 15.58 14.51 14 14.45 13.80 Standard deviation 1.59 1.61 1.3 1.2 1.10 Minimum atomic number 11.79 9.87 6.74 12.10 10.88 Maximum atomic number 20.19 14.14 15.80 17.00 15.81 E, Experiment; C, control
Days 15 28 42 90
15 1 0.061 1.04E-05 3.91E-05
28 1 0.014 0.032
42
90
1 0.696
1
90 C
E
C
14.47 0.99 12.47 18.72
13.91 1.19 12.41 16.79
14.13 0.87 12.66 16.84
252
F. Korkusuz et al.: Fracture callus assessment by QCT
Table 4. Callus atomic numbers on days 15, 28, 42, and 90 Days
15
Sample
E
Mean atomic number Standard deviation Minimum atomic number Maximum atomic number
28 C
E
42 C
E
15.58 12.81 14.41 14.73 13.80 1.59 0.07 0.16 0.14 1.10 11.79 1.87 6.74 7.07 10.80 24.09 27.36 20.80 18.34 18.56
90 C
14.47 0.99 12.47 20.65
E
C
13.60 13.90 1.19 0.71 10.42 12.64 17.54 16.84
E, Experiment; C, control
Days 15 28 42 90
15
28
42
90
1 0.001 0.091 0.114
1 0.055 0.031
1 0.864
1
gation. Mineralization of the callus is one of the major and crucial steps in this event. The mineralization process in a healing fracture occurs either at the cortical ends (primary mineralization) or in the fracture hematoma (secondary mineralization). Mechanical stability improves as mineralization proceeds. Advanced methods, such as ultrasound,23,26 vibration analysis,7 and improved mechanical tests,7,11,19,24 have been used to assess the mechanical stability of the fracture site more accurately. Nevertheless, none of the above-mentioned methods have proved to be effective in the establishment of a relation between callus mineralization and mechanical stability.2,6,8,19,20,22,24 Conventional radiography, which is the most common tool in clinical practice, is particularly unreliable for the quantitative assessment of bone fracture healing. An b alternative method for the evaluation of callus a mineralization may be dual X-ray absorptiometry Fig. 3a,b. Radiological findings on a day 28 and b day 90. (DEXA); however, its sensitivity and specificity is still Periosteal callus was most apparent on day 28. Total union under investigation.16,19 occurred on day 90. Arrows show the fracture lines The aim of this study was to establish a relation between callus mineralization and quantification of fracture healing in the long bones of rabbits. QCT data philia declined. Ossification at the fracture site was for the fracture site were obtained for this reason, and apparent by the propagation of the trabeculae on day the results were compared with those of conventional 42 (Fig. 5). Histologically, ossification was almost radiography and histology. Further, the atomic numbers completed on day 90. The periosteal findings for healing of elements (ie, calcium and phosphate) present in the were in accordance with those for the bone. Periosteal fracture callus were measured to allow more accurate loose connective tissue thickened and became denser understanding of the mineralization process. Both the and more cellular, starting from day 15. The cells of the cortical ends of the fractured bone and the mineral cambium layer of the periosteum transformed into content of the periosteal callus were assessed by QCT. bone-forming cells, namely, osteoblasts (Figs. 6 and 7). Callus mineralization was most apparent in the early stages of fracture healing, and was detected only by Discussion QCT, but not with conventional radiography. The size and density of the periosteal callus decreased with time Bone fracture healing is known to be affected by and this callus was almost undetectable at the end of the various factors which are still under extensive investiexperiment. The mean density of the periosteal callus
F. Korkusuz et al.: Fracture callus assessment by QCT
253
Fig. 4. A Fracture site at lower magnification. B Higher magnification of area marked with an asterix in A, showing the junction of cortical (C) and young spongy (S) bone at the fracture site on day 15. A H&E, 34; B H&E, 310
Fig. 5. A New bone formation (short arrow) is apparent with premature Haversian systems (arrowheads) on day 42. TB, Trabecular bone. B Active osteoblasts are observed on higher
magnification at the location indicated by the arrow in A. A H&E, 34; B H&E, 310
was significantly higher on day 15 than on days 28, 42, and 90. The size of the periosteal callus had decreased significantly on day 42 and was almost undetectable on day 90. The cortical density of the involved area, on the other hand, showed a significant decrease on days 28 and 42. This may be explained by the osteoclastic resorption of the cortex on those days. Calcium and
phosphate removed from the cortex most probably accumulated in the fracture hematoma and this accelerated the early calcification of the periosteal callus. Compared with the QCT findings, conventional X-ray findings, however, were unable to detect this early cortical resorption and mineral accumulation in the callus. X-rays showed apparent callus formation on
254
Fig. 6. A thin layer of bone-forming cells (arrows) lying on loose connective tissue (LCT) is visible on day 15. H&E, 310
Fig. 7. The periosteum overlying the fracture site is thickened and highly cellular, and rather dense connective tissue (asterisks) is observed on day 28. H&E, 310
day 28, which was contrary to the QCT findings. The possible effects of intramedullary fixation (ie, suppression of intramedullary healing and mineralization) in this experiment should be kept in mind when these results are interpreted. The fracture line was still present on day 90 on three-dimensional images of QCT (data not presented), while conventional X-rays showed almost normal healing. These findings indicate that findings on conventional X-ray should always be assessed with caution, and only estimates of the results of the mineralization process should be made. Periosteal callus mineralization, as quantified by QCT, represents the early stage of fracture healing and its mechanism is still under investigation. Although two hypotheses have been put forward to explain callus
F. Korkusuz et al.: Fracture callus assessment by QCT
mineralization,9 the constant mineral components in the fracture area, as defined by atomic number measurements in the present study, supports the hypothesis of Glimcher.14 There is a milieu of minerals, mainly calcium and phosphates, in the fracture area, most probably removed from the cortical ends of the fractured bone, and mineralizaton occurs very early in the hematoma by the initiation of release of vesicular calcium from the cytoplasm of osteoblasts. Atomic numbers of elements in the cortex and the callus may also largely depend on the inorganic matrix of bone found in that area. Collagen orientation and its effect on hydroxyapatite accumulation are under investigation. One advantage of QCT is its effectivity in regard to elemental atomic number measurements. Although a descriptive element cannot be defined as the sum of the elements found in that area, further studies may determine the relation of atomic numbers and tissue integrity more precisely. The variance in density measurements on days 28, 42, and 90 should be interpreted by further experiments. It can be speculated that the homeostatic mechanism of the bone is active in fracture healing, and the elements found in the fracture area are possibly removed from the cortex by osteoclasts. Then they are recycled in the mineralization process of the callus, by an as yet unknown mechanism. Variables other than callus mineralization, ie collagen orientation, may create a more complicated situation for the quantitative assessment of fracture healing.4,5,14,17,21,25 Histologically, osteoblasts and osteoclasts were active during the period of fracture healing. Their morphology and basophilic staining during the early phase of healing defined the activity of the osteoblasts. The existence of early-ossifying highly vascularized connective tissue close to the compact bone on days 15 and 28 is in accordance with the QCT findings. The atomic numbers and the histological findings of bone fracture healing were in accordance. The average atomic number depends on the tissue type and integrity of the callus. Recent QCT literature, mainly based on osteoporosis studies, defines its effectiveness and sensitivity in density measurements.10 As predicted by Gasser,12 this method is also sensitive in small animals; therefore, the results of that study can be compared with ours. QCT can assess the changes in bone mass and architecture. Although Ferretti10 defines a formula to assess the mechanical quality and spatial distribution of the hard tissue from QCT data, we made an attempt to quantify the periosteal callus and cortical mineral density by QCT and compared these findings with those of conventional radiography and histology. It can be concluded that QCT is an effective tool for the assessment of callus mineralization. Highresolution, three-dimensional imaging and a volumetric
F. Korkusuz et al.: Fracture callus assessment by QCT
approach are the advantages of QCT.18 However, as callus mineralization occurs at an early stage of fracture healing, and as other variables, such as collagen orientation, may play a role in stability, callus mineralization should not be used as the definitive tool of quantitative evaluation. Although a correlation between bone mineral density and vibration transmission could not be established in a recent study,1 biomechanical methods seem to have the potential for use in the monitoring of fracture healing. QCT, on the other hand, is superior to conventional radiography, and, being non-invasive, it may be used in patients for whom immobilization for fracture healing may have to be terminated. Acknowledgments. Quantitative computerized tomographic analysis was performed at the Petroleum Research Laboratory (PAL) of the Department of Petroleum and Natural Gas Engineering, Middle East Technical University. Valuable contributions during the evaluation of histological sections were obtained from Professor Dr. Nur Çakar, Department of Histology and Embryology, Hacettepe University Medical School. Technical assistance for the X-ray pictures was received from Ahmet Yürekli and Haydar Yag˘cı, Medical Center, Middle East Technical University, Ankara, Turkey.
References 1. Akkus¸ O, Korkusuz F, Akın S, et al. Relation between mechanical stiffness and vibration transmission for fracture callus: an experimental study on rabbit tibia. Proc Instn Mech Engrs 1998;212:327–36. 2. Aro HT, Wippermann BW, Hodgson SF, et al. Prediction of properties of fracture callus by measurement of mineral density using micro-bone densitometry. J Bone Joint Surg Am 1989;71: 1020–30. 3. Brooks RA, Di Chiro G. Principles of computer assisted tomography (CAT) in radiographic and radioisotopic imaging. Phys Med Biol 1976;21:689–732. 4. Buckwalter JA, Glimcher MJ, Cooper RR, et al. Bone biology (part I): structure, blood supply, cells, matrix, and mineralization. J Bone Joint Surg Am 1995;77:1256–75. 5. Buckwalter JA, Glimcher MJ, Cooper RR, et al. Bone biology (part II): formation, form, modeling, remodeling, and regulation of cell function. J Bone Joint Surg Am 1995;77:1276–89. 6. Chakkalakal DA, Lippiello L, Wilson R, et al. Mineral and matrix contributions to rigidity in fracture healing. J Biomech 1990;23: 425–34. 7. Cunningham JL, Kenwright J, Kershaw CJ. Biomechanical measurement of fracture healing. J Med Eng Technol 1990;14:92– 101.
255 8. Dimarogonas AD, Abbasi-Jahromi SH, Avioli LV. Material damping for monitoring of density and strength of bones. Calcif Tissue Int 1993;52:244–7. 9. Dunham J. Metabolism of the fracture callus. In: Hall BK, editor. Bone (vol. 5). Boca Raton, Florida: CRC Press; 1992. p. 1–31. 10. Ferretti JL. Noninvasive assessment of bone architecture and biomechanical properties in animals and humans employing pQCT technology. J Jpn Soc Bone Morphom 1997;7:115–25. 11. Foux A, Uhthoff HK, Black RC. Healing of plated femoral osteotomies in dogs. A mechanical study using a new test method. Acta Orthop Scand 1993;64:345–53. 12. Gasser JA. Quantitative assessment of bone mass and geometry by pQCT in rats: in vivo and site specificity of changes at different skeletal sites. J Jpn Soc Bone Morphom 1997;7:107–14. 13. Genant HK, Engelke K, Fuerst T, et al. Noninvasive assessment of bone mineral and structure: state of the art. J Bone Min Res 1996;11:707–30. 14. Glimcher MJ. The nature of the mineral component of bone and the mechanism of calcification. In: Griffin PP, editor. Instructional course lectures 36. Park Ridge: American Academy of Orthopaedic Surgeons, 1987. p. 49–68. 15. Kapıcıog˘lu MIS, Tüzünalp Ö, Arslan A, et al. Sonospectrographic assessment of fracture healing. Proceedings of the 1994 National Biomedical Engineering Days. Istanbul: Bog˘aziçi University Press, 1994. p. 175–7. . 16. Korkusuz F, Karamete K, Irfanog˘lu B, et al. Do porous calcium hydroxyapatite ceramics cause porosis in bone? A bone densitometry and biomechanical study on cortical bones of rabbits. Biomaterials 1995;16:537–43. 17. Lane JM, Werntz JR. Biology of fracture healing. In: Lane JM, editor. Fracture healing. New York: Churchill Livingstone; 1987. p. 49–60. 18. Lang T, Augat P, Majumdar S, et al. Noninvasive assessment of bone density and structure using computed tomography and magnetic resonance. Bone 1998;22:149S–53S. 19. Markel MD, Wikenheiser MA, Morin RL, et al. The determination of bone fracture properties by dual-energy X-ray absorptiometry and single-photon absorptiometry: a comparative study. Calcif Tissue Int 1991;48:392–9. 20. Markel MD, Chao EYS. Noninvasive monitoring techniques for quantitative description of callus mineral content and mechanical properties. Clin Orthop 1993;293:37–45. 21. Mundy GR. Bone remodeling and its disorders. London: Martin Dunitz; 1995. 22. Powell ES, Lawford PV, Duckworth T, et al. Is callus calcium content an indicator of the mechanical strength of healing fractures? An experimental study in rat metatarsals. J Biomed Eng 1989;11:277–81. 23. Ricciardi L, Perissinotto A, Dabala M. Mechanical monitoring of fracture healing using ultrasound imaging. Clin Orthop 1993;293: 71–6. 24. Richardson JB, Cunningham JL, Goodship AE, et al. Measuring stiffness can define healing of tibial fractures. J Bone Joint Surg Br 1994;76:389–94. 25. Schenk RK, Huinziker EB. Histological and ultrastructural features of fracture healing. In: Brighton CT, Friedlaender GE, Lane JM, editors. Bone formation and repair. Park Ridge: American Academy of Orthopaedic Surgeons; 1994. p. 117–46. 26. Yeni YN, Günel U, Korkusuz F, et al. Ultrasonic properties of human bones in senile osteoporosis. Acta Orthop Traumatol Turc 1995;29:294–8.