ELSEVIER
`
`0250-832X(95)00053-4
`
`Dentomaxillofac. Radiol., Vol. 25, No. 3, pp. 146-150, 1996
`Copyright © 1996 Elsevier Science Ltd for the IADMFR. All rights reserved
`Printed in Great Britain
`0250-832»% $15.00 + 0.00
`
`Quantitative computed tomography of
`trabecular bone in the mandible
`
`C. Lindh*, M. Nilsson*, B. Klinge* and A. Petersson*
`* Department of Oral Radiology, Centre for Oral Health Sciences, Lund University, Malmö, f Department of
`Radiation Physics, University Hospital, Malmö, and * Department of Periodontology, Karolinska Institutet,
`Stockholm, Sweden
`
`Received 5 May 1995 and in final form 11 October 1995
`
`Objective. To evaluate the potential use of quantitative computed tomography (QCT) for the
`assessment of bone mineral density of the edentulous mandible prior to implant placement.
`Methods. Ten 2 mm thick CT slices of anterior and posterior edentulous sections from 15
`mandibles were obtained perpendicular to the buccal and lingual plates. The bone mineral
`density, expressed as the amount of calcium hydroxyapatite (mg c m - 3 ) of the trabecular bone,
`was calculated using a method that takes into account the influence of fat.
`Results. The variation of bone mineral density between mandibles was high. Anterior sections
`showed higher values than posterior sections and a variation was found within sections of the
`same mandible.
`Conclusion. CT provides a site-related measure of the bone mineral density in the mandible
`and appears potentially useful as a non-invasive method to determine a parameter that may
`reflect bone quality prior to implant placement. Copyright © 1996 Elsevier Science Ltd for
`IADMFR.
`
`Keywords: Mandible; tomography, X-ray computed; osteoporosis; dental implantation
`
`Dentomaxillofac. Radiol., 1996; 25: 146-150
`
`Bone quality is referred to in many follow-up studies of
`implant treatment as a factor of great importance in the
`outcome 1 - 5. However, it is obvious that the term 'bone
`quality' is open to considerable interpretation and there
`is no real consensus in the literature6. Bone quality
`probably reflects a number of aspects of bone morph-
`ology of which the degree of mineralization may be one.
`The radiological methods that have been developed
`for determining the bone mineral density (BMD) are
`mainly used to study skeletal changes in osteoporosis
`and other metabolic bone diseases7. Densitometrie
`measurements have been made from intra-oral and
`panoramic radiographs of the mandible to find signs of
`osteoporosis8-10. Dual photon absorptiometry11'2 and
`dual energy X-ray absorptiometry13 have been used to
`measure BMD of the mandible and to correlate it with
`that of other parts of the skeleton. Both these methods
`measure an integrated sum of cortical and trabecular
`bone density. However, the consistency of the trabecu-
`lar bone is particularly significant in implant treatment,
`especially when the cortex is thin and the implant
`is inserted mainly
`in
`trabecular bone. Computed
`
`tomography (CT) is the only non-invasive pre-operative
`method where it is possible to obtain information on
`the degree of mineralization of trabecular bone as dis-
`tinct from the cortex. At present, it is mainly used pre-
`operatively to evaluate jaw bone volume and to make
`measurements of bone height and width14-17. Although
`it has been suggested by a number of workers in this field
`that CT scans could provide radiological densitometric
`readings of the bone tissue in Hounsfield units18-20,
`there are no previous reports on the measurement of
`BMD for the purpose of evaluation of bone quality in
`the jaws prior to implant treatment.
`Quantitative computed tomography (QCT) has the
`major advantage of enabling trabecular and cortical
`bone density to be evaluated separately. Bone mineral
`content has been assessed in the mandible11'12, but
`QCT of the trabecular bone density has been limited to
`correlating it with the degree of skeletal osteoporosis in
`postmenopausal women21. The aim of the present
`study was to evaluate the possible use of QCT for
`measuring the trabecular BMD in the edentulous man-
`dible prior to implant placement.
`
`Dentsply Sirona Inc. – Exhibit 1037
`
`

`

`QCT of the mandible
`
`147
`
`manually on each image as the trabecular bone 1-2 mm
`from the inner margin of the cortex (Figure 2, left).
`However, if there was a break in the inner cortex, the
`innermost part was included in the ROI (Figure 2,
`right). Every CT image was enlarged in order
`to
`increase the accuracy when the ROI was superimposed
`on the image using a graphic tablet. The enlargement
`procedure did not affect the quantitative data obtained.
`
`Calibration of the CT scanner for BMD determination
`The calibration procedure is described in detail by
`Nilsson et al.22. The CT scanner was calibrated for
`BMD determination using (i) BMD-simulating sub-
`stance samples placed in a phantom simulating the
`skull, and (ii) the same samples placed in free air
`(simulating the mandibles in this study). Due to the
`very hard beam filtration of the Somatom DRG unit
`(added filtration 2.5 mm Al + 0.4 mm Cu), the beam
`hardening in the lower part of the skull phantom is not
`very pronounced. The calibration equation did there-
`fore not significantly differ for these two situations and
`could be described by the equation:
`
`BMD (mg cm - 3) = 0.960 (HU m e a n + 14)
`- 2.3 x 10- 4 (HU m e a n + 14)2
`
`where HU m e a n is the mean Hounsfield unit (HU) value
`in the trabecular part of the mandible. It must be
`stressed that this calibration is valid only for this type of
`CT scanner, for a slice thickness of 2 mm and using a
`specific convolution kernel.
`The Somatom DRG scanner uses caesium iodide
`(Csl) scintillation detectors which are very sensitive to
`variations in temperature. Effects of drifting on the
`sensitivity of the detector, both as a whole and for
`individual detector elements, have to be corrected for
`at least every hour. This procedure, called 'air calibra-
`tion', is performed with no object in the gantry. In this
`
`right
`
`left
`
`A
`
`Outline of
`R O K
`
`Figure 2 Schematic drawings of CT images illustrating how the
`region of interest (ROI) was allocated between the cortical (A) and
`trabecular (B) bone. The margin between A and Β is well defined on
`the left but difficult to identify on the right
`
`Figure 1 Schematic drawing of a mandible illustrating how the CT
`scans were obtained perpendicular to the mandible in the anterior
`and posterior sections as indicated by the tantalum pins
`
`Material and methods
`
`Material
`The material consisted of 15 mandibles from indi-
`viduals aged 62-94 years (mean 79 years) who before
`death had elected to donate their bodies to medical
`research. There was no history of disease or treatment
`that might have altered bone metabolism. The whole
`body was fixed in formalin using a mortal perfusion
`technique. The mandibles were later removed, de-
`gloved and post-fixed in 10% neutralized buffered
`formalin solution. This method preserves both the
`structure and tissues, including fat. The mandibles were
`wholly or partially edentulous, and sections without
`teeth, six anterior and 27 posterior, were examined. To
`obtain reference points, two tantalum pins (1.5 x 0.5
`mm) were inserted at the buccal side of each edentu-
`lous section. In the posterior sections, one pin was
`inserted just beneath the mental foramen and the other
`2 cm distally. In the anterior sections, the pins were
`inserted 1 cm from the midline on each side.
`
`Production of CT images
`CT images were obtained with a Somatom DRG
`scanner (Siemens, Erlangen, Germany) at 125 kV and
`230 mA with a slice thickness of 2 mm using a standard
`convolution kernel. The mandibles were carefully posi-
`tioned with the scanning planes perpendicular to the
`buccal and lingual plates. Ten direct CT images were
`obtained of the section between the tantalum pins
`(Figure 1), except for one mandible (no. 6), where
`there were seven. The accuracy of slice thickness and
`table feed was checked with a weekly quality assurance
`programme. The region of interest (ROI) was defined
`
`

`

`148
`
`C. Lindh et al.
`
`study, air calibration was always carried out immediate-
`ly prior to scanning.
`
`Statistical analysis
`Differences in BMD between mandibles and different
`sections of the mandibles were investigated using
`analysis of variance with repeated measurements. A
`statistically significant difference was considered to be
`present when ρ
`0.05. The measurement of BMD was
`performed twice in 20 CT images. The second measure-
`ment was made at least one week after the first and the
`entire procedure with window-setting and enlargement
`was repeated. The precision of single measurements
`was expressed as the standard deviation SD = VZßP/2n,
`where d is the difference between two measurements
`and η is the number of duplicate measurements.
`The measurement precision was estimated as SD =
`17 mg c m - 3 .
`
`Results
`
`The results of the measurements of BMD are shown in
`Table I. Significant differences (p = 0.0001) in mean
`BMD were found between the 15 mandibles. Measure-
`ments were performed in both anterior and posterior
`sections in six mandibles, and, with the exception of
`one (no. 8), the differences were significant (p $ 0.05):
`values were higher in anterior than in posterior sec-
`tions, as shown in Figure 3. The variation between
`different slices within a section was also large, as can be
`seen from the maximum and minimum values in Table
`I. The mean BMD for the first five slices were com-
`pared with those of the last five in the 12 mandibles
`where measurements were made on both left and right
`posterior sections; those for the first five slices were
`higher in 16 of the 24 sections examined (Figures 4 and
`5). However, significant differences were found in only
`nine of these sections. An example of a CT image with
`a high BMD is shown in Figure 6 (left) and one with
`low BMD is shown in Figure 6 (right).
`
`Table I Mean and standard deviation (sd) of BMD (amount of
`calcium hydroxyapatite; mg cm - 3) in 33 sections of 15 mandibles.
`Minimum and maximum values are given for the different slices
`within the mandibular sections
`
`Mandible
`
`Section
`
`Side
`
`Mean
`value
`
`sd
`
`Minimum
`value
`
`Maximum
`value
`
`1
`
`2
`
`3
`
`4
`
`5
`
`6
`7
`
`8
`
`9
`
`10
`
`11
`12
`
`13
`
`14
`
`15
`
`Posterior R
`Anterior
`Posterior L
`Posterior R
`Anterior
`Posterior L
`Posterior R
`Posterior L
`Posterior R
`Anterior
`Posterior L
`Posterior R
`Posterior L
`Posterior R
`Posterior R
`Anterior
`Posterior L
`Posterior R
`Anterior
`Posterior L
`L
`Anterior
`Posterior
`Posterior R
`Posterior L
`Posterior R
`Posterior R
`Posterior L
`Posterior R
`Posterior L
`Posterior R
`Posterior L
`Posterior R
`Posterior L
`
`294
`580
`348
`177
`457
`341
`207
`238
`134
`349
`207
`479
`410
`169
`176
`444
`107
`194
`233
`213
`224
`354
`205
`207
`11
`162
`132
`227
`209
`197
`181
`462
`487
`
`47
`99
`53
`56
`115
`161
`45
`51
`53
`93
`170
`58
`77
`35
`55
`86
`151
`46
`77
`38
`92
`60
`45
`24
`35
`78
`80
`41
`70
`34
`19
`89
`84
`
`236
`433
`281
`112
`351
`72
`122
`149
`41
`214
`- 3 5
`379
`244
`101
`78
`308
`- 1 4 7
`139
`113
`169
`- 3
`257
`152
`160
`- 5 2
`64
`42
`153
`100
`142
`154
`398
`359
`
`382
`683
`425
`271
`700
`543
`277
`310
`200
`440
`461
`541
`484
`210
`261
`563
`274
`260
`394
`271
`303
`452
`309
`236
`68
`281
`281
`267
`304
`249
`216
`602
`613
`
`Discussion
`
`We determined BMD by single-energy QCT. The
`precision is higher than that of dual-energy QCT but
`
`Figure 3 Histogram showing mean values for BMD in anterior and
`posterior sections of six mandibles:
`left posterior; • anterior; &
`right posterior
`
`Figure 4 Histogram showing mean values for BMD of the first five
`and last five slices of the right posterior sections in 12 mandibles: P9
`first five scans; ΙΊ last five scans
`
`Mandible No.
`
`

`

`QCT of the mandible
`
`149
`
`single measurement was also found to be high. How-
`ever, direct CT scans perpendicular to the buccal and
`lingual bone plates are difficult to obtain clinically
`because of difficulties in patient positioning. We there-
`fore intend to compare the results of BMD measure-
`ments in reformatted axial scans with those from the
`direct images in our study.
`QCT has previously been used to measure mineral
`content in the trabecular bone of lumbar vertebrae22.
`However, results from measurements in other parts of
`the skeleton should not be compared with those in the
`mandible, as the latter seems to be subject to a more
`marked decrease
`in the amount of bone mineral
`through life25. This was confirmed by Klemetti et al.21,
`who found no correlation between BMD in the trabecu-
`lar bone of the mandible, femoral neck or lumbar
`spine. Using mandibular autopsy specimens gave us
`an opportunity to compare BMD with other measure-
`ments such as the trabecular bone volume at the same
`site which we will report subsequently.
`We found higher values of BMD in the trabecular
`bone of anterior compared with posterior sections and
`also marked variations within the same section. This is
`in agreement with Klemetti et al.21. However, they
`found generally higher values of BMD. This could be
`due to the fact that their patients were younger than
`the individuals that we examined. There are also several
`other reports confirming that the bone in the anterior
`part of the mandible is denser than in the posterior and
`that variations in bone density are found in the same
`region26-28. However, it is not clear whether dense
`bone means the high quality required for successful
`implant
`treatment. In a clinical follow-up study,
`Friberg et al.3 found the highest fixture loss in mandibles
`with the densest bone and in maxillae with low-density
`trabecular bone. They suggested that this failure may
`be due to overheating during drilling at bone sites with
`high density.
`The future demands for implant treatment are diffi-
`cult to predict but will probably increase in the partially
`dentate population. This, together with the fact that the
`anterior teeth in the mandible are usually retained the
`longest, makes it important to be able to estimate bone
`quality in the posterior parts of the jaw. In our view,
`bone quality must be described in a more detailed way,
`which should include not only mineral content but also
`
`Figure 6 CT images from a site in the mandible where BMD was
`high (left) and low (right)
`
`8
`7
`5
`Mandible No.
`
`10
`
`12
`
`13
`
`Figure 5 Histogram showing mean values of BMD of the first five
`and last five slices of the left posterior mandibular sections in 12
`last five scans; V* first five scans
`mandibles:
`
`the accuracy is lower due to the errors arising from the
`fat content of the trabecular bone23'24. The method
`used in our study reduces its influence on the derived
`values by taking into account the fact that, with few
`exceptions, trabecular bone is replaced by fatty bone
`marrow with increasing age22. This is illustrated by the
`negative BMD found at four sites, which indicates that
`there is no or very little mineral within these trabecular
`specimens. This affects both the density and atomic
`composition of the region being investigated. The
`effective atomic number of trabecular bone is reduced,
`mainly due to the higher proportion of carbon in fat
`and bone marrow and the relatively lower calcium
`hydroxyapatite [Ca10(PO4)6(OH)2] content. As a con-
`sequence, the BMD is reduced. The change in density
`as recorded by CT will therefore be a function of the
`reduction in both density and mean atomic number
`in individuals with a lower BMD. This will lead to
`a non-linear calibration curve that can be described
`by a two-degree polynominal. Different workers cali-
`brate their CT units with different concentrations of
`bone mineral-simulating substances (for example
`K 2 H P 0 4 or CaHP0 4). However, the liquid in the
`sample is commonly of the same atomic composition,
`regardless of the 'BMD' concentration. Therefore, the
`variation in effective atomic number in such samples
`does not reflect that of human trabecular bone, and
`this will lead to erroneous results, especially for low
`and high BMDs. For instance, using a calibration based
`on K 2 H P 0 4 solutions in water for the Somatom CT
`unit will yield a correct BMD result for a human trabecu-
`lar bone specimen with a mean value of about 320 HU,
`but will give results that are 6% too low and too high at
`40 and 580 HU, respectively.
`Taguchi et al.24 studied the effect of the size of the
`ROI in QCT and found that an ROI < 1 cm2 with a
`2 mm slice thickness gave unacceptable results, but the
`values for ROI > 1 cm2 were consistent. If axial scans
`are used, the buccal and lingual plates at a specific site
`limit the area that it is possible to measure. In contrast,
`none of our perpendicular slices was less than 1 cm2,
`which gave a larger area to measure. The precision of a
`
`

`

`150
`
`C. Lindh et al.
`
`the volume and structure of both cortical and trabecu-
`lar bone. BMD measurements of trabecular bone are
`needed because the skeleton undergoes age-related
`changes which affect trabecular bone, due to its higher
`turnover rate, more than they affect cortical bone29.
`QCT gives a site-related measure of BMD which could
`be an advantage since, as shown in our study, the state
`of the bone varies within a small area.
`
`References
`1. Engquist, B, Bergendal T, Kallus T, Lind6n U. A retrospective
`multicenter evaluation of osseointegrated implants supporting
`overdentures. Int J Maxillofac Implants 1988; 3: 129-34.
`2. Jaffin RA, Berman CL. The excessive loss of Bränemark fixtures
`in type IV bone: a 5-year analysis. J Periodontal 1991; 62: 2-4.
`3. Friberg B, Jemt T, Lekholm U. Early failures in 4641 consecu-
`tively placed Bränemark dental implants: a study from stage 1
`surgery to the connection of completed prostheses. Int J Oral
`Maxillofac Implants 1991; 6: 142-6.
`4. Johns RB, Jemt T, Heath RM et al. A multicenter study of
`overdentures supported by Bränemark implants. Int J Oral
`Maxillofac Implants 1992; 7: 513-22.
`5. Jemt T. Fixed implant-supported prosthesis in the edentulous
`maxilla. A five-year follow-up report. Clin Oral Implant Res
`1994; 5: 142-7.
`6. Friberg B. Bone quality evaluation during implant placement.
`Dissertation. Gothenburg: University of Gothenburg, 1994.
`7. Andresen J, Nielsen HE. Assessment of bone mineral content
`and bone mass by non-invasive radiologic methods. Acta Radiol
`1986; 27; 609-17.
`8. Kribbs P. Comparison of mandibular bone in normal and
`osteoporotic women. J Prosthet Dent 1990 63: 218-22.
`9. Devlin H, Horner K. Measurement of mandibular bone mineral
`content using the dental panoramic tomogram. J Dent 1991; 19:
`116-20.
`10. Mohajery M, Brooks SL. Oral radiographs in the detection of
`early signs of osteoporosis. Oral Surg Oral Med Oral Pathol
`1992; 73: 112-7.
`11. von Wowern N. Dual-photon absorptiometry of mandibles: in
`vitro test of a new method. Scand J Dent Res 1985; 93: 169-77.
`12. Solar P, Ulm CW, Thornton B, Matejka M. Sex-related
`differences in the bone mineral density of atrophic mandibles.
`J Prosthet Dent 1994; 71: 345-9.
`13. Corten FGA, van't Hof MA, Buijs WCAM, Hoppenbrouwers P,
`Kalk W, Corstens FHM. Measurement of mandibular bone
`density ex vivo and in vivo by dual-energy X-ray absorptiometry.
`Arch Oral Biol 1993; 38: 215-9.
`
`14. Schwarz MS, Rothman SLG, Rhodes ML, Chafetz N. Computed
`tomography: part 1. Preoperative assessment of the mandible for
`endosseous implant surgery. IntJ Oral Maxillofac Implants 1987;
`2: 137-41.
`15. Shimura M, Babbush CA, Majima H, Yanagisawa S, Sairenji E.
`Presurgical evaluation for dental implants using a reformatting
`program of computed
`tomography: maxilla/mandible shape
`pattern analysis (MSPA). Int J Oral Maxillofac Implants 1990; 5:
`175-81.
`16. Williams ΜΥΑ, Mealey BL, Hallmon WW. The role of com-
`puterized tomography in dental implantology. IntJ Oral Maxillo-
`fac Implants 1992; 7: 373-80.
`17. Lindh C, Petersson A, Klinge Β. Measurements of distances
`related to the mandibular canal in radiographs. Clin Oral Im-
`plant Res 1995; 6: 96-103.
`18. Berman CL. Complications. Prevention, recognition, treatment.
`Dent Clin North Am 1989; 33: 635-63.
`to installation of
`19. Andersson L, Kurol M. CT scan prior
`osseointegrated implants in the maxilla. Int J Oral Maxillofac
`Surg 1987; 16; 50-5.
`20. Duckmanton NA, Austin BW, Lechner SK, Klineberg IJ. Imag-
`ing for predictable maxillary implants. Int J Prosthodont 1994; 7:
`77-80.
`21. Klemetti E, Vainio P, Lassila V, Alhava E. Trabecular bone
`mineral density of mandible and alveolar height in postmeno-
`pausal women. Scand J Dent Res 1993; 101: 166-70.
`22. Nilsson M, Johnell O, Jonsson K, Redlund-Johnell I. Quantita-
`tive computed tomography in measurement of vertebral trabecu-
`lar bone mass. A modified method. Acta Radiol 1988: 29:
`719-25.
`23. Cann CE. Quantitative CT for determination of bone mineral
`density: a review. Radiology 1988; 166: 509-22.
`24. Taguchi A, Tanimoto K, Ogawa M, Sunayashiki T, Wada T.
`Effect of size of region of interest on precision of bone mineral
`measurements of
`the mandible by quantitative computed
`tomography. Dentomaxillofac Radiol 1991; 20: 25-9.
`25. von Wowern N. Bone mass of mandibles. In vitro and in vivo
`analysis. Dan Med Bull 1986; 33: 23-44.
`26. Klinge Β, Johansson C, Albrektsson T, Hallström Η, Engdahl Τ.
`A new method to obtain bone biopsies at implant sites peri-
`operatively: technique and bone structure. Clin Oral Implant Res
`1995; 6: 96-103.
`27. von Wowern N. Variations in structure within the trabecular
`bone of the mandible. Scand J Dent Res 1977; 85: 613-22.
`28. Orenstein IH, Synan WJ, Truhlar RS, Morris HF, Ochi S. Bone
`quality in patients receiving endosseous dental implants: DICRG
`interim report no. 1. Implant Dent 1994; 3: 90-4.
`29. Kimmel PL. Radiologic methods to evaluate bone mineral
`content. Ann Intern Med 1984; 100; 908-11.
`
`Address: C. Lindh, Centre for Oral Health Sciences, Carl Gustavs
`väg 34, S-214 21 Malmö, Sweden.
`
`