Temporomandibular joint computed tomography: Development of a direct sagittal technique B. van der Kuijl, D.D.S.,* L. M. Vencken, M.D.,** L. G. M. de Bont, D.D.S., Ph.D.,* and G. Boering, D.D.S., Ph.D.* University
of Groningen,
The Netherlands
Radiology plays an important role in the diagnosis of temporomandibular disorders. Different techniques are used with computed tomography offering simultaneous imaging of bone and soft tissues. It is therefore suited for visualization of the articular disk and may be used in patients with suspected internal derangements and other disorders of the temporomandibular joint. Previous research suggests advantages to direct sagittal scanning, which requires special positioning of the patient and a sophisticated scanning technique. This study describes the development of a new technique of direct sagittal computed tomographic imaging of the temporomandibular joint using a specially designed patient table and internal light visor positioning. No structures other than the patient’s head are involved in the imaging process, and misleading artifacts from the arm or the shoulder are eliminated. The use of the scanogram allows precise correction of the condylar axis and selection of exact slice level. (J PROSTHET DENT 1990,64:709-16.)
R adiology
plays an important role in diagnosing temporomandibular joint (TMJ) dis0rders.l Most current imaging techniques can be classified as conventional radiography (including transcranial and transpharyngeal techniques), tomography, and arthrography. Several combinations of techniques are advocated for different TMJ disorders. Because of its role in the physiology of temporomandibular disorders (TMDs),~ there is need for visualization of the articular disk. Conventional radiographic techniques are not suitable for visualization of the articular disk because only bony details are displayed, and arthrography has been regarded as the most appropriate x-ray imaging technique. Initial results of TMJ arthrography were described in 1944,3 and its diagnostic value is well recognized. Nevertheless, TMJ arthrography is an invasive technique that can be difficult to perform routinely. Computed tomography (CT) offers good possibilities for TMJ imaging because this technique can display details of soft-tissue and bone in the same image.4 Although axial and coronal projections have been used for TMJ diagnosis by several authors,5,6 the way the clinician looks at the TMJ and its movement patterns, together with position, shape, and size of the TMJ disk, requires evaluation in the sagittal plane. *Orofacial Research Group. **Department of Neuroradiology, gen, The Netherlands.
University
10/l/20272
THE
JOURNAL
OF PROSTHETIC
DENTISTRY
Hospital, Gronin-
This study describes the development of a technique for CT evaluation of the TMJ using direct sagittal scanning.
PRINCIPLES TOMOGRAPHY
OF COMPUTED
CT produces images of slices of the human body.It can be compared with conventional tomography regarding the use of multiple x-ray projections through the object to form a sharp image of a certain slice of this object. In conventional tomography, the image is composed of a large number of projections through the whole object. By reciprocal movement of the x-ray tube and the film, a certain slice of the object is displayed sharply because the projection of structures within this slice hit the same spot of the film, while the image of structures outside this slice is blurred because of the movement of their projections relatively to the film.’ In CT the image is reconstructed from a large number of projections through only one slice (approximately 1.5 to 12 mm thick) of the object. Radiation is bmited to the slice to be displayed. The system uses a narrow x-ray beam on one side and an array of a large number of x-ray detectors (instead of x-ray film) on the other side of the object. In this way, a projection through the slice is made. Then the system rotates for a certain angle and a projection can be made as in Fig. 1. After every projection, the amount of radiation measured by every single x-ray detector is stored in the computer. After completion of the total number of projections, the system calculates the local x-ray attenuation value
709
VANDERKUIJLETAL
- -
1
D J
Fig. 1. Principle of computed tomography. Multiple projections of thin slice (S) of object (0) are made. Amount of transmitted radiation is measured by X-ray detectors (0) in array opposite x-ray tube (X). Digital scanning data are processed by computer system to reconstruct image of slice.
Fig. 3. Principle of “target scan.” By moving x-ray tube (X) closer to center of rotation (*) of system, greater number of x-ray detectors (0) is used for measurement of area of interest, resulting in higher-quality images.
+lOOO,and is calibrated on -1000 for the density of air and on 0 for the density of water. The higher the H-value, the lighter a pixel will appear. For example, bone, which is rather dense and has a high H-value range of approximately +600 H and higher, will be displayed light; lung tissue, highly air-containing, has a low H-value range and will be displayed rather dark. In this way, differences in x-ray attenuation in the object are displayed as contrasts in the CT image.
MATERIAL CT unit
Fig. 2. Philips Tomoscan T 350 CT unit. Conventional type, movable patient table (T) for axial (transversal) and coronal (frontal) scanning in front of gantry (G).
(density value) from the original data for every part of the slice examined by means of a complex mathematical algorithm. For this purpose, the slice of the object is divided into a matrix of, for example, 256*256 parts called volume elements (voxels). Corresponding with these voxels, the CT image displayed on the monitor is divided into a matrix of, for example, 256*256 parts called picture elements (pixels). The local x-ray attenuation value, expressed in arbitrary Hounsfield units (H-units, called after the inventory of CT4), is displayed with a distinct gray-level. The Hounsfield value (H-value or H) runs from -1000 up to more than
710
AND
METHODS
In this study, a Philips Tomoscan T 350 CT unit (Philips Medical Systems, Eindhoven, The Netherlands), a third generation machine with 576 x-ray detectors, was used (Fig. 2).* System specifications are listed in Table I. Before tomographic exposures are started, a digitized overview image can be made by linear movement of the patient through the gantry with stationary x-ray tube and detector array. This “scanogram,” also called “scout view” in CT systems of other manufacturers,5 resembles conventional radiography. The radiation dose is relatively low because of the sensitive x-ray detectors. The direction of the scanogram projection is determined by the position of the tube and detectors relative to the patient. During the exposure, the computer system stores data on the position of the patient table. Thereby, the scanogram projection can be used to select the position of the following scan planes by defining slice levels in the image presented on the monitor. The system includes a target scan mode in which only a limited part of a slice from a certain level of the patient is reconstructed to an image. This system, also referred to as
TMJ
COMPUTED
TOMOGRAPHY
I. System specifications of Philips Tomoscan T 350 CF unit
Table
Item
Specification
Detectors
572 Xenon gas ionization chambers (20 Bar) Variable by target scanning Variable by target scanning and reconstructive zoom 1.5 to 12 mm Fast, regular, precision, high-precision 360 Degrees
Scanned area diameter Field of view diameter Slice thickness Scanning modes Rotation in high-precision mode Scanning time in high-precision mode Performance spatial resolution Tube tension Tube current Pulse width Pulse repetition Number of pulses in high precision mode Exposure time in high-precision mode Number of measurements in high-precision mode Reconstruction matrix
9.6 set 8 Line pairs/cm at 50% contrast 100 to 120 kV pulsed 100-600 mA 2 msec 125hec 1200 2.4 set 691.200 256*256
“geometrical enlargement,” uses variation of the focus to isocenter distance. With the scanning angle constant, a greater number of detectors is used for scanning of a limited part of the slice (Fig. 3). Fields of view smaller than the field given by a certain target scanning mode can be achieved by the reconstructive zoom technique. This technique reconstructs the desired field of view from the original scanning data. For visualization of fine details, the apparatus can be set to the high-precision mode. In this way, a spatial resolution of 8 line pairs per centimeter at 50% contrast can be achieved.
Direct
sagittal
computed
tomography
For direct sagittal CT imaging of the TMJ, a special patient table and headrest were developed (direct sagittal neuroaccessory [DSNA] set, Philips Medical Systems)g (Fig. 4). This set is based on the principle of maximal lateroflexion of the neck with the patient lying on the side of the joint under examination, It is installed on the opposite site of the gantry because the permanent patient table cannot be removed. When the gantry is tilted to its minimal angulation (-20 degrees), the bottom of the rear side of the gantry aperture is angulated 10 degrees to the horizontal plane. The special patient table is adapted to this
THE
JOURNAL
OF PROSTHETIC
4. Direct sagittal accessory set. T, Movable permanent patient table; P, special patient table for direct sagittal scanning, mounted on rear side of gantry; H, sagittal headrest for direct sagittal scanning; 8, orientation of direct sagittal scanplane. Fig.
DENTISTRY
Fig. 5. Patient positioning for direct sagittal scanning of TMJ, using direct sagittal accessory set, mounted on rear side of gantry. Head is tilted to maximal lateroflexion.
angle to form a flat support for the patient in the scanner (Fig. 5). To position the patient properly, a set of sagittal headrests (left and right side) was developed The correct sag&al headrest is selected and attached to the top of the permanent patient table. The angulation of the sagittal headrest can be adjusted to the estimated angle of the condylar axis. By this means, the angulat,ion of the selected scan planes relative to the condylar axis can be controlled. To facilitate positioning of the patient for direct sagittal scanning of the TMJ, an additional light vizor was installed within the gantry, representing the exact localization of the scan plane (Fig. 6). The sagittal headrest can be moved in and out of the gantry by moving the permanent patient ta-
711
VAN DBR KUIJL
Fig. 6. Detail of Fig. 5. Lateral side of head of patient positioned in special headrest for direct sagittal scanning, connected to movable standard patient table. Internal light visor projects position of scanplane (arrow).
ET AL
Fig. 8. Direct sagittal CT of left TMJ, closed mouth, most lateral slice. Field of view 240 mm, convolution filter 6, window width 3200 H, window level 600 H. A, External auditive meatus; B, base of skull; C, lateral pole of mandibular condyle; E, articular eminence; F, mandibular fossa; M, mastoid bone; P, tip of coronoid process; R, roof of skull.
Table II. Scanning parameters for direct sagittal CT examination of TMJ Parameter
Fig. 7. “Scanogram” view in submentovertex projection. Mandibular condyle is indicated by arrows. Selection of direct sagittal scanplanes is indicated by white dots. Most lateral slice has not to be used because it does not contain lateral pole of condyle.
ble with a remote control switch so that the correct sagittel slice through the head of the patient can be selected. Care must be taken to prevent radiation of the radiosensitive ocular lens and nerve and the thyroid gland. With the tube below the mandible and the detector array above the skull of the patient, a scanogram in submentovertex projection is made and the position of the CT slices is selected (Fig. 7). When the patient is correctly positioned, the condylar axis is perpendicular to the scan plane. If necessary, a second scanogram with corrected angulation of the headrest is made. The first slice is scanned with a 240 mm field of view (FOV) in target scan mode (Fig. 8). After the first reconstruction of the image, a smaller, 90 mm FOV, including the external auditory meatus, condylar proces, and articular eminence, is reconstructed from the original data (Fig. 9). The following slices are made on certain distances medially 712
Tension Current Scantime Exposure time Slicethickness Field of view Gantry angulation Sliceposition Slice increment
Scanogram
CT slices
120 kV 442 mA 1.0 set 1.0 set 3.0 mm (400 mm) -20 degrees
120 kV 201 mA 9.6 set 2.4 set 1.5 mm 90 mm -0 degrees
Variable
Variable
60.0 mm
3.0 mm
until the medial pole of the condyle has been reached and reconstructed with a 90 mm FOV. If necessary, openmouth pictures are made on selected slice levels (Fig. 10). (For description of the scanning parameters, see Table II.) In this study, 41 joints (21 left and 20 right joints) of 22 patients suffering from internal derangement of the TMJ (17 women, age range 19 to 48 years, mean 30.9 years; five men, age range 14 to 48 years, mean 32.7 years) were examined by direct sagittal scanning.
RESULTS Patient positioning was satisfactory in nearly all instances with technique described. Nearly all of the patients had sufficient ranges of movement of back and neck to be properly positioned for our technique of direct sagittal scanning. Older patients had more problems resulting in suboptimal caudocranial and/or antero-posterior angula-
DECEMBER
1990
VOLUME
64
NUMBER
6
TMJ
COMPUTED
TOMOGRAPHY
Fig. 9. Direct sagittal CT of left TMJ, closed mouth, second slice from lateral, enlarged by use of reconstructive zoom. Field of view of 91 mm, convolution filter 4, window width 400 H, window level 100 H. A, External auditive meatus; C, mandibular condyle; E, articular eminence; F, mandibular fossa.
tion of the scan plane or some temporary complaints of pain and stiffness of the neck afterward. The application of the submentovertex scanogram projection proved to be a great value in the examination of 38 of the 41 joints (Fig. 7). The use of the scanogram projection resulted in reliable definition of scan planes, demonstrated by appearance of the anatomic structures as selected in the scanogram image. Nevertheless, interpretation of the submentovertex scanogram was difficult in three of the 41 scanning procedures, mainly because of suboptima1 patient positioning. In some instances, the TMJ was not displayed at all. A second scanogram was made after correction of the patient’s position. The technique described produced high-quality direct sagittal high-resolution images of the TMJ. There was excellent display of bone structures, and soft tissue imaging was quite good. Time needed for the examination of one joint was approximately 15 minutes.
DISCUSSION Direct sagittal computed tomography of the TMJ is of great value in the diagnosis of TMD. Computed tomography displays bone and soft tissue structures. Injection of contrast material is not necessary as it is for arthrography, so there are no problems regarding invasive techniques. The examination seldom fails and can be completed quickly on one apparatus. Because CT is a slice technique, the image is undistorted. Because of the series of slices from the lateral to the medial pole, one gets a good impression of the three-dimensional relationships within the joint. For sagittal imaging of the TMJ, some authors use sagittal reconstructional techniques.iO-l7 This means that the
THE
JOURNAL
OF PROSTHETIC
DENTISTRY
Fig. 10. As Fig. 9, open mouth. Shape of articular disk is well demarcated (arrows). D, Articular disk.
---1
--II
Unfavorable
! I
Fig. 11. Unfavorable patient positioning technique for direct sagittal TMJ scanning. Arm and/or shoulder through gantry results in artifacts in image. 7’, Standard patient table with movable tabletop; P, special patient table for direct sagittal scanning; S, orientation of direct sagittal scan plane.
sagittal image is reconstructed from a series of continuous axial or coronal slices. This technique has a number of disadvantages. The quality of a reconstructed image is lower compared with the quality of the original axial or coronal images or compared with direct sagittal images. Furthermore, a large number of thin, high-resolution slices is needed to reconstruct a sagittal image in which one can orientate properly for good interpretation. This implicates a high radiation dose to the patient, a long examination time, and considerable computing time for reconstruction of the sagittal images. Therefore, in our opinion, direct sagittal scanning is preferable for TMJ imaging. The first problem encountered with direct sagittal CT is the positioning of the patient. CT units are developed for axial (transversal) scanning of the head and body of the
713
VAN DEB KUIJL
patient. In examination of the head, the coronal (frontal) projection is used. For direct sagittal CT, special adaptations to the CT unit must be made (Figs. 4 through 6). The maximal angulation of the gantry of the CT unit is plus or minus 20 degrees. Thus, to achieve the sagittal plane of the head to be scanned, extension, lateroilexion, and rotation of the neck should correct for the remaining 70 degrees, theoretically. The maximal angle that can be achieved in normal individuals is approximately 60 degrees. Therefore, the remaining 10 degrees must be overcome by altering of the angulation of the body-axis of the patient. Because the body-axis angulation of the permanent patient table cannot be changed, a special device constructed with a body-axis of -10 degrees was needed. To verify the position of the patient in routine CT, a light-vizor is mounted in front of the gantry on most CT units. Since the distance between the light vizor and the scanplane is known, the patient can be positioned in front of the gantry and moved to the selected scan plane along this distance. In our direct sagittal scanning system, the external light vizor cannot be used. Because of the rigid special patient table, the patient cannot be moved into the gantry aperture, but must be positioned directly in respect to the radiation beam. For this reason, an internal light vizor, representing the exact location of the (direct sagittal) scan plane is installed. For direct sagittal CT of the TMJ, the scan plane should be corrected for the angulation of the condylar axis. In the development of arthrotomography, several authors describe the use of the submentovertex projection for determination of the thickness of the overlying structures, the dimensions of the condyle, and the angulation of the condylar axis.ls In addition, for direct sagittal CT, the scan plane should be oriented perpendicular to the condylar axis. This can be established in two ways. First, we can make a conventional submentovertex x-ray image before the examination and measure the angulation. This technique has the disadvantage that the patient must be examined on another apparatus before the CT procedure can be started. Second, the CT system offers the possibility to make a submentovertex view of the patient’s head by using the scanogram feature. In this way, the submentovertex view and the scan planes are correlated to each other within the system itself. Another advantage of the use of the scanogram is that one is not seduced to go “searching” for the joint in cases of bad positioning of the patient, causing an unnecessary increase of radiation dose to the patient. The use of the submentovertex scanogram for patient positioning and for corrected sagittal scanning is not described in earlier literature and proved to be valuable in our study. However, patient positioning was critical: when the patient’s head was not sufficiently tilted, interpretation could be nearly impossible because of overprojection of other skull structures. Some authors use a system in which the patient, lying on the back, is placed on a additional patient table in front of 714
ET AL
the gantry of the scanner1gV25(Fig. 10). With one arm and shoulder through the aperture of the gantry, the TMJ to be examined can be positioned correctly with respect to the radiation beam. However, these authors do not describe the possibility of correction of the angulation of the condylar axis with respect to the scan plane in their systems. This can be achieved by using the gantry tilt and by rotation of the patient’s head along the body axis. A scanogram cannot be made with these techniques of patient positioning because the desired movement of the permanent patient table is impaired. Some patient positioning techniques described in the literature are more or less comparable to our direct sagittal set, but are aimed to examine the contralateral joint in this position; thus the neck and the shoulder are in the scan plane.26 The presence of an arm, a shoulder, or the neck in the gantry aperture produces considerable image distortion, seen as severe linear streaking artifacts2’ These artifacts even may produce false-positive visualization of the articular disk in some patients. Careful study of the literature in this respect revealed caudocranial streaking artifacts in most of the images, produced by using a patient positioning technique as described above. Care must be taken not to confuse these effects with artifacts from dental work. To avoid artifacts, the arm and shoulder are carefully prevented from taking part in the imaging process in our technique. After collection of the rough x-ray attenuation data from the detector array in the scanner, the image is reconstructed by the computer system. This process and the evaluation process can be influenced by important factors with considerable consequences for the diagnosis of bone and soft-tissue disorders. An article describing reconstruction and evaluation parameters and the application of different image-processing modalities is in preparation.
REFERENCES 1. Blaschke DD. Radiology of the temporomandibular joint: current status of transcranial, tomographic, and arthrographic procedures. In: Laskin D, Greenfield W, Gale E, et al, eds. The president’s conference on the examination, diagnosis and management of temporomandibular joint disorders. Chicago: American Dental Association, 1982;64-74. 2. Farrar WB, McCarty WL Jr. A clinical outline of temporomandibular joint diagnosis and treatment. 7th ed. Montgomery: Normandie Puhlications, 1983. 3. NQrgaard F. Arthrography of the mandibular joint. Acta Radio1 [Diagn] (Stockholm) 1944;25:679-85. 4. Hounsfield GN. Computerized transverse axial scanning (tomography): Part I. Description of system. Br J Radio1 1973;46:1016-22. 5. Christiansen EL, Thompson JR. Anteriorly displaced temporomandibular joint disk. Report of a case diagnosed by computed tomography. Oral Surg Oral Med Oral Path01 1984;58:355-7. 6. Avrahami E, Horowitz I, Cohn DF. Computed tomography of the temporomandihular joint. Comput Radio1 1984,8:211-6. ‘7. Rosenberg HM. Laminagraphy: methods and application in oral diagnosis. J Am Dent Assoc 1967;74:88-96. 8. Tomoscan operator’s manual. Eindhoven: Philips Medical Systems, 1984. 9. van Waes PFGM, Zonneveld FW, Damsma H, Rabischong P, Vignaud J. Direct multiplanar CT of the petrous bone. Eindhoven: Philips Medical Systems, 1982. DECEMBER
1990
VOLUME
64
NUMBER
6
TMJ
COMPUTED
TOMOGRAPHY
10. Helms CA, Katzberg RW, Morrish R, Dolwick MF. Computed tomography of the temporomandibular joint meniscus. J Oral Maxillofac Surg 1983;41:512-7. 11. Christiansen EL, Thompson JR, Kopp SFO, Haaso AN, Hinshaw DB. Radiographic signs of temporomandibular joint disease: an investigation utilizing x-ray computed tomography. Dentomaxillofac Radio1 1985;14:83-92. 12. Fjellstroem C-A, Olofsson 0. Computed tomography of the temporomandibular joint meniscus. J Maxillofac Surg 1985;13:24-7. 13. Freesmeyer WB, Luckenbach A, Muller Th, Huls A. Vergleichende Untersuchungen zwischen mechanisch und elektronisch registrierter Unterkieferbewegung in Beziehung zur Gelenktopografie. Dtscb Zahnarztl Z 1984;39:870-5. 14. Helms CA, Richardson ML, Vogler III JB, Hoddick WK. Computed tomography for diagnosing temporomandibular joint disk displacement. J Craniomand Pratt 1984;3:23-6. 15. Helms CA, Vogler II JB, Morrish RB, Goldman SM, Capra RE, Proctor E. Temporomandibular joint internal derangements: CT diagnosis. Radiology 1984;152:459-62. 16. Helms CA, Vogler II, JB Morrisb RB. Diagnosis by computed tomography of temporomandibular joint meniscus displacement. J PROSTHET DENT 1984;51:544-7.
17. Katzberg RW, Tallents RH, Hayakawa K, Miller TL, Goske MJ, Wood HP. Internal derangements of the tamporomandibular joint: findings in the pediatric age group. Radiology 1985;154:125-7. 18. Bussard DA, Kerr G, Hutton C, Yune H. Technique and use of “corrected -axis” tomograms of the mandibular condyles. Oral Surg Oral Med Oral Pathol 1980;49:394-7. 19. Cohen HR, Silver CM, Schatz SL, Motamed MM. Correlation of sagittal computed tomography of the temporomandibular joint with surgical findings. J Craniomand Pratt 1985;3:351-7.
Copies of Prosthodontics
20. Manco LG, Messing SG, Busino LJ, Fasulo CP, Sordill WC. Internal derangements of the temporomandibular joint evaluated with direct sagittal CT: a prospective study. Radiology 1985;157:407-12. 21. Manco LG. Messing SG. Splint therapy evaluation with direct sagittal computed tomography. Oral Surg Oral Med Oral Path01 1986;61:5-11. 22. Manzione JV, Katzberg RW, Brodsky GL, Seltzer SE, Mellins HZ. Internal derangements of the temporomandibular joint: diagnosis by direct sagittal computed tomography. Radiology 1984;150:111-5. 23. Manzione JV, Katzberg RW, Manzione TJ. Internal derangements of the temporomandibular joint II: diagnosis by arthrography and computed tomography. Int J Periodont Restor Dent 1984;4:17-27. 24. Manzione JV, Seltzer SE, Katzberg RW, Hammerschlag SB, Chiango BF. Direct sagittal computed tomography of the temporomandibular joint. AJNR 1982;3:677-9. 25. Raustia AM, Pyhtinen J, Virtanen KK. Examination of the temporomandibular joint by direct sagittal computed tomography. Clin Radio1 1985;36:291-6. 26. Sartoris DJ, Neumann CH, Riley RW. The temporomandibular joint: true sagittal computed tomography with meniscus visualization. Radiology 1984;150:250-4. 27. Simon DC, Hess ML, Smilak MS, Beltran J. Direct sagittal CT of the temporomandibular joint. Radiology 1985;157:545. Reprint requests to: DR. B. VAN DER KUIJL OROFACIAL RESEARCH GROUP UNIVERSITY OF GRONINCEN ANT. DEUSINGLAAN 1
9713 AV GRONINGEN THE NETHERLANDS
21 symposium
proceedings
available
Copies of the September 1990 issue of THE JOURNAL OF PIKWTHETIC DENTISTRY (vol. 64, No. 3), which carried Proceedings of F’rosthodontics 21, a national symposium on prosthodontics sponsored by the Federation of Prosthodontic Organizations October l-51989 at Mayo Medical Center, Rochester, Minn., are available for purchase from the publisher, Mosby-Year Book, Inc., at a cost of $6.50. (Foreign postage is not included.) Quantity discounts are available. Please contact Katherine Carter, Director of Subscription Services, Mosby-Year Book, Inc., 11830 Westline Industrial Drive, St. Louis, MO 63146-3318, or call (314)872-8370, ext. 7422.
THE
JOURNAL
OF PROSTHETIC
DENTISTRY
715
of Groningen,
The Netherlands
Radiology plays an important role in the diagnosis of temporomandibular disorders. Different techniques are used with computed tomography offering simultaneous imaging of bone and soft tissues. It is therefore suited for visualization of the articular disk and may be used in patients with suspected internal derangements and other disorders of the temporomandibular joint. Previous research suggests advantages to direct sagittal scanning, which requires special positioning of the patient and a sophisticated scanning technique. This study describes the development of a new technique of direct sagittal computed tomographic imaging of the temporomandibular joint using a specially designed patient table and internal light visor positioning. No structures other than the patient’s head are involved in the imaging process, and misleading artifacts from the arm or the shoulder are eliminated. The use of the scanogram allows precise correction of the condylar axis and selection of exact slice level. (J PROSTHET DENT 1990,64:709-16.)
R adiology
plays an important role in diagnosing temporomandibular joint (TMJ) dis0rders.l Most current imaging techniques can be classified as conventional radiography (including transcranial and transpharyngeal techniques), tomography, and arthrography. Several combinations of techniques are advocated for different TMJ disorders. Because of its role in the physiology of temporomandibular disorders (TMDs),~ there is need for visualization of the articular disk. Conventional radiographic techniques are not suitable for visualization of the articular disk because only bony details are displayed, and arthrography has been regarded as the most appropriate x-ray imaging technique. Initial results of TMJ arthrography were described in 1944,3 and its diagnostic value is well recognized. Nevertheless, TMJ arthrography is an invasive technique that can be difficult to perform routinely. Computed tomography (CT) offers good possibilities for TMJ imaging because this technique can display details of soft-tissue and bone in the same image.4 Although axial and coronal projections have been used for TMJ diagnosis by several authors,5,6 the way the clinician looks at the TMJ and its movement patterns, together with position, shape, and size of the TMJ disk, requires evaluation in the sagittal plane. *Orofacial Research Group. **Department of Neuroradiology, gen, The Netherlands.
University
10/l/20272
THE
JOURNAL
OF PROSTHETIC
DENTISTRY
Hospital, Gronin-
This study describes the development of a technique for CT evaluation of the TMJ using direct sagittal scanning.
PRINCIPLES TOMOGRAPHY
OF COMPUTED
CT produces images of slices of the human body.It can be compared with conventional tomography regarding the use of multiple x-ray projections through the object to form a sharp image of a certain slice of this object. In conventional tomography, the image is composed of a large number of projections through the whole object. By reciprocal movement of the x-ray tube and the film, a certain slice of the object is displayed sharply because the projection of structures within this slice hit the same spot of the film, while the image of structures outside this slice is blurred because of the movement of their projections relatively to the film.’ In CT the image is reconstructed from a large number of projections through only one slice (approximately 1.5 to 12 mm thick) of the object. Radiation is bmited to the slice to be displayed. The system uses a narrow x-ray beam on one side and an array of a large number of x-ray detectors (instead of x-ray film) on the other side of the object. In this way, a projection through the slice is made. Then the system rotates for a certain angle and a projection can be made as in Fig. 1. After every projection, the amount of radiation measured by every single x-ray detector is stored in the computer. After completion of the total number of projections, the system calculates the local x-ray attenuation value
709
VANDERKUIJLETAL
- -
1
D J
Fig. 1. Principle of computed tomography. Multiple projections of thin slice (S) of object (0) are made. Amount of transmitted radiation is measured by X-ray detectors (0) in array opposite x-ray tube (X). Digital scanning data are processed by computer system to reconstruct image of slice.
Fig. 3. Principle of “target scan.” By moving x-ray tube (X) closer to center of rotation (*) of system, greater number of x-ray detectors (0) is used for measurement of area of interest, resulting in higher-quality images.
+lOOO,and is calibrated on -1000 for the density of air and on 0 for the density of water. The higher the H-value, the lighter a pixel will appear. For example, bone, which is rather dense and has a high H-value range of approximately +600 H and higher, will be displayed light; lung tissue, highly air-containing, has a low H-value range and will be displayed rather dark. In this way, differences in x-ray attenuation in the object are displayed as contrasts in the CT image.
MATERIAL CT unit
Fig. 2. Philips Tomoscan T 350 CT unit. Conventional type, movable patient table (T) for axial (transversal) and coronal (frontal) scanning in front of gantry (G).
(density value) from the original data for every part of the slice examined by means of a complex mathematical algorithm. For this purpose, the slice of the object is divided into a matrix of, for example, 256*256 parts called volume elements (voxels). Corresponding with these voxels, the CT image displayed on the monitor is divided into a matrix of, for example, 256*256 parts called picture elements (pixels). The local x-ray attenuation value, expressed in arbitrary Hounsfield units (H-units, called after the inventory of CT4), is displayed with a distinct gray-level. The Hounsfield value (H-value or H) runs from -1000 up to more than
710
AND
METHODS
In this study, a Philips Tomoscan T 350 CT unit (Philips Medical Systems, Eindhoven, The Netherlands), a third generation machine with 576 x-ray detectors, was used (Fig. 2).* System specifications are listed in Table I. Before tomographic exposures are started, a digitized overview image can be made by linear movement of the patient through the gantry with stationary x-ray tube and detector array. This “scanogram,” also called “scout view” in CT systems of other manufacturers,5 resembles conventional radiography. The radiation dose is relatively low because of the sensitive x-ray detectors. The direction of the scanogram projection is determined by the position of the tube and detectors relative to the patient. During the exposure, the computer system stores data on the position of the patient table. Thereby, the scanogram projection can be used to select the position of the following scan planes by defining slice levels in the image presented on the monitor. The system includes a target scan mode in which only a limited part of a slice from a certain level of the patient is reconstructed to an image. This system, also referred to as
TMJ
COMPUTED
TOMOGRAPHY
I. System specifications of Philips Tomoscan T 350 CF unit
Table
Item
Specification
Detectors
572 Xenon gas ionization chambers (20 Bar) Variable by target scanning Variable by target scanning and reconstructive zoom 1.5 to 12 mm Fast, regular, precision, high-precision 360 Degrees
Scanned area diameter Field of view diameter Slice thickness Scanning modes Rotation in high-precision mode Scanning time in high-precision mode Performance spatial resolution Tube tension Tube current Pulse width Pulse repetition Number of pulses in high precision mode Exposure time in high-precision mode Number of measurements in high-precision mode Reconstruction matrix
9.6 set 8 Line pairs/cm at 50% contrast 100 to 120 kV pulsed 100-600 mA 2 msec 125hec 1200 2.4 set 691.200 256*256
“geometrical enlargement,” uses variation of the focus to isocenter distance. With the scanning angle constant, a greater number of detectors is used for scanning of a limited part of the slice (Fig. 3). Fields of view smaller than the field given by a certain target scanning mode can be achieved by the reconstructive zoom technique. This technique reconstructs the desired field of view from the original scanning data. For visualization of fine details, the apparatus can be set to the high-precision mode. In this way, a spatial resolution of 8 line pairs per centimeter at 50% contrast can be achieved.
Direct
sagittal
computed
tomography
For direct sagittal CT imaging of the TMJ, a special patient table and headrest were developed (direct sagittal neuroaccessory [DSNA] set, Philips Medical Systems)g (Fig. 4). This set is based on the principle of maximal lateroflexion of the neck with the patient lying on the side of the joint under examination, It is installed on the opposite site of the gantry because the permanent patient table cannot be removed. When the gantry is tilted to its minimal angulation (-20 degrees), the bottom of the rear side of the gantry aperture is angulated 10 degrees to the horizontal plane. The special patient table is adapted to this
THE
JOURNAL
OF PROSTHETIC
4. Direct sagittal accessory set. T, Movable permanent patient table; P, special patient table for direct sagittal scanning, mounted on rear side of gantry; H, sagittal headrest for direct sagittal scanning; 8, orientation of direct sagittal scanplane. Fig.
DENTISTRY
Fig. 5. Patient positioning for direct sagittal scanning of TMJ, using direct sagittal accessory set, mounted on rear side of gantry. Head is tilted to maximal lateroflexion.
angle to form a flat support for the patient in the scanner (Fig. 5). To position the patient properly, a set of sagittal headrests (left and right side) was developed The correct sag&al headrest is selected and attached to the top of the permanent patient table. The angulation of the sagittal headrest can be adjusted to the estimated angle of the condylar axis. By this means, the angulat,ion of the selected scan planes relative to the condylar axis can be controlled. To facilitate positioning of the patient for direct sagittal scanning of the TMJ, an additional light vizor was installed within the gantry, representing the exact localization of the scan plane (Fig. 6). The sagittal headrest can be moved in and out of the gantry by moving the permanent patient ta-
711
VAN DBR KUIJL
Fig. 6. Detail of Fig. 5. Lateral side of head of patient positioned in special headrest for direct sagittal scanning, connected to movable standard patient table. Internal light visor projects position of scanplane (arrow).
ET AL
Fig. 8. Direct sagittal CT of left TMJ, closed mouth, most lateral slice. Field of view 240 mm, convolution filter 6, window width 3200 H, window level 600 H. A, External auditive meatus; B, base of skull; C, lateral pole of mandibular condyle; E, articular eminence; F, mandibular fossa; M, mastoid bone; P, tip of coronoid process; R, roof of skull.
Table II. Scanning parameters for direct sagittal CT examination of TMJ Parameter
Fig. 7. “Scanogram” view in submentovertex projection. Mandibular condyle is indicated by arrows. Selection of direct sagittal scanplanes is indicated by white dots. Most lateral slice has not to be used because it does not contain lateral pole of condyle.
ble with a remote control switch so that the correct sagittel slice through the head of the patient can be selected. Care must be taken to prevent radiation of the radiosensitive ocular lens and nerve and the thyroid gland. With the tube below the mandible and the detector array above the skull of the patient, a scanogram in submentovertex projection is made and the position of the CT slices is selected (Fig. 7). When the patient is correctly positioned, the condylar axis is perpendicular to the scan plane. If necessary, a second scanogram with corrected angulation of the headrest is made. The first slice is scanned with a 240 mm field of view (FOV) in target scan mode (Fig. 8). After the first reconstruction of the image, a smaller, 90 mm FOV, including the external auditory meatus, condylar proces, and articular eminence, is reconstructed from the original data (Fig. 9). The following slices are made on certain distances medially 712
Tension Current Scantime Exposure time Slicethickness Field of view Gantry angulation Sliceposition Slice increment
Scanogram
CT slices
120 kV 442 mA 1.0 set 1.0 set 3.0 mm (400 mm) -20 degrees
120 kV 201 mA 9.6 set 2.4 set 1.5 mm 90 mm -0 degrees
Variable
Variable
60.0 mm
3.0 mm
until the medial pole of the condyle has been reached and reconstructed with a 90 mm FOV. If necessary, openmouth pictures are made on selected slice levels (Fig. 10). (For description of the scanning parameters, see Table II.) In this study, 41 joints (21 left and 20 right joints) of 22 patients suffering from internal derangement of the TMJ (17 women, age range 19 to 48 years, mean 30.9 years; five men, age range 14 to 48 years, mean 32.7 years) were examined by direct sagittal scanning.
RESULTS Patient positioning was satisfactory in nearly all instances with technique described. Nearly all of the patients had sufficient ranges of movement of back and neck to be properly positioned for our technique of direct sagittal scanning. Older patients had more problems resulting in suboptimal caudocranial and/or antero-posterior angula-
DECEMBER
1990
VOLUME
64
NUMBER
6
TMJ
COMPUTED
TOMOGRAPHY
Fig. 9. Direct sagittal CT of left TMJ, closed mouth, second slice from lateral, enlarged by use of reconstructive zoom. Field of view of 91 mm, convolution filter 4, window width 400 H, window level 100 H. A, External auditive meatus; C, mandibular condyle; E, articular eminence; F, mandibular fossa.
tion of the scan plane or some temporary complaints of pain and stiffness of the neck afterward. The application of the submentovertex scanogram projection proved to be a great value in the examination of 38 of the 41 joints (Fig. 7). The use of the scanogram projection resulted in reliable definition of scan planes, demonstrated by appearance of the anatomic structures as selected in the scanogram image. Nevertheless, interpretation of the submentovertex scanogram was difficult in three of the 41 scanning procedures, mainly because of suboptima1 patient positioning. In some instances, the TMJ was not displayed at all. A second scanogram was made after correction of the patient’s position. The technique described produced high-quality direct sagittal high-resolution images of the TMJ. There was excellent display of bone structures, and soft tissue imaging was quite good. Time needed for the examination of one joint was approximately 15 minutes.
DISCUSSION Direct sagittal computed tomography of the TMJ is of great value in the diagnosis of TMD. Computed tomography displays bone and soft tissue structures. Injection of contrast material is not necessary as it is for arthrography, so there are no problems regarding invasive techniques. The examination seldom fails and can be completed quickly on one apparatus. Because CT is a slice technique, the image is undistorted. Because of the series of slices from the lateral to the medial pole, one gets a good impression of the three-dimensional relationships within the joint. For sagittal imaging of the TMJ, some authors use sagittal reconstructional techniques.iO-l7 This means that the
THE
JOURNAL
OF PROSTHETIC
DENTISTRY
Fig. 10. As Fig. 9, open mouth. Shape of articular disk is well demarcated (arrows). D, Articular disk.
---1
--II
Unfavorable
! I
Fig. 11. Unfavorable patient positioning technique for direct sagittal TMJ scanning. Arm and/or shoulder through gantry results in artifacts in image. 7’, Standard patient table with movable tabletop; P, special patient table for direct sagittal scanning; S, orientation of direct sagittal scan plane.
sagittal image is reconstructed from a series of continuous axial or coronal slices. This technique has a number of disadvantages. The quality of a reconstructed image is lower compared with the quality of the original axial or coronal images or compared with direct sagittal images. Furthermore, a large number of thin, high-resolution slices is needed to reconstruct a sagittal image in which one can orientate properly for good interpretation. This implicates a high radiation dose to the patient, a long examination time, and considerable computing time for reconstruction of the sagittal images. Therefore, in our opinion, direct sagittal scanning is preferable for TMJ imaging. The first problem encountered with direct sagittal CT is the positioning of the patient. CT units are developed for axial (transversal) scanning of the head and body of the
713
VAN DEB KUIJL
patient. In examination of the head, the coronal (frontal) projection is used. For direct sagittal CT, special adaptations to the CT unit must be made (Figs. 4 through 6). The maximal angulation of the gantry of the CT unit is plus or minus 20 degrees. Thus, to achieve the sagittal plane of the head to be scanned, extension, lateroilexion, and rotation of the neck should correct for the remaining 70 degrees, theoretically. The maximal angle that can be achieved in normal individuals is approximately 60 degrees. Therefore, the remaining 10 degrees must be overcome by altering of the angulation of the body-axis of the patient. Because the body-axis angulation of the permanent patient table cannot be changed, a special device constructed with a body-axis of -10 degrees was needed. To verify the position of the patient in routine CT, a light-vizor is mounted in front of the gantry on most CT units. Since the distance between the light vizor and the scanplane is known, the patient can be positioned in front of the gantry and moved to the selected scan plane along this distance. In our direct sagittal scanning system, the external light vizor cannot be used. Because of the rigid special patient table, the patient cannot be moved into the gantry aperture, but must be positioned directly in respect to the radiation beam. For this reason, an internal light vizor, representing the exact location of the (direct sagittal) scan plane is installed. For direct sagittal CT of the TMJ, the scan plane should be corrected for the angulation of the condylar axis. In the development of arthrotomography, several authors describe the use of the submentovertex projection for determination of the thickness of the overlying structures, the dimensions of the condyle, and the angulation of the condylar axis.ls In addition, for direct sagittal CT, the scan plane should be oriented perpendicular to the condylar axis. This can be established in two ways. First, we can make a conventional submentovertex x-ray image before the examination and measure the angulation. This technique has the disadvantage that the patient must be examined on another apparatus before the CT procedure can be started. Second, the CT system offers the possibility to make a submentovertex view of the patient’s head by using the scanogram feature. In this way, the submentovertex view and the scan planes are correlated to each other within the system itself. Another advantage of the use of the scanogram is that one is not seduced to go “searching” for the joint in cases of bad positioning of the patient, causing an unnecessary increase of radiation dose to the patient. The use of the submentovertex scanogram for patient positioning and for corrected sagittal scanning is not described in earlier literature and proved to be valuable in our study. However, patient positioning was critical: when the patient’s head was not sufficiently tilted, interpretation could be nearly impossible because of overprojection of other skull structures. Some authors use a system in which the patient, lying on the back, is placed on a additional patient table in front of 714
ET AL
the gantry of the scanner1gV25(Fig. 10). With one arm and shoulder through the aperture of the gantry, the TMJ to be examined can be positioned correctly with respect to the radiation beam. However, these authors do not describe the possibility of correction of the angulation of the condylar axis with respect to the scan plane in their systems. This can be achieved by using the gantry tilt and by rotation of the patient’s head along the body axis. A scanogram cannot be made with these techniques of patient positioning because the desired movement of the permanent patient table is impaired. Some patient positioning techniques described in the literature are more or less comparable to our direct sagittal set, but are aimed to examine the contralateral joint in this position; thus the neck and the shoulder are in the scan plane.26 The presence of an arm, a shoulder, or the neck in the gantry aperture produces considerable image distortion, seen as severe linear streaking artifacts2’ These artifacts even may produce false-positive visualization of the articular disk in some patients. Careful study of the literature in this respect revealed caudocranial streaking artifacts in most of the images, produced by using a patient positioning technique as described above. Care must be taken not to confuse these effects with artifacts from dental work. To avoid artifacts, the arm and shoulder are carefully prevented from taking part in the imaging process in our technique. After collection of the rough x-ray attenuation data from the detector array in the scanner, the image is reconstructed by the computer system. This process and the evaluation process can be influenced by important factors with considerable consequences for the diagnosis of bone and soft-tissue disorders. An article describing reconstruction and evaluation parameters and the application of different image-processing modalities is in preparation.
REFERENCES 1. Blaschke DD. Radiology of the temporomandibular joint: current status of transcranial, tomographic, and arthrographic procedures. In: Laskin D, Greenfield W, Gale E, et al, eds. The president’s conference on the examination, diagnosis and management of temporomandibular joint disorders. Chicago: American Dental Association, 1982;64-74. 2. Farrar WB, McCarty WL Jr. A clinical outline of temporomandibular joint diagnosis and treatment. 7th ed. Montgomery: Normandie Puhlications, 1983. 3. NQrgaard F. Arthrography of the mandibular joint. Acta Radio1 [Diagn] (Stockholm) 1944;25:679-85. 4. Hounsfield GN. Computerized transverse axial scanning (tomography): Part I. Description of system. Br J Radio1 1973;46:1016-22. 5. Christiansen EL, Thompson JR. Anteriorly displaced temporomandibular joint disk. Report of a case diagnosed by computed tomography. Oral Surg Oral Med Oral Path01 1984;58:355-7. 6. Avrahami E, Horowitz I, Cohn DF. Computed tomography of the temporomandihular joint. Comput Radio1 1984,8:211-6. ‘7. Rosenberg HM. Laminagraphy: methods and application in oral diagnosis. J Am Dent Assoc 1967;74:88-96. 8. Tomoscan operator’s manual. Eindhoven: Philips Medical Systems, 1984. 9. van Waes PFGM, Zonneveld FW, Damsma H, Rabischong P, Vignaud J. Direct multiplanar CT of the petrous bone. Eindhoven: Philips Medical Systems, 1982. DECEMBER
1990
VOLUME
64
NUMBER
6
TMJ
COMPUTED
TOMOGRAPHY
10. Helms CA, Katzberg RW, Morrish R, Dolwick MF. Computed tomography of the temporomandibular joint meniscus. J Oral Maxillofac Surg 1983;41:512-7. 11. Christiansen EL, Thompson JR, Kopp SFO, Haaso AN, Hinshaw DB. Radiographic signs of temporomandibular joint disease: an investigation utilizing x-ray computed tomography. Dentomaxillofac Radio1 1985;14:83-92. 12. Fjellstroem C-A, Olofsson 0. Computed tomography of the temporomandibular joint meniscus. J Maxillofac Surg 1985;13:24-7. 13. Freesmeyer WB, Luckenbach A, Muller Th, Huls A. Vergleichende Untersuchungen zwischen mechanisch und elektronisch registrierter Unterkieferbewegung in Beziehung zur Gelenktopografie. Dtscb Zahnarztl Z 1984;39:870-5. 14. Helms CA, Richardson ML, Vogler III JB, Hoddick WK. Computed tomography for diagnosing temporomandibular joint disk displacement. J Craniomand Pratt 1984;3:23-6. 15. Helms CA, Vogler II JB, Morrish RB, Goldman SM, Capra RE, Proctor E. Temporomandibular joint internal derangements: CT diagnosis. Radiology 1984;152:459-62. 16. Helms CA, Vogler II, JB Morrisb RB. Diagnosis by computed tomography of temporomandibular joint meniscus displacement. J PROSTHET DENT 1984;51:544-7.
17. Katzberg RW, Tallents RH, Hayakawa K, Miller TL, Goske MJ, Wood HP. Internal derangements of the tamporomandibular joint: findings in the pediatric age group. Radiology 1985;154:125-7. 18. Bussard DA, Kerr G, Hutton C, Yune H. Technique and use of “corrected -axis” tomograms of the mandibular condyles. Oral Surg Oral Med Oral Pathol 1980;49:394-7. 19. Cohen HR, Silver CM, Schatz SL, Motamed MM. Correlation of sagittal computed tomography of the temporomandibular joint with surgical findings. J Craniomand Pratt 1985;3:351-7.
Copies of Prosthodontics
20. Manco LG, Messing SG, Busino LJ, Fasulo CP, Sordill WC. Internal derangements of the temporomandibular joint evaluated with direct sagittal CT: a prospective study. Radiology 1985;157:407-12. 21. Manco LG. Messing SG. Splint therapy evaluation with direct sagittal computed tomography. Oral Surg Oral Med Oral Path01 1986;61:5-11. 22. Manzione JV, Katzberg RW, Brodsky GL, Seltzer SE, Mellins HZ. Internal derangements of the temporomandibular joint: diagnosis by direct sagittal computed tomography. Radiology 1984;150:111-5. 23. Manzione JV, Katzberg RW, Manzione TJ. Internal derangements of the temporomandibular joint II: diagnosis by arthrography and computed tomography. Int J Periodont Restor Dent 1984;4:17-27. 24. Manzione JV, Seltzer SE, Katzberg RW, Hammerschlag SB, Chiango BF. Direct sagittal computed tomography of the temporomandibular joint. AJNR 1982;3:677-9. 25. Raustia AM, Pyhtinen J, Virtanen KK. Examination of the temporomandibular joint by direct sagittal computed tomography. Clin Radio1 1985;36:291-6. 26. Sartoris DJ, Neumann CH, Riley RW. The temporomandibular joint: true sagittal computed tomography with meniscus visualization. Radiology 1984;150:250-4. 27. Simon DC, Hess ML, Smilak MS, Beltran J. Direct sagittal CT of the temporomandibular joint. Radiology 1985;157:545. Reprint requests to: DR. B. VAN DER KUIJL OROFACIAL RESEARCH GROUP UNIVERSITY OF GRONINCEN ANT. DEUSINGLAAN 1
9713 AV GRONINGEN THE NETHERLANDS
21 symposium
proceedings
available
Copies of the September 1990 issue of THE JOURNAL OF PIKWTHETIC DENTISTRY (vol. 64, No. 3), which carried Proceedings of F’rosthodontics 21, a national symposium on prosthodontics sponsored by the Federation of Prosthodontic Organizations October l-51989 at Mayo Medical Center, Rochester, Minn., are available for purchase from the publisher, Mosby-Year Book, Inc., at a cost of $6.50. (Foreign postage is not included.) Quantity discounts are available. Please contact Katherine Carter, Director of Subscription Services, Mosby-Year Book, Inc., 11830 Westline Industrial Drive, St. Louis, MO 63146-3318, or call (314)872-8370, ext. 7422.
THE
JOURNAL
OF PROSTHETIC
DENTISTRY
715
