UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Effect of stent marker materials on the localization of cross-sectional images using spiral tomography Orpe, Elaine 2005

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-ubc_2005-0584.pdf [ 9.26MB ]
Metadata
JSON: 831-1.0092136.json
JSON-LD: 831-1.0092136-ld.json
RDF/XML (Pretty): 831-1.0092136-rdf.xml
RDF/JSON: 831-1.0092136-rdf.json
Turtle: 831-1.0092136-turtle.txt
N-Triples: 831-1.0092136-rdf-ntriples.txt
Original Record: 831-1.0092136-source.json
Full Text
831-1.0092136-fulltext.txt
Citation
831-1.0092136.ris

Full Text

Effect of Stent Marker Materials on the Localization of Cross-Sectional Images using Spiral Tomography by Elaine Orpe D . M . D . , The University of Manitoba, 1991 A thesis submitted in partial fulfillment of the requirements for the degree of Master of Science in The Faculty of Graduate Studies Dental Science The University of British Columbia July 2005 © Elaine Orpe 2005 11 Abstract With the increase in endosseous dental implant placement there is an increasing demand for cross-sectional imaging of the jaws to facilitate treatment planning of the patient. While there are many articles recommending different types o f marker materials in guide stents to be used for this imaging, no quantitative comparisons o f the efficacy of such materials have been published. The purpose of this study was to compare different types o f marker materials for radiographic guide stents for conventional spiral tomography. Differences in the ability o f observers to localize cross-sectional images relative to longitudinal images with the different types o f markers were determined. Radiographic guide stents for dried human mandibles and one metatarsal bone were fabricated to accept acrylic cylinders with different marker materials in the same relative position. These were imaged using conventional spiral tomography and observers were asked to localize the cross-sectional images on corresponding longitudinal images using the markers as guides. A pilot study confirmed that there were statistically significant differences in the mean absolute error o f slice localization by observers for different marker materials. The pilot study also confirmed that any type of marker resulted in much less mean absolute error in slice localization than use o f no marker. The definitive study found statistically significant differences in mean error for slice localization not only for different types o f markers, but also for the location of the tomographic slice and for the stents made for different bones. Ill Table of Contents Abstract ii Table of Contents iii List of Tables v List of Figures vi Acknowledgements viii Chapter 1 Introduction 1 Chapter 2 Review of Literature 8 Rationale 16 Chapter 3 Pilot Study 18 3.1 Introduction 18 3.2 Materials and Methods 20 3.3 Results 32 Chapter 4 Definitive Study 37 4.1 Goals 38 4.2 Materials and Methods 39 Chapter 5 Results 46 5.1 Interobserver Variability 46 5.2 Location of the Tomographic Slice 48 5.3 Size and Material of the Marker 50 5.4 Marker Material 52 5.5 Stent 54 Chapter 6 57 6.1 Discussion 57 6.2 Conclusions 59 Bibliography 60 Appendix Measurement 70 1. Resolution 2. Precision 3. Accuracy 4. Validity 5. Discussion 6. Conclusions Reference List of Tables 1. A N O V A results for pilot study. 34 2. Bonferroni probability values for marker materials. 35 3. A N O V A results for observer. 47 4. Bonferroni probability values for observer. 47 5. A N O V A results for location of tomographic slice. 48 6. Bonferroni probability values for location of tomographic slice. 49 7. A N O V A results for marker type. 50 8. Bonferroni probability values for marker type. 51 9. A N O V A results for marker material. 52 10. Bonferroni probability values for marker material. 53 11. A N O V A results for stent. 54 12. Bonferonni probability values for stent. 55 13. Contrast ratio values for line-pair plate image made with the cephalometric attachment of the Comm-Cat. 78 14. Line-pairs per millimetre at 50% contrast ratio for tomographic images. 84 15. Contrast ratios for the second line-pair group using spiral tomography. 85 vi List of Figures 1. Mandibular radiographic/surgical guide stent with gutta percha markers for a partially edentulous patient. 19 2. Maxillary radiographic/surgical guide stent with brass tubing markers for a fully edentulous patient. 20 3. Stent markers for pilot study. 22 4. CommCat complex-motion tomography unit. 24 5. Mandible in position on custom support table. 25 6. Occlusal radiograph of the mandible taken with small ball bearing markers in place on the stents. 26 7. Close up view of mandible and stent with amalgam markers. 27 8. Spiral tomograms made with 1.2 mm gutta percha markers. 28 9. Spiral tomograms made with 3.2 mm gutta percha markers. 28 10. Spiral tomograms made with 3.2 mm brass tubing markers. 29 11. Spiral tomograms made with 1.6 mm ball bearing markers. 29 12 Spiral tomograms made with 5.9 mm ball bearing markers. 30 13. Spiral tomograms with acetate and observer's markings. 31 14. Close-up view of observer's markings. 31 15. Pilot study marker versus mean error graph. 33 16. Pilot study marker material versus mean error graph. 34 17. Markers used for definitive study. 40 18. Occlusal radiograph of one of the mandibles. 42 19. Occlusal-type radiograph taken of metatarsal bone. 42 20. Instructions to observers. 45 21. Observer versus mean error graph. 46 22. Location of tomographic slice versus mean error graph. 48 23. Marker versus mean error graph. 50 Vll 24. Stent marker material versus mean error graph. 52 25. Stent versus mean error graph. 54 26. Contact radiograph of line-pair plate. 70 27. Spiral tomogram of line-pair plate at 1 mm slice thickness. 72 28. Graph of top row of line-pair plate using cephalometric unit. 73 29. Graph of mean pixel values for top row of line-pair plate, spiral tomogram with horizontal line-pair plate. 74 30. Graph of mean pixel value for top row of line-pair plate, circular motion tomogram, vertical line-pair plate. 76 31. Second line-pair set used for contrast ratio. 77 32. Graph of contrast ratio plotted against line-pair group for the image of the line-pair plate made with the cephalometric attachment of the CommCat. 78 33. Graph of contrast ratios for spiral tomograms at 1 mm slice thickness with line-pair plate horizontally oriented. 79 34. Graph of contrast ratios for hypocycloidal tomograms at 1 mm slice thickness with line-pair plate vertically oriented. 79 35. Graph of measurements of the line-pairs to line-pair group on image made with Kodak D-speed intraoral film. 81 36. Graph of mean line-pairs per mm for different image types at 50% contrast ratio. 82 37. Graph of line-pairs per mm at 50% contrast ratio for spiral tomography. 82 38. Graph of line-pairs per mm at 50% contrast ratio for hypocycloidal tomography. 83 39. Graph of line-pairs per mm at 50% contrast ratio for circular tomography. 85 Vll l Acknowledgements M y first thanks must go to Dr . Garnet Packota, for initially inspiring my interest in oral and maxillofacial radiology. I wi l l always be grateful to Dr . Michael J . Pharoah, whose interpretive skills were matched by his inspirational teaching example. M y residency with him was the most enjoyable part o f my prolonged academic training and has instilled me with a deep love of my chosen field o f practice. M y parents and my brother Graham provided me with extensive technical assistance and support, and without their help I could not have accomplished this work. Dr . K e n Hutchison provided expert proofreading and corrections. I would like to thank all o f my observers for the hours o f work they contributed while completing their intensive dental education. W i l l Ruehl provided hours o f technical assistance. I would particularly like to thank my co-supervisor, Dr . Dav id MacDonald. Without his contribution to the U . B . C . Faculty o f Dentistry I would not have been able to complete this work. Most o f all I am grateful to D r . Col in Price for all of his help and guidance, and for agreeing to continue to supervise me long after his retirement. 1 CHAPTER 1 INTRODUCTION Endosseous dental implants fabricated from titanium or titanium alloy have been recognized as an acceptable treatment in North America for over 20 years (Taylor, T. D. and Agar, J. R. 2002; Kassebaum, D. K. et al. 1990) and have been used in Europe for over 40 years (Linkow, L. I. and Dorfman, J. D. 1991). With a better than 90% five-year success rate, their use has been increasing dramatically over the past decade (Mupparapu, M. and Singer, S. R. 2004; Tyndall, A. A. and Brooks, S. L. 2000) . However, there have been reports of serious complications from implant placement ranging from temporary to permanent altered sensation of the mental nerve (Ziccardi, V. B. and Assael, L. A. 2001; Kraut, R. A. and Chahal, O. 2002; Harris, J. M. 2002; Gregg, J. M . 2000; Ellies, L. G. 1992; Dao, T. T. and Mellor, A. 1998; Chaushu, G. et al. 2002; Bartling, R. et al. 1999) to life-threatening hemorrhage or airway obstruction (Worthington, P. 1995; Niamtu, J., I l l 2001; Mordenfeld, A. et al. 1997; Kalpidis, C. D. and Setayesh, R. M. 2004; Isaacson, T. J. 2004; Givol, N . et al. 2000; Givol, N . et al. 2002; Darriba, M. A. and Mendonca-Caridad, J. J. 1997; Boyes-Varley, J. G. and Lownie,J. F. 2002; 1991). A more common untoward situation is an error in implant placement. In this case, the implant though successfully integrated into the bone, is placed in a position that is not an amenable to functional and/or aesthetic restoration 2 (Adrian, E . D . etal. 1992; Almog, D . M . etal. 2001; Kopp, K . C. etal. 2003; Tyndall, A . A . and Brooks, S. L . 2000; Solow R.A. 2001; Takeshita, F. and Suetsugu, T. 1996; Verde, M . A. and Morgano, S. M . 1993). Although most implants are placed successfully with only intraoral and/or panoramic imaging, many authors have recommended cross-sectional imaging such as linear or complex motion conventional tomography or computed tomography with an imaging guide stent. This approach allows practitioners to avoid the above complications, is an aid in treatment planning, and leads to improved communication between the restorative and surgical dentists and the dental laboratory (Tyndall, A . A . and Brooks, S. L . 2000; Adrian, E . D . etal. 1992; Almog, D . M . etal. 2001; Becker, C. M . and Kaiser, D . A . 2000; Engelman, M . J . etal. 1988; Iplikcioglu, H . etal. 2002; Cehreli, M . C. etal. 2000; Cehreli, M . C. and Sahin, S. 2000; Cehreli, M.' C. etal. 2002; Akca, K . etal. 2002; Kopp, K . C. etal. 2003; Mupparapu, M . 2002; Mupparapu, M . and Singer, S. R. 2004; Pesun, I. J. and Gardner, F. M . 1995; Pharoah, M . J. 1993; Petrikowski, C. G. etal. 1989; K u , Y . and Shen, Y . F. 2000; Solow R.A. 2001; Sykaras, N . and Woody, R. D . 2001;Takeshita, F. and Suetsugu, T. 1996; Varvara, G . etal. 2003; Verde, M . A . and Morgano, S. M . 1993; Wat, P. Y. etal. 2002; Gray, C. F. etal. 1996; Gray, C. F. etal. 1998b; Dula, K . etal. 2001) . The American Academy of Oral and Maxillofacial Radiology has published a position paper (Tyndall, A . A . and Brooks, S. L. 2000) and parameters of care document (White, S. C. et al. 2001) recommending cross-sectional imaging in the treatment planning of all patients receiving dental implants. This recommendation has not been accepted universally as the standard of care. The European Association for Osseointegration has produced its own 3 guidelines for the use of diagnostic imaging in implant planning which only recommends the use of cross-sectional imaging in selected cases where the clinician feels that insufficient information is achieved by clinical evaluation and panoramic and intraoral radiographic examination (Harris, D . etal. 2002). A review of the literature indicates some difference in bias for or against the use of cross-sectional imaging in dental implant planning from different branches of the profession and from different regions. In articles addressing this issue published in prosthetic, periodontal, and radiographic journals the authors predominantly favour the use of cross-sectional imaging in planning implant placement. However, this view is not shared by some European radiologists (Harris, D . etal 2002; Frei, C. et al. 2004; Dula, K . et al. 1997; Dula, K . etal. 2001) who recommend reserving cross-sectional imaging for complex cases. To further ensure accurate evaluation of the bone contours at the planned site most authors recommending routine cross-sectional imaging in treatment planning also recommend the use of an imaging guide with radiopaque markers that may be converted to a surgical guide (Tyndall, A . A . and Brooks, S. L. 2000; Adrian, E . D . etal. 1992; Almog, D . M . etal. 2001; Becker, C. M . and Kaiser, D . A. 2000; Engelman, M . J. etal. 1988; Iplikcioglu, H . etal. 2002; Cehreli, M . C. etal. 2000; Cehreli, M . C. and Sahin, S. 2000; Cehreli, M . C. etal. 2002; Akca, K . etal. 2002; Kopp, K . C. etal. 2003; Mupparapu, M . 2002; Mupparapu, M . and Singer, S. R. 2004; Pesun, I. J. and Gardner, F. M . 1995; Pharoah, M . J. 1993; Petrikowski, C. G . etal. 1989; K u , Y . and Shen, Y . F. 2000; Solow R A . 2001; Sykaras, N . and Woody, R. D . 2001; Takeshita, F. and 4 Suetsugu, T. 1996; Varvara, G . etal. 2003; Verde, M . A . and Morgano, S. M . 1993; Wat, P. Y . etal. 2002). The rationale for the use of the above technique is that this will ensure ideal placement of implants with maximum potential for functional and esthetic restorations. Deficiencies in the bone contour will be identified prior to surgery and appropriate corrective measures such as bone grafting can be planned. However, some authors favour restricting the radiographic examination to a panoramic view in most cases (Frei, C. etal. 2004; Pieper, S. P. and Lewis, S. G . 2001; Bartling, R. et al. 1999; Tal, H . and Moses, O. 1991; Verde, M . A. and Morgano, S. M . 1993). One oral surgeon, though the title of his article suggests that routine use of cross-sectional imaging is desirable, in fact only recommended its use for select cases (Kraut, R. A . 2001). The argument of these authors against routine use of cross-sectional imaging in treatment planning of implant cases is that vertical measurements of the bone are adequate using a panoramic image and the bony contours can be evaluated at the time of surgery and corrections made at that time. This may be sufficient to avoid most serious complications such as hemorrhage due to perforation of the lingual cortex of the mandible involving the lingual artery or violation of the inferior alveolar nerve canal. However, an implant that is osseointegrated and does not violate critical anatomical structures may be a technical success, but its position or angulation may be such that it cannot be functionally or aesthetically restored (Adrian, E . D . etal. 1992; Almog, D . M . etal. 2001; Kopp, K . C. et al. 2003; Tyndall, A . A. and Brooks, S. L. 2000; Solow R. A. 2001; Takeshita, F. and Suetsugu, T. 1996; Verde, M . A. and Morgano, S. M . 1993; Choi, M . etal. 2004). 5 Several authors state that panoramic images should not be used for measurements because of distortion and variability in magnification (Lam, E. W. etal. 1995; Reiskin, A. B. 1998; Engelman, M . J. etal. 1988; Frederiksen, N . L. 1995; Kassebaum, D . K . etal. 1992; Miles, D . A. and Van Dis, M . L. 1993; Mupparapu, M . and Singer, S. R. 2004; DelBalso, A. M . etal. 1994; Tyndall, A . A. and Brooks, S. L. 2000; White, S. C. etal. 2001; Gray, C. F. etal. 2001; Gray, C. F. etal. 2003; Bolin, A . etal. 1996) Methods to obtain cross-sectional images of the jaws include lateral cephalometric views for the anterior maxilla or mandible, conventional tomography using linear, spiral, or hypocycloidal motions, conventional (sequential slice) computed tomography, spiral computed tomography, cone-beam or limited cone-beam computed tomography, and magnetic resonance imaging. Tuned aperture computed tomography has been described as another potential method of providing cross-sectional imaging for dental implant treatment planning (Rashedi, B. et al. 2003; Limrachtamorn, S. et al. 2004; Liang, H . et al. 1999), but this is not yet available commercially. Lateral cephalometric films can provide cross-sectional imaging limited to the anterior mandible or maxilla with little distortion if the magnification is known (Lam, E. W. etal. 1995; Engelman, M . J . etal. 1988; Adrian, E . D . etal. 1992; Gray, C. F. et al. 2003; Miles, D . A. and Van Dis, M . L. 1993; Reiskin, A. B. 1998; Kassebaum, D. K . et al. 1992). This type of imaging is readily available, low in cost, and low in radiation dose to the patient. 6 Conventional tomography has been reported to be the method of choice for treatment planning one or two implant sites in simple cases (White, S. C. etal. 2001; Tyndall, A . A . and Brooks, S. L . 2000). The reported effective radiation doses for such cases are relatively low, and the financial cost intermediate. Access to complex-motion tomography may be limited. Many newer dental panoramic units offer conventional linear tomography and two units offer complex motion tomography (Panorex CMT, Imaging Sciences Intl., Hatfield PA; Cranex Tome, Soredex, Milwaukee WI). Linear tomography has been reported to be subject to distortion (Todd, A. D . etal 1993; Butterfield, K . J. et al. 1997). Hypocycloidal conventional tomography is reported to create the best blurring of objects outside the focal trough, but spiral conventional tomography has been demonstrated to provide the best image of the inferior alveolar nerve canal (Serhal, C. B. etal 2001; Lindh, C. etal. 1992). Computed tomography is considered by many to be the best method of imaging for dental implant planning and the method of choice in treatment planning complex cases (Frederiksen, N . L. 1995; Cehreli, M . C. etal 2002; Reiskin, A . B. 1998; Varvara, G . etal. 2003; Friedland, B. and Valachovic, R. W. 1997; Friedland, B. 2003; White, S. C. etal. 2001; Tyndall, A . A . and Brooks, S. L. 2000). Conventional (sequential slice) computed tomography has the highest radiation dose of the methods of cross-sectional imaging for dentistry (Friedland, B. and Valachovic, R. W. 1997; Friedland, B. 2003; Ekestubbe, A. et al. 1992; Ekestubbe, A. etal. 1993; Lecomber, A. R, etal. 2001; Scaf, G . etal. 1997). Spiral computed tomography is reported to have effective doses much lower than conventional computed tomography but higher than that for conventional tomography when only a few sites are treatment planned for implants (Bou, S. C. etal 2001). Newer cone-beam computed tomography 7 scans have reported effective doses several times lower than spiral computed tomography, but more than panoramic radiography (Ludlow, J . B. etal. 2003; Schulze, D . etal. 2004). Limited cone-beam computed tomography is only suitable for a single site or a few adjacent sites, but has an effective dose comparable to panoramic radiography (personal communication with representative of J . Morita USA inc.). Magnetic resonance imaging has the advantage of using no ionizing radiation and providing cross-sectional images of the jaws (Gray, C. F. etal. 1996; Gray, C. F. etal. 1998a; Gray, C. F. etal. 1998b; Gray, C. F. etal. 2003;Tyndall, A . A . and Brooks, S. L. 2000). Currently the cost of the scans is much higher than other cross-sectional imaging modalities and access to scanners is limited (Engelman, M . J . etal. 1988). There is also some concern about distortions due to field inhomogeneity and the presence of certain dental materials such as steel dentin pins, and to magnetic susceptibility difference between bone and soft tissue or between tissue and air. Low field magnetic resonance scanners are less susceptible to such distortion, and have a significantly lower capital cost (Gray, C. F. etal. 1998a; Gray, C. F. etal. 1998b; Gray, C. F. etal. 2003) that may translate into a lower cost to the patient. 8 CHAPTER 2 REVIEW OF LITERATURE The earliest articles identified concerning implant guide stents referred only to surgical guide stents that were not used for imaging (Edge, M . J. and Medina, T. V . 1988; Edge, M . J . 1987; Balshi, T . J . and Garver, D . G . 1987; Johnson, C. M . etal. 1988; Burns, D . R. etal 1988). Although it has been shown that use of a surgical stent does improve accuracy of implant placement from a prosthetic standpoint, this is only true when the patient's anatomy allows the implant to be placed in the preferred location. This approach often leads to surgeons abandoning the stent and placing the implant in the location that the patient's anatomy dictates, which can result in an implant unsuitable for restoration (Adrian, E . D . etal. 1992; Almog, D . M . etal. 2001; Kopp, K . C. etal. 2003; Tyndall, A . A . and Brooks, S. L . 2000; Solow R A . 2001; Takeshita, F. and Suetsugu, T. 1996; Verde, M . A. and Morgano, S. M . 1993). Numerous other articles describe surgical templates for implant placement without reference to imaging and these will not be discussed. Many articles do describe the use of imaging guides for radiographic evaluation of potential implant sites which may be converted to use as a surgical guide. Most of these studies are either case reports or technical notes on fabrication of the authors' recommended imaging guide design. Only two articles were identified that compared different types of imaging guide marker materials (Borrow, J. W. and Smith, J . P . I 996; Almog, D . M . et al 2001), but these did not provide evidence of quantifiable differences between the marker materials 9 and instead gave subjective descriptions of the advantages and disadvantages of the different materials. In 1996 Borrow etal. published an article describing different materials used as markers for patients undergoing multiplanar reformatted computerized tomography for presurgical implant planning. The advantages and disadvantages of the different types of markers for computed tomography imaging is described (Borrow, J. W. and Smith, J. P. 1996). Almog etal. (Almog, D . M . etal. 2001) are the only authors who have compared different types of markers for conventional cross-sectional tomography. No objective studies were performed to compare the efficacy of the different types of stents. The technical method of fabrication of four different types of stents is described with a discussion of the advantages of each of the different types. For all but a modification of one of the methods described, a diagnostic wax up of the planned site is recommended. The types of stents described are as follows: 1. A vacuum-formed template fabricated over a cast of the diagnostic wax up with the edentulous site filled with acrylic resin and a thin lead strip adapted over the buccal occlusal and lingual surfaces. It is noted that at this technique does not provide a long axis indicator for the osteotomy site and may allow for incorrect angulation during surgery. 2. A vacuum-formed template fabricated as above but with the edentulous site filled with acrylic resin and a buccal access groove prepared at the planned site of the implant. This groove has a thin lead strip adapted to its lingual surface. This method does serve as both an imaging guide and an 10 osteotomy guide but is subject to variation in bucco-lingual angulation during surgery. 3. A vacuum-formed template fabricated as above but with the edentulous site filled with acrylic resin and a prepared access hole at the planned implant site filled with gutta percha. The gutta percha is then removed to convert the imaging guide to a surgical stent. This will provide a more accurate surgical guide than the previous two methods. 4. For the last method described, a hole is prepared through the diagnostic wax up and into the cast to a depth of 10 mm. A Guide Right (De Plaque Corp., Victor, N.Y.) pin is placed into the prepared hole and a matching steel sleeve is placed over the pin and an acrylic guide is fabricated capturing the sleeve in place. This provides a very accurate surgical guide provided the cross-sectional imaging confirms that bone is available for the planned trajectory. A second method for fabricating this type of stent is described in which setup disks also provided by De Plaque Corp. may be used in place of the diagnostic wax up to ensure adequate spacing of the implants. It is noted that if diagnostic imaging reveals a discrepancy between the available bone and the planned position or trajectory of the implant then the stent will need to be modified accordingly, but a method for modifying the stent is not explained. Borrow and Smith describe the use of seven different types of stent markers in computed tomography (Borrow, J. W. and Smith, J. P. 1996). Similarly to Almog et al, they describe the advantages and disadvantages of each of the markers, make no attempt to quantify the differences. 1 1 Steel ball bearings are recommended by several authors, particularly in combination with panoramic radiography as the known size of the ball bearing may be used to calculate the magnification factor of the panoramic image (Miles, D . A . and Van Dis, M . L. 1993; Mupparapu, M . and Singer, S. R. 2004; Petrikowski, C. G . etal. 1989; Engelman, M . J . etal. 1988; Pieper, S. P. and Lewis, S. G . 2001; Kraut, R. A. 2001; Tal, H . and Moses, O. 1991; Ismail, Y . H . et al. 1995; Arlin M L 2005). It is also mentioned that distortion of the shape of the ball-bearing with conventional tomography will alert the clinician to the presence of distortion of the image (Poon, C. K . et al. 1992). However, the ball bearing marker does not provide information about the trajectory of the planned implant relative to the available bone. Also, metallic artifact from a ball bearing interferes with computed tomography imaging. Use of a smaller size of ball bearing (2 millimetre diameter versus the more often recommended 5 millimetre diameter) reduces the metal artifact and the problem of the ball bearing appearing on multiple images (Borrow, J. W. and Smith, J. P. 1996). A variant of die use of ball bearings is the use of a ball clasp wire with a known diameter of the ball portion of the clasp wire. This allows an angular reference to be included while still allowing for magnification correction with the panoramic scan (Frederiksen, N . L. 1995; Arlin M L 2005). Many authors recommend doing a diagnostic wax up (Adrian, E . D . et al. 1992; Almog, D . M . etal. 2001; Arlin M L 2005; Becker, C. M . and Kaiser, D . A. 2000; Engelman, M . J. etal. 1988; Floyd, P. P. P. and Palmer R. 1999; Kopp, K . C. etal. 2003; K u , Y . and Shen, Y. F. 2000; Lee, S. Y . and Morgano, S. M . 1994; Lee, S. J. and Toothaker, R. W. 1998; Pesun, I. J. and Gardner, F. 12 M . 1995; Takeshita, F. etal. 1997; Takeshita, F. and Suetsugu, T. 1996; Solow R.A. 2001; Tsuchida, F. etal. 2004; Varvara, G . etal. 2003; Walker, M . and Hansen, P. 1999; Wat, P. Y . etal. 2002) or denture tooth set up (Cehreli, M . C. etal. 2000; Cehreli, M . C and Sahin, S. 2000; Cehreli, M . C. etal. 2002; Shahrasbi, A . H . and Hansen, C. A. 2002; Weingart, D . and Duker, J. 1993) of the desired prosthetic result and a duplication of this in acrylic to be used as a guide stent. Alternatively, a provisional fixed partial denture (Stellino, G . et al. 1995) or removable denture (Tsai, T. P. et al. 2001) may be used as a guide stent, providing that it is well-fitted with the teeth in the desired locations for the definitive prosthesis. Acrylic stents or dentures with prepared holes filled with radiopaque materials provide a long axis to allow correlation of the planned implant site with the patient's anatomy. Materials that have been recommended for filling the holes include gutta percha (Watzinger, F. et al. 1999; Stellino, G . et al. 1995; Varvara, G. etal. 2003; Wat, P. Y . etal. 2002; K u , Y . and Shen, Y . F. 2000; Mupparapu, M . and Singer, S. R. 2004), radiopaque temporary filling material (Tsai, T. P. et al. 2001; Tsuchida, F. etal. 2004), and amalgam (Ismail, Y . H . etal. 1995). Gutta percha and temporary filling material are of suitable radiopacity for use with either conventional or computed tomography, and are easily removed from the guide stent. Amalgam creates excessive metallic artifact with computed tomography and is more difficult to remove from the guide stent or denture. The periphery of the teeth on the guide stent may be made opaque by coating them with barium sulphate (Israelson, H . et al. 1992; Borrow, J. W. and Smith, J. P. 1996) or adapting lead foil to the surface (Adrian, E . D . etal. 1992; Almog, D . M . etal. 2001; Urquiola, J. and Toothaker, R. W. 1997). This may be used in conjunction with a prepared guide hole that may be filled with an opaque 13 material or coated with barium sulfate. One article describes the use of a channel lined with lead foil instead of a guide hole (Almog, D . M . etal 2001). Alternatively the entire tooth may be made radiopaque by mixing barium sulfate (Walker, M . and Hansen, P. 1999; Takeshita, F. and Suetsugu, T. 1996; Borrow, J. W. and Smith, J. P. 1996) or an iodine contrast medium (Siu, A. S. et al. 2003; Borrow, J. W. and Smith, J. P. 1996) with the acrylic used in its. fabrication so that the guide hole will appear radiolucent within it. This works well with CT imaging but not with conventional tomography. Borrow and Smith describes adding a more opaque material such as gutta percha to the guide holes as an alternative approach. The incorporation of barium sulfate into the acrylic is technique sensitive — an appropriate amount of the material must be used to achieve a suitable level of radiopacity, and it may not disperse evenly. The use of an iodine contrast medium allows even radiopacity of the acrylic, but use of this material may have an increased risk of patient sensitivity to the stent. Additionally, prefabricated radiopaque teeth may be used to make the stent in place of the diagnostic wax-up (Besimo, C. E . etal. 2000; Besimo, C. etal 1995; Borrow, J. W. and Smith, J. P. 1996). Some authors recommend that a longitudinal marker be place buccal to the planned implant site, which is particularly useful when there are adjacent restorations that could obscure the marker if it were centred over the edentulous ridge (Kopp, K . C. et al. 2003; Shahrasbi, A . H . and Hansen, C. A . 2002; Borrow, J. W. and Smith, J. P. 1996). The distance from the marker to the crest of the alveolar bone can be measured on the cross-sectional images and this used to adapt the stent for surgery appropriately. 14 Solow describes an elegant imaging stent with radiopaque composite resin channels (Solow R.A. 2001). A clear description of its fabrication is given, with meticulous instructions on modification of the stent to correspond with the patient's anatomy after imaging if necessary. Sethi recommends a unique stent made by using an actual denture set up in wax rims with radiopaque material placed in buccal grooves in the wax for the imaging template (Sethi, A . 1993). This is then duplicated after the scan and the image data used to prepare the guide holes for the surgical stent. Metal tubing such as brass or steel tubes provide very clear images for conventional tomography and are generally thin enough that they do not cause excessive metallic artifact for. computed tomography (Becker, C. M . and Kaiser, D . A. 2000; Cehreli, M . C. etal. 2002; Takeshita, F. and Suetsugu, T. 1996; Takeshita, F. etal. 1997). The metal tube provides a very precise guide for the surgeon. Almog etal. found that prosthodontists and periodontists favour a restrictive guide such as is provided by these types of stents (Almog, D . M . et al. 2002). The more recent paper by Takeshita uses the metal tube in conjunction with barium mixed into the acrylic of the teeth of the stent (Takeshita, F. etal. 1997). The stent described by Cehreli is unique in that it uses telescoping metal guides that can be removed sequentially as the diameter of the drills used during surgery is increased (Cehreli, M . C. etal. 2002). Besimo etal describe the use of a titanium cross-shaped marker for use in computed tomography, with one arm of the cross aligned perpendicular to (a tangent to) the dental arch (Besimo, C. E . et al. 2000). For magnetic resonance imaging, the available contrast materials (usually gadolinium based) are typically in liquid form for use intravenously. A method 15 to seal the contrast material into an acrylic stent with prepared guide holes is described (Gray, C F. etal. 1998a; Gray, C F. etal 1998b; Gray, C F. etal. 2003). A few authors mention incorporating some sort of horizontal alignment guide into the stent to ensure proper head alignment for the scan (Besimo, C. E. et al. 2000; Petrikowski, C. G . etal. 1989; Varvara, G . etal. 2003). This is important for conventional tomography since the images cannot be manipulated to correct the angulation after acquisition and errors in angulation will result in incorrect length measurements (Petrikowski, C. G . et al. 1989; Choi, S. C. etal 2002). For conventional and spiral computed tomography, distortions may occur in reformatted images made at an angle not perpendicular to the scanning plane (Varvara, G . etal. 2003; Verstreken, K . etal. 1996; Kohavi, D . etal. 1997; Dantas, J. A . et al. 2005). However, the image matrix of cone-beam computed tomography scanners uses cuboidal voxels, instead of rectangular voxels as for conventional or spiral computed tomography. This allows for image reconstructions for cone-beam computed tomography to be dimensionally stable along all axes (Marmulla, R. et al. 2005; Lascala, C. A. et al. 2004; Kobayashi, K . et al. 2004; Fortin, T. et al. 2002), and for software to be used to correct the angle of the cross-sectional images to correspond with the planned implant trajectory. 16 RATIONALE There is a lack of published data comparing different stent marker materials for conventional tomography. For the surgeon to be able to use the information obtained by cross-sectional imaging there must be a way to relate the cross-sectional images to the clinical situation. The complex-motion tomography unit used in this study does provide a matched corrected sagittal image for the tomograms to which the cross-sectional images may be fairly accurately located (Thunthy, K. H. 2000), but other systems in common usage such as the Scanora (Soredex, Tuusula, Finland) or Cranex Tome (Soredex, Tuusula, Finland) do not. The intent of this study was to determine whether use of different stent marker materials would affect the ability of observers to localize cross-sectional images. Before undertaking the definitive study a pilot study was performed to determine whether the study design was suitable. To be able to compare the different materials to one another it was necessary to be able to position the different markers in the same relative position to the support structure, and to obtain the images in the same planes. For this reason it was decided that the stents would be fabricated such that the marker materials could be interchanged without removing the stent from the support structure, and the images could then be made without changing the orientation of the stent or markers. A review of the literature failed to reveal information regarding the resolution of the CommCat system, and it was therefore decided to perform a 17 supplementary study comparing the resolution of the tomographic images produced by the CommCat using different tomographic motions and slice thicknesses with a line-pair plate. Cr 18 CHAPTER 3 PILOT STUDY 3.1 Introduction Use of conventional tomography has been advocated for evaluation of potential implant sites to avoid surgical or restorative complications. While most authors recommend use of a radiographic stent, preferable one which can be modified to function as a surgical guide, there are a wide range of marker materials used and advocated. For examples of actual radiographic/surgical stents used for patients see figures 1 and 2. Of the materials identified for use as radiographic markers for conventional tomography, amalgam, gutta percha, and steel ball bearings were selected based on personal experience of frequency of use of these materials by practitioners (figure 3). Temporary filling material (Cavit G, 3 M ESPE, St. Paul, M.N., USA) markers were fabricated as well. However, the gutta percha markers and the temporary filling markers were of similar radiographic density, but drying of the temporary filling material resulted in gaps in the markers. The gaps would add another variable to the study which might have affected the observers' ability to localize the images. For this reason the temporary filling material markers were eliminated from the study. In addition, brass tubing was used because of a strong recommendation for its use from a referring dentist. 1 9 Figure 1. Mandibular radiographic/surgical guide stent with gutta percha markers for a partially edentulous patient. The surface has also been coated with barium sulphate in acrylic resin to allow the outer contours of the planned restorations to be seen on the tomographic images. 20 Figure 2. Maxillary radiographic/surgical guide stent with brass tubing markers for a fully edentulous patient. 3.2 Materials and Methods A dried human mandible was used to support the stents. The posterior region of the mandible is the site for which there appears to be the greatest consensus as to the need for cross-sectional imaging. This is due to the presence of the neurovascular bundle within the inferior alveolar nerve canal and the lingual nerve and artery that may be violated if the lingual plate of the mandible is breached during surgery. The CommCat (Imaging Sciences International, Hatfield PA, USA) is a dedicated dental tomographic unit that is specifically designed to make 21 conventional tomography of the jaws simple and provides a method to choose the tomographic slices relative to the dental arch. The familiarity of the investigator with using the CommCat to image the jaws made the use of mandibles to support the guide stents intuitive (figure 4). It was not intended that the observers would refer to the anatomy of the mandible in making their estimations of the locations of the cross-sectional images. Because the anatomy of the mandible should not have been a factor in the decision making process of the observers, the use of a soft-tissue equivalent material to introduce soft-tissue scatter was not believed to be necessary. The presence of soft-tissue if wet mandibular specimens were used should not have affected the images of the stents tiiemselves as the scatter would be mostly confined to a region well below the markers. The presence of buccal soft tissues can effect image quality at the level of the stents in live patients, as increased image noise is seen in large patients. However, use of full-skull human wet specimens would not have been practical, particularly since the CommCat unit was located at a downtown private practice facility and not on campus. Two light-cured (Triad baseplate material, 3M) acrylic stents were made for each side of a fully edentulous dried human mandible. Two sites were prepared on each stent to accept 6.35 millimetre cylindrical acrylic rods. 32 acrylic rods were fashioned, 16 each of heights 7.5 millimetres and 5.8 millimetres. Matched sets of one of each height of acrylic rod were prepared for 16 different types of stent markers (figure 3). For all stents, the shorter acrylic rod was positioned in the posterior site and the longer rod in the higher site. This 22 corresponds to the situation most often found in patients, where the clearance to the opposing arch is less posteriorly (figures 5 and 6). The stent marker materials were as follows: -amalgam in diameters of 1.2 ,1.6, 2.0, 2.4, and 3.2 millimetres -gutta percha in diameters of 1.2,1.6, 2.0,2.4, and 3.2 millimetres -brass tubing in outer diameters of 2.4 and 3.2 millimetres -stainless steel ball bearings in diameters of 1.6,2.4, 3.5, and 5.9 millimetres Figure 3. Stent markers for pilot study. 23 A CommCat (figure 4) dedicated dental complex-motion tomographic unit was used to take the tomographic images using Fuji Super HRG30 film (Fuji Photo Film Company, Tokyo, Japan) with Kodak Ektavision (Kodak Canada Inc., Toronto O N , Canada) intensifying screens. The films were processed with fresh Nature Care solutions (Patterson Dental Canada Inc., Montreal QB, Canada) in an AT 2000 X R processor (Air Techniques, Hicksville NY, USA). Figure 4. CommCat complex-motion tomography unit. 2 5 Imaging Sciences International custom fabricated a support table for the mandible which could be substituted for the patient chin rest. This had four holes drilled into it, and a polyvinyl siloxane (Exaflex Type O putty, GC America Inc., Alsip IL, USA) positioning jig was made which fitted into the holes while stabilizing the mandible with its occlusal plane horizontal to the floor (Figure 3 and Figure 4). Figure 5. Mandible in position on custom support table. Prior to making any images for the study, the machine calibration protocol provided by Imaging Sciences Int. for the CommCat was performed to ensure the machine was functioning as accurately as possible. 26 An occlusal radiograph of the mandible with the stents in place using small ball bearing markers was obtained (Figure 6). This radiograph was scanned into the CommCat software to map the arch and choose the tomographic slices to be taken. Figure 6. Occlusal radiograph of the mandible taken with small ball bearing markers in place on the stents. The smallest ball-bearing markers were used in one stent and the next to smallest ball-bearing markers in the other. Calibration films for each stent set-up consisting of one corrected sagittal image and one cross-sectional image were taken and repeated until correct slice locations and exposure parameters were achieved. Study films were then produced consisting of one corrected sagittal and 5 cross-sectional images per stent-marker combination. For each series, the markers of the same type were used, with the shorter marker placed more posteriorly. A series with no marker 27 was also made for each stent. The same 5 slice locations were used for all the series made on each stent, with a spacing of 4 millimetres between slices. Figure 7. Close-up view of mandible and stent with amalgam markers. The order for the images to be taken for each stent-marker combination was randomized by listing all of the different possible orders and assigning them numbers from 1-120 (the permutations of possible slice orders). Microsoft Excel was then used to generate a random number table. The images were produced such that, when viewed, each side would appear to be the right side of the mandible, with the corrected sagittal tomographic image first and the cross-sectional tomographic images following. Letters A-E were marked over 28 the cross-sectional images, and the films were coded and randomized, again using Microsoft Excel to generate a random number table (figures 8-12). 4 7 Figure 8. Spiral tomograms made with 1.2 mm gutta percha markers. Acrual slice locations were added after all observers had completed the slice localization task. 73 I A B c o n Figure 9. Spiral tomograms made with 3.2 mm gutta percha markers. 29 r Figure 11. Spiral tomograms made with 1.6 m m ball bearing markers. Figure 12. Spiral tomograms made with 5.9 mm ball bearing markers. An alignment cross-hatching mark was placed above the sagittal image on each film, and acetate tracing film (Cephalometric Tracing Paper, G A C International, Inc., Central Islip N Y , USA) attached to the image with the alignment marks duplicated on the acetate. The observers, who consisted of two retired dentists and a certified dental laboratory technician, were given verbal instructions to draw vertical lines on the acetate over the sagittal image and label them according to where they judged each cross-sectional image to have been taken (figures 13 and 14). A clrafting triangle was aligned with the bottom edge of the film to ensure that the vertical lines were perpendicular to bottom of the film. The observers knew that the slices were spaced 4 millimetres apart, but did not know the order of the images. The observers were provided with a 1.26 magnified scale ruler to correspond with the magnification of the tomographic images. Figure 14. Close-up view of observer's markings. 32 The distance of each marked line from the centre of the sagittal image was measured using a 1.26 magnified scale ruler, and the films were re-sorted. The distance of the marked line for each cross-sectional image from its true slice location was calculated based on the known slice order. The data were entered into Microsoft Excel, and imported into SYSTAT 11.0 (Systat Software Inc., Point Richmond CA, USA) for analysis. The data were analyzed with both inclusion and exclusion of the images made with no markers. 3.3 Results Significant differences were found using A N O V A for markers (p < 0.001), and marker material (p <0.001), and sides of the mandible (p = 0.004), but not for the slice location (p = 0.913) or observer (p = 0.885). The dependent variable for all statistical calculations was the absolute value of the error. Using the Bonferroni adjustment, significant differences were found in mean absolute error for the different sides of the mandible (p = 0.004) but not for observers. Mean absolute error was significantly worse for no marker versus any type of marker (p < 0.001). Mean absolute error for gutta percha was significantly worse than for amalgam (p < 0.001) or brass tubing (p = 0.002). 33 O u u G 9 8 7 6 5 4 3 2 1 0 1: Amalgam 1.2 2: Amalgam 1.6 3: Amalgam 2.0 4: Amalgam 2.4 5: Amalgam 3.2 6: Gutta Percha 1.2 7: Gutta Percha 1.6 8: Gutta Percha 2.0 9: Gutta Percha 2.4 10: Gutta Percha 3.2 j l iL 11: Brass Tubing 1.6 12: Brass Tubing 2.4 13: BaU Bearing 1.6 14: Ball Bearing 2.4 15: Ball Bearing 3.5 16: Ball Bearing 5.6 17: No Marker nfi I i f 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 Marker Figure 15. Pilot study marker versus mean error graph. Error bars indicate standard error. 34 8 7 IT 6 ^ 5 u § 4 W § 3 S 2 1 J 0 1: Amalgam 2: Gutta Percha 3: Brass Tubing 4: Ball Bearing 5: No Marker 1 2 3 4 5 Marker Material Figure 16. Pilot study marker material versus mean error graph. Error bars indicate standard error. Variable Degrees of Freedom F-ratio Probability Side 1 8.442 0.004 Observer 2 0.123 0.885 Slice Location 4 0.244 0.913 Marker 16 20.052 < 0.001 Marker Material 4 64.432 < 0.001 Table 1. A N O V A results for pilot study. 35 Material Amalgam Gutta Percha Brass Tubing Ball Bearing Gutta Percha < 0.001 Brass Tubing 1.000 0.002 Ball Bearing 1.000 0.190 0.655 No Marker < 0.001 < 0.001 < 0.001 <0.001 Table 2. Bonferroni probability values for marker materials. For four markers, the second smallest amalgam marker, the largest amalgam marker, the largest gutta percha marker, and the smaller brass tubing marker, no errors were made in slice localization. Minimal errors were made for the smallest amalgam marker, the larger brass tubing marker, and the largest ball bearing. The greatest errors were made with no marker, and of the markers, the smallest gutta percha marker had the greatest errors, followed by the smallest ball bearing and the third smallest gutta percha marker. Using the Bonferroni correction, the difference in mean absolute error between no marker and any type of marker was highly significandy different (p < 0.001). Significant differences were found also found for mean absolute error between amalgam and gutta percha (p < 0.001) and between gutta percha and brass tubing (p < 0.002). 36 3.4 Discussion Clearly any marker provided better accuracy in slice localization than was possible without a marker. The use of different marker materials also appears to affect observers' ability to localize slices since significant differences were found between some of the different types of markers. The fact that statistically significant differences were found using such a small sample size was evidence that the design of the study was suitable for a larger scale study. Accuracy of slice localization is only one determinant in the choice of a stent marker material. Use of a marker with a long axis is recommended by many authors to ensure that the correct angulation of the implant can be achieved using the stent as a surgical guide. Many surgeons prefer a larger hole in the stent to enable the pilot bur to pass through it without touching the acrylic, and thus risking introduction of acrylic shavings into the surgical site. If the stent is to be used as a surgical guide, and the marker is in the guide hole, then the marker must be removed prior to the surgery. Amalgam is much more difficult to remove than materials such as gutta percha and temporary filling material. Brass tubing markers produced a high level of accuracy in slice localization, provide a long axis indication, and are easily removed for use of the stent in surgery. 37 CHAPTER 4 DEFINITIVE STUDY The definitive study was very similar in design to the pilot study. The pilot study was able to produce the type of analysis planned, and the method of analysis proved satisfactory. Since the pilot study had clearly demonstrated that any type of marker evaluated was greatly superior to no marker at all it was decided to eliminate the no marker control from the definitive study. Because the observers would have many more image sets to evaluate, it was decided to reduce the number of types and sizes of markers evaluated to reduce observer fatigue. To try to determine whether the observers were using anatomical cues from the mandibles to aid in the determination of slice position, a metatarsal bone was included in the definitive study to act as a control. This bone is similar in size and density in cross-section to an edentulous mandible, and would fit in the CommCat machine readily for imaging. Unlike a mandible, however, all of the cross-sectional images appear very similar to one another without the confounding factor of the inferior alveolar nerve canal that may be located at different heights on different views, or the presence of the mandibular ramus that may be seen on posterior cross-sectional views of the mandible. Although there are a few recommendations for use of amalgam as a marker material, it seems that few clinicians are utilizing it for this purpose, possibly because of the difficulty in removing it before conversion to a radiographic 38 guide to be converted to a surgical guide. It therefore was also decided to eliminate the amalgam markers from the definitive study. Three sizes of each of the markers (gutta percha, brass tubing and ball bearings), small, medium and large were used. However, the diameters of the markers could not be matched for each of the materials. Although it may have been preferable to use markers of the same diameters, the commercial availability of sizes of brass tubing and the stainless steel ball bearings was limited. Also, although it is unlikely that a clinician would use very thin brass tubing, they often do use very thin gutta percha markers such as a single gutta percha point. 4.1 Goals The null hypotheses were: •There would be no difference in the ability of observers to localize cross-sectional conventional spiral tomography slice locations on a corresponding longitudinal tomographic image for different stent markers. • There would be no difference in the ability of observers to localize cross-sectional conventional spiral tomography slice locations on a corresponding longitudinal tomographic image for different slice locations. •There would be no difference in the ability of observers to localize cross-sectional conventional spiral tomography slice locations on a corresponding longitudinal tomographic image for stents made on different mandibles or for one made on a metatarsal bone. •There would be no ixiterobserver variation in ability to localize cross-sectional conventional spiral tomography slice locations on a corresponding longitudinal tomographic image. 39 4.2 Materials and Methods Seven acrylic stents (Triad baseplate material, 3M) were made for five mandibles and one metatarsal bone. One mandible had bilateral stents fabricated. Two sites were prepared on each stent to accept 6.35 millimetre cylindrical acrylic rods. The spacing of the two markers varied slightly between different stents. Eighteen acrylic rods were fashioned, nine each of heights 7.5 millimetres and 5.8 millimetres. For all stents, the longer marker was positioned anteriorly and the shorter marker posteriorly to simulate the situation typically found in patients, where the clearance to the opposing arch is less posteriorly. The materials used (figure 17) were: -gutta percha in diameters of 1.2, 2.0, and 3.2 millimetres -brass tubing in outer diameters of 2.4, 3.2, and 4.0 millimetres -stainless steel ball bearings in diameters of 1.6, 3.5, and 5.9 millimetres 40 Figure 17. Markers used for definitive study. A CommCat (Imaging Sciences International, Hatfield PA, USA) dedicated dental complex-motion tomographic unit was used to take the tomographic images using Fuji Super HRG30 film (Fuji Photo Film Company, Tokyo, Japan) with Kodak Ektavision (Kodak Canada Inc., Toronto O N , Canada) intensifying screens. The films were processed with fresh Nature Care solutions (Patterson Dental Canada Inc., Montreal QB, Canada) in an A T 2000 X R processor (Air Techniques, Hicksville N Y , USA). Imaging Sciences International custom fabricated a support table for the mandible which could be substituted for the patient chin rest. This had four holes drilled into it, and a polyvinyl siloxane (Exaflex Type O putty, G C America Inc., Alsip IL, USA) positioning jig was made which fitted into the holes while stabilizing the mandible with its occlusal plane or the superior 41 surface of the metatarsal bone horizontal. This was done to try to ensure that the acrylic cylinders and markers with a long axis would be perpendicular to the floor, and therefore in the same vertical plane as the cross-sectional images. If the markers were not perpendicular to the floor, then the top and bottom of a long marker would not be in the same cross-sectional tomographic image. Prior to making any images for the study, the calibration protocol provided by Imaging Sciences Int. for the CommCat was performed to ensure the machine was performing as accurately as possible. A n occlusal radiograph of each mandible and a similar view of the metatarsal bone were obtained with small ball bearing markers in place on the stents (figure 18 and 19). The occlusal radiographs were then scanned into the CommCat control computer to map the shape of the arch and select the slice locations to be imaged. Figure 18. Occlusal radiograph of one of the mandibles. Figure 19. Occlusal-type radiograph taken o f metatarsal bone. 43 Calibration films consisting of one corrected sagittal image and one cross-sectional image were taken and repeated until correct slice locations and exposure parameters were achieved. It was necessary to add copper filtration to the CommCat tube head to obtain satisfactory images. Study films were then produced consisting of one corrected sagittal and 5 cross-sectional images per stent-marker combination. For each series, the markers of the same type were used, with the shorter marker placed more posteriorly. The same 5 slice locations were used for all the series made on each stent, with a spacing of 4 millimetres between slices. The order for the images to be taken for each stent-marker combination was randomized by producing a list of the 120 possible different orders and numbering them. A random number table was generated using Microsoft Excel. The images were produced such that, when viewed, each image set for the mandibles would appear to be the right side of the mandible, with the corrected sagittal tomographic image first and the cross-sectional tomographic images following. Letters A-E were marked over the cross-sectional images and the films randomized in order using a random number table generated with Microsoft Excel and the films renumbered accordingly. An alignment cross-hatching mark was placed above the sagittal image on each film, and acetate tracing film (Cephalometric Tracing Paper, GAC International, Inc., Central Islip NY, USA) attached to the image with the alignment marks duplicated on the acetate. 44 Ten senior dental students were recruited as observers. A n orientation session was conducted to introduce the observers to tomographic imaging and to explain the protocol to use. Each observer received a package containing a specialized ruler with a magnified millimetre scale correlating to the 1.26 magnification of the tomographic images, a 0.5 mm mechanical pencil, a drafting triangle, and an instruction sheet (figure 20). The observers were instructed to draw vertical lines on the acetate over the sagittal images and label these lines according to where they judged each cross-sectional image to have been taken. The drafting triangle was used to align the drawn lines with the bottom edge of the film to ensure that the vertical lines were perpendicular to bottom of the film. The observers were informed that the slices were spaced 4 millimetres apart, but did not know the order of the images. The film sets were distributed amongst the observers with instructions to return each film set with the marked lines. As each set was returned, the distance of each of the marked lines from the centre of the sagittal image was recorded to the nearest millimetre and then the acetate was removed and a new acetate placed for the next observer. Seven of the observers were able to complete all of the image sets. Results from observers who did not complete all of the data sets were not included in the analysis. After all of the data sets were completed, the distance of each marked slice from the centre of the image was then compared to the actual slice location using the randomization key, and the difference of the marked position to the true position recorded to the nearest millimetre. 45 The data were entered into Microsoft Excel, and imported into SYSTAT 11.0 (Systat Software Inc., Point Richmond CA, USA) for analysis. Observer Instructions 1. Please legibly initial the tracing film. 2. The cross-sectional images are spaced at 4 mm intervals. Because the images are magnified, you should use the magnified scale on the enclosed ruler to determine the spacing of your indicator lines. 3. Once you have determined the location for each of the slices, please use the enclosed triangle to draw the lines perpendicular to the bottom of the film. 4. Write the letter for each cross-sectional view above the corresponding line you drew on the sagittal view. 5. Use the enclosed checklist to mark off which films you have completed. 6. Once you have completed a set of films, please return them to me to be redistributed. Figure 20. Instructions to observers. 46 CHAPTER 5 RESULTS The observer errors were initially recorded as positive or negative to indicate the direction away from the correct position (negative to the right of the correct location and positive to the left of the correct position). The mean errors cancelled each other out, resulting in mean errors of 0, and it was therefore decided to use the absolute value of the error. 5.1 Interobserver Variability Using A N O V A , there were significant differences in the performances of the different observers (p = 0.004, table 3). The most accurate observer was the slowest to complete the task. With the Bonferroni correction observer 7 was significantly better than observer 5 (p = 0.016, table 4). 1.8 1.6 "FT 1.4 ¥ i W 0.8 H I 0.6 ^ 0.4 0.2 0 3 4 5 Observer 7 Figure 21. Observer versus mean error graph. Error bars indicate standard error. 4 7 Degrees of Freedom F-Ratio Probability 6 3.149 0.004 Table 3. A N O V A results for observer. Observer 1 2 3 4 5 6 2 1.000 3 1.000 0.630 4 1.000 1.000 1.000 5 1.000 0.206 1.000 1.000 6 1.000 1.000 1.000 1.000 0.483 7 0.433 1.000 0.066 0.194 0.016 1.000 Table 4. Bonferroni probability values for observer. 5.2 Location of the Tomographic Slice Surprisingly, with A N O V A there was a significant difference in the observers localization of the cross-sectional images depending on their location (p < 0.001, table 5), which may have been due to some anatomical feature — the most posterior slice was most accurately located, and for this view the ramus of the mandible may have been partially visible for the images made using mandibles. The second slice, that corresponded to a slice made through the more anterior marker, was the next most accurately localized. This marker was the longer of the two, or, in the case of the ball-bearings, was positioned higher. The greater height of this marker may have made it easier to distinguish than 48 the lower marker that may have been partially obscured by ghosting from the. higher marker and therefore easy to confuse with the slice made between both markers. The most anterior slice was most the most inaccurately identified. With the Bonferroni correction, it was shown that the observers' localization of the most anterior slice was significantly worse than for any of the other slice locations (table 6). V-l O 2.5 2 J 1.5 0.5 0 1 2 3 4 • 5 Location of Tomographic Slice Figure 22. Location of tomographic slice versus mean error graph. Error bars indicate standard error. Degrees of Freedom F-Ratio Probability 4 9.835 < 0.001 Table 5. A N O V A results for location of tomographic slice. 49 Slice Location 1 2 3 4 2 < 0.001 3 0.012 1.000 4 0.020 0.919 1.000 5 < 0.001 1.000 0.085 0.056 Table 6. Bonferroni probability values for location of tomographic slice. 5.3 Size and Material of the Marker Since only two types of markers shared a size, the effect of size was considered with the material to obtain results for comparison of each of the different size/shape combinations. There was no significant difference between the observers' localization of cross-sectional images using the largest size of gutta percha marker and any of three sizes of brass tubing. The ability of the observers to localize the cross-sectional images on the corresponding sagittal images was good with all of the brass tubing markers and the largest gutta percha marker (figure 23). 50 VI O vi w a <u 2.5 n 1.5 0.5 0 1: Gutta Percha 1.2 2: Gutta Percha 2.0 3: Gutta Percha 3.2 4: Brass Tubing 2.4 5: Brass Tubing 3.2 6: Brass Tubing 4.0 1 2 3 4 5 Marker 7: Ball Bearing 1.6 8: Ball Bearing 3.5 9: Ball Bearing 4.0 7 8 Figure 23. Marker versus mean error graph. Error bars indicate standard error. Degrees of Freedom F-ratio Probability 8 11.561 < 0.001 Table 7. A N O V A results for marker type. 51 Marke r T y p e G.P. 1.2 G.P. 2.0 G.P. 3.2 B.T. 2.4 B.T. 3.2 B.T. 4.0 B.B. 1.6 B.B. 3.5 G.P. 2.0 1.000 G.P. 3.2 <0.001 <0.001 B.T. 2.4 <0.001 <0.001 1.000 B.T. 3.2 <0.001 <0.001 1.000 1.000 B.T 4.0 <0.001 <0.001 1.000 1.000 1.000 B.B. 1.6 0.257 0.854 0.056 0.040 0.276 0.007 B.B. 3.5 0.031 0.133 0.418 0.318 1.000 0.077 1.000 B.B. 5.9 0.207 0.707 0.071 0.051 0.341 0.010 1.000 1.000 Table 8. Bonferroni probability values for marker type. 5.4 M a r k e r Mate r ia l When only the marker material was considered without the influence of size, A N O V A demonstrated significant differences were present between materials (p < 0.001). With the Bonferroni correction, the observers' localizations using gutta percha and ball bearings were not significandy different from one another (p = 0.457) but were significandy better using brass tubing was (p < 0.001). 52 1.8 1.6 o 1 pq 0.8 | 0.6 ^ 0.4 0.2 0 1: Gutta Percha 2: Brass Tubing 3: Ball Bearing Marker Material Figure 24. Stent marker material versus mean error graph. Error bars indicate standard error. Degrees of Freedom F-Ratio Probabilty 2 23.684 < o ; o o i Table 9. A N O V A results for marker material 53 M a r k e r Ma te r i a l Gutta Percha Brass Tubing Brass Tubing < 0.001 Ball Bearing 0.475 < 0.001 Table 10. Bonferroni probability values for marker material 5.5 Stent Using A N O V A , there was significant difference in the observers ability to localize slices for the different stents (p < 0.001). With the Bonferroni correction, the slice localizations made using the stent for the metatarsal bone were significandy more accurate (p < 0.001 to p = 0.040) than for all of the other stents except for the third mandible stent (p = 1.0). The localizations for the third mandibular stent were significandy more accurate than for the second (p < 0.001) and fifth (p = 0.018) mandibular stents. The localizations for the second mandibular stent were also significandy less accurate than for the fourth (p = 0.016) and sixth (p = 0.050) mandibular stents. 54 O VI V rxi a CCS <U 2.5 2 1.5 1 0.5 0 2 3 4 5 Stent 7 Figure 25. Stent versus mean error graph. Error bars indicate standard error. Degrees of Freedom F-Ratio Probability 6 9.319 < 0.001 Table 11. A N O V A results for stent. 5 5 Stent 1 2 3 4 5 6 2 0.033 3 0.161 < 0.001 4 1.000 0.016 0.293 5 1.000 0.261 0.018 1.000 6 1.000 0.050 0.110 1.000 1.000 7 0.020 < 0.001 1.000 0.040 0.001 0.012 Table 12. Bonferroni probability values for stent. 56 CHAPTER 6 6.1 DISCUSSION Statistical analysis of the results of the study allowed all of the null hypotheses to be rejected. There were significant differences between the mean absolute error for localization for the different markers whether they were considered independently or for material only. The difference in the ability of the observers to localize slices based on their position was unexpected. The differences in the mean error for different stents may reflect both differences in the spacing of the markers for the stents, as well as differences in the underlying bone. Although the observers were not supposed to use the anatomical features to aid in slice localization, the greatest accuracy in slice location was found for images made on the metatarsal bone, which did not provide anatomical cues to aid in slice localization. This suggests that although the observers were to make the estimates of slice localization on all images based solely on the imaged markers, they in fact did attempt to use anatomical cues to guide them, but did so incorrectly. This cannot be attributed solely to the inexperience of the observers as this was true for all observers and one of the observers was a surgeon with familiarity with tomographic images (a foreign-ttained dentist taking the International Dental Degree Completion Program). Although one observer was significantly more accurate than another using the Bonferroni correction for the A N O V A analysis, this more accurate observer was not the surgeon. The most accurate observer did take much longer to complete the task than any other observer, and the amount of time spent by this observer may not be reflective of actual clinical practice. 57 There were differences between the accuracy of localization of cross-sectional slices for different types of markers. The mean magnitudes of error were quite small. The brass tubing markers and the largest gutta percha marker had mean errors of less than one millimetre while the ball bearing markers had mean , errors close to one and a half millimetres, and the smaller gutta percha markers had mean errors of nearly two millimetres. The clinical significance of errors of this magnitude warrants further investigation. From consultation with prosthodontists and implant surgeons, it was indicated that errors of two millimetres could be clinically significant although data to support this view was not identified. Although the mean errors of magnitude were quite small, individual errors of up to twenty-one millimetres were made. This may not occur in practice, however, where the order of the slices would be known and a systematic error in which the practitioner incorrecdy identified all of the slices to be either anterior or posterior to the actual locations would be more likely to occur. The importance of this study is that although there are many authors recommending the use of radiographic stents in implant planning and surgery, no study had attempted to correlate the type of stent marker used to its accuracy in localizing the slice. A n in vitro study by Choi et al (Choi, M . et al. 2004) was able to demonstrate that the dimensions of guide holes of a surgical stent influences the accuracy of implant placement. In particular, length of the guide channel was shown to be the most important factor in controlling the implant placement. However, no quantitative studies for the effect of markers on the accuracy of imaging could be identified. 58 The ability of the surgeon to relate the cross-sectional images to the patient's anatomy is important to validate the use of imaging in implant planning. This may be less of an issue for imaging with computed tomography, which may become the method of choice for all implant imaging since the introduction of cone-beam computed tomography. Although the image quality is superior with cone-beam computed tomography compared to conventional tomography, radiation dose and cost to the patient remains less with conventional tomography for imaging of one to two adjacent sites. Access to conventional tomography will likely remain better than that for computed tomography in many geographic locations, and therefore the results of this study are of importance. Further research into the clinical significance of errors in slice localization may help to clarify the importance of the findings of this study. With the introduction of dental cone-beam CT scanners it is anticipated that these will become the preferred method for cross-sectional imaging for dental implant planning in the majority of cases. There is a similar lack of published data comparing the efficacy of different stent marker materials for computed tomography as for conventional tomography. As such, it would be useful to perform a similar study to this one but with the use of cone-beam computed tomography in place of conventional spiral tomography. 59 6.2 CONCLUSIONS -The mean error of slice localization was affected by different marker materials. The magnitude of such errors could be clinically significant and may warrant further investigation. -The anatomy of the underlying bone may have influenced the observers. -The differences due to interobserver variability were small. The only significant interobserver variability found was between two observers, the more accurate of whom spent much more time completing the task than would be expected in clinical practice. 60 Bibliography Endosseous implants and hemorrhage. J.Am.Dent.Assoc. 122[2], 16,18. 1991. Adrian, E . D. , Ivanhoe, J. R., and Krantz, W. A. Trajectory surgical guide stent for implant placement. J.Prosthet.Dent. 67[5], 687-691. 1992. Akca, K . , Iplikcioglu, H . , and Cehreli, M . C. A surgical guide for accurate mesiodistal paralleling of implants in the posterior edentulous mandible. J.ProsthetDent. 87[2], 233-235. 2002. Almog, D . M . , Torrado, E. , and Meitner, S. W. Fabrication of imaging and surgical guides for dental implants. J.ProsthetDent. 85[5], 504-508. 2001. Almog, D . M . , Torrado, E. , Moss, M . E. , Meitner, S. W., and LaMar, F. Use of imaging guides in preimplant tomography. Oral Surg. Oral Med. Oral Pathol. Oral RadioLEndod. 93[4], 483-487. 2002. Arlin M L . Optimal placement of osseointegrated implants. J.Can.DentAssoc. 56[9], 873-876. 2005. Balshi, T. J. and Garver, D . G . Surgical guidestents for placement of implants. J.Oral Maxillofac.Surg. 45[5], 463-465. 1987. Bartling, R., Freeman, K . , and Kraut, R. A. The incidence of altered sensation of the mental nerve after mandibular implant placement. J.Oral Maxillofac.Surg. 57 [12], 1408-1412. 1999. Becker, C. M . and Kaiser, D . A. Surgical guide for dental implant placement. J.ProsthetDent. 83[2], 248-251. 2000. Besimo, C , Lambrecht, J. T., and Nidecker, A. Dental implant treatment planning with reformatted computed tomography. Dentomaxillofac.Radiol. 24[4], 264-267. 1995. Besimo, C. E. , Lambrecht, J. T., and Guindy, J. S. Accuracy of implant treatment planning utilizing template-guided reformatted computed tomography. Dentomaxillofac.Radiol. 29[1], 46-51. 2000. 61 Bolin, A. , Eliasson, S., von Beetzen, M . , and Jansson, L. Radiographic evaluation of mandibular posterior implant sites: correlation between panoramic and tomographic determinations. Clin.Oral Implants.Res. 7[4], 354-359. 1996. Borrow, J. W. and Smith, J. P. Stent marker materials for computerized tomograph-assisted implant planning. Int.J.Periodontics.Restorative.Dent. 16[1], 60-67. 1996. Bou, S. C , Jacobs, R , Gijbels, F., Bosmans, H . , Hermans, R , Quirynen, M . , and van Steenberghe, D . Absorbed doses from spiral CT and conventional spiral tomography: a phantom vs. cadaver study. Clin.Oral Implants.Res. 12, 473-478. 2001. Boyes-Varley, J. G . and Lownie, J. F. Haematoma of the floor of the mouth following implant placement. SADJ. 57 [2], 64-65. 2002. Burns, D. R , Crabtree, D . G. , and Bell, D . H . Template for positioning and angulation of intraosseous implants. J.Prosthet.Dent. 60[4], 479-483. 1988. Butterfield, K . J., Dagenais, M . , and Clokie, C. Linear tomography's clinical accuracy and validity for presurgical dental implant analysis. Oral Surg.Oral Med.Oral Pathol.Oral Radiol.Endod. 84[2], 203-209. 1997. Cehreli, M . C , Asian, Y. , and Sahin, S. Bilaminar dual-purpose stent for placement of dental implants. J.ProsthetDent. 84[1], 55-58. 2000. Cehreli, M . C , Calis, A. C , and Sahin, S. A dual-purpose guide for optimum placement of dental implants. J.Prosthet.Dent. 88[6], 640-643. 2002. Cehreli, M . C. and Sahin, S. Fabrication of a dual-purpose surgical template for correct labiopalatal positioning of dental implants. IntJ.Oral Maxillofac.Implants. 15[2], 278-282. 2000. Chaushu, G. , Taicher, S., Halamish-Shani, T., and Givol, N . Medicolegal aspects of altered sensation following implant placement in the mandible. IntJ.Oral Maxillofac.Implants. 17[3], 413-415. 2002. Choi, M . , Romberg, E. , and Driscoll, C. F. Effects of varied dimensions of surgical guides on implant angulations. J.Prosthet.Dent. 92[5], 463-469. 2004. Choi, S. C , Ann, C. H. , Choi, H . M . , Heo, M . S., and Lee, S. S. Accuracy of reformatted CT image for measuring the pre-implant site: analysis of the image distortion related to the gantry angle change. Dentomaxillofac.Radiol. 31 [4], 273-277. 2002. 62 Dantas, J. A . , Montebello Filho, A. , and Campos, P. S. F. Computed tomography for dental implants: the influence of the gantry angle and mandibular positioning on the bone height and width. Dentomaxillofacial Radiology 34[1], 9-15. 1-1-2005. Dao, T. T. and Mellor, A . Sensory disturbances associated with implant surgery. IntJ.Prosthodont. 11 [5], 462-469. 1998. Darriba, M . A . and Mendonca-Caridad, J. J. Profuse bleeding and hfe-mreatening airway obstruction after placement of mandibular dental implants. J.Oral Maxillofac.Surg. 55[11], 1328-1330. 1997. DelBalso, A . M . , Greiner, F. G. , and Licata, M . Role of diagnostic imaging in evaluation of the dental implant patient. Radiographics 14[4], 699-719. 1994. Dula, K . , Mini, R., Lambrecht, J. T., van der Stelt, P. F., Schneeberger, P., Clemens, G. , Sanderink, H . , and Buser, D . Hypothetical mortality risk associated with spiral tomography of the maxilla and mandible prior to endosseous implant treatment. Eur.J.Oral Sci. 105[2], 123-129. 1997. Dula, K . , Mini, R., van der Stelt, P. F., and Buser, D . The radiographic assessment of implant patients: decision-making criteria. IntJ.Oral Maxillofac.Implants. 16[1], 80-89. 2001. Edge, M . J. Surgical placement guide for use with osseointegrated implants. J.Prosthet.Dent. 57[6], 719-722. 1987. Edge, M . J. and Medina, T. V . Surgical location guide for use with osseointegrating implants. Mil.Med. 153[11], 585-586. 1988. Ekestubbe, A. , Thilander, A. , and Grondahl, H . G . Absorbed doses and energy imparted from tomography for dental implant installation. Spiral tomography using the Scanora technique compared with hypocycloidal tomography. Dentomaxillofac.Radiol. 21 [2], 65-69. 1992. Ekestubbe, A. , Thilander, A. , Grondahl, K . , and Grondahl, H . G . Absorbed doses from computed tomography for dental implant surgery: comparison with conventional tomography. Dentomaxillofac.Radiol. 22[1], 13-17. 1993. Ellies, L. G . Altered sensation following mandibular implant surgery: a retrospective study. J.ProsthetDent. 68[4], 664-671. 1992. Engelman, M . J., Sorensen, J. A . , and Moy, P. Optimum placement of osseointegrated implants. J.ProsthetDent. 59[4], 467-473. 1988. 63 Floyd, P. P. P. and Palmer R. Radiographic techniques. Br.DentJ. 187(7], 359-365. 1999. Fortin, T., Champleboux, G. , Bianchi, S., Buatois, H . , and Coudert, J. L . Precision of transfer of preoperative planning for oral implants based on cone-beam CT-scan images through a robotic drilling machine. Clin.Oral Implants.Res. 13[6], 651-656. 2002. Frederiksen, N . L. Diagnostic imaging in dental implantology. Oral Surg.Oral Med.Oral Pathol.Oral Radiol.Endod. 80[5], 540-554. 1995. Frei, C , Buser, D. , and Dula, K . Study on the necessity for cross-section imaging of the posterior mandible for treatment planning of standard cases in implant dentistry. Clin.Oral Implants.Res. 15[4], 490-497. 2004. Friedland, B. Implant radiography. J.Mass DentSoc. 51 [4], 14-17. 2003. Friedland, B. and Valachovic, R. W. Advanced imaging techniques assist in implant planning. J.Mass DentSoc. 45[4], 44-50. 1997. Givol, N . , Chaushu, G. , Halamish-Shani, T., and Taicher, S. Emergency tracheostomy following Hfe-direatening hemorrhage in the floor of the mouth during immediate implant placement in the mandibular canine region. J.Periodontol. 71 [12], 1893-1895. 2000. Givol, N . , Taicher, S., Halamish-Shani, T., and Chaushu, G . Risk management aspects of implant dentistry. IntJ.Oral Maxillofac.Implants. 17[2], 258-262. 2002. Gray, C. F., Redpath, T. W., Bainton, R., and Smith, F. W. Magnetic resonance imaging assessment of a sinus lift operation using reoxidised cellulose (Surgicel) as graft material. Clin.Oral Implants.Res. 12[5], 526-530. 2001. Gray, C. F., Redpath, T. W., and Smith, F. W. Pre-surgical dental implant assessment by magnetic resonance imaging. J.Oral Implantol. 22[2], 147-153. 1996. Gray, C. F., Redpath, T. W., and Smith, F. W. Low-field magnetic resonance imaging for implant dentistry. Dentomaxillofac.Radiol. 27[4], 225-229. 1998a. Gray, C. F., Redpath, T. W., and Smith, F. W. Magnetic resonance imaging: a useful tool for evaluation of bone prior to implant surgery. Br.DentJ. 184[12], 603-607. 6-27-1998b. 64 Gray, C F., Redpath, T. W., Smith, F. W., and Staff, R. T. Advanced imaging: Magnetic resonance imaging in implant dentistry. Clin.Oral Implants.Res. 14[1], 18-27. 2003. Gregg, J. M . Neuropathic complications of mandibular implant surgery: review and case presentations. Ann.R.Australas.Coll.Dent.Surg. 15,176-180. 2000. Harris, D. , Buser, D. , Dula, K . , Grondahl, K . , Hans, D. , Jacobs, R., Lekholm, U . , Nakielny, R., van Steenberghe, D. , and van der, S. P. E .A.O. guidelines fo the use of diagnostic imaging in implant dentistry. A consensus workshop organized by the European Association for Osseointegration in Trinity College Dublin. Clin.Oral Implants.Res. 13[5], 566-570. 2002. Harris, J. M . Trigeminal nerve injuries. J.Am.Dent.Assoc. 133[12], 1608. 2002. Iplikcioglu, H . , Akca, K . , and Cehreli, M . C. The use of computerized tomography for diagnosis and treatment planning in implant dentistry. J.Oral Implantol. 28[1], 29-36. 2002. Isaacson, T. J. Sublingual hematoma formation during immediate placement of mandibular endosseous implants. J.Am.Dent.Assoc. 135[2], 168-172. 2004. Ismail, Y . H . , Azarbal, M . , and Kapa, S. F. Conventional linear tomography: protocol for assessing endosseous implant sites. J.ProsthetDent. 73[2], 153-157. 1995. Israelson, H . , Plemons, J. M . , Watkins, P., and Sory, C. Barium-coated surgical stents and computer-assisted tomography in the preoperative assessment of dental implant patients. Int.J.Periodontics.Restorative.Dent. 12[1], 52-61. 1992. Johnson, C. M . , Lewandowski, J. A. , and McKinney, J. F. A surgical template for aligned placement of the osseointegrated implant. J.Prosthet.Dent. 59[6], 684-688. 1988. Kalpidis, C. D . and Setayesh, R. M . Hemorrhaging associated with endosseous implant placement in the anterior mandible: a review of the literature. J.Periodontol. 75[5], 631-645. 2004. Kassebaum, D . K . , Nummikoski, P. V. , Triplett, R. G. , and Langlais, R. P. Cross-sectional radiography for implant site assessment. Oral Surg.Oral Med.Oral Pathol. 70[5], 674-678. 1990. Kassebaum, D. K . , Stoller, N . H . , and Goshorn, B. I. Radiographic techniques for presurgical assessment of dental implant sites. Gen.Dent. 40[6], 502-510. 1992. 65 Kobayashi, K . , Sliirnoda, S., Nakagawa, Y. , and Yamamoto, A. Accuracy in measurement of distance using limited cone-beam computerized tomography. IntJ.Oral Maxrllofac.Implants. 19[2], 228-231. 2004. Kohavi, D. , Bar-Ziv, J., and Marmary, Y. Effect of axial plane deviation on cross-sectional height in reformatted computed tomgraphy of the mandible. Dentomaxillofac.Radiol. 26[3], 189-191. 1997. Kopp, K . C , Koslow, A. H . , and Abdo, O. S. Predictable implant placement with a diagnostic/surgical template and advanced radiographic imaging. J.Prosthet.Dent. 89[6], 611-615. 2003. Kraut, R. A. A case for routine computed tomography imaging of the dental alveolus before implant placement. J.Oral Maxillofac.Surg. 59[1], 64-67. 2001. Kraut, R. A. and Chahal, O. Management of patients with trigeminal nerve injuries after mandibular implant placement. J.Am.Dent.Assoc. 133[10], 1351-1354. 2002. K u , Y . and Shen, Y . F. Fabrication of a radiographic and surgical stent for implants with a vacuum former. J.Prosthet.Dent. 83[2], 252-253. 2000. Lam, E . W., Ruprecht, A. , and Yang, J. Comparison of two-dimensional orthoradially reformatted computed tomography and panoramic radiography for dental implant treatment planning. J.ProsthetDent. 74[1], 42-46. 1995. Lascala, C. A. , Panella, J., and Marques, M . M . Analysis of the accuracy of linear measurements obtained by cone beam computed tomography (CBCT-NewTom). Dentomaxillofac.Radiol. 33[5], 291-294. 2004. Lecomber, A. R , Yoneyama, Y. , Lovelock, D . J., Hosoi, T., and Adams, A. M . Comparison of patient dose from imaging protocols for dental implant planning using conventional radiography and computed tomography. Dentomaxillofac.Radiol. 30[5], 255-259. 2001. Lee, S. J. and Toothaker, R. W. Computed tomographic scan template for maximum accuracy of reformatted CT scan images. J.Prosthodont. 7[4], 261-264. 1998. Lee, S. Y . and Morgano, S. M . A diagnostic stent for endosseous implants to improve conventional tomographic radiographs. J.Prosthet.Dent. 71 [5], 482-485. 1994.. 66 Liang, H . , Tyndall, D . A. , Ludlow, J. B., and Lang, L. A . Cross-sectional presurgical implant imaging using tuned aperture computed tomography (TACT). DentomaxiUofac.Radiol. 28[4], 232-237. 1999. Limrachtamorn, S., Edge, M . J., Gettleman, L., Scheetz, J. P., and Farman, A. G. Array geometry for assessment of mandibular implant position using tuned aperture computed tomography (TACT). Dentomaxillofac.Radiol. 33[1], 25-31. 2004. Lindh, C , Petersson, A. , and Klinge, B. Visualisation of the mandibular canal by different radiographic techniques. Clinical Oral Implants Research 3[2], 90-97. 6-1-1992. Linkow, L. I. and Dorfman, J. D . Implantology in dentistry. A brief historical perspective. N.Y.State Dent.J. 57[6], 31-35. 1991. Ludlow, J. B., vies-Ludlow, L. E. , and Brooks, S. L. Dosimetry of two extraoral direct digital imaging devices: NewTom cone beam CT and Orthophos Plus DS panoramic unit. Dentomaxillofac.Radiol. 32[4], 229-234. 2003. Marmulla, R., Wortche, R., Muhling, J., and Hassfeld, S. Geometric accuracy of the NewTom 9000 Cone Beam CT. Dentomaxillofac.Radiol. 34[1], 28-31. 2005. Miles, D. A. and Van Dis, M . L. Implant radiology. Dent.Clin.North Am. 37[4], 645-668. 1993. Mordenfeld, A. , Andersson, L., and Bergstrom, B. Hemorrhage in the floor of the mouth during implant placement in the edentulous mandible: a case report. IntJ.Oral Maxillofac.Implants. 12[4], 558-561. 1997. Mupparapu, M . Imaging of the jaws for routine implant placement. Oral Surg.Oral Med.Oral Pathol.Oral Radiol.Endod. 93[5], 509-510. 2002. Mupparapu, M . and Singer, S. R. Implant imaging for the dentist. J.Can.Dent.Assoc. 70[1], 32. 2004. Niamtu, J., III. Near-fatal airway obstruction after routine implant placement. Oral Surg.Oral Med.Oral Pathol.Oral Radiol.Endod. 92[6], 597-600. 2001. Pesun, I. J. and Gardner, F. M . Fabrication of a guide for radiographic evaluation and surgical placement of implants. J.Prosthet.Dent. 73[6], 548-552. 1995. Petrikowski, C. G. , Pharoah, M . J., and Schmitt, A . Presurgical radiographic assessment for implants. J.Prosthet.Dent. 61 [1], 59-64. 1989. 67 Pharoah, M . J. Imaging techniques and their clinical significance. IntJ.Prosthodont. 6[2], 176-179. 1993. Pieper, S. P. and Lewis, S. G . A case against routine computed tomography imaging of the dental alveolus before implant placement. J.Oral Maxillofac.Surg. 59[1], 68-70. 2001. Poon, C. K , Barss, T. K . , Murdoch-Kinch, C. A. , Bricker, S. L. , Miles, D . A. , and Van Dis, M . L . Presurgical tomographic assessment for dental implants. 1. A modified imaging technique. Int.J.Oral Maxillofac.Implants. 7[2], 246-250. 1992. Rashedi, B., Tyndall, D. A. , Ludlow, J. B., Chaffee, N . R., and Guckes, A. D . Tuned aperture computed tomography (TACT) for cross-sectional implant site assessment in the posterior mandible. J.Prosthodont. 12[3], 176-186. 2003. Reiskin, A. B. Implant imaging. Status, controversies, and new developments. Dent.Clin.North Am. 42[1], 47-56. 1998. Scaf, G. , Lurie, A . G. , Mosier, K . M . , Kantor, M . L., Ramsby, G . R., and Freedman, M . L . Dosimetry and cost of imaging osseointegrated implants with film-based and computed tomography. Oral Surg.Oral Med.Oral Pathol.Oral Radiol.Endod. 83[1], 41-48. 1997. Schulze, D. , Heiland, M . , Thurmann, H. , and Adam, G. Radiation exposure during midfacial imaging using 4- and 16-slice computed tomography, cone beam computed tomography systems and conventional radiography. Dentomaxillofac.Radiol. 33[2], 83-86. 2004. Serhal, C. B., van Steenberghe, D. , Quirynen, M . , and Jacobs, R. Localisation of the mandibular canal using conventional spiral tomography: a human cadaver study. Clinical Oral Implants Research 12[3], 230-236. 6-1-2001. Sethi, A . Precise site location for implants using CT scans: a technical note. IntJ.Oral Maxillofac.Implants. 8[4], 433-438. 1993. Shahrasbi, A . H . and Hansen, C. A. Surgical oral radiographic guide with a removable component for implant placement. J.Prosthet.Dent. 87[3], 330-332. 2002. Siu, A . S., L i , T. K . , Chu, F. C , Comfort, M . B., and Chow, T. W. The use of lipiodol in spiral tomography for dental implant imaging. Implant.Dent. 12[1], 35-40. 2003. 68 Solow R.A. Simplified radiographic-surgical template for placement of multiple, parallel implants. J.ProsthetDent. 85[1], 26-29. 2001. Stellino, G. , Morgano, S. M . , and Imbelloni, A . A dual-purpose, implant stent made from a provisional fixed partial denture. J.Prosthet.Dent. 74[2], 212-214. 1995. Sykaras, N . and Woody, R. D . Conversion of an implant radiographic template into a surgical template. J.Prosthodont. 10[2], 108-112. 2001. Takeshita, F. and Suetsugu, T. Accurate presurgical determination for implant placement by using computerized tomography scan. J.ProsthetDent. 76[6], 590-591. 1996. Takeshita, F., Tokoshima, T., and Suetsugu, T. A stent for presurgical evaluation of implant placement. J.Prosthet.Dent. 77, 36-38. 1997. Tal, H . and Moses, O. A comparison of panoramic radiography with computed tomography in the planning of implant surgery. Dentomaxillofac.Radiol. 20[1], 40-42. 1991. Taylor, T. D . and Agar, J. R. Twenty years of progress in implant prosthodontics. J.ProsthetDent. 88[1], 89-95. 2002. Thunthy, K . H . Interrelationship between cross-sectional and sagittal imaging in computer-assisted dental implant tomography. Dentomaxillofac.Radiol. 29[2], 65-69. 2000. Todd, A. D. , Gher, M . E. , Quintero, G., and Richardson, A. C. Interpretation of linear and computed tomograms in the assessment of implant recipient sites. J.Periodontol. 64[12], 1243-1249. 1993. Tsai, T. P., Chang, S. H . , Hong, H . H . , and Shen, Y . F. Interim denture used as a guide for presurgical radiographic evaluation of dental implant therapy. J.ProsthetDent. 86[3], 324-325. 2001. Tsuchida, F., Hosoi, T., Imanaka, M . , and Kobayashi, K . A technique for making a diagnostic and surgical template. J.Prosthet.Dent. 91 [4], 395-397. 2004. Tyndall, A . A . and Brooks, S. L. Selection criteria for dental implant site imaging: a position paper of the American Academy of Oral and Maxillofacial radiology. Oral Surg.Oral Med.Oral Pathol.Oral Radiol.Endod. 89[5], 630-637. 2000. 69 Urquiola, J. and Toothaker, R. W. Using lead foil as a radiopaque marker for computerized tomography imaging when implant treatment planning. J.ProsthetDent. 77[2], 227-228. 1997. Varvara, G. , Esposito, P., Franchini, F., Perinetti, G. , and Caputi, S. A positioning device for computed tomography: a clinical report. J.Prosthet.Dent. 89[2], 123-126. 2003. Verde, M . A . and Morgano, S. M . A dual-purpose stent for the implant-supported prosthesis. J.ProsthetDent. 69[3], 276-280. 1993. Verstreken, K. , Van Cleynenbreugel, J., Marchal, G. , Naert, I., Suetens, P., and van Steenberghe, D . Computer-assisted planning of oral implant surgery: a three-dimensional approach. Int.J.Oral Maxillofac.Implants. 11 [6], 806-810. 1996. Walker, M . and Hansen, P. Dual-purpose, radiographic-surgical implant template: fabrication technique. Gen.Dent. 47[2], 206-208. 1999. Wat, P. Y. , Chow, T. W., Luk, H . W., and Comfort, M . B. Precision surgical template for implant placement: a new systematic approach. Clin.Implant.Dent.Relat Res. 4[2], 88-92. 2002. Watzinger, F., Birkfellner, W., Wanschitz, F., Millesi, W., Schopper, C , Sinko, K. , Huber, K . , Bergmann, H . , and Ewers, R. Positioning of dental implants using computer-aided navigation and an optical tracking system: case report and presentation of a new method. J.Craniomaxillofac.Surg. 27[2], 77-81. 1999. Weingart, D . and Duker, J. A tomographic technique for the depiction of atrophied alveolar ridges prior to endosseous implant placement. Dentomaxillofac.Radiol. 22[1], 38-40. 1993. White, S. C , Heslop, E . W., HoUender, L. G. , Mosier, K . M . , Ruprecht, A. , and Shrout, M . K . Parameters of radiologic care: A n official report of the American Academy of Oral and Maxillofacial Radiology. Oral Surg.Oral Med.Oral Pathol.Oral Radiol.Endod. 91[5], 498-511. 2001. Worthington, P. Medicolegal aspects of oral implant surgery. Aust.ProsthodontJ. 9 Suppl, 13-17. 1995. Ziccardi, V . B. and Assael, L. A . Mechanisms of trigeminal nerve injuries. Adas.Oral Maxillofac.Surg.CUn.North Am. 9[2], 1-11. 2001. 70 Appendix MEASUREMENT 1. Resolution Resolution refers to how fine a detail can be detected by the measurement system (Brunette, D. M . 1996). The spatial resolution of the CommCat is reported to be at least 1 millimetre (personal communication with technologist at Imaging Sciences Intl.). The spatial resolution of a radiographic system is often reported as line-pairs per millimetre (lp/mm). Standardized radiographic line pair plates are available to test the spatial resolution of radiographic imaging systems. The line-pair plates are made of dense metal with sets of gaps and bars. One gap and one bar is a line pair. Figure 26. Contact radiograph of line-pair plate (Buckbee-Meers Inc., Minneapolis MI, USA) made with Kodak Ultraspeed (D-speed) Intraoral Film (Eastman Kodak Company, Rochester N Y , USA). 71 Images were produced with a line-pair plate using three different types of motion (circular, spiral and hypocycloidal) at different offsets and with the line-pair plates oriented either horizontally or vertically. This revealed that the resolution for spiral images with no offset was at least 1.4 line-pairs per mm for slice thicknesses of 1, 2, or 3 rnillimetres with no offset and the plate in either horizontal or vertical orientation. Tomograms were also taken for the above slice thicknesses and either vertical or horizontal line-pair plate organization for offsets of the image plane at 1, 2, and 3 mm intervals to either side of the line-pair plate. For comparison purposes, images were also made using the conventional cephalometric attachment of the CommCat using the same film-screen combination and development parameters as the tomographic images. The cephalometric attachment to the CommCat uses an x-ray tube head identical to its tomographic assembly, but the source to film distance is greater. The resulting image is magnified much more for the tomographic image than for the image made with the cephalometric machine. This magnification is reflected in the greater number of pixel columns needed to display the line-pair plate data for tomographic images compared to the image made with the cephalometric attachment. The magnification for the CommCat tomographic system is 1.26. A l l of the images were scanned using an H P Scanjet 6100 c scanner with transparency adapter (Hewlett-Packard USA, Houston T X , USA) then imported into Image J (Rasband, W.S., ImageJ, U . S. National Institutes of Health, Bethesda, Maryland, USA, http://rsb.info.nih.gov/ij/, 1997-2005). A long 7 2 rectangular region of interest was selected to include all 8 line pair sets in the top row. The selected region was converted to text data based on 8 bit shades of gray and imported into Microsoft Excel for analysis. Figure 27. Spiral tomogram of line-pair plate at 1 mm slice thickness. The mean for each row of pixel gray shade values was calculated and plotted as a line graph. 73 Cephalometric Machine c/3 <U J3 > ID PL, G <u 250 240 230 220 210 200 190 180 170 101 201 301 401 Columns of Pixels 501 Figure 28. Graph of top row of line-pair plate using the same film-screen combination and processing parameters as the tomographic images but using a conventional cephalometric unit. Comparison of this graph to the following one demonstrates the greater number of line-pairs that may be resolved with the cephalometric image compared to a tomographic image. 74 Spiral Tomogram at 1 mm Slice Thickness, No Offset, Horizontal Line-Pair Plate 101 201 301 401 501 Columns of Pixels 601 701 Figure 29. Graph of mean pixel values for top row of line-pair plate, spiral tomogram with horizontal line-pair plate. The section of the graph corresponding to the second line pair group was selected and the distance from the top of its right peak (one of the spaces) to the bottom of the right valley (one of the bars) was compared to the difference between the top of its right peak and the base (the metal area between line pair groups) to calculate the contrast ratio as a percentage. The graph can be seen to have a curved shape with the central peaks and valleys located higher than those at either end of die range. For the horizontal line-pair plate tomograms, this may be pardy due to a slight mismatch of the pre and post object collimation resulting in the image being lighter at the edges. The pre-object collimation of the CommCat unit uses four computer-controlled lead 75 shields between the source and the object end of the tube head to control the size of the X-ray beam vertically and horizontally. The post-object collimation uses two computer controlled lead shields located at the lateral edges of the beam to reduce scatter. If the post-object collimation is inadvertently set slightly smaller than the beam size determined by the pre-object collimation, then the primary image will be affected at the edges. Visual examination of the tomograms and the resultant pixel value graphs appear to confirm that this was the case. However, although the effect was less pronounced for tomograms made with the line-pair plate in the vertical orientation, there was still a similar distortion of the graph of the pixel values. Decreased optical density at the edges of the tomographic image may be inherent to complex motion tomograms made with a wide angle of motion and tight collimation. This effect was much less for the circular motion tomograms than for the spiral or hypocycloidal motion tomograms, and was barely evident for the circular tomograms made with the line-pair plate oriented vertically. 76 Circular Tomogram at 2 mm Slice Thickness, No Offset, Vertical Line-Pair Plate 1 101 201 301 401 501 601 701 Columns o f Pixels Figure 30. Graph of mean pixel values for top row of line-pair plate, circular motion tomogram taken, 2 mm slice thickness, and the line-pair plate oriented vertically. The fading of the edges of the image precluded using the first line-pair set to generate the contrast ratio graphs. The original intent had been to use the second line-pair group for the contrast ratios, before analysis of the images revealed that the resolution of the system was so poor that the second line-pair group could not be distinguished with some of the offset tomograms. 77 Spiral T o m o g r a m at 1 m m Slice T h i c k n e s s , N o Offset, H o r i z o n t a l L i n e - P a i r Plate 255 w 250 £ 245 j| 240 T3 235 £ 230 G 225 <a 220 215 210 Contrast Ratio = 78.5% T T T T n T i T r r n T i T i T i T i T r r r i T n T n T n T T T T T T T r n T T Y i T i 1111111111111111111111111 rm 1111111111111111111111111111 1 9 17 25 33 41 49 57 65 73 81 89 97 105 Columns of Pixels Figure 31. Second line-pair set used for contrast ratio. Contrast ratios for the entire top row and two groups of line-pairs for the second row for the image made with the cephalometric attachment to the CommCat were obtained, and these were entered into Microsoft Excel and plotted as a line graph. This gives an illustration of the relationship of contrast ratio to resolution. 78 Contrast Ratio versus Line-Pair Group for Cephalometric Unit 100 i 0-4 0 2 4 6 8 10 Line-Pair Group Figure 32. Graph of contrast ratio plotted against line-pair group for the image of the line-pair plate made with the cephalometric attachment of the CommCat. Line-Pair Group 1 2 3 4 5 6 7 8 9 10 Contrast Ratio 80% 79% 73% 53% 45% 39% 30% 24% 14% 8% Table 13. Contrast ratio values for line-pair plate image made with the cephalometric attachment of the CommCat. 79 Spiral Tomogram at 1 mm Slice Thickness, No Offset, Horizontal Line-Pair Plate 100 -, 0 2 4 6 8 10 Line-Pair Group Figure 33. Graph of contrast ratios for spiral tomograms at 1 mm slice thickness with line-pair plate horizontally oriented. Hypocycloidal Tomogram at 1 mm Slice Thickness, No Offset, Vertical Line-Pair Plate Line-Pair Group Figure 34. Graph of contrast ratios for hypocycloidal tomograms at 1 mm slice thickness with line-pair plate vertically oriented. 80 The regression line equation for each graph of contrast ratio versus line-pair group was solved using Microsoft Excel. Y was substituted with 50 to obtain the line-pair group number for 50% contrast ratio. Using the cephalometric machine equation for example: 50 = -8.6242x + 91.933 x = 4.87 (line-pair group) Once the line-pair group number was obtained, the following equation was used in Microsoft Excel to derive the line-pairs per millimetre for that line-pair group: L p / m m = 2/ (2.3812 - 0.41761 x + 0.036667 x2 + 0.0016604 x3 + 0.000029904 x 4) Where x = line-pair group For the cephalometric machine example, this gives a line-pairs per millimetre of 1.92 at 50% contrast ratio. The equation to derive the relationship of line-pair group to resolution was obtained by imaging the line-pair plate using Kodak D-speed intraoral film (figure 26) with the line-pair plate in contact with the film. A General Electric G E 100 intraoral x-ray machine (General Electric Company, Milwaukee, WI, USA) was used at a source to object distance of one metre. The resultant image was digitized to a high resolution Kodak C D by a commercial laboratory. The image of the line-pair plate was imported into N I H image (public domain program developed at the U.S. National Institutes of Health, available for download at http://rsb.info.nih.gov/nih-image/). N I H image was used to measure the distance of two line-pairs for each line-pair group. Microsoft Excel was used to plot the measurements in millimetres against the line-pair groups, 81 and a fourth order polynomial equation was obtained to describe the curve (figure 35). '0 -! , , , 1 0 5 10 15 20 Line-pair group Figure 35. Graph of measurements of the line-pairs to line-pair group on image made with Kodak D-speed intraoral film. 82 C ci h c c u a I K CJ DH •s. .a I u c 0 Circular Hypocycloidal Spiral Cephalometric Image Type Figure 36. Graph of mean line-pairs per mm for different image types at 50% contrast ratio. 1 Horizontal line-pair plate t§ • Vertical line-pair plate Slice Thickness Figure 37. Graph of line-pairs per mm at 50% contrast ratio for spiral tomography. 83 O •c u Z U o i r , w .a OH I CJ 5 Horizontal line-pair plate Vertical line-pair plate 0 r - f l 1 2 3 Slice Thickness Figure 38. Graph of line-pairs per mm at 50% contrast ratio for hypocycloidal tomography. c u ed e z o u c L P , a i u p-l 00 i (U a 0 I Horizontal line-pair plate I Vertical line-pair plate r-TI Figure 1 2 3 SUce Thickness 39. Graph of line-pairs per mm at 50% contrast ratio for circular tomography. 84 Spiral Hypocycloidal Circular 1 mm slice thickness Horizontal lpp 1.47 1.29 1.25 1 mm slice thickness Vertical lpp 1.88 1.80 1.55 2 mm slice thickness Horizontal lpp 1.53 1.68 1.32 2 mm slice thickness Vertical lpp 1.82 1.99 1.75 3 mm slice thickness Horizontal lpp 1.70 1.42 1.40 3 mm slice thickness Vertical lpp 1.91 1.82 1.71 Table 14. Line-pairs per millimetre at 50% contrast ratio for tomographic images. For the tomograms taken at offsets to either side of the line-pair plate, insufficient line-pair groups were visible to produce contrast ratio graphs. The contrast ratios for the second line pair group using spiral tomography are charted below. The contrast ratios for offsets with a negative number are worse than those for positive offsets of the same millimetre distance. This indicates that the central image plane was slighdy off the line-pair plate. However, since none of the offset values were worse than those for the same image type with no offset, the calibration of the machine must have been accurate to within half a millimetre. 85 1 mm slice 2 mm slice 3 mm slice Horizontal l.p plate No offset 79% . 68% 75% Vertical l.p. plate N o offset 74% 78% 79% Horizontal l.p plate 1 mm offset 45% 63% 77% Vertical l.p. plate 1 mm offset 66% 74% 80% Horizontal l.p. plate -1 mm offset 30% 59% 60% Vertical l.p. plate -1 mm offset 38% 64% 75% Horizontal l.p. plate 2 mm offset 28% 0%. 59% Vertical l.p. plate 2 mm offset 36% 36% 58% Horizontal l.p. plate -2mm offset 0% 0% 25% Vertical l.p. plate -2 mm offset 57% 35% 54% Table 15. Contrast ratios for second line-pair group using spiral tomography. 86 2. P r e c i s i o n Precision requires that repeated measurements of the same quantity are close to each other (Brunette, D . M . 1996). The measurements made by all the observers for the metatarsal bone were within 1 millimetre of the true value. Further, no errors were found for four types of markers used in the pilot study, which used the same measuring system as the definitive study. These observations suggest that the method of measurement is precise for the imaging modality used. 3. A c c u r a c y This refers to how close the measured value is to the true value (Brunette, D . M . 1996). The accuracy of the CommCat imaging system itself is reported to be within 1 millimetre, and may be higher with a properly calibrated machine (personal communication with an Imaging Sciences Intl. technician). The fact that for the pilot study, using the same measurement protocol as the definitive study, no errors from the known location of the slices were found using four of the different marker materials suggests that the measurement method was accurate. The finding for the resolution study that the negative offset contrast ratios were worse than for the positive value offsets of the same magnitude shows that the image plane was not perfectly centred on the line-pair plate, which will affect the accuracy of the measured values. Another problem can arise if the line-pair plate is positioned obliquely to the image plane, which could cause some of the line-pair groups to be at greater distances from the image plane. The same set up was used for all of the image sets taken, with the plate adjusted only once to change it from horizontal to 87 vertical alignment. The finding that higher resolution was obtained for all types of tomographic motion with the line-pair plate in the vertical plane could be due to a systematic error if there was an error in plate alignment. It is possible that the finer line-pair groups of the top row of the line-pair plate were angled slighdy away from the image plane for the horizontal plate position but within it for the vertical alignment images. 4. V a l i d i t y This is the relationship of what is actually measured to what the observer is attempting to measure (Brunette, D . M . 1996). The gold standard used was the "known" location of the cross-sectional images recorded when the images were made. According to the manufacturers of the CommCat imaging system, this would be accurate to within 1 millimetre. The finding that offset images for the resolution study were all worse than images taken with no offset confirms that the machine was calibrated to within half a millimetre. The fact that we were able to achieve no error for four types of markers in the pilot study and no more than 1 millimetre of error for the metatarsal bone in the definitive study suggests that the measurement method was valid. 5. D i s c u s s i o n Previous publication of evaluation of the resolution of the CommCat complex motion imaging system could not be identified. O f interest is that the mean resolution of hypocycloidal or spiral tomography was very close to that of the static image taken using the same film-screen combination. The mean resolution found with the cephalometric system at a 50% contrast ratio was 1.92 line-pairs per millimetre. Because each line pair consists of a bar and a space, the actual resolution is therefore 0.26 mm (one half the size of the line-pair). For the spiral 8 8 tomographic system the mean resolution was 1.72 line-pairs per millimetre, or 0.29 mm. This was the case even though the offset image sets showed that the line-pair plate was not perfecdy aligned with the image plane for the tomographic images. The poorer resolution found with the line-pair plate in the horizontal orientation versus vertical orientation may require additional investigation. For hypocyloidal tomography this finding could be pardy explained by the asymmetry of the motion, but the motion should be symmetrical for circular tomography and close to symmetrical for spiral tomography. The presence of post-object scatter shields laterally but not above or below the image could be partially responsible for the difference. Also, the pre-object collimation was not symmetrical, with the image being narrower horizontally than vertically. However, there is also the possibility that a systematic error occurred due to positioning of the line-pair plate slighdy further outside the image plane for the images taken with the line-pair plate horizontally oriented versus those taken with it in vertical orientation. 6. C o n c l u s i o n s -Comparing 50% contrast ratio of radiographic images produced with the line-pair plate versus the line-pair plate numbers produced graphs with very good linearity. This validates the use of this method. -The spatial resolution of a spiral tomogram taken with the CommCat complex-motion tomography system was slighdy poorer than that of a stationary radiograph made with the same film-screen combination. -Spiral tomograms had a greater spatial resolution than hypocycloidal tomograms and circular tomograms had the poorest spatial resolution. 89 -Greater spatial resolution was found for tomographic images with the line-pair plate vertically positioned versus tomograms made with the line-pair plate horizontally oriented. 90 Reference Brunette, D . M . Critical Thinking. 63-70. 1996. Carol Stream, IL, Quintessence Publishing Co, Inc. 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0092136/manifest

Comment

Related Items