THREE DIMENSIONAL MAGNETIC RESONANCE IMAGING OF THE NATURAL DENTITION RYAN JEFFERY COX

Transcription

1 THREE DIMENSIONAL MAGNETIC RESONANCE IMAGING OF THE NATURAL DENTITION by RYAN JEFFERY COX CHUNG HOW KAU, CHAIR ANDRÉ FERREIRA AMJAD JAVED CHRISTOS VLACHOS A THESIS Submitted to the graduate faculty of The University of Alabama at Birmingham, in partial fulfillment of the requirements for the degree of Master of Science BIRMINGHAM, ALABAMA 2012

2 THREE-DIMENSIONAL MAGENTIC RESONANCE IMAGING OF THE NATURAL DENTITION RYAN JEFFERY COX DEPARTMENT OF ORTHODONTICS ABSTRACT The purpose of this study was to determine if ultra short echo time magnetic resonance imaging (UTE-MRI) technology could be used to image teeth in orthodontics. The objectives of this study were to determine the accuracy and resolution of UTE-MRI on dental morphology and its application to the field of orthodontics with respect to appliance affect on morphology. Teeth were collected from the Orthodontic clinic at the University of Alabama at Birmingham and the Institutional Review Board of the University of Alabama at Birmingham approved the study. High-resolution 3-Tesla UTE-MRI was preformed on sixty extracted human premolar teeth with fixed ceramic orthodontic appliances at the University of Ulm, Germany. Linear measurements of the tooth morphology and orthodontic bracket dimensions were acquired with calipers and compared with virtual MRI images. Three ceramic bracket types including 3M Clarity, American Radiance, and Ormco Ice were studied. Both the caliper and UTE MRI measurements were highly reliable and accurate. Comparisons between the two methods showed no statistically significant differences in any of the bracket or tooth dimensions with a P-value > In general, the differences in the values ranged from -0.01mm to 0.06mm. A visual evaluation scale was used to assess the quality of the obtained UTE-MRI images when assessing the delineation of dental hard tissues. The visual evaluation scale (VES) of the images revealed that enamel, dentin, pulp and ceramic orthodontic appliances could be ii

3 subjectively delineated at a high level with the use of UTE-MRI. The following can be drawn from this prospective study; (1) Ceramic orthodontic appliances, without metal components, cause no dental image distortions and are readily visible on the UTE-MRI scans. (2) The measurement comparing linear tooth measurements to virtual MR images demonstrated that MRI has statistically and clinically significant accuracy on external tooth and bracket measurements. (3) Visual evaluation of the images revealed that enamel, dentin, pulp and ceramic orthodontic appliances could be subjectively delineated at a high level with the use of UTE-MRI. (4) Metallic slots in ceramic appliances cause severe image distortions. These distortions are localized and should not affect surrounding tissues in full volume MRI. iii

4 ACKNOWLEDGMENTS First and foremost, I would like to express my appreciation to the Lord for his guidance and hand in helping me achieve this long sought after goal. For without Him, none of this would have been possible. In addition, I would like to thank my loving and supportive wife Amy for her encouragement in my continuous endeavor to become an excellent orthodontist. Her patience and reinforcement served as an inspiration in my pursuit. I would also like to acknowledge and show my deepest gratitude to the members of my committee for their roles in completing this thesis. I especially would like to thank Dr. Chung How Kau for the inspiration, funding, support, and guidance he offered in completing, publishing, and the expense of this work. I would also like to thank the Department of Internal Medicine II at The University of Ulm for their hospitality and their willingness to contribute to this research. I would like to thank Dr. Volker Rasche for his generosity and kindness while I was in Ulm, Germany. This is a collaborative effort between the Department of Orthodontics of the University of Alabama at Birmingham and the Department of Internal Medicine II at the University of Ulm. iv

5 TABLE OF CONTENTS v Page ABSTRACT... ii ACKNOWLEDGEMENTS... iv LIST OF TABLES... vii LIST OF FIGURES... viii LIST OF ABBREVIATIONS... ix CHAPTER 1 SUMMARY OF INVESTIGATIONS INTRODUCTION LITERATURE REVIEW Magnetic Resonance Imaging Three Dimensional Imaging Modalities in Orthodontics Computerized Tomography (CT) Cone Beam Computerized Tomography (CBCT) Three-Dimensional Facial Scanning Laser Scanning Stereo-Photogrammetry Structured Light Technique dMDface TM System Ultrasound Four-Dimensional Imaging Dental Application of Magnetic Resonance Imaging Soft Tissue Imaging Temporomandibular Joint Abnormalities, Disease, and Tumors Hard Tissue Imaging Impacted Teeth Dental Impressions Detection of Carious Lesions Magnetic Resonance Imaging Techniques... 26

6 Indirect Techniques Direct Techniques Difficulty of Direct Proton Magnetic Resonance Imaging Direct Versus Indirect Magnetic Resonance Imaging Costs Involved Dental Material Artifacts with Magnetic Resonance Imaging Cone Beam CT Versus Magnetic Resonance Imaging Contraindications Limitations to MRI Technology...31 Image Manipulation AIMS AND NULL HYPOTHESIS Aim Hypothesis MATERIALS AND METHODS Subjects Inclusion Criteria Methods Imaging Unit Software Analysis Parameters Measured Statistics and Reliability Assessment RESULTS DISCUSSION CONCLUSION FUTURE DIRECTIONS Critique Future Directions LIST OF REFERENCES vi

7 LIST OF TABLES Tables Page 1 Tissue contrast Tissue signal intensity in MRI Means, standard deviations, and significance (Student t test) of bracket measurements between digital calipers and Ultra-short Echo Time Magnetic Resonance Imaging (n=30) Means, standard deviations, and significance (Student t test) of tooth measurements between digital calipers and Ultra-short Echo Time Magnetic Resonance Imaging (n=45) Comparisons of the linear measurements of the teeth samples to digital measurements vii

8 LIST OF FIGURES Figure Page 1 Diagrammatic representation of magnetic resonance imaging with patient in supine position Actual image of MRI machine Sample teeth mounted in agarose gel Tooth and bracket sample measured from occlusal to apical at the greatest distance parallel to the long axis of the subject to the nearest 0.01mm Digital Calipers measuring real dimensions of tooth sample Traditional Spin Echo UTE MRI showing clear separation of enamel from dentin Inversion of Figure 7; UTE MRI showing clear separation of enamel from dentin Ideal UTE MRI image of ceramic orthodontic bracket Metallic slots in ceramic appliances and severe localized image distortions Subsequent Slices 0.19mm viii

9 LIST OF ABBREVIATIONS CAD/CAM CBCT CT MPR MRI SPI STRAFI SWIFT TMJ TSE UTE Computer-Aided Design and Manufacturing Cone-Beam Computerized Tomography Computerized Tomography Multi-Planar Reconstruction Magnetic Resonance Imaging Single Point Imaging Stray Field Imaging SWeep Imaging with Fourier Transformation Temporomandibular Joint Turbo-Spin Echo Ultra-Short Echo Time ix

10 CHAPTER 1 SUMMARY OF INVESTIGATIONS Aim The purpose of this study was to determine if ultra short echo time magnetic resonance imaging technology could be used to image teeth in orthodontics. The objectives of this study were to determine the accuracy and resolution of ultra short echo time magnetic resonance imaging on dental morphology and its application to the field of orthodontics with respect to appliance affect on morphology. Materials and Methods Teeth were collected from the Orthodontic clinic at the University of Alabama at Birmingham (UAB) and the International Review Board of UAB approved the study. High-resolution 3-Tesla ultra short echo time magnetic resonance imaging was preformed on sixty extracted human premolar teeth with fixed ceramic orthodontic appliances at the University of Ulm in Germany. Real measurements of the tooth morphology and orthodontic appliance morphology were acquired with calipers and compared with nonreal MRI images. Three ceramic bracket types including 3M Clarity, American Radiance, and Ormco Ice were studied. The spin echo and high spatial resolution multi slice TSE were only used for visual comparison to the UTE slices. A visual evaluation scale was 1

11 used to assess the quality of the obtained ultra-short echo time magnetic resonance images when assessing the delineation of dental hard tissues. Results Both the caliper and UTE MRI measurements were highly reliable and accurate. Comparisons between the two methods showed no statistically significant differences in any of the bracket or tooth dimensions with a P-value > In general, the differences in the values ranged from -0.01mm to 0.06mm. The visual evaluation scale (VES) of the images revealed that enamel, dentin, pulp and ceramic orthodontic appliances could be subjectively delineated at a high level with the use of ultra short echo time magnetic resonance imaging. When compared using the visual evaluation scale, UTE clearly outperformed the standard and turbo spin echo. Conclusions The following can be drawn from this prospective study; (1) Ceramic orthodontic appliances, without metal components, cause no dental image distortions and are readily visible on the ultra short echo time magnetic resonance imaging scans. (2) The measurement comparing real tooth measurements to non-real MR images demonstrated that MRI has statistically and clinically significant accuracy on external tooth and bracket measurements. 2

12 (3) Visual evaluation of the images revealed that enamel, dentin, pulp and ceramic orthodontic appliances could be subjectively delineated at a high level with the use of ultra short echo time magnetic resonance imaging. (4) Metallic slots in ceramic appliances cause severe image distortions. These distortions are localized and should not affect surrounding tissues in full volume MRI. 3

13 CHAPTER 2 INTRODUCTION Imaging in orthodontics has seen rapid evolution due to the development of threedimensional (3-D) imaging modalities. After their development in the 1990s, custombuilt craniofacial machines began to appear on the market in the early 2000s. Other sophisticated imaging, with a variety of applications to the maxillofacial dental and skeletal regions, has followed. Orthodontists currently have the option to image their patients with a multitude of imaging modalities, which include computed tomography (CT), cone beam computed tomography (CBCT), cephalometric head films, panoramic films, intraoral and extraoral photography, soft-tissue laser scanning, stereophotogrammetry, and magnetic resonance imaging (MRI). Although there have been many developments and significant research focusing on imaging modalities for skeletal structures and soft-tissue morphology, the use of MRI is diminutive in comparison. 1 The use of non-ionizing MRI has not been considered as an imaging modality for routine orthodontic diagnosis due in part to its poor performance in imaging dental and skeletal hard tissue. 2-4 Traditional MR imaging is better adapted for imaging the soft-tissue components, including temporomandibular joint (TMJ) pathology, 5,6 tumors, 7,8 muscles, and attachments. 8 This may be attributed to the higher water content available in these tissues. However, recent literature has been published 4

14 exemplifying the use of MRI in dental hard-tissue applications, showing quality images without the exposure to ionizing radiation. 1-4,9-16 Magnetic resonance imaging began with the ideas of an Irish physicist, Sir Joseph Larmor, 17 and the first in vivo medical MRI was published in 1977 by Damadian. 18 The magnet is the biggest and most important part in an MRI system. Most MRI machines are graded on the strength of the magnet, which is measured in Tesla units. To put these numbers into perspective, 1 Tesla is the equivalent of 20,000 times the magnetic field strength of Earth. 8 Currently, MRI units are in the range of 1.5 to 3 Tesla for in vivo applications, but the magnetic strength may be much higher for research. Magnetic resonance imaging essentially utilizes hydrogen atoms to capture an image. By aligning individual atoms with magnetization, the machine applies a radiofrequency pulse to depolarize the atoms, capturing the signal emitted on excitation with a receiver coil. 19 A basic depiction of MRI technology can be seen in Figure 1 and an image of an actual MRI machine can be seen in Figure 2. Hydrogen is found in abundance in soft tissue, but is lacking in most hard tissues. Because the head and neck region are primarily dental and skeletal hard tissues, MRI has been perceived to have limited applications. A chance finding during research of four-dimensional cardiac MRI, an MRI sequence capable of hard tissue differentiation was discovered. 4,20 Multiple MRI sequences exist and are selected for each diagnostic procedure to help enhance the contrast of the tissue that is being studied. Each sequence consists of different radiofrequency pulses and associated mathematical gradients. Along with these parameters, timing of signal capture also plays a critical role when imaging hard tissue. With further testing and optimization of the sequence, it was realized that the new 5

15 technique of ultra-short echo time magnetic resonance (UTE-MRI) could maximize hard tissue capture, and hence, differentiate enamel from dentin and pulpal tissue. 3,4,20 Standard MRI sequences are not capable of such differentiation and characterize enamel, dentin, and skeletal tissues equally. The application of MRI in dentistry first appeared in 1981 in a paper entitled NMR: dental imaging without x-rays? 21 Since that time, the advancement of MRI in dental applications has been relatively insignificant. Current applications include assessment of the TMJ 5,6 and evaluation of soft tissue tumors of the head and neck. 7,8 Recent case reports have shown that MRI technology in orthodontics may be used to take indirect impressions of the oral cavity, 11,16 assess impactions of the dentition, 9 detect root resorption, 9 detect demineralization and caries, 1,3,10,13 and diagnose and treatment plan the three-dimensional patient. Unfortunately, imaging patients with metallic orthodontic appliances in active treatment has been shown to cause strong local image artifacts Metallic materials can either amplify or weaken the magnetic field depending on the magnetic properties. Metals, regardless of properties, cause an absence of spins in the image and thus result in signal voids where the image will appear black. 22 The metal will also affect radio signals during the image retrieval, potentially causing ring artifacts, shadowing and geometric distortions. 22 Nevertheless, the use of ceramic appliances and the development of composite wires may void this as a concern in imaging in the future. MRI has traditionally not been used clinically in dentistry and orthodontics because dental hard tissue has a relatively short T2/T2* relaxation time well below 200µs. 20 Bracher et al have shown that UTE has now made image acquisitions of echo times as low as a few µs possible. 3 Enamel has a T2 time of 70µs and dentin has a T2 of 150µs in a 1.5 6

16 Tesla field, which with the use of UTE, dental hard tissue assessment with MRI appears feasible. 20 Since MR imaging is non-invasive, non-destructive and offers no ionizing radiation, various craniofacial and dental MRI applications and sequences have become a topic of interest and pioneering research over the last few years. Various forms of spin echo imaging have been preformed on the dentition for caries research, but are only capable of achieving high-resolution visualization of moisture entering grossly decayed teeth. Unfortunately, it is also relatively slow requiring capture times over one hour. 10 The Fast Spin echo can capture the same image in around ten minutes. 10 Regrettably, these slower imaging techniques do not delineate the dental hard tissues, other than pulpal tissue, offering only an indirect appearance of the tooth surface. 10 On the other hand, direct proton imaging sequences, also as known as solid-state imaging, have further developed the capability of MRI to delineate dental hard tissues. Single point imaging (SPI) and strayfield imaging (STRAFI) are two of these techniques capable of such delineation, but limitations including long data collection times and high levels of acoustic noise associated with the scanning process render them unsuitable for in-vivo applications. 10 Another imaging sequence that has shown recent advancement in imaging dental hard tissues is SWeep Imaging with Fourier Transformation (SWIFT). Idiyatullin et al have recently shown that with the SWIFT imaging sequence, tissues with short T2 relaxation times such as dentin and enamel can be differentiated and that demineralized areas of the tissue appear as hyper-intense areas in the image. 13 Tymofiyeva has suggested that the SWIFT and UTE techniques are the two techniques that show the most promising in-vivo results. 1 Conversely, these direct proton-imaging techniques are inferior to the 7

17 indirect technique at capturing the tooth surface. 1 This would make a combination of direct and indirect imaging the most accurate technique for dental applications. Accuracy of the minimum distance from the pulp and the carious lesion has also been demonstrated with MRI. 1 Tymofiyeva et al and Bracher et al in separate studies were able to demonstrate that carious detection was not limited to gross open decay but direct proton measurement with solid state MRI. 1,16 With MRI being able to show the pulp, orthodontist can also have a better idea of pulpal related pathology in treatment planning including pulpal necrosis or internal resorption. Tymofiyeva also suggest the use of MRI technology to image carious lesions under MRI compatible ceramic crowns, which could translate into the use of MR imaging around an orthodontic bracket to visualize demineralization around MRI compatible orthodontic appliances. 1 Reperfusion of transplanted teeth for orthodontic purposes may also be used for assessing the dental pulp in treatment. 25 In the literature two different explanations exist on why we obtain the images we do and that is due to indirect and direct imaging. This study made use of both direct proton imaging and indirect imaging using a signal providing agarose gel. Using indirect methods of MR imaging with intraoral contrast mediums and an intraoral prototype RF receiver coil, Tymofiyeva et al were able to develop an MRI based technique for dental impressions. 11 In the records process of orthodontics Alginate or PVS impressions are a standard for obtaining a three-dimensional model of the patients dentition, unfortunately this impression process is one of the most feared procedures for the patient. Some patients even have a strong enough gag reflex that impressions are removed from the records and diagnosis process altogether. Challenges include accurate enough impressions to fabricate 8

18 orthodontic appliances. 11 Fortunately, the literature has shown that in-vitro the tooth surface can be reconstructed with a mean deviation below 30 µm and prosthodontic copings can be created with accuracy below 100 µm. 11 Although the impression material is replaced with a fluid in the patients mouth, patients seem to be be more comfortable with the fluid since they are laying face down during the imaging process. MRI has the potential to provide orthodontists with full volume imaging with the elimination of ionizing radiation. 9 This can be very useful to orthodontists for evaluating growth, superimposing serial images for the understanding of treatment changes with various mechanics, imaging of the temporomandibular joint and related pathology, supervision of eruption and the evaluation and determination of impactions or ankylosis. Impacted canines can complicate orthodontic treatment and the duration of treatment. Localizing canines can be accomplished with standard periapical radiographs with tube shifts to locate canines with the buccal object rule as well as with CBCT. Unfortunately, the ionizing radiation involved limits the diagnosis and supervision of the pediatric population we treat. Tymofiyeva found that MRI could yield a clear separation between impacted teeth and the surrounding tissue, and the position and angulation of impacted teeth three-dimensionally. 9 Such a technology would allow orthodontists to periodically image a patient to monitor improvements in impaction and determine the appropriate timing for intervention. It has also been suggested that this technology would give adequate information to the orthodontist regarding root resorption and transposed or overlapping teeth. 9 It is the objective of this study to evaluate the applicability of three-dimensional ultra short echo time (3D-UTE) magnetic resonance imaging on dental morphology and its 9

19 application to the field of orthodontics with respect to appliance affect on morphology by an ex-vivo assessment of ceramic appliances on the natural dentition. No studies thus far have investigated the accuracy of linear dental and orthodontic appliance measurements generated from a non-ionizing imaging source compared to the true anatomical dimensions. 10

20 The signals from the receiver coil are transferred to the imaging software for manipulation The magnetic,ield of the MRI is used to align the hydrogen atoms Radiofrequency waves (RF) or pulses are absorbed by the atoms The signals are captured in a receiver coil near the tissues being imaged The atoms emit a signal upon excitation Figure 1: Diagrammatic representation of magnetic resonance imaging with patient in supine position. 11

21 Figure 2: Actual image of MRI machine. 12

22 CHAPTER 3 LITERATURE REVIEW Magnetic Resonance Imaging Magnetic resonance imaging (MRI) is a technique that requires a specialized educational background to comprehend its complex physical principles. Therefore, it is beyond the scope of this text to provide an inclusive explanation of this technology. For further appreciation of magnetic resonance the author suggests reading dedicated text on the function and physics of MRI. Basic explanations of the principles of MRI described below are essential for appreciation of the technology. Magnetic resonance imaging began with ideas from an Irish physicist, Sir Joseph Larmor. 17 Larmor introduced the Larmor frequency, which became an important equation in MRI. 17 Medical MRI began as nuclear magnetic resonance with the first successful nuclear magnetic resonance (NMR) experiments being completed in 1946 separately by Felix Bloch and Edward Purcell. The first MR image was published in 1973, and Dr. Raymond Damadian published the first studies performed on humans in The presence of a hydrogen atom is necessary to produce an image with magnetic resonance. A single hydrogen atom contains a proton and an electron making the electrical charge neutral. The protons have a spin or an angular momentum around a constant rotational axis, like a globe, and also a magnetic moment due to the proton s electrical 13

23 charge. 19 Because of the magnetic moment, the proton functions like a miniature magnet. When inserted into a magnetic field, the spins of the protons will align in the direction parallel to the magnetic field with a minority aligned 180 to the field, also known as antiparallel. 8 While aligned with the magnetic field the spins will go through a process called precession, which is a process that has been compared to the spin of a top under influence of gravity. 19 Gravity will slowly pull on a spinning top causing a change in its rotational axis, much like the process of precession on a spin in a magnetic field. The speed with which precession occurs is based on the strength of the magnetic field, which is directly proportional to the Tesla unit. To put these numbers into perspective, one Tesla is the equivalent of 20,000X the magnetic field strength of Earth. 8 Currently, MRI units are in the range of 1.5 to 3 Tesla for in-vivo applications, but can range much higher for research. The rate of rotational axis change or wobble is also known as the Larmor frequency. 8 During the MRI scan the spins will first align parallel to the magnetic field, defined as longitudinal magnetization. Once the spins are parallel a radiofrequency pulse, equivalent to the Larmor frequency, is applied which excites the spins to tip exactly 90 to a transverse magnetization. After excitation the MR signals will fade back to the original longitudinal magnetization by longitudinal and transverse relaxation. Longitudinal or T1 relaxation occurs as excess energy from the excitation is dissipated to the peripheral environment and the spin returns to its original longitudinal axis. Transverse or T2 relaxation occurs when the spins loose phase coherence and exchange energy with each other. MR signal can also fade due to T2* relaxation which is caused by inhomogeneities at tissue borders which can be eliminated by spin echo. 19 T2 recovery will happen in the first ms whereas T1 relaxation will occur in 0.5-5s

24 Images are obtained from an MRI due to the fact that tissues will have a difference in their proton density, T1, and T2, which will give contrast between the tissues. Tissue contrast can be respected in table 1, whereas tissue signal intensity can be seen in Table 2. Using different MR pulse sequences, which can be created by differing the multiple tissue parameters, can modify tissue contrast. The MR image is obtained as a slice that has been excited several times. The interval between the excitation pulses is defined as the repetition time (TR) and can have an affect on T1 weighting. The time between the excitation pulse and the collection of the MR signal is defined as the echo time (TE) and effects T2 weighting. Table 1: Tissue Contrast Short T1 Tissue appears Bright Stronger MR Signal Long T1 Tissue appears Dark Weaker MR Signal Short T2 Tissue appears Dark Weaker MR signal Long T2 Tissue appears Bright Stronger MR signal 15

25 Table 2: Tissue signal intensity in MRI Tissue T1 weighting T2 weighting Air No Signal No signal Blood - Flowing No Signal No Signal Bone Marrow (fat) Bright Bright Cortical Bone Dark Dark Disk Dark Dark/Medium Fluid (inflammation) Medium Bright Muscle Medium Medium Three Dimensional Imaging Modalities in Orthodontics In orthodontics many techniques are used for imaging and recording the patient s morphology. Unfortunately, the most commonly used techniques are only twodimensional representations of a three-dimensional patient. Recent technological advancements have resulted in several modalities capable of giving the orthodontists a three-dimensional record. Computerized Tomography (CT) Computerized tomography began in the medical field in 1967 and has gone through five generations of scanners. 26 In the current generation, unlike fan-beam CT, there is a 16

26 moving radiation source with a fixed detector. 26 Computerized tomography creates a three-dimensional image by taking multiple two-dimensional images around a single axis. These two dimensional images can be reconstructed using multi-planar techniques to achieve an accurate three-dimensional image. Computerized tomography has also been used for surface imaging applications without exacting color. Computerized tomography has numerous limitations and few applications to orthodontics. The CT unit requires extensive space and is considerably more expensive than traditional radiography techniques. 26 The modality is also time consuming and costineffective due to the process of obtaining a final image of stacked slices. 26 This technique can be utilized in orthodontics for craniofacial anomalies, but exposes the patient to a higher radiation dose, limiting its routine use. Cone Beam Computerized Tomography (CBCT) Cone beam computerized tomography was developed to streamline CT technology and offer faster, less expensive, and convenient CT imaging with a smaller radiation dose. Unlike stacking multiple sliced images, CBCT uses a single rotation and a cone of X-rays to capture a entire region of interest in a single rotation. 26 It has been shown that this technology delivers 20% of the radiation of traditional CT and is equivalent to a full mouth periapical radiographic series. 27 Although the spatial resolution of the images is less than that of traditional CT, CBCT offers a full field image with better than diagnostic quality. 17

27 Three-Dimensional Facial Scanning With the recognition of the value of three-dimensional records, numerous noninvasive and non-ionizing methods of surface imaging were developed. Surface imaging affords the clinician an exacting image of the patient s facial soft tissue that cannot be acquired in alternative modalities. The following modalities exist with their own advantages and disadvantages: Laser scanning Developed from automotive and aerospace engineering, laser scanning is a valuable modality that allows the visualization of microscopic surface changes. 28 Laser technology uses an optical source to project a scattered laser image onto the face, which is captured by a charged couple device and interpreted by a computer software system. By triangulating the distances between the reflected laser beam and the scanned surface, the scanner can detect all three dimensions of the subject. 29 This modality is reliable and reproducible when it is used in three-dimensional imaging of the human face. 28 Stereo-photogrammetry Stereo-photogrammetry is a technique that utilizes two or more cameras that are configured at different angles to obtain three-dimensional coordinates of facial morphology. 29 Using the triangulation principles and a stereo-pair of cameras this technique has the advantage of rapid data acquisition, which has proven to be reliable and reproducible. 30,31 18

28 Structured light technique The structured light technique utilizes the principles of triangulation and a projected light pattern to capture an image of the subject. When the light pattern is projected onto the targeted surface the light distorts and bends and is simultaneously captured by cameras. 29 When the cameras are at a fixed distance they capture the reflected and distorted light pattern and translate it to three-dimensional coordinates. 29 3dMDface TM system A proprietary structured light technique that combines the technology of stereophotogrammetry and the structured light technique. 29 With the use of three cameras per side, the combination of four color and two infra-red cameras captures a high quality threedimensional picture. The combination of cameras is optimally configured at various angles and is synchronized to capture an image from a highly structured light system in around 1.5ms. 29 This technique is the most accurate at less than 0.5mm. 29 Ultrasound Ultrasound holography technology utilizes a 3.5 to 7.0 MHz high-frequency wave from a hand held device to image an area of the subject in interest. The repeated ultrasound beams are capable of scanning thin areas of tissue and reflect back to the device for imaging. Penetration can be set on the device, as well as capture time. There is currently no known risk to the patient with this technology. 19

29 Four-Dimensional Imaging Imaging in the fourth dimension has been proposed with the use of computerized tomography and magnetic resonance imaging. Sequences have been developed that are fast enough to capture the beating heart as well as the flow of blood. With further research these techniques may have applications in cardiac imaging and may have an application in the temporomandibular joint. Dental Application of Magnetic Resonance Imaging The application of magnetic resonance imaging in dentistry first appeared in 1981 in a paper entitled NMR: dental imaging without x-rays?. 21 Since that time the advancement of MRI for dental applications has been relatively insignificant. Current applications include the assessment of the form and function of the temporomandibular joint, and the evaluation of soft tissue tumors of the head and neck. Several papers have appeared in recent literature discussing the use of MRI for the assessment of impacted teeth, the planning of dental osseous implants, estimation of facial bone structure, detection of osteoporosis, indirect MRI imaging for dental impressions, and for the detection of root resorption. Researchers are also developing various applications to establish MRI as an imaging modality for the detection of carious lesions. Soft Tissue Imaging Temporomandibular Joint. Diseases of the temporomandibular joint can develop in hard and/or soft tissue. In many cases, hard tissue conditions are recognized and documented with standard panoramic or cephalometric images. Occasionally, like with 20

30 any disease process, additional imaging including conventional computerized tomography (CT) or dental cone-beam CT are prescribed to provide detailed information of the condition. Unfortunately, these imaging modalities do not provide insight into conditions of the disk or capsule. Clinical examinations for orthodontic patients ideally screen for TMJ abnormalities. When signs and symptoms of TMJ malfunction or deformity are diagnosed, an initial attempt is made to treat the clinical symptoms with conventional treatment. The most common condition to present clinically is disk displacement that may be caused by morphology or position. The etiology of this condition is complicated and its severity varies. When the symptoms do not respond to standard protocols or more information is needed to diagnose or justify surgical treatment, further imaging is implicated. Once popular contrast injected radiographs (arthography) have largely been replaced with MRI for additional imaging. Dedicated local TMJ coils have been developed to enhance imaging. Slices of the joint are typically seen in the coronal and saggital planes to help with visualization and interpretation of the displaced disk. Properly directed saggital views will distinguish antero-posterior discrepancies while coronal slices will discern displacements medially or laterally. When evaluating the position of the disk the patient should be imaged while in open and closed posture. In the closed image, the disk is generally situated on the anterior aspect of the articular surface of the condyle. In the opened position the condyle will lie directly over the body of the condyle in the reduced position. Abnormalities of the discal position are apparent in both positions and can present as displacement in all directions, but anterior disk displacement is most common. 21

31 In the closed position the disk will appear anterior to the articular surface of the condyle and when the patient opens it can either reduce back onto the disk or remain anteriorly displaced. 32 To obtain adequate information of the joint several image sequences are preformed. T1 weighted images are used to evaluate the morphology of the disk as well as its position in relation to the condyle and fossa. With the T1 sequence, the disk and cortical bone will appear dark and the medullary bone will have a brighter signal. Typically this protocol will be accompanied by a separate T2 protocol that will highlight joint inflammation and effusion. In the T2 sequence, the disk will again appear dark, but if inflammation or effusion is present these fluids will emit a higher signal. 8 Abnormalities, Disease, and Tumors. The breadth of soft tissue abnormalities, diseases and tumors and their diagnostic characteristics in magnetic resonance imaging could encompass an entire textbook, so for the sake of brevity the content presented will encompass the most commonly found pathology. One of the most important and recognized conditions in the field of orthodontics is the developmental abnormalities associated with cleft lip and palate. Although ultrasound is the gold standard for early identification of prenatal clefting, research has shown that ossified bone may disrupt assessment. 33 Therefore, magnetic resonance imaging has distinguished itself as a supplemental modality in prenatal clefting diagnosis. The use of this imaging modality has been successful at identifying the standard craniofacial deformities common to clefting, but has also shown to be useful at concomitantly diagnosing parallel developmental abnormalities in the cardiac, skeletal and other various tissues. Although many 22

32 pathological conditions of the maxillofacial complex are typically first identified and delineated with a screening panoramic image, MRI has been supplemental for distinguishing tumors, inflammation, and or abnormalities. Due to the high water content of tumors and inflammation, they appear to have a high signal on T2 weighted MR images. The image qualities of these pathological conditions can also be increased with the use of an intravenous contrast medium such as gadolinium and the prescription of the proper MR sequence to enhance the target tissues. MRI has been successful used to help delineate and diagnose pleomorphic adenomas, lymphoma, squamous cell carcinoma, hemangiomas, and neoplasms of the salivary glands. Hard Tissue Imaging Impacted Teeth. Diagnosis of impacted teeth in orthodontics is essential in patient treatment planning. Unfortunately, with current options we are often combining clinical examinations with two-dimensional radiographs or exposing our patients to ionizing radiation with a cone-beam CT to locate and diagnose impactions. A study performed by Tymofiyeva et al has shown that MRI can yield a clear separation between impacted teeth and the surrounding tissue, the position and angulation of impacted teeth an all dimensions. MRI will yield full volume images without ionizing radiation, allowing for periodic evaluation of the impactions without the risk of radiation exposure. 9 Unlike indirect impression scans, no contrast medium is needed to visualize the impactions due to the signal giving tissue surrounding the impactions. Unfortunately, roots that extend into the sinus cannot be distinguished in dental MRI due to tooth/air interface. Also, there is almost no distinction between the teeth and cortical bone due to the thin PDL and insufficient 23

33 resolution. The teeth can be rendered by manual separation from the cortical bone with imaging software. Advantages to orthodontics include determining location of impactions and root resorption diagnosis. Unfortunately, imaging patients with metallic orthodontic braces has shown to cause strong image artifacts and making canine impaction determination unfeasible. 9 The author has shown in unpublished research that ceramic orthodontic appliances will not cause this strong image artifact. Dental Impressions. In orthodontics, as in dentistry, patients typically dislike the impression process. MRI based indirect impressions may provide a safe and accurate alternative to such procedures. With traditional MRI there is limited signal from the crown of a tooth, therefore MRI impressions rely on indirect visualization from an applied contrast medium in the oral cavity. With this, a negative impression of the external surface of dental hard tissue can be obtained and developed into a dental cast, model, or direct production of a dental restoration. 11,16 Current research has recently demonstrated that tooth surface digitization with a full body MRI scanner is capable of producing accurately sufficient digital impressions for fabrication of dental restorations. 11 In their studies they were able to recognize the inefficiency of standard MRI measuring times and create an intraoral radiofrequency receiver coil that expedited the in vivo process. This dedicated intraoral receiver provides a high sensitivity in the area of the dentition and allows the addition of contrast media in the patient s oral cavity. Although subjects were limited in this study, they were able to obtain accuracy below 100 µm and fabricate two separate zirconium oxide copings with computer-aided design and manufacturing (CAD/CAM) based on the MRI data. The 24

34 copings were fitted with GC Fit Checker. In a similar in-vitro project Schmid et al were able to achieve tooth surface accuracy with a mean deviation below 30µm. Their research preceded Tymofiyeva et al, and suggested the development of an intraoral RF receiver coil. This intraoral device reduced the distance from the conductor to the teeth and achieved the highest possible coil sensitivity in the dentition. Unfortunately, current prosthodontic recommendations would consider 100µm inferior to the 50µm standard. Further development and research will be done to enhance the in-vivo technique, nevertheless these published results demonstrate exceptional viability of this technique. 11 In orthodontics this procedure can be advantageous in eliminating the standard clinical practice of impression taking of two arches and possible retakes for production of casts or digital models. To obtain clinically significant accuracy, a high resolution is required but hindered by such limitations as patient movement and dental material artifacts. Motion can be eliminated with chin and forehead rests, while dental material artifacts may need further research to determine how to limit such artifacts. 11 Detection of Carious Lesions. Standard radiographs only give clinicians twodimensional ideas of the extent of decay, while a three-dimensional image may afford the ability to determine accurate lesion size and distance to the pulpal tissues. According to Bracher et al, magnetic resonance imaging will not have a general application in dentistry unless it has similar performance to two-dimensional bitewings in the detection of carious lesions. 3,20 Bracher et al suggests that the main inadequacy of MRI for the identification of carious lesions is due to its intrinsic inability to assess mineralized hard tissues of the human dentition. 3 Seventy percent of dentin consists of the mineral 25

35 hydroxylapatite and enamel is composed of ninety-six percent. Since both of these have high mineral content, the concentration of free protons is low and can only attribute to weak magnetization and a poor MRI image. The relaxation times are also very low in a one Tesla MRI with dentin relaxing in under 1ms and enamel in under 250µs. 3 Therefore a very fast sequence is necessary to capture and detect differences between dental tissue and disease. The optimized UTE sequence used by Bracher et al demonstrated that dental demineralization and caries were detectable with the technique and resulted in detection similar to standard bitewings. 3 With UTE MRI the extent of the lesions appeared larger than standard x-ray films which can more accurately tell us the depth of the lesion. 3 This is an expected result with our understanding of how much mineralized tissue must be decayed prior to obtaining a carious image on standard x-ray. Magnetic Resonance Imaging Techniques Indirect Techniques Indirect methods of MR imaging can be accomplished with a standard spin echo sequence when the teeth are embedded in a dedicated signal providing material, such as gadolinium based oral contrast medium. These methods have been able to show carious lesions only if the signal-providing medium penetrates into a lesion, require significant decay. The indirect technique has also been established as a modality for dental impressions. Using a turbo spin sequence one can expect quality images with a measurement time less than ten minutes

36 Direct Techniques There has been a variety of direct proton imaging techniques published in the MRI literature regarding hard or mineralized tissue. These techniques include multinuclear imaging, solid-state imaging, constant time imaging, single point imaging, stray field imaging (STRAFI), Sweep Imaging with Fourier Transformation (SWIFT), and Ultra short echo time (UTE) imaging. 3 Solid-state imaging techniques such as STRAFI and SPI are very time consuming and are not suitable for in vivo applications. 13 Stray field imaging has been used to capture 3D images of carious lesions in extracted human teeth in-vitro. Among all these direct techniques, UTE and SWIFT appear to be the most feasible to obtain in vivo images with reasonable measurement times. 1 Sweep Imaging with Fourier Transform combines simultaneous excitation of spins and acquisition of the excited signal, and is good at studying objects with very short T2 times. 13 Acquisition time with this sequence is comparable to regular fast gradient echo techniques. This sequence has been used in-vivo and in-vitro providing images of teeth, demineralization and surrounding tissue in around ten minutes. Intraoral RF receiver coils can be used in an in-vivo application with SWIFT to increase accuracy and resolution. 11 Boujraf et al has shown the first in-vivo direct imaging of dental hard tissues by applying UTE for the delineation of pulp, dentin and enamel in volunteer patients. 20 Boujraf et al demonstrated that 3D UTE MRI is an adequate imaging technique to identify dental demineralization and carious lesions. 20 This in-vivo study demonstrated that the extent of the lesion in UTE imaging appeared larger than observed in x-ray, which is in accordance with Tymofiyeva et al. 34 He concluded that MRI might be a more accurate assessment of the real extent of the lesion and especially a better estimation of the distance 27

37 of the lesion to the pulp. Standard x-ray typically underestimates a lesions extent since it requires a certain amount of mineral loss before detection. In-vitro applications of the UTE sequence have shown imaging with delineation of the dental hard tissues at high spatial resolution with sufficient contrast and resolution to clearly delineate enamel, dentin, pulp and root material. 20 Difficulty of Direct Proton Magnetic Resonance Imaging Low concentrations of protons to contribute to magnetization and very short spinspin relaxation constants T2 of the magnetization. Dentin is 70% mineral hydroxyapatite, 20% organic and 10% water. Enamel is 96% hydroxyapatite and less than 4% water and organic material. The spin-spin relaxation constants T2 of the proton magnetization are very short in dentin (12µs-1ms) and extremely short in enamel (14-240µs). 1,35,36 Direct Versus Indirect Magnetic Resonance Imaging Solid-state direct proton MRI has the advantage of differentiating between dental hard tissues. The indirect technique is a better choice when the tooth surface reconstruction is needed. A combination may be the most beneficial when imaging the patient. Costs Involved With the current clinical MRI strength, the system is not known to produce any significant adverse effects in man. Disadvantages include relatively long acquisition times, (which makes plausible motion artifacts and if additional sequences are necessary for optimization of separate tissues the patient will need to remain in the exact same position 28

38 for superimposition), claustrophobia, noise, and image artifacts from ferromagnetic items. An additional disadvantage in orthodontics comes when we consider patient positioning. If traditional MRI machines are used, patients will have a soft tissue pull from gravity that will be 90 degrees to the gravitational pull at natural head position, which may allow excessive relaxation of the mandible and attached tissues. Another disadvantage to any three dimensional technique according to Kau, is that we can obtain a three dimensional image, but it will not be accurate over the full field of view. 26 We have also seen a lack in in-vivo applications of dental MRI do to the need for high magnetic fields, long measurement times and a lack of dedicated hardware. 1 Contraindications include cardiac pacemakers, implanted cardiac defibrillators, aneurysm clips, nuerostimulators, metallic foreign bodies, especially in the eye, and other implantable devices. Some companies are currently producing devices that are MRI safe, but this should be referenced prior to scan. Fortunately, Dental MRI can provide the clinician with good soft tissue contrast, none of the risks of ionizing radiation, and multi-planar imaging can be preformed. Tissue contrast agents can also be made available to aid in patient diagnosis. Regrettably, cost, benefit and risk ratios have limited decision of which imaging equipment we use. MRI is relatively safe but costly. Dental Material Artifacts with Magnetic Resonance Imaging The presence of any metal in the image volume will cause artifacts, shadowing, or voids in the image outcome. This can be enhance when the materials proximity to the center of the scan is decrease or with variations in MRI pulse sequences. 29

39 Some composite materials may cause local image distortions due to the presence of metal oxides such as iron oxide. Solutions to reduce metal artifacts are being sought in MRI. 9,23 Cone Beam CT Versus Magnetic Resonance Imaging In current three-dimensional imaging for dental purposes, only the cone-beam CT gives off ionizing radiation. Although the total radiation exposure to the patient is typically 20% of a traditional medical CT, sources have quoted that it can be as high as 44 times the typical panoramic dose. 37 The resulting effective radiation dose is dependent on the quality of the image desired and can be adjusted with kvp and ma. This rather high radiation dose voids CBCT as a screening and supervision supplement when we recognize that the benefit should outweigh the cost to the patient. Tymofiyeva compared the economic feasibility of MRI vs CBCT at the University of Wuerzburg, Germany. 1 His comparison as of 2009 showed that a CBCT dantomaxillofacial examination cost 175 euro while a similar exam with an MRI would cost 206 euro. 1 Unfortunately the cost of an MRI is based on time, and can be lengthened or shortened. He showed that MRI is inferior in spatial resolution, but MRI can provide the information necessary for localization of impacted teeth. He also noted that metal artifacts are noted in both imaging modalities. 1 Traditional spin echo based MRI methods only allow for high-resolution visualization of moisture entering into heavily decayed teeth and can take one hour for data collection. 10 A Fast spin echo can reduce this time to ten minutes but separation of the tooth s tissues is not possible due to their T2 times

40 Boujraf et al states that acquisition times equivalent to x-ray may never be feasible, but times of less than 15 minutes for a complete assessment of the status of the teeth and the head and neck appears feasible. 20 Compared to CBCT, MRI has better soft tissue contrast but is less suited for identifying bony irregularities. MRI is also inferior to CBCT in spatial resolution. 26 CBCT is excellent in imaging hard tissue structures and some soft tissue components, but it does not have the capability to display muscle structures and attachments. 26 These have to be obtained with magnetic resonance imaging, which incidentally does not exposure the patient to ionizing radiation. 26 Children are at a higher risk of radiation effects because their tissues are more radiosensitive and they are more likely to live to see the detrimental side effects. 26 Contraindications Contraindications include cardiac pacemakers, implanted cardiac defibrillators, aneurysm clips, metallic foreign bodies, neurostimulators, and other implantable devices. 1 Some companies are currently producing devices that are MRI safe, but the device in question should be referenced prior to taking a scan. Limitations to MRI Technology At present, there is no dedicated MRI machine in orthodontics. This is in part the result of an undiscovered usage of the technology. Tymofiyeva et al compared the economic feasibility of MRI vs CBCT at the University of Wuerzburg, Germany. 1 This comparison as of 2009 showed that a CBCT dento-maxillofacial examination cost

41 euro, while a similar exam with an MRI would cost 206 euro. 1 Unfortunately, the cost of an MRI is based on time, and can be lengthened or shortened. The article also showed that MRI is inferior in spatial resolution, but MRI can provide the information necessary for diagnosis. 1 It was also noted that metal artifacts are noted in both imaging modalities. 1 Bracher et al states that acquisition times equivalent to x-ray or CBCT may never be feasible, but times of less than 15 minutes for a complete assessment of the status of the teeth and the head and neck appears feasible. 3 The greatest hindrances to this rising technology include relatively long image acquisition times, high development and operating costs, and poorer spatial resolution compared to CBCT. 1,3 Additional disadvantages of MRI include claustrophobia, noise, and image artifacts from ferromagnetic items. 1 In orthodontics, patient positioning in a natural position also becomes a dilemma. If traditional MRI machines are used, patients will have a softtissue pull from gravity that will be 90 degrees to the gravitational pull at natural head position, which may allow excessive relaxation of the mandible and attached tissues. Another disadvantage to any 3-D technique is the accuracy of the 3-D image over the full field of view. 26 Lastly, there is a lack of in-vivo applications of dental MRI due to the need for high magnetic fields, long measurement times, and a lack of dedicated hardware. 1 Image Manipulation Once an image has been captured, it may be exported into a DICOM format. This format may be used by most available 3-D software and manipulated with its tools. 32

42 Multi-planar views and soft-tissue contrasting are possible in these DICOM formats and are very similar to CBCT derived images. 33

43 CHAPTER 4 AIM AND NULL HYPOTHESIS Aim The purpose of this study was to determine if Ultra-short echo time magnetic resonance imaging technology could be used to image teeth accurately in orthodontics. The objectives were to determine the accuracy and resolution of ultra-short echo time magnetic resonance imaging on dental morphology and its application to the field of orthodontics with respect to appliance affect on morphology. Hypothesis The null hypothesis in this research is: The use of ultra-short echo time magnetic resonance imaging is no different in the dimensional accuracy of the image when compared to the real measurements. 34

44 CHAPTER 5 MATERIALS AND METHODS Subjects Sample teeth were collected from the Orthodontic Clinic at the University of Alabama at Birmingham. Ceramic brackets were obtained with donations from 3M Unitek, American Orthodontics, and Ormco. The sample included 60 human premolars, 15 3M Unitek Clarity, 15 Ormco Ice, and 15 American Orthodontics Radiance ceramic brackets. Inclusion Criteria The sample teeth had to meet the following criteria; (1) Teeth should be human premolars; teeth numbers 4, 5, 12, 13, 20, 21, 28, or 29. (2) Teeth should be void of caries. (3) Teeth should be void of restorations. (4) Teeth should have closed apicies. 35

45 The brackets had to meet the following criteria; (1) The brackets should be made of ceramic. (2) The brackets should be designed for the same tooth and be from the same manufacturers lot. Methods High resolution Ultra-short Echo Time (UTE) magnetic resonance imaging was preformed on sixty extracted human premolar teeth with fixed ceramic orthodontic appliances. The ceramic upper right 18-slot premolar brackets included 3M Unitek Clarity, American Orthodontic Radiance, and Ormco Ice. Fifteen brackets from each manufacturer were bonded to the buccal surface of fifteen human premolar teeth with 3M Unitek Transbond XT. All excess composite was removed prior to final cure. The teeth were then mounted in an agarose gel to replicate an oral based medium that surrounded the surface of the dentition and the ceramic bracket, as well as to decrease air/surface interface artifacts. The agarose gel provided a good indirect image, or negative, of the external surface. Imaging Unit The MRI images were completed on a 3 Tesla whole body MRI system (Achieva, Phillips Medical Systems, Best, The Netherlands) equipped with gradient hardware capable of a maximum amplitude of 40mT/m using a maximum slew rate of 200mT/ms. 20 A prototypical two times two-element carotid artery coil sized 120x50mm (Phillips Research 36

46 Europe, Hamburg, Germany) was used to acquire data. 20 Prior to this study, an optimization/calibration procedure of the 3D-UTE technique was carried out. 20 A detailed overview of the MRI acquisition parameters can be found by referencing research previous preformed. 20 After a fast initial survey in the mid-saggital direction along the long axis of the tooth, three high resolution surveys were aquired. Each survey was planned in a saggital orientation along the tooth s long axis. The diagnostic scans performed were a standard spin echo, a high spatial resolution multi slice TSE acqusition, and a 3D UTE scan. The series of scans of the tooth/bracket samples were completed in one hour and fifteen minutes. To reduce research use of the clinical MRI scanner, two samples were mounted in each sample vial as seen in Figure 3. 37

47 Figure 3: Sample teeth mounted in agarose gel Software Analysis The MRI images were analyzed on Osirix imaging software. The three dimensional MRI data was multi planar reformatted (MPR) along the long axis of the tooth with the other axes from the buccal to lingual and the mesial to distal. The same clinician, who was trained on the software, preformed all measurements, and repeated measurements 38

48 were made seven days following the initial measurements for intra-examiner reliability. Brackets were measured from the mesial to the distal and the occlusal to apical at the greatest diameter parallel and perpendicular to the occlusal plane and recorded to the nearest 0.01mm as seen in figure 4. The teeth were measured from mesial to distal, buccal to lingual, and occlusal to apical at the greatest diameter parallel or perpendicular to the occlusal plane to the nearest 0.01mm, also seen in figure 4. Figure 4: Tooth and bracket sample measured from occlusal to apical at the greatest distance parallel to the long axis of the subject to the nearest 0.01mm Parameters Measured Linear measurements of the tooth morphology and orthodontic appliance morphology were made with digital calipers to the nearest 0.01mm measuring at the largest diameter (Mitutoyo Absolute Digimatic Calipers) as seen in figure 5. These linear measurements were compared with the virtual digital MRI images. 39

49 Figure 5: Digital Calipers measuring real dimensions of tooth sample Statistics and Reliability Assessment Reliability and accuracy were assessed by using paired Student t tests. A P value of 0.05 was used to assign statistical significance. Intra-examiner reliability was assessed by measuring two of each of the samples seven days after the first measurements and comparing these measurments to the first. 40

50 CHAPTER 6 RESULTS The imaging protocol was completed on all sixty of the sample teeth. Severe localized distortion due to metallic slots in ceramic brackets was observed in all three scans including standard spin echo, high spatial resolution multi slice Turbo-spin echo acquisition, and the three-dimensional Ultra-short echo scan. These distortion artifacts were limited to the coronal half of the tooth and would be localized in an in-vivo application. Therefore, the ceramic brackets with metallic slots from 3M Unitek had to be excluded from the study. Linear measurements of the orthodontic appliance and tooth morphology compared with digital MRI images were dimensionally accurate within 0.1mm in all three planes of space. These results can be seen in Table 3 and Table 4. Paired student t-tests were preformed to compare the MRI digital images to the linear caliper measurements in the mesial-distal and occlusal-gingival measurements of the brackets and the mesial-distal, buccal-lingual, and occlusal-gingival measurements of the teeth. When comparing the means of digital to linear measurements in all samples the results were non significant with all P-values > The UTE MRI values tended to slightly over estimate 41

51 the real values on the bracket measurements. These results can be seen in Table 3 and Table 4. Comparisons of the linear measurements of the teeth samples to digital measurements can be seen in Table 5. Therefore the teeth and brackets are dimensional accurate using the UTE MRI technology when compared to real measurements. Intraexaminer reliability measurements showed no differences between the measurements. Figure 6 shows the image of a single tooth imaged with traditional Spin Echo MRI and Figure 7 shows the image of a single tooth imaged with the 3D UTE MRI sequence. An inverted image of Figure 7 can be seen in Figure 8. The ceramic bracket can be accurately visualized in the slice in Figure 9. Only the ceramic appliances with metal slots (3M Unitek Clarity Twin) had to be excluded from this study due to image artifacts from the metal slot resulting in degraded image quality. Therefore it was re-confirmed that metallic slots (Figure 10) in ceramic appliances cause severe image distortions. These distortions appear localized and may not affect surrounding tissues in full volume MRI. Figure 11 shows a UTE MRI slice followed by a subsequent slice at 0.19mm. 42

52 Table 3. Means, standard deviations, and significance (Student t test) of bracket measurements between digital calipers and Ultra-short Echo Time Magnetic Resonance Imaging (n=30) Digital Calipers UTE MRI Measurement (mm) Mean SD Mean SD Δ Mean DC-UTE MRI Sig Ormco Mesial-Distal NS Ormco Occlusal-Gingival NS American Mesial-Distal NS American Occlusal-Gingival NS Sig, significance; NS, not significant Table 4. Means, standard deviations, and significance (Student t test) of tooth measurements between digital calipers and Ultra-short Echo Time Magnetic Resonance Imaging (n=45) Digital Calipers UTE MRI Measurement (mm) Mean SD Mean SD Δ Mean DC-UTE MRI Sig Ormco Mesial-Distal NS Ormco Occlusal-Gingival NS Ormco Buccal-Lingual NS No Bracket Mesial-Distal NS No Bracket Occlusal-Gingival NS No Bracket Buccal-Lingual NS American Mesial- Distal NS American Occlusal-Gingival NS American Buccal-Lingual NS Sig, significance; NS, not significant 43

53 Table 5: Comparisons of the linear measurements of the teeth samples to digital measurements Tooth Measurements Occlusal Gingival Real Caliper Measurements UTE MRI Measurements

54 Figure 6: Traditional Spin Echo Figure 7: UTE MRI showing clear separation of enamel from dentin Figure 8: Inversion of Figure 7; UTE MRI showing clear separation of enamel from dentin Figure 9: Ideal UTE MRI image of ceramic orthodontic bracket 45

55 Figure 10: Metallic slots in ceramic appliances and severe localized image distortions Figure 11: Subsequent Slices 0.19mm 46