We present here the first in vivo optical coherence tomography (OCT) images of human dental tissue. A novel dental optical coherence tomography system has been developed. This system incorporates the interferometer sample arm and transverse scanning optics into a handpiece that can be used intraorally to image human dental tissues. The average imaging depth of this system varied from 3 mm in hard tissues to 1.5 mm in soft tissues. We discuss the application of this imaging system for dentistry and illustrate the potential of our dental OCT system for diagnosis of periodontal disease, detection of caries, and evaluation of dental restorations.
©1998 Optical Society of America
Currently, dentists evaluate the oral health of a patient through three main avenues: visual/tactile examination, periodontal probing, and radiographic imaging. Probes are placed between the soft tissue and tooth to assess periodontal conditions. The depth of probe penetration (probable pocket depth) is measured and the location of the soft tissue attachment is estimated from a fixed reference point on the tooth. Periodontal probing can be painful for the patient and diagnostically imprecise. Probing errors result from variations in insertion force, inflammatory status of tissue, diameter of probe tips, and anatomical tooth contours [1, 2]. Radiographs reveal morphological characteristics of teeth and alveolar bone that can not be identified in a visual examination. While radiographs are highly sensitive to detecting regions of carious demineralization and alveolar bone loss, they have several limitations. Radiographs cannot distinguish active from inactive disease. Periodontal disease is not identified until significant bone loss has occurred. Since radiographs are two dimensional, precisely locating the position of a carious lesion or osseous defect is impossible. Finally, radiography uses harmful ionizing radiation and provides no information on soft tissue state. Currently, there is no method to reliably quantify the soft tissue changes that occur in gingivitis and periodontal diseases.
The goal of our dental optical coherence tomography (OCT) system is to produce in vivo images of dental microstructure that can be used to make both qualitative and quantitative assessments of oral tissue health. In particular, we hope to derive from these images clinically important anatomical features such as location of the soft tissue attachment, morphological changes in gingival tissue, tooth decay, and structural integrity of dental restorations. Optical coherence tomography is a well known technique for creating noninvasive, high resolution (< 20 μm) images of biological microstructure . Clinically, OCT systems have been developed for ophthalmic , dermatological , and endoscopic  applications. Previously, we performed the first in vitro OCT imaging studies of porcine dental tissue . Based on the successful implementation of our OCT system in animal models, we then designed and built a handpiece for intraoral imaging of human dental tissue. In this article we present the details of this system and discuss its limitations in terms of image contrast, imaging depth, and the occurrence of image artifacts. In addition, we show the first intraoral OCT images of the human oral cavity and discuss these images in terms of their relevance to oral diagnosis.
2. Dental OCT system design
A schematic of the OCT instrumentation is shown in Figure 1. It is based on a white light fiber optic Michelson interferometer. Output from a low coherence light source is split at the 2 × 2 fiber optic coupler and directed toward the sample and reference arms of the interferometer. Reflections from the mirror and backscattered light from the sample are recombined at the coupler and propagated to the detector and light source. An interferometric signal is detected when the distance to the reference and sample arm reflections is matched to within the source coherence length. A scanning retro-reflector varies the path length of the reference arm for each transverse location on the sample. Loss in signal intensity caused by birefringence effects in the optical fiber is corrected using polarization paddles.
A cross sectional image is produced by transversely scanning the beam across the sample and collecting a reflectance versus depth profile at each transverse location. The reflectance intensities are recorded digitally on a grey-scale image as a function of transverse and axial distances. An example of how a OCT image is created is provide in the attached QuickTime movie (Fig. 2). The upper portion of the movie corresponds to a scan taken at a given transverse position in the tissue, or A-Scan image. The axes correspond to depth in the tissue versus intensity of the collected reflections. A greyscale intensity plot, or B-Scan image, consists of a sequential series of these scans. This particular image shows recession of the gingival tissue from an anterior tooth (Fig. 2). The total scan time for each image was approximately 45 seconds. The system had a lateral resolution of 50 μm and an average total lateral scan distance of 12 μm. The lateral resolution of the system is limited by the size of the focused beam on the tissue (20 μm) and determined by the speed with which the sample arm collection optics are laterally scanned. A 15 mW fiber amplified source was used that has a central wavelength of 1310 nm and a spectral bandwidth of 47 nm. The coherence length of this source corresponds to a free space axial resolution of 15 μm. The signal to noise ratio of the OCT system was measured to be 110 dB.
An optical handpiece was developed for making intraoral OCT scans (Fig. 3). Comfortable access to the human oral cavity and a strategy for remotely scanning the sample arm collection optics to perform cross-sectional imaging were the primary considerations for design of this handpiece. The collection optics include a 0.46 NA gradient refractive index (GRIN) lens/ angle prism combination mounted on the end of the sample arm single mode fiber optic. Light emerging from the fiber is focused by the GRIN lens to a 20 μm spot on the tissue and reflected 90 degrees by an internal reflection from a prism. The fiber optic/GRIN lens assembly is linearly scanned parallel to the facial surface of the dental tissue. This “side-firing" handpiece can easily access the posterior portion of the oral cavity. This access is critical since, clinically, a significant fraction of disease occurs in or around the posterior teeth.
Since the handpiece was in direct contact with the dental tissue, a disposable plastic sleeve was placed over the end for infection control. The mid-facial surface of teeth in three volunteers was imaged using the dental OCT system. The resulting images were then analyzed for the visibility of clinically relevant anatomical features.
3. In vivo Optical Coherence Tomography imaging of dental structures
The images produced using our OCT imaging system showed a significant amount of structural detail in both the hard and the soft dental tissues. A cartoon highlighting the cross-sectional anatomy of the imaged dental tissues is shown below (Figure 4a). The upper part of the drawing corresponds to the top, or crown, of the tooth. The dentin shell (D), made up of a mixture of collagen and hydroxyapatite salts, is covered in the crown by an enamel cap (E). The interface between the dentin and enamel is called the dento-enamel junction (DEJ). Cementum forms the thin outer layer (~10 μm) at the bottom, or root of the tooth. The gingival margin (GM) is located at the point where the tooth transitions to gingival tissue. The sulcus (S) is defined as the shallow groove between the gingival epithelium (EP) and the tooth. The connective tissue attaches the free gingival margin to the tooth surface, while the periodontal ligament connects the alveolar bone (AB) to the cementum and thus anchors the tooth to the jaw bone.
Representative OCT images of anterior and posterior teeth are shown in Figures 4b and 4c respectively. No post-processing has been used in creating these images. The axial dimensions shown are in terms of optical pathlength. This scale needs to be divided by the refractive index of the relevant tissue to obtain true physical dimensions (~1.3 for oral mucosa, ~1.6 for enamel, and~1.5 for dentin), resulting in an axial compression of the images. The OCT system generates images with an axial resolution of 15 μm (free space) as defined by the source coherence length. Our data was acquired at 3 pixels per 15 μm. The resolution of the computer screen, however, is limited to 30 μm per pixel. The maximum imaging depth of our system varied according to the optical properties each tissue type, attaining physical distances of up to 3 mm in hard tissues and up to 1.5 mm in soft tissues.
The OCT images represent a 20 μm cross-section of the tooth at the cervical region (Fig. 4b, 4c). The plastic sleeve used for infection control (IC) is visible as a dark line before the facial surface. Motion artifacts were only present in 5% of the images. Motion artifacts that did occur were primarily low frequency modulations between adjacent A-scans due to patient breathing. The sharp contour directly below the gingival margin in Figure 4c is probably a result of such an artifact. Several structural components of the gingival tissue including the sulcus (S), the epithelium (EP), and the connective tissue layer (CT) were visible in the collected OCT images. The porous alveolar bone (AB) was presumptively identified in these images as the region underlying the dark epithelial layer. Because progression of periodontal disease results in alterations of the connective tissue attachment and alveolar bone loss, identification of these anatomical features using an OCT imaging technique is an important first step in developing a useful diagnostic technique.
Hard tissue structures in the tooth were also identified in the OCT images. The imaging depth of the OCT system in enamel (E) was higher than in the dentin (D), with the dento-enamel junction (DEJ) visible in all the images. We expected this large difference since the scattering coefficient of enamel has been measured by previous investigators to be an order of magnitude lower at wavelengths approaching 1.3 μm than the scattering coefficient of dentin . Swirls of alternating dark and light bands appeared in the enamel layer of several images (Fig. 4c). These bands did not correspond to any known anatomical structure in the enamel, and therefore were assumed to be artifacts induced by tissue birefringence. This hypothesis was strengthened by the fact that these artifacts appeared primarily in contoured posterior teeth.
We imaged several extracted teeth using our bulk-optic birefringence sensitive OCT system to determine if these bands were indeed image artifacts caused by birefringence effects. Although the details of this technique are described in great detail elsewhere [9, 10], the operating principle is provided here as background. Since OCT is an interferometric technique, signal intensity is a function of the relative polarization states of the sample and reference arm reflections. A sample that is birefringent, with an anisotropic refractive index, modifies the polarization state of light passing through it. Conventional OCT systems do not have the ability to compensate for this change, resulting in image artifacts that correspond to the amount of mismatch between the polarization state of the sample and reference arm reflections. Our birefringence insensitive OCT system eliminates these artifacts by measuring the polarization state of the light of the returned light as it passes through the sample. Two images of extracted teeth collected using our birefringence sensitive OCT system are shown (Fig. 5a, 5b). The artifacts present in the enamel layer (Fig. 5a) are identical to those which would have been obtained on a standard OCT system using polarized light oriented at 45 degrees to the axis of tissue birefringence. The polarization insensitive OCT image (Fig. 5b) does not contain these artifacts and has improved depth of penetration because of the elimination of signal fading due to mismatched polarization states in the reference and sample arm.
The amount of birefringence in enamel depends on the relative orientation between the enamel prisms and the incident probe light. The enamel prisms run like the spokes of a wheel from the dentoenamel junction to the outer enamel surface . These prisms, made up of hydroxyapatite crystals, can be as long as 3–4 mm near the cervix of the tooth and average 6 μm in diameter. Neighboring prisms are separated by 0.1 to 0.2 μm wide glycoprotein prism sheaths. Enamel is birefringent because the index of refraction for light polarized along the axis of the prisms is different from the index of refraction for light polarized perpendicular to the axis. When the enamel surface is perpendicular to the incident light from the optical handpiece, the two axes of polarization of the incident light are both perpendicular to the axes of the prisms and minimal birefringence effects occur. As the contour of the tooth changes, however, the prisms become more perpendicular to the probe light and birefringence-induced artifacts appear in the OCT images. The effect of these artifacts, unfortunately, is to obscure structural features in the enamel that might otherwise be visible, such as small pre-carious or carious lesions. The successful implementation of OCT for caries detection will therefore probably rely on some technique for eliminating these artifacts, such as a fiber optic birefringence insensitive OCT system.
There are some clinical applications, however, such as evaluation of dental restorations, where the OCT birefringence artifacts may be useful as a contrast agent. Dental restorations provide a barrier restricting oral fluids and bacteria from entering the tooth. An inadequate marginal seal can result in a further loss of tooth structure and dissemination of bacteria . The most commonly used methods for evaluating the margins and structural integrity of restorations are visual, radiographic, and tactile examination . We imaged a composite restoration (CR) from both the facial and occlusal surface of the tooth using the in vivo OCT system (Fig. 6). The composite material has relatively homogeneous scattering properties and is not significantly birefringent compared to the surrounding enamel microstructure (E) as viewed from the occlusal, or top surface of the tooth (Fig 6a). The birefringence artifacts in this image, therefore, serve to clearly delineate the interface between the enamel and the restoration. We have presumptively identified from the OCT image the interface between the restoration and underlying dentin (I). While the transition between the restoration and dentin is not clear, the interface between the enamel and composite restoration is visible for the full length of the enamel layer (Fig. 6b). These preliminary studies indicate OCT is potentially a powerful technique for visualizing structural and marginal restoration defects before significant leakage occurs, minimizing tooth loss and decreasing the number of unnecessary replacement restorations. We are currently undertaking further studies using extracted teeth to obtain a clearer understanding of the features present in these OCT images.
The development of a compact OCT system and optical handpiece have, in conclusion, allowed us to produce the first in vivo cross-sectional images of dental microstructure. These images demonstrate clearly the potential of OCT in a variety of clinical dental applications including diagnosis of periodontal disease, detection of caries, and evaluation of dental restorations. Our immediate future research projects will focus on modifying the existing OCT system to improve image acquisition time, eliminate artifacts due to tissue birefringence and infection control, and improve registration of the handpiece against the tissue surface. We will also correlate OCT image characteristics with standard assessments of oral health in patient oriented clinic studies.
This work was performed under the auspices of the U.S. Department of Energy by Lawrence Livermore National Laboratory under Contract No. W-7405-Eng-48 and funded by a National Institute of Dental Research (NIDR) RO1 DE11154-03 grant “Diagnostic Optimal Imaging of Periodontal Tissues.”
References and Links
1. L. Mayfield, G. Bratthall, and R. Attstrom, “Periodontal probe precision using four different periodontal probes,” J. Clin. Periodontology 23, 76–82 (1996). [CrossRef]
2. B. L. Pihlstrom, “Measurement of attachment level in clinical trials - probing methods,” J. Periodontology 63, 1072–1077 (1992). [CrossRef]
3. D. Huang, E. A. Swanson, C. Lin, J. S. Shuman, W. Stinson, W. Chang, M. R. Hee, T. Flotte, K. Gregory, C. A. Puliafito, and J. G. Fujimoto, “Optical coherence tomography,” Science 254, 1178–1181 (1991). [CrossRef] [PubMed]
4. M. R. Hee, C. A. Puliafito, C. Wong, and J. S. Duker, “Quantitative assessment of macular edema with optical coherence tomography,” Arch. Opth. 113, 1019–1029 (1995). [CrossRef]
6. G. J. Tearney, S. A. Boppart, B. E. Bouma, M. E. Brezinski, N. J. Weissman, J. F. Southern, and J. G. Fujimoto, “Scanning single-mode fiber optic catheter-endoscope for optical coherence tomography,” Opt. Lett. 21, 543–545 (1996). [CrossRef] [PubMed]
7. B. W. Colston, M. J. Everett, L. B. Da Silva, L. L. Otis, P. Stroeve, and H. Nathel, “Imaging of hard and soft tissue structure in the oral cavity by optical coherence tomography,” Appl. Opt. 37, 3582–3585 (1998). [CrossRef]
8. D. Fried, R. E. Glena, J. D. B. Featherstone, and W. Seka, “Nature of light scattering in dental enamel and dentin at visible and near-infrared wavelengths,” Appl. Opt. 34, 1278–1285 (1995). [CrossRef] [PubMed]
9. M. J. Everett, K. Schoenenberger, B. W. Colston, and L. B. Da Silva, “Birefringence characterization of biological tissue using optical coherence tomography,” Opt. Lett. 23, 228–230 (1998). [CrossRef]
10. M. R. Hee, D. Huang, E. Swanson, and J. Fujimoto, “Polarization-sensitive low-coherence reflectometer for birefringence characterization and ranging,” J. Opt. Soc. Am. B 9, 903–908 (1992). [CrossRef]
11. K. F. Leinfelder and J. E. Lemons, Clinical Restorative Materials and Techniques. (Lea & Febiger, Philadelphia, 1988).
12. J. F. Cvar and G. Ryge, Criteria for the clinical evaluation of dental restoration materials. (US Dept HEW PHS, Dental Health Center: San Francisco, 1971).
13. L. Moss-Salentijn and M. Hendricks-Klyvert, Dental and Oral Tissues. (Lea & Febiger, Philadelphia, 1990).