From Volume 38, Issue 10, December 2011 | Pages 679-690
This article aims to describe the current status of 3-dimensional (3D) imaging in dental practice. Advances in this field have made 3D imaging far more accessible in all dental fields. This paper describes methods of imaging dental hard and soft tissues and their clinical uses. In addition, the potential advantages and disadvantages of various systems are discussed, as well as expected future developments.
Technology in dentistry continues to develop and improve at what seems to be an ever increasing pace. Whereas in the past we relied on conventional analogue records, such as study and working models, along with 2D radiography, dentistry is now well and truly embracing the digital age. In particular, advances in 3D digital technology mean that we are able to record, construct and reconstruct both 3D intra- and extra-oral images.
This paper aims to provide an overview of the current uses of 3D imaging in dentistry, exploring the new advances along with their advantages and disadvantages. In particular, it will focus on 3D radiography, CAD/CAM, model and facial scanning.
Until relatively recently the only way to get good quality 3D radiographic scans was to use conventional computerized tomography (CT). However, not only were the machines not readily available, except in hospital radiography departments, but their use was associated with a high radiation dose, high cost, and relatively poor resolution when used for dental purposes. The dental images produced could also be difficult to interpret.
The introduction of dedicated dental cone beam computed tomography (CBCT) machines has overcome many of these obstacles. These new machines comprise a linear array of multiple detectors, allowing multiple radiographic slices of the patient to be taken simultaneously.
In the late 1990s, Italian and Japanese groups,1,2 working independently of each other, developed new tomographic scanners, known as CBCT, specifically for dental and maxillofacial use. CBCT differs from medical CT imaging in that the whole 3D volume of data is acquired in a single sweep of the scanner, using a simple and direct relationship between sensor and source. The X-ray beam is cone-shaped and captures a cylindrical or spherical volume of data, known as the field of view. The size of this primary field of view is dependent on the particular make of the machine. There are ‘large volume’ scanners, eg i-CAT (i-CAT, Imaging Sciences International, Hatfield, PA, USA) and NewTom (Quantitative Radiology, Verona, Italy), both of which are capable of capturing the entire maxillofacial skeleton. In addition, the field of view of the i-CAT can be adjusted to capture only the maxilla or mandible. There are also limited volume CBCT scanners which, as the name suggests, capture smaller volumes of data, for example a 30 mm x 40 mm cylinder, and might include just 2–3 teeth (Accuitomo, J Morita Corporation, Osaka, Japan). Not only are these 3D images of a similar size to periapical radiographs, but they are of similar image quality. CBCT has a much lower radiation dose compared to conventional CT and, in some cases, it is a similar dose to that of a DPT or set of bite-wings. However, the effective dose from the large volume machines can range from 2.7 up to 25 times of a conventional DPT.3
Unlike conventional CT, the contrast resolution of CBCT is limited to the densities of calcified structures, such as bone. Although the interface between soft tissue and air is readily identifiable, differentiation between various soft tissues can be difficult. As with all radiographs, but in particular 3D imaging, CBCT is affected by metallic objects, such as amalgam restorations, alloy crowns, implants and, to a lesser extent, root-filling materials.
Sophisticated software is applied that allows the huge volume of data collected to be broken down and processed and reconstructed into a format that closely resembles that produced by medical CT scanners. The data can also be used to construct panoramic radiograph and lateral cephalostat images (Figure 1).
Processing and reconstruction times vary but usually take less than 5 minutes. From the image the clinician can select the relevant sagittal, coronal or axial slices, thereby eliminating superimposition of anatomical structures, so that roots or periapical tissues can be visualized separately and in all three planes without superimposition of zygomatic buttress, alveolar bone and adjacent roots. Enhancement of the image is also possible, eg magnification, visual enhancements (grey scale, brightness and contrast levels); measurements and annotations can also be made (Figure 2).
As a result of these capabilities, a number of potential dental applications of CBCT scanning have been suggested and embrace almost every dental specialty.
In vitro studies have been performed to assess the use of CBCT in the diagnosis of both caries and periodontal disease. For caries diagnosis, it has been found to be equal to conventional bite-wings for proximal caries detection.4
However, the cost, both in terms of equipment and time, means that it is unlikely to supersede conventional bite-wing radiography in the near future. Periodontal applications of CBCT images include 3D visualization of intrabony defects, dehiscence/fenestration defects and molar furcation involvement5(Figure 3). It may be that CBCT has the potential to replace intra-oral imaging for the assessment of periodontal architecture, however, clinical studies would be needed in order to support this.
Endodontic imaging can also be achieved using CBCT. Figure 4 demonstrates imaging nerve canals prior to preparation and obturation.
The assessment of the jaws prior to the placement of dental implants is probably the most common dental indication for CBCT at present in restorative dentistry. Unfavourable implant placement may adversely affect the long-term success and aesthetics of the implant-supported prosthesis and therefore careful presurgical planning is essential to achieve optimal results. Standard planning for placement of dental implants has traditionally involved using mounted study models to construct a model-based surgical guide. Both conventional dental radiographs and CBCT scans will help the clinician determine how the underlying bone relates to the potential implant. However, 3D imaging and software advances have not only allowed the clinician to treatment plan and place implants more predictably,6 but the implants can also be placed ‘virtually’ (Figure 5).
Once the virtual position has been determined, the computer can then be used to create a customized surgical template to guide implant placement. In this way, implants can be placed in optimal positions more accurately, predictably and safely as vital structures (adjacent tooth roots, inferior dental nerve) can be identified and avoided.
Traditionally in orthodontic practice, treatment planning and the monitoring of treatment progress has been based on information gathered from 2D radiographs, such as the lateral cephalogram, DPT and nasal occlusal radiograph. Although in most cases the information gathered is perfectly adequate, there are instances where 3D information would be of benefit. Examples include the assessment of asymmetries, the extent of the buccal bone and the condition of roots prior to rapid maxillary expansion, and in the determination of the position of unerupted canines and condition of adjacent tooth roots.
It is this latter function, ie the locating of the position of impacted teeth, in particular permanent canines, which is perhaps the most discussed use of CBCT in orthodontics. The prevalence of impacted canines is 0.9–3% and, of these, approximately 85% are palatal. The conventional method of identifying their position is to use parallax, which is reliant on the tube shift between two radiographs (usually two periapicals or a nasal occlusal and DPT radiographs). Although in most instances this method works well, there are times when the position of the impacted canine is still not easily determined.7 With CBCT the precise localization of the impacted tooth can be visualized, in particular the proximity of adjacent incisors roots and any associated root resorption, the inclination of the long axis of the tooth to the occlusal plane, the amount of bone covering the tooth and the size of its follicle can also be accurately assessed (Figures 6 and 7). Certainly, the sensitivity of CT in assessing the presence of root resorption has been found to be much higher than with conventional plane films, where identification of labial or palatal resorption can be problematic.8
Although CBCT supplies useful information, owing to the radiation doses involved, its use must still be justified. White and Pae have published an algorithm for orthodontic patient selection9 (Figure 8). In summary, CBCT was recommended in cases of facial asymmetry or disharmony, sleep apnoea and impacted canines, otherwise the usual lateral cephalometric and panoramic views were recommended.
It is perhaps in surgery that CBCT will have its widest and most immediate application. As with the location of impacted teeth in orthodontics, in surgery CBCT is useful locally in identifying the mandibular nerve and evaluation of lower third permanent molars and in their relationship to the mandibular nerve (Figure 9).
It is also used for the assessment of bony pathology, such as cysts and tumours, the temporomandibular joint and fractures within the facial skeleton.
Several examples of the use of CBCT in diagnosing the extent of dento-alveolar trauma have been documented.10,11 The obvious advantage of CBCT is that, from just one scan, the exact nature, severity and number of hard and soft tissue injuries can be visualized from any number of angles. CBCT can also be used with other diagnostic and planning software used for orthodontic assessment and analysis, eg Dolphin 3D (Dolphin imaging, Chatsworth, Calif), to enhance the planning process in orthognathic surgery (Figure 10). The 3D hard tissue image can be combined with 3D soft tissue facial scans captured using a facial scanner so that a prediction can be made as to the effect of any planned jaw surgery.
A CBCT scan can also be used in craniofacial surgery to produce accurate laser generated or stereolithographic models in epoxy or acrylic resin of part or all of the jaws. These in turn can be used to create diagnostic and surgical implant guidance stents (Virtual Implant Placement, Implant Logic Systems, Cedarhurst, New York; Simplant, Materialise, Leuven, Belgium), to assist in the computer-aided design and manufacture of implant prosthesis and provide surgical simulations for osteotomies and distraction osteogenesis procedures (Maxilim, Materialise, Leuven, Belgium).
CAD/CAM (Computer Aided Design/Computer Aided Manufacture) was first introduced in dentistry in the late 1970s and, at the time, there were concerns regarding the quality of the restorations produced and also the cost in comparison to conventional laboratory procedures.12 This has changed with time to the point where restorations produced using CAD/CAM are beginning to supersede the more traditional casting techniques. CAD/CAM is an all-embracing term and there are several different systems on the market within dentistry which utilize, in particular, different methods of image capture and subsequent manufacture of the restoration.
Three-dimensional surface imaging is performed using 3D scanners; these collect digital data on the shape and size of the object in question. This data is processed to create a ‘point cloud’ representing the 3D co-ordinates of the digitized surface.
The density of the point cloud determines the accuracy of the imaged surface. Missing data within the point cloud, for example, an undercut can pose difficulties when accurately trying to replicate the original imaged surface. However, to improve the visualization process, a computerized polygonization of the point cloud is used to fabricate a virtual surface from the data. The resulting polygon mesh is composed of a series of flat polygons, but still does not provide the smooth surface we are used to viewing. For this to occur, complex algorithms are used to surface render the polygon mesh. There are two main categories of 3D scanner; contact and non-contact.
Most contact scanners are CMM (co-ordinate measuring machines) that move a measuring probe over a surface and determine the co-ordinates of the points that make up that surface. The stylus tip of the measuring probe records x, y and z co-ordinates, creating the point cloud. This type of scanner can only be used on hard surfaces, such as dental stone, as a soft impression surface will deform or wear when in contact with the probe. The Incise contact probe scanner (Renishaw, Wooton-Under-Edge, Gloucestershire, UK) and Procera (Nobel Biocare, Goteborg, Sweden) are examples of contact scanning in CAD/CAM. A conventional impression is taken and the stone model is scanned using a contact probe scanner. The 3D information is then sent to a milling facility for a zirconia coping to be made, upon which a final crown is constructed by a technician.
The advantage of a non-contact scanner is that an impression or even the dental tissues can be scanned directly. An example of a non-contact scanning is a 3D laser scanner. This employs the principle of triangulation, as used in cartography, to obtain a 3D surface image. A laser beam is projected on to the surface of an object and a camera-like device, such as a charged-couple device (CCD) or position sensitive detector is used to record the location of the point at which the laser beam strikes the object. Since the angle of the laser and camera are known, the position of the surface can be calculated using triangulation. Beam reflection and stray light can, however, affect accuracy. In addition, most laser scanners use a series of parallel laser beams to record the surface, as this reduces the data capture time. However, wherever there is a sudden change in the object's surface, for example an undercut or void, the first laser line passes out of sight of the detector or camera and, when the second laser line arrives at the void, it is mistaken for the first. This can create errors when the image is reconstructed by the computer software. In the case of the original Cerec (Sirona, Bensheim, Germany), conventional (line-by-line) laser scanning of the tooth was carried out. This was not good for undercut areas and therefore several scans from different perspectives were required in order to create a model of the tooth with minimum ‘voids’. This has now been superseded by the Cerec AC introduced in 2009. This technology uses BlueCam LED in order to scan the image. The advantages of this are that it provides consistent illumination of the area, with no darker areas at the periphery of the scan, and only one scan of the tooth is required. Other advantages include a faster scanning speed and higher precision due to the shorter wavelength of the blue light.
Alternatively, the impression is sent to the dental laboratory where it is scanned directly, for example with a non-contact structured light method, eg inEos (Sirona, Bensheim, Germany). The structured light technique uses a well characterized simple image which is projected on to the surface to be scanned. The image is usually black and white lines or a checkerboard of black and white squares. Since the surface to be scanned is not completely flat, the image appears distorted and does not appear as lines or squares, but is a composite of the original image and the surface topography. The pattern images captured by the CCD camera are digitized and a complex algorithm is used to decipher the surface topography from the original surface and projected image. The advantages of such a system are speed and accuracy.
Finally, recent developments in dental CAD/CAM have meant that impressions are no longer necessary in certain situations. An intra-oral scan is taken with a handheld scanner and the 3D information is again sent to a facility that will construct the final restoration or framework for the final restoration using a milling machine, eg Lava COS wand, 3M ESPE, Minnesota, USA (Figure 11).
The Lava COS wand contains a highly complex optical system comprising multiple lenses and blue LED cells. The system captures the visual images in just seconds and the information can be modelled in real time. This means that the images being captured in the mouth can be displayed concurrently on to the touch screen monitor, allowing the clinician to assess when sufficient information has been captured.
Other examples of these complete systems include Cerec AC and E4D (E4D, Richardson, TX, USA). An intra-oral scan is again taken with a handheld scanner and the final restoration is milled at the chairside before being fitted to the patient, usually at the same visit. Sending an impression to the laboratory has the inherent drawbacks associated with patient comfort and postage of the impression. However, there is the advantage of not having a large initial outlay for components, as with the chairside system. These complete systems are expensive to purchase but have the advantage of immediacy.
CAD/CAM restorations are associated with optimum aesthetics. The modern systems offer an extensive range of materials, including oxide ceramics, giving excellent aesthetic results that cannot be manufactured by conventional means.13 In addition, some systems have the additional benefits of allowing a single visit procedure.
CAD/CAM systems are improving all the time, with improved precision of final restorations and increased user-friendliness of equipment. The quality of the restorations is consistently high, owing in part to the uniformity in production technique. There are still drawbacks, notably the high cost of the technology, but opinion is no longer divided on the benefits of CAD/CAM and the market continues to expand.
Digital study models for use in orthodontics were introduced commercially in 1999 by OrthoCad (Cadent, Fairview, NJ). Since this time several other companies have entered the market, including OrthoCad, 3-Shape (Copenhagen, Denmark), E-models (Geodigm Corporation, Chanhassen, Minn), Orthographics (Ortho Cast, High Bridge, NJ) and DigiModel (OrthoProof). As in restorative dentistry, digital study models can be produced using a number of different methods. These include:
In the majority of cases, impressions are still required to be sent to the laboratory for scanning. Depending on the scanning system, the impressions may be scanned directly, or the plaster models are poured and then scanned. Related software then allows the clinician to view the digital model chairside (Figure 12).
There are several proposed advantages to digital models over conventional plaster models. Replacement of conventional plaster models with their digital equivalent not only reduces the need for dedicated storage facilities, but also reduces the risk of model damage. Files are easily backed up to ensure no loss of valuable data.
Another major advantage is the ability to share files easily and quickly with colleagues. The models can easily be viewed from several geographic locations rather than having to duplicate and post conventional plaster models.14 These files are also well suited to modern practice management systems, allowing digital study models to be kept alongside other digital records, eg photographs and radiographs. When required, the virtual digital models can still be reverse engineered back to stereolithic hard copies.15
Using dedicated software, the digital image can be viewed on-screen from any angle and can also be manipulated in order to create a diagnostic setup.16 However, the digital model needs to be accurate, quick and easy to use if it is to supersede the plaster model. Several studies have now been carried out examining the accuracy and reproducibility of digital models versus conventional models. Whereas studies have found good intra-operator agreement for linear measurements performed on both digital and plaster models, early studies found that measurements made using digital callipers on plaster models were slightly more accurate than those made on digital models.17,18,19 More recently, accuracy and reproducibility was found to be better with digital models than with conventional plaster models, which can be attributed to recent advances in soft- and hardware in this rapidly changing field.20 Similar results have been found when comparing digital versus plaster for Bolton analyses and PAR scoring. Although statistically significant differences were found between the two types of model, they were not clinically significant and so it is unlikely to affect treatment planning, or final outcome scoring.17,21 Interestingly, sending alginate impressions by post and therefore delaying pouring by 3–5 days has not been found to affect the accuracy and quality of the plaster models subsequently produced.20
There are, of course, downsides to any new technology, notably cost implications. There is also reliance on the laboratory having sufficient expertise in information technology to support the clinician.
For now, the majority of systems still require impressions to be taken. However, this is a rapidly changing field and, as technology improves, digital models have the potential to replace fully the conventional plaster model.
Assessment of the face has traditionally been made using 2D photographic records, where multiple views of the face need to be obtained. The advent of affordable digital cameras, PCs with increasing computing power and storage capabilities, and developments in software mean that it is now possible to obtain 3D facial images in milliseconds and with just one exposure. Like 3D image acquisition for study models, facial scans can be created using one of several different techniques.
This is a technique whereby data is captured using two or more cameras fixed at a set distance in order to record images of the same object from different aspects. The images are then combined to produce a 3D image.
Eye-safe laser beams are projected on to the face and the reflected light is captured using an appropriate digital camera. A pair of scanners is required to take simultaneous scans of left and right sides of the face. An example of laser scanners on the market is the Konica Minolta VIVID range (Ramsey, NJ), which uses a slit beam laser. This product has a stated accuracy of 0.1 mm and scans are usually completed in a few seconds. Clinical studies have found the reproduction of facial morphology to be accurate to within 0.85 mm.22
Here a grid or pattern of light is projected on to the face. The distortion of the projected pattern on the surface is recorded by two or more cameras and the data is reconstructed to produce a 3D image.
A combination of stereophotogrammetry and structured light techniques can also be employed and there are currently a number of systems on the market. The 3dMDface™ (3dMD, London, UK) system uses a random light pattern projection on to the object and the images are recorded by two colour and four infra-red cameras (Figure 13).
This is suitable for use with children as it captures the 3D image within 1.5 milliseconds and accuracy has been found to be comparable to direct clinical measurements. Several other systems have been developed such as Di3D (Dimensional imaging, Hillington Park, Glasgow, UK) which also reports highly accurate results (0.4 mm and 0.2 mm accuracy levels, respectively).23
3D facial scanning has a number of possible applications in dentistry but in particular Maxillofacial Surgery and in Orthodontics.
Within the field of cleft lip and palate, 3D facial imaging has been used to help record and measure facial asymmetry and to evaluate the effects of surgery, such as the improvement in facial symmetry following alveolar bone grafting.24,25
In the case of orthognathic surgery, 3D facial scans can be used either in isolation or, preferably, in combination with 3D radiographic views, both pre- and post-operatively. Pre-operatively it can be used to help in the diagnosis of facial deformity and to help predict the effects of various surgical approaches (Figure 14).
Post-operatively it can be used to assess the soft tissue changes that occur as a result of surgery and even to assess the effect of various drug regimes used during surgery on subsequent facial swelling.26 Most recently has been the possibility of linking 3D hard tissue imaging with 3D facial imaging to create a virtual 3D patient model to allow orthognathic assessment and planning in 3D, for example using the Maxilim (Medicim NV, Belgium)27 and Anatomage systems (San Jose, USA).
Owing to the non-invasive nature of 3D facial scanning, it is possible to make multiple consecutive scans without risk to the patient, making it an ideal tool for assessing facial growth over time.28 Similarly, the short- and long-term effects of orthodontic treatment, such as functional appliance therapy,29 and extraction versus non-extraction fixed appliance therapy,30 can be studied.
This paper gives an overview of advances in scanning techniques available for use in dentistry. Some techniques, eg CAD/CAM, have been available for quite some time now and continue to evolve. Other techniques are newer and have yet to achieve the same degree of popularity. However, technology continues to develop at a pace and all of us are likely to see digital 3D technologies playing some part in our working lives in the future, if not already.
