Intraoral scanners in dentistry: a review of the current literature

The advantages and disadvantages of optical impressions with respect to conventional physical impressions (i.e. impressions made with trays and materials) are presented below and summarised in Table 1.

The main feature an IOS should have is accuracy: a scanner should be able to detect an accurate impression [1‐8]. In metrics and engineering, accuracy is defined as the ‘closeness of agreement between a measured quantity value and a true quantity value of a measurand’ (JCGM 200:2012, ISO 5725–1, 1994). Ultimately, accuracy is the sum of trueness and precision [4‐8]. Trueness, usually expressed in terms of bias, is the ‘closeness of agreement between the expectation of a test result or a measurement result and a true value’ [4‐8]. Precision is defined as the ‘closeness of agreement between indications or measured quantity values obtained by replicate measurements on the same objects under specified conditions’ [4‐8]. Ideally, an IOS should have high trueness (it should be able to match reality as closely as possible). An IOS should therefore be as true as possible, that is, be able to detect any impression detail and permit the establishment of a virtual 3D model as similar as possible to the actual model, and that little or nothing deviates from reality. The only means of calculating the trueness of an IOS is to overlap its scans with a reference scan obtained with a powerful industrial machine (industrial optical scanner, articulated arm, coordinate measuring machine) [4‐8]. After the overlapping of these images/models, powerful reverse-engineering software can be used to generate colorimetric maps displaying the distances/differences between the surfaces of the IOS and the reference model at micrometric level [4]. Precision can be calculated more easily, simply by overlapping different scans/models taken with the same IOS at different times and again evaluating the distances/differences at micrometric level. Technically, an IOS could have high trueness but low precision, or vice versa. In both cases, the optical impressions would be unsatisfactory: this would negatively affect the entire prosthetic workflow, where reducing the marginal gap is the prosthodontist’s major task. Trueness and precision mainly depend on the scanner acquisition/processing software, which performs the most difficult task: ‘building’ the 3D virtual models [1, 2, 4‐8]. The resolution of acquisition, that is, the minimum difference an instrument is capable of measuring (i.e. sensitivity of the instrument) is also important; however, it depends on the cameras inside the scanner, which are generally very powerful.

To date, the scientific literature considers the accuracy of optical impressions clinically satisfactory and similar to that of conventional impressions in the case of single-tooth restoration and fixed partial prostheses of up to 4–5 elements [18, 19, 21, 24, 35‐49]. In fact, the trueness and precision obtained with the optical impressions for these types of short-span restorations are comparable to those obtained with conventional impressions [35‐49]. However, optical impressions do not appear to have the same accuracy as conventional impressions in the case of long-span restorations such as partial fixed prostheses with more than 5 elements or full-arch prostheses on natural teeth or implants [6‐8, 35‐50]. The error generated during intraoral scanning of the entire dental arch does not appear compatible with the fabrication of long-span restorations, for which conventional impressions are still indicated [6‐8, 35‐49].

However, the latest-generation scanners are characterised by very low errors in full-arch impressions [4], and in this sense the data in the literature must be interpreted critically, as preparing and publishing a scientific article generally takes time, whereas manufacturers release new powerful software for mesh construction very frequently.

To date, only a few studies have compared the trueness and precision of different IOS [4, 50‐58]. Almost all are in vitro studies based on models [4, 50‐58], as it is currently not possible to calculate the trueness of IOS in vivo; in addition, these studies have quite different experimental designs [50‐58]. Some focused on the accuracy of the IOS in dentate models [50, 52, 53, 55‐57], while others evaluated the accuracy of the IOS in oral implantology [4, 51, 54, 58]. Regardless, the upshot of these studies is that different IOS have different accuracy; therefore, some devices seem to have more indications for clinical use (for making impressions for fabricating long-span restorations) while others appear to have more limited clinical applications (for making single or short-span restorations) [50‐58]. It is very difficult to compare the results (in terms of trueness and precision) of these studies, as scanners have different image-capture technologies and may therefore require different scanning techniques [4, 54, 59, 60]; unfortunately, little is known about the influence of scanning technique on the final results [59‐61], and the scientific literature should address this topic in the coming years.

Trueness and precision, however, are not the only elements that can differentiate the devices currently available commercially [1, 2, 4, 7, 34, 54, 59, 62]. A whole series of elements (necessity of opacization with powder, scanning speed, tip size, ability to detect in-colour impressions) differentiate IOS in terms of their clinical use [1, 2, 4, 54, 62]. In particular, scanning systems can differ based on the possibility of whether there is a free interface with all available CAD software (open versus closed systems) and the purchase/management costs [1, 2, 4, 54, 62].

The need for powder and opacization is typical of the first-generation IOS; the more recently introduced devices can detect optical impressions without using powder [2, 4, 34, 62, 63]. Technically, a scanner that allows the clinician to work without opacization should be preferred; in fact, powder may represent an inconvenience for the patient [2, 4, 34, 62, 63]. In addition, applying a uniform layer of powder is complex [2, 34, 62, 63]. An inappropriate opacization technique may result in layers of different thicknesses at various points of the teeth, with the risk of errors that reduce the overall quality of the scan [2, 34, 62, 63].

Scanning speed is certainly a matter of great importance for an IOS [2, 4, 50, 54, 62]. IOS have different scanning speeds, and the latest-generation devices are generally faster than the oldest ones. However, the literature has not clarified which device can be more efficient: in fact, the scanning speed does not depend only on the device, but largely on the experience of the clinician [2, 4, 34, 50, 54, 62].

The size of the tip plays a role as well, especially in the case of second and third molars (i.e. the posterior regions of the maxilla/mandible) [2, 4, 12‐18, 34, 62]. A scanner with a tip of limited dimensions would be preferable for the patient’s comfort during the scan; however, even scanners with more voluminous tips allow excellent scanning in posterior areas [2, 4, 12‐18, 34, 62].

The possibility of obtaining in-colour 3D models of the dental arches represents one of the latest innovations in the field of optical scanning [1, 2, 4, 28, 34, 64]. To date, only a few IOS can make in-colour impressions. Generally, colour is simply added to the 3D models derived from the scan, overlaying these with high-resolution photographs. The information on colour is meaningful especially in communication with the patient, and is therefore of less clinical importance [1, 2, 4, 28, 34, 64]; in the future, it is possible that IOS will include functions that are now the prerogative of digital colorimeters.

Finally, an IOS should be able to fit in an ‘open’ workflow and should have an affordable purchase and management price [1, 2, 4, 54]. Ideally, an IOS should have two outputs: a proprietary file with legal value, and an open-format file (e.g.. STL,. OBJ,. PLY). Open-format files can be immediately opened and used by all CAD prosthetic systems [1, 2, 4, 54]. In such cases, the literature generally refers to an ‘open system’ [1, 2, 4, 54]. The advantage of these systems is versatility, together with a potential reduction of costs (there is no need to buy specific CAD licenses or to pay to unlock the files); however, a certain degree of experience may be required, initially, to interface the different software and milling machines [1, 2, 4, 54, 62]. This problem does not arise in the case of IOS within a ‘closed system’. Such scanners have as output only the reference proprietary (closed) file, which can be opened and processed only by a CAD software from the same manufacturing company. The inability to freely dispose of. STL files, or the need to pay fees to unlock them, certainly represents the main limits of closed systems [1, 2, 4, 54, 62]. However, the inclusion within an integrated system may encourage workflow, especially in the case of less experienced users. In addition, some closed systems offer a complete, fully integrated digital workflow, from scanning to milling, and provide chair-side solutions. Finally, converting files (e.g. the conversion of proprietary files to open formats) may result in loss of quality and information [2, 62].

The most important features an IOS should have are summarised in Table 2.

IOS are of great utility and are applied in various fields of dentistry, for diagnosis and for fabricating restorations or custom devices in prostheses, surgery and orthodontics [65‐132]. IOS are in fact used for acquiring 3D models for diagnostic purposes [2, 4, 6]; these models can be useful for communicating with the patient [2, 6]. Diagnosis and communication are not, however, the only fields of application for IOS. In prostheses, IOS are used to make impressions of preparations of natural teeth [6‐8, 65‐88] for fabricating a wide range of prosthetic restorations: resin inlays/onlays [65, 66], zirconia copings [67, 68], single crowns in lithium disilicate [69‐74], zirconia [19, 75‐77], metal-ceramic [78] and all-ceramic [79‐81] as well as frameworks and fixed partial dentures [82‐87]. Several studies [69‐81] and literature reviews [88] have shown that the marginal gap of ceramic single crowns made from intraoral scans is clinically acceptable and similar to that in crowns produced from conventional impressions. The same considerations can be extended to short-span restorations such as fixed partial dentures of three to five elements [36, 82‐87], obviously considering the differences stemming from the different accuracies of various IOS. To date, the literature does not support the use of IOS in full-arch impressions: several studies and literature reviews have shown that the accuracy of IOS is not yet sufficient in such challenging clinical cases [7, 8, 35, 37, 39].

In prosthodontics, IOS can be successfully used to capture the 3D position of dental implants and to fabricate implant-supported restorations [4, 14, 17, 18, 21, 24, 47, 51, 54, 58]. The 3D position of the implants captured with the IOS is sent to the CAD software, where the scanbodies are coupled with an implant library, and the desired prosthetic restorations can be drawn within minutes; this restoration then can be physically realised by milling through a powerful CAM machine using ceramic materials [89‐119]. At present, implant-supported single crowns [21, 22, 89‐104], bridges [104‐113] and bars [114‐116] can be successfully fabricated from optical impressions. Similar to what the literature has found for natural teeth [6‐8, 35, 37], the only apparent limitation to the use of IOS in implant prosthodontics is that of long-span restorations on multiple implants (such as long-span bridges and fixed full arches supported by more than four implants): at least, this is what emerges from the most important reviews [39, 117, 118] and from different in vitro studies on trueness and precision, which indicate that conventional impressions are the best solution for these challenging clinical situations [4, 49, 54, 58].

At present, only a few studies have addressed the use of IOS for fabricating partially [119, 120] and completely [57, 121] removable prostheses; in particular, this last application still presents some issues due to the absence of reference points and the impossibility of registering soft tissue dynamics. However, IOS can be successfully used for digital smile design applications [122], post and core fabrication [123] and for fabricating obturators, in complex cases [124, 125].

Dentogingival model scanning can be superimposed onto files from cone beam computed tomography (CBCT) too, via specific software to create a virtual model of the patient [126‐130]. This model is used for planning the positioning of the implants and to draw one or more surgical stents useful for placing the fixtures in a guided manner [126‐130]. The use of IOS in this sense has supplanted the old technique of double scanning with CBCT only, which was based on radiologic scans of the patient and of the patients’ plaster models. In fact, the scanning resolution of CBCT is lower than that of IOS; the use of IOS therefore allows the detection of all details of the occlusal surfaces with greater accuracy. This can make the difference in, for example, the preparation of tooth-supported surgical templates. However, care should be taken, as the use of IOS in guided surgery is only in its infancy.

Finally, IOS represent a very useful tool in orthodontics for diagnosis and treatment planning [3, 5, 6, 12, 15, 16, 25, 27, 131, 132]. In fact, optical impressions can be used as a starting point for the realisation of a whole series of customised orthodontic devices, among which aligners should be mentioned [3, 5, 6, 12, 15, 16, 25, 27, 131, 132]. In the coming years, it will be probable that almost all orthodontic appliances will be designed from an intraoral scan, so they will be entirely ‘custom’ and adapted to the patient’s specific clinical needs [3, 5, 6, 12, 15, 16, 25, 27, 131, 132].

The most important clinical indications and contraindications on the use of IOS are summarised in Table 3.