Assessing tooth wear progression in non-human primates: a longitudinal study using intraoral scanning technology

Technological and methodological considerations

The results of this study provide compelling evidence supporting the applicability of intraoral scanning technology in assessing dental tissue loss in non-human primates. Our results show that the equipment and software development for assessing dental tissue loss in humans can be reliably used in other primate species, in this case a captive baboon sample. Furthermore, our findings indicate an increase in tissue loss with longer time periods, aligning with expectations based on dental wear progression. Contrary to potential concerns regarding the accuracy of these methods in non-contemporary human and non-human primate samples, due to faster levels of tissue loss (Kaidonis, 2008; d’Incau, Couture & Maureille, 2012), our study suggests that such concerns may be unwarranted.

While efforts to minimize alignment issues between scans, such as those made by the creators of WearCompare and similar software, have been largely successful, alignment errors cannot be fully removed due to the absence of fixed reference points in the dentition as teeth wear (O’Toole et al., 2019b; O’Toole et al., 2020). To mitigate this, O’Toole et al. (2020) implemented strict alignment criteria, requiring over 75% of alignment points to be within 25µm of each other. In the present study, the limited wear on the buccal and lingual sides of the teeth meant all alignments were above this threshold. However, as tooth wear progresses over time, achieving precise alignment becomes increasingly challenging. The observed decrease in alignment between the two-year and three-year interval groups, coupled with the decrease in alignment with increase tissue loss (Fig. 4), suggests that the current alignment techniques may surpass the O’Toole et al. (2020) defined alignment limits after approximately four to six years of wear, for the techniques and sample used in the present study. However, a larger time series data set is required to find the actual limit of these alignment techniques. Additionally, the ability to align scans accurately will not just vary between samples due to difference in tooth wear, but also in relation to the resolution and accuracy of the scans themselves (discussed in more detail below).

It is noteworthy that the alignment of the baboon scans in our study exceeded that of clinical human samples used by O’Toole et al. (2020), as evidenced by fewer individuals being excluded due to alignment issues. Several factors may contribute to this phenomenon. Firstly, although the baboons exhibit substantially more occlusal wear than the human sample, it is possible that erosion, which is likely less prevalent in these baboons compared to attrition or abrasion, contributes to preservation of the buccal and lingual tooth positions over time. Secondly, human molars possess a more bulbous morphology than baboon molars, potentially affecting the ease of aligning their crowns. Thirdly, regular tooth brushing in contemporary human samples, especially if brushing follows an erosive challenge, may remove sufficient buccal/lingual tissue over a three-year period (Addy & Hunter, 2003; Voronets & Lussi, 2010; Bartlett et al., 2013), making alignment more challenging compared to non-human primates. Lastly, it is important to note that differences in scanners used, and how 3D models are generated (i.e., the software and algorithms used) and the samples themselves (e.g., artefacts or dental movement/pathologies), may make direct comparisons difficult.

Further research with additional samples is warranted to explore these potential hypotheses. However, even without tooth brushing, movement of food by the tongue and cheeks against the teeth is sufficient to cause a small amount of wear on non-occlusal surfaces (Fox, Juan & Albert, 1996; Ungar & Teaford, 1996), and so in every sample there will be a limit to the interval possible for these techniques. Our study, in conjunction with the findings of O’Toole et al. (2020), suggests that studies aiming to investigate wear progression over extended periods should consider using multiple scans, such as conducting a new scan of the dentition every two to three years. It is also worth noting that it has been possible to assess very low rates of tooth wear in human samples. Consequently, in heavily worn non-human primate samples, shorter intervals between scans may provide valuable insights into short-term wear progress, including seasonal changes in diet or behavior.

Despite the high trueness and precision (and therefore accuracy) of intraoral scanners, ongoing research aims to further enhance scanning technology by addressing potential differences in resulting 3D models based on various factors, including scanner model, color variation, distance from the object, scanning pattern, and ambient temperature (Flügge et al., 2016; Abduo & Elseyoufi, 2018; Kihara et al., 2020; Amornvit, Rokaya & Sanohkan, 2021; Revilla-Leon et al., 2021; Revilla-León et al., 2022; Afrashtehfar, Alnakeb & Assery, 2022). This field of technological advancement is rapidly evolving, with each new model, and software update, bringing significant progress. The model used in the present study, the Medit i700, is widely regarded as extremely reliable for complete arch scanning, with trueness comparable to that of high-end desktop scanners (Borbola et al., 2023). Variation in trueness and precision has however been found to vary in the literature, including in association with scanning conditions (Borbola et al., 2023; Giuliodori, Rappelli & Aquilanti, 2023; Tran et al., 2023; Sorrentino et al., 2024).

In this study, we used the high-resolution (HD mode) setting for all scans. According to the manufacturer, this mode enhances model detail—or resolution—without affecting the specified accuracy (10.9 µm ± 0.98). While precision, trueness and accuracy of intraoral scanners are often reported by manufactures and in publications, resolution receives much less attention (Desoutter et al., 2024). Indeed, it is typically difficult to get information on the resolution of intraoral scanners from manufacturers (e.g., personal communication, 2024, with Medit Support: “the creation of the mesh and mapping for triangulations, this process is proprietary, protected by our patent, and involves the implementation of the software and the functioning of the AI”). How resolution should be interpreted for 3D surface models also varies in the literature, although intraoral scanners produce models consistent with the range of resolutions of 3D dental models created using other scanning techniques (i.e., 10–100 µm; Medina-Sotomayor, Pascual-Moscardó & Camps, 2018; Berthaume, Winchester & Kupczik, 2019; Desoutter et al., 2024).

Studies examining how scanning accuracy and resolution, and associated processing procedures, influence estimates of tissue loss are warranted, similar to those performed in other areas of dental and osteological research (e.g., Profico et al., 2016; Veneziano, Landi & Profico, 2018; Berthaume, Winchester & Kupczik, 2019; Morley & Berthaume, 2023). Direct comparisons between different techniques are also warranted, including to assess if any consistent variation in tissue loss is found in the same samples using different methods (both variation among intraoral scanners, but also with surface models generated through different techniques), and/or processing techniques. Ultimately, a standardized method for researchers to assess accuracy and resolution individually for each study, such as proposed by Desoutter et al. (2024), will be crucial before detailed comparisons across scanners, software and processing techniques can be made.

This study predominantly uses scans of dental casts. Although both methodologies produce high resolution 3D models (as presently defined), molding teeth followed by casting, followed by scanning with an intraoral scanner, will undoubtedly lead to some (even if slight) reduction of scan quality (e.g., through the inclusion/accumulation of artefacts). Additionally, as discussed above for creating virtual 3D models of teeth, molding and casting comes with its own variation in end product depending on materials used and methods invoked (Fiorenza, Benazzi & Kullmer, 2009). But the less-reflective surface of casts may also increase scan detail/quality, even if intraoral scanners can reliably collect data from enamel. Therefore, as observed in clinical contexts over recent years, it is imperative to compare results between research groups using different techniques to confirm any inter-study variation in non-human primate dental tissue loss data. This also includes for different types of samples, including wet (i.e., scans taken on living individuals), dry (i.e., scans of osteological specimens), and casts. The size of the specimen is also something that requires further research, not just with the perspective of tissue loss progression, but also potential scan alignment influences (Fig. 4; see also Balolia & Massey, 2021).

Despite these limitations, resulting variation in tissue loss data between methods is likely to be small. For example, differences in intraoral scanner used, alternative methods can yield similar tissue loss results (e.g., García et al., 2022; Branco et al., 2023; Schlenz et al., 2023). Indeed, any resulting differences should be expected to be minimal, given the demonstrated high accuracy of intraoral scanners in assessing tissue loss, assessed in relation to Micro-CT scanning (Mitrirattanakul et al., 2023). Furthermore, intraoral scanners have proven highly reliable in detecting small amounts of dental tissue loss in a clinical setting (e.g., García et al., 2022; Marro et al., 2022; Schlenz et al., 2022).

Other surface scanners, such as the commonly used Artec 3D Space Spider, have been evaluated for their suitability in dental scanning. In two recent studies comparing intraoral scanners with an Artec 3D Space Spider scanner, one suggested the former as better for scanning teeth (Piedra-Cascón et al., 2021), whilst the other concluded although both produce similar results, the Artec 3D Space Spider should be considered the “gold standard” for creating 3D models (Winkler & Gkantidis, 2020). However, significant negatives of scanners such as the Artec 3D Space Spider have also been pointed out, including its large size and subsequent difficulty in scanning dentitions. Additionally, and most importantly, to produce the “gold standard” results in the Winkler & Gkantidis (2020) study, a coating was required to be sprayed over the specimens, and the authors suggest that without such methods accurate readings are not possible. Such a procedure is unlikely to be permitted for many osteological collections. Therefore, the major benefit of an intraoral scanner is that it is designed with dentitions in mind, and therefore collects images reliably without coating teeth. The small size also makes it easy to move around a dentition whether you are scanning a live individual, a cast, or an osteological specimen. Additionally, in many applications of these methods, a living individual will be involved, at least for one of the time points, meaning a larger scanner, such as an Artec 3D Space Spider would not be feasible.

The versatility of intraoral scanners extends beyond dental tissue loss assessment, as highlighted by recent studies in archaeological osteological collections (Colleter et al., 2023). These scanners offer non-destructive 2D and 3D imaging capabilities, eliminating the need for direct anthropometric measurements and reducing specimen handling. In addition to being less destructive compared to traditional methods involving calipers or molding materials, the 3D models generated by these scanners can be shared, significantly minimizing the handling of fragile specimens (Omari et al., 2021; Colleter et al., 2023). Moreover, the scanning process for osteological or cast dentitions typically takes less than five minutes, making it is also vastly quicker than Micro-CT scans for generating surface 3D models. Intraoral scanners also excel at distinguishing color and texture variations, thereby offering potential use for paleopathology studies. Given their diverse capabilities, intraoral scanners can serve as additional or complimentary tools in primatology, osteoarcheology, and biological anthropology research contexts. Their non-destructive nature, speed, and accuracy may make them useful for various applications beyond dental tissue loss assessment.

Deciding on which intraoral scanner to purchase or use can be a difficult process, with scanners costing from under $10,000 to over $50,000. Although intraoral scanners generally offer high precision and accuracy, variations exist among different devices (e.g., Amornvit, Rokaya & Sanohkan, 2021). Each new model released by manufacturers tends to improve these metrics, yet it is the additional features that often distinguish one generation from the next. These enhancements can include the transition from corded to cordless designs, improved color accuracy, faster scanning speeds, streamlined calibration processes for color recognition, reduced warm-up times, smaller devices or tips, expanded software applications, and improved resolution of the 3D models generated. Many of these advanced functionalities may not be directly applicable, or important, in osteological or archaeological studies due to specific research requirements. A significant advantage of the scanner utilized in our research is the use of free software that supports the export of scan data in multiple formats. Nevertheless, some commercial scanners come with mandatory subscription fees and software that may not be ideally suited for scanning in a non-clinical setting. The high initial expense of these devices might currently limit their widespread adoption in non-clinical fields, but prices are decreasing. Moreover, the growing use of these devices in clinical and laboratory settings presents opportunities for collaborative research. This approach also allows for the separate purchase of scanner tips, thus mitigating the risk of cross-contamination.

Tissue loss in captive baboons

The average tissue loss values produced in this study can serve as a foundational reference for future investigations, whether on captive primates or their wild counterparts. A recent study by O’Toole et al. (2020), based on a clinical study of contemporary humans that had high levels of erosive tooth wear, found relatively uniform levels of tooth wear across the dental arcade. The only molar in O’Toole et al.’s (2020) study with comparable data is the upper first molar, which showed an average total volume of 2.53 mm³ over a mean period of three years. In relative terms, accounting for tooth size, this equates to a mean volume loss per mm² of surface analyzed of 0.03 mm³. Therefore, in our study, the tissue loss observed in the upper second molars of baboons over a similar three-year interval is typically over four times greater in both absolute and relative terms. The average rate of dental tissue loss through wear is likely even lower in most contemporary humans, considering the participants in O’Toole et al. (2020) were considered at risk for high levels of erosive tooth wear. Indeed, although using slightly different techniques, a recent study by Schmid et al. (2024) found less tissue loss in individuals that have undergone orthodontic treatment than the baboons in the present study, even after a period of over nine years. These observations align with the prevalent soft food diet typical in most parts of the world today. In contrast, the baboons in our study showed macroscopically visible tissue loss even in the two-year interval group. This higher rate of tooth wear in the SNPRC baboons, compared to the humans in the dentistry studies, may be attributed to the high-fiber content of the monkey chow lab feed. However, differences in the way baboons and humans masticate foods (including jaw movements, forces and occlusion), may also be important to consider and compare in future studies. Additionally, social behaviors and differences in dental morphology may also be crucial factors to consider.

The comparison of tissue loss between carious and non-carious teeth offers insights into the interplay between pathological conditions and wear dynamics in primate populations. The observed reduction in tissue loss in the opposing molar of some individuals with carious lesions suggests the presence of compensatory mechanisms within the masticatory system. However, in one case, severe tissue loss is evident in the opposing molar, higher than in the carious tooth, and indicates potential atypical wear patterns that may be related to the carious lesion in the opposing tooth. This particular case looks very similar to lesions that may be described as erosion (or a combination of erosion and attrition), in a clinical human setting, although a severe form of dental caries cannot be ruled out. Similarly, in the two cases where clear and large antemortem fractures have occurred (one before the first mold and the other between the two molds), the tissue loss across the occlusal surface is substantially different than in other individuals with no fractures. These observations underscore the complex interrelationships between different types of dental tissue loss, such as erosion, attrition, abrasion, fracture, and caries, and the need for further studies to elucidate these dynamics over time.

Despite all baboons in this study sharing the same diet, we observed significant variations in tooth wear and pathologies. This suggests that differences in physiology, morphology, or behavior—or more likely, a complex combination of factors—might explain these patterns. However, even with a uniform diet, the social dynamics and hierarchy within the group is also important to consider, as such factors may be associated with different amounts of food intake. The way individual baboons handle food can significantly impact tooth wear. This includes their eating duration each time, whether they drop food to the ground (which could introduce dust or grit), and their interaction with water and food (notably, some baboons have been observed softening food by dropping it into water). Some individuals may place non-food items in their mouth and there is also potential for stress-induced tooth grinding.

Physiological and morphological aspects, such as saliva production, the structure and mechanical properties of dental tissues, and genetic factors influencing chewing, will all also have a direct effect on dental tissue loss progression (Lucas, 2004; Buzalaf, AR & Kato, 2012; Mulic et al., 2012; Hara & Zero, 2014; Towle et al., 2023a; Towle et al., 2023b; Towle et al., 2023c). Distinguishing between these various factors can be challenging, as illustrated by the extreme difference in tooth wear of two male baboons shown in Fig. 5. Both these individuals show slight asymmetry in tooth wear between the left and right posterior teeth, but this does not explain the observed differences between these individuals. This highlights that significant variation in tooth wear can occur even with a uniform diet, pointing to the importance of considering food processing habits, masticatory differences, and idiosyncratic behaviors in understanding these patterns. A study utilizing a larger sample size that considers these factors, as well as genetic inputs (e.g., subspecies of P. hamadryas, or in a pedigree analysis that incorporates tooth wear) may allow some understanding into the main driving forces that explain such wide variation in tooth wear despite a uniform diet.

The prominent tissue loss observed on functional cusps supports their additional susceptibility to wear, although this study focused on moderately worn molars. Assessment of tissue loss progression in later stages of wear, including the examination of different types of dental tissue loss such as fractures, is warranted (Towle et al., 2021b; Towle et al., 2023a). It is conceivable that the wear on non-functional cusps may intensify over time, consistent with alterations observed in the plane of occlusion over an individual’s lifespan. Assessing wear progression will allow this to be directly assessed, not only in wild primates, but also in controlled laboratory settings using wear simulation machines. Such techniques can also be used to look at the association between tissue loss and other variables (e.g., microwear and underlying enamel microstructure; e.g., Hua et al., 2015; Towle et al., 2023b). Variation in bite force across the dental arcade, and particularly during the eruption of permanent teeth, can influence wear dynamics over time, potentially reducing tooth wear on individual teeth as more tooth contacts are made (Lucas, 2004).

Analyses of tooth wear progression across primate populations and samples have predominantly involved evaluating tooth wear at a single time point (e.g., osteological specimens), either by correlating tissue loss with the age of individuals, or by contrasting wear patterns across different teeth within the same dentition (e.g., Ozaki et al., 2010; Elgart, 2010; Galbany et al., 2011; Morse et al., 2013; Galbany et al., 2014; Spradley, Glander & Kay, 2016; Pampush et al., 2018; Ungar et al., 2021). These approaches essentially construct a time-series analysis by comparing individuals of varying ages, assuming a continuity in wear patterns across each population/sample. Fewer studies have used more direct longitudinal methodology, tracking wear progression in the same individuals over time, by recording specific wear scores/parameters or the ratio/percentage of dentine exposure, on multiple occasions (Froehlich, Thorington & Otis, 1981; Teaford & Glander, 1996; Phillips-Conroy, Bergman & Jolly, 2000; Cuozzo et al., 2010). This study builds on these studies by introducing more comprehensive techniques for evaluating tissue loss through time. These advancements may be particularly beneficial in research comparing tooth wear with other dentition variables, such as recent studies on dental topography and chewing efficiency (e.g., Venkataraman et al., 2014; Pampush et al., 2018).

The dynamic between dentine and enamel removal once dentine becomes exposed on the occlusal surface can vary substantially and is influenced by factors such as the tooth crown’s position, the mastication cycle, and diet (Lucas, 2004). Moreover, the specific material being masticated affects tissue loss patterns, as the ability of one object to indent another is determined by their relative hardness. However, volume loss is also influenced by other factors, such as toughness, Young’s modulus, and surface characteristics affecting friction (Kendall, 2001). Enamel wear is also heavily influenced by the orientation of the enamel prisms relative to the tooth surface, with prisms perpendicular to the contact surface wearing down more rapidly (Teaford & Walker, 1983; Walker, 1984). External factors such as the presence of dust, silica, and grit significantly impact the patterns of dental tissue loss. All these factors not only affect occlusal wear, but also interproximal wear (Wood, Abbott, & Graham, 1983; Murphy, 1959). Taken together, these factors and variables highlight the complexity of assessing tooth wear in primates, and the tissue loss techniques described in this study may help tease apart the different components involved.

Laboratory experiments utilizing wear simulations and tissue loss assessment methods can also provide valuable insights into understanding dental tissue loss dynamics from a fossil or archaeological perspective. By investigating how tissue loss varies across the dental arcade over time and in different species or groups, researchers can discern patterns that may have genetic, dietary, or environmental origins (Tobias, 1980; Macho & Berner, 1994). For instance, if certain patterns of tissue loss are consistently observed in specific species or populations, it may suggest underlying genetic factors influencing masticatory or tooth morphology, or dietary habits shaping wear dynamics. The techniques described here, originally developed for clinical dentistry contexts, offer a promising avenue for investigating the effects of tooth wear on functional morphology in non-human primates and other species (Ungar & Williamson, 2000). These methods are not limited to assessing occlusal wear but have also demonstrated accuracy in evaluating other types of wear, such as non-carious cervical lesions (Denucci et al., 2024). By conducting research in both laboratory and wild primate settings, scientists can gain a deeper understanding of the development of atypical tooth wear observed in fossil hominin samples, such as “toothpick” grooves (Hlusko, 2003).

The ability to reliably quantify dental tissue loss through non-invasive means offers an innovative approach to investigating wear patterns, adaptive responses, and potential ecological shifts within primate populations. These methods allow us to assess individual cusp wear over time and shed light on how underlying tooth morphology influences wear variation among individuals and populations. Research integrating tooth morphology and cusp size/presence (Selig, 2024) presents a promising avenue for studying the genetic underpinnings of tooth wear in primates. In this context, wear could be considered a phenotype in phylogenetic, hybridization and interpopulation studies. For example, previous studies on baboons, such as those by Hlusko & Mahaney (2003), Hlusko et al. (2004), Ackermann, Rogers & Cheverud (2006) and Hlusko, Lease & Mahaney (2006), have demonstrated strong associations between cusp size and shape, overall molar size, and the presence of supplemental teeth with genetic factors. By examining a much larger sample of these baboons with uniform diets and analyzing variation across individual tooth crowns for dental tissue loss data, valuable genetic insights could be gleaned.

Moreover, other large cast and osteological collections offer opportunities for similar studies, including investigations into tooth wear differences between wild and captive primates. Scanning additional dentitions of wild and captive individuals may also be feasible, either as part of general health checkups, or alongside other research objectives in a primatological context. While full arch scanning may take several minutes to complete, targeting specific areas, such as mandibular left posterior teeth, can significantly reduce scanning time to just seconds.