Specimens

Nine horses of different breeds (eight warmbloods, one standardbred) were examined to acquire CT and high-field MRI scans of selected equine cheek teeth, their periodontal tissues and adjacent structures. The horses were clinic-owned (TiHo Hannover, Clinic for horses, Germany). The study population was divided into three groups according to age classes: Group A “young” (2–5 years), B “middle-aged” (6–15 years) and C “old to senile” (≥ 16 years), each with n = 3 horses. The age of the selected horses ranged from 2.4 to 21.7 years (median 8.5 years). All horses were subjected to euthanasia on human grounds for non-dental reasons and for health purposes not related to this study. When evaluating the CT and MR images, the Triadan system was used for numbering the cheek teeth [31]. A total of 36 maxillary cheek teeth were examined. The sample pool included eighteen Triadan 08 s and eighteen Triadan 09 s. Owing to eruption time, dental age was determined by subtracting the earliest possible eruption age of the tooth from the age of the horse [32]. An eruption time of 3.5 years for the 08 s and six months for the 09 s was assumed [33]. The population included selected teeth between 0.9 and 20.95 years (median age 7.95 years).

Diagnostic imaging techniques and image evaluation

The CT image acquisition was performed with a Brilliance™ CT – Big Bore Oncology Scanner (Philips Medical System, Best, The Netherlands). Images of the head were acquired using a modified dental scanning protocol with 140 kV and 300 mAs. A Philips Achieva™ 3.0TX-Series® was used for MRI acquisition. Surface coils (Philips SENSETM FlexM® and Philips SENSETM FlexL®) were positioned around the region of interest (ROI), between the rostral margin of the facial crest and the orbital cavity. The MRI sequences obtained were: T1 weighted (T1w), T2 weighted (T2w), proton density weighted (PDw) and PD fat-suppressed short-tau-inversion-recovery (STIR) images.

Visualisation and measurements

The CT and MR images were examined and visualisation of the dental and adjacent structures depicted was performed. All examinations were observed by M.H., an experienced radiologist, A.B.-Z., Diplomate EVDC (Equine), and C.S., a trained veterinarian.

Special attention was paid to age-related changes. The positional relations between the dental roots and the floor of the paranasal sinuses were described. Pulpar sizes and the positional relation between the dental alveoli and the infraorbital canal (IOC) were portrayed. Each pulp was measured in the mid-tooth section of the cheek tooth selected (Fig. 1a) after each tooth’s half-length was determined in the transverse MRI scans. The distance between the dental alveoli and the IOC was portrayed in transverse sections of the CT images. Therefore, the line between the dental alveolus above the pulp position five and the inner border of the IOC was measured (Fig. 1b). Pulpar sizes and the distance between the IOC and the dental alveoli were measured by one of the examiners (C.S.) using easyVET image editor (easy Vet, IFS Informationssysteme GmbH, Hannover, Germany). All measurements were repeated three times under the same conditions and the median value for each pulp or each distance (between the dental alveoli and IOC) was obtained. Moreover, the IOC’s shape was assessed.

Statistical analysis

Data were collected on spreadsheets (Excel® 2010, Microsoft® Corporation Redmond, Washington, USA). SAS® software (SAS Institute, Cary North Carolina, USA) was used for statistical analysis and GraphPad Software, Inc.® (La Jolla, California, USA) was used for graphical and statistical representations. Data were tested for normal distribution with the Kolmogorov-Smirnov test. Correlation between dental ages and pulp sizes or alveolar-infraorbital distance was tested by Spearman’s rho. A Wilcoxon matched pairs signed rank test was chosen to calculate whether there were significant differences in the pulp sizes or alveolar-infraorbital distances between both sides of the maxilla. The Kruskal-Wallis test and Dunn’s multiple comparison test were performed to validate significant differences of pulp dimensions between different age groups. A p value <0.05 was considered statistically significant.

Imaging modalities

Anatomical landmarks and imaging advantages comparing CT and MR images regarding the maxillary 08 s and 09 s and their adjacent structures were evaluated and summarised in Tables 2 and 3. Acquisition by CT proved to be the best imaging technique for the representation of the bony non-dental structures (alveolar bone, cortical and cancellous maxillary bone and the IOC). These bony structures were well delineated against the low attenuated soft tissues. Dental enamel, dentine and cementum were also detected as structures of specific density, marked with different greyscales by CT imaging, but the differentiation between them was not as clear as the line between the alveolar bone and the periodontal space due to their different opacity. Regarding the endodontic system, CT images revealed not only hypodense pulp horns and chambers, but also hypodense funnel-like infundibula. The infundibula could not be detected on the MR images. Marked hypoattenuating stripes were visible adjacent to the infundibula in some of the CT slices, that could be determined as gas. Compared with slightly blurred pulpar CT imaging, the pulp tissue appeared as bright, sharp, hyperintense pulp on the MR images. Other soft tissues, such as the periodontal ligament, respiratory mucosa and the infraorbital nerve, were also better visualised with MRI compared to CT images.

Age-related pulpar sizes

Negative correlation between dental age and pulpar dimensions was found (r = −0.9). Pulpar size decrease can be explained by dental aging in 71% of the teeth examined: the older the selected cheek teeth of clinically healthy horses became, the smaller the measurements of the pulps (Fig. 2). The sizes of all five pulps varied with age between 14.31 and 1.7 mm for the 08 s and between 14.05 and 1.3 mm for the 09 s. Other pulpar sizes are documented in Table 5. Comparing the pulp horns, pulp 4 was the largest and pulp 5 was the smallest, except for two teeth (Triadan 109) of age group C, where pulp 3 showed the smallest size. Whilst young teeth showed wide pulpar size variations, the range of size differences regarding all pulps became smaller with increasing age; this is visualised by the boxes in Fig. 4. There was no significant difference in the pulpar dimensions between both sides of the maxilla. Significant differences were measurable for the same pulp horn position in all teeth by comparing of the pulp sizes between age group A and C. Pulp 4 also showed a significantly smaller pulpar size between group A and B (Fig. 3). The percentage reduction of mean pulpar size comparing both age classes was between 40 (for all P5) and 56% (for all P4). Pulpar size decrease was smaller considering the same comparisons between the pulps of group B and C, with values ranging from 42 (P3, P2) to 53% (Fig. 4).

	P1	P2	P3	P4	P5
n	35	36	34	36	36
Mean (mm)	4.89	5.81	5.57	7.32	3.45
s.d. (mm)	2.59	3.26	3.42	4.78	1.65
Median (mm)	3.89	3.75	5.24	5.67	3.17
Range (mm)	7.15	8.76	10.09	13.02	5.30
xmin – xmax (mm)	2.19–9.34	2.16–0.91	1.3–11.39	1.28–14.31	1.57–6.87

Common pulp chambers

The common pulp chambers (CPCs) were evaluated in dorsally orientated PDW sequences of MR images. A CPC was encountered in all cheek teeth of age group A (n = 12). The median dental age for these cheek teeth was 2.025 years and the maximum age was 3.75 years. Cheek teeth of the other age groups showed no CPC.

Positional relations between the maxillary sinus and the dental alveoli

Only two dental alveoli (upper 08 s of age group A) out of all 36 cheek teeth were not located below the maxillary sinus. Positional relations of the structures could be seen in both imaging modalities. Whilst CT images highlighted the bony alveoli, the MRI scans delineated the differentiation of sinus and alveolus through the hyperintense periodontal ligament (PDL) and the sinus mucosa. Eight of the 34 teeth roots located within the boundaries of the sinus cavity were only located below the maxillary sinus with their caudal aspect (08 s), including pulp horn 2 and 4. Fifty percent of these teeth (n = 4) partly contacting the sinus floor contained a CPC. The percentage of teeth that were positioned below the maxillary sinus with all five pulps (n = 26) consists of 08 s (n = 8) and 09 s (n = 18).

Age-related changes of the infraorbital canal and its distance to the dental alveoli

Age-related changes of the IOC were analysed in transverse CT scans showing the positional relation between tooth and bony canal, on the one hand, and the canals’ shape, on the other (Fig. 1b). Where the bony IOC was not in contact with the alveolar bone, it was oval-shaped and held in position by a free bone bridge. If the cheek tooth alveolus and the IOC were in close contact, the bony canal was irregularly shaped. Age-related positional changes of the IOC could be measured (Fig. 5). Two teeth roots (both 08 s) of the youngest horse in group A showed direct contact with the IOC, but the distance between dental alveolus and IOC increased with dental age. The increase in distance in 88% could be explained by dental aging (R² = 0.88). There was no significant difference between both sides of the maxilla.

Imaging modalities

As detailed evaluation of the internal and periodontal structures of the cheek teeth is important in deciding on treatment when pulpar or apical infection occurs [35], CT and MR images have been acquired in the present study, aiming to highlight the best imaging modality for each structure. The overview of the two techniques on the same head allows a direct comparison of the potential of both. There are several studies published in equine medicine describing the qualities of CT or MRI for the diagnosis of head pathologies without comparison of both techniques [1–6] and with lower MRI field strength of 0.5 to 1.5 T [17, 25, 36].

The image quality of MRI in the present study was comparable to that of CT, but was better for bony tissues in CT and for soft tissues in MRI. The present results correspond with the findings of Gerlach et al. [36] and Kaminsky et al. [10], who proved the excellent quality of MRI in portraying the dental pulp, PDL, bone-marrow, gingiva, facial soft tissue, sinus mucosa, infraorbital nerve and vessels, owing to a high water content and hydrogen atoms. There was no detectable MRI signal from hard dental tissues, cortical bone, lamina dura and the IOC, whereas these structures could be visualised with good detail resolution by CT acquisition. The best MR image quality in the present study could be achieved with T2w scans for an anatomical overview and with PDw or STIR scans for more detailed images; this is in line with the findings of Kraft and Gavin [37], where T2w, STIR and PDw scans appeared superior to T1w scans in all images evaluated. This is related to the thinner slice thickness of the 3D T1w sequences, which are more often affected by artefacts compared to the transversal and dorsal orientated sequences of the STIR, PD and T2 weighted images with thicker slices.

Compared to low-field tomography, 3.0 T MRI has a higher signal-to-noise ratio, leading to better resolution. Consequently, sequences can be taken with a lower recording time and equal image quality compared to low-field MRI. The time for general anaesthesia for patients can be shortened and anaesthesia risk is reduced [38, 39]. The time required for imaging examination in the present study differed greatly between MRI and CT and between the different MRI scans. The CT was around 18 times faster than the entire MRI examination, due to the different imaging techniques and various MRI scans. The long MRI examination times were chosen to acquire images of excellent quality in the current study. The number of image alignments (transverse, dorsally orientated), resolution, or FOV might be decreased for clinical use to reduce scan time.

Due to the differences in the alignment of teeth [40], the MR images obtained could include artefacts, because image alignment was chosen for the entire skull and not for each tooth. Finally, some pulps and PDLs appeared blurry only due to image angulation (T2w, PDw, STIR) in the current study, that might be mistaken for pathology [17]. The alignment for each tooth would take more time and anaesthetic risk in clinical patients would increase due to longer examination times [41]. T1w images with 3D datasets offer an exception: through MPR, this technique provides the possibility of producing image planes in alignment with the teeth depicted. The time for image acquisition in T1w sequences is prolonged in high-field MRI, as T1 relaxation time is longer [38]. Therefore, a short acquisition time for T1w images, as was applied in the current study, is always accompanied by worse image quality compared to the T2w scans.

In contrast to MRI, CT provides the possibility of different angulations through subsequent MPR and, thus, the potential to evaluate the dental and periodontal changes in alignment of every single tooth. Nevertheless, CT also shows limitations, as it is less useful in identifying early changes in the pulp [42] due to its inability to visualise soft tissues as excellently as T2w or PDw images.

Selected cheek teeth

As described in the literature, dental pathologies such as apical infections, dental fractures and infundibular caries occur predominantly in certain maxillary cheek teeth [43]. The upper 08 s and 09 s show clinical signs most frequently [44–46] and, therefore, they were chosen for examination in the current study.

Selected age groups

The planning of surgical procedures in cases of dental and associated pathologies requires accurate anatomical knowledge and imaging of the cheek teeth and adjacent structures, which are illustrated in this study. Selected equine groups of different ages were chosen in our study as dental changes can be influenced by morphological, functional and mechanical changes, which occur with increasing age [47]: Equine cheek teeth show second dentition and real longitudinal growth up to the horse’s age of five years (Group A in the current study); afterwards, the tooth length decreases as teeth are pulled out of the alveoli in the oral direction and the cheek teeth abrasion continues (Group B in our study: 6–15 years) [48, 49]. In older horses (Group C from an age of 15 years onwards), cheek teeth have an exposition to higher dental forces [50]. All these aspects might influence changes of dental structures (e.g. pulpar size) or positional relations between dental and periodontal structures, which were examined in the current study.

Regarding the preselected horse groups (resulting in preselected dental age groups), the present evaluations should be treated with considerable caution: due to the small size of the study population, age distribution and image interpretation might not comply with the entire equine population.

Age-related variances