Alpacas commonly develop dental abnormalities, such as tooth root abscesses, mandibular osteomyelitis, malocclusion, fractured teeth, uneven teeth, overgrown teeth, worn teeth, and persistent deciduous teeth.1 Traditionally, radiography has been used to identify which teeth are involved and determine the extent of osteomyelitis, sequestra, and draining tracts in alpacas with dental disorders.1–3
For alpacas, acquisition of optimal radiographs of the skull requires sedation and a minimum of 4 views.1,3 Standard radiographic views of the skull include lateral, dorsoventral, left-to-right lateroventral-laterodorsal oblique, and left-to-right laterodorsal-lateroventral oblique views, and an additional intraoral view of the mandible may be necessary for some patients.1,3 Small dental x-ray sensors are routinely used for intraoral imaging in small animals, but to our knowledge, use of those systems for intraoral imaging in alpacas has not been described. The acquisition of diagnostic radiographs of the skull of alpacas can be laborious because rotation of the animal from sternal to lateral recumbency and insertion of a mouth gag for oblique and intraoral images is generally required.1,3 In a study3 involving cadaveric skulls of alpacas and llamas, CT was superior to radiography because of its increased contrast resolution, lack of superimposition of soft tissue and osseous structures, and the ability to reconstruct multiplanar and 3-D volume rendering images. From a clinical perspective, imaging the skulls of alpacas with CT rather than radiography allows the animals to be positioned in a natural body position (sternal vs dorsal recumbency), reduces the need for repositioning, and potentially decreases procedure time.
Both single and multislice CT scanners are now readily available to veterinarians.4 A CT scan can be acquired in either sequential (axial) or helical (spiral) mode. In dogs, the optimal 4-slice CT protocol for evaluation of dentition is sequential with a 1-mm slice thickness.4 Helical image quality is reported to be equivalent to sequential image quality when CT scanners with > 16 slices are used.4–6 Determination of the optimal CT slice thickness for evaluation of dentition in alpacas is important owing to the frequency of dental disorders in this species. The optimized CT protocol should prioritize visualization of the tooth roots and surrounding alveolar bone because those are the structures that are most frequently affected by dental disease.
Literature regarding the positioning of alpacas for CT scans of the skull is scarce. The literature that is available describes imaging results for detached cadaveric heads3 or is in the form of clinical reports7,8 of individual animals, in which the CT examination was performed under inhalant anesthesia. In another study,9 the skulls of llamas that were anesthetized with inhalant anesthesia and positioned in dorsal recumbency were scanned with a single-slice CT scanner at a slice thickness of 5 mm. The llamas of that study9 were supported by the normal CT couch and easily positioned within the gantry.
Historically, alpacas are anesthetized with inhalant agents for major surgical procedures or when restraint or immobilization is required for a prolonged period. Compared with other species, the trachea of camelids is small relative to the overall body size, which in conjunction with a deep oral cavity, makes oral intubation difficult.10 Camelids, including alpacas, tend to be calm and quiet when recovering from injectable anesthesia and generally do not attempt to stand until they are awake and fully functional.10 Ketamine is the most common injectable anesthetic used in camelids and is often administered in combination with an α2-adrenergic receptor agonist (eg, xylazine) and opioid (eg, butorphanol). Alpacas can be effectively restrained with injectable anesthetics for short periods (< 30 minutes) with variable levels of systemic analgesia.10
Owing to the rapid speed with which images can be acquired with 64-slice CT scanners and the minimal discomfort for alpacas when restrained in natural sternal recumbency, we believe that injectable anesthetic protocols should be sufficient to provide the depth of anesthesia and analgesia necessary for CT examination of the skull. The development of a positioning protocol for the acquisition of diagnostic CT scans of the skull of alpacas under injectable anesthesia should decrease both handling and anesthesia times and may be more cost effective for clients because injectable anesthesia is generally less expensive than inhalant anesthesia.
The purpose of the study reported here was to determine an optimal protocol for the acquisition of CT images of the dentition of alpacas with the animals anesthetized with injectable anesthetics and positioned in sternal recumbency. Six CT protocols were compared. We hypothesized that optimal CT images would be obtained by use of a 64-slice CT scanner in helical mode with a slice thickness of 1.25 mm.
Materials and Methods
Animals
All study procedures were reviewed and approved by the Purdue University Animal Care and Use Committee. Three university-owned adult male alpacas with body weights ranging from 50 to 65 kg were used for the study. Each alpaca was presumed healthy on the basis of results of a physical examination.
Anesthesia and positioning for CT examination
All CT scans were performed with the alpacas anesthetized with injectable anesthetics. Food but not water was withheld from each alpaca for 12 hours prior to the scanning procedure. Each alpaca was anesthetized with a combination of ketamine (4 to 6 mg/kg), xylazine (0.4 to 0.6 mg/kg), and butorphanol (0.1 mg/kg); the 3 drugs were combined in 1 syringe and injected IM in a semimembranosus or semitendinosus muscle. The CT examination was initiated 5 minutes after drug administration. Each alpaca was positioned in sternal recumbency on the CT couch with its head facing the gantry and legs folded in a natural cush position. A small mouth block was placed between the incisors and dental pad to separate the dental arcades. The neck and head were extended into the gantry so that the head was positioned within the isocenter of the gantry. The mandible was positioned parallel to the couch (Figure 1). Anesthetic depth was monitored by assessment of respiratory rate, heart rate, and eyelid movement.
Representative photographs of a healthy adult male alpaca positioned in sternal recumbency on a CT couch with its legs folded in a natural cush position (A) and its head positioned within the isocenter of the gantry of a 64-slice CT scanner (B) such that its mandible was parallel to the couch during a study to determine the optimized CT protocol for assessment of dentition. Each of 3 alpacas was anesthetized with a combination of ketamine (4 to 6 mg/kg), xylazine (0.4 to 0.6 mg/kg), and butorphanol (0.1 mg/kg) that was injected IM into a semimembranous or semitendinosus muscle. In panel B, notice that a small mouth block was used to separate the dental arcades during image acquisition.
Citation: American Journal of Veterinary Research 79, 3; 10.2460/ajvr.79.3.311
Experimental protocol
All CT scans were obtained in the transverse plane with a 64-slice CT scannera by use of a standard kVp (120 kVp) and variable mAs, a tube rotation speed of 1.0 seconds, a 512 × 512 matrix, and variable slice thickness (1.25, 2.5, and 5 mm) in sequential and helical modes (standard pitch) and 50% reconstruction intervals/0.625, 1.25, and 2.5-mm interslice gaps. All CT scans were acquired in a medium-frequency (standard) algorithm and reconstructed in a high-frequency (bone) algorithm. Each alpaca underwent 6 scan protocols, 3 sequential scans with slice thicknesses of 1.25 (S1.25), 2.5 (S2.5), and 5 (S5) mm and 3 helical scans with slice thicknesses of 1.25 (H1.25), 2.5 (H2.5), and 5 (H5) mm.
Image quality was evaluated independently by 3 board-certified veterinary radiologists (BGC, CKL, and HGH) who were unaware of (blinded to) the CT protocol used and subject identification. All CT scans were evaluated by means of dedicated imaging softwareb with fixed bone window width (3,600 Hounsfield units) and window level (600 Hounsfield units) settings. For each of the 6 scan protocols, the principle investigator (CVF) selected 5 representative images on the basis of dentition location (ie, tooth roots of molar, premolar and incisor teeth) such that 2 molar, 2 premolar, and the mandibular incisor teeth were evaluated. The 5 images selected from each protocol allowed for evaluation of the same tooth roots whenever possible with slight differences in the acquired slices produced by differences in slice thickness. Thus, a total of 30 images were assessed for each alpaca. For each location, the selected images from the 6 protocols were displayed on 1 screen in a random order such that there were 2 rows of 3 images. For each image, tooth root visibility, tooth root sharpness, and extent of image noise artifact were subjectively evaluated on a scale of 1 to 3 (Appendix). Tooth root visibility was considered excellent if the entire circumference of the root and its alveolar bone margin (lamina dura) could be identified. Tooth root sharpness was considered very sharp if the dental and alveolar margins were clearly demarcated. Image noise artifact was evaluated on the basis of the perceptible degree of graininess.
Data analysis
For each of the 3 CT criteria (tooth root visibility, tooth root sharpness, and extent of image noise artifact) evaluated, the extent of agreement among the 3 radiologists was assessed by calculation of the Cohen κ coefficient. Mixed-effects logistic models for ordered responses were used to evaluate the effects of acquisition mode (sequential or helical), slice thickness (1.25, 2.5, or 5 mm), and the interaction between acquisition mode and slice thickness on each of the 3 CT criteria. Each model included a random effect for alpaca to account for repeated measures within subjects. Finally, for each CT protocol (acquisition mode–slice thickness combination), the means for the 3 CT criteria were summed to create a total score, and the protocol with the greatest total score was considered the optimized CT protocol. Values of P < 0.05 were considered significant for all analyses.
Results
Alpacas
The injectable anesthetic protocol used provided adequate restraint for the duration of the CT examination for all 3 alpacas. Each alpaca became sedate enough to be positioned on the CT couch within 5 minutes after IM administration of the injectable anesthetic. The alpacas were easily managed on the couch with minimal support, and the head was easily positioned within the isocenter of the gantry. The anesthetic protocol also provided good muscle relaxation, and no motion artifact was detected on any of the evaluated images. The mean duration was 15 seconds for the helical mode scans and 1 minute for the sequential mode scans, and the mean duration from initiation of the first protocol to completion of the sixth protocol was 9 minutes. All animals recovered from anesthesia without complications.
CT image evaluation
All 6 CT protocols resulted in diagnostic images for evaluation (Figure 2). Inter-rater agreement was greatest for tooth root sharpness (κ = 0.539), followed by tooth root visibility (κ = 0.332) then image noise artifact (κ = 0.010). Slice thickness significantly affected both tooth root visibility (P < 0.001) and tooth root sharpness (P < 0.001) but did not affect image noise artifact (P = 0.332). Acquisition mode also significantly affected tooth root visibility (P = 0.033) and tooth root sharpness (P = 0.001) as well as image noise artifact (P < 0.001). The mean scores for tooth root visibility, tooth root sharpness, and image noise artifact for each protocol were summarized (Table 1). When the means for only tooth root visibility and tooth root sharpness were summed together for each protocol, the S1.25 protocol had the greatest score followed by the H1.25 protocol, but the summed scores for those 2 protocols did not differ significantly (Figure 3). When the means for all 3 CT criteria were summed together, the total scores for all sequential-mode protocols (S1.25, S2.5, and S5) were greater than those for all helical-mode protocols (H1.25, H2.5, and H5). However, the image noise score did not differ significantly by slice thickness within either acquisition mode (Figure 4).
Magnified CT images of the third molar tooth root of the right maxillary arcade of a healthy adult male alpaca. Each image was obtained in the transverse plane by a 64-slice CT scanner in the sequential mode at a slice thickness of 1.25 (S1.25 protocol; A), 2.5 (S2.5 protocol; B), and 5 (S5 protocol; C) mm and in the helical mode at a slice thickness of 1.25 (H1.25 protocol; D), 2.5 (H2.5 protocol; E), and 5 (H5 protocol; F) mm. Notice that the S1.25 (A) and H1.25 (D) protocols provided the best visualization of the lamina dura and clear demarcation of the dental and alveolar bone margins and that tooth root visualization and sharpness decreased as slice thickness increased.
Citation: American Journal of Veterinary Research 79, 3; 10.2460/ajvr.79.3.311
Magnified CT images obtained in the transverse plane of the third molar tooth root of the right maxillary arcade of an adult alpaca acquired by use of the S1.25 (A) and H1.25 (B) protocols described in Figure 2. For both images, the contrast has been decreased to enhance visualization of the tooth root. Notice that visualization of the lamina dura and demarcation of the dental and alveolar bone margins are clearer and sharper in the image obtained with the S1.25 protocol than in the image obtained with the H1.25 protocol. See Figure 2 for remainder of key.
Citation: American Journal of Veterinary Research 79, 3; 10.2460/ajvr.79.3.311
Magnified CT images obtained in the transverse plane of the right zygomatic arch of a healthy adult alpaca acquired by use of the S1.25 (A) and H1.25 (B) protocols described in Figure 2. The image of the zygomatic arch obtained with the H1.25 protocol contains alternating white and black lines (arrows) that represent windmill artifact. That artifact was observed in all CT images acquired with the helical mode. See Figure 2 for remainder of key.
Citation: American Journal of Veterinary Research 79, 3; 10.2460/ajvr.79.3.311
Mean ± SD scores for tooth root visibility, tooth root sharpness, and image noise artifact as assigned independently by 3 board-certified veterinary radiologists to CT images of the dentition for 3 healthy adult alpacas acquired by a 64-slice CT scanner in the sequential mode at a slice thickness of 1.25 (S1.25 protocol), 2.5 (S2.5 protocol), and 5 (S5 protocol) mm and in the helical mode at a slice thickness of 1.25 (H1.25 protocol), 2.5 (H2.5 protocol), and 5 (H5 protocol) mm.
| Protocol | Tooth root visibility | Tooth root sharpness | Image noise artifact |
|---|---|---|---|
| S1.25 | 3.0 ± 0.0 | 3.0 ± 0.0 | 2.5 ± 0.6 |
| S2.5 | 2.4 ± 0.5 | 2.0 ± 0.5 | 2.6 ± 0.5 |
| S5 | 1.6 ± 0.6 | 1.2 ± 0.4 | 2.3 ± 0.8 |
| H1.25 | 2.9 ± 0.3 | 2.9 ± 0.3 | 2.1 ± 0.9 |
| H2.5 | 2.0 ± 0.7 | 1.7 ± 0.6 | 2.0 ± 0.8 |
| H5 | 1.6 ± 0.6 | 1.2 ± 0.4 | 1.9 ± 0.7 |
Each feature was subjectively scored on a scale of 1 to 3 by each of the 3 radiologists. The scoring system was designed such that the numeric score was positively correlated with image quality (ie, the greater the mean score, the better the image quality).
Discussion
In the present study, diagnostic CT images of the dentition for 3 healthy male alpacas were acquired by use of an injectable anesthesia protocol with the animals positioned in sternal recumbency on the CT couch in a natural cush position. All 3 alpacas tolerated the injectable anesthesia protocol (IM injection of a combination of ketamine, xylazine, and butorphanol) well, and the protocol provided adequate anesthesia for completion of the 6 CT scan protocols evaluated. Thus, inhalant anesthesia and restraint in dorsal recumbency are not necessary to obtain diagnostic images of the dentition or skull of alpacas.
Results of the present study indicated that the optimized CT protocol for the acquisition of diagnostic images of the dentition for alpacas by means of a 64-slice CT scanner was a sequential mode with a slice thickness 1.25 mm (S1.25 protocol). Compared with the other 5 protocols evaluated (sequential mode with a slice thickness of 2.5 [S2.5 protocol] and 5 [S5 protocol] mm and helical mode with a slice thickness of 1.25 [H1.25 protocol], 2.5 [H2.5 protocol], and 5 [H5 protocol] mm), the S1.25 protocol provided images of superior quality for evaluation of tooth roots. Images acquired in the sequential mode contained less image noise artifact than images acquired in the helical mode. Also, images acquired with a slice thickness of 1.25 mm provided better visualization of the tooth roots and demarcation of the dental and alveolar bone margins, compared with images acquired with a slice thickness of 2.5 or 5 mm.
Image noise artifact is defined as image graininess caused by variation in the number of x-ray photons emerging from the patient that are recorded by the photon detectors.11 As the slice thickness decreases, there is an improvement in z-axis resolution but an increase in image noise artifact because fewer photons reach the x-ray detector.12 Helical scanning uses a different start and end point for each slice and requires an additional data processing step to create an image, which is known as interpolation and can result in reconstruction artifacts.13 Sequential scanning does not require interpolation. Results of the present study indicated that image noise artifact was not significantly affected by slice thickness but was significantly affected by image acquisition mode (sequential or helical).
In human medicine, the quality of skull images acquired in helical and sequential modes by CT scanners with ≥ 64 slices has been evaluated.5,6 In 1 study,5 neuroradiologists subjectively preferred the image quality provided by sequential scans for evaluation of certain regions of the skull; however, the investigators of the study concluded that the image quality of sequential scans was likely equivalent to that of helical scans from a clinical standpoint. In another study6 in which a 128-slice CT scanner was used to acquire images of the skulls of human patients by means of a helical scan with a slice thickness of 5 mm and a sequential scan with a slice thickness of 4.8 mm, the image quality was virtually equivalent between the helical and sequential scans. The fairly large slice thicknesses (4.8 to 5 mm) used in that study6 may have benefited image quality owing to a decrease in image noise artifact but adversely affected spatial resolution.5 Spatial resolution can be improved by the acquisition of sequential scans with a thinner (< 4.8 mm) slice thickness.5 Multiplanar reconstruction for helical scans is superior to that for sequential scans.5 In dogs, a thin slice thickness is recommended for the acquisition of CT images of dental structures so that small structures can be evaluated and to ensure minimal volume averaging.4,14 Also, a high-frequency reconstruction algorithm is typically used to assess CT images of dental structures in dogs because it allows for sharp margin delineation and an increase in spatial resolution.4
All helical scans acquired in the present study contained an aliasing artifact called windmill artifact,15,16 which is caused by inadequate sampling along the longitudinal plane.14 Windmill artifact consists of black-and-white striations that spin off of high-contrast features with curvature along the longitudinal, or z-axis.15 The amplitude of windmill artifact decreases as the number of detector rows increases.16 Although windmill artifact was not specifically evaluated in the present study, its presence may have affected the overall subjective assessment of image noise by the radiologists who evaluated the CT images.
The present study had several limitations. Only 5 static areas rather than the entire skull were evaluated for each alpaca. In a clinical setting, the entire CT examination would be evaluated dynamically. All CT scans were acquired in a medium-frequency algorithm and reconstructed in a high-frequency algorithm to mimic a clinical setting. Thick-section reformatting of thin slices is recommended to reduce windmill artifact; however, it was not performed in the present study because we wanted to directly compare the raw (unformatted) helical images with the sequential images. Also, images acquired by the use of only 1 CT machine and reconstructed by the use of only 1 software package were evaluated; therefore, the results may not be generalizable to images acquired by the use of different CT scanners or software programs.
Results of the present study indicated that CT examination of the skull of alpacas can be performed with injectable anesthesia and the animal positioned in sternal recumbency. The optimal protocol for the acquisition of images of dentition by means of a 64-slice CT scanner was a sequential scan with a slice thickness of 1.25 mm and use of a high-frequency reconstruction algorithm. If multiplanar reconstruction images are desired, we recommend an additional helical scan.
Acknowledgments
Supported by the Purdue University Veterinary Clinical Science Graduate Student Competitive Research Fund. The funding source did not have any involvement in the study design, data analysis and interpretation, or writing and publication of the manuscript.
The authors declare that there were no conflicts of interest.
Footnotes
Lightspeed 64-slice CT Scanner, GE Healthcare Inc, Princeton, NJ.
OsiriX imaging software, version 5.4, OsiriX Foundation, Geneva, Switzerland.
References
1. Niehaus AJ. Dental disease in llamas and alpacas. Vet Clin North Am Food Anim Pract 2009;25:281–293.
2. Niehaus AJ, Anderson DE. Tooth root abscesses in llamas and alpacas: 123 cases (1994–2005). J Am Vet Med Assoc 2007;231:284–289.
3. Rostami A, Geissbühler U, Schellenberger F, et al. Computed tomographic and radiographic examination of dental structures in South American camelid specimen of different ages. BMC Vet Res 2014;10:4.
4. Esmans MC, Soukup JW, Schwarz T. Optimized canine dental computed tomographic protocol in medium-sized mesaticephalic dogs. Vet Radiol Ultrasound 2014;55:506–510.
5. Abdeen N, Chakraborty S, Nguyen T, et al. Comparison of image quality and lens dose in helical and sequentially acquired head CT. Clin Radiol 2010;65:868–873.
6. Davis AJ, Ozsvath J, Vega E, et al. Continuous versus sequential acquisition head computed tomography: a phantom and clinical image quality comparative study. J Comput Assist Tomogr 2015;39:876–881.
7. Hardefeldt LA, Rylander H, Iskandar BJ, et al. Diagnosis and surgical treatment of an intracranial cyst in an alpaca cria. J Am Vet Med Assoc 2012;240:1501–1506.
8. Britt LG, Middleton JR, Warhover TT, et al. Acanthomatous ameloblastoma of the maxilla of an adult aplaca. Vet Radiol Ultrasound 2005;46:65–68.
9. Hathcock JT, Pugh DG, Cartee RE, et al. Computed tomography of the llama head: technique and normal anatomy. Vet Radiol Ultrasound 1995;36:290–296.
10. Abrahamsen EJ. Chemical restraint, anesthesia, and analgesia for camelids. Vet Clin North Am Food Anim Pract 2009;25:455–494.
11. Jones JC, Wilson ME, Bartels JE. A review of high resolution computed tomography and a proposed technique for regional examination of the canine lumbosacral spine. Vet Radiol Ultrasound 1994;35:339–346.
12. Zarelli M, Schwarz T, Puggioni A, et al. An optimized protocol for multislice computed tomography of the canine brain. Vet Radiol Ultrasound 2014;55:387–392.
13. Dowsett D, Kenny PA, Johnston RE. Computed tomography. In: The physics of diagnostic imaging. 2nd ed. Boca Raton, Fla: CRC Press, 2006;381–435.
14. Porat-Mosenco Y, Schwarz T, Kass PH. Thick-section reformatting of thinly collimated computed tomography for reduction of skull-base-related artifacts in dogs and horses. Vet Radiol Ultrasound 2004;45:131–135.
15. Silver MD, Taguchi K, Hein IA, et al. Windmill artifact in multislice helical CT. In: Medical imaging. Bellingham, Wash: International Society for Optics and Photonics, 2003:1918–1927.
16. Oliveira CR, Ranallo FN, Pijanowski GJ, et al. The Vetmousetrap: a device for computed tomographic imaging of the thorax of awake cats. Vet Radiol Ultrasound 2011;52:41–52.
Appendix
Subjective scoring system used by 3 board-certified veterinary radiologists to independently evaluate tooth root visibility, tooth root sharpness, and image noise artifact on CT images of the dentition for 3 healthy adult male alpacas obtained by each of 6 protocols during a study to determine the optimized CT protocol for assessment of dentition in alpacas.
| Score | Tooth root visibility | Tooth root sharpness | Image noise artifact |
|---|---|---|---|
| 1 | Poor | Blurry | High |
| 2 | Good | Sharp | Medium |
| 3 | Excellent | Very sharp | Low |
All images were acquired with a 64-slice CT scanner in the sequential mode at a slice thickness of 1.25 (S1.25 protocol), 2.5 (S2.5 protocol), and 5 (S5 protocol) mm and in the helical mode at a slice thickness of 1.25 (H1.25 protocol), 2.5 (H2.5 protocol), and 5 (H5 protocol) mm.
