Computed tomography (CT) has become more readily available for post-mortem examination, offering an alternative to cetacean cranial measurements obtained manually. Measurement error may result in possible variation in cranial morphometric analysis. This study aimed to evaluate the accuracy and reliability of cetacean cranial measurements obtained by CT three-dimensional volume rendered images (3DVRI). CT scans of 9 stranded cetaceans were performed. The acquired images were reconstructed using bone reconstruction algorithms. The reconstructed crania obtained by 3DVRI were visualized after excluding other body structures. Accuracy of cranial measurements obtained by CT 3DVRI was evaluated by comparing with that obtained by manual approach as standard of reference. Reproducibility and repeatability of cranial measurements obtained by CT 3DVRI were evaluated using intraclass correlation coefficient (ICC). The results demonstrated that cranial measurements obtained by CT 3DVRI yielded high accuracy (88.05%– 99.64%). High reproducibility (ICC ranged from 0.897 to 1.000) and repeatability (ICC ranged from 0.919 to 1.000 for operator 1 and ICC range from 0.768 to 1.000 for operator 2) were observed in cranial measurements obtained by CT 3DVRI. Therefore, cranial measurements obtained by CT 3DVRI could be considered as virtual alternative to conventional manual approach. This may help the development of a normative reference for current cranial maturity and discriminant analysis studies in cetaceans.
Citation: Yuen AHL, Tsui HCL, Kot BCW (2017) Accuracy and reliability of cetacean cranial measurements using computed tomography three dimensional volume rendered images. PLoS ONE 12(3): e0174215. https://doi.org/10.1371/journal.pone.0174215
Editor: Matthew C. Mihlbachler, New York Institute of Technology, UNITED STATES
Received: September 4, 2016; Accepted: March 6, 2017; Published: March 22, 2017
Copyright: © 2017 Yuen et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: The Agriculture, Fisheries and Conservation Department of Hong Kong Special Administrative Region has placed restrictions on the data that prevent it from being shared publicly. Readers may contact Dr. NG Wai-Chuen (Marine Conservation Officer [West]) from the Agriculture, Fisheries and Conservation Department (AFCD) of Hong Kong Special Administrative Region to request the data. Interested readers can contact Dr. NG Wai-Chuen at the following email address: email@example.com.
Funding: This project was financially supported by the Hong Kong Research Grants Council [Grant number: UGC/FDS17/M07/14] and received by BK. URL: http://www.ugc.edu.hk/eng/rgc/result/sf/fds1415_abstract.htm#FDS17M07. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Osteological specimens, particularly crania, represent ideal structures for the study of evolutionary and geographic variation in vertebrate morphology. In cetaceans, the morphometric assessment of cranial measurements parameters may provide additional information to identify sex, age [1–2], evolutionary history, geographical variation [3–6] and sexual dimorphism [3,5–6]. Cetacean cranial morphometrics often rely on manual measurement , requiring extensive time and manpower for skeletal excarnation [8–10]. Bone loss may be subsequently induced as consequence of weathering and improper preparation, hindering cranial landmark identification, and thus the accuracy and reliability of corresponding morphometrics. To precisely interpret structural variation of cetacean crania, methods offering accurate and reliable measurements of anatomical specimens must be available.
Cross-sectional imaging techniques, such as computed tomography (CT), have become more readily available for post-mortem examination. A growing body of anatomical data has become available electronically for pathological and biological profile investigation, offering an alternative to traditional cranial morphometrics of cetaceans. In human medicine, CT three-dimensional volume rendering (3DVR) is a non-invasive image reformation technique offering insight of complex, multi-spatial orientation of bony structures [8–11]. It allows operators to bypass the process of skeletal excarnation . The reconstructed CT three-dimensional volume rendering images (3DVRI) can be sectioned in any plane and rotated in space, allowing three-dimensional (3D) insight into the anatomy of the cranium . In humans, Stull and his team  demonstrated low average percentage difference between dried bone, the CT 3DVRI of bone before excarnation, and CT 3DVRI of bone after excarnation. In cetaceans, measurement error—either because of uncertainty in landmark identification or alterations in the objects’ true dimensions as a consequence of imaging—could induce possible variability in cranial measurements. To properly utilize this new growing resource, studies must be undertaken to ensure that cetacean cranial measurements obtained by CT 3DVRI are accurate and reliable.
The present study aimed to evaluate the accuracy of cetacean cranial measurements obtained by CT 3DVRI, compared to those obtained by the conventional manual measurement on dried crania as a standard of reference. The present study also examined the inter- and intra-operator variability of cranial measurements obtained by CT 3DVRI.
Materials and methods
A total of 9 cetaceans (8 Neophocaena phocaenoides and 1 Sousa chinensis) stranded and deceased in the Hong Kong waters in 2012–2015 were included in the present study. Neonates, calves and carcasses with collapsed crania were excluded. The condition of the carcasses was classified using Simthsonian condition codes  and ranged from code 2 to 4. Thirty one cranial measurements were collected from each specimen, with reference to the measurement definitions from Yao and her team  (Figs 1 and 2, Table 1).
For abbreviations, refer to Table 1.
For abbreviations, refer to Table 1.
All CT scans (2 whole body scans, 1 upper body scan and 6 cranial scans) were performed with a Toshiba 16-row multislice spiral CT scanner Alexion™ prior to excarnation. The scans were operated at 100-120kV, 50-170mA, and 1mm slice thickness. Scan field of view (sFOV) was ranged from 15.3cm to 43.2cm. The acquired data sets were reconstructed in 3DVRI using the bone reconstruction algorithm of the inbuilt software (Alexion V4.86ER003). The reconstructed crania obtained by 3DVRI were visualised after excluding other body structures. Two operators (AY and BK) took the cranial measurements obtained by CT 3DVRI in different orientations using the rotation function in the computer graphics system . Cranial measurements obtained by CT 3DVRI were collected by 2 operators (AY and BK) twice. To avoid recall bias, there was a time interval of at least 30 minutes between measurements on the same cranium. Each operator was blinded to the results of the other operator. Operator 1 (AY) had about 2 years of experience in CT scanning and cetacean cranial morphometrics, whereas operator 2 (BK) had more than 4 years of experience.
Following CT scanning, excarnation was performed, removing soft tissue from the crania. They were then air dried at room temperature before proceeding with manual measurement. All manual measurements were taken by AY, thereby eliminating the potential problem of inter-operator bias. Manual measurements were made by using 15cm, 30cm, and 60cm calipers. All measurements were taken to the nearest 0.5mm.
To evaluate the accuracy of the cranial measurements obtained by CT 3DVRI, the measurements taken by AY on 3DVRI were compared to those obtained by the conventional manual approach. The absolute percentage error was calculated. The absolute percentage error of each cranial measurement obtained by CT 3DVRI was defined as the percentage difference between cranial measurement obtained by CT 3DVRI and cranial measurement of dried cranium:
To analyze the intra-operator variability (repeatability) and inter-operator (reproducibility) variability of cranial measurements obtained by CT 3DVRI, intraclass correlation coefficient (ICC) and 95% confidence intervals (C.I.) were used. An ICC > 0.7 was used to indicate sufficient general reliability of the measurements [15–16]. All statistical analyses were performed with the use of SPSS 23.0 software (Statistical Package for the Social Sciences Inc., Chicago, Illinois 60606, USA).
All procedures in this study were reviewed and approved by the Agriculture, Fisheries and Conservation Department of Hong Kong Special Administrative Region [AF GR CON 09/68 PT.9].
In comparing cetacean cranial measurements obtained by CT 3DVRI with those obtained by manual measurement (Table 2), the mean percentage errors ranged from 0.36% to 11.95%, indicating that measurements obtained by 3DVRI yielded high accuracy (ranging from 88.05% to 99.64%).
The repeatability of 31 cetacean cranial measurements obtained by CT 3DVRI was shown in Table 3. The ICC values for operator 1 varied from 0.919 to 1.000 with 95% C.I. ranging from 0.685 to 1.000, indicating repeatability of 91.9% to 100%. The ICC values for operator 2 varied from 0.768 to 1.000 with 95% C.I. ranging from 0.265 to 1.000, indicating repeatability of 76.8% to 100%. Overall, results demonstrated that cetacean cranial measurements obtained by CT 3DVRI yielded a high repeatability.
The reproducibility of 31 cetacean cranial measurements obtained by CT 3DVRI was shown in Table 4. The ICC values were ranging from 0.897 to 1.000 with 95% C.I. ranging from 0.608 to 1.000, indicating reproducibility of 89.7% to 100%. Results demonstrated that cetacean cranial measurements obtained by CT 3DVRI yielded a high reproducibility.
CT scanning was widely used for evaluation of osteological structures because of its documented high spatial resolution, contrast resolution, and signal-to-noise ratio. The image quality, measurement accuracy and reliability of 3DVRI were largely depended on the slice thickness, sFOV and reconstruction algorithm . The data acquired in the present study was with slice thickness of 1 mm. Slice thicknesses greater than 1.25 mm may require more interpolation between slices when rendering CT 3DVRI, which results in a loss of fidelity of the representative anatomy . Such data loss can lead to inaccuracies in cetacean cranial measurements. Bone reconstruction algorithm with minimized sFOV was also used in the present study to produce adequate 3DVRI resolution for accurate cranial measurements.
The results of the present study show that cetacean cranial measurements obtained by CT 3DVRI are highly accurate (88.0–99.6%) when compared with cranial measurements obtained by manual measurement. Of the 31 cranial measurements included in the present study, relatively lower accuracy (88%) was observed in length of antorbital process of left lacrimal (LAO). The cause of the slight deviation was related to 3DVR software interpolates data between scan slices.
The present study also demonstrated high repeatability and reproducibility in cranial measurements obtained by CT 3DVRI. Among the 31 cranial measurements, the lowest repeatability was observed in width of premaxillaries at midlength of rostrum (WPM) and distance from anterior most end of junction between nasals to hindmost point of margin of supraoccipital crest (DNS), while the lowest reproducibility was observed in length of left orbit-from apex of preorbital process of frontal to apex of post orbital process (LPO), distance from anterior most end of left jugal to posterior margin of last tooth on upper left rostrum (JG), greatest length of left pterygoid (GLP), WPM and DNS. Measurements of WPM, DNS, LPO and JG might be subject to the effect of interpolation between slices when rendering CT 3DVRI, which resulted in loss of cranial landmarks or sutures definition, affecting precision of caliper placement. Variation in GLP measurement may be accounted by the relatively low bone density in cetacean pterygoid hamulus, which leads to the failure in reconstructing the intact pterygoid as a result of inability to display actual Hounsfield units during conversion from CT number to electron density. Nonetheless, most cranial characters presented in 3DVRI were recognizable and thus, cranial measurements were considered to be accurate and reliable.
CT 3DVRI data set was recorded in Digital Imaging and Communication in Medicine (DICOM) format in the present study. DICOM data sets can be permanently stored in Picture Archiving and Communication System, which can be recalled at will to the same specimen if additional information is required. Its digitally transferable feature can also facilitate second opinion by other professionals or institutes worldwide even if the specimens cannot not be physically provided or available.
CT scanning of cetaceans’ cranium can be a valuable resource to identify and document osteological structures, as well as soft tissues that were not preserved in osteological collections. Spatial relationship between specific organs, soft tissues and bones can be selectively investigated in situ in their true dimensions and positions [19–20]. CT 3DVRI could be a valuable approach for obtaining morphological information from cetaceans. Since cross-sectional imaging technologies are now more readily available in studies of animals’ taxonomy and life history, there are increasing collaboration between museum and imaging centres to digitalize specimens [21–22]. Digitalized specimens would allow for investigations of anatomical characters and might increase the accessibility of phenotypical data [21–22].
To conclude, the results of the present study demonstrate that cetacean cranial measurements obtained by CT 3DVRI were accurate and reliable. CT 3DVRI could be considered a promising alternative to traditional manual cranial measurements in providing information for biological profiles investigation in cetaceans.
The authors would like to thank the Agriculture, Fisheries and Conservation Department of the Hong Kong Special Administrative Region Government for the continuous support in this project. Sincere appreciation is also extended to veterinarians, staff and volunteers from Ocean Park Hong Kong, Ocean Park Conservation Foundation Hong Kong, and Tung Wah College for their great efforts with the stranding response and skeletal excarnation in this project. Special gratitude is owed to technicians from Hong Kong Veterinary Imaging Center for operating the CT unit in the present study. Wholeheartedly thanks to Dr. Chiou-Ju Yao (Department of Biology, National Museum of Natural Science, Taichung, Taiwan) for her advice and guidance on cetacean cranial measurements in this project. Special thanks to Dr. Michael Bradley for English editing of this manuscript.
- Conceptualization: AY HT BK.
- Data curation: HT.
- Formal analysis: AY HT BK.
- Funding acquisition: HT BK.
- Investigation: AY HT BK.
- Methodology: AY BK.
- Project administration: BK.
- Resources: HT.
- Software: HT.
- Supervision: BK.
- Validation: HT.
- Visualization: HT.
- Writing – original draft: AY BK.
- Writing – review & editing: BK.
- 1. Jordan FF, Murphy S, Martinez E, Amiot C, Helden A, Stockin KA. Criteria for assessing maturity of skulls in the common dolphin, Delphinus sp., from New Zealand waters. Mar Mamm Sci. 2015;31: 1077–97.
- 2. Rommel S. Osteology. In: Leatherwood S, Reeves RR, editors. The Bottlenose Dolphin. San Diego: Academic Press; 1990. pp. 309–329.
- 3. Thompson KF, Ruggiero K, Millar CD, Constantine R, Helden AL. Large-scale multivariate analysis reveals sexual dimorphism and geographic differences in the Gray's beaked whale. J Zool. 2014;294:13–21.
- 4. Wang JY, Chou LS, White BN. Osteological differences between two sympatric forms of bottlenose dolphins (genus Tursiops) in Chinese waters. J Zool. 2000;252:147–62.
- 5. Perrin WF. Variation of spotted and spinner porpoise (genus Stenella) in the eastern Pacific and Hawaii. Bull Scripps Inst Oceanogr. 1975;21:1–206
- 6. Murphy S, Herman JS, Pierce GJ, Rogan E, Kitchener AC. Taxonomic status and geographical cranial variation of common dolphins (Delphinus) in the eastern north Atlantic. Mar Mamm Sci. 2006;22:573–99.
- 7. Yao CJ, Yamada TK, Chen YJ, Chou LS. Cranial variation in the pantropical spotted dolphin, Stenella attenuata, in the Pacific Ocean. Zoolog Sci. 2008;25:1234–46. pmid:19267651
- 8. Stull KE, Tise ML, Ali Z, Fowler DR. Accuracy and reliability of measurements obtained from computed tomography 3D volume rendered images. Forensic Sci Int. 2014;238:133–40. pmid:24713466
- 9. Brough AL, Bennett J, Morgan B, Black S, Rutty GN. Anthropological Measurement of the Juvenile Clavicle Using Multi-Detector Computed Tomography—Affirming Reliability. J Forensic Sci. 2013;58:946–51. pmid:23682704
- 10. Grabherr S, Cooper C, Ulrich-Bochsler S, Uldin T, Ross S, Oesterhelweg L, et al. Estimation of sex and age of “virtual skeletons”–a feasibility study. Eur Radiol. 2009;19:419–29. pmid:18766348
- 11. Cavalcanti MGP, Rocha SS, Vannier MW. Craniofacial measurements based on 3D-CT volume rendering: implications for clinical applications. Dentomaxillofac Radiol. 2014;33:170–6.
- 12. Fatterpekar GM, Doshi AH, Dugar M, Delman BN, Naidich TP, Som PM. Role of 3D CT in the Evaluation of the Temporal Bone. Radiographics. 2006;26:S117–32. pmid:17050510
- 13. Geraci JR, Lounsbury VJ. Marine mammals ashore: a field guide for strandings. 2nd ed. Texas: Texas A&M University Sea Grant College Program; 1993.
- 14. Lopes PM, Moreira CR, Perrella A, Antunes JL, Cavalcanti MG. 3-D volume rendering maxillofacial analysis of angular measurements by multislice CT. Oral Surg Oral Med Oral Pathol Oral Radiol Endod. 2008;105:224–30. pmid:18230391
- 15. Chien PF, Khan KS. Evaluation of a clinical test. II: Assessment of validity. BJOG. 2001;108:568–72. pmid:11426889
- 16. Kha KS, Chien PF. Evaluation of a clinical test. I: assessment of reliability. BJOG. 2001;108:562–7. pmid:11426888
- 17. Whyms BJ, Vorperian HK, Gentry LR, Schimek EM, Bersu ET, Chung MK. The effect of CT scanner parameters and 3D volume rendering techniques on the accuracy of linear, angular, and volumetric measurements of the mandible. Oral Surg Oral Med Oral Pathol Oral Radiol Endod. 2013; 115: 682.
- 18. Ford JM, Decker SJ. Computed tomography slice thickness and its effects on three-dimensional reconstruction of anatomical structures. Journal of Forensic Radiology and Imaging. 2016;4: 43–46.
- 19. Lautenschlager S, Bright JA, Rayfield EJ. Digital dissection–using contrast-enhanced computed tomography scanning to elucidate hard- and soft-tissue anatomy in the Common Buzzard Buteo buteo. J ANAT. 2014; 224: 412–431. pmid:24350638
- 20. Montie EW, Manire CA, Mann DA. Live CT imaging of sound reception anatomy and hearing measurements in the pygmy killer whale, Feresa attenuata. J Exp Biol. 2011; 214: 945–955. pmid:21346122
- 21. Lauridsen H, Hansen K, Wang T, Agger P, Andersen L. Inside Out: Modern imaging techniques to reveal animal anatomy. PLoS ONE. 2011; 6: e17879 pmid:21445356
- 22. Van De Kamp T, dos Santos Rolo T, Vagovič P, Baumbach T, Riedel A. Three-dimensional reconstructions come to life–interactive 3D PDF animations in functional morphology. PLoS ONE. 2014; 9: e102355. pmid:25029366