Abstract
The study aims were to investigate the correlation between vertebral shape and hand–wrist maturation and to select characteristic parameters of C2–C5 (the second to fifth cervical vertebrae) for cervical vertebral maturation determination by mixed longitudinal data. 87 adolescents (32 males, 55 females) aged 8–18 years with normal occlusion were studied. Sequential lateral cephalograms and hand–wrist radiographs were taken annually for 6 consecutive years. Lateral cephalograms were divided into 11 maturation groups according to Fishman Skeletal Maturity Indicators (SMI). 62 morphological measurements of C2–C5 at 11 different developmental stages (SMI1–11) were measured and analysed. Locally weighted scatterplot smoothing, correlation coefficient analysis and variable cluster analysis were used for statistical analysis. Of the 62 cervical vertebral parameters, 44 were positively correlated with SMI, 6 were negatively correlated and 12 were not correlated. The correlation coefficients between cervical vertebral parameters and SMI were relatively high. Characteristic parameters for quantitative analysis of cervical vertebral maturation were selected. In summary, cervical vertebral maturation could be used reliably to evaluate the skeletal stage instead of the hand–wrist radiographic method. Selected characteristic parameters offered a simple and objective reference for the assessment of skeletal maturity and timing of orthognathic surgery.
Orthognathic surgery is generally performed in patients affected by dental–skeletal facial abnormalities, which can affect their mastication, aesthetic appearance, self-esteem, confidence and lifestyle . Increasing numbers of adolescents and young adults are undergoing facial skeletal surgical procedures.
Body image can affect behaviour and emotional development, especially for female patients . Their primary motives for seeking surgery are aesthetic and they hope to solve the problem as soon as possible . Orthognathic surgery should be carried out once growth and development are complete, so that the minimum relapse and maximum stability can be assured. This raises a question about when growth and development end. What is the earliest sign that establishes the end of growth and development so that orthognathic surgery can be started?
Clinically, it is often recommended that jaw surgery should be carried out after the age of 18 years, because that is the age at which growth and development are considered to have finished. It has long been recognized, however, that an individual’s chronological age does not necessarily correlate well with their maturational age .
Traditionally, maturational status has been evaluated by hand–wrist radiographs, which are considered the gold standard , even though their clinical usefulness has been questioned . There are also concerns about extra radiation exposure. The shape of the cervical vertebrae is related to skeletal maturation, permitting skeletal age assessment from cephalograms, instead of subjecting the patient to the additional radiation of a hand–wrist radiograph .
It is not known whether the cervical vertebral maturation (CVM) method provides the same accurate and simple information as an assessment based on hand–wrist radiographs. If hand–wrist and cervical maturation methods were highly correlated, there would be no justification in taking a hand–wrist radiograph for skeletal maturation determination. Comparisons of skeletal maturation between hand–wrist and cervical vertebrae indicators have been made , but no comparison based on longitudinal data was reported. Longitudinal data presented on a yearly basis are of great value to orthodontists interested in the detailed study of facial growth .
Regarding the morphometric issues of cervical vertebrae, shape measurements have either been confined to the height–width ratio of limited vertebral bodies (the third to fourth cervical vertebra) , or just used or referred to the atlas reported by L amparski . Using the limited cervical measurements might affect the correlation obtained because much useful information will be missed.
Cervical vertebrae have fewer indicators of skeletal maturity compared with the hand–wrist, but they have obvious morphological changes at different developmental stages. The use of an atlas is convenient because changes in cervical vertebral bodies can be evaluated with regard to growth in the atlas. However, an atlas cannot be used to evaluate growth in an objective and detailed manner because it uses all vertebrae as a group. The observation of different vertebral bodies is sometimes contradictory. This leads to confusion between various indicators and inaccurate or subjective results . A more comprehensive shape analysis and more specific parameters could provide additional information to show the underlying morphometric relationship and increase the predictive power.
The purpose of this study, using mixed longitudinal samples, was twofold. First, to determine whether the morphological changes seen in the cervical vertebrae were as useful as those from hand–wrist radiographs to determine the growth stage in Chinese subjects. Second, to select characteristic parameters from the second to fifth cervical vertebrae (C2–C5) to evaluate skeletal maturation objectively from lateral cephalometric radiographs.
Materials and methods
This study was a retrospective review of available data. Longitudinal population data were derived from Beijing University Research Center of Craniofacial Growth and Development. More than 900 patients born in 1977–1978 were reviewed initially. The final study population consisted of 87 adolescents (32 males, 55 females) aged 8–18 years with normal occlusion because of the problems of longitudinal radiographical recordings. They formed two groups depending on the age at which the recordings began. Observation began for group 1 (43 adolescents, 16 males, 27 females) at 8–9 years, and for group 2 (44 adolescents, 16 males, 28 females) at 12–13 years of age. Sequential lateral cephalograms (LCR) and hand–wrist films were taken once a year, for 6 consecutive years . Informed consent was obtained from all 87 adolescents and their parents. The study protocol was reviewed and approved by the Institutional Review Board.
The 87 subjects fulfilled the following criteria: deciduous, mixed, or permanent dentition; normal occlusion (<3 mm overjet, and overbite less than one-third coverage of mandibular incisor); harmonious facial profile and lip competence at rest; and no orthodontic treatment or extractions of permanent teeth.
Most previous studies were based on Lamparski’s atlas by direct image-reading . In this study, Lamparski’s atlas was not used. The samples were divided into 11 maturation groups according to the Fishman Skeletal Maturity Indicators (SMI) , which was thought to be the gold standard for skeletal maturation evaluation .
The SMI system uses only 4 stages of bone maturation, all found at 6 anatomical sites located on the thumb, third finger, fifth finger and radius. The 11 discrete adolescent SMI, covering the entire period of adolescent development, are (in chronological order): (1) width of epiphysis as wide as diaphysis in the third finger, proximal phalanx; (2) width of epiphysis as wide as diaphysis in the third finger, middle phalanx; (3) width of epiphysis as wide as diaphysis in the fifth finger, middle phalanx; (4) ossification of adductor sesamoid of thumb; (5) capping of epiphysis in the third finger, distal phalanx; (6) capping of epiphysis in the third finger, middle phalanx; (7) capping of epiphysis in the fifth finger, middle phalanx; (8) fusion of epiphysis and diaphysis in the third finger, distal phalanx; (9) fusion of epiphysis and diaphysis in the third finger, proximal phalanx; (10) fusion of epiphysis and diaphysis in the third finger, middle phalanx; (11) fusion of epiphysis and diaphysis in radius.
LCRs of all subjects, coupled with hand-wrist radiographs, totalled 522, of which 4 radiographs did not reach even SMI1 and 7 radiographs were discarded because they were fuzzy. The remaining 511 LCRs were divided into 11 maturation groups by a calibrated technician according to SMI assessed from their hand–wrist radiographs, as shown in Table 1 . 62 morphological characteristic parameters of C2–C5 at 11 different developmental stages (SMI1–11) were measured and analysed.
| SMI | N | Average age (year) | Age range (year) | ||
|---|---|---|---|---|---|
| Females ( X ± SD) | Males ( X ± SD) | Females | Males | ||
| 1 | 31 | 9.01 ± 1.47 | 9.81 ± 1.41 | 7.00–10.49 | 7.83–11.25 |
| 2 | 28 | 9.67 ± 1.02 | 10.83 ± 0.93 | 8.50–11.07 | 9.65–11.80 |
| 3 | 30 | 10.01 ± 1.50 | 11.29 ± 1.15 | 8.52–12.08 | 10.15–12.50 |
| 4 | 28 | 10.71 ± 0.45 | 11.61 ± 0.54 | 10.28–11.97 | 11.03–13.17 |
| 5 | 28 | 11.18 ± 1.19 | 12.28 ± 0.60 | 10.07–13.17 | 12.00–13.57 |
| 6 | 32 | 11.92 ± 1.46 | 12.81 ± 0.67 | 10.25–14.17 | 12.08–14.33 |
| 7 | 28 | 12.29 ± 0.91 | 13.65 ± 0.95 | 11.28–14.33 | 12.25–15.75 |
| 8 | 30 | 12.98 ± 0.61 | 14.19 ± 1.23 | 12.05–13.75 | 13.00–16.07 |
| 9 | 38 | 13.64 ± 1.20 | 15.17 ± 0.79 | 12.25–15.50 | 14.10–16.20 |
| 10 | 52 | 14.93 ± 0.90 | 16.22 ± 1.10 | 13.88–16.33 | 14.93–17.50 |
| 11 | 49 | 16.41 ± 1.39 | 17.60 ± 0.50 | 13.98–17.92 | 16.00–18.18 |
The parameters correlated with SMI were selected using a new modelling method based on a nonparametric method, called locally weighted scatterplot smoothing (LOESS), which is very flexible and can account for a much wider range of component patterns than any single parametric model . The characteristic parameters for assessing cervical vertebral maturation were selected by correlation coefficient analysis and variable cluster analysis.
All points and lines shown in Figs 1 and 2 and defined in Table 2 were traced with a pencil by one observer under optimal conditions and then measured with micrometer callipers. The ratios of these parameters, as defined in Table 2 , were calculated.
| Parameter | Description |
|---|---|
| D2 | Vertical distance of C2d to the connection of C2a and C2p |
| D3–5 | Vertical distance of C3–5d to the connection of C3–5lp and C3–5la |
| @2 | Antero-superior angle of C2d–C2p connection to C2p–C2a connection |
| @3 | Antero-superior angle of C3d–C3lp connection to C3lp–C3la connection |
| @4 | Antero-superior angle of C4d–C4lp connection to C4lp–C4la connection |
| @5 | Antero-superior angle of C5d–C5lp connection to C5lp–C5la connection |
| AI2–3 | Distance between C2a and C3ua |
| PI2–3 | Distance between C2p and C3up |
| AI3–4 | Distance between C3la and C4ua |
| PI3–4 | Distance between C3lp and C4up |
| AI4–5 | Distance between C4la and C5ua |
| PI4–5 | Distance between C4lp and C5up |
| PH3–5 | Vertical distance of C3–5up to the connection of C3–5lp and C3–5la |
| H3–5 | Vertical distance of C3–5um to the connection of C3–5lp and C3–5la |
| AH3–5 | Vertical distance of C3–5ua to the connection of C3–5lp and C3–5la |
| AD3–5 | Distance between C3–5la and C3–5ua |
| PD3–5 | Distance between C3–5lp and C3–5up |
| UW3–5 | Vertical distance of C3–5ua to the connection of C3–5up and C3–5lp |
| W3–5 | Vertical distance of C3–5am to the connection of C3–5up and C3–5lp |
| LW3–5 | Vertical distance of C3–5la to the connection of C3–5up and C3–5lp |
| AH3–5/H3–5 | Ratio of AH3–5 to H3–5 |
| H3–5/PH3–5 | Ratio of H3–5 to PH3–5 |
| AH3–5/PH3–5 | Ratio of AH3–5 to PH3–5 |
| AH3–5/W3–5 | Ratio of AH3–5 to W3–5 |
| H3–5/W3–5 | Ratio of H3–5 to W3–5 |
| PH3–5/W3–5 | Ratio of PH3–5 to W3–5 |
| UW3–5/LW3–5 | Ratio of UW3–5 to LW3–5 |
| AD3–5/PD3–5 | Ratio of AD3–5 to PD3–5 |
The data were analysed using statistical software SPSS Version 13.0 for Windows. The arithmetic mean and standard deviation were calculated for all 62 variables. The statistical analyses used included locally weighted scatterplot smoothing (LOESS), correlation coefficient analysis (CC) and variable cluster analysis.
Intraobserver reliability and reproducibility of the linear measurements were checked on 20 randomly selected cephalometric radiographs that were retraced and redigitized 2 weeks later. The method error did not exceed 0.2 mm for all linear variables.
Results
After statistical analysis of LOESS smoothing ( Fig. 3 ), among 62 parameters, 44 parameters were positively correlated with SMI, 6 were negatively correlated and 12 were not correlated ( Table 3 ). Those that were not relevant were parameters relating to cervical vertebral width (UW, W, LW, UW/LW), although the ratios of cervical vertebral width such as H4/W4, H3/W3, had a strong correlation with SMI (0.9107 and 0.9091, respectively) as shown in Table 3 . Parameters of inter-vertebral distance had a negative correlation with SMI, meaning that the vertebral space became smaller with growth.
