European Journal of Orthodontics 31 (2009) 438–442 © The Author 2009.

or 2009. Published by Oxford University Press on behalf of the European Orthodontic Society.
doi:10.1093/ejo/cjp015 All rights reserved. For permissions, please email: [email protected].
Advance Access publication 23 April 2009

Reliability of growth prediction with hand–wrist radiographs

Damian Verma*, Timo Peltomäki** and Andreas Jäger*
*Clinic of Orthodontics, Rheinische Friedrich-Wilhelms-University, Bonn, Germany and **Clinic of Orthodontics
and Pediatric Dentistry, University of Zurich, Switzerland

SUMMARY The aim of this study was to investigate the validity of hand–wrist radiographic analysis in
estimating the amount of remaining craniofacial growth. The material compromised cephalograms of 22
males and 27 females with a Class I malocclusion. The median age of the females at the beginning (T1)
was 11 years 10 months and of the males 12 years 6 months and at the end (T2) of treatment 14 years 7
months and 15 years 3 months, respectively. Statural height was measured and a lateral cephalogram was
obtained for every patient at T1 and T2. A hand–wrist radiograph was taken only at T1. The cephalograms
were scanned and analyzed. Relative dimensional growth changes in statural height as well as of the
length of the cranial base (N–S), the maxilla (Ptm–A), and the dimensions of the mandible (Co–Gn, Go–Gn,
and Co–Gn) from T1 to T2 were determined and statistically compared (Pearson’s correlation coefficients)
with the growth prediction assessed with the help of hand–wrist radiographs according to Greulich and
Pyle.
The results showed a highly significant correlation between statural growth increases and growth
prediction assessed from the hand–wrist radiographs (females: r = 0.68; males: r = 0.7). Concerning
craniofacial structures, the increase in mandibular corpus showed the highest correlation with growth
prediction (females: r = 0.21; males: r = 0.52), but this association would not allow a reliable growth
prediction. There was no significant correlation between growth increases of the cranial base, the maxilla,
the ramus, and the effective length of the mandible and growth prediction assessed with the help of
hand–wrist radiographs.
As each patient has an individual growth pattern and different craniofacial structures show individual
growth potential, it is questionable if quantitative craniofacial growth prediction with the help of hand–
wrist radiographs is reliable. However, in an individual case for the assessment of the timing of the
growth process, a hand–wrist radiograph can contribute to treatment planning.

Introduction
other treatment decisions, it would be favourable if the
As most orthodontic patients are growing individuals, remaining craniofacial growth of a patient could be precisely
orthodontists have to consider the remaining craniofacial predicted.
growth of each patient for successful treatment planning. Growth prediction can be assessed using physiological
During treatment, growth can cause significant changes in parameters such as peak growth velocity in standing height,
the dental and skeletal structures as well as in the profile. pubertal markers, dental development, and radiological
Abundant studies have shown that such changes are complex analysis of skeletal maturation. While physiological markers
due to the fact that each patient has an individual growth do not allow precise growth prediction, evaluation of
pattern (Björk, 1951, 1963; Nanda, 1955; Kraus et al., 1959; skeletal maturation with the help of radiographs is considered
Bambha and Van Natta, 1963; Johnston et al., 1965; Mitani, to be the more reliable approach. The most preferred method
1977; Fishman, 1979; Bishara et al., 1984; Bishara and is the use of hand–wrist radiographs (Chapman, 1972;
Jakobsen, 1985). Accordingly, onset duration, velocity, Grave and Brown, 1976; Houston et al., 1979). Commonly,
direction, and the amount of remaining growth differ hand–wrist radiographs are analyzed using the comparison
significantly among individuals with the same chronological method according to the atlas of Greulich and Pyle (1959),
age. which was based on a longitudinal growth study. The atlas
Remaining craniofacial growth can contribute to consists of plates of typical hand–wrist radiographs at 6
orthodontic correction as it is expected in patients with a monthly intervals of chronological age. To determine the
Class II malocclusion. On the other hand, remaining growth skeletal age of a particular patient, the hand–wrist bones are
can also have an adverse effect on a skeletal malocclusion, compared with those presented as standards in the atlas. The
particularly in patients with a Class III or open bite prediction of mature height is assessed based on the current
malocclusion. Therefore, in borderline cases, where the skeletal age and body height according to growth prediction
clinician has to decide whether orthognathic surgery is tables attached to the atlas. Thus, the atlas was originally
essential, in extraction cases, before the application of introduced to predict mature statural height by determining
extraoral forces to correct skeletal discrepancies, or due to the skeletal age using analysis of the hand and wrist.

Downloaded from https://academic.oup.com/ejo/article-abstract/31/4/438/403720

by guest
on 20 May 2018
 RELIABILITY OF HAND–WRIST RADIOGRAPHS 439

The use of hand–wrist radiographs to examine skeletal spine on the T2 lateral cephalogram [cervical vertebrae
maturity has been criticized as the patient is exposed to maturation indicator 4 (deceleration) or above,
additional radiation. Therefore, analysis of the cervical according to Hassel and Farman, 1995].
spine (Lamparski, 1972; Hassel and Farman, 1995) or of the Taking these criteria into account, 49 patients were
frontal sinus (Ruf and Pancherz, 1996) on lateral included in the study. The female group comprised 27
cephalograms have been recommended as alternative patients with a median age of 11 years 9 months at T1 (range
methods. However, both are rather vague and do not allow 9 years 7 months to 13 years 6 months) and 14 years 9
a precise prediction of the remaining growth potential. months at T2 (range 12 years 8 months to 15 years 11 months;
Orthodontists have regularly taken hand–wrist Figure 1). The male group consisted of 22 patients with a
radiographs of their patients to determine remaining median age of 12 years 6 months (range 9 years 7 months to
craniofacial growth before the beginning of treatment. As 14 years 8 months) and 15 years 2 months (range 13 years 5
the assessment of skeletal maturity with hand–wrist months to 20 years 3 months) at T1 and T2, respectively
radiographs was initially introduced to predict statural (Figure 1). Cephalograms were digitized and analyzed
height, several studies have questioned if such an approach (Figure 2) with the WinCeph® program (CompuGroup
could be applied to craniofacial structures. The results of Holding AG, Koblenz, Germany).
those studies are contradictory. While some authors conclude Relative dimensional growth changes between T1 and T2
that there is an association between peak velocity of (percentage) of statural height and length of the cranial base
craniofacial growth and peak velocity of statural height (N–S), the maxilla (Ptm–A), and the mandible (ramus: Co–
(Nanda, 1955; Johnston et al., 1965; Hunter, 1966, Go, corpus: Go–Gn, and effective length: Co–Gn) were
Bergersen, 1972), others (Seide, 1959; Bambha, 1961; calculated (Figure 3).
Fishman, 1982) could not find a significant relationship Growth changes were statistically compared with the
between skeletal maturity of the hand and wrist and growth growth prediction based on the hand–wrist radiograph at T1
of the craniofacial structures. However, most of these according to Greulich and Pyle (1959). Pearson’s correlation
investigations focused on time course and velocity of coefficients were calculated using a computer program
growth and neglected to investigate how precisely the (PlotIt® 3.2, Scientific Programming Enterprises, Haslett,
amount of remaining growth potential can be determined. Michigan, USA; Table 1).
The aim of the present study was to evaluate the prediction Female and male patients were considered separately due
reliability of remaining growth of different craniofacial to different growth curves.
structures with the use of hand–wrist radiographs.

Results
Subjects and methods Growth increments of all craniofacial structures were for
The records of 485 well-documented patients treated at the both genders in the range previously described in the
Clinic of Orthodontics and Pediatric Dentistry of the literature (Bhatia and Leighton, 1993).
University of Zurich, Switzerland, represented the database.
To produce representative results, only patients who fulfilled
the following criteria were selected.
1. No syndromes or specific disease.
2. Skeletal Class I malocclusion at the beginning of
treatment (T1), which was confirmed with the help of
the lateral cephalogram.
3. No growth modifying appliances used during treatment.
4. A hand–wrist radiograph available before the pubertal
growth spurt (T1).
5. Lateral cephalograms available at T1 and after (T2) the
pubertal growth spurt.
6. Statural height documented each time a lateral
cephalogram was obtained.
7. Patients had remaining growth potential at T1. This was
confirmed with the help of the hand–wrist radiograph
(females: standard 21 or less, males: standard 25 or
less, according to Greulich and Pyle, 1959).
8. Patients had clearly passed the pubertal growth peak at Figure 1 Age distribution of the female and male patients when the first
T2. This was confirmed by analysis of the cervical (T1) and second (T2) lateral cephalograms were obtained.

Downloaded from https://academic.oup.com/ejo/article-abstract/31/4/438/403720

by guest
on 20 May 2018
 440 D. VERMA ET AL.

Female group
Statural height of the female patients increased by 8 per
cent from T1 to T2 (Figure 3A). Comparison of this increase
with growth prediction of statural height based on the
Greulich and Pyle analysis showed a highly statistically
significant correlation (r = 0.68, P < 0.001; Table 1).
As the cranial base and maxilla increased by only 2 per
cent from T1 to T2, no statistically significant correlation
(r = 0.01, P = 0.96 for N–S and r = −0.16, P = 0.94 for
Ptm–A) with the growth prediction according to Greulich
and Pyle (1959) could be calculated.
The length of the mandibular corpus increased on average
by 5 per cent and ramus height by 6 per cent. Again, only
low correlation coefficients were found (r = 0.06 for Co–
Go, r = 0.21 for Go–Gn, and r = 0.29 for Co–Gn) which
were not statistically significant (P = 0.78 for Co–Go, P =
Figure 2 Lateral cephalometric radiographic points measured in the 0.29 for Go–Gn, and P = 0.29 for Co–Gn).
study: cranial base (N–S), the sagittal length of the maxilla (Ptm–A), the
mandibular ramus (Co–Go), the sagittal length of the mandibular corpus
(Go–Gn), and the effective length of the mandible (Co–Gn).
Male group
Male patients showed an increase of 15 per cent (Figure 3B)
in statural height during T1–T2. Compared with growth

Figure 3 Measurements of body height, cranial base (N–S), sagittal length of the maxilla (Ptm–A), mandibular ramus
(Co–Go), sagittal length of the mandibular corpus (Go–Gn), and effective length of the mandible (Co–Gn) at the first
(T1) and second (T2) lateral cephalogram for the female (A) and male (B) groups.

Downloaded from https://academic.oup.com/ejo/article-abstract/31/4/438/403720

by guest
on 20 May 2018
 RELIABILITY OF HAND–WRIST RADIOGRAPHS 441

Table 1 Statistical analysis of the measurements: Pearson’s To obtain reliable results, a test group should only include
correlation coefficient (r) and statistical significance between growth similar patients. This postulation is probably the greatest
prediction assessed with hand–wrist radiographs (Greulich and Pyle
problem in a clinical study as every patient shows individual
method) and growth changes during the observation period for the
body height, cranial base (N–S), sagittal length of the maxilla features. To minimize this problem, this research only
(Ptm–A), mandibular ramus (Co–Go), sagittal length of the mandibular included patients who would show physiological growth
corpus (Go–Gn) ,and effective length of the mandible (Co–Gn). during the observation period. Therefore, unlike the study
of Hunter (1966), which included patients with retarded and
Females (n = 27) Males (n = 22) accelerated growth and those treated with extraoral forces,
none of the subjects in the present investigation had a
Height 0.68*** 0.70*** specific condition, showed severe occlusal discrepancies at
N–S 0.01 0.11 the beginning of treatment, or had been treated with extraoral
Ptm–A −0.16 −0.16
Co–Gn 0.06 −0.009
or intermaxillary forces. All these aspects would have
Go–Gn 0.21 0.52* influenced normal growth.
Co–Gn 0.21 0.28 Overall, the results demonstrated that during T1–T2,
males grew more than females, both for statural height and
*P < 0.05; ***P < 0.001. the different craniofacial structures. In both groups, there
was a highly significant correlation between statural height
growth during T1–T2 and growth prediction assessed with
prediction assessed with the hand–wrist radiographs, a high the help of hand–wrist radiographs. This would confirm that
and statistically significant correlation was found (r = 0.7, the prediction method of Greulich and Pyle (1959) is reliable
P < 0.001; Table 1). for statural height even in today’s population, despite the
Within the male group, the cranial base increased by 5 fact that the atlas is based on children born between 1920
per cent and the maxilla by 4 per cent during T1–T2. and 1930. The correlation coefficients of r = 0.7 for males
When compared with the growth prediction according to and r = 0.68 for females are comparable with a previous
Greulich and Pyle (1959), no significant correlation was study (Moore et al., 1990). At T2, the average age of the
obtained. The increase in growth of the mandibular corpus females was 14 years 7 months and that of the males 15
(Go–Gn) was 7 per cent and showed a significant years 3 months. Even if most natural growth had taken place
correlation with the growth prediction assessed with the during this time, it must be assumed that the majority of the
hand–wrist radiographs (r = 0.52, P = 0.01). The patients in the sample still had some minor remaining
mandibular ramus grew 14 per cent but showed no growth potential. If the measurements had been taken in
statistically significant correlation (r = −0.009, P = 0.97) adulthood, the correlation coefficients might have been
with growth prediction. even stronger.
In the female group, both the cranial base and the maxilla
showed only minor growth, while in the male group, there
Discussion
were weak increases for the cranial base and for the maxilla
A successful orthodontic treatment plan requires consideration (5 and 4 per cent, respectively). These growth increments
of the remaining craniofacial growth in direction, velocity, were far less than those for stature. Consequently, significant
and quantity. A common method to predict the quantity of correlation coefficients are difficult to determine. A reliable
remaining growth has been to analyze skeletal maturity using growth prediction for the cranial base and the maxilla cannot
hand–wrist radiographs and to use the growth prediction be obtained with the help of hand–wrist radiographs.
tables in the atlas of Greulich and Pyle (1959). Therefore, the In the present study, growth of the mandible was analyzed
present study was based on analysis of hand–wrist radiographs separately in the vertical (Co–Go) and horizontal (Go–Gn)
with the comparison method according to those authors. As dimension. In the female group, growth increments in these
the atlas shows standards only with 6 month intervals, the dimensions were rather similar but less than that of statural
accuracy of the analysis of the hand–wrist radiographs to height. Again, no significant correlations could be found. It
determine skeletal age is generally limited. In addition, must therefore be concluded that for females hand–wrist
radiographs cannot in every case be assigned to the standards radiographs should not be used to predict the mature size of
with absolute congruence. Measurements of the different the mandible.
craniofacial structures were carried out by determining The results in the male group were slightly different.
specific cephalometric landmarks on the lateral cephalograms. Here, the sagittal length of the mandibular corpus increased
This method includes errors as morphological structures can by only 7 per cent, whereas ramus height showed a
be distorted due to X-ray beam geometry. In addition, the significant growth increase. However, a significant
localization of cephalometric landmarks shows intra- and correlation existed only between mandibular corpus length
interindividual variation (Sekiguchi and Savara, 1972; and growth prediction. This is in concordance with the
Midtgård et al., 1974; Kamoen et al., 2001). findings of Silveira et al. (1992) and Tofani (1972).

Downloaded from https://academic.oup.com/ejo/article-abstract/31/4/438/403720

by guest
on 20 May 2018
 442 D. VERMA ET AL.

The different results between the female and male group Bishara S E, Jakobsen J R 1985 Longitudinal changes in three facial types.
American Journal of Orthodontics 88: 466–502
concerning mandibular growth would confirm the statement of
Bishara S E, Peterson L, Bishara E C 1984 Changes in facial dimensions
Smith (1980) who concluded that only hand–wrist radiographs and relationships between the ages of 5 and 25 years. American Journal
of male patients would provide valuable information for of Orthodontics 85: 238–252
orthodontics, while those of females would not be useful. Björk A 1951 The significance of growth changes in facial pattern and their
In the present study, incremental growth changes of the relationship to changes in occlusion. The Dental Record 71: 197–205
anterior cranial base (S–N), length of the maxilla (Ptm–A), Björk A 1963 Variation in the growth pattern of the human mandible:
mandibular ramus (Co–Go), and mandibular corpus (Go– longitudinal radiographic study by the implant method. Journal of
Dental Research 42: 400–411
Gn) were examined. However, incremental growth is only
Chapman S M 1972 Ossification of the adductor sesamoid and the
one aspect of facial growth. Phenomena such as adolescent growth spurt. Angle Orthodontist 42: 236–245
displacement, rotation, or remodelling of different skeletal Enlow D H 1990 Facial growth, 3rd edn.W.B. Saunders, Philadelphia
structures were not taken into account, but have a major Fishman L S 1979 Chronological versus skeletal age, an evaluation of
impact when facial growth is studied (Enlow, 1990). craniofacial growth. Angle Orthodontist 49: 181–189
Unfortunately, these growth features cannot be predicted Fishman L S 1982 Radiographic evaluation of skeletal maturation. A
with certainty. Considering the complexity of facial growth, clinically oriented method based on hand-wrist films. Angle Orthodontist
52: 88–112
it must be concluded that any growth prediction of the
Grave K C, Brown T 1976 Skeletal ossification and the adolescent growth
craniofacial complex cannot be obtained with precision. spurt. American Journal of Orthodontics 69: 611–619
Greulich W W, Pyle S I 1959 Radiographic atlas of skeletal development
of hand and wrist. 2nd edn. Stanford University Press, Stanford
Conclusion
Hassel B, Farman A G 1995 Skeletal maturation evaluation using cervical
The possibility to precisely predict remaining craniofacial vertebrae. American Journal of Orthodontics and Dentofacial
Orthopedics 107: 58–66
growth would allow the orthodontist to establish a successful
Houston W J B, Miller J C, Tanner J M 1979 Prediction of the timing of the
treatment plan and anticipate the treatment outcome. adolescent growth spurt from ossification events in hand-wrist films.
Hand–wrist radiographs, which are commonly used for British Journal of Orthodontics 6: 142–152
this purpose, seem not to provide such information as each Hunter C J 1966 The correlation of facial growth with body height and
patient displays an individual growth pattern and different skeletal maturation at adolescence. Angle Orthodontist 36: 44–54
craniofacial structures show individual growth potential. Johnston F E, Hufham H P, Moreschi A F, Terry G D 1965 Skeletal
maturation and cephalofacial development. Angle Orthodontist 35: 1–11
Whether the routine use of hand–wrist radiographs for
Kamoen A, Dermaut L, Verbeeck R 2001 The clinical significance of error
quantitative craniofacial growth prediction justifies the measurement in the interpretation of treatment results. European Journal
additonal radiation exposure to the patient should be questioned. of Orthodontics 23: 569–578
However, in an individual case, for the assessment of the Kraus B S, Wise W J, Frei R H 1959 Heredity and the craniofacial complex.
timing of the growth process, a hand–wrist radiograph can American Journal of Orthodontics 45: 172–217
contribute to treatment planning. Lamparski D 1972 Skeletal age assessment utilizing cervical vertebrae.
Thesis. University of Pittsburgh, Pennsylvania
Midtgård J, Björk G, Linder-Aronson S 1974 Reproducibility of
Address for correspondence cephalometric landmarks and errors of measurements of cephalometric
cranial distances. Angle Orthodontist 44: 56–61
Dr Damian Verma Mitani H 1977 Occlusal and craniofacial growth changes during puberty.
Poliklinik für Kieferorthopädie American Journal of Orthodontics 72: 76–84
Zentrum für Zahn- Moore R N, Moyer B A, DuBois L M 1990 Skeletal maturation and
Mund- und Kieferkrankheiten craniofacial growth. American Journal of Orthodontics and Dentofacial
Orthopedics 98: 33–40
Welschnonnenstrasse 17
Nanda R S 1955 The rates of growth of several facial components measured
53111 Bonn from serial cephalometric roentgenograms. American Journal of
Germany Orthodontics 41: 658–673
E-mail: [email protected] Ruf S, Pancherz H 1996 Development of the frontal sinus in relation to
somatic and skeletal maturity. A cephalometric roentgenographic study
at puberty. European Journal of Orthodontics 18: 491–497
References Seide L J 1959 The relationship of dentofacial and skeletal maturation to
Bambha J K 1961 Longitudinal cephalometric roentgenographic study of malocclusion. American Journal of Orthodontics 45: 801–816
face and cranium in relation to body and height. Journal of the American Sekiguchi T, Savara B S 1972 Variability of cephalometric landmarks used
Dental Association 63: 776–799 for face growth studies. American Journal of Orthodontics 61: 603–618
Bambha J K, Van Natta P 1963 Longitudinal study of facial growth in Silveira A M, Fishman L S, Subtelny J D, Kassebaum D K 1992 Facial
relation to skeletal maturation during adolescence. American Journal of growth during adolescence in early, average and late maturers. Angle
Orthodontics 49: 481–492 Orthodontist 62: 185–190
Bergersen E O 1972 The male adolescent facial growth spurt: its prediction Smith R J 1980 Misuse of hand-wrist radiographs. American Journal of
and relation to skeletal maturation. Angle Orthodontist 42: 319–338 Orthodontics 77: 75–78
Bhatia S N, Leighton B C 1993 A manual of facial growth. Oxford Tofani M I 1972 Mandibular growth at puberty. American Journal of
University Press, Oxford Orthodontics 62: 176–195

Downloaded from https://academic.oup.com/ejo/article-abstract/31/4/438/403720

by guest
on 20 May 2018