Forensic Dentistry

Forensic Dentistry

Age estimation of indian adults from

orthopantomographs
Sudhanshu Saxena(a)

Department of Public Health Dentistry,

Peoples College of Dental Sciences and
Research Centre, Peoples group, Bhanpur,
Bhopal, Madhya Pradesh, India.

(a)

Abstract: The aim of this study was to develop a method for estimating
the chronological age of Indian adults based on the relationship between
age and various morphological variables of canine teeth, obtained using orthopantomographs. Orthopantomographs of 120 selected patients
were digitized, and radiographic images of the right maxillary canine in
each case were processed using a computer aided drafting program. Pulp/
tooth area ratio, pulp/root length ratio, pulp/tooth length ratio, pulp/root
width ratio at the cemento-enamel junction level, pulp/root width ratio
at midroot level, and pulp/root width ratio at the midpoint between the
cemento-enamel junction and the midroot of the canine were calculated by measuring various features on the images. Pearsons correlation,
multiple linear regression, one way analysis of variance (ANOVA), and
Students t-test were used for statistical analysis. Regression equations
were developed to estimate age from morphological variables. The observed minus the estimated age in the total study sample ranged from
-2.2 to +1.5 years, in males from -0.9 to +0.8 years, while in females it
was from -1 to +0.8 years. Differences between observed and estimated
ages of subjects were not statistically significant. In conclusion there is a
linear relationship of pulp/root width ratio at mid-root level and pulp/
tooth area ratio of the right maxillary canine with chronological age in
the Indian population. Age of subjects can therefore be estimated with a
good degree of accuracy using regression equations.
Descriptors: Forensic Dentistry; Age Determination by Teeth; Adult.

Introduction
Correponding author:
Sudhanshu Saxena
Department of Public Health Dentistry,
Peoples College of Dental Sciences and
Research Centre, Peoples group, Bhanpur
Karond bypass road
Bhopal - Madhya Pradesh - India
ZIP: 462037
E-mail: [email protected]

Received for publication on Oct 11, 2010

Accepted for publication on Dec 15, 2010

It is often necessary to estimate an individuals age due to certain

questions related to legal requirements in paleodemographic research or
in a forensic context.
Although several parts of the body can be used for age estimation,
the poor condition of the remains in particularly severe crashes or fires
in cases of those recently dead; or of moisture and burial conditions in
the case of historic subjects, make many parts of the body unusable. For
these reasons, the teeth are the part of the body most frequently used for
identification and age estimation.1
In children, age determination from teeth is a relatively simple, accurate procedure and is based on the stages of development and eruption
of teeth. However, in adults it is a challenge to medico-legal science. 2
Up to now, a multiplicity of methods have been applied to this problem,

Braz Oral Res. 2011 May-Jun;25(3):225-9

225

Age estimation of indian adults from orthopantomographs

including methods which analyze the various forms

of tooth modification such as wear, dentin transparency, tooth cementum annulations, racemization of
aspartic acid and apposition of secondary dentin.
Some of the methods are very complex and destructive and are therefore not normally used; wear and
the apposition of secondary dentin are the currently
available nondestructive methods. Tooth wear is
influenced by various external factors (masticatory function, type of food, timing and sequence of
tooth eruption), tooth form, position of teeth, thickness and hardness of enamel and predisposition to
enamel hypoplasia. However, apposition of secondary dentin is a continuing, regular process which is
only modified by caries or particular abrasion. Secondary dentin has been studied by sectioning and
radiography.1
The study of tooth radiographs is a nondestructive and simple process which can be applied to both
living and deceased persons, in contrast to other
time consuming, expensive, less reliable and destructive methods which may not be acceptable for
ethical, religious, cultural or scientific reasons.3 Further, procedures such as digitization of panoramic
radiographs and computer assisted image analysis
avoid the bias inherent in observer subjectivity and
improve reliability, accuracy and precision.4
Any tooth can be used to assess age. Canine
teeth are a good candidate for age estimation because they are often present in old age. They are less
likely than other anterior teeth to suffer wear as a
result of particular work and are the single-rooted
teeth with the largest pulp area and thus are the easiest to analyze.1
Based on the above background, this study was
conducted to develop a method for estimating the
chronological age of Indian adults based on the relationship between age and various morphological
variables of canine teeth, obtained using orthopantomographs.

Materials and Methods

Based on the results of a pilot study, to achieve
a 5% level of significance and 80% power, a study
with a sample size of 120 participants was required,
calculated using Power and Precision v. 4 (Biostat

226

Braz Oral Res. 2011 May-Jun;25(3):225-9

Inc., Englewood, USA). One hundred and twenty

orthopantomographs (Planmeca Oy, Helsinki, Finland) from patients of both sexes seen at the Peoples
College of Dental Sciences and Research Centre,
Bhopal, Madhya Pradesh state, India, were collected
based on the following criteria:
a. Patients aged between 21 to 60 years
b. Fully erupted right maxillary canine
c. Fully formed root of right maxillary canine
d. Right maxillary canine free from any pathology
such as caries, periodontitis, abrasion, erosion,
fracture or impaction
e. Absence of restoration and endodontic filling of
the right maxillary canine
f. Absence of malaligned or rotated right maxillary
canine
Orthopantomographs were taken as part of routine treatment that was being rendered to the patient. Participation in the study was voluntary. Relevant data including date of birth and gender were
entered into the proforma after obtaining informed
consent. Ethical clearance was obtained from the
ethical committee of the Peoples College of Dental
Sciences and Research Centre.

Methods
Orthopantomographs of selected patients were
obtained and digitized using an ASTRA 4000U
scanner (UMAX Technologies Inc., Dallas, USA).
Radiographic images of canines (RIC) were processed using the computer-aided drafting program,
Auto CAD 2005 (Autodesk Inc., San Rafael, USA).
Twenty points around the edge of the tooth outline and ten points around the pulp outline of the
right maxillary canine were identified. Measurements of the canine tooth area and the pulp area
from the radiographic images were then evaluated.
The tooth length, pulp length and root length
were measured. The width of the root and pulp at
3 different levels, one at the cemento-enamel junction (CEJ), the second at midroot level and the third
at the level of the midpoint between the CEJ and
the midroot level were measured. All measurements
were carried out by a single observer. To test intraobserver reproducibility, a random sample of thirty

Saxena S

RIC were re-examined after an interval of one week.

The following morphological variables were calculated from the above measurements:
AR = pulp/tooth area ratio
p= pulp/root length ratio
r = pulp/tooth length ratio
a = pulp/root width ratio at CEJ level
c = pulp/root width ratio at midroot level
b = pulp/root width ratio at midpoint level between CEJ level and midroot level
The observed age was calculated by subtracting
the date of birth from the date of the radiograph.
Statistical analysis was performed using SPSS
v. 17 (SPSS Inc., Chicago, USA). Pearsons correlation coefficients were calculated for the relationship
between observed age and predictive variables. A
multiple linear regression model for age estimation
was developed by selecting those variables which
contributed significantly to age estimation. One way
analysis of variance (ANOVA) was used to compare
differences between observed and estimated age in
different age groups. Students t-test was used to
compare the morphological variables of males with
those of females and to compare observed age with
estimated age. A p value<0.05 was considered statistically significant.

Results
Kappa statistics showed that the intra-examiner
agreement score (k=0.94) was almost perfect.
In the present study, 60 males and 60 females (total 120 subjects) were equally divided into four age
groups, i.e. 21-30 years, 31-40 years, 41-50 years,
and 51-60 years.
The morphological variables did not show any
significant difference between the sexes (p > 0.05).
Pearsons correlation coefficients between observed
age and predictive morphological variables in males,
females and in the total sample population were
highly significant and inversely correlated for AR
(pulp/tooth area ratio) and c (pulp/root width ratio
at midroot level) (p<0.001).
The subjects ages were modeled as a linear function of the morphological variables (predictors). A
multiple linear regression procedure was therefore

applied to optimize this model.

Table 1 shows the regression analysis of the total
sample, while Table 2 shows the results in males and
Table 3 in females. Among all the morphological
variables tested, only AR (pulp/tooth area ratio) and
c (pulp/root width ratio at midroot level) showed
significant results. Hence, a regression model utilizing AR and c was used (Table 4) which yields
the following linear regression formula to estimate
chronological age:
For subjects of unknown gender:
Age=72.48 -203.74(AR) -51.69(c)
For male subjects:
Age=73.33 -209.97(AR) -54.06(c)
For female subjects:
Age=72.12 -179.12(AR) -68.66(c)
Table 1 - Regression analysis with all morphological variables (predictors) and age as the dependent variable in the
total study sample (n=120).
Regression
coefficient

Significance

Standard
error

t value

p value

71.65

0.60

120.30

<0.001

AR

-204.52

3.78

-54.16

<0.001

0.09

0.44

0.21

0.833

0.59

0.70

0.84

0.402

0.63

0.59

1.07

0.285

1.48

1.02

1.46

0.148

-52.21

3.66

-14.24

<0.001

Model
Constant

Dependent variable: Age

Table 2 - Regression analysis with all morphological variables (predictors) and age as the dependent variable in
males (n=60).
Regression
coefficient

Model
Constant

73.05

Significance

Standard
error

t value

p value

0.67

109.13

<0.001

6.0

-29.19

<0.001

AR

-175.09

-0.11

0.38

-0.28

0.780

-0.92

0.75

-1.22

0.226

-0.95

1.41

-0.67

0.504

-0.72

1.44

-0.50

0.617

-70.54

4.83

-14.60

<0.001

Dependent variable: Age

Braz Oral Res. 2011 May-Jun;25(3):225-9

227

Age estimation of indian adults from orthopantomographs

Table 3 - Regression analysis with all morphological variables (predictors) and age as the dependent variable in females (n=60).
Standard
error

t value

p value

72.54

0.59

122.52

<0.001

AR

-208.10

3.11

-66.82

<0.001

1.05

0.57

1.83

0.073

-0.39

0.73

-0.53

0.597

0.49

0.42

1.17

0.247

0.23

0.88

-0.26

0.792

-56.37

3.45

-16.36

<0.001

Model

Regression
coefficient

Significance

Constant

Dependent variable: Age

Table 4 - Regression analysis with selected morphological

variables (predictors) and age as the dependent variable.
Group

Standard
error

t value

p value

72.48

0.19

388.39

<0.001

AR

-203.74

3.71

-54.93

<0.001

-51.69

3.63

-14.24

<0.001

Constant

73.33

0.21

349.76

<0.001

AR

-209.97

2.99

-70.24

<0.001

-54.06

3.27

-16.54

<0.001

Constant

72.12

0.16

437.99

<0.001

AR

-179.12

5.00

-35.83

<0.001

-68.66

4.53

-15.16

<0.001

Model

Regression
coefficient

Males

Total

Constant

Females

Significance

Dependent variable: Age

The observed minus the estimated age in the

total sample population ranged from -2.2 to +1.5
years. There was no significant difference between
observed and estimated age for any of the age groups
(Table 5). In addition, the differences of observed
and estimated age between age groups were also
not statistically significant (ANOVA value=2.259,
p=0.085).
The observed minus the estimated age in males
ranged from -0.9 to +0.8 years, while in females it
ranged from -1 to +0.8 years. There was no significant difference between observed and estimated age
for either males or females (Table 6).

228

Braz Oral Res. 2011 May-Jun;25(3):225-9

Table 5 - Comparison of observed and estimated age in

subjects grouped according to age.
Age
groups
(years)
21-30

Age in years (meanSD)

Observed age
24.862.98

Estimated age
24.643.25

t-test
value

p value

0.267

0.790 (NS)

31-40

35.252.73

35.302.64

0.061

0.952 (NS)

41-50

44.603.02

44.463.11

0.171

0.865 (NS)

51-60

55.092.55

54.812.60

0.432

0.667 (NS)

Over all

39.9511.58

39.8011.56

0.100

0.921(NS)

*NS = Not significant

Table 6 - Comparison of observed and estimated age of

subjects according to gender.
Age in years (meanSD)
Observed age

Estimated age

t-test
value

Male

40.2011.15

40.0411.13

0.080

0.937 (NS)

Female

39.7012.08

39.5612.05

0.062

0.951 (NS)

Over all 39.9511.58

39.8011.56

0.100

0.921 (NS)

Gender

p value

*NS = Not significant

Discussion
The results of the present study showed no significant difference between observed and estimated
age. Age was estimated using regression equations
based on pulp/tooth area ratio and pulp/root width
ratio at mid-root level.
In 1925 Bodecker established that the apposition
of secondary dentin correlated with age. 5 In 1995,
Kvaal et al. presented a method for age estimation
which was based on investigation of periapical radiographs, 3 while Paewinsky et al. verified the applicability of this method on orthopantomograms.6
In 2004 Cameriere et al. found a linear relationship between the pulp/root width ratio at mid-root
level, the pulp/tooth area ratio, and chronological
age.4 This finding was consistent with the findings
in present study. Cameriere et al. stated that the ratio between pulp and tooth area correlated best with
age.4 With this background they developed regression equations for age estimation using the pulp/
tooth area ratios of canines from the maxilla and
the mandible separately.1
In their study, Babshet et al. found that Cameri-

Saxena S

eres formula, based on the Italian population, is not

as applicable to the Indian population.7 I also found
the same result in my pilot study using this formula.
These results indicate that racial and cultural factors might play an important role in age estimation.
Therefore I decided to develop regression equations
for the Indian population. I used both pulp/root
width ratio at mid-root level and pulp/tooth area ratio in regression equations to make the equation as
close to ideal as possible.
The present study showed no significant influence of gender on age estimation using canine tooth
measurements, which was similar to the findings of
previous studies.1,4,8 Nevertheless I developed separate formulas for males and females to predict age
as accurately as achievable.
The present study showed 99.7% variance and
0.60 standard error of the estimate, when selected
variables were used. This result was similar to results of the study conducted by Cameriere et al.,
who reported 97% variance with 2.37 years standard error of the estimate.9

Though my results are promising, I cannot generalize them to other populations. Further research
should be aimed at analyzing a larger sample, including not only age and gender but also racial and
cultural parameters. This method of age estimation
cannot be applied to multirooted teeth, as accurate
measurements are difficult to perform on these. Similarly, as the curved arch of the jaw is projected on
to a flat film, there will always be a certain amount
of distortion when measuring the image presented
there.8 To create a uniform amount of distortion in
all the radiographs used I chose a particular tooth,
the right maxillary canine.

Conclusion
Within the limitations of the study it can be concluded that there is a linear relationship of pulp/root
width ratio at mid-root level, and pulp/tooth area
ratio of the right maxillary canine with chronological age in the Indian population. In unknown subjects age can be estimated using regression equations based on the above mentioned variables.

References
1. Cameriere R, Ferrante L, Belcastro MG, Bonfiglioli B, Rastelli
E, Cingolani M. Age estimation by pulp/tooth ratio in canines
by peri-Apical Xrays. J Forensic Sci. 2007Jan;52(1):166-70.
2. Jain RK, Rai B. Age estimation from permanent molars attrition of Haryana population. Indian J Forensic Odontol. 2009
Apr;2(2):59-61.
3. Kvaal SI, Kolltveit KM, Thomsen IO, Solheim T. Age estimation of adults from dental radiographs. Forensic Sci Int.
1995Jul 28;74(3):175-85.
4. Cameriere R, Ferrante L, Cingolani M. Variations in pulp/
tooth area ratio as an indicator of age: a preliminary study. J
Forensic Sci. 2004 Mar;49(2):317-9.
5. Meinl A, Tangl S, Pernicka E, Fenes C, Watzek G. On the
applicability of secondary dentin formation to radiological age estimation in young adults. J Forensic Sci. 2007
Mar;52(2):438-41.

6. Paewinsky E, Pfeiffer H, Brinkmann B. Quantification of

secondary dentine formation from orthopantomograms--a
contribution to forensic age estimation methods in adults. Int
J Legal Med. 2005Jan;119(1):27-30.
7. Babshet M, Acharya AB, Naikmasur VG. Age estimation in
Indians from pulp/tooth area ratio of mandibular canines. Forensic Sci Int. 2010 Apr 15;197(1-3):125.e1-4. Epub 2010Jan
27.
8. Singaraju S, Sharada P. Age estimation using pulp/tooth area
ratio: a digital image analysis. J Forensic Dent Sci. 2009
May;1(1):37-41.
9. Cameriere R, Cunha E, Sassaroli E, Nuzzolese E, Ferrante L.
Age estimation by pulp/tooth area ratio in canines: study of
a Portuguese sample to test Camerieres method. Forensic Sci
Int. 2009 Dec 15;193(1-3):128.e1-6. Epub 2009 Oct 24.

Braz Oral Res. 2011 May-Jun;25(3):225-9

229