THE ANATOMICAL RECORD 300:1609–1617 (2017)

Frontal Sinus Development and

Juvenile Age Estimation
KAITLIN MOORE* AND ANN ROSS
Sociology & Anthropology, North Carolina State University, North Carolina

ABSTRACT
Assessment of development is an important component of age estima-
tion in juveniles. One area that has not been fully investigated as a possi-
ble aging method is the development of the frontal sinus. The frontal
sinuses form when the ectocranial table of the frontal bone separates
from the endocranial table forming an air pocket in the bone. The endo-
cranial table ceases growth with the brain, while the ectocranial table is
displaced anteriorly as the facial bones continue growth. In order to
examine growth and the utility of the frontal sinuses for age estimation,
392 radiographs were examined ($5159 and #5233) from the Juvenile
Radiograph Database at North Carolina State University and the Patricia
Database from Mercyhurst University. The sample included individuals
who ranged in age from 0 to 18 years. Anterior view (or AP) radiographs
were examined and were grouped based upon the presence or absence of
the frontal sinus. Individuals were grouped into four age categories. A
one-way ANOVA was performed to test whether developmental phase was
related to age. Results from the ANOVA show that developmental phase
is significantly related to age (P <.0001). An ordinal logistic regression
was conducted to examine whether developmental phase could be used to
predict age. The results of the logistic regression suggest that develop-
mental phase is an accurate indicator of age (P <.0001, df 5 1, Chi-
Squared 5 537.2428); however, the age ranges can be quite wide and
should be utilized alongside other established methods of age estimation.
Anat Rec, 300:1609–1617, 2017. V C 2017 Wiley Periodicals, Inc.

Key words: frontal sinus; development; age estimation;

juvenile

The frontal sinus is an air filled cavity located on the
frontal bone of the skull. It consists of bilateral sinuses
separated by a bony septum, which is located in the pos-
terior portion of the supercilliary ridge between the
endocranial and ectocranial tables of the frontal bone
(Elnow and Hans, 1996; Dixon et al., 1997; Standring,
2008). The frontal sinus is lined with a moist mucosal
membrane that acts as a humidifier to warm air as it is
breathed in (McLaughlin et al., 2001; Verma et al.,
2015).
Unlike the other sinuses, the frontal sinus is absent at
birth (Enlow and Hans, 1996; Ponde et al., 2003; Standr-
ing, 2008; Tatlisumak et al., 2008; da Silva et al., 2009;
Tabor et al., 2009). The frontal sinus begins to form as two
small pea shaped pockets located above the orbits on
either side of the osteometric point glabella (Gagliardi
et al., 2004). The sinuses can begin forming in individuals
as young as two years of age; however, they often are not
visible radiographically until approximately six years of
age (Enlow and Hans, 1996; Dixon et al., 1997; Tatlisu-
mak et al., 2008; da Silva et al., 2009; Verma et al., 2015).
Growth continues through puberty, extending beyond the
middle part of the supercilliary ridge and towards the
*Correspondence to: Kaitlin M. Moore, North Carolina State
University, Sociology and Anthropology, 10 Current Drive Cam-
pus Box 8107, Raleigh, North Carolina 27695. E-mail:
kmmoor10@ncsu.edu
Received 10 May 2016; Revised 8 December 2016; Accepted
20 December 2016.
DOI 10.1002/ar.23614
Published online 19 May 2017 in Wiley Online Library
(wileyonlinelibrary.com).

C 2017 WILEY PERIODICALS, INC.

V
1610 MOORE AND ROSS

Fig. 1. Radiograph depicting an individual with frontal sinuses that are absent (birth to 5 years). Radiograph from Patricia (Pediatric Radiology
Interactive Atlas).
 FRONTAL SINUS DEVELOPMENT IN JUVENILE 1611

Fig. 2. Radiograph of an individual during the initial stage of frontal sinus development (age 6–7). The sinuses are circled in red. Radiograph
from Patricia (Pediatric Radiology Interactive Atlas).

medial portion of the orbital roof (Ponde et al., 2003; Mar-
ques et al., 2014; Verma et al., 2015). Growth ceases at
about 20 years of age, when the shape and size of the
sinus become stable (Tabor et al., 2009).
The mechanisms that are responsible for the develop-
ment and expansion of the frontal sinus are highly
regulated. Development is influenced by both brain
expansion and the growth of the midfacial region (Enlow
and Hans, 1996; Dixon et al., 1997). The skull grows
rapidly from birth to age seven in order to compensate
for rapid brain growth (Standring, 2008). During this
time, the endocranial table of the frontal bone is
1612 MOORE AND ROSS

Fig. 3. Radiograph depicting an individual during the initial stage of frontal sinus expansion (age 8–10). The sinus is circled in red. Radiograph
from Patricia (Pediatric Radiology Interactive Atlas).
 FRONTAL SINUS DEVELOPMENT IN JUVENILE
signaled by surrounding tissue to grow to provide space
for the expansion of the brain (Enlow and Hans, 1996).
The tables of the frontal bone begin to separate. When
brain growth slows, at approximately six years of age,
the endocranial table of the frontal bone ceases to grow
and adopts the general shape of the brain (Enlow and
Hans, 1996; Dixon et al., 1997). This growth period
accounts for the beginning stages of frontal sinus
development.
A second period of rapid growth in the skull takes
place at the onset of puberty (Standring, 2008). While
there is growth in every direction of the skull, this
period is generally associated with rapid growth of the
midfacial region (Standring, 2008). During this time, the
nasomaxillary complex of the face is remodeled and dis-
placed anteriorly and inferiorly, and the ectocranial table
of the frontal bone is similarly displaced (Enlow and
Hans, 1996; Dixon et al., 1997; Standring, 2008). This
displacement causes further separation between the two
tables of bone. The medullary bone between the tables is
resorbed, allowing for the growth of the sinus (Enlow
and Hans, 1996). Expansion continues until the sinus
occupies all the space between the two tables of bone
that was not being functionally utilized (Enlow and
Hans, 1996).
Frontal sinus shape is unique to each individual
(Schuller, 1921). Many studies have examined the
uniqueness of its shape, all of which have concluded
that no two are alike (Schuller, 1921; Tang et al., 2009).
Studies including monozygotic twins have also found
that the frontal sinus shape differs considerably (Uth-
man et al., 2010). One study examined the size and
shape of the frontal sinus among kin groups and found
that less than 2% of the individuals had similarly
shaped frontal sinuses (Chaiyasate et al., 2007; Camer-
iere et al., 2008; Tabor et al., 2009).
Though much research has been done on the heritabil-
ity of and identification from frontal sinus shape, little
research has been done examining frontal sinus develop-
ment and its association with age. Research that does
exist tends to focus on volume of the frontal sinus in
relation to age and utilizes adults rather than juveniles.
The purpose of this study is to more fully explore the
development of the frontal sinus and to test whether it
could be a skeletal parameter for juvenile age
estimation.
MATERIALS AND METHODS
Radiographs from three different locations were uti-
lized for this study. The samples from the North Caro-
lina State University Juvenile Radiographic Database
(NIJ 2010-DN-BX0K214) were comprised of radiographs
obtained from the North Carolina Office of the Chief
Medical Examiner. The database contains individuals
from birth to 18 years of age who died between 1974 and
2002. The database was amassed by the second author
and the data are freely available at https://sites.google.
com/a/ncsu.edu/radiograph-database-project/.
The Patricia database (Pediatric Radiology Interactive
Atlas), compiled by Mercyhurst University, was also uti-
lized for this study (NIJ 2008-DN-BX-K152). This data-
base includes radiographs from over 20 locations across
the country, including two clinical locations. Patricia
contains radiographs of individuals from birth to 21
1613
years of age. For this study, only juveniles under the age
of 18 were considered. This database can also be
accessed via the web at http://math.mercyhurst.edu/
~sousley/databases/radiographic_database/.
Several radiographs were received directly from the
North Carolina Office of the Chief Medical Examiner in
Raleigh, North Carolina. These radiographs were
obtained to increase the sample size of individuals
between the age of five and ten years, all of which were
from cases prior to 2013.
Radiographs were considered for inclusion in this
study if they were anterior views and if there was no
trauma to the frontal bone. Both males and females
were considered for inclusion, as were individuals of
varying ancestral origins. Radiographs were excluded if
the image was overexposed, making it difficult to distin-
guish between anatomical structures. A total of 392 indi-
viduals were included from these three sources (# 5 233,
$ 5 159).
The radiographs that were selected for this study first
underwent a visual assessment. The age, sex, and ances-
tral groups were recorded as were the presence or
absence of the frontal sinus. If the frontal sinus was pre-
sent, the shape and extent of the expansion of the sinus,
both laterally and superiorly, were recorded. The radio-
graphs were then scored numerically based on the size
and shape of the frontal sinus.
Individuals with no sign of frontal sinus development
were scored as a 0. Radiographs from these individuals
had frontal bones that appeared solid and uniform with
no areas of radiolucency visible above the orbits (Fig. 1).
The ages of these individuals ranged from birth to age
five.
Individuals in the first stage of frontal sinus develop-
ment also had frontal bones that appeared solid and uni-
form, but with two small, kidney bean shaped areas of
radiolucency on the medial portion of the superior bor-
der of the orbits near the anatomical landmark glabella.
This stage of development was scored as 1 and appeared
in individuals between six and seven years of age (Fig.
2).
Individuals between the ages of eight and ten years
old were in the initial stage of frontal sinus expansion.
In these individuals, the two areas of radiolucency had
expanded medially and laterally, but were still in the
glabella region of the frontal bone (Fig. 3). These indi-
viduals were scored as 2.
The final stage of development encompassed individu-
als from 11 to 18 years of age. Individuals in this stage
showed fully formed frontal sinuses and were scored as
3. The sinus was larger and more cavernous and
expanded laterally and superiorly on the frontal bone
(Fig. 4).
Statistical Analysis
A one-way analysis of variance (ANOVA) was per-
formed on the sample with phase as the independent
variable and age as the dependent variable. In addition,
a Tukey-Kramer adjustment for multiple comparisons
was used to determine if the mean ages of each phase
category were significantly different from one another.
Finally, the data were fit in a logistic regression to
assess if developmental phase was an accurate predictor
of age. A logistic regression was utilized because it can
1614 MOORE AND ROSS

Fig. 4. Radiograph of an individual with a fully formed frontal sinus (age 11–18). The sinus is circled in red. Adapted from Patricia (Pediatric
Radiology Interactive Atlas).
 FRONTAL SINUS DEVELOPMENT IN JUVENILE 1615
TABLE 1. Distribution of sample by sex and ancestry TABLE 2. Classification matrix of actual phase by
predicted phase
Ancestry Male Female Total
Actual Phase Predicted
European Americans 114 84 198
African Americans 68 45 113 0 1 2 3
Hispanic Americans 31 16 47 0 98% 0.74% 0.74% 0.37%
Native Americans 6 5 11 1 43% 33% 23% 0%
Asian Americans 3 2 5 2 12.50% 33% 50% 0
Indian Americans 1 0 1 3 0 0 2.38% 97.60%
Unknown 10 7 17
Total 233 159 392

TABLE 3. Prediction profile results showing the prob-

abilities of age and phase
Probability of Phase by age
Age 0 1 2 3
0–1 100%
2–4 95%
5 83% 15%
6 61% 32%
7 37% 46% 18%
8 15% 44% 36%
9 11% 28% 55%
10 12% 58% 28%
11 42% 53%
12 22% 76%
13 9% 90%
14 96%

Only ages 0–14 are presented.

These individuals spanned different ages and different

Fig. 5. Box plot depicting the spread and mean ages for each
developmental phase.
be applied to both continuous and categorical variables
and does not require an underlying assumption of nor-
mality to be met (Fournier and Ross, 2015).
The data were first fit with an ordinal logistic regres-
sion with both sex and age as dependent variables and
developmental phase as the independent variable. Sex
was not found to have a significant effect on the develop-
mental phase, so the model was refit, excluding sex as a
parameter. All statistical analyses were performed in
JMP Pro 12.0.1 (SAS 2015).
RESULTS
Simple statistics were calculated using Microsoft
Excel VR (Table 1). Bilateral absence of the frontal sinus,
or absence of both sides of the sinus, was observed in
eight individuals who were over age six. Seven of these
individuals were between the ages of six and eight
years. Because it is generally agreed upon that frontal
sinuses become visible in radiographs around age six,
the absence of the sinus could not be confirmed in these
seven individuals as they could have been developmen-
tally delayed. However, one individual was well into
their teens and thus, was the only individual in which
bilateral absence of the frontal sinus could be confirmed.
Unilateral absence of the frontal sinus was observed
in 0.2% of the sample (nine individuals). In all nine of
these individuals, the right side of the sinus was absent.
ancestral groups.
In this sample, there were 271 individuals who had
no signs of frontal sinus development, 21 individuals in
the beginning stages of growth, 16 showing the initial
spread of the frontal sinus, and 84 individuals with fully
formed frontal sinuses. Figure 5 depicts a box plot of the
age distribution for each developmental phase.
The ANOVA procedure indicated a statistically signifi-
cant relationship between the age and the development
stage of the frontal sinus (df 5 2, SS 5 13420.032,
4473.33, F Ratio 5 1325.666, Prob > F <.0001).
The Tukey-Kramer adjustment showed that phase 0 is
significantly different than phases 1–3 (P-value5
<.0001). However, phase 1 is not significantly different
than phase 2 (P -value 5 0.1179), but it is significantly
different from phase 0 and phase 3 (P-value 5 <.0001),
and phase 3 is significantly different from phase 0–2
(P-value 5 <.0001).
The logistic regression showed that sex did not have a
significant effect on developmental phase (P-val-
ue 5 0.9310). The model was refit, excluding sex as a
parameter. The logistic regression results suggested that
this model is useful for predicting age in juvenile indi-
viduals (ChiSq 5 537.24, df 5 1, P-value <.0001). The
area under the Receiver Operating Characteristic (ROC)
curve was .99 for each of the developmental phases, indi-
cating a strong separation between the phases. The mis-
classification rate for the model was 0.07%. Table 2
presents the classification matrix, which shows that for
phase 0 the misclassification rate is 0.01%, for phase 3 it
is 0.02% with the highest misclassification rates falling
in phases 1 and 2, which is most likely because of the
relatively small sample size in these two phases.
1616 MOORE AND ROSS
A prediction profiler, which measures the predictive
response as one variable is changed (in this case age)
while the other (phase) is held constant, was also used
in the logistic regression to provide probabilities for each
age being a specific phase (Table 3). The prediction pro-
filer showed that juveniles do not develop frontal sinuses
between 0 and 4 years of age. They begin to develop a
frontal sinus between 5 and 6, they transition from
phase 0 to 3 between 7 and 9 years of age. At 10 years
of age they transition between phase 1 and 3 and transi-
tion between phase 2 and 3 between the ages of 11 and
13. Finally there is a 96% probability that the frontal
sinus is fully developed by age 14.
DISCUSSION
Frontal sinus development and growth is strongly
associated with specific age related periods of skull
growth. These periods of growth are detectable with the
phases of frontal sinus development as examined in this
study. Frontal sinus absence of phase 0 was observed
after birth and during the initial growth period of the
skull. Though the endocranial table of the frontal bone
was being remodeled to support brain expansion, the
sinus was not radiographically visible. The sinus only
became visible in radiographs towards the end of the
first period of skull growth. Coincidentally, this is the
time when the endocranial table ceases growing and con-
forms to the general shape of the brain. This first phase
of development was not observed in individuals younger
than six years of age, which conforms to what most of
the clinical literature reported about frontal sinuses not
being visible in radiographs of juveniles younger than
six years of age (Enlow and Hans, 1996; Dixon et al.,
1997; da Silva et al., 2009; Verma et al., 2015).
The second phase of frontal sinus development as
defined in this study coincides to the timespan between
the primary and secondary periods of skull growth. This
occurs in the years leading up to puberty, during a time
in which frontal sinus growth is slow and steady (Ponde
et al., 2003).
The third phase of development, fully formed frontal
sinuses, occurs when an individual enters puberty. As
this is the period of time when much of the facial growth
occurs, this phase shows the most dramatic change
between the developmental phases.
While this study provides valuable information
regarding the use of the frontal sinuses to estimate age,
there were some shortcomings.
The sources used for this study obtained the majority
of their radiographs through medical examiners, indicat-
ing that these individuals were either homicide victims
or experienced death under suspicious circumstances.
According to the Crimes Against Children Research Cen-
ter, there are two major periods in life during which
juveniles experience a high risk of homicide. In young
children, individuals under the age of one year have the
highest homicide rate with a slight decline until age six
(Finkelhor and Ormrod, 2001). These homicides are usu-
ally committed by family members. The second period of
high homicide rates in juveniles begins at age 12 with a
steep increase during the late teen years (Finkelhor and
Ormrod, 2001). High rates of homicide in teens are asso-
ciated with poverty, gang activity, and drug-related vio-
lence (Finkelhor and Ormrod, 2001). As these time
periods experience the highest rates of homicides, they
make up the bulk of the radiographs in this sample.
Juveniles aged 6 to 12 years have homicide rates lower
than any other subset of the US population (Finkelhor
and Ormrod, 2001). Because this age range has the low-
est homicide rates, there were very few postmortem
radiographs from this group included in this study
sample.
Future directions would include radiographs or medi-
cal scans of clinical data. This would increase the sample
size overall, but more importantly would increase the
number of radiographs for the 6 to 12 year old age
range. A larger sample size would allow for a more in
depth analysis of the age ranges and perhaps would
allow for the identification of more phases of develop-
ment with narrower associated age ranges.
Although the results of this study indicate that frontal
sinus development can be an accurate indicator of age,
the ranges observed in this study span several years.
Stages 1 and 2 provide the narrowest age ranges, span-
ning two and three years, respectively. The largest age
range is seen in stage 3, which spans eight years. This
may limit the usefulness of frontal sinus development as
an age indicator in juveniles older than 11. However,
this study does show that by age 14 the frontal sinuses
are fully formed in this sample and the older age bracket
could be tightened.
The primary use of the frontal sinus in medico-legal
contexts has been the identification of unknown individ-
uals. In this arena, it has proven to be extremely useful,
boasting up to 100 percent accuracy in identifying indi-
viduals (Besana and Rogers, 2010). However, the growth
of the frontal sinus and its link to age has largely been
ignored. This study shows that frontal sinus develop-
ment can be utilized to accurately predict age in suba-
dults. This may prove useful in cases of fragmented or
incomplete remains that make traditional methods of
age estimation difficult. Until further work can be done
to narrow the age ranges associated with each develop-
mental stage, this method should be used with caution
and alongside existing methods of juvenile age
estimation.
ACKNOWLEDGMENTS
The authors would like to thank the North Carolina
Office of the Chief Medical Examiner and specifically
Mr. Clyde Gibbs for providing the radiographs used in
this study. And we thank Dr. Beth Murnane, the second
author’s son’s pediatrician, for fruitful discussions of
frontal sinus development in children.
LITERATURE CITED
Besana JL, Rogers TL. 2010. Personal identification using the fron-
tal sinus*. J Forensic Sci 55:584–589.
Cameriere R, Ferrante L, Molleson T, Brown B. 2008. Frontal sinus
accuracy in identification as measured by false positives in kin
groups. J Forensic Sci 53:1280–1282.
Chaiyasate S, Baron I, Clement P. 2007. Analysis of paranasal sinus
development and anatomical variations: A CT genetic study in
twins. Clin Otolaryngol 32:93–97.
da Silva RF, Prado FB, Caputo IGC, Devito KL, de Luscena Botelho
T, Junior ED. 2009. The forensic importance of frontal sinus
radiographs. J Forensic Leg Med 16:18–23.
 FRONTAL SINUS DEVELOPMENT IN JUVENILE
Dixon AD, Hoyte DA, Ronning O. 1997. Fundamentals of craniofa-
cial growth. Boca Raton: CRC Press
Enlow DH, Hans MG. 1996. Essentials of facial growth. Philadel-
phia: Saunders.
Finkelhor D, Ormrod R. 2001. Homicides of Children and Youth.
US Department of Justice Office of Juvenile Justice and Delin-
quency Prevention.
Fournier NA, Ross AH. 2015. Sex, ancestral, and pattern type vari-
ation of fingerprint minutiae: A forensic perspective on anthropo-
logical dermatoglyphics. Am J Phys Anthropol 160: 625–632.
Gagliardi A, Winning T, Kaidonis J, Hughes T, Townsend G. 2004.
Association of frontal sinus development with somatic and skele-
tal maturation in aboriginal australians: A longitudinal study.
HOMO 55:39–52.
Marques JAM, de Oliveira Musse J, Gois BC, Galv~ ao LCC, Paranhos
LR. 2014. Cone-beam computed tomography analysis of the frontal
sinus in forensic investigation. Int J Morphol 32:660–665.
McLaughlin RB, Rehl RM, Lanza DC. 2001. Clinically relevant frontal
sinus anatomy and physiology. Otolaryngol Clin North Am 34:1–22.
Ponde JM, Metzger P, Amaral G, Machado M, Prandini M. 2003.
Anatomic variations of the frontal sinus. Minim Invasive Neuro-
surg 46:29–32. doi:10.1055/s-2003-37956 [doi]
1617
Schuller A. 1921. Das rontgenogram der stirnhohle: ein hilfsmittel
f€
ur die identit€atsbestimmung von schadeln. Monatsschr Ohren-
heilkd Laryngorhinol 5:1617–1620.
Standring S. 2008. Gray’s Anatomy: The basis of clinical practice.
Edinburgh: Churchill Livingstone/Elsevier.
Tabor Z, Karpisz D, Wojnar L, Kowalski P. 2009. An automatic rec-
ognition of the frontal sinus in x-ray images of skull. IEEE Trans
Biomed Eng 56:361–368.
Tang J, Hu D, Jiang F, Yu X. 2009. Assessing forensic applications
of the frontal sinus in a chinese han population. Forensic Sci Int
183:104. e1–104. e3.
Tatlisumak E, Ovali GY, Asirdizer M, Aslan A, Ozyurt B, Bayindir
P, Tarhan S. 2008. CT study on morphometry of frontal sinus. Cli-
nAnat 21:287–293.
Uthman AT, Al-Rawi NH, Al-Naaimi AS, Tawfeeq AS, Suhail EH.
2010. Evaluation of frontal sinus and skull measurements using
spiral CT scanning: An aid in unknown person identification.
Forensic Sci Int 197:124. e1–124. e7.
Verma P, Verma KG, Khosa R, Kumar S, Basavaraju S,
Patwardhan N. 2015. Combined use of frontal sinus and nasal
septum patterns as an aid in forensics: A digital radiographic
study. North Am J Med Sci 7:47.