Yale University
EliScholar – A Digital Platform for Scholarly Publishing at Yale
Yale Medicine Thesis Digital Library
School of Medicine
January 2019
Ossification Of The Phalanges Of The Foot And Its
Relationship To Peak Height Velocity And The
Calcaneal System
Mekka Garcia
Follow this and additional works at: https://elischolar.library.yale.edu/ymtdl
This Open Access Thesis is brought to you for free and open access by the School of Medicine at EliScholar – A Digital Platform for Scholarly
Publishing at Yale. It has been accepted for inclusion in Yale Medicine Thesis Digital Library by an authorized administrator of EliScholar – A Digital
Platform for Scholarly Publishing at Yale. For more information, please contact elischolar@yale.edu.
Recommended Citation
Garcia, Mekka, “Ossification Of The Phalanges Of The Foot And Its Relationship To Peak Height Velocity And The Calcaneal System”
(2019). Yale Medicine Thesis Digital Library. 3496.
https://elischolar.library.yale.edu/ymtdl/3496
Ossification of the Phalanges of the Foot and its Relationship
to Peak Height Velocity and the Calcaneal System
A Thesis Submitted to the Yale University School of Medicine in
Partial Fulfillment of the Requirements for the Degree of Doctor of
Medicine
By
Mekka R. Garcia
2019
Abstract
Background: There are multiple skeletal maturity grading systems, but none of them
utilizes the phalanges of the foot. To minimize radiation, it would be ideal if one could
assess the skeletal maturity of a foot based on bones seen on routine foot x-rays, if
guided growth is being considered as a treatment option, as in hallux valgus. We
developed a system that in combination with the calcaneal system, can closely predict
skeletal maturity and help with the timing of surgical interventions of the foot.
Methods: We selected 94 healthy children from the Bolton-Brush study, each with
consecutive radiographs from age ten to fifteen years old. Using the AP view, we
analyzed the ossification patterns of the phalanges and developed a six stage
classification system. We then determined the Peak Height Velocity (PHV) for each
subject and defined its relationship with our system. Our system was then compared to
the previously established calcaneal system.
Results: We calculated an Intraclass correlation coefficient (ICC) range of 0.957-0.985
with an average of 0.975 and interclass reliability coefficient of 0.993 indicating that this
method is reliable and consistent. Our system showed no significant difference between
sexes, with respect to PHV, which makes it a reliable surrogate for determining bone age
in pediatric and adolescent patients.
Conclusions: Our system has a strong association with the calcaneal system. It is reliable
and correlated more strongly with PHV than chronological age. The system requires
knowledge of the ossification markers used for each stage but is easily used in a clinical
setting.
Acknowledgements
I would like to thank the Yale University StatLab and the co-authors of this work for their
guidance and support.
Allen D. Nicholson1, Afam M. Nduaguba1, James O. Sanders2, Raymond W. Liu3, Daniel R.
Cooperman1
1 Department of Orthopaedics and Rehabilitation, Yale School of Medicine, PO Box
208071, New Haven, CT 06510
2 Department of Orthopaedics, University of Rochester School of Medicine, 601
Elmwood Ave, Box 665, Rochester, NY 14642
3 Department of Orthopaedics, Case Western Reserve University School of Medicine,
11100 Euclid Ave, Cleveland, OH 44106
Table of Contents
Introduction……………………………………………………………………………………….………1
Materials and Methods………………………………………………………………………………6
Results………………………………………………………………………………………………………..9
Discussion…………………………………………………………………………………………………20
References……………………………………………………………………………………………….24
1
Introduction
Skeletal maturity staging is used to determine a child’s current growth velocity
and growth potential. Many pediatric diseases including growth, endocrine and
chromosomal disorders depend on skeletal maturity systems to assess the amount of
growth that a child has remaining, whether bone age and chronological age are
concordant.1 If a discrepancy is seen, an underlying disorder may be suggested. While
assessment of bone age is widely used by most pediatric physicians, current systems
remain complex and traditionally involve comparing the patient’s radiographs to an
atlas composed of hundreds of standard images.2-4
First described by Todd5, methods of evaluating “skeletal age” have become a
standard in pediatrics through the Greulich and Pyle atlas.3 The Todd and the Greulich
and Pyle atlases describe a method of determining “skeletal age” from hand
radiographs. These atlases made use of the Bolton-Brush study conducted in Cleveland,
Ohio by Dr. T. Wingate Todd from 1926 to 1942.3,6 This study followed the growth and
development of 4,435 healthy children who had serial radiographs of the skull, left
shoulder, elbow, wrist and hand, hip, knee, and foot in addition to anthropometric data
such has height and weight. The Greulich and Pyle hand atlas grouped the children of
the Bolton-Brush study by sex and chronological age and then by the “average”
appearance of the bones of the hand for each age group.3 The atlas involves comparison
of hand radiographs to references in the atlas to establish the “skeletal age” of a child
and remains one of the most commonly used systems. Given that males and females
mature at different chronological ages, the definition of skeletal age as defined by
2
Greulich and Pyle becomes problematic.7 Another issue with the Greulich and Pyle atlas
and other similar atlases is the primarily white demographics of the children used for
the database.8-10 However, many studies have shown that the Greulich and Pyle atlas is
generally applicable to modern children although less so around puberty.7,10 Moreover,
the atlas has been shown to yield a large inter-observer error.35
Developed in 1975, the Tanner-Whitehouse II (TW-II) method assesses skeletal
maturity using radiographs of the hand.11 Since Tanner incorporated the concept of
peak height velocity into the TW-II method, there have been several studies showing
that skeletal maturity is more closely related to the timing of PHV than chronological
age.12-16 The Tanner-Whitehouse method established twenty regions in the hand, each
divided into distinct morphologic stages. Based on the appearance of each ossification
centers and the sex of the child, a score is assigned to each region and then added for
an overall maturity score. In 2001, Tanner updated his skeletal maturity system to
include population growth charts of modern children with the morphological grading
staying the same. This system, Tanner-Whitehouse III (TW-III), requires scoring of
multiple ossification centers and access to complex scoring tables. The complexity of the
TW-III can cause significant inter-rater variability.17
The Sanders method is derived from the Greulich and Pyle method as well as the
TW-III atlas and uses a series of eight reference descriptors to evaluate maturity and has
been used to determine the prognosis of adolescent idiopathic scoliosis (AIS) curve
progression.7 Using the Sanders hand scores and AIS curvature, Sanders estimated the
likelihood of AIS curve progression that would require surgery (>50 degrees) by
3
following twenty-two girls with AIS for two years through their growth spurt. One
disadvantage of the Sanders method is that most the hand stages occur after PHV has
been reached, therefore making it difficult to predict how much time children have until
they reach PHV, the time of maximal curve progression in scoliosis.
In addition to the hand atlases, there are multiple widely-used systems for other
ossification centers in the body. Perhaps the most widely used skeletal maturity system
after the Greulich and Pyle hand atlas, the Risser system uses the iliac apophysis to
predict remaining vertebral growth.18 Due to the availability of pelvic radiographs in
scoliosis patients, the RIsser system remains commonly used. However, the maturation
of the iliac apophysis begins after PHV, preventing the prediction of maturity before
PHV. Moreover, the Risser system has been shown to poorly correlate with scoliosis
curve acceleration, preventing an accurate prognosis of curve progression.19-24 Another
method to utilize the ossification of the hip is the Oxford method which grades nine
ossification centers that surround the hip and has been used for evaluation slipped
capital femoral epiphysis (SCFE).25 Researches have found a narrow window of bone
age, as determined by the Oxford method, where SCFE occurs.26 Others have shown that
the modified Oxford method, which consists of five ossification centers, are strongly
predictive of contralateral SCFE.27,28
A maturation system exists that uses radiographic imaging of the foot and ankle.
It was created by Hoerr, Pyle and Francis. Like its predecessor atlases of the hand and
knee, the atlas made use of osseous landmarks as bone age indicators.2-4 A recent
evaluation of the Hoerr atlas found a strong correlation between estimated “bone age”
4
and chronological age.29 However, it is widely known that individuals of the same
chronological age can differ in their skeletal age, yielding a wide spectrum of peak
height velocity (PHV). This necessitates a more thorough system of skeletal maturity
where PHV is incorporated and can therefore act as a surrogate measure. Such a system
would be invaluable in determining the timing of surgical interventions.
For example, in determining treatment options for hallux valgus,
hemiepiphysiodesis requires the patient to be skeletally immature, and osteotomies are
typically performed after the patient is skeletally mature. If a maturity system could
determine the amount of time before and after PHV is reached, it would be helpful in
evaluating the timing of epiphysiodesis.30-32 Greene et al. established normative values
on growth of the first metatarsal but did not correlate their findings to PHV.30 A six-stage
system of calcaneal apophyseal ossification, as previously described, allows for the
identification of the period of growth before and after PHV and is highly reliable, but
requires lateral views of the foot.33
The Shorthand Bone Age (SBA) developed by Heyworth and colleagues is derived
from the commonly used Greulich and Pyle method.34 Another method worth
mentioning is the Sauvegrain Method, which uses elbox x-rays. It has been shown to
reliably correlate with the timing of the PHV.35 Our group also recently developed a
system utilizing the calcaneal apophysis that resembles the Risser system. This system is
able to predict the skeletal maturity of children before and after PHV has been
reached.33 In our study, given that a standard AP view is routinely used for evaluating
the severity of foot pathologies, we wanted to explore the utility of the phalanges of the
5
foot for assessing skeletal maturity. To avoid additional radiation exposure, our purpose
was to generate a skeletal maturity system using existing radiographic images of foot
pathologies and utilize the phalanges of the foot as a surrogate for peak height velocity,
which is a useful marker in timing of surgical correction in foot pathologies. We also
compare our system to the calcaneal system and explore the utilization for a combined
skeletal maturity system.
6
Materials and Methods
The children in this study are the same as those used for the Greulich and Pyle
atlas of the hand and the calcaneal apophyseal ossification system, in which serial AP
and lateral foot radiographs of 94 children (49 females, 45 males) were followed for
over a decade with consecutive radiographs, made at least yearly from age ten to
fifteen, which is the age range most associated with PHV. These radiographs were part
of the Bolton-Brush collection in Cleveland, Ohio, collected by Dr. T. Wingate Todd from
1929 to 1942. These children were part of a prospective, longitudinal study of growth in
healthy children, some of whom entered the study in infancy, and many of whom were
followed to the end of growth. They had serial x-rays taken of their skull and left
shoulder, elbow, wrist and hand, hip, knee and foot on multiple occasions. In addition,
other anthropometric data were gathered whenever the children received x-rays such
as height and weight. Heights were measured using a stadiometer with standardized
measurement technique allowing consistency in measurements over time and between
observers.
The Brush Inquiry consists of 4435 children recruited primarily through the
Cleveland area schools and through referral from family physicians. The families were
above average in economic and educational status. The majority of the children were
white (92.2%) and the remainder were mostly black (7.7%).3,5-6 Those who were selected
for the study had no gross physical or mental defects, along with the permission of their
parents to participate until the conclusion of the study.
7
We had 732 AP and 738 lateral x-rays available. We evaluated all AP x-rays that
clearly demonstrated all of the epiphyses of the proximal phalanges of the 2nd through
the 5th toes and the distal phalanx of the 1st toe and had a matching lateral x-ray taken
the same day that demonstrated the entire calcaneal apophysis. This resulted in 728
matching sets of x-rays.
Using the AP view, we examined the degree of ossification and fusion of the
digital proximal epiphyses of the proximal phalanges and the distal phalanx of the first
ray. The lateral x-ray of the foot taken at the same session as the AP view was graded
using the previously described calcaneal system.33 All x-rays were reviewed by the
author of the thesis (MRG). One hundred randomly selected sets of x-rays were
evaluated by a board certified pediatric orthopedic surgeon (DRC), an orthopedic
resident with considerable experience in developing and using bone maturity
assessment tools (ADN), and another orthopedic resident without such experience
(AMN). These three authors first graded 100 randomly selected sets of x-rays, then had
a consensus building session led by MRG, then graded another 100 x-rays. Intraclass
correlation coefficient (ICC) (two-way mixed model & absolute agreement) and
interclass reliability coefficient were calculated using IBM SPSS and all other statistics
and graphs were generated using Excel 2016 (Microsoft, Redmond, Washington).
The subjects in this study had their heights taken each time x-rays were
obtained. The peak height velocity was calculated using these serial height
measurements after the approach of Tanner and Davies using a cubic spline to
determine the maximum velocity which was defined as peak height velocity.36 The
8
subject’s age for this value was used for the point in time of peak height velocity. Under
the mentorship of Dr. Cooperman and guidance of the co-authors, the lead author
graded the images, calculated statistics and wrote the published article. This work has
been published in Journal of Children’s Orthopaedics.38
9
Results
The ossification of the proximal phalanges and the distal phalanx of the great toe
occurs in an orderly sequence. The ossification center of the distal epiphysis of the great
toe is first seen at 14 and 18 months of age for females and males, respectively. By 18
and 24 months, the ossification centers of all phalanges in females and males can be
seen.2
The orderly ossification and fusion of the foot phalanxes was divided into six
stages, shown in Figure 1A-F. In MEKKA 0, at least one of the proximal phalanx has an
epiphysis that is not as wide as its corresponding metaphysis. This finding is often most
noticeable for the fifth proximal phalanx. MEKKA 1 is characterized by all digital
proximal epiphyses of the proximal phalanges being as wide or wider than the
metaphysis, or “covered,” as is the proximal epiphysis of the distal phalanx of the first
digit. The epiphyseal plates of proximal phalanges 2-5 are concave in shape. MEKKA 2 is
marked by the formation of a small, medial epiphyseal “hook” over the metaphysis of
the distal phalanx of the first ray. Presence of the “hook” is diagnostic of stage 2, even if
other criteria of stages 0 and 1 are not met. MEKKA 3 is marked by the initiation of
fusion of the proximal epiphysis of the second, third, and/or fourth proximal phalanges.
Fusion typically starts in the center of the physis. It is still possible to see the epiphyseal
“hook” on the first ray. In MEKKA 4, fusion of the proximal epiphysis of the first and/or
fifth proximal phalanges is seen. Again, it is still possible to see the epiphyseal “hook” on
the first digit. In MEKKA 5, fusion of the digital proximal epiphysis and the proximal
epiphysis of the distal phalanx of the first digit are complete. The complete ossification
10
of the growth plate is represented by cortical continuity with or without the presence of
a residual white physeal scar.
Figure 1. The Metaphysis, Epiphysis hooK sKeletal Assessment (MEKKA) of the
phalanges
Reprinted with permission from Garcia MR, Nicholson AD, Nduaguba AM, Sanders JO,
Liu RW and Cooperman DR. Ossification of the phalanges of the foot and its relationship
to peak height velocity and the calcaneal system. Journal of Children’s Orthopaedics
2018; 12:84-90.38
A) MEKKA 0: not all digital epiphyses of the proximal phalanges are covered
B) MEKKA 1: all digital epiphyses of the proximal phalanges and the proximal
epiphysis of the distal phalanx of the first digit are covered. It is ‘covered’ when
the epiphysis is as wide or wider than the metaphysis. It can also be noted that
the epiphyseal plate of the proximal phalanges 2 to 5 are concave
11
C) MEKKA 2: capping of the metaphysis of the first digit by the epiphysis as
represented by a small, medial epiphyseal ‘hook’ over the metaphysis
D) MEKKA 3: initiation of fusion of the proximal epiphyses of the second, third
and/or fourth proximal phalanges. The ‘hook’ on the first ray may still be visible.
E) MEKKA 4: initiation of fusion of the proximal epiphyses of the first and/or fifth
proximal phalanges. Again, the ‘hook’ on the first digit may still be visible.
F) MEKKA 5: complete ossification of all digital epiphyses of the proximal phalanges
and the proximal epiphysis of the distal phalanx of the first digit
Table I shows the years before the PHV and the range of ages relative to PHV for each
MEKKA stage. The presence of the “hook” of MEKKA 2 appears at a mean age of 1.79
years before PHV. Fusion of the proximal epiphysis of digits 2, 3, and/or 4 is initiated in
MEKKA 3 at a mean age of 0.59 years before PHV, while complete fusion of all digital
proximal epiphysis and the proximal epiphysis of the big toe occurs at a mean age of
3.90 years after PHV.
12
Figure 2 shows a comparison of the MEKKA and calcaneal system with respect to peak
height velocity (PHV). PHV is achieved in MEKKA 3 and between calcaneal stages 3 and
4. There was an overlap in the initiation of fusion between the two stages (a total of 33
children that is both MEKKA 3 and calcaneal stage 4). The timing of the MEKKA stages
with respect to the PHV between sexes are not statistically different as seen in Table II.
The chronological age for each MEKKA stage is shown in Table III.
13
Figure 2. Comparison of the MEKKA and calcaneal system with respect to peak height
velocity (PHV).
Reprinted with permission from Garcia MR, Nicholson AD, Nduaguba AM, Sanders JO,
Liu RW and Cooperman DR. Ossification of the phalanges of the foot and its relationship
to peak height velocity and the calcaneal system. Journal of Children’s Orthopaedics
2018; 12:84-90.38
14
A box-and-whisker plot shows the age with respect to the PHV for the MEKKA and
calcaneal stages. The black lines represent the range for each stage, while the blue box
represents the middle 50% of the data. The blue line inside each box represents the
median, while the black diamond in the middle represents the mean. Negative numbers
represent years before PHV and positive numbers represent years after. Both MEKKA
and calcaneal 5* represent the first appearance of complete fusion.
15
The trend, in general, between the mean and median and first and third quartile, seems
to overlap which suggests a normal distribution of the data. There is a distribution of
MEKKA stages corresponding to specific calcaneal scores and vice versa as shown in
Figure 3. The MEKKA and calcaneal stages were combined and for each combined stage
with a sample size >5, the mean number of years before or after PHV were plotted
(Table IV and Figure 4).
An Intraclass correlation coefficient (ICC) range of 0.957-0.985 with an average of 0.975
and interclass reliability coefficient of 0.993 were calculated.
16
Figure 3. Distribution of calcaneal scores within MEKKA scores (A) and vice versa (B).
17
Reprinted with permission from Garcia MR, Nicholson AD, Nduaguba AM, Sanders JO,
Liu RW and Cooperman DR. Ossification of the phalanges of the foot and its relationship
to peak height velocity and the calcaneal system. Journal of Children’s Orthopaedics
2018; 12:84-90.38
18
Figure 4. A box-and-whisker plot of the combined MEKKA and calcaneal stages with
respect to peak height velocity (PHV).
Reprinted with permission from Garcia MR, Nicholson AD, Nduaguba AM, Sanders JO,
Liu RW and Cooperman DR. Ossification of the phalanges of the foot and its relationship
to peak height velocity and the calcaneal system. Journal of Children’s Orthopaedics
2018; 12:84-90.38
19
A box-and-whisker plot of the combined metaphysis, epiphysis hook skeletal assessment
(MEKKA) (‘M’) and calcaneal (‘C’) stages show the years before (negative) and after
(positive) the peak height velocity (PHV). The black lines represent the range for each
stage, while the blue box represents the middle 50% of the data. The line inside each
box represents the median, while the black diamond in the middle represents the mean.
M0/C0 represents the immaturity up to 8+ years before PHV while M5/C5 represents
full maturity of up to 8+ years after PHV. Combined stages with sample size of <20 were
excluded from the graph but are shown in Table 4.
20
Discussion
Many methods for assessment of skeletal maturity have been developed. The
most common system in use today remains the hand-wrist skeletal ages of Greulich and
Pyle.3 The atlas established a series of reference hand radiographs that can be used to
establish bone age through comparison to the reference radiographs. However, it has
been shown to yield a large inter-observer error and has been problematic when applied
to different ethnic groups.34-35 Moreover, this work was done prior to the understanding
of PHV delineated by Tanner, and height velocity was not considered.36-37
An atlas of skeletal development of the foot was developed by Hoerr et al. using
the same Bolton-Brush study cohort as Greulich and Pyle.3 Both atlases were created
before the concept of PHV was described by Tanner.14-15 The Hoerr atlas established
standard plates outlining the timing of appearance and patterns of ossification centers
of the foot. Hoerr et al. described the proximal phalanges of the second, third, and
fourth digits as the first to fuse, similar to the findings in our study. Moreover, they
found that the calcaneal apophysis is the last ossification center to appear and the last
to fuse, although they don’t comment on the process of calcaneal ossification.2 Hoerr et
al. found the epiphyseal-diaphyseal fusion to be complete by age 17.5 years in males
and 15.0 years in females. Similarly, our data showed a mean age of 16.5 years in males
and 15.2 years in females for complete fusion. We also showed that regardless of sex,
ossification of the proximal phalanges and distal phalanx of the great toe, with respect
to PHV, does not differ significantly. The consistency of the MEKKA ossification method
makes it a suitable system for assessing PHV and skeletal maturity of the foot. For
