close

Вход

Забыли?

вход по аккаунту

?

Comparability in skeletal maturation research.

код для вставкиСкачать
Comparability in Skeletal Maturation Research
WILLIAM M. MOORE
Medical Director, Ross Laborntories, Columbus, Ohio 43216
ABSTRACT
Comparability is a fundamental issue in skeletal maturation
research. Since the introduction of the first edition of the Greulich-Pyle Atlas
and the Tanner-Whitehouse method, a number of methodologic reports have
appeared regarding potential sources of error, reliability and replicability i n
the assessment of skeletal maturity from hand-wrist radiographs. Some of these
reports are mentioned and two recent examples of methodologic studies are
cited. Maximum reliability of skeletal assessments can be expected only when
there is strict adherence to carefully standardized investigative procedures.
Technical as well as human factors must be taken into account to insure minimal variation in findings within and between laboratories over time. Single or
serial skeletal radiographs uniformly taken on properly identified subjects constitute a valuable permanent record of biologic maturation. While the film
image can be considered as objective evidence of skeletal maturity, a subjective
element is introduced in observing and reporting the presence or absence of
particular ossification centers, or rating a n ossification pattern against a
standard. Intra- and interobserver skeletal maturity assessment replicability
relates to such factors as motivation, training, assessment method, and quality
control procedures. Suggestions are presented to facilitate comparability in
skeletal maturation research, including the possibility of preparation and distribu tion of sets of standardized skeletal radiographs for periodic determination and improvement of assessor reliability.
Since the introduction of the first edition of the Greulich-Pyle Atlas (Greulich
and Pyle, ' 5 0 ) and the Tanner-Whitehouse method (Tanner, Whitehouse and
Healy, '62), a number of methodologic
reports have appeared regarding potential sources of error, reliability, and replicability in the assessment of skeletal
maturity from hand-wrist radiographs. It
seems appropriate in the Symposium on
the Assessment of Skeletal Maturity to
mention some of these reports, to cite
some recent examples of methodologic
studies, and to raise for discussion some
suggestions designed to improve comparability in skeletal maturation research.
Mainland in the mid-1950's described
systematic and variable errors in the assessment of radiographs (Mainland, '53,
'54). Briefly, he reported that one observer demonstrated a systematic error in
tending to underestimate skeletal ages
when compared with previous assessments
by experts of the same radiographs. He
concluded that the skeletal assessment
method would probably be sufficiently reAM. J. PHYS.ANTHROP., 35: 411416.
liable in the comparison of average skeletal ages of groups of children, but i t
would be of doubtful value in the assessment of a single radiograph or of a child's
progress. Regarding variable errors, he
indicated that there was no significant
difference in variable error associated
with age of child, sex, differences between
skeletal and chronologic age, or differences between radiographs of the same
child, except as related to poor reproductions of some films. In both communications an appeal for more data was issued.
The second edition of the Greulich-Pyle
Atlas carries an effective response to the
points made by Mainland (Greulich and
Pyle, '59).
Early in the 1960's Acheson and colleagues investigated the reIiability of assessing skeletal maturity from radiographs and presented findings related to
the use of the Greulich-Pyle Atlas (Acheson et al., '63), the bone-specific approach
(Acheson et al., '64), and a comparison of
the Greulich-PyIe Atlas and the TannerWhitehouse methods (Acheson et al., '66).
41 1
41 2
WILLIAM M. MOORE
Several papers also appeared involving
lateral comparisons of skeletal maturity
in the human hand and wrist (Dreizen
et al., ’57; Roche, ’63) and, more recently, studies have been done on factors influencing the replicability of assessments
of skeletal maturity (Roche, Davila, Pasternak and Walton, ’70) including the
effect of training (Roche, Rohinann,
French and Davila, ’70).
EXAMPLES OF RECENT METHODOLOGIC
STUDIES
To avoid the appearance of favoring
either the Greulich-Pyle or the TannerWhitehouse methods of skeletal maturity
assessment, I will refer to recent methodologic studies involving the application of
both methods. In the first example, two
investigators with extensive experience in
skeletal maturation assessment, independent of each other and with no background
information other than the sex of the
child, made estimates of skeletal age for
each individual center in hand-wrist radiographs of 157 Chinese children, utilizing the Greulich-Pyle Atlas. The children
were born between January, 1960 and
March, 1967 and the films were obtained
between September, 1969 and March,
1970. While the data are not yet fully
analyzed, certain findings are worth mentioning. Less than 6 % of individual center
readings were more than 12 months apart.
The frequency distribution of maximum
single center interobserver skeletal age
difference is in table 1. Parenthetically,
the first 20 films rated in this double-blind
study showed a maximum interobserver
rating difference of 18 months, whereas
increasing the sample size led to several
differences of as much as 30 months
and a single difference of 42 months.
Another aspect of this study relates to
a difference in technique of the two investigators. That is, one deliberately randomized the sequence of bone assessment
while the other deliberately read bones in
“blocks” such as all metacarpals, all proximal phalanges, etc. It is apparent that
the investigator who randomized bone assessments has a lower frequency of all
bones in particular “block” being assigned the same skeletal age. However,
the investigator using the randomized approach assigned a zero when a center was
missing, whereas the other investigator
read the end of the shaft. This difference
in technique would tend to reduce the frequency of all bones in a “block’ being
assigned the same skeletal age under the
randomized approach. This factor, however, probably could not account for the
relatively higher frequencies of all bones
being assigned the same skeletal age utilizing the “block’ approach (table 2).
In another example, the Tanner-Whitehouse method of assessing skeletal ma-
TABLE 1
Maximztm interobserver difference in skeletnl age for single bones zismg the Greuhclt-Pyle
Atlas o n Chiizese boys a n d gcrls
~
Maximum
difference
Boys
Total
Girls
Number
Per cent
Number
Per cent
Number
Per cent
3
6
9
12
15
18
21
24
27
30
33
36
39
42
2
4
12
19
16
21
6
1
4
2
2.3
4.6
13.6
21.6
18.2
23.8
6.8
0
0.0
4
15
14
20
13
3
5.8
21.8
20.3
29.0
18.8
4.4
1.1
0
4.6
2.3
0
0
0
1
0
0
0
0.0
0.0
0.0
1.1
0.0
0.0
0.0
0.0
0.0
0.0
0.0
1.3
5.1
17.2
21.0
22.9
21.0
5.7
1.3
2.5
1.3
0.0
0
0
0
0
0
0
2
8
27
33
36
33
9
2
4
2
1
0.6
Total
88
100.0
69
100.1
157
99.9
months
0.0
0.0
413
COMPARABILITY IN SKELETAL MATURATION RESEARCH
TABLE 2
Frequency of nll bones w i t h i n the same “block” b e m g assigned t h e same skeletal age
in (I s a m p l e of 157 C h i n e s e children
(Grezclich-Pyle M e t h o d )
Metacarpals
no.
Randomized
approach
“Block”
approach
Prox. phalanges
5%
3
1.9
3
1.9
1
0.6
29
18.5
11
7.0
44
28.0
18
11.5
5,
TABLE 3
Rating summations
Assessor
Average
Standard
deviation
140.80
136.24
136.09
135.00
133.71
133.49
133.18
133.11
133.09
45.65
46.15
44.16
45.89
46.87
45.20
45.55
46.98
42.48
Q
Current study
L
Y
.
Z
1
From. Malina, ’68;96.
no.
no.
0.0
Comparison of the r n t m g s of 28 bones averaged
over 45 widlographs with thow w p o i t r d
b y A c h e s o n e t nl. (’64) 1
( T n n i i e r - W h i t e h o u s eM e t h o d )
0
Distal phalanges
0
turity was used. In this instance, the investigator, prior to the assessment of a
semi-longitudinal sample of white and
Negro elementary school children, assessed the films used by Acheson and colleagues (Acheson et al., ’64) in their
study of the reliability of assessing skeletal maturity from hand-wrist films (Malina, ’68). Table 3 indicates the standing
of this investigator i n comparison with
the participants in the Acheson study.
The unweighted summations of ratings
for all 28 bones of the hand and wrist,
when rank ordered, allowed the investigator to judge his standing among the
other assessors. The investigator then proceeded to assess approximately 2200 films
without knowledge of the age and sex
of the subjects. Subsequently, he reassessed every fifteenth film throughout the
series and recorded the average unweighted summations of ratings for all
28 bones for each assessment (table 4)
and the average sums of 7 round bones,
13 long bones, and skeletal age for first
and second assessments (table 5 ) . The
N
P
M
Mid. phalanges
no.
5%
%
degree of replicability was high with correlations for replicate readings yielding
r’s of f 0 . 9 8 2 , +0.981, and +0.989 for
round bones, long bones and skeletal age,
respectively (Malina, ’68, ’70).
DISCUSSION
I n view of the multiple potential sources
of error and factors affecting skeletal maturity assessments, maximum reliability
can be expected only when there is strict
adherence to carefully standardized investigative procedures. Technical as well as
human factors must be taken into account to insure minimal variation in findings within and between laboratories over
time. Single or serial skeletal radiographs
uniformly taken on properly identified
subjects constitute a valuable permanent
record of biologic maturation. While the
film image can be considered as objective
evidence of skeletal maturity, a subjective element is introduced in observing
and reporting the presence or absence of
particular ossifications centers, or rating
a n ossification pattern against a standard.
Comparability of findings is frequently a
fundamental issue in skeletal maturation
research when as assessor rates the same
film at different times, or different films
at the same or different times. The need
for comparability of findings is increased
when multiple assessors rate the same or
different films at different times. To insure comparability, irrespective of the assessment method, assessors must have a
good and constant level of motivation,
satisfactory training and experience, and
a sensitivity for the need of periodic quality control procedures.
Improving comparability in skeletal
maturation research. To facilitate comparability in skeletal maturation research, the possibility of preparation and
414
WILLIAM M. MOORE
TABLE 4
Average unweighted s u m m a t i o n s of ratings for rill 28 bones on two separate assessments
(Tanner-Whitehouse Method)
First assessment
1st 50 films
2nd 50 films
3rd 50 films
Total 150 films
Second assessment
Average
Standard
deviation
Average
Standard
deviation
129.72
135.14
128.00
130.95
22.71
24.18
26.37
24.49
128.58
133.64
126.04
129.42
21.99
24.35
25.94
24.20
From: Malina. ’68; 97.
1
TABLE 5
Averages, standard deviations, a n d product-moment correlation coefficients f o r the
s u m of seven round bones, t h e slim of thirteen long bones, a n d
skeletal age on tulo separate assessments 1
(Trinner-Whitehouse Method)
First assessment
Average
Second assessment
Standard
deviation
Average
Standard
deviation
r
74.07
74.03
82.61
76.81
0.989
0.981
0.981
0.982
1st 50 films
2nd 50 films
3rd 50 films
Total 150 films
192.54
202.84
184.98
193.45
Round bones
76.23
186.80
84.22
193.06
176.26
92.47
185.37
84.32
1st 50 films
2nd 50 films
3rd 50 films
Total 150 films
195.72
216.90
196.00
202.87
Long bones
81.29
89.49
86.32
85.77
195.98
209.52
194.10
199.87
76.44
87.28
85.40
82.90
0.981
0.979
0.984
0,981
1st 50 films
2nd 50 films
3rd 50 films
Total 150 films
9.67
10.25
9.62
9.85
Skeletal age
2.15
1.89
2.35
2.14
9.62
10.02
9.49
9.71
2.10
1.81
2.34
2.09
0,990
0.983
0.992
0.989
1
From: Malina, ’68; 98.
distribution of sets of standardized skeletal radiographs for periodic determination
and improvement of assessor reliability
deserves consideration. This approach has
proved useful in the establishment and
maintenance of comparability in microbiological and clinical chemistry laboratories in the United States and throughout the world by the presentation of
standard “unknowns” for determinations
by such agencies as the Center for Disease Control of the U . S. Public Health
Service and the various reference laboratories of the World Health Organization.
An important feature of such a program
is that there be prompt feedback of results of the standardization tests, together with appropriate comments on how
reliability might be improved.
During the Intensive Course in Human
Biology directed by Gabriel Lasker and
Morris Goodman and held at the Wayne
State Medical School in Detroit, Michigan, in conjunction with the meeting of
the AAPA-SSHB in April, 1968, previously
rated hand-wrist radiographs were available for graduate student training. It
would seem that implementation of occasional follow-up exercises of this type
involving not only graduate students, but
also principal investigators, would promote greater comparability of skeletal
maturation assessments, both within and
between laboratories. This suggestion is
predicated on a minimum cost basis in
view of the fact that films can be easily
and faithfully reproduced and mailed from
one center to another, thereby precluding
COMPARABILITY IN SKELETAL MATURATION RESEARCH
expensive transportation and subsistence
costs which would be necessitated by the
movement of investigators from one laboratory to another. The process could
work the other way, however, by having
a single individual or small group of individuals travel from one laboratory to another either to standardize their own techniques or to assist in the standardization
of others. Most participants in this symposium can probably recall instances
where either one or the other, or both of
these procedures have been followed. If
some of these beneficial or not so helpful
experiences could be shared during the
discussion, it might be easier to arrive at
a conclusion as to the feasibility of adopting these suggestions for improving comparability i n skeletal maturation research.
ACKNOWLEDGMENTS
I thank Dr. Ann Sproul of the Child
Health and Development Studies, Oakland, California, Dr. Marjorie M. C. Lee
of the University of Nebraska College of
Dentistry, Lincoln, Nebraska, as well as
Dr. Robert M. Malina of the University
of Texas, Austin, Texas, for the use of
previously unpublished data and findings.
LITERATURE CITED
Acheson. R. M., G. Fowler, E. I. Fry, M. Janes,
K. Koski, P. Urbano and J. J. Van Der Werff
Ten Bosch 1963 Studies i n the reliability of
assessing skeletal maturity from X-rays. Part I.
Greulich-Pyle Atlas. Hum. Biol., 35: 317-349.
Acheson, R. M., J. H . Vicinus and G. B. Fowler
1964 Studies i n the reliability of assessing skeletal maturity from X-rays. Part 11. The BoneSpecific Approach. Hum. Biol., 36: 211-228.
Studies in the reliability of assessing
skeletal maturity from X-rays. Part 111. Greu-
415
lich- Pyle Atlas and Tanner-Whitehouse Method
Contrasted. Hum. Biol., 38: 204-218.
Dreizen, S . , R. M. Snodgrasse, H . Webb-Peploe,
G. S. Parker and T. D. Spies 1957 Bilateral
symmetry of skeletal maturation i n the human
hand and wrist. Amer. J. Dis. Child., 93: 122127.
Greulich, W. W., and S. I. Pyle 1950 First Ed.
1959 Second Ed. Radiographic Atlas of Skeletal Development of the Hand and Wrist. Stanford University Press. Stanford, California.
Mainland, D. 1953 Evaluation of the skeletal
age method of estimating children’s development. I. Systematic errors i n the assessment of
roentgenograms. Ped., 1 2 : 114-129.
1954 Evaluation of the skeletal age
method of estimating children’s development.
11. Variable errors i n the assessment of roentgenograms. Ped., 13: 165-1 73.
Malina, R. M. 1968 Growth, maturation and
performance of Philadelphia Elementary School
Children. Ph.D. Dissertation i n Physical Anthropology, Graduate School of Arts and Sciences,
University of Pennsylvania.
1970 Skeletal maturation studies longitudinally over one year in American Whites
and Negroes six through thirteen years of age.
Hum. Biol., 42: 377-390.
Roche, A. F. 1963 Lateral comparisons of the
skeletal maturity of the human hand and wrist.
Amer. J. Roentgen., 89: 1272-1280.
Roche, A. F., G. H. Davila, B. A. Pasternak and M. J.
Walton 1970 Some factors infiuencing the
replicability of assessments of skeletal maturity
(Greulich-Pyle). Amer. J. Roentgen., 99: 29%
306.
Roche, A. F., C. G. Rohmann, N. Y. French and
G. H. Davila 1970 Effect of training on replicability of assessments of skeletal maturity
(Greulich-Pyle). Amer. J. Roentgen., 98: 51 1515.
Tanner, J. M., R. H. Whitehouse and M. J. R.
Healy 1962 A new system for estimating
skeletal maturity from the hand and wrist with
standards derived from a study of 2,600 healthy
British children. 11. The Scoring System. Pans,
International Children’s Centre.
Документ
Категория
Без категории
Просмотров
0
Размер файла
366 Кб
Теги
skeletal, research, comparability, maturation
1/--страниц
Пожаловаться на содержимое документа