Keywords
arthroplasty, replacement, hip - hip prosthesis - radiography - pelvis - radiographic
magnification
Introduction
Preoperative surgical planning is fundamental to predict difficulties during the procedure,
sizes of components, and positioning of implants in hip arthroplasty.[1]
[2]
[3]
[4]
[5]
[6] Adequate surgical planning reduces the number of complications associated with discrepancy
between members, poor positioning, early implant loosening, instability, periprosthetic
fracture, and loss of bone mass.[1]
[2]
[3]
[4]
[5]
[6] It was traditionally performed with printed radiographs and use of transparencies
with magnification established by the prosthesis' manufacturer, with usual variations
in the range of 100 to 130% of the real size.[5] However, with the development of digital radiographic imaging, computerized planning
has become a practical alternative with proven reproducibility.[7]
To correct the magnification of digital radiography, radiopaque markers such as spheres,[8]
[9]
[10] disks (coins),[11] and plates[12]
[13] have been used in the past 10 years.[9]
[10]
[14]
[15]
[16]
[17] The nearer the position in relation to the joint studied, the smaller will be the
variation in the marker's magnification,[17] decreasing the errors in calibration. For hip replacement, the markers are commonly
positioned in the region around the joint, such as in the greater trochanter[9]
[11]
[12]
[14]
[15]
[17] and between the thighs.[9]
[10]
[12]
[13]
[14]
[16]
[17]
Calibration is an essential first step of the digital planning, performed by the examiner
to identify a known distance between two points or by the diameter of a circle circumscribed
around three peripheral points of the reference object. The gold standard calibration
is calculated from the head of the prosthesis implanted in postoperative patients.
There is no standardization either of the best position or the best type of marker
for hip arthroplasty planning.
The objective of the present paper is to evaluate the precision and differences between
two types of metal markers—sphere, and coin—for radiographic calibration in preoperative
planning of hip arthroplasty. Our hypothesis was that the sphere positioned alongside
the trochanter would be the best method for calibration due to the anatomical proximity
to the hip joint and the facility of calibration, but the coin marker would have similar
performance when keeping its largest diameter always visible in the radiograph.
Materials and Methods
This prospective transversal study was carried out by the Hip Surgery Group of the
Department of Orthopedics and Traumatology of our institution.
The study was conducted with a convenience sample of 33 patients of both genders,
with total and/or partial hip prosthesis, either primary or revision, who underwent
pelvic radiography as part of the postoperative follow-up in April and May 2017. The
radiographs in which the diameter of the head of the prosthesis was unknown from hospital
records or that were outside of the radiographic standards were excluded from the
study.[18]
Procedures
All the patients were submitted to pelvic radiography in anteroposterior incidence
(AP), performed with the patient in dorsal decubitus, with internal rotation of the
hips between 15 and 20° and incident radius at the midline, just above the pubic symphysis.[18] For standardization, the focus film distance was 100 cm, determined with a tape
measure and Optimus 50 light indicator (Philips Medical Systems, Hamburg, Germany).
To standardize analysis of the radiographs, the coccyx was aligned with the pubic
symphysis (with distance between them of 2.5 cm for women and 1.5 cm for men), as
well as symmetry of the obturator foramens.[18]
Before the radiography, a single physician positioned four spheres or four coins in
four positions around the pelvis. The diameter of both markers was 25 mm, confirmed
by a Mitutoyo p-06 analog pachymeter (Mitutoyo Sul Americana, Suzano, SP, Brazil)
and a WesternPRO (Pasadena, CA, EUA) carbon fiber composite caliper (USA), both certified
by Brazil's National Institute of Metrology (INMETRO, in the Portuguese acronym).
The markers were held in place with adhesive tape or a clear polyvinyl chloride (PVC)
tube with length of ¾ inch, as described in a previous study,[19] and a flexible PVC electrical conduit of the same length. The hose and conduit were
sliced lengthwise to allow controlled slippage of the sphere to adjust the position.
The markers were positioned at the following sites: 1–Right greater trochanter, on
the skin, with the sphere in the electrical conduit and the coin attached with tape
at the same level ([Figs. 1A] and [1B]); 2 - Between the thighs, in the plane of the greater trochanter, with the sphere
in the hose placed as proximate as possible and the coin at the same level ([Figs. 2A] and [2B]); 3–At the level of the anterosuperior border of the pubic symphysis, both attached
with adhesive tape ([Fig. 3A]); 4–Between the exam table and proximal segment of the left thigh, 4 cm distal to
the greater trochanter, both attached with adhesive tape ([Fig. 3B]).
Fig. 1 Positioning of the sphere at the right greater trochanter (1A); Positioning of the
coin at the right greater trochanter (1B).
Fig. 2 Positioning of the sphere between the thighs, in the plane of the greater trochanter,
side view (2A) and top view (2B).
Fig. 3 Positioning of the sphere and coin in the pubic symphysis (3A); Positioning of the
sphere and coin on the exam table, on the left side of the patient (3B).
Analysis of the Images
We only considered radiographs that enabled digital measurement of the head of the
prosthetic component and at least two spherical markers and coins ([Figs. 4A], [4B], [4C] and [4D]) in the same image. Markers not visualized in the radiograph were not analyzed.
Fig. 4 Four radiographs of the sample analyzed. [Figure 4A], note the absence of markers at the greater trochanter (broken line circle). In
the male patients ([Figs. 4A] and [4B]), the markers between the thighs are at the lower limit of the image, but the abdominal
apron due to the large abdominal hernia in [Figure 4B] prevents the correct positioning of the markers, and the coin is hard to visualize
between the thighs (white arrow). The overlap of the coin with the femoral shaft in
[Figure 4C] (circle) does not prevent calibration but hampers the technique. The partial visualization
of the coin when placed vertically at the greater trochanter or between the thighs,
as depicted in [Figure 4D] (broken-line arrow), prevents measurement of the larger diameter, making calibration
impossible.
The digital images were analyzed by two doctors who had received previous training
for use of the AGFA HealthCare – IMPAX Orthopaedic Tools (AGFA Healthcare, Greenville,
SC, EUA) planning software. The prosthesis head was used as a calibration parameter
based on three marginal points and confirmed with a circle,[8]
[9]
[10]
[17] since the diameter is known and the location is centered in the joint.
The diameters of the markers after radiographic calibration were measured by the same
technique of identifying three peripheral points and forming a circle to confirm the
marker's diameter, both for the sphere[9] and the coin.[11] The diameter of the sphere was evaluated considering the whole circle filled, while
for the coin, the largest diameter of the image was considered.
Each evaluator measured all the images of the coins and spheres visualized in each
position for subsequent comparison between the measures. All the analyses were performed
with a single HP Pavilion DV7 computer (The Hewlett-Packard Company, Palo Alto, CA,
USA).
Statistical Analysis
The anthropometric characteristics of the sample were described by calculating the
means and standard deviations as well as the absolute and relative frequencies.
The one-sample Student t-test was used to compare the mean at each position for each
of the examiners' evaluation with the reference value of 25 mm. The paired Student
t-test was used, and the mean, standard deviation, coefficient of variation, minimum
and maximum of each of the examiners' evaluation were calculated at each position
to judge the similarity between the markers.
The descriptive statistics in percentage of the real value (25 mm) were evaluated
by the mean, median, standard deviation, minimum, maximum, and confidence interval.
The accuracy of each marker at each position was also calculated. The equality of
two proportions test was applied to characterize the relative frequency distribution
of the qualitative variables. The results were compared by analysis of variance (ANOVA)
and the Pearson and chi-squared tests.
All the analyses were performed with the programs IBM-SPSS for Windows version 20.0
(IBM Corp., Armonk, NY, USA), Minitab 16 (Minitab LLC, State College, PA, USA), and
Microsoft Excel 2010 (Microsoft Corp., Redmond, WA, USA), with a 5% significance level,
following the literature of reference.[20]
[21]
[22]
Results
The mean values of both markers at the greater trochanter and between the thighs were
near the real value of 25 mm. There was no difference between the mean values at the
greater trochanter position and the actual value, but in the position with the marker
between the thighs, only the coin obtained a statistical difference and was considered
more precise. [Table 1] reports the evaluations of the two examiners in comparison with the real size of
25 mm.
Table 1
|
Examiner 1
|
Mean (mm)
|
SD (mm)
|
CV
|
Min (mm)
|
Max (mm)
|
N
|
CI
|
P
-value
|
|
Coin
|
Greater trochanter
|
24.71
|
0.61
|
2%
|
24
|
26
|
14
|
0.32
|
0.104
|
|
Between the thighs
|
25
|
0.65
|
3%
|
24
|
26
|
20
|
0.28
|
1
|
|
Pubic symphysis
|
27.7
|
0.47
|
2%
|
27
|
28
|
33
|
0.16
|
< 0.001
|
|
Exam table
|
22.91
|
0.29
|
1%
|
22
|
23
|
33
|
0.1
|
< 0.001
|
|
Sphere
|
Greater trochanter
|
25.07
|
0.27
|
1%
|
25
|
26
|
14
|
0.14
|
0.336
|
|
Between the thighs
|
25.58
|
0.5
|
2%
|
25
|
26
|
33
|
0.17
|
< 0.001
|
|
Pubic symphysis
|
27.88
|
0.48
|
2%
|
27
|
29
|
33
|
0.17
|
< 0.001
|
|
Exam table
|
23
|
0.25
|
1%
|
22
|
24
|
33
|
0.09
|
< 0.001
|
|
Examiner 2
|
Mean (mm)
|
SD (mm)
|
CV
|
Min (mm)
|
Max (mm)
|
N
|
CI
|
P
-value
|
|
Coin
|
Greater trochanter
|
24.79
|
0.58
|
2%
|
24
|
26
|
14
|
0.3
|
0.189
|
|
Between the thighs
|
25
|
0.73
|
3%
|
24
|
26
|
20
|
0.32
|
1
|
|
Pubic symphysis
|
27.7
|
0.59
|
2%
|
26
|
29
|
33
|
0.2
|
< 0.001
|
|
Exam table
|
23.15
|
0.44
|
2%
|
22
|
24
|
33
|
0.15
|
< 0.001
|
|
Sphere
|
Greater trochanter
|
24.79
|
0.58
|
2%
|
24
|
26
|
14
|
0.3
|
0.189
|
|
Between the thighs
|
25.27
|
0.57
|
2%
|
24
|
26
|
33
|
0.2
|
0.01
|
|
Pubic symphysis
|
27.97
|
0.53
|
2%
|
27
|
29
|
33
|
0.18
|
< 0.001
|
|
Exam table
|
22.97
|
0.3
|
1%
|
22
|
24
|
33
|
0.1
|
< 0.001
|
Comparison of the findings of the examiners for each marker at the different positions,
as shown in [Table 2], reveals similarity of the following measurements: coin at the greater trochanter,
between thighs, and pubic symphysis positions; and sphere at the pubic symphysis and
exam table positions.
Table 2
|
Coin
|
Mean (mm)
|
SD (mm)
|
CV (mm)
|
Min (mm)
|
Max (mm)
|
N
|
CI
|
P
-value
|
|
Greater trochanter
|
Examiner 1
|
24.71
|
0.61
|
2%
|
24
|
26
|
14
|
0.32
|
0.336
|
|
Examiner 2
|
24.79
|
0.58
|
2%
|
24
|
26
|
14
|
0.3
|
|
|
Between the thighs
|
Examiner 1
|
25
|
0.65
|
3%
|
24
|
26
|
20
|
0.28
|
1
|
|
Examiner 2
|
25
|
0.73
|
3%
|
24
|
26
|
20
|
0.32
|
|
|
Pubic symphysis
|
Examiner 1
|
27.7
|
0.47
|
2%
|
27
|
28
|
33
|
0.16
|
1
|
|
Examiner 2
|
27.7
|
0.59
|
2%
|
26
|
29
|
33
|
0.2
|
|
|
Exam table
|
Examiner 1
|
22.91
|
0.29
|
1%
|
22
|
23
|
33
|
0.1
|
0.003
|
|
Examiner 2
|
23.15
|
0.44
|
2%
|
22
|
24
|
33
|
0.15
|
|
|
Sphere
|
Mean (mm)
|
SD (mm)
|
CV (mm)
|
Min (mm)
|
Max (mm)
|
N
|
CI
|
P
-value
|
|
Greater trochanter
|
Examiner 1
|
25.07
|
0.27
|
1%
|
25
|
26
|
14
|
0.14
|
0.04
|
|
Examiner 2
|
24.79
|
0.58
|
2%
|
24
|
26
|
14
|
0.3
|
|
|
Between the thighs
|
Examiner 1
|
25.58
|
0.5
|
2%
|
25
|
26
|
33
|
0.17
|
0.006
|
|
Examiner 2
|
25.27
|
0.57
|
2%
|
24
|
26
|
33
|
0.2
|
|
|
Pubic symphysis
|
Examiner 1
|
27.88
|
0.48
|
2%
|
27
|
29
|
33
|
0.17
|
0.263
|
|
Examiner 2
|
27.97
|
0.53
|
2%
|
27
|
29
|
33
|
0.18
|
|
|
Exam table
|
Examiner 1
|
23
|
0.25
|
1%
|
22
|
24
|
33
|
0.09
|
0.572
|
|
Examiner 2
|
22.97
|
0.3
|
1%
|
22
|
24
|
33
|
0.1
|
|
[Table 3] demonstrates a statistical difference of the measures between the markers (coin
vs. sphere) at all positions except on the exam table (p-value = 0.083).
Table 3
|
Examiner 1
|
Mean (mm)
|
SD (mm)
|
CV (mm)
|
Min (mm)
|
Max (mm)
|
N
|
CI
|
P
-value
|
|
Greater trochanter
|
Coin
|
24.71
|
0.61
|
2%
|
24
|
26
|
14
|
0.32
|
0.019
|
|
Sphere
|
25.07
|
0.27
|
1%
|
25
|
26
|
14
|
0.14
|
|
Between the thighs
|
Coin
|
25
|
0.65
|
3%
|
24
|
26
|
20
|
0.28
|
0.002
|
|
Sphere
|
25.55
|
0.51
|
2%
|
25
|
26
|
20
|
0.22
|
|
Pubic symphysis
|
Coin
|
27.7
|
0.47
|
2%
|
27
|
28
|
33
|
0.16
|
0.032
|
|
Sphere
|
27.88
|
0.48
|
2%
|
27
|
29
|
33
|
0.17
|
|
Exam table
|
Coin
|
22.91
|
0.29
|
1%
|
22
|
23
|
33
|
0.1
|
0.083
|
|
Sphere
|
23
|
0.25
|
1%
|
22
|
24
|
33
|
0.09
|
[Table 4] identifies the percentage values in relation to the true size of 25 mm to allow
understanding the range of variation measured at each position of the markers.
Table 4
|
Examiner 1
|
Mean
|
Median
|
SD
|
Min
|
Max
|
N
|
CI
|
|
Coin
|
Greater trochanter
|
1.14%
|
0.00%
|
2.44%
|
−4.00%
|
4.00%
|
14
|
1.28%
|
|
Between the thighs
|
0.00%
|
0.00%
|
2.60%
|
−4.00%
|
4.00%
|
20
|
1.14%
|
|
Pubic symphysis
|
−10.79%
|
−12.00%
|
1.87%
|
−12.00%
|
−8.00%
|
33
|
0.64%
|
|
Exam table
|
8.36%
|
8.00%
|
1.17%
|
8.00%
|
12.00%
|
33
|
0.40%
|
|
Sphere
|
Greater trochanter
|
−0.29%
|
0.00%
|
1.07%
|
−4.00%
|
0.00%
|
14
|
0.56%
|
|
Between the thighs
|
−2.30%
|
−4.00%
|
2.01%
|
−4.00%
|
0.00%
|
33
|
0.68%
|
|
Pubic symphysis
|
−11.52%
|
−12.00%
|
1.94%
|
−16.00%
|
−8.00%
|
33
|
0.66%
|
|
Exam table
|
8.00%
|
8.00%
|
1.00%
|
4.00%
|
12.00%
|
33
|
0.34%
|
The accuracy calculated for each marker at the various positions by each examiner
is shown in [Table 5]. With the coin on the exam table, utilizing a diameter of 23 mm as a reference,
examiner 1 obtained accuracy of 90.9% (30 of the 33 radiographs) and examiner 2 achieved
accuracy of 78.8% (27 of the 33 radiographs).
Table 5
|
Accuracy (25 mm)
|
Examiner 1
|
Examiner 2
|
Total
|
|
N
|
Accuracy
|
N
|
Accuracy
|
|
Coin
|
Greater trochanter
|
8
|
57.10%
|
9
|
64.30%
|
14
|
|
Between the thighs
|
12
|
60.00%
|
10
|
50.00%
|
20
|
|
Pubic symphysis
|
0
|
0.00%
|
0
|
0.00%
|
33
|
|
Exam table
|
0
|
0.00%
|
0
|
0.00%
|
33
|
|
Sphere
|
Greater trochanter
|
13
|
92.90%
|
9
|
64.30%
|
14
|
|
Between the thighs
|
14
|
42.40%
|
20
|
60.60%
|
33
|
|
Pubic symphysis
|
0
|
0.00%
|
0
|
0.00%
|
33
|
|
Exam table
|
0
|
0.00%
|
0
|
0.00%
|
33
|
There was greater loss of images of both markers at the greater trochanter position,
corresponding to 19 losses in 33 radiographs (57.6%). In that position, for the female
sample, there were 11 losses in 18 radiographs (61.1%), while the figure for the males
was 8 losses in 15 radiographs (53.3%). For the coin placed between the thighs, there
were 13 losses in 33 radiographs (39.4%), with the breakdown by gender being 8 losses
in 15 radiographs (53.3%) for men and 5 losses in 18 radiographs (27.8%) for women.
Discussion
This is a transversal study with the objective of assessing the accuracy, precision,
and differences in the use of two metal markers (sphere and coin) to calibrate radiography
of the pelvis. Considering accuracy, precision, coefficient of variation, visual inspection
of the images, and ease of positioning, the coin placed between the thighs was the
best marker, with a mean of 25 mm, but suffered from large loss of images (39.4%).
The coin on the exam table presented a mean of 23 mm, smaller than the real size,
but obtained a low coefficient of variation (1%) and appeared in all the images. Therefore,
we suggest the use of coins in the two positions.
The best type of radiographic marker should be widely available, practical to position,
reliably visible in the images, and easy to measure and precisely calibrate. The coin's
dimensions are standardized, and it is available to any radiology service that does
not have another type of marker. The identification of markers when positioned at
the greater trochanter and pubic symphysis is susceptible to human error, unlike the
exam table position, which is easy to achieve and free of this type of error.
In the pubic symphysis and exam table positions, all the images were easily visualized.
Despite obese patients requiring special care when positioning the markers, which
should be placed below the abdominal fold in the pubic symphysis position and below
the thigh when on the exam table, no image losses occurred with these markers. The
images of the coin between the thighs in 13 radiographs (39.4%) and of both markers
at the greater trochanter in 19 radiographs (57.6%) were only partially visible or
not visible, so no calculations were possible. There was a difference between the
genders, with greater loss of images in female patients at the greater trochanter
(61.1%), probably due to the pattern of posterolateral fat accumulation in the hips,
while the loss was greater for male subjects when the marker was positioned between
the thighs (53.3%), probably caused by the genital volume in this region. The circular
images of the sphere and coin on the exam table and at the pubic symphysis were easier
to measure. Considering the coefficient of variation lower than 3% for both markers
at all positions, they can be considered stable in relation to the mean.
With respect to precision, the measures of coin between the thighs (p = 1.000), sphere at the greater trochanter (p = 0.336), and coin at the greater trochanter (p = 0.189) were the closest to the actual size of 25 mm. At the greater trochanter,
the accuracies of the coin and sphere were, respectively, between 57.1 and 63.3% and
between 64.3 and 92.9%. Between the thighs, the accuracy measures of the coin and
sphere were between 50 and 60% and 42.4 and 60%, respectively. Even though not precise
at 25 mm, the result for the coin and sphere on the exam table was reliable, with
the best confidence interval (between 0.09 and 0.1), denoting low variation of the
mean, and the accuracy for 23 mm was between 78.8% and 93.9%.
Markers, when located at the level of the pubic symphysis and on the exam table, have
been found to be subject to variation due to the radiographic magnification of the
emission cone of objects farther from the film in comparison to those that were closer.[12]
[23] For calibration with these markers, it is necessary to understand this influence
so that errors do not occur outside the margin between +3% and -3%, which is acceptable
for adequate preoperative planning.[23]
Wimsey et al.[8] used a caliper to compare the positioning of a coin between the thighs and the measurement
of the anterior superior iliac spines in the patient. The results of the magnification
calculation were favorable to the use of the coin, with accuracy of 98.9% and error
< 0.5 mm in the digital measurements of the marker. When positioned between the thighs,
the average coin size in our sample was 25 mm for both examiners, the coefficient
of variation was at most 3%, and the accuracy was between 50 and 60%. The maximum
error was 1 mm (4%), probably overestimated by the fact that the software used for
measurement did not have millimetric precision.
In the present study, the markers were positioned by a physician with experience in
identification of bony prominences, unlike in the great majority of previous studies.[8]
[9]
[10]
[11]
[12]
[13]
[14]
[15]
[16]
[17]
[23] The simplest position was on the exam table, since the identification of bony prominences
can be difficult, especially in patients who are obese or have deformities. The et
al.[23] reported the importance of correct positioning, demonstrating that the marker placed
at the level of the trochanter at 1 cm anterior or posterior to the hip represents
a 1% difference in magnification.
Some limitations of the present study should be mentioned. The sample loss at the
greater trochanter probably occurred due to the difficulty of adapting this method
in the fat pads lateral and posterior in relation to the hip, especially in female
patients. Alterations in the surface anatomy of patients who have undergone multiple
surgeries was another factor that might have influenced the precision of the positioning
of the markers. Bias in the calibration precision and calculation of the diameter
of the markers might also have occurred because the software used is not capable of
measuring decimal values of millimeters.
Conclusion
We recommend using the coin placed between the thighs and suggest the concomitant
use of another coin, or a sphere, on the exam table, considering the 8% reduction
in relation to its real size, and to avoid failure to visualize the marker in the
image.
The methods presented in the current study are useful for radiographic calibration
in preoperative planning and can estimate the radiographic magnification with a safety
margin between -3% and +3%.[23] We believe the coin can be used instead of a sphere without sacrificing precision
when applied clinically for radiographic calibration in preoperative planning. It
is necessary to understand the size variations of the markers around the hip before
performing the radiographic calibration for accurate correction of the magnification
and to achieve better precision in preoperative planning. In those patients who are
morbidly obese, have an abdominal apron, hip deformity, or previous surgery, attention
should be paid to the correct palpation of the bone structures to minimize marker
positioning errors.
