Abstract
Background: Recreation of normal anatomical
relationships may be important to optimize the outcome of proximal
humeral arthroplasty. With use of computerized tomographic data
and three-dimensional computer modeling, we concurrently studied
both extramedullary and intramedullary humeral morphology, including
canal shape, and related these findings to the design of proximal
humeral prostheses.
Methods: Sixty cadaveric humeri (thirty pairs:
fifteen from male donors and fifteen from female donors) were studied.
Three-dimensional computer models were built from canal and periosteal
contours extracted from computerized tomographic data and multiple
measured anatomical parameters, including humeral canal axis, humeral
head center, and hinge point offset; greater tuberosity and bicipital groove
offset; humeral head center, radius, thickness, retroversion, and
inclination; and size and torsion of sections of the canal.
Results: On the average, the humeral head center
was offset both medially (seven millimeters) and posteriorly (two
millimeters) from the humeral axis. The humeral head hinge point
did not line up with the axis but instead was laterally offset by
an average of seven millimeters. The average humeral head thickness
was nineteen millimeters. The humeral head thickness and length
were proportionately linked. There was marked variability in all
of these parameters. Humeral head inclination averaged 41 degrees
but was less variable than previously described, with 95 percent
of our sample within the range of 35 to 46 degrees. The proximal
section of the humeral canal was retroverted, and the retroversion
was found to be similar to that of the humeral head on statistical
analysis. Version of the middle and distal sections of the canal,
however, was dissimilar to that of the proximal section of the canal.
Proximal humeral retroversion was found to be extremely variable
and averaged 19 degrees. The accuracy, reliability, and repeatability
of the computer-based-model measurements were found to be excellent.
Conclusions: Measurements of external proximal
humeral morphology made with three-dimensional computer models of
cadaveric specimens derived from the Midwestern United States agreed,
in general, with those described for different populations evaluated with
different measuring techniques. Proximal humeral morphology was
extremely variable as highlighted by the large ranges of measurements
seen for all variables. Examination of the intramedullary morphology
showed that there is an internal version, with measurements dependent
on the canal distance distal to the anatomical neck.
Clinical Relevance: Because of the marked variabilities
seen in proximal humeral morphology, newer prosthetic designs are
now allowing surgeons to control multiple prosthetic variables.
An understanding of the normal values for proximal humeral morphology
can serve as an important guideline for component selection, especially
when the normal anatomy is distorted. Additionally, variations in intramedullary
version may have important consequences for future designs of press-fit
proximal humeral replacement.
Recently introduced designs for prosthetic replacement of
the proximal part of the humerus have emphasized the importance
of accurate recreation of the normal three-dimensional anatomy1-5,7-12,17. However, to date, very
few studies have detailed the external three-dimensional anatomy
of the proximal part of the humerus. To our knowledge, no investigators
have directly measured intramedullary proximal humeral morphology
or correlated it to extramedullary morphology. Authors of available
studies have employed either reamers to create an internal (canal)
axis1,9,10 or noncontinuous surface
measurements from which to interpolate and create a three-dimensional
model3,18.
The purpose of the present study was to employ modern techniques
to describe concurrently both extramedullary and intramedullary
humeral morphology, including canal shape, and to relate these findings
to the design of proximal humeral prostheses. Computerized tomographic
images and computer modeling were used to nondestructively create
accurate three-dimensional models from which extramedullary and
intramedullary morphology could be analyzed.
Three-Dimensional Modeling
Sixty humeri from thirty unembalmed, fresh cadavera were obtained
from donors in the Midwestern United States. Humeral pairs were
excluded from the study if the donor had any apparent osseous abnormalities,
including moderate or severe arthritis, or if osseous abnormalities
were detected radiographically in either humerus of the pair. None
of the donors had had any surgical procedure performed on the humeri
or scapulae. There were fifteen women and fifteen men, ranging in
age from twenty to eighty-two years (mean [and standard deviation],
54 ± 18 years) at the time of death. Eight
pairs were from donors between the ages of twenty and forty years;
eight, from donors between forty-one and sixty; twelve, from donors
between sixty-one and eighty; and two, from donors older than eighty years.
Donor height ranged from 149 to 194 centimeters (mean, 171.2 ± 12 centimeters). Three donors were black and twenty-seven
were white; none were Hispanic.
Anteroposterior and lateral radiographs were made of the humeri.
Four humeri at a time were placed into a specially designed holder,
and computerized tomographic images were made with use of a Siemens
Somatom Plus S scanner (Siemens Medical Systems, Erlangen, Germany).
The axial slice thickness was two millimeters at two-millimeter
intervals in the proximal and distal thirds of the humerus and at
five-millimeter intervals in the mid-third (mid-part of the shaft)
of the humerus. The field of view was approximately ten by ten centimeters,
and all computerized tomographic images were reconstructed with
use of a bone algorithm. The entire humerus was included to enable
model alignment in space.
The computerized tomographic data were then transferred to a
computer-aided-design workstation. Initial work was performed with
use of a UNIX-based SurgiCAD workstation running EMS (Event Management
Software) (Intergraph, Huntsville, Alabama). Later work was performed with
use of an Intel-Windows NT-based workstation running Materialise
(Ann Arbor, Michigan) and UG (Unigraphics Solutions, Maryland Heights,
Missouri) software. All image manipulation, contour extraction,
model building, model alignment, and analyses were performed with
use of computer-aided-design workstations. Inner (canal) and outer
(periosteal) contours of the humeri were extracted with use of the
SurgiCAD or Materialise software. This software was used to build the
solid models, and EMS or UG software was used to align and analyze
the models (Fig. 1).
Measurements of Proximal Humeral Morphology
Prior to morphological analysis, each solid humeral model was
aligned in space. The long axis of each humerus was rigidly rotated
to a vertical orientation in the coronal and sagittal planes, and
the head center was rotated into the coronal plane. The models were
rotated about the canal axis, creating anterior and medial views.
The humeri were also rotated 90 degrees about the axial plane to
create superior views. These different views were used for the analysis
described below.
The following parameters were calculated for each humerus: (1)
humeral canal axis, calculated with use of least-squares fit16 to describe a line through the canal
distal to the surgical neck and continuing to the mid-part of the humeral
shaft; (2) humeral head center, calculated with use of the least-squares
method to fit a sphere to the three-dimensional head data; and (3)
humeral head radius, calculated in the anterior, superior, and medial
perspectives by least-squares-method fitting of a circle to the
articular surface, in each view (Fig. 2).
Humeral head inclination was defined by the angle between the
canal axis and the least-squares-fit line describing the articular
margin (anatomical neck or inclination line). The neck-shaft angle could
be calculated by adding 90 degrees to the inclination angle. It
was defined by the angle between the canal axis and a line (neck-shaft
line) passing through the head center and perpendicular to the articular
inclination line. Humeral head thickness was defined by the distance
along the neck-shaft line between the medial and lateral edges of
the articular surface (Fig. 2).
Humeral head center offsets from the canal axis were calculated
in both the anterior and the superior perspective. In the anterior
perspective, along the canal axis, the horizontal distance was measured
between the intersection of the canal axis with the superior articular
surface and the intersection of the inclination line with the superior
articular surface. This distance has been called the hinge point3. In the superior perspective, the
medial-lateral and anterior-posterior distances were measured as shown
in Figure 3.
The position of the greater tuberosity was also measured (Fig. 2). The medial
edge of the greater tuberosity was defined by a vertical line (paralleling
the canal axis) starting at the point of inflection between the superolateral
edge of the articular surface and the superomedial edge of the greater
tuberosity.
Additional parameters measured in the superior perspective included
humeral head retroversion, defined as the angle between the horizontal
and a line connecting the most anterior points of the trochlea and
capitellum, and the medial-lateral and anterior-posterior distances
from the head center to the trough of the bicipital groove (Fig. 3).
Three axial canal sections were selected from each right-sided
humeral model (thirty humeri). These sections were equally spaced
from the surgical neck to the mid-part of the shaft. An ellipse
was fitted (with the least-squares method) to the axial canal sections,
and the major and minor axis lengths and major axis version angle
were measured (Fig. 4).
Statistical Analysis
Two-tailed t tests were used to determine differences between
right and left humeri and between humeri from male and female donors
with regard to head retroversion, head inclination, head radius, head
thickness, medial and posterior head center offsets, hinge point
offset, distance between the greater tuberosity and the articular
surface, and lateral and anterior biceps offset. We also tested
for differences between male and female donors with regard to humeral
length and lengths of the major and minor axes of the ellipses.
The same methods were used to test for differences between head
retroversion and the three canal axis angles described above. This
comparison demonstrated the extent of the retroversion deformity
along the proximal part of the shaft. Linear regression analyses
were performed for each of the following morphological parameters:
humeral length, head radius, head thickness, head inclination, head
retroversion, medial and posterior head center offsets, and greater tuberosity
width. All statistical testing was performed with use of JMP software
(SAS Institute, Cary, North Carolina), with the level of significance
set at p £ 0.05.
Measurement of Accuracy, Reliability, and Repeatability
To test accuracy, reliability, and repeatability, original and
repeat measurements of head radii in the anterior perspective were
made for twenty randomly selected right humeri. Measurements were made
of the actual bones with use of precision calipers and of the computer-assisted-design
models (soft-copy measurements) with use of Materialise software.
Accuracy was defined as the difference between caliper and soft-copy
measurements. Reliability (precision) was defined as the average
difference between repeated measurements of the same humeri. Repeatability
was the measure of reliability relative to the variation among specimens. Tests
were performed with JMP statistical software.
Humeral Length, Head Radius, and Head Thickness
The means, standard deviations, and ranges for these parameters,
as measured in the anterior perspective, are listed in Table I. The frequency distribution of head thickness measurements
is illustrated in Figure 5. Humeri from male donors were significantly longer
(mean [and standard deviation], 35 ± 2 centimeters)
than humeri from female donors (mean, 31 ± 2
centimeters), according to the two-tailed unpaired t test. Only
right humeri were measured for length. Humeri from male donors also
had a significantly larger head radius (mean, 24 ± 2
millimeters compared with 22 ± 2 millimeters
for female donors) and head thickness (mean, 20 ± 2
millimeters compared with 18 ± 2 millimeters
for female donors), according to the two-tailed unpaired t test.
Correlation coefficients from linear regression analysis of these
parameters are listed in Table II. Humeral length was correlated
with head radius (r = 0.72) and head thickness (r = 0.76). Head
radius was correlated with head thickness (r = 0.71).
Humeral Head Retroversion and Inclination
The means, standard deviations, and ranges for humeral head retroversion
and inclination are listed in Table I. The frequency distributions of
the head retroversion and inclination measurements are illustrated in Figure 6 and Figure 7. There was
no significant difference between male and female donors for these
two parameters, according to the two-tailed unpaired t test. However, there
was a significant difference in head inclination between the right
and left sides.
Humeral Head Center and Hinge Point Offsets
The head center was found to be medially and posteriorly offset
(mean, seven and two millimeters, respectively) and the hinge point
was found to be laterally offset (mean, seven millimeters) from
the canal axis (Table I). There was a significant difference
between male and female donors for medial head offset. Otherwise,
there was no significant difference between male and female donors
or between right and left humeri for these offset measurements,
according to the two-tailed t test. Medial head center offset was
correlated with humeral length (r = 0.55).
Greater Tuberosity and Bicipital Groove Parameters
The medial edge of the greater tuberosity was a mean of 10 ± 2 millimeters (range, five to fifteen millimeters) lateral
to the canal axis and 16 ± 2 millimeters
(range, thirteen to twenty-one millimeters) lateral to the head
center. The distance from the medial edge of the greater tuberosity
to the medial articular surface averaged 39 ± 5
millimeters (range, thirty to forty-nine millimeters). The mean
vertical distance from the superior aspect of the greater tuberosity
to the superior aspect of the humeral head was 6 ± 2
millimeters (range, three to eight millimeters). There was no significant
difference between right and left humeri for these measurements,
according to the two-tailed paired t test. The only significant male-female
difference, according to the two-tailed unpaired t test, was for
the distance from the medial edge of the greater tuberosity to the
medial articular surface, which averaged 41 ± 3
millimeters for the men and 37 ± 3 millimeters
for the women (Table I).
The medial-lateral width of the greater tuberosity, as measured
in the anterior perspective, averaged 9 ± 2
millimeters (range, six to twelve millimeters). This distance was
correlated with humeral length (r = 0.59) and with medial head center
offset (r = 0.53) (Table II).
The lateral offset of the bicipital groove from the head center
averaged 16 ± 3 millimeters (range, ten to
twenty-three millimeters), whereas the anterior offset measured
15 ± 3 millimeters (range, nine to twenty-one
millimeters). The lateral offset was significantly larger in the
men (18 ± 3 millimeters compared with 15 ± 3 millimeters in the women) and the right humeri (18 ± 3 millimeters compared with 15 ± 3
millimeters in the left humeri), according to the two-tailed t test.
There was no male-female difference between the anterior offsets,
according to the two-tailed unpaired t test. The right anterior
offset was significantly larger than the left (16 ± 3
millimeters compared with 14 ± 3 millimeters),
according to the two-tailed paired t test.
Canal Shape, Size, and Version
The means, standard deviations, and ranges of the major and minor
axis lengths and major axis angles (version) in the canal ellipses
are listed in Table III. The angle values were negative,
indicating retroversion. The major and minor axis lengths in the most
proximal section were significantly larger in the men than in the
women, according to the two-tailed unpaired t test. While the men
had slightly larger axis lengths at the other two section levels, no
significant male-female difference was detected with the two-tailed
unpaired t test. No right-left differences were noted with the two-tailed
paired t test.
The two-tailed paired t test showed significant similarity between
the humeral head retroversion and the major axis version of the
most proximal canal section (p = 0.039). In other words, humeral head
retroversion matched the retroversion of the ellipses fit to the
proximal part of the canal. Humeral head retroversion was significantly
different from the version of the two more distal sections, according
to the two-tailed paired t test. The two-tailed paired t test also
showed that the angle measurements at the three section levels were
significantly different from each other. In addition, no significant
male-female differences were found in the analyses (two-tailed unpaired
t test) of the angle measurements in the three sections (Table III).
Measurement of Accuracy, Reliability, and Repeatability
Both caliper (actual bone) and soft-copy (computer-aided-design
model) measurements were highly reliable, with the means (and standard deviations)
of the original and repeat caliper measurements being 20.5 ± 1.7 and 20.4 ± 1.5 millimeters,
respectively, and the means of the original and repeat soft-copy
measurements being 20.1 ± 1.7 and 20.0 ± 2.1 millimeters, respectively. The repeatability
of both the caliper and the soft-copy measurements was very high,
with the variation between the original and repeat measurements
representing less than 0.1 percent of measurement variation among humeri.
Soft-copy measurements were slightly biased, underestimating caliper
measurements by 0.4 millimeter (standard deviation of the difference,
0.7 millimeter). This represents 2.6 percent in total measurement
variation.
If recreation of normal anatomy is the goal of prosthetic replacement
of the proximal part of the humerus, it is important to have a three-dimensional understanding
of normal extramedullary and intramedullary humeral morphology.
This knowledge can affect prosthetic sizing, positioning, and design.
The extramedullary position of the prosthesis is important for joint
kinematics, while intramedullary anatomy can influence prosthetic fixation
and articular surface position. This link between the two anatomical
considerations necessitates concurrent understanding of both extramedullary
and intramedullary morphology, in order to better approximate normal
anatomy with proximal humeral arthroplasty.
Using radiographs, surface scans, or direct measurements, previous
investigators have primarily studied the extramedullary morphology
of the proximal part of the humerus1,3,6,9,10,12.
Ballmer et al.1 were, to our knowledge,
the first to relate external morphology to an internal axis dependent
on native intramedullary morphology. Pearl and Volk9,10 extended this work, measuring
retroversion, lateral canal offset, and greater tuberosity width,
all relative to the center of a surgically reamed canal. Approximations
based on the axis of a reamed canal have a practically relevant
attribute; however, they may be inaccurate if the canal is not cylindrical
and the line-to-line fit of the reamer is not consistent.
Recently, Boileau and Walch3,18 measured
extramedullary morphology and related it to an approximated canal
axis generated from measurements of the periosteal surface. In that study,
three-dimensional reconstructions were created from surface measurements
made at 10-degree increments every five millimeters along the proximal
part of the humerus. Although a large number of surface measurements
were made, this method left gaps of information that required interpolation
during computer reconstruction. For a straight cylinder with uniformly
thick walls, surface-based approximations of the canal axis would be
accurate. However, the humeral shaft is curved and cortical thickness
varies, potentially producing inaccuracies in such surface-based
approximations.
Our use of three-dimensional high-resolution computer models
based on computerized tomographic data had several advantages over
previous study methods. These included (1) analysis of both extramedullary
and intramedullary morphology, including canal shape; (2) nondestructive
creation of a reproducible intramedullary axis; (3) use of volumetric
continuous data, which reduced model interpolation and increased
accuracy; and (4) analysis of cortical morphology. While proximal
humeral arthroplasty was the primary focus of this study, we included
morphology of the greater and lesser tuberosities and the bicipital
groove so that this analysis would be relevant to fracture reconstruction
as well.
The computer-based measurement technique that is described here
was reliably reproducible and accurate. Application of three-dimensional
reconstruction from computerized tomographic images has precedence
in total joint arthroplasty for degenerative or developmentally
dysplastic hips15. Previous applications
of this technology have aided in evaluation preceding total hip
arthroplasty and in development of custom-made hip prostheses13,14,19. This technique can thus be
important not only for preoperative planning when the anatomy may
be severely distorted, but also in the development of prosthetic
designs based on normal anatomy.
Computer-based measurements, when compared with caliper measurements
of the actual specimens, underestimated the humeral head radius
by only 0.4 millimeter. With regard to extramedullary parameters,
our data confirmed previously recorded measurements of proximal
humeral morphology1,3,6,10,12.
We found that the humeral head was not a perfect sphere because
the radius of curvature was smaller when the head was viewed from
the superior perspective than when it was viewed from the anterior perspective.
Our radial dimensions agreed with those in previous literature (Table IV). We also
confirmed that humeral length was correlated with head radius and
head thickness and that head radius was correlated with head thickness3,7,10. This finding implies, as previously
suggested by Pearl and Volk10 and
by Boileau and Walch3, that appropriate
prosthetic sizing should proportionately link head radius and thickness.
Head retroversion was quite variable, with 95 percent of our
sample in the range of 9.5 to 31 degrees, and it was not correlated
with other parameters; thus, a proportional sizing link is not possible.
These results agreed with those of prior studies3,4,12 and
substantiate the importance of individual reconstruction of native
retroversion.
Humeral head inclination was variable, but less so, with 95 percent
of our sample in the range of 35 to 46 degrees. As with retroversion,
head inclination was not correlated with other proximal humeral parameters;
thus, its reconstruction must be tailored to the individual. Third-generation
prosthetic designs can permit individual replication of head inclination3. However, approximating the location
of the true anatomical neck can be difficult secondary to distortion
by osteophytes. An alternative approach would be to produce two
inclination angles for common sizes of prostheses. Given a fixed
inclination angle of 41 degrees and acceptance of an error of 3
degrees when matching patients' native angles with the prosthesis,
65 percent of our sample would fit into this group. With two inclination
sizes of 38 and 44 degrees and allowance of a 3-degree mismatch
with patients' native angles, 95 percent of our sample could be
accommodated. Thus, 95 percent of our sample would have a match
within 3 degrees if two inclination sizes were available.
Our direct measurements of the magnitude of medial and posterior
offset of the head center relative to the canal axis were similar
to the findings of previous investigators who approximated these measurements
(Table IV).
We found that medial and posterior head offsets did not have good
correlation with other measured parameters. The variability of these
two parameters was relatively small, with 95 percent confidence
intervals spanning eight millimeters. However, these two parameters
are important as small shifts in the location of the prosthetic
head center may produce deleterious effects on prosthetic position10.
The offset of the hinge point was consistently lateral, with
a range of four to eleven millimeters. This offset was larger than
the one described by Boileau and Walch3,18.
This difference may have been due to differences in measuring technique:
we measured the canal axis directly, whereas Boileau and Walch approximated
the canal axis from measurements of the periosteal surface. Canal
offsets dictate relative prosthetic position, but the canal morphology
dictates more than medial-lateral and anterior-posterior prosthetic
location. It also may dictate version (retroversion) and prosthetic
stem size. This is especially true for cementless designs. The rotational
orientation (version) and dimensions of the proximal part of the
canal can play a critical role in positioning. To our knowledge,
our study is the first to define this intramedullary morphology.
A surgically relevant finding in our study was that the long
axes of ellipses fit to the proximal half of the humeral canal were
retroverted. The proximal portion of the canal, in general, was
aligned with the native retroversion of the articular surface. The most
proximal portion of the canal was less retroverted (in alignment
with native retroversion) than the middle portion of the canal (not
aligned with native retroversion). These findings have two important
implications for proximal humeral arthroplasty. First, a cementless
stem with proximal canal fit would align itself with the internal
contours of the proximal part of the humerus and be seated in anatomical
retroversion. However, if the humeral neck resection is not in anatomical
retroversion, the prosthetic collar may not sit evenly on the cut
bone surface. Second, as the degree of canal retroversion increases
from proximal to distal, there may be a mismatch in version alignment
for longer-stemmed cementless prostheses that are press-fit distally.
For this reason, the distal portion of a long-stem prosthesis may
need to be cylindrical.
As the proximal canal cross-sections were fit by an ellipse and
not a circle, an elliptical design proximally for cementless prostheses
may improve fixation, torsional and axial stability, and osseous integration.
Alignment of the prosthesis with the ellipse axis of the proximal
part of the canal would align the prosthetic head with the native
retroversion. Our data suggest that an elliptical shape for the
middle or distal portion of the prosthesis may not be desirable.
An elliptical shape for the distal portion of the prosthesis would
be less advantageous because of the high individual variability
of the major axis angles of the distal part of the canal and the
lack of correlation between these angles and the axis angles of
the proximal part of the canal. A cylindrical shape of the distal
part of the prosthesis may provide less line-to-line fit, but it would
not make inappropriate version necessary.
In conclusion, computer-based modeling was an effective tool
for analyzing humeral morphology with specific reference to proximal
humeral arthroplasty. The findings confirmed and expanded previously
recorded extramedullary measurements while providing new information
on intramedullary morphology. This information should be important
not only in the development of new prosthetic designs but also as
an effective guideline for component selection during prosthetic
arthroplasty with modular systems.
Ballmer, F. T.; Sidles, J. A.; Lippitt, S. B.; and Matsen, F. A. III: Humeral prosthetic arthroplasty: surgically relevant considerations. J. Shoulder and Elbow Surg. ,2: 296-304, 1993.2296
1993
Bigliani, L. U.; Kelkar, R.; Flatow, E. L.; Pollock, R. G.; and Mow, V. C.: Glenohumeral stability. Biomechanical properties of passive
and active stabilizers. Clin. Orthop. ,330: 13-30, 1996.33013
1996
[PubMed]
3.Boileau, P. and Walch, G.: : The three-dimensional geometry of the proximal humerus.
Implications for surgical technique and prosthetic design. J. Bone and Joint Surg. ,79-B(5): 857-865, 1997.79-B(5)857
1997
Friedman, R. J.: Biomechanics of the shoulder following
total shoulder replacement. In Surgery of the Shoulder, pp. 263-266.
Edited by M. Post, B. F. Morrey, and R. J. Hawkins. St. Louis, Mosby-Year
Book, 1990
Harryman, D. T.; Sidles, J. A.; Harris, S. L.; Lippitt, S. B.; and Matsen, F. A. III: The effect of articular conformity and the size of the
humeral head component on laxity and motion after glenohumeral arthroplasty.
A study in cadavera. J. Bone and Joint Surg. ,77-A: 555-563, April 1995.77-A555
1995
Iannotti, J. P.; Gabriel, J. P.; Schneck, S. L.; Evans, B. G.; and Misra, S.: The normal glenohumeral relationships. An anatomical study
of one hundred and forty shoulders. J. Bone and Joint Surg.,74-A: 491-500, April 1992.74-A491
1992
Iannotti, J. P., and Williams, G. R. : Total shoulder arthroplasty. Factors influencing prosthetic
design. Orthop. Clin. North America, ,29: 377-391, 1998.29377
1998
Jobe, C. M., and Iannotti, J. P.: : Limits imposed on glenohumeral motion by joint geometry. J. Shoulder and Elbow Surg. ,4: 281-285, 1995.4281
1995
Pearl, M. L., and Volk, A. G.: Retroversion of the proximal humerus in relationship to
prosthetic replacement arthroplasty. J. Shoulder and Elbow Surg. ,4: 286-289, 1995.4286
1995
Pearl, M. L., and Volk, A. G. : Coronal plane geometry of the proximal humerus relevant
to prosthetic arthroplasty. J. Shoulder and Elbow Surg.,5: 320-326, 1996.5320
1996
Rietveld, A. B.; Daanen, H. A.; Rozing, P. M.; and Obermann, W. R. : The lever arm in glenohumeral abduction after hemiarthroplasty. J. Bone and Joint Surg. ,70-B(4): 561-565, 1988.70-B(4)561
1988
Roberts, S. N.; Foley, A. P.; Swallow, H. M.; Wallace, W. A.; and Coughlan, D. P.: The geometry of the humeral head and the design of prostheses. J. Bone and Joint Surg. ,73-B(4): 647-650, 1991.73-B(4)647
1991
Robertson, D. D.; Walker, P. S.; Granholm, J. W.; Nelson, P. C.; Weiss, P. J.; Fishman, E. K.; and Magid, D.: Design of custom hip stem prostheses using three-dimensional
CT modeling. J. Comput. Assist. Tomog.,11: 804-809, 1987.11804
1987
Robertson, D. D.; Walker, P. S.; Fishman, E. K.; Mintzer, C. M.; Poss, R.; Magid, D.; Granholm, J. W.; Brooker, A. F.; and Essinger, J. R.: The application of advanced CT imaging and computer graphics
methods to reconstructive surgery of the hip. Orthopedics,12: 661-667, 1989.12661
1989
[PubMed]
Robertson, D. D.; Essinger, J. R.; Imura, S.; Kuroki, Y.; Sakamaki, T.; Shimizu, T.; and Tanaka, S.: Femoral deformity in adults with developmental hip dysplasia. Clin. Orthop. ,327: 196-206, 1996.327196
1996
[PubMed]
Sokal, R. R. and Rohlf, F. J.: Biometry.
Ed. 3, pp. 460-461. New York, William H. Freeman, 1995
Soslowsky, L. J.; Flatow, E. L.; Bigliani, L. U.; and Mow, V. C.: Articular geometry of the glenohumeral joint. Clin. Orthop. ,285: 181-190, 1992.285181
1992
[PubMed]
Walch, G., and Boileau, P.: Morphological
study of the humeral proximal epiphysis. In Proceedings of the European
Society for Surgery of the Shoulder and the Elbow. J. Bone and Joint
Surg. 74-B (Supplement I): 14, 1992
Walker, P. S., and Robertson, D. D.: : Design and fabrication of cementless hip stems. Clin. Orthop.,235: 25-34, 1988.23525
1988
[PubMed]