Ninety-three cadaveric proximal humeri and matching scapulae with no e
vidence of shoulder disease were obtained. High-resolution roentgenogr
ams were made. They were then digitized, and a custom computer program
was used to obtain periosteal and endosteal dimensions (humeral canal
width, shaft width, tuberosity offset, head offset, radius of curvatu
re, head diameter, canal flare index, glenoid height and depth, are of
enclosure, radius of curvature, and depth of cancellous bone, among o
thers). Statistical analyses included correlations among pairs of dime
nsions, regression analysis, and confidence intervals to estimate each
geometric parameter. To represent the degree of conformity and constr
aint between the humeral head and glenoid, a conformity index (radius
of head/radius glenoid) and constraint index (arc of enclosure/360) we
re calculated for each shoulder. Most of the parameters measured appro
ximated a Gaussian distribution. Pairwise correlations of the geometri
c parameters showed that many were significantly associated at the 5%
level (p < 0.05) with a high correlation coefficient (r > 0.4), meanin
g that the variations between the parameters were not purely random bu
t rather they were related and that this relationship was useful in a
predictive sense. With these key parameters identified, one can design
components with an optimum fit or match an existing design to a patie
nt's anatomy with a high degree of accuracy. No correlation was found
between the radii of curvature for the humeral head and glenoid in eit
her the coronal or sagittal plane or between the humeral head diameter
and arc of enclosure. The mean conformity index was 0.72 in the coron
al and 0.63 in the sagittal plane. Only 16% of the specimens had a con
formity index greater than 0.9, indicating the radii of curvature matc
hed. The vast majority had a more curved humeral head and flatter glen
oid. Also, there was more constraint to the glenoid in the coronal ver
sus sagittal plane (constraint index = 0.18 vs 0.13). These anatomic f
eatures help prevent superior-inferior translation of the humeral head
but allow translation in the sagittal plane. This database of basic a
natomic geometry defines the anatomic relationships of the proximal hu
merus and glenoid cavity that allow for a precise bone-implant fit and
assesses the match between the shape of existing components and the p
atient's anatomy.