CTAn Parameter Description

Morphometric
parameters measured
by Skyscan™ CT-
analyser software.
Page 1 of 49
2 Bruker-MicroCT CT-Analyser: morphometric parameters in 3D and 2D
Introduction
3D morphometric parameters are calculated by CT-analyser either in
direct 3D based on a surface-rendered volume model in 3D space.
Additionally, morphometry in 2D can be done on individual binarised
crossection images.
There are two types of analysis for both 3D and 2D measurements.
Firstly, all objects in the selected region can be analysed together, and the
integrated results calculated, such as total volume or surface of all
objects, or mean thickness of all objects, etc. Secondly, individual
“discreet” objects can be analysed, defined as groups of connected solid
voxels. Morphometric parameters in 3D and 2D can be carried out on all
these individual objects.
Therefore four categories of analysis can be performed, comprising both
integrated analysis of all selected objects (with a single result output for
all objects collectively), and individual analysis of all discreet objects (with
a separate result output for each object); and both of these are calculated
in both 2D and 3D.
All calculations are performed over the selected region/volume of interest.
Consistent and accurate selection of the regions or volumes of interest is
fundamentally important to obtaining accurate and meaningful data. Some
clarification of terminology for this is useful. The term “region of interest”,
or ROI, will refer to a selected region of a single crossectional image. 2D
analysis is performed within a ROI. The “volume of interest” refers to the
collective sum of all ROIs over a contiguous set of crossectional image
slices, representing a selected 3D volume. Morphometric analysis in 3D is
performed in a VOI.
Note however that 2D type analysis can also be performed on a VOI, by
integrating or summing the results of 2D analysis over many slices. This is
sometimes called “slice-by-slice” analysis. It can yield calculations (or
estimates) of 3D parameters such as surfaces and volumes and
thicknesses of objects, but it should be understood that these results are
based on integration of many 2D analyses of separate crossections, and
this is not the same as true direct 3D analysis, which is performed on a
surface-rendered 3D volumetric model. So, within a 3D volume of
interest, both true 3D analysis, and 2D “slice-by-slice” analysis, can be
performed in CT-analyser. (Note further – you do not see this 3D surface
rendered model of objects during 3D analysis – it is created “in the
background” but not displayed. Creation and display of 3D models is also
provided in the CT-analyser and CT-volume programs, but these are
separate procedures not connected to morphometric analysis.)
Parameter names follow two alternative nomenclatures, “General
Scientific” or “Bone ASBMR”, the latter based on Parfitt et al. (1987).
Parfitt’s paper proposed a system of symbols for bone histomorphometry,
and the principles of Parfitt’s system are applied here to both the Bone
(ASBMR) and the General Scientific parameter names. In this document,
both the General Scientific and the Bone ASBMR name will be given at the
Page 2 of 49
title of each parameter. In the contents pages, however, only the General
Scientific names are given, for conciseness.
Within Skyscan CT-analyser four alternative dimensional units are
selectable: mm, µm, inch or pixel. For clarity in this document, all
dimension units are given as mm.
All measurements of morphometric parameters in 3D and 2D are
performed on segmented or binarised images. Segmentation or
“thresholding” must be done prior to morphometric analysis.
A final note – in the CT-analyser preferences (general tab) you can also
specify numerical parameter reporting in scientific notation (e.g. 1.2345E-
001) or non-scientific (e.g. 0.12345). Scientific notation is useful for
preserving the same decimal point accuracy level for parameter values
over different orders of magnitude.
Page 3 of 49
SUMMARY: Morphometric analysis in 3D

3D integrated analysis of all objects in 3D individual analysis of all objects in
VOI (all selected image levels) the VOI (over all selected image levels)
Parameter Symbol Unit Parameter Symbol Unit
3 3
VOI volume TV mm Object volume Obj.V mm
3 3
Object volume Obj.V mm Object surface Obj.S mm
3
Percent object volume Obj.V/TV % Volume of pores Po.V mm
2 2
VOI surface TS mm Surface of pores Po.S mm
2
Object surface Obj.S mm Porosity Po %
2
Intersection surface i.S mm Number of pores N.Po
-1
Object surface / volume Obj.S/Obj. mm
ratio V Centroid x Crd.X mm
-1
Object surface density Obj.S/TV mm Centroid y Crd.Y mm
-1
Fragmentation index Fr.I mm Centroid z Crd.Z mm
Centroid (x) Crd.X mm Structure model index SMI
Centroid (y) Crd.Y mm Volume-equivalent ESDv mm
sphere diameter
Centroid (z) Crd.Z mm Surface-equivalent ESDs mm
sphere diameter
Structure model index SMI Sphericity Sph
Structure thickness St.Th mm
Structure linear density St.Li.Dn mm
-1
Structure separation St.Sp mm
Degree of anisotropy DA
Eigenvalue 1
Eigenvalue 2
Eigenvalue 3
Fractal dimension FD
Number of objects Obj.N
Number of closed pores Po.N(cl)
3
Volume of closed pores Po.V(cl) mm
2
Surface of closed pores Po.S(cl) mm
Closed porosity (percent) Po(cl) %
3
Volume of open pores Po.V(op) mm
Open porosity (percent) Po(op) %
3
Total volume of pores Po.V(tot) mm
Total porosity (percent) Po(tot) %
Default results filename [prefix]_3D.txt/csv Default results [prefix]_i3D.txt/csv

filename
Page 4 of 49
SUMMARY: Morphometric analysis in 2D
2D integrated analysis of all objects in 2D individual analysis of all objects in

VOI (all selected image levels, line-by- the ROI (from the single current image
line) level only)
Parameter Symbol Unit Parameter Symbol Unit
3 2
VOI volume TV mm Area Ar mm
3
Object volume Obj.V mm Perimeter Pm mm
Percent object volume Obj.V/TV % Form factor FF
2
VOI surface TS mm Area-equivalent circle ECDa mm
diameter
2
Object surface Obj.S mm Roundness Rd
-1
Object surface / volume Obj.S/Obj. mm Euler number EN
ratio V
2
Mean total crossectional T.Ar mm Porosity Po %
ROI area
Mean total crossectional T.Pm mm Extent Ext
ROI perimeter
2
Mean total crossectional Obj.Ar mm Orientation Or °
object area
Mean total crossectional Obj.Pm mm Eccentricity Ecc
object perimeter
Mean number of objects Obj.N Centroid x mm
per slice
2
Average object area per Av.Obj.Ar mm Centroid y mm
slice
4
Average object equivalent Av.Obj.EC mm Moments of inertia x MMIx mm
circle diameter per slice D
4 4
Mean polar moment of MMI mm Moments of inertia y MMIy mm
inertia (polar)
4
Mean eccentricity Ecc Polar Moment of inertia MMIp mm
4
Crossectional thickness Cs.Th mm Product of inertia pI mm
4
Structure thickness (plate St.Th(pl) mm Maximum principal MMI(max) mm
model) moment of inertia
4
Structure separation (plate St.Sp(pl) mm Minimum principal MMI(min) mm
model) moment of inertia
-1
Structure linear density St.Li.Dn(pl) mm Major axis-length of Ma mm
(plate model) inertia
Structure diameter (rod Tb.Dm(rd) mm Minor axis-length of Mi mm
model) inertia
Structure separation (rod Tb.Sp(rd) mm Major diameter d(max) mm
model)
-1
Structure linear density St.Li.Dn mm Minor diameter d(min) mm
(rod model) (rd)
-1
Mean fragmentation index Fr.I mm Aspect ratio AR
Closed porosity (%) Po(cl) % Mean Thickness Av.Th mm
Centroid (x) Crd.X mm Perimeter-equivalent ECDp mm
circle diameter
Centroid (y) Crd.Y mm Hydraulic diameter HD mm
Centroid (z) Crd.Z mm
Mean fractal dimension FD
2
Total intersection surface i.S mm
Default results filename [prefix]_2D.txt/csv Default results [prefix]_3D.txt/csv

filename
Page 5 of 49
A: 3D MORPHOMETRIC PARAMETERS
A1: 3D morphometric parameters integrated for
the whole volume of interest (VOI)
Nomenclature General Scientific Bone ASBMR

Parameter name Total VOI Tissue volume
volume
Parameter symbol TV TV
Unit mm3 mm3
Total volume of the volume-of-interest (VOI). The 3D volume

measurement is based on the marching cubes volume model of the VOI.
Please note that in the case of Bone ASBMR nomenclature, the word
“tissue” simply refers to the volume of interest. It does not mean any kind
of recognition of any particular density range as biological tissue, soft,
hard or otherwise.

Parameter name Object volume Bone volume
Parameter symbol Obj.V BV
Unit mm3 mm3
Total volume of binarised objects within the VOI. The 3D volume
measurement is based on the marching cubes volume model of the
binarised objects within the VOI.

Parameter name Percent object Percent bone
volume volume
Parameter symbol Obj.V/TV BV/TV
Unit % %
The proportion of the VOI occupied by binarised solid objects. This

parameter is only relevant if the studied volume is fully contained within a
well-defined biphasic region of solid and space, such as a trabecular bone
Page 6 of 49
region, and does not for example extend into or beyond the boundary of
the object – such as the cortical boundary of a bone sample. The
meaningfulness of measured percent volume depends on the criteria
applied in selecting the volume of interest. Where the ROI / VOI
boundaries are loosely drawn in the surrounding space around an object
for instance, then % object volume has no meaning.

Parameter name VOI surface Tissue surface
Parameter symbol TS TS
Unit mm2 mm2
The surface area of the volume of interest, measured in 3D (Marching

cubes method).

Parameter name Object surface Bone surface
Parameter symbol Obj.S BS
Unit mm2 mm2
The surface area of all the solid objects within the VOI, measured in 3D
(Marching cubes method).

Parameter name Intersection Intersection
surface surface
Parameter symbol i.S i.S
Unit mm2 mm2
Intersection surface is the surface of the VOI intersected by solid binarised

objects, that is, the part of the VOI boundary surface that runs through
solid objects. This parameter is useful, for example, in evaluating bone
growth at a defined boundary – for example at a fixed distance away from
an orthopaedic bone implant.
Page 7 of 49

Parameter name Object surface / Bone surface /
volume ratio volume ratio
Parameter symbol Obj.S/Obj.V BS/BV
Unit mm-1 mm-1
The ratio of solid surface to volume measured in 3D within the VOI.

Surface to volume ratio or “specific surface” is a useful basic parameter
for characterising the thickness and complexity of structures.

density density
Parameter symbol Obj.S/TV BS/TV
Unit mm-1 mm-1
The ratio of surface area to total volume measured as described above in

3D, within the VOI.

Parameter name Fragmentation Trabecular bone
index pattern factor
Parameter symbol Fr.I Tb.Pf
Unit mm-1 mm-1
This is an inverse index of connectivity, which was developed and defined

by Hahn et al. (1992) for application to trabecular bone. It was applied by
these authors originally to 2D images of trabecular bone, and calculates
an index of relative convexity or concavity of the total bone surface, on
the principle that concavity indicates connectivity (and the presence of
“nodes”), and convexity indicates isolated disconnected structures
(struts). Tb.Pf is calculated in 3D, by comparing volume and surface of
binarised solid before and after a single voxel image dilation. It is defined:
Page 8 of 49
 S  S2 
Tb.Pf   1 
 V1  V2 
Where S and V are solid surface and volume, and the subscript numbers 1
and 2 indicate before and after image dilation.
Where structural / trabecular connectedness results in enclosed marrow
spaces, then dilation of trabecular surfaces will contract the surface. By
contrast, open ends or nodes will have their surface expanded by surface
dilation. As a result, lower Tb.Pf signifies better connected trabecular
lattices while higher Tb.Pf means a more disconnected trabecular
structure. A prevalence of enclosed cavities and concave surfaces can
push Tb.Pf to negative values – as with the structure model index (SMI) –
see below. This parameter Tb.Pf or fragmentation index is best considered
as a relative index for comparing different scanned objects; its absolute
value does not have much meaning.

Parameter name Centroids X, Y, Z Centroids X, Y, Z
Parameter symbol Crd.X, CrdY, CrdZ Crd.X, CrdY, CrdZ
Unit mm mm
The centroid is the 3D XYZ coordinate of the average Cartesian vectorial

position of all voxels within the VOI.

Parameter name Structure model Structure model
index index
Parameter symbol SMI SMI
Unit (none) (none)
Structure model index indicates the relative prevalence of rods and plates
in a 3D structure such as trabecular bone. SMI involves a measurement of
surface convex curvature. This parameter is of importance in osteoporotic
degradation of trabecular bone which is characterised by a transition from
plate-like to rod-like architecture. An ideal plate, cylinder and sphere have
SMI values of 0, 3 and 4 respectively. (Conversely, cylindrical and
spherical cavities have SMI of -3 and -4 respectively.)
Page 9 of 49
The calculation of SMI is based on dilation of the 3D voxel model, that is,
artificially adding one voxel thickness to all binarised object surfaces
(Hildebrand et al. 1997b). This is also the basis of the Tb.Pf parameter
(see above) which explains why changes in both parameters correlate
very closely with each other. SMI is derived as follows:
 S 'V 
SMI  6   2 
 S 
where S is the object surface area before dilation and S’ is the change in
surface area caused by dilation. V is the initial, undilated object volume.
It should be noted that concave surfaces of enclosed cavities represent
negative convexity to the SMI parameter, since dilation of an enclosed
space will reduce surface area causing S’ to be negative. Therefore
regions of a solid (such as bone) containing enclosed cavities – such as
regions with relative volume above 50% – can have negative SMI values.
As a consequence, the SMI parameter is sensitive to percent volume. Note
also that artificial corners and edges created by the intersection of an
object with the volume of interest boundary will also affect the measured
SMI, increasing its value.

Parameter name Structure Trabecular
thickness thickness
Parameter symbol St.Th Tb.Th
Unit mm mm
With 3D image analysis by micro-CT a true 3D thickness can be measured

which is model-independent. Local thickness for a point in solid is defined
by Hildebrand and Ruegsegger (1997a) as the diameter of the largest
sphere which fulfils two conditions:
the sphere encloses the point (but the point is not necessarily the
centre of the sphere);
the sphere is entirely bounded within the solid surfaces.
The key advantage of the local thickness measurement is that the bias
from the 3D orientation of the structure is kept to a minimum (Ulrich et al.
1999).
Distance transform methods described by Remy and Thiel (2002) are the
basis for the implementation by CT-analyser of local thickness
measurement. The method starts with a “skeletonisation” identifying the
Page 10 of 49
medial axes of all structures. Then the “sphere-fitting” local thickness

measurement is made for all the voxels lying along this axis.
Histomorphometrists typically measure a single mean value of bone Tb.Th
from a trabecular bone site. However a trabecular bone volume – or any
complex biphasic object region – can also be characterised by a
distribution of thicknesses. CT-analyser outputs a histogram of thickness
(and separation also) with an interval of two pixels. Thickness distribution
is a powerful method for characterising the shape of a complex structure.

Parameter name Structure linear Trabecular
density number
Parameter symbol St.Li.Dn Tb.N
Unit mm-1 mm-1
Structure linear density or trabecular number implies the number of

traversals across a trabecular or solid structure made per unit length on a
random linear path through the VOI.
Again the complexities of model dependence associated with 2D
measurements are eliminated by true 3D calculation of St.Li.Dn / Tb.N
from 3D micro-CT images. This parameter is measured in CT-analyser in
3D by application of the equation for the parallel plate model (fractional
volume/thickness), but using a direct 3D measurement of thickness. Note
that the optional stereology analysis (not included in this report) includes
measurements of thickness, separation and number/linear density based
on the mean intercept length (MIL) analysis which represents an
alternative basis for these architectural measurements.
Furthermore, another alternative definition of trabecular number, based
on 3D measurements of the spacing of trabeculae, is:
1
Tb.N 
Tb.Th  Tb.Sp 
Since both Tb.Th and Tb.Sp are measured in 3D by sphere-fitting, this

latter equation could be considered a calculation directly from a 3D-
measured trabecular spacing.
Page 11 of 49

separation separation
Parameter symbol Sr.Sp Tb.Sp
Unit mm mm
Trabecular separation is essentially the thickness of the spaces as defined

by binarisation within the VOI. Skyscan CT-analyser software measures
Tb.Sp directly and model-independently in 3D by the same method used
to measure trabecular thickness (see above), just applied to the space
rather than the solid voxels.

Parameter name Degree of Degree of
anisotropy anisotropy
Parameter symbol DA DA
Unit (none) (none)
Isotropy is a measure of 3D symmetry or the presence or absence of

preferential alignment of structures along a particular directional axis.
Apart from percent volume, DA and the general stereology parameters of
trabecular bone are probably the most important determinants of
mechanical strength (Odgaard 1997). Mean intercept length (MIL) and
Eigen analysis are used to calculate DA, and these involve some quite
advanced engineering mathematics. However the essentials of the MIL
eigen analysis can be summarised in normal English.
Consider a region or volume containing two phases (solid and space), both
having complex architecture, such as a region of trabecular bone. We can
study this volume to determine isotropy. If the volume is isotropic, then a
line passing through the volume at any 3D orientation will make a similar
number of intercepts through the solid phase. A bag of marbles would be
isotropic. However a packet of spaghetti would be non-isotropic, or
anisotropic, since lines going along the direction of the spaghetti would
make few intercepts along the spaghetti rods while lines crossing at right-
angles would make many intercepts. Figure 1 illustrates the difference in
the number of intercepts for lines from different directions through an
anisotropic, aligned group of structures.
Mean intercept length (MIL) analysis measures isotropy (it is usual to talk
of measurement of the negative quantity anisotropy). Mean intercept
length is found by sending a line through a 3D image volume containing
binarised objects, and dividing the length of the test line through the
analysed volume by the number of times that the line passes through or
Page 12 of 49
intercepts part of the solid phase. Note that in this MIL calculation the
intercept length may correlate with object thickness in a given orientation
but does not measure it directly. Therefore it will give an accurate result if
analysing a volume containing a sufficiently large number of objects, but
is not suitable for analysis of single or small numbers of objects. For the
MIL analysis, a grid of test lines is sent through the volume over a large
number of 3D angles. The MIL for each angle is calculated as the average
for all the lines of the grid. The spacing of this grid can be selected in CT-
analyser preferences (the “advanced” tab).
Figure 1. A group of aligned long structures has a high anisotropy: test lines
make few intercepts through the solid objects in the direction of the long axis of
the structures, but perpendicular to the structures the lines make many more
intercepts (numbers of intercepts are shown for each line).
This requires that a spherical region is first defined within which the MIL
analysis will be done and anisotropy measured, since the test lines must
all cross the sphere centre and have an equal distribution of lengths,
covering all 3D angles but distributed at random. In CT-analyser you can
actually set a spherical volume of interest (VOI). However if a non-
spherical VOI is set, the MIL analysis fits a sphere enclosing the VOI.
Page 13 of 49
Figure 2. For the MIL analysis, a grid of lines is sent through the volume over a
large number of 3D angles (just two are illustrated here). The MIL for each angle
is calculated as the average for all the lines of the grid.
The next step involves visualisation of the 3D distribution of MIL lengths

as an ellipsoid. All the MIL lines are drawn passing through one point, and
the length of each line is the bone phase MIL for that line. This process is
called a polar plot of MIL. In 3D this creates a dense pin-cushion like
effect with lines in all directions at different lengths. Figure 3 shows in a
simple diagram the appearance of an MIL distribution in 3D. Any
asymmetry in the MILs with respect to 3D angle - which will represent the
anisotropy of the bone in the spherical region - will make the line
distribution depart from an overall spherical shape and become elongated
in the direction where the solid structures have the longest MIL (such as
the axis of the spaghetti packet).
Clearly the MIL "pin-cushion" is a complex object, and a method is needed
to extract some summary numerical parameters defining the orientation
and isotropy / anisotropy of the MIL distribution. This is where the
anisotropy tensor analysis steps in. (Tensor means matrix.) This method
is probably best attributed to Harrigan and Mann (1984) and describes the
MIL distribution as an ellipsoid. An ellipsoid is a 3D ellipse.
As shown in figure 1, an ellipsoid has three axes. These describe the
longest orientation, and the length and width (major and minor axes) of
the ellipse section at right-angles to the longest orientation. The ellipsoid
can be asymmetric in one axis only, like a rugby ball, or in two axes, like
a bar of soap.
Page 14 of 49
Vector 2
Vector 3
Vector
1
Figure 3. An ellipsoid (3D ellipse) is fitted to the 3D distribution of MILs (mean

intercept lengths) measured over a full range of 3D stereo-angles. This ellipsoid is
fitted statistically and has 3 vectors which are orthogonal (at right-angles to each
other). A tensor (matrix) of 9 (3x3) eigenvectors describes the directions of the
three vectors.
An ellipsoid is fitted to the MIL “pin-cushion” 3D polar plot. This is a

statistical fit, finding the ellipsoid which most closely describes the 3D
shape of the MIL distribution. MIL analysis therefore also outputs values
indicating the strength of fit of the ellipsoid and associated error, such as
the correlation coefficients.
A tensor or matrix is a way of describing an ellipse by a 3x3 matrix of
numbers. Technically this is a second order tensor. The tensor describing
the anisotropy ellipsoid is an orthogonal tensor, since it describes the
ellipsoid axes which are orthogonal (at right angles) to each other. The
end result of the anisotropy tensor analysis is the eigen analysis, eigen
meaning characteristic. This comes in two parts. You have the 3x3 matrix
of eigenvectors which describe the 3D angles of the three axes of the
ellipsoid as described above - one column of 3 numbers for each vector.
And the three eigenvalues are each an index of the relative MIL values
(distance between bone intercepts) in each of the three axes described by
the eigenvectors.
Finally, you can derive from the tensor eigen analysis a single parameter
measuring anisotropy: this is the degree of anisotropy (DA), and is
traditionally expressed as the maximum eigenvalue divided by the
minimum eigenvalue. Values for DA calculated in this way vary from 1
(fully isotropic) to infinity (fully anisotropic). Mathematically this is a
cumbersome scale. A more convenient mathematical index of anisotropy
is calculated as:
Page 15 of 49
  min eigenvalue  
DA  1    
  max eigenvalue  
Here DA is 0 for total isotropy and 1 for total anisotropy. (Both values are
reported by CT-analyser).

Parameter name Eigenvalues 1, Eigenvalues 1,
2, 3 2, 3
Parameter symbol none none
Unit none none
The three eigenvalues are each an index of the relative MIL values
(distance between intercepts) in each of the three directions of the three
MIL analyses. These three directions are also expressed as the
eigenvectors, and are orthogonal to each other.

Parameter name Fractal Fractal
dimension dimension
Parameter symbol FD FD
Unit none none
Fractal dimension is an indicator of surface complexity of an

object, which quantifies how that object’s surface fills space.
For examples of fractal objects, “fractal art” is abundant on
the internet. True fractal objects have surface shapes which
are repeated over many spatial scales. So the closer you look
(i.e. the higher the magnification or “zoom in”) the more self-
similar structure you see. A typical example is a fern leaf in
which each side-branch is very similar to the whole fern leaf,
and likewise each side-finger of each side branch also looks
the same as the whole fern leaf, and so on. A true fractal
object essentially has fractional, non-integer dimension, i.e. a
line “trying” to fill a plane, or a plane trying to fill a 3D space,
having dimension somewhere between 2 and 3.
Page 16 of 49
Fractal dimension is calculated using the Kolmogorov or “box counting”

method. It is calculated in both 2D and 3D in Skyscan CT-analyser. For
the 3D calculation of FD, the volume is divided into an array of equal
cubes, and the number of cubes containing part of the object surface is
counted. This is repeated over a range of cube sizes such as 2-100 pixels.
The number of cubes containing surface is plotted against cube length in a
log-log plot, and the fractal dimension is obtained from the slope of the
log-log regression. Fractal characteristics of trabecular bone, and methods
for measurement of fractal dimension, are discussed by Chappard et al.
(2001).
For more details on fractal dimension please refer to:
http://en.wikipedia.org/wiki/Fractal_dimension

Parameter name Number of Number of
objects objects
Parameter symbol N. Obj. N. Obj.
Unit none none
The total number of discreet binarised objects within the VOI is reported.
A discreet 3D object is a connected assemblage of solid (white) voxels
fully surrounded on all sides in 3D by space (black) voxels.

Closed Pores Closed Pores
Parameter symbol Po.N(cl) Po.N(cl)
Unit none none
The total number of discreet binarised closed pores within the VOI is
reported. A closed pore in 3D is a connected assemblage of space (black)
voxels that is fully surrounded on all sides in 3D by solid (white) voxels.
Page 17 of 49

Parameter name Volume of Volume of
Parameter symbol Po.V(cl) Po.V(cl)
Unit mm3 mm3
The total volume of all closed pores within the VOI, as defined above
(under “number of closed pores”) is reported.

Parameter name Surface of Surface of
Parameter symbol Po.S(cl) Po.S(cl)
Unit mm2 mm2
The total surface area of all closed pores within the VOI, as defined above
(under “number of closed pores”) is reported.

Parameter name Closed Porosity Closed Porosity
(percent) (percent)
Parameter symbol Po Po
Unit % %
Percent closed porosity is the volume of closed pores (as defined above)
as a percent of the total of solid plus closed pore volume, within the VOI.
(Please note – this is a “material porosity”, and is calculated differently
from of open porosity and total porosity, where the denominator is total
VOI volume.)
Page 18 of 49

Parameter name Volume of Open Volume of Open
Pore Space Pore Space
Parameter symbol Po.V(op) Po.V(op)
Unit mm3 mm3
The total volume of all open pores within the VOI, is reported. An open
pore is defined as any space located within a solid object or between solid
objects, which has any connection in 3D to the space outside the object or
objects.

Parameter name Open Porosity Open Porosity
(percent) (percent)
Parameter symbol Po(op) Po(op)
Unit % %
Percent open porosity is the volume of open pores (as defined above) as a
percent of the total VOI volume.

Parameter name Total Volume of Total Volume of
Pore Space Pore Space
Parameter symbol Po.V(tot) Po.V(tot)
Unit mm3 mm3
The total volume of all open and closed pores within the VOI, is reported.
Page 19 of 49

Parameter name Total Porosity Total Porosity
(percent) (percent)
Parameter symbol Po(tot) Po(tot)
Unit % %
Total porosity is the volume of all open plus closed pores (as defined
above) as a percent of the total VOI volume.

Parameter name Euler number Euler number
Parameter symbol EN EN
Unit none none
The calculation of Euler number in 3D is based on software code kindly

provided by Erasmus University, Rotterdam, The Netherlands (Dr Erwin
Waarsing) using the Conneulor method from Aarhus University, Denmark
(Dr. Anders Odgaard; see Gunderson et. al. 1993).
Note also that up to CTAn version 1.9.3.2 Euler number was calculated in
the 2D analysis only (both in the integrated and individual object 2D
analyses) . From CTAn version 1.9.3.3 onwards Euler connectivity and
number are also calculated in 3D, and included in the standard
(integrated) 3D analysis.
The Euler-Poincare number –abbreviated to “Euler number” – is an
indicator of connectedness of a 3D complex structure. The Euler number is
characteristic of a three-dimensional structure which is topologically
invariant (it is unchanged by inflation or compression or distortion of the
structure). It measures what might be called “redundant connectivity” –
the degree to which parts of the object are multiply connected (Odgaard
et al. 1993). It is a measure of how many connections in a structure can
be severed before the structure falls into two separate pieces.
(Topologically the object can be compressed into a sphere and the
redundant connections appear as “handles”.) The components of the Euler
number are the three Betti numbers: 0 is the number of objects, 1 the
connectivity, and 2 the number of enclosed cavities. The Euler-Poincare
formula for a 3D object X is given:
 ( X )  0  1  2
Page 20 of 49
Euler analysis provides a measure of connectivity density, indicating the

number of redundant connections between trabecular structures per unit
volume. Trabecular connectivity can contribute significantly to structure
strength (Odgaard 1997). One useful and fast algorithm for calculating the
Euler connectivity in 3D is the “Conneulor” (Gunderson et al. 1993).
Page 21 of 49
A2: 3D morphometric parameters calculated for all

individual binarised 3D object within the volume of
interest (VOI)

Unit mm3 mm3
Total volume of each binarised 3D object within the VOI. The 3D volume
measurement is based on the marching cubes volume model.

Unit mm2 mm2
Total surface of each binarised 3D object within the VOI. The 3D surface
measurement is based on the marching cubes model.

Parameter name Volume of Pores Volume of Pores
Parameter symbol Po.V Po.V
Unit mm3 mm3
The total volume of all pores within each discreet 3D object.
Page 22 of 49

Parameter name Surface of Pores Surface of Pores
Parameter symbol Po.S Po.S
Unit mm2 mm2
The total surface area of all pores within each discreet 3D object.

Parameter name Percent Porosity Percent Porosity
Unit % %
Percent porosity is the volume of closed pores within each discreet 3D

object as a percent of the total volume of that object (including any closed
pore volume).

Pores Pores
Parameter symbol Po.N Po.N
Unit none none
The number of pores within each discreet 3D object.

Unit mm mm
The centroid is the XYZ coordinate of the average Cartesian vectorial

position of each discreet 3D object, within the VOI.
Page 23 of 49

Parameter name Structure model Structure model
index index
Parameter symbol SMI SMI
Unit
Structure model index indicates the relative prevalence of rods and plates
in a 3D structure such as trabecular bone. SMI involves a measurement of
surface convexity. Ideal values of SMI for an exact rectangular plate,
cylinder and sphere are 0, 3 and 4 respectively. (Conversely, cylindrical
and spherical cavities have SMI of -3 and -4 respectively.)
The calculation of SMI is based on dilation of the 3D voxel model, that is,
artificially adding one voxel thickness to all binarised object surfaces
(Hildebrand et al. 1997b). SMI is derived as follows:
 S 'V 
SMI  6   2 
 S 
where S is the object surface area before dilation and S’ is the change in
surface area caused by dilation. V is the initial, undilated object volume.
It should be noted that concave surfaces of enclosed cavities represent
negative convexity to the SMI parameter, since dilation of an enclosed
space will reduce surface area causing S’ to be negative. Therefore
regions of a solid (such as bone) containing enclosed cavities – such as
regions with relative volume above 50% – can have negative SMI values.
As a consequence, the SMI parameter is sensitive to percent volume. Note
also that artificial corners and edges created by the intersection of an
object with the volume of interest boundary will also affect the measured
SMI, increasing its value.

Parameter name Volume- Volume-
equivalent equivalent
sphere diameter sphere diameter
Parameter symbol ESDv ESDv
Unit mm mm
The diameter of the sphere that would have the same volume as the
discreet 3D object. ESDv is expressed:
Page 24 of 49
where V is the object volume.

Parameter name Surface- Surface-
equivalent equivalent
sphere diameter sphere diameter
Parameter symbol ESDs ESDs
Unit mm mm
The diameter of the sphere that would have the same surface area as the
discreet 3D object. ESDs is expressed:
where S is the object surface area.

Parameter name Sauter diameter Sauter diameter
Parameter symbol SD SD
Unit mm mm
The Sauter mean diameter is the diameter of the sphere that would have
the same volume-to-surface (V/S) ratio as the discreet 3D object. SD is a
fluid dynamics parameter, introduced by the German scientist J. Sauter in
the late 1920’s; it is expressed:
Page 25 of 49
where ESDv and ESDs are as defined just above. A Wikipedia entry on SD
is at:
http://en.wikipedia.org/wiki/Sauter_mean_diameter

Parameter name Sphericity Sphericity
Parameter symbol Sph Sph
Unit
Sphericity is a measure of how spherical a 3D object is. This parameter

was defined by Wadell in 1935. Sphericity, Ψ or Sph, of a particle is the
ratio of the surface area of a sphere (with the same volume as the given
particle) to the surface area of the particle:
√ ( )
where V and S are the object volume and surface area respectively.
For complex, non-spherical objects the surface area of the volume-
equivalent sphere will be much smaller than the particle surface area, thus
Sph will be low. The maximum value possible is 1, which would be
obtained for a sphere.
A Wikipedia entry on Sph is at:
http://en.wikipedia.org/wiki/Sphericity
Page 26 of 49
B: 2D MORPHOMETRIC PARAMETERS
B1: 2D morphometric parameters calculated “slice-
by-slice”, integrated for the whole volume of
interest (VOI)

Parameter name Total VOI Tissue volume
volume
Parameter symbol TV TV
Unit mm3 mm3
Total volume of the volume-of-interest (VOI). The 2D measurement is the

total number of voxels of (solid and space) in the VOI times the voxel
volume. Please note that in the case of Bone ASBMR nomenclature, the
word “tissue” refers to the volume of interest. It does not mean any kind
of recognition of any particular density range as biological tissue, soft,
hard or otherwise.

Unit mm3 mm3
Total volume of binarised objects within the VOI. The 2D measurement is

equal to the number of voxels of binarised solid objects within the VOI
times the voxel volume.
Page 27 of 49

Parameter name Percent volume Percent bone
volume
Parameter symbol Obj.V/TV BV/TV
Unit % %
The percentage of the VOI occupied by binarised solid objects. This

parameter is only relevant if the studied volume is fully contained within a
well-defined biphasic region of solid and space, such as a trabecular bone
region, and does not for example extend into or beyond the boundary of
the object – such as the cortical boundary of a bone sample. The
meaningfulness of measured percent volume depends on the criteria
applied in selecting the volume of interest. Where the ROI / VOI
boundaries are loosely drawn in the surrounding space around an object
for instance, then % object volume has no meaning.

Parameter name VOI surface Tissue surface
Parameter symbol TS TS
Unit mm2 mm2
The surface area of the volume of interest, measured in 2D, using the
Pratt algorithm (Pratt 1991). Note that there are two components to
surface measured in 2D for a 3D multilayer dataset; first the perimeters of
the binarised objects on each crossectional level, and second, the vertical
surfaces exposed by pixel differences between adjacent crossections.

Unit mm2 mm2
The surface area of all the solid objects within the VOI, measured in 2D,
using the Pratt algorithm (Pratt 1991). Note that there are two
components to surface measured in 2D for a 3D multilayer dataset; first
the perimeters of the binarised objects on each crossectional level, and
Page 28 of 49
second, the vertical surfaces exposed by pixel differences between

adjacent crossections.

Parameter name Object surface / Bone surface /
volume ratio volume ratio
Parameter symbol Obj.S/Obj.V BS/BV
Unit mm-1 mm-1
The ratio of solid surface to volume (both as defined above) measured in

2D within the VOI. Surface to volume ratio or “specific surface” is a useful
basic parameter in characterising the complexity of structures.

Parameter name Mean total Mean total
crossectional crossectional
ROI area tissue area
Parameter symbol T.Ar T.Ar
Unit mm2 mm2
The mean of the crossectional ROI area for all slices in the selected range
of the VOI. Note that by using the “shrink-wrap” function, the ROI
boundary can become effectively the outer boundary of the object, while
not including any internal spaces or structure in the measurement. This
allows for example measurement of inner and outer surfaces of porous or
hollow objects, such as the periosteal (outer) and endosteal (inner)
cortical bone surfaces.
Page 29 of 49

ROI perimeter tissue perimeter
Parameter symbol T.Pm T.Pm
Unit mm mm
The mean of the crossectional ROI perimeter for all slices in the selected
range of the VOI. Note that by using the “shrink-wrap” function, the ROI
perimeter can become effectively the outer perimeter of the object, while
not including any internal spaces or structure in the measurement. This
allows for example measurement of inner and outer surfaces of porous or
hollow objects.

object area bone area
Parameter symbol Obj.Ar B.Ar
Unit mm2 mm2
The mean of the crossectional object or bone area for all slices in the
selected range of the VOI.

object perimeter bone perimeter
Parameter symbol Obj.Pm B.Pm
Unit mm mm
The mean of the crossectional object or bone perimeter for all slices in the
selected range of the VOI.
Page 30 of 49

Parameter name Mean number of Mean number of
objects per slice objects per slice
Parameter symbol Obj.N Obj.N
Unit (none) (none)
The mean number of discreet 2D objects in a single crossectional image

slice. This parameter can be an indicator of structural connectivity – in
interconnected structures (e.g. trabecular bone or other networks,
scaffolds or foams) high connectivity results in highly interconnected
objects which are few in number in a single crossection – by contrast
fragmentation results in larger numbers of smaller objects (e.g.
Vermeirsch et al. 2004, 2007).

Parameter name Average object Average object
area per slice area per slice
Parameter symbol Av.Obj.Ar Av.Obj.Ar
Unit mm2 mm2
The mean area of discreet 2D objects in a single crossectional image slice.

This parameter can be an indicator of structural connectivity – in
interconnected structures (e.g. trabecular bone or other networks,
scaffolds or foams) high connectivity results in highly interconnected
objects which are few in number in a single crossection – by contrast
fragmentation results in larger numbers of smaller objects (for application
of this parameter to bone morphometry see for example Vermeirsch et al.
2004, 2007).
Page 31 of 49

Parameter name Average object Average object
equivalent circle equivalent circle
diameter per diameter per
slice slice
Parameter symbol Av.Obj.ECD Av.Obj.ECD
Unit mm mm
The mean equivalent circle diameter of all discreet 2D objects in a single

crossectional image slice, defined as the diameter of the equivalent circle
having the same area as the measured object.

Parameter name Mean polar Mean polar
moment of moment of
inertia inertia
Parameter symbol MMI(polar) MMI(polar)
Unit mm4 mm4
This 2D crossectional function is a basic strength index and indicates the

resistance to rotation of a crossection about a chosen axis (assuming
uniform material stress-strain strength properties).
Moment of inertia is the rotational analogue of mass for linear motion and
must be specified with respect to a chosen axis of rotation. For a point
mass (represented by an image pixel) the moment of inertia (I) is simply
the mass (m) times the square of perpendicular distance (r) to the
rotation axis, I = mr2. Moment of inertia for a crossection – for example of
cortical bone – is the integral of all the solid (bone) pixels.
For more detail on moment of inertia please refer to:

http://en.wikipedia.org/wiki/Moment_of_inertia
Page 32 of 49

Parameter name Mean Mean
eccentricity eccentricity
Parameter symbol Ecc Ecc
Unit (none) (none)
Eccentricity is a 2D shape analysis of discreet binarised objects. Objects

are approximated as ellipses, and
eccentricity is an elliptic parameter
indicating departure from circular
shape by lengthening (a circle is an
ellipse with an eccentricity of zero).
An ellipse is defined as having two
focal points (F1 and F2), a centre (C)
and major and minor axes. The major
axis is defined as 2a where a is the
“semi major axis”; likewise, 2b is the
minor axis. Eccentricity e is a function
of the (semi)major and minor axes
such that:
b2
e  1
a2
Higher eccentricity means generally elongated objects, while decreased

values means an approach toward circular shape.

Parameter name Crossectional Crossectional
thickness thickness
Parameter symbol Cs.Th Cs.Th
Unit mm mm
Crossectional thickness is calculated in 2D on the basis of the plate model

where Th = 2 / (surface/volume) . Surface in 2D integrated “slice-by-
slice” analysis is calculated as described above (Obj.S) from both
horizontal and vertical pixel boundaries. Crossectional thickness is a
calculation of thickness in which the vertical surfaces, between pixels of
adjacent crossections, are excluded; only the perimeters of crossections
are used to calculate surface. This definition of surface excludes error
caused by large artificially cut surfaces that sometimes occur at the top
Page 33 of 49
and bottom of VOIs, such as a VOI representing a short section of cortical

bone with large exposed crossections at each end. (By contrast, the
“structure thickness” or “trabecular thickness” 2D parameters – below –
include surface measured both vertically and horizontally: thus this latter
thickness calculation yields smaller thickness values.)

thickness (plate thickness (plate
and rod models) and rod models)
Parameter symbol St.Th (pl) or (rd) Tb.Th(pl) or (rd)
Unit mm mm
Calculation – or estimation – of St.Th or Tb.Th from 2D measurements

requires an assumption about the nature of the structure (a “structural
model”). Three simple structure models, the parallel plate, the cylinder
rod and the sphere model, provide the range of values within which a
hypothetical “true” thickness will be located. Thickness defined by these
models (Parfitt et al. 1987) is given:
Parallel plate model:

2
Tb.Th 
BS / BV 
Cylinder rod model:

4
Tb.Th 
BS / BV 
Sphere model:
6
Tb.Th 
BS / BV 
Where BS/BV is the surface to volume ratio, mm-1.

Note that for the above equations, if the surface measurements are made
from one or a few 2D crossectional slices (such as in histological analysis),
then the numerator (2, 4 or 6) should be divided by a correction factor of
1.199. This approximates a correction from 2D to 3D. However as
mentioned above (object surface, Obj.S) the “2D” surface measurement in
Page 34 of 49
CT-analyser includes vertical surfaces between solid and space voxels in

adjacent image slices. It is thus in fact an essentially 3D measurement
based on a simple cubic voxel model. Therefore the thickness model
based estimated should not here include the 1.199 correction factor.

separation separation
(plate and rod (plate and rod
models) models)
Parameter symbol Sr.Sp Tb.Sp
Unit mm mm
Trabecular separation is essentially the thickness of the spaces as defined

by binarisation within the VOI. It can also be calculated from 2D images
with model assumptions. Applying the structural models as for thickness,
structure separation is calculated:
 1 
Tb.Sp     Tb.Th
 Tb.N 
Cylinder rod model:
     TV  
Tb.Sp  Tb.Dm         1
  4   BV  

where TV is total volume of VOI and BV is bone (or solid) volume (Parfitt
et al. 1987). Note that each of the above definitions takes the Tb.Th (or
Tb.Dm) value derived from the corresponding plate or rod model.
Page 35 of 49

Parameter name Structure linear Trabecular
density (plate number (plate
and rod models) and rod models)
Parameter symbol St.Li.Dn(pl) Tb.N
Unit mm-1 mm-1
Structure linear density or trabecular number implies the number of

traversals across a solid structure, such as a bone trabecula, made per
unit length by a random linear path through the volume of interest (VOI).
Plate and rod model 2D-based definitions of Tb.N again take the
corresponding Tb.Th values:
( BV / TV )
Tb.N 
Tb.Th
Cylinder rod model:
 4   BV 
  
    TV 
Tb.N 
Tb.Dm

Parameter name Fragmentation Trabecular bone
index pattern factor
Parameter symbol Fr.I Tb.Pf
Unit mm-1 mm-1
This is an inverse index of connectivity, which was developed and defined

by Hahn et al. (1992) for application to trabecular bone. It was applied by
these authors originally to 2D images of trabecular bone, and it calculates
an index of relative convexity or concavity of the total bone surface, on
the principle that concavity indicates connectivity (and the presence of
“nodes”), and convexity indicates isolated disconnected structures
(struts). Tb.Pf is calculated in 2D by comparing area and perimeter of
binarised solid before and after a single voxel image dilation. It is defined:
Page 36 of 49
 P  P2 
Tb.Pf   1 
 A1  A2 
Where P and A are the solid perimeter and area, respectively, and the
subscript numbers 1 and 2 indicate before and after image dilation.
Structural / trabecular connectedness results in enclosed spaces and
concave perimeters, and dilation of such perimeters will contract them,
resulting in reduced perimeter. By contrast, open ends or nodes which
have convex perimeters will have their surface expanded by surface
dilation, thus causing increased perimeter. As a result, lower Tb.Pf
signifies better connected trabecular lattices while higher Tb.Pf means a
more disconnected trabecular structure. A prevalence of enclosed cavities
and concave surfaces can push Tb.Pf to negative values. This parameter is
best considered as a relative index for comparing different scanned
objects; its absolute value does not have much meaning.

Parameter name Closed Porosity Closed Porosity
(percent) (percent)
Parameter symbol Po(cl) Po(cl)
Unit % %
Closed porosity is measured in the 2D slice-by-slice analysis in CT-

analyser. Binarised objects are identified containing fully enclosed spaces,
and closed porosity is the area of those spaces as a percent of the total
area of binarised objects. Note that the denominator of total object area
includes the closed pores. Closed porosity measurement ignores space
which is not fully surrounded by solid. Note – closed porosity in 2D is
usually much larger than the equivalent parameter measured in 3D; a
space region is more likely to be surrounded by solid in a single
crossectional plane in 2D than in all directions in 3D.
Page 37 of 49

Unit mm mm
The centroid is the XYZ coordinate of the average Cartesian vectorial

position of all voxels within the VOI.

Parameter name Fractal Fractal
dimension dimension
Parameter symbol FD FD
Unit none none
Fractal dimension is an indicator of surface complexity of an

object, which quantifies how that object’s surface fills space.
For examples of fractal objects, “fractal art” is abundant on
the internet. True fractal objects have surface shapes which
are repeated over many spatial scales. So the closer you look
(i.e. the higher the magnification or “zoom in”) the more self-
similar structure you see. A typical example is a fern leaf in
which each side-branch is very similar to the whole fern leaf,
and likewise each side-finger of each side branch also looks
the same as the whole fern leaf, and so on. A true fractal
object essentially has fractional, non-integer dimension, i.e. a
line “trying” to fill a plane, or a plane trying to fill a 3D space,
having dimension somewhere between 2 and 3.
Fractal dimension is calculated using the Kolmogorov or “box counting”
method. It is calculated in both 2D and 3D in Skyscan CT-analyser. For 2D
calculation of FD, the crossection image is divided into an array of equal
squares, and the number of squares containing part of the object surface
is counted. This is repeated over a range of square sizes such as 2-100
pixels. The number of squares containing surface is plotted against square
length in a log-log plot, and the fractal dimension is obtained from the
slope of the log-log regression. Fractal characteristics of trabecular bone,
and methods for measurement of fractal dimension, are discussed by
Chappard et al. (2001). Fractal dimension is a useful indicator of
complexity in structures.
Page 38 of 49

Parameter name Intersection Intersection
surface surface
Parameter symbol i.S i.S
Unit mm2 mm2
Intersection surface is the surface of the ROI intersected by solid binarised

objects, that is, the part of the ROI boundary surface that runs through
solid objects. This parameter is useful, for example, in evaluating bone
growth at a fixed distance away from an orthopaedic bone implant. The
intersection surface is calculated for each separate crossection level, but is
also integrated over all VOI levels in the data summary at the head of the
2D analysis report. Note that the intersection surface calculated in 2D
does not include vertical (between-crossection) surfaces, only perimeters
of objects within crossections. This can make the 2D calculation of i.S.
useful for excluding artificial cut surfaces at the top and bottom of the
VOI, as was discussed for the parameter crossectional thickness.
Page 39 of 49
B2: 2D morphometric parameters calculated for

each individual binarised 2D object within the
region of interest (ROI) for the single selected
image level only

Parameter name Area Area
Parameter symbol Ar Ar
Unit mm2 mm2
The area of the individual 2D object, calculated using the Pratt algorithm
(Pratt 1991).

Parameter name Perimeter Perimeter
Parameter symbol Pm Pm
Unit mm mm
The perimeter of the individual 2D object, calculated using the Pratt

algorithm (Pratt 1991).

Parameter name Form factor Form factor
Parameter symbol FF FF
Unit none none
The form factor of the individual 2D object. FF is calculated by the

equation below:
 4  A 
FF   
 Pm 
2
Page 40 of 49
where A and Pm are object area and perimeter respectively. Elongation of

individual objects results in smaller values of form factor.

Parameter name Area Equivalent Area Equivalent
Circle Diameter Circle Diameter
Parameter symbol ECDa ECDa
Unit mm mm
The area-equivalent circle diameter of a discreet 2D object is defined as

the diameter of the circle having the same area as the measured object. It
is calculated as
where A is the area of the object.

Parameter name Roundness Roundness
Parameter symbol R R
Unit none none
Roundness of a discreet 2D object is defined as:
( ( ) )
where A is the object area. The value of R ranges from 0 to 1, with 1

signifying a circle.
Page 41 of 49

Parameter name Euler number Euler number
Parameter symbol EN EN
Unit none none
The Euler-Poincare – or the abbreviated “Euler number” – is also an

indicator of connectedness of a complex object. The Euler number
measures what might be called “redundant connectivity” – the degree to
which parts of the object are multiply connected (Odgaard et al. 1993). It
is a measure of how many connections in a structure can be severed
before the structure falls into two separate pieces. (Topologically the
object can be compressed into a circle and the redundant connections
appear as “handles”.) The components of the Euler number are the three
Betti numbers: 0 is the number of objects, 1 the connectivity, and 2 the
number of enclosed cavities. The Euler-Poincare formula for a 3D object X
is given:
 ( X )   0  1   2
In calculating the Euler number of an individual object,  0 will always be 1.

The values of 1 and  2 will therefore determine the single object’s Euler
number.

Parameter name Porosity Porosity
Unit % %
Porosity for individual objects in 2D is defined as closed porosity as

described elsewhere – the volume of any spaces fully surrounded in the
given crossectional plane by solid, as a percent of the volume of solid plus
closed pores.
Page 42 of 49

Parameter name Extent Extent
Parameter symbol Ext Ext
Unit none none
The parameter “extent” for an object is calculated as:

area
Ext 
width  height
Extent is the area of the object as a fraction of the

area of the horizontally aligned rectangle bounding
the object (see the image below).

Parameter name Orientation Orientation
Parameter symbol Or Or
Unit degree degree
The orientation of the major axis of the object, in degrees. Orientation is

defined in the upper hemisphere only, with values from 0-180 degrees.

Parameter name Eccentricity Eccentricity
Parameter symbol Ecc Ecc
Unit (none) (none)
Eccentricity is a 2D shape analysis of

discreet binarised objects. Objects are
approximated as ellipses, and
eccentricity is an elliptic parameter
indicating departure from circular
shape by lengthening (a circle is an
ellipse with an eccentricity of zero). An
ellipse is defined as having two focal
points (F1 and F2), a centre (C) and
major and minor axes. The major axis
Page 43 of 49
is defined as 2a where a is the “semi major axis”; likewise, 2b is the

minor axis. Eccentricity e is a function of the (semi)major and minor axes
such that:
b2
e  1
a2
Higher eccentricity means generally elongated objects, while decreased

values means an approach toward circular shape.

Parameter name Centroids X, Y Centroids X, Y
Parameter symbol Crd.X, CrdY Crd.X, CrdY
Unit mm mm
The centroid of an individual object in 2D is the XY coordinate of the

average Cartesian vectorial position of all the pixels within the object.

Parameter name Moment of Moment of
inertia inertia
parameters parameters
Parameter symbol MMI MMI
Unit mm4 mm4
Moment of inertia is a 2D crossectional function indicating mechanical

strength in terms of the resistance to rotation of a crossection about a
chosen axis (assuming uniform material stress-strain strength properties).
For individual 2D objects, the X, Y, polar, product of inertia, and the
minimum and maximum principle moments of inertia are calculated.
Moment of inertia is the rotational analogue of mass for linear motion and
must be specified with respect to a chosen axis of rotation. For a point
mass (represented by an image pixel) the moment of inertia (I) is simply
the mass (m) times the square of perpendicular distance (r) to the
rotation axis, I = mr2. Moment of inertia for a crossection – for example of
cortical bone – is the integral of all the solid (bone) pixels.
Page 44 of 49
Note that the maximum and minimum MMI values are the two
orthogonally opposed MMI values for which the product of inertia equals
zero.
For more detail on moment of inertia please refer to:
http://en.wikipedia.org/wiki/Moment_of_inertia

Parameter name Major axis- Major axis-
length of inertia length of inertia
Parameter symbol Mj Mj
Unit mm mm
The length of the major axis of an approximation of the object as an

ellipse. The major axis is two times the elliptical parameter a, the “semi-
major axis” already described under eccentricity. Mj is also called the
“major length of inertia”.
For more details on the major axis and the ellipse, please refer to:
http://en.wikipedia.org/wiki/Major_axis

Parameter name Minor axis- Minor axis-
length of inertia length of inertia
Parameter symbol Mi Mi
Unit mm mm
The length of the minor axis of an approximation of the object as an

ellipse. The minor axis is two times the elliptical parameter b, the “semi-
minor axis” already described under eccentricity. Mi is also called the
“minor length of inertia”.
For more details on the minor axis and the ellipse, please refer to:
http://en.wikipedia.org/wiki/Minor_axis
Page 45 of 49

Parameter name Major diameter Major diameter
Parameter symbol d(max) d(max)
Unit mm mm
The major diameter is defined for an individual 2D object as the distance

between the two most distant pixels in that object.

Parameter name Minor diameter Minor diameter
Parameter symbol d(min) d(min)
Unit mm mm
The minor diameter is defined for an individual 2D object as longest chord

through the object that can be drawn in the direction orthogonal to that of
the major diameter (see above).

Parameter name Aspect Ratio Aspect Ratio
Parameter symbol AR AR
Unit none none
The aspect ratio of a 2D object is defined as the ratio of the major to the
minor diameter, d(max)/d(min), these two parameters defined as stated
above.
Page 46 of 49

Parameter name Mean Thickness Mean Thickness
Parameter symbol Av.Th Av.Th
Unit mm mm
Mean thickness is calculated for an individual 2D object by a 2D

implementation of the sphere-fitting local thickness method used for 3D
thickness measurement, using circles not spheres.

Parameter name Perimeter Perimeter
Equivalent Equivalent
Circle Diameter Circle Diameter
Parameter symbol ECDp ECDp
Unit mm mm
The perimeter-equivalent circle diameter of a discreet 2D object is defined

as the diameter of the circle having the same perimeter as the measured
object. It is calculated as:
where P is the object perimeter.

Parameter name Hydraulic Hydraulic
diameter diameter
Parameter symbol HD HD
Unit mm mm
The hydraulic diameter, HD, is a term used for analysing flow in

noncircular tubes.
Page 47 of 49
Where A and P are the object crossectional area and perimeter

respectively.
A Wikipedia article on HD can be found at:
http://en.wikipedia.org/wiki/Hydraulic_diameter
Please note that several of these definitions of individual 2D objects

conform to definitions established by the ASTM (American Society for
Testing and Measurement) in the document Designation: F 1877 – 98
(Reapproved 2003) entitled: “Standard Practice for Characterization of
Particles”.
Page 48 of 49
References
ASTM (American Society for Testing and Measurement) document Designation: F
1877 – 98 (Reapproved 2003): “Standard Practice for Characterization of
Particles”.
Chappard D, Legrand E, Haettich B, Chales G, Auvinet B, Eschard J-P, Hamelin J-
P, Basle M-F, Audran M (2001) Fractal dimension of trabecular bone: comparison
of three histomorphometric computed techniques for measuring the architectural
two-dimensional complexity. J. Pathol. 195: 515–521.
Gunderson HJG, Boyce RW, Nyengaard JR, Odgaard A (1993) The Conneulor:
unbiased estimation of connectivity using physical disectors under projection.
Bone 14: 217-222.
Hahn M, Vogel M, Pompesius-Kempa M, Delling G (1992) Trabecular bone pattern
factor – a new parameter for simple quantification of bone microarchitecture.
Bone 13: 327-330.
Harrigan TP and Mann RW (1984) Characterisation of microstructural anisotropy
in orthotropic materials using a second rank tensor. J. Mater. Sci. 19: 761-767.
Hildebrand T and Ruegsegger P (1997a) A new method for the model
independent assessment of thickness in three dimensional images. J. Microsc.
185: 67-75.
Hildebrand T and Ruegsegger P (1997b) Quantification of bone microarchitecture
with the structure model index. Comp. Meth. Biomech. Biomed. Eng. 1: 15-23.
Lorensen WE, Cline HE (1987) Marching cubes: a high resolution 3D surface
construction algorithm. Computer graphics 21 (4): 163-169.
Odgaard A (1997) Three-dimensional methods for quantification of cancellous
bone architecture. Bone 20 (6): 315-328.
Odgaard A, Gundersen HJ (1993) Quantification of connectivity in cancellous
bone, with special emphasis on 3-D reconstructions. Bone 14 (2): 173-182.
Pratt WK. Digital image processing, 2nd ed. New York: Wiley; 1991. Chapter 16.
Parfitt AM, Drezner MK, Glorieux FH, Kanis JA, Malluche H, Meunier PJ, Ott SM,
Recker RR (1987) Bone Histomorphometry: standardization of nomenclature,
symbols and units. Journal of Bone and Mineral Research 2 (6): 595-610.
Remy E, Thiel E 2002 Medial axis for chamfer distances: computing look-up tables
and neighbourhoods in 2D or 3D. Pattern Recognition Letters 23: 649–661.
Sauter J 1928 Die Grössenbestimmung der in Gemischnebeln von
Verbrennungskraftmaschinen vorhandenen Brennstoffteilchen" VDI-
Forschungsheft Nr. 279 (1926) und Nr. 312.
Ulrich D, van Rietbergen B, Laib A, Rüegsegger P (1999) The ability of three-
dimensional structural indices to reflect mechanical aspects of trabecular bone.
Bone 25 (1): 55-60.
Vermeirsch H, Nuydens R, Salmon PL, Meert TF (2004) Bone cancer pain model in
mice: Evaluation of pain behavior, bone destruction and morphine sensitivity.
Pharmacology, Biochemistry and Behavior 79: 243-251.
Vermeirsch H, Biermans R, Salmon PL, Meert TF (2007) Evaluation of pain
behavior and bone destruction in two arthritic models in guinea pig and rat.
Pharmacol. Biochem. Behav. 87 (3): 349-359.
Wadell, Hakon 1935 Volume, Shape and Roundness of Quartz Particles. Journal of
Geology 43 (3): 250–280. doi:10.1086/624298.
Page 49 of 49

CTAn Parameter Description

Uploaded by

Copyright:

Available Formats

CTAn Parameter Description

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

CTAn Parameter Description

Uploaded by

Copyright:

Available Formats

Morphometric

SUMMARY: Morphometric analysis in 3D

Default results filename [prefix]_3D.txt/csv Default results [prefix]_i3D.txt/csv

SUMMARY: Morphometric analysis in 2D

2D integrated analysis of all objects in 2D individual analysis of all objects in

Default results filename [prefix]_2D.txt/csv Default results [prefix]_3D.txt/csv

Nomenclature General Scientific Bone ASBMR

Total volume of the volume-of-interest (VOI). The 3D volume

Nomenclature General Scientific Bone ASBMR

Nomenclature General Scientific Bone ASBMR

The proportion of the VOI occupied by binarised solid objects. This

Nomenclature General Scientific Bone ASBMR

The surface area of the volume of interest, measured in 3D (Marching

Nomenclature General Scientific Bone ASBMR

Nomenclature General Scientific Bone ASBMR

Intersection surface is the surface of the VOI intersected by solid binarised

Nomenclature General Scientific Bone ASBMR

The ratio of solid surface to volume measured in 3D within the VOI.

Nomenclature General Scientific Bone ASBMR

The ratio of surface area to total volume measured as described above in

Nomenclature General Scientific Bone ASBMR

This is an inverse index of connectivity, which was developed and defined

Nomenclature General Scientific Bone ASBMR

The centroid is the 3D XYZ coordinate of the average Cartesian vectorial

Nomenclature General Scientific Bone ASBMR

Nomenclature General Scientific Bone ASBMR

With 3D image analysis by micro-CT a true 3D thickness can be measured

medial axes of all structures. Then the “sphere-fitting” local thickness

Nomenclature General Scientific Bone ASBMR

Structure linear density or trabecular number implies the number of

Since both Tb.Th and Tb.Sp are measured in 3D by sphere-fitting, this

Nomenclature General Scientific Bone ASBMR

Trabecular separation is essentially the thickness of the spaces as defined

Nomenclature General Scientific Bone ASBMR

Isotropy is a measure of 3D symmetry or the presence or absence of

The next step involves visualisation of the 3D distribution of MIL lengths

Figure 3. An ellipsoid (3D ellipse) is fitted to the 3D distribution of MILs (mean

An ellipsoid is fitted to the MIL “pin-cushion” 3D polar plot. This is a

Nomenclature General Scientific Bone ASBMR

Nomenclature General Scientific Bone ASBMR

Fractal dimension is an indicator of surface complexity of an

Fractal dimension is calculated using the Kolmogorov or “box counting”

Nomenclature General Scientific Bone ASBMR

Nomenclature General Scientific Bone ASBMR

Nomenclature General Scientific Bone ASBMR

Nomenclature General Scientific Bone ASBMR

Nomenclature General Scientific Bone ASBMR

Nomenclature General Scientific Bone ASBMR

Nomenclature General Scientific Bone ASBMR

Nomenclature General Scientific Bone ASBMR

Nomenclature General Scientific Bone ASBMR

Nomenclature General Scientific Bone ASBMR

The calculation of Euler number in 3D is based on software code kindly

Euler analysis provides a measure of connectivity density, indicating the

A2: 3D morphometric parameters calculated for all

Nomenclature General Scientific Bone ASBMR

Nomenclature General Scientific Bone ASBMR

Nomenclature General Scientific Bone ASBMR

The total volume of all pores within each discreet 3D object.