Quantitative Topographic Anatomy of the
Femoral ACL Footprint – A Micro-CT Analysis
D.G. Norman1* A. Getgood2 J. Bird3 T. Spalding4 & M.A. Williams5
1
School of Engineering, University of Warwick, 2University of Western Ontario, 3University Hospital Lewisham, 4University Hospitals Coventry and Warwickshire NHS Trust, 5WMG
INTRODUCTION
Anterior Cruciate Ligament (ACL) Reconstruction surgery is the sixth most common orthopaedic surgery in
the USA and as a result has become the most studied musculoskeletal system1. Despite this, full recovery
after surgery is unsatisfactory with only ≈70% of patients regaining full knee mobility2. Attempting to solve
this, clinicians have investigated surgical methods that reconstruct the knee’s original anatomy after
rupture3. To accomplish this, physicians need to determine where to attach the newly grafted ligament to
the Lateral Intercondylar wall - the ACL footprint. The approach is obviously to place the new ligament in
the same location as the old one but this can’t be identified after an ACL rupture so other landmarks are
needed. According to Wolff’s Law, there should be thickening
(and hence a ridge) around the original ligament attachment area .
These ridges, when visible, are often used to place the ligament
anatomically. Nevertheless, these ridges, the Lateral Intercondylar
Ridge and the Lateral Bifurcate Ridge, (locations illustrated in a
diagram published by Ferretti et al 2007) are often not visible and
controversy over their actual existence has sparked1. The aim of
this project was to use micro-CT to investigate the existence of
these ridges to a degree of resolution not explored yet in the literature and validate,
or otherwise, their existence. To ensure objectivity, the surface topology was extracted
at a high resolution and then used to create relief maps along with other visual methods.
Seven cadaveric human
knees (66-82 years old)
were micro-CT scanned
at 60μm and then
reconstructed using
Nikon’s proprietary
software before being
imported into VGStudio
Max where bone was
isolated from tissue
SUMMARY
Understanding the anatomy of the femoral anterior cruciate ligament (ACL) footprint is important to allow correct
placement of the ACL graft. The lateral intercondylar and bifurcate ridges have been proposed as bony
landmarks to guide tunnel placement, however, it has been proposed that these ridges can be difficult to find
arthroscopically and may be an inconsistent landmark. A new technique utilising high resolution micro CT is
described, allowing detailed 3-dimensional quantitative analysis of the femoral footprint of the ACL. Using these
3D images, a qualitative assessment of the landmarks was undertaken by a group of experienced ACL surgeons
to determine whether these landmarks can be dependably used to guide correct femoral tunnel placement. Seven
cadaveric knees where run through a high-resolution micro-CT scanner following arthroscopic localisation of the
centre of the ACL footprint. The images were manipulated using a new extraction technique, leaving the detail of
the footprint intact, allowing quantitative analysis of the footprint. These images were displayed on an immersive
3D visualisation wall where experienced ACL surgeons were asked to comment on the presence and morphology
of the osseous landmarks. Surface models were created of the distal femur with detailed anatomy of the femoral
ACL footprint found to be within 70 microns. A significant variability of the presence and morphology of the
osseous landmarks was found within the sample. Qualitative assessment of the ACL footprint also showed
significant variability (lateral intercondylar ridge and bifurcate ridge; Fleiss’ Kappa 0.01 and 0.07 respectively).
This is the first study to demonstrate the utilisation of micro-CT in the study of the ACL footprint. A novel method
of extracting large volumes of data to usable files has been described. Within the study sample, significant
variability was observed in the quantitative and qualitative assessment of the osseous landmarks of the femoral
ACL footprint. This study has demonstrated that there is a variability in the presence of these landmarks,
.therefore other methods to guide placement should be also be used so as to reproducibly place the femoral
- _tunnel in the desired position.
METHODOLOGY
Due to the size of the data (≈40Gb) it was difficult to carry out any
analysis. Therefore surface models of the data were created whilst
maintaining the resolution. This was done in two processing
streams A and B (see diagram) with ‘A’ being the ACL footprint and
‘B’ being the rest of the femur. Volumes were extracted separately
as mesh data and then wrapped using a ‘shrink-wrap function’ to
extract a closed surface. Finally, the two surfaces are optimised
and merged together to create the finished surface model.
The surface models were then rendered with a
bony texture and were displayed on an
immersive 3D visualisation wall. Then ten
expert knee surgeons were asked to rate the
presences of both the ridges (e.g. “Is the
lateral Intercondylar ridge present in this
knee?”) on each of the 7 models using a 5
point Likert scale. Analysis of agreement
(using Kappa statistics) was then conducted
Validation of the surface extraction method was
done by laser scanning a sample femur (baseline
scan) and then micro-CT scanning the same femur.
The micro-CT scanned femur then underwent the
developed surface extraction method. The two
scanned surfaces were then compared using a
deviation analysis which gives a coloured
representation of the differences between the ‘actual’
and ‘processed surface’. This allows us to quantify
the error of the developed method.
To investigate the morphology of the ACL
footprint the surface was turned into a
relief map. To accomplish this the
surface of the ACL footprint is duplicated
and smoothed flat to create a baseline
surface. The baseline reference surface
is overlaid and matched to the original
surface. Finally, a coloured deviation
analysis is run, which coloured the
original surface based on its distance
from the corresponding baseline surface.
This creates relative relief maps
representing the output data.
RESULTS
Relief maps of the ACL footprint allow an objective
approach to identifying the presence of osseous landmarks.
On observing all of the relief maps of the ACL footprint, a
significant variability of the presence and morphology of the
osseous landmarks existed within this sample indicating
that they can’t be used reliably for tunnel placement
Data size reduced
from 40Gb to 20Mb
Surface accurate
to 70μm±70μm
These models developed from
micro-CT data were found to be
accurate to 70μm±70μm even
though the data size was
reduced 2000x! Applications for
the finished surface models
developed as part of this
method include; 3D visualization, 3D Printing, Quantitative
analysis, Finite Element Analysis, CAD
compatibility for
medical device design etc.
The presence of the
osseous landmarks in this
sample were identified
from both the relief maps
and our surgeons
responses and compared
with previous findings in
the literature. Our results
disfavour the current
consensus
Various Statistical methods used to calculate the
agreement between our participants yet all
showed that there was no statistical agreement
between them! Both graphs showing surgeons
ratings of the presence of each osseous
landmarks also show a range of responses
indicating little agreement.
Surgeons didn’t agree on the presents of the
osseous landmarks in our sample of knees!
Surface
models
could be
visualised
on a 3D
powerwall
3 Inter-rater agreement
tests, for multiply raters,
giving the coefficients
for ‘fair’ agreement for
10 raters for all of ACL
footprints examined for
both osseous
landmarks.
Individual ratings, mean and standard deviation for participants ratings between Absolutely
Yes’ to ‘Absolutely No’ for the presence of the lateral Intercondylar ridge in each knee.
Individual ratings, mean and standard deviation for participants ratings between Absolutely
Yes’ to ‘Absolutely No’ for the presence of the lateral bifurcate ridge in each knee.
3D printing
could now be
done using the
developed
surface models
CONCLUSIONS
This present study has described a technique for quantitatively analysing the femoral ACL
footprint anatomy, utilising micro-CT imaging. This method been shown to be accurate and could
be used to investigate anatomical features of many regions of interest of the musculoskeletal
system. Quantitative and qualitative assessment of the femoral footprint showed significant
variability of the presence and orientation of osseous landmarks within the subjects in this study.
This does question whether these landmarks can be accurately utilised to guide surgical
placement of the femoral tunnel in ACL reconstruction. However, further work must be done to
investigate age, gender and race variation in a larger sample population before firm conclusions
can be made. We recommend that during ACL reconstruction, accurate femoral tunnel placement
should be guided by a range of previously described techniques, to get the optimal result.
1
REFERENCES
Ziegler, C.G et al., 2010. Arthroscopically Pertinent Landmarks for Tunnel Positioning in Single Bundle and Double Bundle Anterior Cruciate Ligament Reconstruction. The American Journal of Sports Medicine, 39(6) pp.743-5
2
Shafizadeh S, Huber H, Grote S, Hoeher J, Paffrath T, Tiling T et al (2005) Principles of fluoroscopic-based navigation in ACL reconstruction. Oper Tech Orthop
15:70–75
3
Van Eck, F.C., 2010. Does the lateral intercondylar ridge disappear in ACL deficient patients? KSSTA, 18, pp.1184-1188
4
Ferretti, M. et al., 2007. Osseous Landmarks of the Femoral Attachment of the ACL: An Anatomic Study. Arthroscopy: Arthroscopic. 23(11),
5
Farrow, L.D. et al., 2007. Morphology of the Femoral Intercondylar Notch. Journal of Bone and Joint Surgery, 89, pp. 2150–2155 pp.1218-25
6
Huchinson, M.R., Ash, S.A., 2003. Resident’s Ridge: assessing the cortical thickness of the lateral wall and roof of the intercondylar notch. Arthroscopic.
FURTHER WORK
Secured funding from
1. Consider Cortical thickening
2. Quantify and map thickening under ligament representing
graft forces
3. Further analysis of osseous landmarks
4. Validate ruler technique using multimodality imaging
techniques
5. Compare NHS and Micro-CT scans
ACKNOWLEDGEMENTS
Thanks to Manpreet Dhillon, Pete Thompson, Ercihan Kiraci,
Abdul-Hadi Abulrub and John Thornby for their assistance. Also
thanks to the WMG internship for their financial support. This
work has been presented at the BASK, the ESSKA and the
_BCUR conference.
Vs