Cement mantle morphology in TKA
+1Gebert de Uhlenbrock, A; 1Püschel, V; 1Bishop, N; 1Sellenschloh, K; 2van Stevendaal, U; 1Morlock, M M
+1 TUHH Hamburg University of Technology, Germany 2 Philips Research Europe, Hamburg, Germany
a.uhlenbrock@tuhh.de
METHODS:
Eleven human knee joints treated with total knee arthroplasty (TKA)
were harvested during autopsy (mean time in situ 4.2years (0-8years),
mean age 77±3years) and CT-scanned (Philips, voxel size 0.15mm ×
0.15mm × 0.4 mm) with a calibrated phantom both with and without the
implant in place (Figure 1A). The extraction of the implant was realized
by heating (<80°C) during a constant pull-out force of 150N, leaving the
cement layer intact (Figure 1B). The scans were filtered (Philips),
reconstructed 3-dimensionally based on the Hounsfield units (HU) and
were aligned along the stem axis (Avizo 5.1; Figure 1C).
Four specimens with titanium implants were used to determine the
accuracy of the method and to choose a suitable threshold, enabling
segmentation between bone and cement. The threshold between cement
and implant was set based on the known volume of each implant. These
specimens were cut, using a diamond-coated bandsaw, twice
transversely through the stem and once in the frontal plane 5mm
posterior to the stem axis. The cement thickness was measured at
various distinct positions in each section and compared with the cement
thickness in the reconstructed CTs for varying thresholds (400HU,
600HU and 800HU, n=52, Figure 1D).
Intact cement layer
Implant
Cement
Retrieved
implant
A
lateral
Phantom
z
B
Destinctiv thickness
medial
Stem thickness
x
C
HU 800
HU 600
HU 400
D
Figure 1: (A) CT scan of the bone with implant and phantom; (B)
implant after pull-out; (C) reconstruction of the aligned implant and
cement mantle; (D) bone cement section and the corresponding
segmented cement mantle for various HU.
All 11 specimens were then used to quantify the cement mantle in
relation to the mean bone mineral density (BMD, SPSS15, α<0.05).
Three of these specimens were cobalt-chromium implants, which had to
be evaluated without the implant due to metallic artifacts in the CTs. The
cement thickness and the interdigitation beneath the tibial tray, as well
as the cement volume and the total cement thickness around the stem
were determined every 5mm along the stem for four regions
(MatlabR14; Figure 2). The boundary between cement layer thickness
and cement interdigitation was defined by the cut bony plateau of the
tibia. Interdigitation was evaluated as far as 10mm below this boundary.
Cement layer thickness [mm]
Cement interdigitation [mm]
3.5
0
-1
-2
anterior
3.0
anterior
-3
2.5
medial
-4
lateral 2.0 medial
lateral
-5
-6
1.5
-7
-8
1.0
posterior
0.5
posterior
-9
-10
-11
0.0
Figure 2: Cement layer thickness and interdigitation distributions
beneath the tibial tray for a representative specimen.
RESULTS:
The accuracy of the cement mantle thickness estimation by CT
ranged from 10-20% for 400 and 600HU (Figure 3). The threshold used
for further investigation was 600HU.
deviation of the
measured thickness [%]
INTRODUCTION:
Cement anchorage is the “gold standard” in fixation of total knee
arthoplasty (TKA) implants. The quality and morphology of the cement
mantle have a large influence on the long-term clinical performance [1].
Defects in the cement or regions of thin cement mantle compromise the
strength of fixation. X-ray is so far the standard tool to assess the cement
mantle in situ. However, x-ray images can only be used to describe
radiolucent lines and are not suitable for assessing the shape, volume
and thickness of the cement mantle. With the use of polymeric replicas
of implants the cement mantle morphology can be assessed 3-dimensionally based on CT scans, which has been investigated for hips [2].
In this study a methodology assessing the cement morphology for
titanium TKA implants in situ was developed and its accuracy
determined. This method was then used to assess the cement mantle
morphology of functional TKA in relation to the bone quality.
100
HU 800
HU 600
HU 400
80
60
40
20
0
0-3
3-5
5-10
>10
measured thickness ranges [mm]
Figure 3: Accuracy of the measured cement thickness for the
investigated thresholds (mean and standard deviation, n=52).
The mean total cement volume was 14.6±6.7ml, of which 40% was
comprised by the cement layer below the tibial tray and 23%
interdigitation of up to 10mm below the bony plateau. The remaining
37% comprise the cement mantle surrounding the stem. The median
cement layer thickness below the tray was 1.9±0.4mm and the median
interdigitation was 1.5±0.3mm, ranging from 1.1 to 1.8mm. The cement
layer thickness below the tray deviated by more than 0.5mm from the
lateral to the medial side in 4/11 samples. At least 75% of each tibial
plateau area had an interdigitation greater than 0.5mm. The mean
cement mantle thickness around the stem was 2.7mm, and was greatest
for the most proximal and most distal sections in 10/11 samples
(~3.1mm).
Mean BMD was 99±30mg/cm³. There was no correlation between
BMD and the parameters investigated (p>0.05).
DISCUSSION:
The evaluation of the cement mantle morphology with an implant in
situ has been inadequate so far, due to artifacts introduced in the CT
images by scatter around the metal. In this study the CTs were filtered
using an exclusive technique (Philips) to eliminate such artifacts around
titanium implants. This method was demonstrated to be suitable for the
evaluation of the cement mantle to ~20%, which exceeds criteria
proposed by others [3]. This allowed a preliminary assessment of
retrieved samples.
All the retrieved specimens showed an interdigitation of at least
0.5mm (median 1.5mm) across the whole plateau, which may be
sufficient for good fixation, since all samples were well-functioning. It is
unclear whether the low BMD values already existed pre-implantation or
whether they were due to the development of osteolysis. The unexpected
lack of an inverse relationship between interdigitation and BMD may
suggest that the bone quality decreased over time. The tilt in the cement
layer thickness below the tray appeared to be sufficiently low to allow
proper function but still could represent some malpositioning.
The method developed could be used to assess the mechanical
integrity of the cement mantle in vivo and such in the evaluation of
cementing techniques or implant designs. Unfortunately, cobalt-chrome
implants still produce too many artifacts to be included in such studies.
REFERENCES:
[1] Hofmann et al., JoA 2006 21: 353-7; [2] Scheerlinck et al., JOR 2005
23: 698-704. [3] Scheerlinck et al., JBJS 2006 88-B:19-25.
Poster No. 2099 • 56th Annual Meeting of the Orthopaedic Research Society