Validation of a measuring technique with computed tomography for cement penetration into trabecular bone underneath the tibial tray in total knee arthroplasty on a cadaver model

In total knee arthroplasty (TKA), up to ninety percent of the orthopaedic surgeons use cement to fixate the tibia component [1]. Despite this fixation, one of the main reasons for late failure of TKA is loosening of the tibia component [2]. This loosening is believed to be caused by micro motion at the cement-bone interface. The degree of fixation at the cement-bone interface, and hence the prevention of micro motion, is thought to depend on the penetration depth of cement into the trabecular bone underneath the tibial tray [3, 4]. In a post-mortem retrieval analysis of fixation strength of cemented tibial trays, Gebert de Uhlenbrock et al. concluded that fixation is improved by greater cement penetration [5]. Other studies on tibia cement penetration in TKA have shown that cement penetration between 3 and 5 mm beneath the tibial tray appears to be the optimal depth [6, 7]. Cement penetration to a depth of less than 3 mm results in a weak cement-bone interface, which predisposes for micro motion. However, penetration deeper than 5 mm can lead to heat-induced necrosis of bone and does not increase the strength of the interface [8, 9].

While these studies have demonstrated the importance of optimal cement penetration, measuring the cement mantle surrounding implants in vivo remains a challenge. In current clinical practice, cement penetration is measured in vivo by means of conventional radiographs [8]. However, the downside of conventional radiographs is that they merely provide a two dimensional view of the cement-bone interface. Moreover, it is difficult to shoot the X-ray beam exactly parallel to the prosthesis. Consequently, the measurement of the penetration depth will not always take place directly under the prosthesis, which is the place where an optimal interdigitation between bone and cement is required for fixation of the tibial tray when using the most common cementing technique, i.e. surface cementing.

Various authors used Computed Tomography (CT) imaging to examine the cement mantle around implants [5, 10‐14]. Goodheart et al. [11] and Mann et al. [13] used micro-CT image processing, a technique which cannot be used in vivo. CT images, on the other hand, can be used in vivo. The disadvantage of using CT images used to be that the metal prosthesis produced artefacts in the images, which made a good interpretation difficult. In 2009, Lui et al. [12] attempted to develop an algorithm for reducing metal artifacts in CT images of implanted metal orthopaedic devices. Since then, with the recent technical improvements of the CT equipment and software, this problem might have been solved, and CT images might have become a valid measuring technique for tibia cement penetration in TKA. We hypothesized that it might be possible to examine a cross section (i.e. a transversal CT slice) of the cancellous bone, and to assess the cement distribution on that cross section by calculating the percentage of the cement-penetrated cancellous bone versus the non-penetrated cancellous bone. By examining cross sections at consecutive depths, we might measure the width of the cement mantle as well as the density of the cement penetration at the required depth.

So far no measuring technique has been validated for measuring tibia cement penetration in TKA in vivo by means of CT imaging. Hence, the objective of this study was to develop and validate an in vivo measuring technique which provides a truthful indication of the percentage distributed and penetrated cement in the total scanned surface underneath a tibia prosthesis. We assessed the accuracy of using CT scan images to measure cement distribution, i.e. the amount of cement on a cross section of the bone, and cement penetration, i.e. the amount of cement in the depth of the bone just underneath the tibial tray, and we evaluated the influence of metal elements on the depiction of cement on the CT reconstruction.

For this study, we used 3 femoral heads which had been removed during total hip arthroplasty. All patients consented to the use of the redundant femoral heads. We prepared these femoral heads on the day of the resection. One femoral head was used to measure the radio-density range of pure PMMA cement and of cement-penetrated trabecular bone, as well as to validate the measured percentage of penetrated cement by comparing it to photographs of the femoral head slices sawed off at the required depth. The other two femoral heads were used as cadaver models for measuring cement penetration depth after cementing a tibia prosthesis into each of them, and one of these two femoral heads was used again for examining the influence of metal tibia prosthesis on the CT images.

The CT slices were analyzed with Osirix® (version 3.2.1 32-bit). OsiriX® is an image processing software dedicated to DICOM images produced by a CT scan and has been specifically designed for navigation and visualization of multimodality and multidimensional images. The photographs of the sawed-off slices were analysed with Adobe Photoshop® CS2 (version 9.0), a graphics editing program that can select an area of an image and sample it for future use.

Our findings confirm the assumptions of other studies which used CT-scans for femoral implants and defined the Hounsfield Units for (cancellous) bone, PMMA and prosthesis in the same region as we found with our study [10, 14].

Our results showed no overlap in the radio-density ROIs of trabecular bone, cement-penetrated trabecular bone and the prostheses. This indicates that when the radio-density ROI of cement-penetrated trabecular bone was registered, there was no trabecular bone or prosthesis in the areas under investigation. Thus it is possible to see the differences between trabecular bone, cement-penetrated trabecular bone or prosthesis in every CT-slice underneath the tibia prosthesis. This enables clinicians to get a good impression of the three-dimensional extent of cement penetration into the trabecular bone after cementing the tibia component of a TKA.

A comparison between the CT slices and the corresponding optical pictures showed different percentages of penetrated cement of the total surfaces of the transversal slices (Table 3). Nevertheless, the percentages of measured cement-penetrated trabecular bone in the CT slices of the specimen were within the range of percentages that could be expected based on the measurements with the photographs (p = 0.04). While the percentages of penetrated cement measured in the corresponding CT slices were higher than the percentages of penetrated cement measured in the photographs, the percentages of penetrated cement in the CT slices 0.8 mm beneath the corresponding CT slices were slightly lower, both 0.6%, than the percentages of penetrated cement measured in the photographs. The difference in percentage of penetrated cement of the total surface can be explained by the fact that the corresponding CT slices did not exactly correspond with the cut surface of the transversal slices. The corresponding CT slices were defined using the tantalum beads as an orientation point, but the CT scans made 0.8 mm slices, implying a possible fault margin of maximally 0.8 mm, which must have resulted in a different cement percentage of the surface. The CT slice of the area slightly above the cut surface of the slice will show a higher percentage of penetrated cement, because the cement in that area met with less resistance and could penetrate to a greater extent.

In conclusion, the trabecular cement penetration in TKA can be measured reliably by means of transversal CT slices in combination with the automatically marked radio-density ROI of cement-penetrated trabecular bone and Osirix® computer software. The metal prostheses had no significant effect on the measured radio-density values in the CT images, and there was no overlap between the automatically marked radio-density ROIs of trabecular bone and cement-penetrated trabecular bone. The percentages of cement penetration measured in the CT slices were within the range of percentages that could be expected based on the measurements in the optical pictures. This reproducible technique is, until now, the only way to measure the three-dimensional extent of cement penetration into the trabecular bone after cementing the tibia component of a TKA. Moreover, this technique will make it easier to compare the results of different cementing methods in TKA.