A
Dynamic CT–Based Quantification of Patellofemoral Contact Area Following Cartilage Removal: An Ex Vivo StudyManouAcke1✉EmailManou.Amaryllis.Acke@vub.be
Dr.
BenyameenKeelson1,2Prof. Dr.
ThierryScheerlinck4Nicolas
Van
Vlasselaer3Prof. Dr.
GertVan
Gompel1Prof. Dr.
ErikCattrysse3Prof. Dr.
NicoBuls11
A
Department of RadiologyVrije Universiteit Brussel (VUB), Universitair Ziekenhuis Brussel (UZ Brussel)Laarbeeklaan 1011090Jette, BrusselsBelgium2Department of Electronics and Informatics (ETRO)Vrije Universiteit Brussel (VUB)Pleinlaan 21050Etterbeek, BrusselsBelgium
3Department of Physical TherapyVrije Universiteit Brussel (VUB)Laarbeeklaan 1011090Jette, BrusselsBelgium
4Department Orthopaedics and TraumatologyVrije Universiteit Brussel (VUB), Universitair Ziekenhuis Brussel (UZ Brussel)Laarbeeklaan 1011090Jette, BrusselsBelgium
Authors: Manou Acke1*, Dr. Benyameen Keelson1,2, Prof. Dr. Thierry Scheerlinck4, Nicolas Van Vlasselaer3, Prof. Dr. Gert Van Gompel1, Prof. Dr. Erik Cattrysse3, Prof. Dr. Nico Buls1
1Department of Radiology, Vrije Universiteit Brussel (VUB), Universitair Ziekenhuis Brussel (UZ Brussel), Laarbeeklaan 101, Jette, 1090, Brussels, Belgium.
2Department of Electronics and Informatics (ETRO), Vrije Universiteit Brussel (VUB), Pleinlaan 2, Etterbeek, 1050, Brussels, Belgium.
3Department of Physical Therapy, Vrije Universiteit Brussel (VUB), Laarbeeklaan 101, Jette, 1090, Brussels, Belgium.
4Department Orthopaedics and Traumatology, Vrije Universiteit Brussel (VUB), Universitair Ziekenhuis Brussel (UZ Brussel), Laarbeeklaan 101, Jette, 1090, Brussels, Belgium.
* e-mail of corresponding author: Manou.Amaryllis.Acke@vub.be
A
Abstract
Purpose
Knee osteoarthritis (OA) is a musculoskeletal disorder characterised by abnormal joint motion, joint space narrowing, and cartilage degeneration. Early diagnosis is important to slow disease progression. Dynamic CT imaging adds three-dimensional functional information to an otherwise static diagnosis. This study investigates the diagnostic potential of inter-articular contact area during knee motion across three stages of cartilage degeneration.
Methods
A
Cartilage was progressively removed from the patellofemoral joint of a cadaveric knee to simulate stages of cartilage degeneration: first laterally, then at the crista, and finally medially. After each intervention, dynamic CT was performed during induced cyclic flexion-extension motion of the cadaveric knee. We quantified contact area dynamically by measuring distances between intact cartilage layers of the patella and femur. Contact areas across stages (baseline, lateral, crista, medial) were compared using the Wilcoxon Signed Rank test at six knee flexion angles (0°, 5°, 10°, 15°, 20°, 25°). Reproducibility was assessed by repeating the dynamic CT acquisition post-cartilage removal three times.Results
Contact area increased significantly at each stage of cartilage degeneration compared to the previous stage (p < 0.05), progressing from 0 mm² to 84.2 mm², 176.0 mm², and 308.7 mm². For reproducibility, the standard error of the mean across three repetitions showed an average deviation of 6%.
Differences varied across knee angles, with the least distinction observed in full extension. Spatially, the contact areas corresponded anatomically to the regions of cartilage removal.
Conclusion
Progressive cartilage removal in a cadaveric knee results in a significant increase in contact area.
A
Introduction
Knee osteoarthritis (OA) is a common musculoskeletal disorder, affecting 16% of the global population [1]. The prevalence of these disorders is expected to rise further due to the aging global population, underscoring the need for effective prevention, diagnosis, and treatment strategies. Early diagnosis can slow disease progression and enhance patients' quality of life [11]. Knee osteoarthritis is caused by uneven pressure distribution and abnormal joint movement. Over time, these factors lead to cartilage thinning, joint space narrowing, and increased bone-on-bone friction [1, 16]. Quantifying the extent of these structural and kinematic changes can provide an early diagnosis for osteoarthritis.
A range of assessment methods have been introduced for the diagnosis and quantification of knee OA, broadly categorized into image-based and non-image-based approaches.
Non-image-based methods include electromyography (EMG) for assessing muscle activity and motion capture systems for evaluating joint kinematics [4, 19]. However, these approaches are often limited by inaccuracy or impracticality [14, 22]. Invasive marker-based techniques are more accurate but unsuitable for clinical application [14, 22].
Image-based techniques, such as radiography, ultrasound, MRI, fluoroscopy, and CT, are routinely employed to diagnose musculoskeletal disorders (MSDs) by investigating structural changes in the bones, muscles, and surrounding soft tissues [11].
These methods can also quantify metrics, including cartilage thickness and surface-to-surface distance, which measures the gap between opposing cartilage or bone surfaces during motion. 3D imaging can identify contact areas, surface areas where cartilage layers are in contact. Increased contact area indicates potential cartilage degeneration. CT imaging offers superior spatial resolution for bone and fast acquisition times. However, unlike MRI, CT imaging cannot reliably differentiate cartilage from joint space fluid. To address this, studies either use iodine contrast to visualise cartilage [15] or assume a constant cartilage thickness [8, 13]. Heatmaps can be used to visualize spatial variations in cartilage thickness and inter-bone distances, aiding interpretation of these degenerative changes [2, 19].
Most clinical imaging relies on either static 3D CT or MRI imaging, or dynamic 2D fluoroscopy imaging. While the former lacks kinematic information, the latter is limited by anatomical superposition and the lack of depth information. Currently, no clinical imaging modality can simultaneously assess joint morphology and function in three dimensions over time.
Dynamic CT or time-resolved CT imaging was initially developed for cardiopulmonary applications, but gained recognition in musculoskeletal research for its ability to capture real-time joint motion with high spatial and temporal resolution [3, 4, 9, 23]. It enables non-invasive, high-speed imaging to quantify joint kinematics and morphological changes during motion [16, 23]. Dynamic CT allows simultaneous assessment of morphology and motion in 4D [16, 23].
This study assessed induced cartilage removal in a cadaveric knee phantom by surgically removing cartilage in three progressive steps. Changes in contact area were evaluated using dynamic CT during joint flexion and extension. To estimate contact areas, a CT-based, model-driven approach was introduced, simulating healthy cartilage on both patella and femur and computing the inter-surface distance.
Materials and methods
Experimental set-up
A Thiel-embalmed cadaveric left leg was used for this study to enable the preservation of both bone and soft tissue structures. The preservation of the soft tissue envelope allows to represent biomechanics and cartilage degeneration of the human knee.
The study was performed in accordance with the ethical standards as laid down in the 1964 Declaration of Helsinki and its later amendments or comparable ethical standards. Ethical approval for the study was granted by the ethics committee of VUB–UZ Brussels University Hospital (B.U.N. 1432023000100).
Cartilage was sequentially removed by an expert orthopaedic surgeon using a milling cutter and evaluated across four conditions: a baseline with no intervention, followed by progressive damage at the lateral facet, crista, and medial facet. Final removal at the medial facet resulted in total cartilage loss within the patellofemoral joint.
Dynamic CT imaging captured each condition of the knee during a cyclic flexion-extension motion to generate proximity maps showing changes in intra-articular contact areas during motion. Reproducibility was evaluated by repeating the CT acquisitions three times. The baseline and crista scenarios are illustrated on Fig. 1.
Fig. 1
(a) The anatomy that will be imaged for the baseline CT acquisition with cartilage still fully intact. (b) The anatomy with cartilage removed at the lateral facet and crista.
A wooden frame was foreseen to manipulate the leg in a cyclic flexion-extension while remaining in the CT scanner’s field of view (Fig. 2).
Fig. 2
The knee is manipulated in a linear vertical motion while the ankle and hip can freely move along a linear railing system. This allows the knee to remain in the CT scanner’s field of view (red).
Dynamic CT images were obtained while the leg underwent cyclic flexion-extension motion after each intervention stage. The scans were acquired using a 256-slice wide-beam CT (GE Revolution Healthcare) with 16 cm z-coverage, a 50 cm field of view, tube voltage of 80 kV, tube current of 452 mA, slice thickness of 0.625 mm, and pixel size of 0.98 mm.
Reproducibility was assessed by repeating the dynamic CT acquisition three times after medial cartilage removal and calculating the standard error of the mean.
Contact area is defined as the distance between two opposing healthy cartilage layer models. Prior to the cartilage removal, these models were derived from a contrast-enhanced CT scan after injecting 10 ml of contrast agent (Iodixanol 320, Visipaque) into the healthy cadaveric knee, allowing visualization of intact cartilage (Fig. 3).
Fig. 3
The cartilage layer was segmented on both the Patella (red) and Femur (green). This segmentation will be used to calculate the distance between the two cartilage layers and to visualize the areas where the cartilage layers are in contact.
Image processing
The femur and patella were segmented and registered using a multi-atlas approach with SimpleITK-Elastix, following a previously published in-house image processing pipeline for joint motion evaluation [12]. Surface proximity maps of the joint were generated to visualize the intra-articular joint space using the Pyvista package in Python.
Proximity maps visualize contact area by calculating the distance between intact patellar and femoral cartilage layers. A cartilage-to-cartilage distance of zero indicates initial cartilage contact; a positive distance reflects a normal joint space without contact, and a negative value suggests active cartilage loss.
A one-tailed Wilcoxon signed-rank test was used to assess whether contact area increased between the different steps of cartilage degeneration.
The Wilcoxon Signed Rank test was used to compare contact areas at six knee angles (0°, 5°, 10°, 15°, 20°, 25°) to assess whether dynamic CT-derived proximity maps increase significantly between healthy and increasing levels of degenerated cartilage (baseline, lateral, crista, medial).
Results
The cartilage thicknesses of intact patellar and femoral cartilage layers of the cadaveric knee are visualised on Fig. 4.
Fig. 4
Cartilage thickness distribution of the femoral and patellar cartilage layers.
The mean (± standard deviation) cartilage thickness was 1.82 ± 0.53 mm for the femur and 1.98 ± 0.68 mm for the patella, adding up to a combined cartilage thickness of 3.8 ± 0.86 mm in the patellofemoral joint.
The resulting proximity maps of contact areas with healthy cartilage layer models are visualised in Fig. 5, where intra-articular contact areas are displayed as a heatmap on the patellar cartilage layer, which is shown in white, while bone structures are rendered in grey. These visualisations offer spatial context to the intra-articular contact areas shown as bar plots in Fig. 6a.
For the baseline, condition, no intra-articular contact area is observed. Following cartilage removal at the lateral facet, contact areas emerge at each flexion angle, localised to the lateral side of the patella, with a mean of 84.2 mm2 (p = 0.016). The crista intervention further increases the contact area to a mean of 176.0 mm2 (109.0% increase, p = 0.031), which also spatially extends into the central region. After complete cartilage removal, including the medial side, the contact area becomes more widespread with an average area of 308.7 mm2 (75.4% increase, p = 0.016), with the medial aspect most affected at 25 degrees of flexion.
The repeated measurements for the medial condition exhibit consistent results, as illustrated in Fig. 6b. The confidence interval shown in Fig. 6a, based on the standard error of the mean from three trials, shows an average deviation of 6%.
At 0° flexion, all four conditions show minimal and indistinguishable contact area variation.
Fig. 5
Proximity maps illustrate contact area in function of flexion angle across increasing levels of cartilage degeneration (baseline, lateral, crista, medial). The inter-cartilage distances within this contact area are represented as a heatmap.
Fig. 6
Contact area was calculated in function of flexion angle across increasing levels of cartilage degeneration (baseline, lateral, crista, medial). To assess reproducibility, dynamic CT acquisition was repeated three times after medial cartilage removal; the resulting variability is shown as error bars.
discussion
This study assessed cartilage contact area in a cadaveric knee across three stages of cartilage removal by intervention. After each intervention, dynamic CT captured knee motion, and proximity maps visualized contact area during 25° dynamic flexion-extension motion. This angle range was constrained to 25° by the physical limitations of the CT gantry. Contact area was computed from contrast-enhanced CT by extracting healthy cartilage layers and computing the area of contact between these two layers.
The resulting area showed a clear trend across all experimental conditions: contact area increased with progressive cartilage removal. In the baseline (healthy) condition, no contact was observed at any flexion angle, consistent with the known absence of cartilage contact in a healthy knee. At full extension (0° angle), contact areas were similar across conditions. This suggests that static scans at this angle may not capture relevant differences between pathological states, reinforcing the value of dynamic CT for detecting kinematic changes.
Our method also showed spatial correspondence. Contact appeared at the lateral facet following lateral cartilage removal, expanded to the central region with crista removal, and extended medially after medial removal. This indicates that contact area maps could potentially provide spatial information about the location of cartilage thinning.
In literature, contact areas can also be defined by calculating bone-to-bone distances and assuming a constant cartilage thickness [6, 21]. In this study, we found a combined thickness (patella and femur) of 3.80 ± 0.86 mm, indicating some thickness variation across the articulating surfaces. Reported average combined patellofemoral cartilage thickness varies from an average of 5.5 mm for adults [5] to 9.0 mm for adolescents [18]. Robinson et. al. [17] found that cartilage stiffness decreases with OA severity, potentially explaining increased contact area. This underscores dynamic CT’s value in assessing cartilage contact during motion. Further research is needed to compare fixed threshold methods with calculations that account for anatomically realistic variations in cartilage thickness.
The increase in cartilage contact area with progressive degeneration aligns with other findings in literature. Eckstein et al. [7] reported increasing total subchondral bone areas (tAB) using static MRI imaging, but only for later stages of OA. Turmezei et al. [20] found a significant reduction in joint space width on static CT imaging, even for early stages of OA. In the wrist, D’Agostino et al. [6] observed a similar trend, where later stage arthritic patients exhibited greater contact areas and joint space narrowing compared to controls. In the first MTP joint, similar patterns were noted by Jones et al. [10], where hallux rigidus (HR) led to increased joint coverage on static CT imaging.
This study has following limitations. It was an ex-vivo study based on a single specimen, which may affect the broader applicability of the results. Cartilage removal was done using a milling cutter, which might not fully replicate the characteristics of natural cartilage degeneration. Additionally, the analysis was limited to flexion angles up to 25 degrees due to constraints of the CT gantry.
conclusion
This study demonstrated that progressive cartilage degeneration in a cadaveric knee specimen leads to a significant increase in contact area. Contact area varies during motion and is least distinguishable in full extension, highlighting the value for dynamic imaging. These findings underscore the potential of dynamic CT-based contact area analysis for assessing cartilage health. Future research should further investigate the diagnostic value of contact area for early osteoarthritis detection and its capacity to provide spatial context to cartilage degeneration.
A
A
Author Contribution
M.A. did project development, data collection, data analysis and manuscript writingB.K. supervised data analysis T.S. did project development, performed the surgical interventionN.V.V. provided technical support for the device in Figure 2G.V.G. did project development, data management, manuscript editingE.C. did project development and assisted surgical interventionN.B. did project development, manuscript editingAll authors reviewed the final manuscript.
Dr. Benyameen Keelson supervised data analysis
Prof. Dr. Thierry Scheerlinck project development, performed the surgical intervention
Nicolas Van Vlasselaer provided technical support for the device in Fig. 2
Prof. Dr. Gert Van Gompel did project development, data management, manuscript editing
Prof. Dr. Erik Cattrysse did project development and assisted surgical intervention
Prof. Dr. Nico Buls project development, manuscript editing
All authors reviewed the final manuscript.
A
Funding information -
not applicable
A
Data Availability
Due to ethical regulations related to the use of human cadaveric material, the data supporting this study are not openly available. However, they can be obtained from the corresponding author upon reasonable request.
References
1.
Adachi, T. et al. Accuracy Verification of Four-Dimensional CT Analysis of Knee Joint Movements: A Pilot Study Using a Knee Joint Model and Motion-Capture System. Cureus 15(2). p. e35616. https://doi.org/10.7759/cureus.35616
2.
Ateshian, G. A. et al. Quantitation of articular surface topography and cartilage thickness in knee joints using stereophotogrammetry. J. Biomech. 24 (8), 761–776. https://doi.org/10.1016/0021-9290(91)90340-S (Jan. 1991).
3.
Blum, A., Nov. & et al 3D reconstructions, 4D imaging and postprocessing with CT in musculoskeletal disorders: Past, present and future. Diagn. Interv Imaging. 101 (11), 693–705. https://doi.org/10.1016/j.diii.2020.09.008 (2020).
4.
Buzzatti, L. et al. Dynamic CT scanning of the knee: Combining weight bearing with real-time motion acquisition. Knee 44, 130–141. https://doi.org/10.1016/j.knee.2023.07.014 (Oct. 2023).
5.
Cohen, Z. A., Aug. & et al Templates of the cartilage layers of the patellofemoral joint and their use in the assessment of osteoarthritic cartilage damage. Osteoarthr. Cartil. 11 (8), 569–579. https://doi.org/10.1016/S1063-4584(03)00091-8 (2003).
6.
D’Agostino, P. et al. In vivo biomechanical behavior of the trapeziometacarpal joint in healthy and osteoarthritic subjects. Clin. Biomech. 49, 119–127. https://doi.org/10.1016/j.clinbiomech.2017.09.006 (Nov. 2017).
7.
Eckstein, F. et al. (eds). Magnitude and regional distribution of cartilage loss associated with grades of joint space narrowing in radiographic osteoarthritis – data from the Osteoarthritis Initiative (OAI). Osteoarthritis Cartilage. 18(6). pp. 760–768. Jun. (2010). https://doi.org/10.1016/j.joca.2009.12.009
8.
Foumani, M. et al. In-vivo dynamic and static three-dimensional joint space distance maps for assessment of cartilage thickness in the radiocarpal joint. Clin. Biomech. 28 (2), 151–156. https://doi.org/10.1016/j.clinbiomech.2012.11.005 (Feb. 2013).
9.
Gondim Teixeira, P. A. et al. Evidence-based recommendations for musculoskeletal kinematic 4D-CT studies using wide area-detector scanners: a phantom study with cadaveric correlation. Eur. Radiol. 27 (2), 437–446. https://doi.org/10.1007/s00330-016-4362-y (Feb. 2017).
10.
Jones, M. T., Sep. & et al Quantification of First Metatarsal Joint Surface Interactions in Hallux Rigidus Using Distance and Coverage Mapping: A Case-Control Study. Foot Ankle Int. 45 (9), 1038–1046. https://doi.org/10.1177/10711007241258447 (2024).
11.
Kalia, V. et al. Functional Joint Imaging Using 256-MDCT: Technical Feasibility. Am. J. Roentgenol. 192 (6), W295–W299. https://doi.org/10.2214/AJR.08.1793 (Jun. 2009).
12.
Keelson, B. et al. Nov.. Automated Motion Analysis of Bony Joint Structures from Dynamic Computer Tomography Images: A Multi-Atlas Approach. Diagnostics. 11(11). pp. Art. no. 11. (2021). https://doi.org/10.3390/diagnostics11112062
13.
Kupper, J. C. & Jul. (. Dynamic In-Vivo Knee Cartilage Contact With Aging. Unpublished/Online. (2020). http://hdl.handle.net/1880/115477
14.
Lam, W. W. T. et al. A systematic review of the applications of markerless motion capture (MMC) technology for clinical measurement in rehabilitation. J. Neuroeng. Rehabil. 20 (1), 57. https://doi.org/10.1186/s12984-023-01186-9 (May 2023).
15.
Myller, K. A. H. et al. Computational evaluation of altered biomechanics related to articular cartilage lesions observed in vivo. J. Orthop. Res. 37 (5), 1042–1051. https://doi.org/10.1002/jor.24273 (2019).
16.
Oka, Y., Dec. & et al Impact of Controlling Abnormal Joint Movement on the Effectiveness of Subsequent Exercise Intervention in Mouse Models of Early Knee Osteoarthritis. Cartilage 13 (2 Suppl), 1334S–1344S. https://doi.org/10.1177/1947603519885007 (2021).
17.
Robinson, D. L. et al. Mechanical properties of normal and osteoarthritic human articular cartilage. J. Mech. Behav. Biomed. Mater. 61. 96–109. https://doi.org/10.1016/j.jmbbm.2016.01.015 (Aug. 2016).
18.
Sidharthan, S., Apr. & et al Patterns of Articular Cartilage Thickness in Pediatric and Adolescent Knees: A Magnetic Resonance Imaging–Based Study. Arthrosc. Sports Med. Rehabil. 3 (2), e381–e390. https://doi.org/10.1016/j.asmr.2020.09.029 (2021).
19.
Siegler, S. et al. Jul.. Analysis of surface-to-surface distance mapping during three-dimensional motion at the ankle and subtalar joints. J Biomech. 76. pp. 204–211. (2018). https://doi.org/10.1016/j.jbiomech.2018.05.026
20.
Turmezei, T. D. et al. Quantitative Three-dimensional Assessment of Knee Joint Space Width from Weight-bearing CT. Radiology 299 (3), 649–659. https://doi.org/10.1148/radiol.2021203928 (Jun. 2021).
21.
Van Royen, K. et al. Automated analysis of trapeziometacarpal joint kinematics using four-dimensional computed tomography. J. Hand Surg. Eur. 49 (9), 1095–1103. https://doi.org/10.1177/17531934241229948 (Oct. 2024).
22.
Wade, L. et al. Applications and limitations of current markerless motion capture methods for clinical gait biomechanics. PeerJ e12995. https://doi.org/10.7717/peerj.12995 (Feb. 2022). 10.
23.
Wong, M. T. et al. Four-dimensional computed tomography: musculoskeletal applications. Can. J. Surg. 65 (3), E388–E393. https://doi.org/10.1503/cjs.023420 (Jun. 2022).