Upright CT of the knee: the effect of weight-bearing on joint alignment
- 462 Downloads
To prospectively compare patellofemoral and femorotibial alignment in supine non-weight-bearing computed tomography (NWBCT) and upright weight-bearing CT (WBCT) and assess the differences in joint alignment.
NWBCT and WBCT images of the knee were obtained in 26 patients (mean age, 57.0 ± 15.9 years; range, 21-81) using multiple detector CT for NWBCT and cone-beam extremity CT for WBCT. Two musculoskeletal radiologists independently quantified joint alignment by measuring femorotibial rotation, tibial tuberosity-trochlear groove distance (TTTG), lateral patellar tilt angle, lateral patellar shift, and medial and lateral femorotibial joint space widths. Significant differences between NWBCT and WBCT were sought using Wilcoxon signed-rank test (P-value < 0.05).
Significant differences were found for femorotibial rotation (the NWBCT mean changed from 2.7° ± 5.1 (reader 1)/2.6° ± 5.6 (reader 2) external rotation to WBCT 0.4° ± 7.7/0.2° ± 7.5 internal rotation; P = 0.009/P = 0.004), TTTG (decrease from NWBCT (13.8 mm ± 5.1/13.9 mm ± 3.9) to WBCT (10.5 mm ± 5.0/10.9 mm ± 5.2; P = 0.008/P = 0.002), lateral patellar tilt angle (decrease from NWBCT (15.6° ± 6.7/16.9° ± 7.4) to WBCT (12.5° ± 7.7/15.0° ± 6.2; P = 0.011/P = 0.188). The medial femorotibial joint space decreased from NWBCT (3.9 mm ± 1.4/4.5 mm ± 1.3) to WBCT (2.9 mm ± 2.2/3.5 mm ± 2.2; P = 0.003/P = 0.004). Inter-reader agreement ranged from 0.52-0.97.
Knee joint alignment changes significantly in the upright weight-bearing position using CT when compared to supine non-weight-bearing CT.
• Cone-beam extremity CT offers upright weight-bearing examinations of the lower extremities.
• Knee alignment changes significantly in an upright position compared to supine position.
• Tibial tuberosity-trochlear groove distance (TTTG) is less pronounced in a weight-bearing position.
• The weight-bearing position leads to a decrease of the lateral patellar tilt angle.
KeywordsCone-beam extremity CT Upright weight-bearing CT Knee alignment Tibial tuberosity-trochlear groove (TTTG) Lateral patellar tilt angle
The knee joint has been well investigated using conventional weight-bearing radiographs, however, X-rays are limited to 2D imaging. Computed tomography (CT) with multi-planar reformations or magnetic resonance tomography (MR) in different planes are necessary for precise evaluation of certain parameters of the femorotibial and patellofemoral alignment, such as femorotibial rotation and tibial tuberosity-trochlear groove (TTTG) distance. Only a few studies have addressed investigations of the knee joint using weight-bearing MR [1, 2, 3, 4, 5, 6, 7, 8, 9, 10], mainly in the supine position with special devices [4, 7, 8] and regarding the patellofemoral joint. As such, the understanding of the physiologic alignment of the patellofemoral and femorotibial joints in the upright weight-bearing position is still limited. Certain cone-beam CT systems allow examinations of the knee joint in the upright weight-bearing position [11, 12]. In daily practice, cross sectional images of the knee joint are limited to the supine, non-weight-bearing position. Changes in joint alignment on cross sectional images between non-weight-bearing and weight-bearing positions are unknown. Thus, the purpose of this study was to prospectively compare patellofemoral and femorotibial joint alignment in standard supine non-weight-bearing CT (NWBCT) and upright weight-bearing CT (WBCT) and assess the differences in joint alignment.
Materials and methods
Patient and CT technique
Institutional review board approval and informed consent of all patients were obtained. NWBCT and WBCT images of the knee were obtained in 26 patients (mean age, 57.0 ± 15.9 years; range, 21-81 years; 15 women; mean age, 58.6 ± 14.2 years; range, 21-76 years; 11 men; mean age, 54.8 ± 18.5 years; range, 23-81 years) using a standard 64-slice multi-detector CT (Brilliance 64, Philips Healthcare, Best, Netherlands) for NWBCT and a novel cone-beam extremity CT for WBCT (Planmed Verity Extremity Scanner, Planmed Oy, Helsinki, Finland). Imaging was performed on the same day.
All patients included in this study were referred from an orthopaedic department with a clinical indication for a CT of the knee due to pain. Indications for the CT were osteoarthritis of the knee (n = 17), follow up of anterior cruciate ligament (ACL) graft (n = 3) and others (n = 6), including healed fracture of the intercondylar eminence (n = 1), avascular necrosis of the medial femoral condyle (n = 4), and a soft tissue tumour (n = 1). Patients were included if they were able to fully bear weight on the examined leg. Patients were excluded if they were under the age of 18 and if they did not give informed consent.
Two musculoskeletal radiologists independently performed the following measurements on NWBCT and WBCT: femorotibial rotation, TTTG, lateral patellar tilt angle, lateral patellar shift, and medial and lateral femorotibial joint space widths.
The medial and lateral femorotibial joint space widths were measured on coronal images at the midpoint of the medial and lateral femorotibial compartment.
Descriptive statistics were used to report the quantitative data. The Wilcoxon signed-rank test was utilized to assess significant changes between NWBCT and WBCT. Inter-reader agreement was assessed using an intra-class correlation coefficient (ICC). According to Rosner , the inter-reader reliability by means of ICC is classified as follows: > 0.75 is excellent, 0.4-0.75 is fair to good, and < 0.4 is poor. A P-value of less than 0.05 was considered statistically significant. For all analyses, SPSS software (release 17.0; SPSS, Chicago, IL, USA) was used.
Quantitative analysis of the knee joint alignment in supine non-weight-bearing computed tomography (NWBCT) and upright weight-bearing computed tomography (WBCT)
Femorotibial rotation [°]
2.7 ± 5.1 ER
0.4 ± 7.7 IR
2.6 ± 5.6 ER
0.2 ± 7.5 IR
13.8 ± 5.1
10.5 ± 5.0
13.9 ± 3.9
10.9 ± 5.2
Lateral patellar tilt angle [°]
15.6 ± 6.7
12.5 ± 7.7
16.9 ± 7.4
15.0 ± 6.2
Lateral patellar shift [mm]
0.8 ± 3.7
1.2 ± 4.5
0.9 ± 3.4
1.3 ± 3.7
Medial joint space width [mm]
3.9 ± 1.4
2.9 ± 2.2
4.5 ± 1.3
3.5 ± 2.2
Lateral joint space width [mm]
4.7 ± 1.5
4.5 ± 2.1
5.6 ± 1.5
5.2 ± 2.1
Inter-reader agreement of the knee joint alignment in supine non-weight-bearing computed tomography (NWBCT) and upright weight-bearing computed tomography (WBCT)
Lateral patellar tilt angle
Lateral patellar shift
Medial joint space width
Lateral joint space width
Routine cross sectional imaging, such as CT or MR, only allows examinations of the lower extremity in the supine position. Only a few MR systems offer upright weight-bearing examinations of the lower extremities [2, 3, 10], mainly provided by special devices to simulate weight-bearing either in the supine position [4, 7, 8] or in a semi-upright position [1, 5, 6]. For CT, usually, no real physiological weight-bearing examinations are possible . The recent introduction of a new scanner design, the cone-beam extremity CT used for this study, allows upright weight-bearing images of the lower extremity under physiological loading conditions [11, 12, 20]. We assessed the knee joint alignment in two positions, the supine non-weight-bearing and the upright weight-bearing position. Clinically significant differences in the weight-bearing position were found for the femorotibial rotation, the TTTG, the lateral patellar tilt angle, and the medial femorotibial joint space width. Excellent inter-reader agreement for a majority of the measurements demonstrates the high reproducibility of the measurement methods.
The lateral patellar tilt decreased significantly in the WBCT (12.5°) compared to NWBCT (15.6°) for reader 1. So far, no study has evaluated the lateral patellar tilt angle, with a line through the medial and lateral borders of the patella and a line through the anterior border of the femoral condyles, in the upright position. Powers et al.  compared the patellofemoral alignment of six patients with patellofemoral pain using kinematic MR in a non-weight-bearing sitting position and upright position with knee flexion from 0° to 45° with results equal to our study. Contraction of the quadriceps muscle, especially the vastus medialis muscle, which is a dynamic stabilizer of the patella, leads to a decrease of the lateral patellar tilt angle . It is known from the literature that the internal tibial rotation in WBCT might as well contribute to the decrease of the lateral patellar tilt. Lin et al. discovered no change of the lateral patellar tilt angle with any change in femoral or tibial rotation using the anterior border of the femoral condyle for measurement, while using the posterior border of the femoral condyle, the patellar tilt angle decreased with increasing internal tibial rotation in relation to the femur . This knowledge was the main reason for choosing the anterior border of the femoral condyle as a reference line for the lateral patellar tilt angle in our study. Hence, the decrease of the lateral patellar tilt angle in the weight-bearing position is supposed to be a true finding, as the femur is rotating underneath the patella . In earlier years, van Kampen and Huiskes  reported that femoral and tibial rotation influences patellar tracking, especially in knee extension and the first degrees of knee flexion.
The change of the femorotibial rotation from external rotation in NWBCT to internal rotation in WBCT may lead to a decrease of the TTTG in WBCT; both mechanisms were consistent with the results of our study. Izadpanah et al.  likewise reported a decrease of the TTTG from the supine to the upright position using MR. However, the TTTG distances, especially in the upright position were different to our results: In our study, the mean TTTG in the upright position was 10.5 – 10.9 mm, while in the study of Izadpanah et al. , the mean TTTG distance was only 6.3 mm. Assessment of the TTTG with different modalities such as MR and CT may vary, as shown by Schoettle et al.  and Camp et al. . The main reason for different TTTG values within or between modalities is the technique of TTTG measurement as well as the degree of extension or flexion of the knee, as shown by Dietrich et al. . Determination of precise axial images for assessment of the trochlea groove as well the tibial tuberosity is challenging. Identifying the most proximal image with a complete cartilaginous trochlea is difficult due to the thin cartilage layer and the low contrast differences between cartilage and synovium. Avoiding error in measurement, we chose a modified method, selecting the trochlea depth for the TTTG three centimetres above the joint line according to Pfirrmann et al. . Furthermore, we have remarked inaccuracies in selecting the most anterior point of the tibial tuberosity, as suggested by Schoettle , since this anatomical landmark is mostly flat. Therefore, we chose the midpoint of the tibial tuberosity at the level of the patella tendon insertion. Izadpanah et al.  selected the depth of the trochlea groove at the most proximal axial image with complete cartilaginous coverage as described earlier by Schoettle et al. , which is slightly more proximal than 3 cm above the joint line. Nevertheless, our mean TTTG of 10.5 - 10.9 mm in the upright position is in the range of normal TTTG values, which is less than 15 – 20 mm [24, 25].
Although not significant, a slightly higher lateral patellar shift was seen with WBCT. This finding is in accordance with the results of a kinematic MR study by Powers et al. . They also reported no significant difference of the lateral patellar shift between non-weight-bearing sitting and the upright weight-bearing position with a more pronounced shift in the weight-bearing position. The mean lateral patellar shift in the upright position (1.2 - 1.3 mm) was slightly higher compared to the results of Tennant et al. with 0.4 mm . Again, internal femorotibial rotation in the upright position may influence the lateral patellar shift, a similar phenomenon seen with the lateral patellar tilt angle. These results are consistent with observations of Powers et al. and Tennant et al. [2, 9].
The significant decrease of the medial joint space width in the upright position is a phenomenon of weight-bearing and might be pronounced in varus-aligned knees or in severe osteoarthritis. With increasing age, the prevalence of osteoarthritis of the knee joint increases, hence, osteoarthritis was evident in the majority of our rather elderly study population with a mean age of 57 years. A decrease of all joint space widths of the lower extremity in the upright weight-bearing position is a known phenomenon, although the decrease of the lateral femorotibial joint space width was not statistically significant, it was slightly less than in the supine position. Certainly, severe valgus-osteoarthritis knees may have a more pronounced lateral joint space narrowing under weight-bearing conditions.
The significant differences in knee joint alignment with upright weight-bearing imaging illustrate a potential method and the need for a natural and physiological position to evaluate the knee joint using cross-sectional imaging for a better understanding of the biomechanics and a precise planning for surgical therapy. Weight-bearing CT, in contrast to X-ray, has the potential to more precisely assess knee joint alignment. However, X-ray is a common method and has the advantages of cost-efficiency and widespread availability, a tool sufficient for baseline examination of the knee joint. The significantly lower values of the lateral patellar shift and the TTTG raises the question of new thresholds for examinations under physiological upright weight-bearing conditions evaluating patellofemoral diseases.
Limitations of our study are the limited number of patients (26), mainly presented with osteoarthritis of the knee joint; changes between non-weight-bearing and weight-bearing might be related to the underlying disease of this population. Hence, our results, especially concerning patellofemoral changes, may not apply to healthy volunteers or patients with patellofemoral disorders. As this is a preliminary study using cone-beam CT in the upright weight-bearing position, further studies are needed to establish threshold values in a larger population, in healthy volunteers and in patients with specific patellofemoral diseases. Examination of different degrees of knee flexion may help illustrate patellar alignment, mimicking physiological conditions such as walking or climbing stairs.
In conclusion, the knee joint alignment differs significantly in the upright weight-bearing position compared to the supine non-weight-bearing position using CT.
The scientific guarantor of this publication is Anna Hirschmann, MD. The authors of this manuscript declare no relationships with any companies whose products or services may be related to the subject matter of the article. The authors state that this work has not received any funding. One of the authors has significant statistical expertise; however, no complex statistical methods were necessary for this paper. Institutional Review Board approval was obtained. Written informed consent was obtained from all patients in this study. Methodology: prospective, diagnostic study performed at one institution
- 10.Izadpanah K, Weitzel E, Vicari M, et al (2014) Influence of knee flexion angle and weight bearing on the Tibial Tuberosity-Trochlear Groove (TTTG) distance for evaluation of patellofemoral alignment. Knee Surg Sports Traumatol Arthrosc 22(11):2655-2661Google Scholar
- 12.Demehri S, Muhit A, Zbijewski W, et al. (2015) Assessment of image quality in soft tissue and bone visualization tasks for a dedicated extremity cone-beam CT system. Eur Radiol. doi:10.1007/s00330-014-3546-6
- 18.Rosner B. Fundamentals of biostatistics, 7th ed. Boston: Brooks/Cole, Cengage Learning, 2011:xvii, 859 p.Google Scholar