1 Introduction

Total knee arthroplasty (TKA), also referred to as total knee replacement (TKR), is one of the most successful procedures in orthopaedics, providing substantial pain relief and functional improvement for patients with advanced osteoarthritis. Optimal outcomes depend on accurate restoration of coronal, sagittal, and axial alignment of the femoral and tibial components, along with balanced soft tissues.^1,2 While coronal and sagittal alignment have historically been emphasised, increasing attention has turned to rotational alignment, given its impact on implant longevity, patellofemoral tracking, and patient satisfaction.

Excessive internal rotation of the femoral component has been shown to compromise stability and patellar tracking, leading to pain, stiffness, and arthrofibrosis.^3–5 Similarly, malrotation of greater or equal to 4° internally or greater than 8° externally has been associated with complications such as patellofemoral maltracking, dislocation, wear, loosening, and even early implant failure.^6–8 Although internal malrotation is typically more detrimental, excessive external rotation can also destabilize the medial flexion gap and impair patellofemoral mechanics.

Achieving reliable rotational alignment remains challenging. Surgeons may adopt a fixed degree of external rotation (for example, 3°), use gap-balancing methods, or rely on bony landmarks. However, each approach has limitations. The surgical transepicondylar axis (sTEA) is considered the most reliable reference but can be difficult to identify intraoperatively, and the medial sulcus is not always visible, even on computed tomography (CT).⁹ The anatomical transepicondylar axis (aTEA) is easier to identify but subject to intra- and inter-observer variability.¹⁰ By contrast, the posterior condylar axis (PCA) is consistently identifiable intraoperatively, and its relationship to the aTEA defines the condylar twist angle (CTA).

Preoperative measurement of the CTA offers a patient-specific approach to femoral rotation, avoiding reliance on arbitrary external rotation values. Prior work using CT-based CTA has demonstrated improved accuracy of femoral component positioning.¹¹ However, CT cannot account for cartilage thickness, which may introduce error. Magnetic resonance imaging (MRI) offers a potential advantage by capturing both bony and cartilaginous landmarks, allowing for more accurate CTA measurement.

Based on this rationale, we hypothesized that preoperative MRI-derived CTA, when reproduced intraoperatively, would improve the accuracy of femoral component rotational alignment in TKA.

2 Methods

2.2

2.2 Definitions

'Anatomical transepicondylar axis' (aTEA) - line drawn between the most prominent points of the medial and lateral epicondyles of the femur. 'Surgical transepicondylar axis' - more representative of the function axis, it's from the line drawn from medial epicondylar sulcus to lateral epicondyle. Not everyone has an identifiable sTEA. 'Posterior condylar axis' (PCA) - a line connecting the posterior margins of the medial and lateral femoral condyles. 'Condylar twist angle' - angle between the PCA and aTEA. It is usually externally rotated to the posterior condylar angle. 'Trochlear line' is a line outlining the deepest parts of trochlear groove, where the patella glides during movement. A graphical representation of the above is available in Fig. 1.

Close

Fig. 1 CT scan of right knee showing key landmarks This shows a CT Scan picture of a right knee, depicting the important landmarks used in total knee arthroplasty (see 'definitions' for more information). The medial sulcus seen in this CT Scan is not appreciable in all cases, therefore sTEA cannot be drawn in all cases. Original image from our institution, used with consent.

Export to PPT

2.3

2.3 Perioperative management and surgical technique

Preoperatively, an MRI scan was obtained of the knee (see Fig. 3). Under guidance of an expert radiologist, the anatomical transepicondylar axis (aTEA), posterior condylar axis (PCA) including the cartilage thickness and the angle between these two; the condylar twist angle (CTA) were calculated. Definitions of relevant axes are detailed below, with a labelled image available in Fig. 1.

All surgeries were performed using a medial parapatellar approach by the primary author and another senior Orthopaedic surgeon. The PCA was identified intraoperatively, while the aTEA was less reliable to visualise. The preoperative CTA was reproduced using a posterior condylar referencing jig (see Fig. 2), which permitted external rotation of the femoral component in 3°, 4°, or 6° increments. When the exact CTA was not available, the nearest lower external rotation was selected. Femoral component sizing was performed with reference to the anterior femoral cortex. The distal femoral cut was made at 5–7° of valgus.

Close

Fig. 2 Intraoperative image of knee with rotation guide This shows the rotation guide hinged at posterior condyles. Rotation was set at 6° (equal to CTA) Subsequently posterior cut was made at 6° external rotation to PCA. Original image from our institution, used with consent.

Export to PPT

Close

Fig. 3 Preoperative magnetic resonance imaging (MRI) axial slice of a right distal femur showing condylar twist angle (CTA). The anatomical transepicondylar axis (aTEA) is drawn between the most prominent points of the medial and lateral epicondyles (upper line). The posterior condylar axis (PCA) is identified by connecting the posterior margins of the medial and lateral femoral condyles (lower line). The angle between the PCA and the aTEA represents the CTA, which in this case measures 7°. Magnetic resonance imaging allows assessment of both bone and cartilage, enabling more accurate patient-specific CTA measurement for guiding femoral component rotation in total knee arthroplasty (TKA). Original image from our institution, used with consent.

Export to PPT

Patients followed a standardised rehabilitation protocol. At one month postoperatively, computed tomography (CT) scans were obtained to reassess the PCA and aTEA. The postoperative CTA was calculated as the angle between these two landmarks (see Fig. 4). The error in femoral rotation was defined as the difference between preoperative (planned) CTA and postoperative (achieved) CTA. A postoperative CTA of 0° indicated perfect alignment with the aTEA.

Close

Fig. 4 Postoperative computed tomography (CT) scan of a right distal femur showing femoral component alignment The anatomical transepicondylar axis (aTEA, upper line) and the posterior condylar axis (PCA, lower line) are marked. In this case, the femoral component is aligned parallel to the aTEA, with a measured angle of 0°, indicating accurate reproduction of the preoperatively planned condylar twist angle (CTA). Postoperative CT validation confirms that MRI-derived CTA measurements can be reliably reproduced intraoperatively to achieve optimal rotational alignment in total knee arthroplasty (TKA). Original image from our institution, used with consent.

Export to PPT

3 Results

A total of 23 knees in 12 patients met the inclusion criteria and underwent TKA guided by preoperative MRI - derived CTA, with the results seen in Table 1. The mean age was 66.5 years (range 49–81, interquartile range [IQR] 60–77). Fourteen knees (60.9 %) belonged to female patients, who were younger at the time of surgery (mean 61.7 years) compared with males (mean 71.0 years). Most patients (22 of 23 knees, 95.7 %) underwent bilateral replacement.

On preoperative MRI, the CTA averaged 6.47° (range 3–9°, IQR 5–8°). Male patients demonstrated a higher mean CTA (7.44°) compared with females (5.85°). A CTA of 7° was most common, observed in 5 knees (21.7 %).

The mean error between the planned and achieved condylar twist angle (CTA), validated on postoperative computed tomography (CT), was 0.60°. Overall, 22 knees (95.7 %) achieved alignment within 1° of the anatomical transepicondylar axis (aTEA). This group included 8 knees (34.8 %) with perfect parallel alignment (0°), 4 knees (17.4 %) with minimal deviations of 0.4–0.5°, and 10 knees (43.5 %) with a 1° deviation. Only 1 knee (4.3 %) showed a deviation of 2°.

Clinically, all patients achieved symmetrical flexion gaps with stable intraoperative balancing. Patellofemoral tracking was satisfactory in all cases, and no lateral retinacular release was required.

4 Discussion

This study evaluated the use of magnetic resonance imaging (MRI) - derived condylar twist angle (CTA) to guide femoral component rotation in total knee arthroplasty (TKA). It provides novel data directly comparing preoperative planning with postoperative computed tomography (CT) validation of rotational alignment. The mean error between expected and achieved CTA was only 0.60°, demonstrating near-perfect rotational alignment. Gender-related differences in CTA were also observed, consistent with previously described anatomical variation. These findings suggest that patient-specific imaging may help reduce malrotation and its sequelae, including knee pain and revision surgery.

Demographic characteristics in our series were broadly consistent with the literature, with a mean patient age of 66.5 years and a predominance of women. Women underwent TKA almost a decade earlier than men, in contrast to prior studies reporting no significant sex-related age difference.¹² Females also had a lower CTA than males, consistent with Jabalameli et al.¹³ who reported mean CTA values of 5.1° in women and 7.2° in men. This difference likely reflects sexual dimorphism of distal femoral morphology, where narrower condyles and a more medially oriented trochlear groove in females reduce the divergence between the posterior condylar axis (PCA) and anatomical transepicondylar axis (aTEA), producing a lower CTA.¹⁴

All knees in our study demonstrated the aTEA in external rotation relative to the PCA, with CTA values ranging from 3° to 9° (mean 6.47°). These results align with prior anatomical series, including Mullaji et al.¹⁵ in an Indian population (mean 5°, range 1.3–9.1°) and İncesoy et al.¹⁶ in a Turkish cohort (median 7°, range 0–13°). Such variation reflects the multifactorial nature of rotational anatomy. Sex and ethnicity influence skeletal morphology,¹⁷ while osteoarthritis severity and coronal deformity can alter the functional relationship between the PCA and aTEA through asymmetric bone loss and cartilage wear.^18,19 Technical factors also contribute, including the obscuring effect of osteophytes,²⁰ the choice of imaging modality (CT captures only bone margins and not cartilage),^21,22 and the reliability of epicondylar landmark identification.⁹ These considerations highlight that every knee is unique in its rotational profile, and reliance on a single fixed external rotation value risks malalignment. Individualising femoral rotation according to patient-specific CTA may therefore help achieve balanced flexion gaps, stable patellofemoral tracking, and long-term implant stability.

With postoperative CTA of 0° considered perfect alignment, our mean error was 0.60°. This compares favorably with a similar study by Sharma et al.,¹¹ who used CT-based preoperative CTA measurements and reported a mean error of 0.77°. This discrepancy may reflect difficulties in marking the aTEA on MRI or minor error in jig positioning, as the prongs may not always sit flush on the PCA. The limited range of available external rotation jigs (3°, 4°, and 6°) also introduced error in patients with higher CTA values. Attempts to correct for high CTA risk lateral femoral cortex notching, as noted by Koudela et al.^20,23 who suggested limiting external rotation to between 0° and 7°. Despite these challenges, all knees in our study had implants externally rotated relative to PCA, with flexion stability, balanced gaps, and required no lateral release.

Strengths of this study include the prospective use of MRI to plan femoral rotation, intraoperative application of these measurements, and postoperative CT validation by an experienced surgeon in a uniform surgical environment. Few prior studies have reported such objective confirmation of intraoperative alignment accuracy. This approach is especially useful in the emerging areas of patient-specific instrumentation (PSI) and robotic-assisted TKA; where MRI can incorporate cartilage thickness to provide more reliable patient-specific targets and reduce the risk of malrotation.^23–25

Limitations include the relatively small sample size (12 patients, 23 knees), single-centre design, and potential technical errors such as jig fit, residual osteophytes, or inter-observer variation in MRI-based CTA assessment. Preoperative MRI may also not be practical or feasible in all centres due to cost or availability.

5 Conclusion

Preoperative magnetic resonance imaging (MRI) measurement of the condylar twist angle (CTA), reproduced intraoperatively, enabled highly accurate restoration of femoral component rotation in total knee arthroplasty (TKA). The mean rotational error was under 1°, with balanced flexion gaps, stable patellofemoral tracking, and no need for lateral release. Sex-based differences in CTA underscore the importance of patient-specific planning rather than fixed external rotation values. By incorporating cartilage thickness and individual anatomical variation, MRI-based CTA provides a reproducible method to reduce malrotation and its complications. Larger multicentre studies with longer follow-up are required to validate these results and assess their effect on long-term function and implant survival.

Patient consent statement

Informed consent for treatment and open access publication was obtained or waived by all participants in this study.

Author contributions according to credit taxonomy

JJ: conceptualisation, data curation, formal analysis, investigation, methodology, project administration, supervision, visualisation.

MY: formal analysis, writing original draft, writing review & editing.

RR: Writing review & editing.

Data availability statement

Data is available on reasonable request by contacting the corresponding author.

Permission to reproduce material from other sources

Not applicable.

Ethical statement

DR. RAM MANOHAR LOHIA HOSPITAL, NEW DELHI (Institutional Ethics Committee) issued approval 1–40/9/2014/IEC/Thesis/PGIMER-RMLH/1472. DR. RAM MANOHAR LOHIA HOSPITAL, NEW DELHI (Institutional Ethics Committee) issued approval 1–40/9/2014/IEC/Thesis/PGIMERRMLH/1472 on October 27, 2014.

Funding statement

No funding was received.