Individualizing a Total Knee Arthroplasty with Three-Dimensional Planning

Abstract

Total knee arthroplasty (TKA) is evolving from mechanical alignment to more individualized alignment options in an attempt to improve patient satisfaction. Thirteen-year survival of kinematically aligned prostheses has recently been shown to be similar to mechanically aligned TKA, allaying concerns of long-term failure of this newer individualized technique. There is a complex inter-relationship of three-dimensional knee and limb alignment for a TKA. This article will review planning parameters necessary to individualize each knee, along with a discussion of how these parameters are related in three dimensions. Future use of computer software and machine learning has the potential to identify the ideal surgical plan for each patient. In the meantime, the material presented here can assist surgeons as newer individual alignment planning becomes a reality.

Alignment philosophies in total knee arthroplasty (TKA) are changing from mechanical alignment for all patients to individualized alignment in an attempt to improve patient outcomes and satisfaction.¹ An array of newer alignment and technique philosophies are currently in the orthopaedic literature and include kinematic alignment (KA), restricted kinematic alignment (rKA), inverse kinematic alignment, functional alignment, and modified mechanical alignment along with techniques performed with mechanical instruments, modified mechanical instruments, navigation, sensors, balancers, robotics, and augmented virtual reality. This multitude of alignment and technique philosophies amounts to more complexity for the busy surgeon trying to decide the best alignment method and technique to perform TKA.²

To make sense of this shifting paradigm, planning must focus on the individual patient's anatomy and pathoanatomy, along with the alignment goal to be achieved with the surgical procedure. Postoperative outcomes and satisfaction can then be correlated with the surgical procedure to help improve results. The following discussion is intended to present the inter-related components of the planning process. In this review, we focus on three-dimensional (3D) planning and specific interaction of each of these dimensions, considering

1. Rationale for individualized planning of total knee arthroplasty

2. Limb alignment vs knee alignment

3. Coronal alignment of the femoral implant

4. Potential effect of increasing degrees of femoral valgus on patellar tracking

5. Axial alignment (rotation) of the femoral implant

6. Relationship of axial and coronal femoral implant alignment to balance flexion gap

7. Effect of axial and coronal femoral alignment on preexisting patellar subluxation

8. Sagittal alignment of the femoral implant

9. Sagittal alignment of the tibial implant

10. Axial rotation of the tibial implant

11. Relationship of limb alignment to knee alignment

12. Relationship of tibial coronal alignment to femoral coronal alignment and limb alignment

Rationale for Individualized Planning of Total Knee Arthroplasty

In 2008, Howell described an individualized alignment method using the shape match method, referencing only knee alignment but not limb alignment.³ The goal was to restore the individual patient's pre-arthritic anatomy, which would align the knee to the kinematic axes of the knee (Figure 1), as originally described by Hollister and later represented by Eckhoff with the cylindrical axis, which they advocate more closely corresponds to the actual flexion-extension axis of the knee.^4-6

Figure 1.

Diagram of the three kinematic axes of the knee.

Using patient-specific guides developed from the MRI of the knee, better outcomes were reported in various studies of KA and no soft-tissue releases were performed.^7,8 A recent 13-year follow-up of this method has shown similar survival of the prosthesis to mechanical alignment, allaying concerns of early failure of the prosthesis with this alignment method.⁹ Kinematically aligned total knee arthroplasties with patient-specific guide instruments have shown a similar risk of revision when compared with all other total knee arthroplasties in a review of the Australian and New Zealand Joint Replacement Registries.¹⁰

Limb Alignment Versus Knee Alignment

Mechanical alignment plans the coronal femoral joint angle at 90° to the mechanical axis of the femur and the tibial joint angle at 90° to the mechanical axis of the tibia. The resultant limb alignment (hip-knee-ankle angle) is, therefore, a straight line. Insall¹¹ thought this would help the survival of the prosthesis by producing symmetric loading of the femur and tibia. Because the coronal femoral and tibial joint angles are always the same in mechanical alignment, there has been very little emphasis on individual knee alignment to date.

A 20-year follow-up of Mayo Clinic total knee arthroplasties showed that a neutral limb alignment did not provide better implant survivorship than those found in the outlier group.¹² In an analysis of alignment comparing kinematically and mechanically aligned total knee arthroplasties, we did not find a difference in average overall limb alignments between the two groups.¹³ We found differences in knee alignment, especially femoral valgus and tibial varus angles, resulting in more joint obliquity relative to the mechanical axis, which may be associated with the better results in the kinematic group.¹⁴

Coronal Alignment of the Femoral Implant

When a surgeon uses conventional instruments for mechanically aligned TKA, the mechanical axis of the femur is not visualized. To reference the mechanical axis of the femur, the surgeon must either obtain long leg hip-to-ankle radiographs or hip-to-ankle CT or MRI scan or use intraoperative technology, such as computer navigation, robotic navigation, or more recently augmented reality. A surgeon can then define an individual patient's mechanical lateral distal femoral angle (mLDFA). The mLDFA varies markedly from patient to patient and is the key to producing the individualized alignment plan.¹⁵ In the cylindrical axis model presented by Eckhoff, the most distal aspect of the cylinder is the mLDFA and the most posterior aspect of the cylinder is the posterior condylar axis (PCA). This allows the femoral implant to be aligned with the axes of rotation of the femur, which was the basis of the shape match method of KA (Figure 2).

Figure 2.

Illustration and images showing the mLDFA. The image on the left shows the location of the mLDFA as the most distal aspect of the cylindrical axis. The image on the right illustrates the mLDFA preoperative and postoperatively along with the 6-month postoperative patient-reported outcome scores for this surgery. mLDFA = mechanical lateral distal femoral angle

The shape match method of alignment was an unrestricted KA. The distal femoral valgus angle relative to the mechanical axis can range from 10° valgus to 5° varus.¹⁶ Higher femoral valgus joint angles move the trochlea of the prosthesis medially, which may adversely affect patellar tracking.

Potential Effect of Increasing Femoral Valgus on Patellofemoral Tracking

One of the unintended effects of unrestricted KA observed in the authors' 13-year follow-up study was patellar subluxation and lateral patellar facet impingement in some patients who had satisfactory preoperative patellar tracking.⁹ Tran et al¹⁷ noted increased patellar complications with unrestricted KA. Riviere previously noted that KA does not restore the native trochlear anatomy.¹⁸ Wang showed that the undercoverage of the trochlear resection surface in KA is mainly correlated with the degree of valgus of the distal femoral joint line (Figures 3 and 4).¹⁹

Figure 3.

Illustration showing with an increasing valgus position of the femoral implant (illustration on the right), the center of the trochlea is moved medially (black dot), which may result in patellofemoral instability.

Figure 4.

Images showing one of the unintended effects of unrestricted kinematic alignment observed in the authors' 13-year follow-up study (6) was patellar subluxation and lateral patellar facet impingement in patients who had satisfactory preoperative patellar tracking.

Axial Alignment (Rotation) of the Femoral Implant

Moreland²⁰ illustrated the axial changes needed to correct the flexion balance of the knee when the alignment method was changed from “anatomic” alignment to mechanical alignment (Figure 5). Assuming an average 3° coronal valgus femoral joint angle and 3° coronal varus tibial joint angle for anatomic alignment, femoral axial rotation was set to zero degrees relative to the PCA. When mechanical alignment was used, the femoral coronal joint angle was zero degrees, tibial coronal joint angle was zero degrees, and femoral axial rotation was set to 3° of external rotation to attain balance of the flexion gap. Because the preoperative femoral joint angle (90—mLDFA) is highly variable, resection to balance the flexion gap in mechanical alignment or some form of rKA is also variable. However, there is a relationship between coronal and axial alignment that can help preplan a balanced TKA.²¹ León‐Muñoz and co-authors noted a linear relationship between the coronal alignment and the rotational geometry of the distal femur.²² Lee et al.²³ showed using radiographs and a navigation system, the mLDFA was predictive of the axial alignment of the femoral implant.

Figure 5.

The image on the left shows an anatomic resection of the tibia in slight varus, with a balanced flexion gap by symmetric resection of the posterior condyles of the femur, parallel to the posterior condylar axis. The image on the right shows the tibial resection at 90° to the mechanical axis, with a balanced flexion gap by externally rotating the femur, resecting more bone at the posteromedial femoral condyle compared with the posterolateral femoral condyle.

This resection can be calculated with trigonometry and achieved with a measured resection surgical technique. Approximately 1 mm additional posterior medial femoral resection will correlate with an additional 1° of femoral external rotation (Appendix A, http://links.lww.com/JG9/A324). Sensors and balancers can also be used; however, it is unclear whether these improve outcomes.^24,27

Relationship of Axial and Coronal Femoral Implant Alignment to Balance Flexion Gap

To prevent notable outliers and to prevent patellar tilt or subluxation, some surgeons have adopted a rKA philosophy, for example, limiting the femoral joint angle to within 5 degrees of the mechanical axis.²⁸ To balance the flexion gap when the femoral joint angle is limited, external rotation relative to the PCA is included in the planning process, similar to how external rotation in mechanical alignment is used to balance the flexion gap.^28-30

Effect of Axial Alignment on Preexisting Patellar Subluxation or Tilt

Axial rotation can also have a role in individualizing a knee arthroplasty plan for patients with significant preoperative patellar subluxation or tilt because as the femoral prosthesis is externally rotated 1 to 2°, the trochlear flange is moved laterally, potentially improving patellofemoral tracking (Figure 6). Kawahara et al³¹ showed ideally, the anterior femoral flange should be shifted 2 mm to 2.5 mm laterally relative to the distal and posterior aspects to provide optimal femoral bone coverage. In addition, self-seating of the tibial implant in these cases can move the tibial tubercle medially attempting to improve the Q-angle and allow for correction of preexisting patellofemoral subluxation without releases.^32,35 An algorithm for helping align the patella is presented in Table 1.

Figure 6.

Diagram and images showing with external rotation of the femoral implant (top), the center of the trochlea is moved laterally, which can improve patellofemoral tracking (bottom)

Table 1.

Steps to Help Align the Patella, Including Knees with Preoperative Subluxation

Steps to help align the patella, including knees with preoperative subluxation
1. Template each proposed patellar cut
2. Measure the patellar thickness before cutting
3. Use a patellar cutting guide adjusted to match the thickness of the implant
4. Measure all four quadrants of the patella once it is cut to be sure the cut is not oblique
5. Ensure after the cut that bone plus implant thickness is equal to initial measurement of the patella
6. Add 1 to 2° external femoral rotation off the posterior condylar axis if the patella is subluxated preoperatively—float (self-seat) the tibia to match

Sagittal Alignment of the Femoral Implant

Modern prosthetic implants have an array of femoral sizes.³⁶ An individualized three-dimensional plan will provide posterior femoral resections to match the posterior thickness of the prosthesis, allowing for bone and cartilage wear in the arthritic knee. The posterior resections are critical to balancing the flexion gap. Once the proper posterior resections are planned, attention is paid to placing the anterior flange of the prosthesis onto the femur without notching. If a patient is in between femoral sizes, the smaller size is selected and the femur is flexed just enough to avoid notching. Limiting excessive femoral flexion may help reduce patellar subluxation. Howell and coauthors showed flexion of the femoral implant reduces the proximal reach of the trochlea by moving the trochlea distally on the femur and may lead to patellofemoral instability.³⁷ The trochlea also has a 3D relationship to the mLDFA because it has been recently noted that the distal trochlear sulcus angle (the angle between the medial and lateral trochlear facets) and the mLDFA have a strong positive correlation. This means that basing coronal femoral alignment on the mLDFA potentially improves trochlear alignment and patellar tracking.³⁸ Posterior-stabilized implants have the additional consideration of postimpingement with increasing sagittal flexion of the femoral implant.

Sagittal Alignment of the Tibial Implant

Individual knees have different degrees of posterior tibial slope relative to the tibial mechanical axis. A study of over 13,000 patients showed that posterior tibial slope is highly variable in patients with arthritis with a range from 5° anterior slope to 25° posterior slope.³⁹

In general, we match the slope of each patient up to a restricted amount, which is a maximum of 2° greater than the manufacturer's recommended slope. We avoid inadvertently planning or executing the slope markedly less than the patient's anatomy, or the manufacturer's recommended slope, because this might limit flexion of the knee.

Axial Rotation of the Tibial Implant

The tibial tubercle location is highly variable from medial to lateral.⁴⁰ If the femoral implant is externally rotated 1 to 2° to improve preexisting patellar subluxation, the range-of-motion technique (self-seating) of the tibial implant to match the posterior femoral implant can allow 1 to 2° of external rotation to the tibial implant, moving the tubercle slightly toward the midline and reducing the Q-angle.^31,34 However, caution should be used when self-seating the tibia if the knee is not balanced because increased forces in the medial or lateral compartment may prevent the implant from assuming the correct rotational orientation.

Relationship of Limb Alignment to Knee Alignment

Bellemans et al⁴¹ reviewed MRI findings for 500 men and women. When the patients were placed into varus <3° limbs, valgus >3°, or neutral limb alignment categories, the average knee alignment parameters can be categorized. Neutral limbs had closely matching mLDFA and medial proximal tibial angle (MPTA), varus limbs had higher mLDFA and lower MPTA, and valgus limbs had lower mLDFA with higher MPTA. In other words,

Neutral limbs: Femoral joint angle = Tibial joint angle

Varus limbs: Tibial joint angle > Femoral joint angle

Valgus limbs: Femoral joint angle > Tibial joint angle

These relationships can be helpful when designing a preoperative plan allowing for residual varus or valgus limbs based on the preoperative limb alignment (Figure 7). These three categories can be expanded with the Coronal Plane Alignment of the Knee classification, which can facilitate research or computerized planning involving coronal alignment.⁴² Notably, coronal alignment is only one of the three dimensions that should be considered for planning a TKA.

Figure 7.

Radiographs of a single patient with windswept deformity of the lower extremities. The right varus limb was left in 2° postoperative varus. The left valgus limb was left in 3° postoperative valgus. Western Ontario and McMaster Universities Arthritis Index score (96 worst, 0 best) for both knees at 6 months was 0. Both postoperative joint lines are parallel to the ground on these standing views. No soft-tissue releases were used to attain this alignment.

Relationship of Tibial Coronal Alignment to Femoral Coronal Alignment and Limb Alignment

The mLDFA can be determined preoperatively by long leg radiograph, MRI, or CT scan. The MPTA is set to the same angle for patients with preoperative neutral limb alignment (hip-knee-ankle ± 3°). For patients with varus and valgus limbs, the mLDFA and MPTA can be adjusted individually to allow for residual limb varus or valgus. One published restricted kinematic algorithm limits limb alignment to within 3 degrees of the mechanical alignment of the limb.²⁸

Conclusion

There is a complex inter-relationship of three-dimensional alignment for individualized total knee arthroplasties. Each of the alignment parameters can be quantitated and correlated with patient-reported outcomes. Statistical analysis and use of machine learning can assist in the planning process.

Machine learning is considered a subset of artificial intelligence that involves the use of various computer algorithms. Machine learning allows the computer to learn and continually improve analysis of data. Large sets of inputs and outputs are used to train the machine to make autonomous recommendations or decisions.^43,44 Future use of computer software and machine learning has the potential to identify the ideal surgical plan for each patient. In the meantime, the material presented here can assist surgeons as newer individual alignment planning becomes a reality.