91³Ô¹ÏÍø

Data Preparation and Image Processing

30 knee osteoarthritis patients underwent CT scanning, with 2 patients disqualified due to poor CT scanning results, leaving a cohort of 28 anonymized datasets. A power analysis was performed for standard sized TomoFix HTO plates (DePuy Synthes) using experimentally measured variation in stiffness and strength. This input helped determine the suitability of the patient cohort size. The TomoFix HTO plate is a widely implanted HTO device and was used as the ¡®generic device¡¯ in this study.

The CT data was imported into Simpleware software for segmentation, and five key landmarks of interest were identified. The required osteotomy correction angle was calculated using MATLAB so that the altered mechanical axis passed through a point 62.5% of the distance from the medial to the lateral tibial plateau. Virtual HTO surgery was performed on each patient to alter the mechanical axis of the knee by creating an opening wedge osteotomy, Ansys SpaceClaim was used for this procedure, with guidance from an orthopedic surgeon specializing in knee surgery.

After the virtual surgeries were performed, each virtual patient was duplicated. One copy had the osteotomy stabilized using the generic plate, and the other using the personalized plate (TOKA, Orthoscape), thus forming the two groups of the trial. For the generic group, the TomoFix high tibial plate was micro-CT scanned (Nikon Metrology) and processed in Simpleware software. For the personalized group, patient-specific implant geometries were created using specialized planning software (Renishaw plc), taking into account the surface of the tibia and the degree of correction for each patient. All simulated knees were virtually implanted with both implant types, therefore generating a total of 56 intervention cases as FE models for simulation in Ansys software.

FE Simulation

The FE models were created, and simulation carried out based on a validated methodology and model, including the method for representing the screw and plate, as well as contact interactions between the components. Patient-specific material properties were applied from each patient¡¯s CT data, using heterogeneous linear elastic properties. Other factors, such as normal contact stiffness, were determined based on experimental testing and Ansys default contact settings. The progression of bone healing was represented by increasing the Young¡¯s modulus of the osteotomy region, with data from previous studies used to quantify the extent of osteotomy gap healing at different time points.

Meshing parameters were based on a previous validated study, and checks made to ensure suitability for the simulation. Muscle forces and joint reaction forces for normal physiological activities were also calculated using a subject-specific musculoskeletal model, with forces automatically registered to the individual patient geometries using a custom transformation and scaling script in MATLAB. Three common activities were considered: fast walking gait, chair rise, and squat, and five key instances selected for each activity based on locations of peak tibial contact force, implemented as load steps 1 to 15 in the FE models.

Simulations were run for both patient models for the physiological activities, three screw configurations, and osteotomy gap bone healing stages at different time intervals: healing stage 1 covered the immediate post-operative period (not simulated but included in the ClinicalTrials.gov entry, and included for consistency), healing stage 2 was 2-weeks after surgery, healing stage 3 was 6 weeks after surgery, and healing stage 4 was 12 weeks after surgery.

The principal outcome variable was the maximum Von Mises stress within the plates, within the bone adjacent to the screws used to fix the plates, and within the inter-fragmentary movement at the osteotomy site. Fatigue failure was taken into account as a variable for all metal implants, based on fatigue testing. Statistical analysis of the effect of the three screw configurations was also conducted.

Results

Solving of the FE models was computationally intense, requiring 500,477 core hours on a high-performance computing cluster. The results indicated that the effects of screw configuration on the Von Mises stress in each type of plate were not statistically significant, with a large degree of overlap between the confidence intervals, while the effect of the healing stage was very dramatic in reducing the maximum Von Mises stresses in both sets of plates. In addition, differences in bone strain around the screw insertions for the two plates were small and not significant. However, the differences in inter-fragmentary movement between the two devices was more significant at healing stage 4 (12 weeks post-operatively) due to increased micro-motion in the personalized device.

As the main failure mechanism for these devices is fatigue failure, the key result was the odds ratio indicative of the relative risk of failure between the personalized and generic groups at different healing stages. While the personalized HTO plates consistently had larger stress values than their generic counterparts, the delta values were relatively small. More importantly, the personalized HTO plates had no difference in their risk of failure compared to generic plates, whilst being more mechanically efficient and less stiff. These results demonstrated that 3D printed, patient-specific plates can address critical challenges such as accuracy, stiffness, and geometric conformity, which limit the current use of HTO surgery, while also enabling simpler and quicker surgeries.

Conclusions

The breakthrough use of in silico trials for orthopaedic safety testing and informing clinical trials showed no increased risk of failure for the personalized design versus the generic design.

The results support further clinical studies looking at the benefits of patient-specific over standard realignment surgery, and the potential of this method for treating early knee arthritis. Future in silico trials could broaden the scope of the study beyond mechanical safety and stability, for example in terms of relative efficacy.

This study formed part of regulatory evidence accepted by MHRA in approval for a subsequent clinical patient trial (the PASHiOn clinical trial). The safety phase in 5 patients using 3D printed HTO plates led to positive outcomes, with one patient being able to participate in a competitive bike ride within 6 months. The intervention involved X-ray and CT scans of patient shin bones (tibia) being used to digitally plan HTO surgery, from which a patient-specific 3D printed surgical guide and stabilization plate were generated. The surgical guide enabled the surgeon to make the cuts and create the appropriate opening wedge which is then stabilized by the plate.

A 25-patient single-group trial at the Rizzoli Institute in Bologna was also conducted and demonstrated successful treatment of patients with the device. The personalized HTO system demonstrated accurate correction in terms of both coronal and sagittal alignment, with excellent patient reported outcomes at a one-year follow-up. An ongoing randomized control trial in the UK () will now compare the novel personalized method with a traditional method.

Professor Richie Gill from the University of Bath commented on the innovative aspect of the research and the contribution of in silico trials to ongoing testing with a larger patient cohort:

"The big difference is there isn't that guesswork which is happening at the time of surgery. I think more surgeons are interested in doing this type of joint-preserving surgery. You can do it at a sooner point in the disease progression than you can with a knee replacement, so we are hoping to make it more widely available."

Learn More

• Read the full on this study
• See how personalized HTO is already
• Read papers from the Rizzoli Orthopaedic Institute in Bologna on the first clinical trial with 25 patients: 
• Except where otherwise noted, content on this page is licensed under a .