Insights from RNOH and UCL's Research in Statistical Shape Modeling and Patient-Specific Designs

Anna Di Laura, Ph.D., Senior Researcher at RNOH and UCL, shares insights on their work with Simpleware software: "Our collaboration with Simpleware on the development of their SSM tool is allowing us to overcome major challenges in surgical planning of hip replacement surgery and implant design. The use of SSM helps improve surgical outcomes by optimising implant positioning and alignment based on a detailed understanding of both common and patient-specific anatomical features. This enables a shift towards more accurate and faster 3D planning options for hip replacement surgery."

The recent International Society for Technology in Arthroplasty (ISTA) 2024 Annual Congress in Nashville, Tennessee, featured four papers showcasing the impressive results achieved by the RNOH and UCL team using Simpleware software. These methods, applied by Dr. Johann Henckel and Prof. Alister Hart, senior members of the team, demonstrate promising clinical impact to aid complex surgeries.

Total hip arthroplasty (THA) is often needed to restore hip joint function when acetabular distortion or bone loss are present. Restoration of the centre of rotation (CoR) allows for an improved kinematic of the hip and preservation of acetabular bone stock, decreasing the risk of long-term loosening. However, correcting limb length inequality whilst maintaining hip stability is a significant challenge. Statistical shape modeling (SSM) can help guide the restoration of the CoR, the orientation of the prosthetic component, adequate offsets, and an equal leg length by predicting information that is missing or impossible to capture with traditional modeling. This project aimed to improve the surgical planning of primary hip surgery by applying SSM to patients who had previously undergone THA.

A retrospective cohort study was carried out involving 50 patients, with an SSM built based on 100 healthy pelvic bones segmented automatically in Simpleware Medical, and used to virtually reconstruct the native pelvic morphology in all patient cases. The SSM-based models were then compared to the postoperative computed tomography (CT)-based reconstructions to evaluate the differences in CoR between the diseased hip and the SSM-based reconstruction. The model was validated for accuracy using healthy anatomy.

Results demonstrated that the median (interquartile range (IQR)) difference in CoR between the diseased hip and its SSM-based reconstruction was 3 mm (IQR: 3 to 6 mm). When validating the model, the median difference in CoR between the healthy hemipelvises and their SSM-based models was 2 mm (IQR: 2 to 4 mm), and the median difference in CoR was 1 mm (IQR: 1 to 2 mm) for X, 1 mm (IQR: 1 to 2 mm) for Y, and 1 mm (IQR: 1 to 1 mm) for Z. The restoration of the hip CoR and leg length are key factors in achieving good clinical outcomes. The team had previously showed that SSM modeling can be used to reconstruct the absent bony landmarks of patients with significant bony defects aiding planning and implant design. This is this first study of using an SSM for patients who underwent primary THA aiding the reconstruction of the native anatomy showing the potential impact of the tool to aid future preoperative planning spanning from primary, complex primary, and revision THA.

Related Papers

• De Angelis, S., Henckel, J., Bergiers, S., Hothi, H., Di Laura, A., & Hart, A., 2023. Statistical shape modeling of the large acetabular defect in hip revision surgery. Journal of Orthopaedic Research, 41(9), 2016-2025.

Advances in additive manufacturing (AM) and 3D computed tomography (CT) have led to a relatively new class of custom 3D printed titanium implants that help surgeons to manage large acetabular defects in revision hip surgery. However, there is no reference for a clinically significant migration threshold of such implants and bony defects, or long-term outcome studies of using custom 3D printed implants. This study reviewed component migration three years post-acetabular reconstruction surgery through bone-to-bone registration of sequential CT imaging. The objectives of the study were to assess implant position three years after surgery in comparison with one-year postoperative CT imaging and to assess clinical outcomes.

A single-centre cohort study was carried out of 20 patients from a single surgeon, with all patients having had large acetabular defect classified as Paprosky type 3B, who received a 3D printed custom acetabular cup including a dual mobility bearing type in all cases. CT images were rendered in Simpleware software to produce 3D reconstructions of the patients’ bony pelvis for relative comparison of two imaging points at one-year and 3-year post-operatively. Bone-to-bone registration enabled an assessment of implant movement over time, in terms of the difference in centre of rotation (CoR) in X (medial-lateral, ML), Y (inferior-superior, IS), and Z (anterior-posterior, AP) planes. Meticulous patient follow-up by the operating surgeon was undertaken to closely monitor for complications.

The outcome measures for the study were the change in the centre of rotation (CoR) between CT scans carried out after one year, and three-year post-operatively, and the related clinical outcomes.  The mean follow-up time for patients was 64 months (42 to 84 months; 3 to 7 years). After three years from surgery, the deviation of CoR was studied as a mean for different planes, with postoperative imaging revealing no acute periprosthetic fractures, and no evidence of implant loosening or failure of metalwork. Bone in-growth was visible from 12 months onwards, as observed in the acetabular wall and roof, where the implant presents its maximum porosity and around the flanges. Overall, the migration was lower 3 years post-op compared to the 1-year-follow-up in the three planes and significantly lower in the inferior-superior direction. The 3D analysis revealed local bony shifts around the implants which suggested bone remodelling was taking place.

When looking at implant-related complications, there were 2 cases that dislocated but were successfully treated and 2 episodes of infection were recorded, the patients were being treated with antibiotics. There was no mechanical failure, and there were no fractures.

Although it is challenging to monitor 3D printed custom-made titanium cups, advances in medical image analysis can help track changes for in-vivo performance. The study demonstrated that clinically well-fixed 3D printed custom implants used to reconstruct massive acetabular defects show minimal migration three years after surgery, the degree of movement lowers over time.

The use of accurate pre-operative planning and custom 3D printed implants therefore shows promising results in improving complex hip revision surgeries.

Related Papers

• De Angelis, S., Di Laura, A., Ramesh, A., Henckel, J., & Hart, A., 2024. The role of bone remodeling in measuring migration of custom implants for large acetabular defects.Journal of Orthopaedic Research, 42(8), 1791-1800.
• Ramesh, A., Di Laura, A., De Angelis, S., Henckel, J., & Hart, A, 2024. Bone remodeling after revision total hip arthroplasty for large acetabular defects. Journal of Orthopaedic Research, 42(12), 2784-2795.

Acetabular osteolysis leading to component loosening is a major problem affecting long-term survival of hip arthroplasty, with the reconstruction of large acetabular defects a challenge due to the extent, location of osteolysis, and metal artefacts when visualising bone on CT images. As off-the-shelf solutions often fail to provide an adequate fitting for the host bone, excessive reaming of the bone is needed for implant positioning, which can lead to a high discrepancy between planned and achieved position of the component.

In this study, researchers aimed to quantitatively evaluate the position of a series of next generation custom-made 3D printed implant through the difference between planned and achieved CoR, and calculation of cup inclination and version angles. A retrospective cohort study was carried out using nine patients with Paprosky type III defects that had received a custom-made titanium 3D printed cup from a single manufacturer, including a built-in dome custom augment to fill in defects for the patient-specific anatomy.

The evaluation found that imaging techniques using CT scanning and 3D image reconstruction can help optimize implant design by more precisely filling the defect, with the custom augments serving as a guide to implant positioning that facilitated the insertion of screws into the most optimal areas of the bone. As a result, the need for extensive bone preparation to accommodate the implant was reduced, and the iliac flanges fitted against the bone. With this approach, next generation implant design is feasible, with minimal discrepancy between planned and achieved positions. Further work is needed to improve the imaging, manufacturing, and surgical implantation workflow.

Related Papers

• Henckel, J., Ramesh, A., Hothi, H., Richards, R., Di Laura, A., & Hart, A., 2023. The accuracy and precision of acetabular implant measurements from CT imaging. Frontiers in Bioengineering and Biotechnology, 11, 1150061.
• Di Laura, A., Henckel, J., & Hart, A., 2023. Custom 3D-printed implants for acetabular reconstruction: Intermediate-term functional and radiographic results. JBJS Open Access, 8(2), e22.
• Durand‐Hill, M., Henckel, J., Di Laura, A., & Hart, A. J., 2020. Can custom 3D printed implants successfully reconstruct massive acetabular defects? A 3D‐CT assessment. Journal of Orthopaedic Research, 38(12), 2640-2648.

The final position of a femoral stem component in uncemented total hip arthroplasty (THA) depends on both the shape of the femur and the design of the prosthetic stem. Various design approaches are currently used for an uncemented femoral stem, but despite their satisfactory long-term survivorship, the anatomically-shaped stems fail to cover the wide range of shape variability of the intramedullary canal of the proximal femur. Delivering the intended stem version in uncemented THA is therefore an unmet need, with this study aiming to better understand the shape variations that characterise the intramedullary femoral canal and control the fit of the prosthesis. Principal Component Analysis (PCA) was used to identify the main modes of variation.

A retrospective cohort study was carried out using 64 preoperative pelvic CT scans of patients who had undergone 3D planned hip replacement surgery. Simpleware software was used to generate 3D reconstructions of the patient’s femoral canal using an appropriate Hounsfield Unit (HU) threshold, and the SSM was built on these models. The input data was then standardised for length and orientation, and a coordinate system was defined for each femoral canal, based on the posterior condylar axis and the intertrochanteric crest; this method allowed for alignment in a fixed reference system before mean shape generation. Each input femoral canal was mapped onto the mean shape using point mapping, with PCA used to extract the directions of variation (eigenvectors) and extent of change (eigenvalues) from the covariance matrix of the data. Outcome measures were the principal modes of variation and the variance of each mode (the contribution of each of the main modes of variation).

Five modes of variation were identified, including canal size, proximal torsion around the calcar, femoral version, varus/valgus orientation, and distal femoral twist, in order of decreasing contribution to the overall variance. These results help to better describe the variability in the intramedullary femoral canal shape in 3D, with distinguishing features identified as size, femoral torsion fixed at the calcar, proximal femoral torsion, varus/valgus orientation, and distal femoral twist. A unique stem design accounting for all these patient characteristics does not yet exist, explaining the current discrepancy in prosthetic stems; this study may then be used to improve the planning and the delivery of anteversion of an uncemented stem.

Related Papers

• Moralidou, M., Di Laura, A., Henckel, J., Hart, A. J., 2023. Can version of the proximal femur be used for CT planning uncemented femoral stems? Medical Engineering & Physics, 116, 103985.
• Belzunce, M. A., Henckel, J., Di Laura, A., Hart, A., 2020. Uncemented femoral stem orientation and position in total hip arthroplasty: a CT study.Journal of Orthopaedic Research, 38(7), 1486-1496.
• Ramesh A, Henckel J, Hart A, Di Laura A., 2024. Understanding the variability of the proximal femoral canal: A computational modeling study. Journal of Orthopaedic Research, 1-10.