Advancements in Leg Malalignment Imaging

This thesis aims to improve the treatment of knee osteoarthritis (OA) by enhancing radiographic methods used in knee osteotomy, especially for patients with leg malalignment. It identifies biomechanical overload caused by malalignment as a distinct and important phenotype of knee OA, often overlooked in current clinical approaches. This recognition is particularly important given the challenges in treating younger, physically active patients for whom traditional OA treatments are insufficient. In this context, leg realignment through knee osteotomy emerges as a promising, joint-preserving surgical option that may be more effective than currently appreciated.

The first part of the thesis examines clinical practices in the Netherlands, revealing significant inconsistencies in knee osteotomy treatment protocols and preoperative planning among orthopedic surgeons. Despite widespread use of whole leg radiographs, best-practice planning methods were underutilized. The research confirms that physical examination is inadequate for diagnosing malalignment and that accurate, standardized imaging is essential. It demonstrates how leg rotation beyond 9 degrees and knee flexion during radiographic acquisition can compromise diagnostic accuracy. A novel radiograph acquisition protocol was developed and tested, showing high reproducibility and reliability, with measurement errors remaining under 1°.

In the second part, the focus shifts to improving the interpretation of radiographs through advanced technologies. An automated analysis tool, Orthopedic Digital Image Analysis (ODIA), was introduced to efficiently and consistently assess key OA parameters such as joint space width and subchondral bone density. This enables large-scale data analysis with high accuracy. Additionally, a mathematical model was developed to optimize preoperative planning by predicting how variations in hinge position and sawcut angle affect tibial slope correction. Further, the thesis explored the use of synthetic CT images derived from MRI to improve 3D landmark detection. These images accurately identified joint centers within clinically acceptable limits. Building on this, a patient-specific, resorbable osteotomy wedge was designed and tested. This 3D-printed scaffold, made from bioresorbable materials, successfully maintained structural integrity and supported cell growth, demonstrating its potential for personalized surgical correction.

The final part of the thesis focuses on better characterizing OA phenotypes related to malalignment. Using automated 3D segmentation of CT scans and ODIA, patients were stratified based on bone shape and intra-articular alignment. The study found that bones tend to shift toward a varus shape over time, contributing to disease progression even in patients with initially normal alignment. These findings suggest that detailed imaging and precise classification of malalignment types can enhance the prediction of OA progression and support more targeted treatment strategies.

Overall, this thesis establishes a foundation for more accurate diagnosis, planning, and personalization in knee osteotomy, promoting it as a powerful, underutilized tool in managing knee OA.