Recently we developed a personalised X-ray reconstruction-based planning and post-operative treatment evaluation system called iLeg for total knee arthroplasty or lower extremity osteotomy. Based on a patented X-ray image calibration cage and a unique 2D-3D reconstruction technique, iLeg can generate accurate patient-specific 3D models of a complete lower extremity from two standing X-rays for true 3D planning and evaluation of surgical interventions at the knee joint. The goal of this study is to validate the accuracy of this newly developed system using digitally reconstructed radiographs (DRRs) generated from CT data of 12 cadavers (24 legs). Our experimental results demonstrated an overall reconstruction accuracy of 1.3±0.2mm.

Tracked B-mode ultrasound (US) potentially provides a non-invasive and radiation-free alternative to percutaneous pointer digitization for intra-operative determination of the anterior pelvis plane (APP). However, most of the published approaches demand a direct access to the corresponding landmarks, which can only be presumed for surgical approaches with the patient in supine position. In order to avoid any change of the clinical routine for total hip arthroplasties (THAs), we propose a new method to determine the pelvic orientation, which could be performed in lateral position.

Our proposed method is based on the acquisition of ultrasound images of the ipsilateral hemi-pelvis, namely the posterior superior iliac spines (PSISs) and iliac crest region. The US images are tracked by a navigation system and further processed to extract three-dimensional point clouds. As only one side of the pelvis is accessible, we estimate the symmetry plane (midsagittal plane) of the pelvis based on additionally digitized bilateral anterior superior iliac spine (ASIS) landmarks. This symmetry plane is further used to mirror the ipsilateral US-derived points to the contralateral side of the pelvis and to register and instantiate a pelvic SSM constructed from 30 CT-scans.

The proposed registration method was evaluated using two plastic pelvis models and two cadaveric pelvises together with special custom-made silicone phantoms to simulate the missing soft-tissue. In each trial, the required data were collected with the pelvis rigidly fixed in lateral decubitus position together with ground truth APP landmarks. A registration error of 3.48° ± 1.10° was found for the anteversion angle, while the inclination angle could be reconstructed with a mean error of 1.26° ± 1.62°.

The performed in-vitro experiments showed reasonably good results, taking the sparsity of the input point clouds into consideration.

The integration of statistical shape models (SSMs) for generating a patient-specific model from sparse data is widely spread. The SSM needs to be initially registered to the coordinate-system in which the data is acquired and then be instantiated based on the point data using some regressing techniques such as principal component analysis (PCR). Besides PCR, partial least squares regression (PLSR) could also be used to predict a patient-specific model. PLSR combines properties of PCR and multiple linear regression and could be used for shape prediction based on morphological parameters.

Both methods were compared on the basis of two SSMs, each of them constructed from 30 surface models of the proximal femur and the pelvis, respectively. Thirty leave-one-out trials were performed, in which one surface was consecutively left out and further used as ground truth surface model. Landmark data were randomly derived from the surface models and used together with the remaining 29 surface models to predict the left-out surface model based on PCR and PLSR, respectively. The prediction accuracy was analysed by comparing the ground truth model with the corresponding predicted model and expressed in terms of mean surface distance error.

According to their obtained minimum error, PCR (1.62 mm) and PLSR (1. 63 mm) gave similar results for a set of 50 randomly chosen landmarks. However PLSR seems to be more susceptible to a wrong selection of number of latent vectors, as it has a more variation in the error.

Although both regression methods gave similar results, decision needs to be done, how to select the optimal number of regressors, which is a delicate task. In order to predict a surface model based on morphological parameters using PLSR, the choice of the parameters and their optimal number needs to be carefully selected.