A surface registration-based approach for assessment of 3D angles in guided growth interventions in the growing femur

Postoperative assessment of surgical interventions for correcting femoral rotational deformities necessitates a comparative analysis of femoral rotation pre- and post-surgery. While 2D assessment methods are commonly employed, ongoing debate surrounds their accuracy and reliability. To address the limitations associated with 2D analysis, we introduced and validated a 3D model-based analysis method for quantifying the angular and rotational impact of corrective rotational osteotomy in the growing femur.

The method is based on surface registration of the pre- and post-intervention 3D femoral models. To this end, 3D triangulated surface models were generated using CT images for the right femurs of 11 skeletally immature pigs, each scanned at two distinct time points with a 12-week interval between scans. In our validation procedures, femoral corrective rotational osteotomy of the post-12-week femur was simulated at varying angles of 5, 10, 15 and 20 degrees in three dimensions. Subsequently, a surface 3D/3D registration-based approach was applied to determine the 3D femoral angulation and rotation between the two models to assess the method's detection accuracy of the predefined twist angles as ground truth references.

The results document the precision and accuracy of the registration-based method in evaluating rotation angles. Consistently high accuracy was observed across all angles, with an accuracy rate of 92.97% and a coefficient of variance of 8.14%.

This study has showcased the potential for improving post-operative assessments with significant implications for experimental studies evaluating the effects of correcting rotational deformities in the growing femur.

Not applicable.