Accuracy and Reliability of Synthetic Computed Tomography for Model-Based Tracking of Biplane Videoradiography Data.

Abstract

Biplane videoradiography (BVR) accurately measures hip kinematics in vivo. One downside of BVR is that it typically requires a pelvic computed tomography (CT) scan to reconstruct surfaces for model-based tracking (MBT), increasing the radiation burden to participants. Synthetic CT (sCT) from magnetic resonance imaging (MRI) provides similar volumetric image-based data. The purpose of this study was to assess the effect sCT has on the accuracy and reliability of in vivo hip kinematics from BVR data. Three asymptomatic participants received CT and MRI scans and were imaged with BVR while standing and during gait. Kinematics were calculated from femur and pelvis surfaces reconstructed from CT and sCT scans. Accuracy and inter-/intra-reviewer reliability of hip kinematics using sCT were assessed against CT as the reference using mean absolute error (MAE), mean difference (MD), precision, and inter-/intra-class correlation coefficients (ICCs). Kinematics were represented with and without correction for standing posture. MAE, MD, and precision of posture-corrected sCT kinematic results were 0.85, 0.60, and 0.71° for rotations and 0.40, 0.23, and 0.39 mm for translations, respectively. Accuracy of posture-corrected joint angles was within 0.17° of prior MBT validation in the hip, while translation accuracy and precision were greater herein. Kinematic results using CT and sCT had similar reliability. The results of this study establish novel reliability metrics for in vivo hip kinematics from BVR. Findings indicate that sCT can be used in lieu of CT to measure in vivo hip kinematics, reducing radiation burden of BVR hip kinematic protocols by up to 73%.