JMIR Publications

ABSTRACT

Background:

Spinopelvic motion, the cornerstone of the sagittal balance of the human body, is pivotal in patient-specific total hip arthroplasty. Purpose: This study aims to develop a novel model using back propagation neural network (BPNN) to predict how the pelvic parameters change when one sits down based on standing lateral spinopelvic radiographs.

Methods:

Healthy volunteers of 18 to 30 years old were screened for symptoms and radiographic signs for the spine, pelvic and hip diseases. On their standing and sitting radiographs, 18 spinopelvic parameters were taken, such as pelvic incidence (PI), pelvic tilt (PT) and so on. First, standing parameters correlated with sitting PT and SS were identified via Pearson correlation. Then, with these parameters as inputs and sitting PT and SS as outputs, the BPNN-based prediction network was established. Finally, the prediction results were evaluated based on relative error (RE), prediction accuracy (PA), and normalized root mean squared error (NRMSE).

Results:

The study included 145 volunteers of 23.1±2.3 (19- 29) years old, of whom 51 were men and 94 were women. Pearson analysis revealed sitting PT was correlated with six standing measurements and sitting SS with five. The best BPNN model achieved 78.48% and 77.54% accuracy in predicting PT and SS, measures changing per pelvic tilt. As for PI, a constant for pelvic morphology, the prediction accuracy was 95.99%. Conclusion: In this small cohort study, the BPNN model yielded desirable accuracy in predicting sitting spinopelvic parameters. The study provides new insights and tools for characterizing spinopelvic parameters throughout the motion cycle.