Novel mathematical model for preoperatively predicting pelvic tilt in patients with thoracolumbar kyphosis due to ankylosing spondylitis after three-column osteotomy.

Objectives: This study aimed to introduce a novel mathematical model for preoperative precalculated pelvic tilt (PT) in patients with thoracolumbar kyphosis due to ankylosing spondylitis (AS) after three-column osteotomy.

Methods: A total of 20 patients with AS, including 19 men and one woman, who underwent three-column osteotomy from April 2017 to April 2021, in the study hospital were retrospectively reviewed. Spinopelvic parameters, including global kyphosis, pelvic incidence, sacral slope, PT, sagittal vertical axis, horizontal distance between hip axis and hilus pulmonis (HDHH), measured on preoperative, postoperative, and final follow-up radiographs were analyzed. A coordinate diagram was created on the lateral spine radiographs with the hip axis as the origin. The distances and angles between the osteotomy apex, hilus pulmonis, and hip axis were measured, and a mathematical model was established using basic vector functions. The planned osteotomy angle was substituted into the mathematical model to obtain precalculated postoperative PT. Paired sample t-test was performed to determine the differences between planned osteotomy angle and actual osteotomy angle and between predictive postoperative PT and actual postoperative PT.

Results: Seven patients underwent single-level osteotomy, and 13 patients underwent two-level osteotomy. No significant difference was observed between the planned osteotomy angle and the actual osteotomy angle. No statistically significant difference was found between the precalculated postoperative PT and the actual postoperative PT.

Conclusions: The novel mathematical model was reliable in predicting postoperative PT in patients with AS undergoing three-column osteotomy.