A prospective study to validate the effectiveness of push-traction film (PTF) in assessment of curve flexibility in adolescent idiopathic scoliosis (AIS).The correction rate(CR) of post-operation (Post-op), supine side-bending (SB), suspension and fulcrum bending (FB) radiographs were compared with PTF in instrumented main thoracic (MT) and thorocolumbar/lumbar (TL/L) curves.Correlation and regression analyses confirmed that PTF was a more stable and accurate method to predict flexibility.Introduction There is no agreement amongst surgeons about the most advantageous method in flexibility evaluation of scoliosis.As all methods available provide the orthopedic force from one direction and use a single torque, it is difficult for them to achieve the postoperative correction, nor could they meet the needs for different types of curves.Methods.Precisely controlled bidirectional (push and traction) orthopedic forces were applied for curve flexibility evaluation in 31 consecutive AIS patients.The correction rate(CR) of post-operation (Post-op), supine side-bending (SB), suspension and fulcrum bending (FB) radiographs were compared with PTF in instrumented main thoracic (MT) and thorocolumbar/lumbar (TL/L) curves.Correlation and linear regression analyses were also been done to find the best predictor among the four methods.Results.In MT group, CR of PTF was significantly higher than that of SB (P=0.010) and suspension (P=0.000) but not significantly different from that of FB (P=0.335).In ML/L group,CR of PTF was significantly higher than that of suspension (P=0.000), but not significantly different from that of SB (P=0.681) and FB (P=0.382).There was no significant difference between CR of PTF and Post-op both in MT (P=0.122) and ML/L(P=0.068) groups.Correlation and linear regression analyses showed that PTF provided the highest correlation of the four methods, with the postoperative angle both in MT(r=0.957) and MT/L group (r=0.779).Conclusion.Although it did not achieve the best CR among the four methods studied, correlation and regression analyses confirmed that PTF was a more stable and accurate method to predict flexibility.We believe that further exploration of a more rational push-traction force ratio would help obtain a better flexibility.Significants: This was the first report about using precisely controlled bidirectional correction forces for curve flexibility evaluation.Correlation and regression analyses confirmed that PTF was a more stable and accurate method to predict flexibility.