Comparison of flexibility assessments in adolescent idiopathic scoliosis (Flexis Study)

Abstract

Introduction: The goals of surgical correction of adolescent idiopathic scoliosis (AIS) are to achieve a balance spine with a parallel fusion block. Preoperative curve flexibility assessment is integral in the radiographic evaluation and surgical decision making. The aim of this prospective study was to compare five radiographic techniques in flexibility assessment and correlate them to surgical correction. Material and Methods: A prospective comparative observational study was carried out in two academic institutions in Hong Kong and Turkey. Consecutive patients undergoing AIS surgical collections were recruited between June 2016 to August 2017. Preoperative radiographic evaluation included standing posteroanterior and lateral whole spine, supine, supine side bending, fulcrum bending (FB) in both sites; awake traction in Hong Kong; and supine traction under general anaesthesia (STUGA) in Turkey. Surgical correction and fusion levels were determined by the surgeons’ usual technique, and postoperative radiographs were taken. For each radiographic assessment, correction rate and flexibility percentages were calculated, and correction index was the correction rate as a percentage of the flexibility. Statistical differences were calculated. Results: 76 patients were recruited into the study, but 2 patients had anterior surgery and were excluded from analysis. 74 patients (65 females and 9 males) with an average age of 13.91 years (range, 11-18) at the time of operation. 51 patients had a thoracic major curve (34 cases Lenke 1; 10 Lenke 2; and 7 Lenke 3), and 23 patients had a thoracolumbar or lumbar major curve (21 cases Lenke 5 and 2 Lenke 6). For thoracic curves, mean preoperative standing frontal Cobb was 57.8o (range, 34.3o-79.7o), and the mean postoperative Cobb was 15.5o (range, 1-40.7o) with a mean correction rate of 72.5% (range, 27.1%-97.9%). For lumbar curves mean preoperative Cobb was 50.3o (range, 35o-88o), and the mean postoperative Cobb was 8.0o (range, 0.1o-22o) with a mean correction rate of 84.5% (range, 59.1%-99.8%). The correction indices showed that the most predictive dynamic assessment for AIS correction was FB, compared with STUGA (p ¼ 0.2), awake traction (p ¼ 0.004), supine side bending (p ¼ 0.003), and supine (p ¼ 0.000) for thoracic curves. For lumbar curves, FB, STUGA and supine side bending were similar in prediction (p > 0.05), which were better than supine film (p ¼ 0.000). Conclusion: Although correction rate can give an indication of the amount of correction achieved per case, it is inaccurate to compare between different cases as it does not take into account the intrinsic flexibility of the curve, and does not necessarily reflect on either the surgical technique or the instrumentation strategy. Correction index takes into account of the curve’s intrinsic flexibility, and the closer one gets to 100%, the more effective is the final surgical correction in taking up this flexibility. Our study showed that FB is most predictive in thoracic curves undergoing posterior instrumented fusion. In lumbar curves which are usually more flexible, no statistical significant difference was found amongst different dynamic radiographs in predicting surgical correction. An accurate flexibility assessment will allow improve pre-operative planning for the need of additional release and fusion level determination.