Serial Study Of Cosmetic Shoulder Balance In Scoliosis

Objective: To evaluate the discrepancy between cosmetic shoulder balance andradiographic material, the efficiency of newly-defined radiographic parameter toassess cosmetic shoulder balance, and phenomenon of spontaneous cosmetic shoulderbalance correction after surgery in adolescent idiopathic scoliosis.Background: Shoulder balance is an important aspect of physical deformity inadolescent idiopathic scoliosis, and cosmetic defect is critical to self-confdence.Meanwhile, radiographs clearly revealing the inner skeletal asymmetry especiallyspine, provide information for deformity evaluation and treatment choice. It would beideal that radiological measures make effective clinical prediction to cosmetic feature.However, correlation analysis used in several reports showed none of radiographicparameters getting strong correlation with cosmetic shoulder balance (r<0.8),dramatic improvements in the radiological deformity did not necessarily transfer intopatients’satisfaction.Methods:1.Atotal of120AIS patients with main thoracic curve admitted to ourhospital were reviewed. Spinal radiographs and shoulder photographs were taken forall the patients in neutral standing position. Radiographic parameters clavicleangle(CA), clavicle difference(CD), clavicle-rib cage intersection(CRCI), coracoidprocess height(CPH), frst rib–clavicle height(FRCH), trapezius length(TL) andcosmetic shoulder different(CSD) were analyzed using correlation and subtractionanalysis.2. Ayes-or-no questionnaire and receiver operating characteristic (ROC)curve analysis were made to evaluate the effectiveness of radiographic parameters topredict subjective balance-or-not judgment of cosmetic shoulder.3. The newly-defined radiographic parameter coracoid process clavicle difference(CPCD) wasmeasured from radiographs, and compared with CA, CD, radiological shoulderheight(RSH) and shoulder height difference(SHD) by correlation, subtraction andROC analysis.4. Radiographic parameters and CSD were measured form56LenkeⅠcurves, t-test was made to explored the difference of CSD after surgry immediately and at last follow-up. Person correlation was used between6factors and spontaneouscosmetic shoulder balance correction (SCSC).5. Three groups were dividedaccording to preoperative shoulder balance: right shoulder lifted, balanced shoulder,left shoulder lifted. Cosmetic shouder balance and SCSC were studied when differentupper instrumented vertebrea was chosen.Results:1. CD and CSD had a excellent correlation（r＝0.871，p＜0.05）, while CA,CRCI, CPH moderate correlation, FRCH negative correlation, and TL had none.2.CD and CSD had the lowest D-value0.57cm±0.52cm, while3patients had theD-value more than1cm.3. Area under curve (AUC) for CA and CD were0.742—0.806；CRCI，CPH，FRCH，TL were0.509—0.678. Two cut-off points were foundfor each radiographic parameter.4. CPCD and CSD had the highest correlationcoefficient r＝0.978, no difference between the two, and the largest AUC0.955and0.945.5. There were significant difference between CSD after surgry immediatelyand at last follow-up, t=3.57，P=0.0008（P＜0.05）.6.51of the56LenkeⅠpatientshave CSD＜1.0cm,4of the other had not noticed the unlevel cosmetic shoulder,1patient had left shoulder lifted and barely satisfied shouder balanced.Conclusion:1. Although CD and CSD had a high correlation which was differentfrom period studies, discrepancy did exist between radiographic and cosmeticparameters.2. Two different subjective critera of shoulder balance may provide morechoices for the upper instrumented vertebrae.3. The newly-defined radiographicparameter CPCD predicted cosmetic shoulder better.4. Cosmetic shoulder maychange into more balanced status for most patients after surgery, and the SCSC hadcorrelation with CB.5. Less upper thoracic vertebrae may be instrumented accordingto SCSC.