Reliability analysis for digital adolescent idiopathic scoliosis measurements

Timothy R. Kuklo*, Benjamin K. Potter, Michael F. O'Brien, Teresa M. Schroeder, Lawrence G. Lenke, David W. Polly

*Corresponding author for this work

Research output: Contribution to journalArticlepeer-review

64 Scopus citations


Objective: Analysis of adolescent idiopathic scoliosis (AIS) requires a thorough clinical and radiographic evaluation to completely assess the three-dimensional deformity. Recently, these radiographic parameters have been analyzed for reliability and reproducibility following manual measurements; however, most of these parameters have not been analyzed with regard to digital measurements. The purpose of this study is to determine the intra- and interobserver reliability of common scoliosis radiographic parameters using a digital software measurement program. Methods: Thirty sets of preoperative (posteroanterior [PA], lateral, and side-bending [SB]) and postoperative (PA and lateral) radiographs were analyzed by three independent observers on two separate occasions using a software measurement program (PhDx, Albuquerque, NM). Coronal measures included main thoracic (MT) and thoracolumbar-lumbar (TL/L) Cobb, SB MT Cobb, MT and TL/L apical vertical translation (AVT), C7 to center sacral vertical line (CSVL), T1 tilt, LIV tilt, disk below lowest instrumented vertebra (LIV), coronal balance, and Risser, whereas sagittal measures included T2-T5, T5-T12, T2-T12, T10-L2, T12-S1, and sagittal balance. Analysis of variance for repeated measures or Cohen three-way κ correlation coefficient analysis was performed as appropriate to calculate the intra- and interobserver reliability for each parameter. Results: The majority of the radiographic parameters assessed demonstrated good or excellent intra- and interobserver reliability. The relationship of the LIV to the CSVL (intraobserver κa = 0.48-0.78, fair to excellent; interobserver κa = 0.34-0.41, fair to poor), interobserver measurement of AVT (ρ = 0.49-0.73, low to good), Risser grade (intraobserver ρ = 0.41-0.97, low to excellent; interobserver ρ = 0.60-0.70, fair to good), intraobserver measurement of the angulation of the disk inferior to the LIV (ρ = 0.53-0.88, fair to good), apical Nash-Moe vertebral rotation (intraobserver ρ = 0.50-0.85, fair to good; interobserver ρ = 0.53-0.59, fair), and especially regional thoracic kyphosis from T2 to T5 (intraobserver ρ = 0.22-0.65, poor to fair; interobserver ρ = 0.33-0.47, low) demonstrated lesser reliability. In general, preoperative measures demonstrated greater reliability than postoperative measures, and coronal angular measures were more reliable than sagittal measures. Conclusions: Most common radiographic parameters for AIS assessment demonstrated good or excellent reliability for digital measurement and can be recommended for routine clinical and academic use. Preoperative assessments and coronal measures may be more reliable than postoperative and sagittal measurements. The reliability of digital measurements will be increasingly important as digital radiographic viewing becomes commonplace.

Original languageEnglish
Pages (from-to)152-159
Number of pages8
JournalJournal of Spinal Disorders and Techniques
Issue number2
StatePublished - Apr 2005
Externally publishedYes


  • Adolescent idiopathic scoliosis
  • Digital measurement
  • Interobserver reliability
  • Intraobserver reliability
  • Radiographic parameter


Dive into the research topics of 'Reliability analysis for digital adolescent idiopathic scoliosis measurements'. Together they form a unique fingerprint.

Cite this