Comparison of manual and digital measurement of radiographic parameters in patients with adolescent idiopathic scoliosis (AIS).
To assess the reliability of digital measures as compared to manual measures in the evaluation of AIS.
Radiographic parameters are critical to the evaluation of patients with AIS, and are frequently used to monitor curve progression and guide treatment decisions. The reliability of many of the more common radiographic measures has only recently been elucidated for both manual and digital measures. However, a comparative analysis of manual versus digital measures has been performed only for coronal Cobb angles. The inter-technique reliability of these parameters will have increasing importance as digital radiographic viewing and analysis become commonplace.
There were 2 independent, blinded observers that measured 30 complete sets of preoperative (posterior-anterior, lateral, and both side-bending) and postoperative (posterior-anterior and lateral) radiographs on 4 different occasions. For the first 2 iterations, manual measurements were taken using the same pencil and protractor. For the last 2 iterations, measurements of digitized radiographs were taken on a software measurement program (PhDx, Albuquerque, NM). Coronal measures included the main thoracic and thoracolumbar/lumbar standing and side-bending Cobb angles, apical vertebral translation, coronal balance, T1 tilt angle, lowest instrumented vertebrae angle, angulation of the disc inferior to the lowest instrumented vertebrae, apical Nash-Moe vertebral rotation, and Risser grade. Sagittal parameters included T2–T5 and T5–T12 regional thoracic kyphosis, T2–T12 thoracic kyphosis, T10–L2 thoracolumbar junction sagittal curvature, T12–S1 lumbar lordosis, and global sagittal balance. The technique-dependent measurement variability and the inter-technique (manual vs. digital), intraobserver reliability were evaluated for each radiographic parameter (within 3°).
Digital measurement showed decreased intraobserver variability for many (9 of 15) of the radiographic parameters assessed. Likewise, digital measures indicated good or excellent correlation with the absolute values obtained with manual measurement for many (10 of 15) parameters. All but 1 of those parameters having moderate-to-poor correlation had been previously shown to have poor reliability, regardless of measurement technique. Statistically significant differences between measurement variability were noted for 6 measures, including 2 favoring digital and 4, manual. Significant differences in the absolute values were noted for 5 measures, determined at a difference of 3°. However, the differences in both parameter variability and absolute values tended to be small and of little clinical significance for manual versus digital measurement.
Digital measurement showed improved measurement precision and good correlation with manual measurements for the majority of AIS parameters. Absolute differences between manual and digital measurements were generally small. Therefore, digital measures are acceptable as a valid technique for scoliosis evaluation. The importance of digital versus manual measurement reliability will increase as digital radiographic viewing becomes more prevalent.
The reliability of radiographic measures is crucial for adolescent idiopathic scoliosis assessment. Digital measurement showed decreased measurement variability and good or excellent correlation with manual measurement values for the majority of common scoliosis radiographic parameters. The reliability of digital measurements will be increasingly critical as digital radiography becomes more prevalent, and comparison of digital and manual measurements becomes required more frequently for patient and study evaluation.
From the Departments of *Orthopaedics and Rehabilitation, Walter Reed Army Medical Center, Washington, DC, and †Orthopaedic Surgery, Miami Children’s Hospital, Miami, FL.
Acknowledgment date: May 27, 2004. First revision date: August 4, 2004. Acceptance date: August 6, 2004.
The manuscript submitted does not contain information about medical device(s)/drug(s).
Corporate/Industry funds were received in support of this work. Although one or more of the author(s) has/have received or will receive benefits for personal or professional use from a commercial party related directly or indirectly to the subject of this manuscript, benefits will be directed solely to a research fund, foundation, educational institution, or other nonprofit organization which the author(s) has/have been associated.
Address correspondence and reprint requests to Teresa M. Schroeder, MBA, 1314 Richmond Grand Ave., Orlando, FL 32820. E-mail: Teresa@humanframeworks.net