Study Design. Measurement reliability
study of adult spinal deformity (ASD) patient radiographs using intraclass correlation coefficients (ICC) and variance.
The aim of the study was to compare picture archiving and communication systems (PACS) to dedicated spine measurement
Summary of Background Data.
Accurate radiographic measurement
of sagittal alignment
is essential for evaluating ASD. PACS measurements often necessitate rudimentary techniques and estimations of anatomic landmarks and angles. Though SMS has been studied and validated, no studies directly compare PACS to SMS.
Eleven independent observers (7 spine surgeons, 4 researchers) digitally measured 20 ASD radiographs for pelvic incidence (PI), pelvic tilt (PT), lumbar lordosis (LL), PI-LL, thoracic kyphosis (TK), and sagittal vertical axis (SVA). Round 1 used PACS basic line/angle tools; Round 2 used a validated SMS that automatically calculates spino-pelvic parameters
from 6 user-identified landmarks. Means, coefficient of variance (CV), and intraclass correlation coefficients (ICC) were analyzed.
PACS measurements were significantly greater than SMS (PI, PT, PI-LL: P
< 0.0001), though within clinical and measurement
margins of error. Excluding TK, the variations in measurement
(CV) were significantly greater for PACS (14–34%) vs.
SMS (11–23%). Reliability
was greater in SMS than PACS for PI, PT, PI-LL, LL, and SVA. The greatest differences in intraclass correlation coefficients (ICC) between PACS and SMS were in PI (PACS: 0.647; SMS: 0.810) and PI-LL (PACS: 0.921; SMS: 0.970). Among surgeons, the differences between PACS and SMS were augmented, and SMS had higher intraclass correlation coefficients (ICC) than PACS for all parameters
(mean intraclass correlation coefficients [ICC] 0.931 vs.
0.861). Among surgeons, PI had the lowest reliability
(PACS: 0.505; SMS: 0.752) and SVA had the highest (PACS: 0.985; SMS: 0.994).
SMS provides significantly more reliable measurements than PACS, especially among surgeons. Consistent use of SMS in the evaluation and surgical planning of ASD patients appears necessary given the significant differences in values, variance, and reliability
between PACS and SMS.
Level of Evidence: 3