Surface topography imaging may provide an alternative to regular radiographic evaluation of spinal deformities and may assist in evaluating the result of their surgical correction. The aim of our study was to measure the accuracy and repeatability of measurements from a commercial imaging system (Quantec, UK) in patients with scoliosis. 8 subjects were measured on 3 different occasions and 6 images were obtained during each visit, 3 with arms hanging freely and 3 with hand fixation. Skin markers were placed over the spinous process of the first thoracic vertebra (T1), the line of spinous processes, and the posterior superior iliac spines. Thoracic kyphosis and lumbar lordosis curves were measured from a sagittal plane view. Thoracic and lumbar curves, pelvic tilt and vertical alignment of the spine were measured from a frontal plane view. Initially, significant errors were found in the accuracy of the equipment but following rectification of problems with the software, accuracy was better than ±5% (or ±2 mm, if larger) over the region of interest. Variance analysis of the data demonstrated that for a single reading, over half the repeatability error was attributable to the visit (probably inconsistency in placing skin markers), and the remainder with the individual image (patient movement and inconsistency in analyzing an image). Use of hand holds to attempt to stabilise the upper body did not improve the repeatability. For our normal procedure, taking the mean of measurements from 3 images, we deduced the 95% confidence interval to be:- frontal plane: upper curve 8.0°, lower curve 6.5°, pelvic tilt 5.5°, vertical alignment 2.0°; sagittal plane: thoracic kyphosis 7.0°, lumbar lordosis 6.5°.