702 Impact of camera deviation on curvature assessments using goniometer and 2D photographies

Abstract

Camera deviations during 2D image acquisition might affect CA accuracy. Our aim was to investigate the impact of camera angle deviations on CA results. We performed an experiment in which 2D pictures were taken from 5 models with a known uniplanar curvature (40°, 45°, 60°, 90°, 120°). The model was kept on a fixed point and camera rotated around. Photographies were taken for every 10° increase in camera deviation from the optimal position (perpendicular to curvature plane). Camera rotated to a maximum of 90° deviation in both vertical and horizontal planes. Pictures were analyzed 2 times by 2 different urologists using a standard goniometer. We also calculated for each picture, using trigonometry, the expected apparent curvature (AC) and the corresponding expected picture assessment error (PAE= AC- model curvature). Intra-observer, inter-observer and observer-AC concordance correlation coefficient (CCC) were calculated. 100 pictures were analyzed. Intra-observer CCC was high (CCC:0.99) for both urologists. Inter observer and observer-AC correlation coefficient were also high (CCC:0.996 and CCC:0.992; respectively). When camera rotated in the horizontal axis, PAE tended to underestimate curvature for models with curvature of less 90°, and to overestimate curvature of the 120° model. When camera rotated in the vertical axis, PAE had an inverse effect. Analyzing its magnitude, irrespectively of model curvature, PAE showed a tendency to increase exponentially with higher deviation, reaching up to 100% for deviation of 80°. Nevertheless, PAE was always <5% for camera deviations of 0-20°