Interobserver reliability in evaluation of pedicle screw positions inserted with a modified technique

Abstract

The aim of this study was to evaluate a new technique for insertion of pedicle screws. The position of the screws was assessed on postoperative plain radiographs and computed tomography (CT) scans, and the interobserver reliability in evaluation of the pedicle screw positions was studied. The technique was applied to insert 201 pedicle screws in 27 patients with various spine conditions. The positions of the screws were evaluated blindly by two independent orthopaedic surgeons and two independent radiologists. Interobserver reliability was evaluated separately for analysis of plain radiographs and CT scans, as well as for the different spinal segments and for the different spine conditions treated. The rate of malpositioned screws was between 6.5% and 32.8% in plain radiographs and between 3.5% and 6.5% in CT scans according to the different observers. In plain radiographs, the rates of malpositioned screws in the upper thoracic, lower thoracic and lumbosacral spine segments were between 3.8%-39.6%, 10.0%-36.3%, 4.4%-23.5%, respectively. In CT scans, the rates of malpositioned screws in the upper thoracic, lower thoracic and lumbosacral spine segments were between 3.8%-13.2%, 2.5%-8.8%, and 0%, respectively. Interobserver reliability was found to be poor in radiographs and fair in CT scans. The technique used for insertion of pedicle screws was found to be simple and reproducible. Assessment of the screw positions with only plain radiographs was not found reliable. A detailed and standard classification system should be developed in order to improve interobserver reliability in assessing the positions of the screws.