Observer Variability of Radiographic Measurements of C2 (Axis) Fractures

STUDY DESIGN: Radiographic measurement study. OBJECTIVES: To assess the inter- and intraobserver reliability of angulation and displacement measurements for odontoid fractures and traumatic spondylolistheses (Hangman fractures) of C2, and to compare the reliability of measurements made on computerized tomography (CT) versus those made using plain radiographs. SUMMARY OF BACKGROUND DATA: Measurements of various parameters of C2 injuries can influence treatment decision making. However, there is a paucity of data regarding the reproducibility of these measurements. METHODS: A series of 19 cases, representing 10 odontoid fractures and 9 Hangman fractures, were prepared and distributed to surgeon raters on 2 occasions. Each rater was provided with a pictorial diagram illustrating the prescribed measurement technique(s). For odontoid fractures, raters were asked to measure angulation, translation, atlanto-dens interval (ADI), and posterior ADI. For Hangman fractures, angulation and C2-C3 translation were assessed. Measurements were made using plain films and CT images. Reliability was reflected by calculation of intraclass correlation coefficients and Pearson correlation coefficients. RESULTS: Intrarater analysis for both odontoid and Hangman fractures demonstrated high reproducibility. Interobserver evaluation showed that the posterior ADI was the most reproducible method, whereas all other measurements for odontoid fractures were found to have poor reproducibility. For Hangman fractures, the endplate method of angulation measurement had higher interrater reproducibility than the posterior tangent method. The interrater reliability for C2-C3 translation using plain films was higher than for measurements made using CT images. CONCLUSION: Despite widespread acceptance of their importance, radiographic measurements for C2 fractures have, at best, only moderate reliability. Even in the ideal situation created in this study, there is limited agreement. Further work to substantiate the importance and improve the reliability of these measurement parameters is needed.