RESUMO
<p><b>OBJECTIVE</b>To compare the inter- and intra-observer reliability of the GATA and SMU classification systems for spinal tuberculosis and assess the clinical value of SMU classification.</p><p><b>METHODS</b>One hundred patients with spinal tuberculosis treated in our hospital from January 2004 to December 2011 were randomly selected for analysis, including 54 males and 46 females with a mean age of 45 years (range, 16-68 years). All the patients had X-ray, CT and MRI examinations. Five observers experienced in spinal tuberculosis independently assigned the classification using the GATA and SMU classification systems, and the assignment was repeated 3 months later to test its reproducibility. Kappa value was used to determine the intra- and inter-observer reliability.</p><p><b>RESULTS</b>For GATA and SMU classification systems, the inter-observer percentage of agreement averaged (59.9∓4.84)% (κ=0.412∓0.058) and (81.6∓6.06)% (κ=0.753∓0.068), and the intra-observer percentage of agreement was (75.6∓5.27)% (κ=0.624∓0.078) and (89.8∓2.28)% (κ=0.862∓0.037), respectively.</p><p><b>CONCLUSION</b>The SMU classification system of spinal tuberculosis has a higher inter-observer and intra-observer reliability than the GATA classification system, but its clinical value needs to be further tested in future clinical trials.</p>