Abstract: Objective To compare the confidence and repeatability of Castellvi classification system and novel classification system for lumbosacral transitional vertebrae and study the clinical value of novel classifcation system. Methods Eighty patients (47 males and 33 females) aged 16 -55 years (mean 36.7 years) were included in this study. Arterioposterior and lateral lumbar vertebrae X-ray flm, transitional vertebrae 3D CT reconstruction flm, and lumbosacral vertebrae MRI flm were taken for each patient. The flms were classifed by 4 experienced spine surgeons according to the two classifcation systems and repeated after two weeks. The confdence and repeatability of the two classifcation systems were compared. The consistency of Kappa test was calculated. Results The confdence of the two classifcation systems was 84.83% (Kappa coeffcient=0.789) and 87.67% (Kappa coeffcient=0.892), respectively. The repeatability of the two classifcation systems was 90.75% (Kappa coeffcient=0.895) and 93% (Kappa coeffcient=0.914), respectively. Conclusion The novel classifcation system is more objective, succinct, and easier to keep it in heart than the Castellvi classifcation system, and can thus be used in treatment of vertebral diseases.