Comparative study of the clinical application of TLICS and AO-TLICS scores

Abstract

Objective To evaluate the reliability of the TLICS and AO-TLICS scoring system for the thoracolumbar fracture and the effectiveness of the final treatment scheme. Methods The clinical data of 56 adult patients with acute traumatic thoracolumbar fracture diagnosed in Department of Orthopedics, Chui Yang Liu Hospital Affiliated to Tsinghua University from June 2015 to June 2017 were analyzed retrospectively. There were 35 males and 21 females, aged (32.3±9.8) years, with an age range of 21-53 years. Two senior deputy chief orthopaedic physicians retrospective analyzed DR plain films, CT, MRI images and clinical records of thoracolumbar spine by independent blind method. According to TLICS and AO-TLICS scoring system(include injury morphology, neurologic status, posterior ligament complex (PLC) damage or M1 modifier, the severity scores of thoracolumbar fracture were classified and calculated. The score was divided into TLICS score group and AO-TLICS score group, each group of scores was the above 56 patients. Three months later, the two physicians repeated the above analysis process and compared the consistency of the scores between the observer and the observer′s own control. The consistency of the two groups was compared, and the accuracy, sensitivity and specificity of the final treatment were compared. Cohen kappa test was used for consistency comparison according to the score and the mean value was taken. Measurement data were expressed as mean±standard deviation(Mean±SD), and t test was used for inter-group comparison. Results In the TLICS score group, the Cohen kappa coefficients of the observer′s own control in fracture type, neurological status assessment, and PLC injury were 0.810, 0.966, and 0.698, respectively, and the total scores were consistent (k=0.727). Cohen kappa coefficients between the observers in fracture morphology, neurological status assessment, and PLC injury were 0.725, 0.931, and 0.594, respectively, and the total scores were consistent (k= 0.615). In the AO-TLICS score group, the Cohen kappa coefficients of the observer′s own control in terms of fracture type, neurological status assessment, and M1 correction factor were 0.760, 0.892, and 0.711, respectively, and the total scores were consistent (k=0.666). Cohen kappa coefficients of the observers in terms of fracture type, neurological status assessment and M1 correction factor were 0.707, 0.836 and 0.604, respectively, and the total score was consistent (k=0.592). According to the TLICS score, the correct rate of treatment was (84.2±2.8)%, the sensitivity was (84.1±2.1)%, and the specificity was (85.9±6.5)%. According to the AO-TLICS score, the correct rate of treatment was (89.0±2.6)%, the sensitivity was (88.3±3.5)%, and the specificity was (89.6±2.4)%. There was a statistically significant difference in the rate of correct rate (t=2.485, P=0.047). Conclusion The advantage of TLICS score is better consistency in interobservers or intraobservers, but the AO-TLICS score system is more comprehensive, and the choice of treatment plan is more instructive. Key words: Spinal fracture; Spinal cord injury; Soft tissue injury