A clinical prediction model for the selection of cage height in oblique lumbar interbody fusion

Aims To develop and validate a preoperative clinical prediction model for the cage height selection in L4-5 oblique lumbar interbody fusion (OLIF) procedure supports intraoperative decision-making. Patients and Methods In this retrospective and prospective study, patients who received L4-5 OLIF were followed to derive and validate a cage height selection prediction model from 2017 to 2021. In the retrospective cohort, the types of cage height were defined as the dependent variable, and fourteen parameters were screened for identifying independent variables. Then decision tree analysis was applied to establish the prediction model of the cage height selection, and an internal validation was determined by the Confusion-matrix chart method and the receiver operating characteristic curve. In the prospective cohort, the external validation was performed based on the prediction model. Results In the retrospective cohort, a total of 284 qualified cases were included in this study. The midpoint intervertebral height (MIVH) (B = 1.48, 95%CI: 1.14–1.83, P = 0.00) and the range of motion (ROM) (B = 0.42, 95%CI: 0.28–0.56, P = 0.00) were demonstrated as independent influential variables for the cage height. Ultimately, the cage height decision tree was established with MIVH and ROM. And the internal validation showed the model with overall sensitivity of 90%, specificity of 94%, positive predictive value of 89%, negative predictive value of 94%, and predictive accuracy of 87% with 55 cases. In the prospective cohort, 112 qualified cases were used for the external validation, which showed the prediction model with sensitivity of 89%, specificity of 94%, positive predictive value of 81%, negative predictive value of 94%, and predictive accuracy of 86%. Conclusion This study developed a clinical prediction model for the selection of cage height in L4-5 OLIF procedure. The model has excellent prediction efficiency in the internal and external validation, which may provide some reference for the selection of cage height.