Prediction of fusion cage subsidence after transforaminal lumbar interbody fusion based on MRI endplate bone quality.

BACKGROUND

Transforaminal lumbar interbody fusion (TLIF) is a significant surgical method for treating degenerative lumbar diseases. Cage subsidence is one of the common postoperative complications of TLIF surgery. Current research lacks imaging prediction methods for cage subsidence in TLIF surgery. Endplate bone quality (EBQ) based on magnetic resonance imaging (MRI) is currently regarded as a potentially effective method for predicting cage subsidence. This study aimed to investigate the correlation between EBQ and cage subsidence.

METHODS

A retrospective analysis was conducted on the data of 102 patients who underwent TLIF surgery between October 2022 and October 2023. Patients' vertebral bone quality (VBQ) and EBQ scores were measured using MRI. Patients were divided into a subsidence group and a non-subsidence group based on whether the postoperative fusion device subsided. The preoperative VBQ, EBQ, bone mineral density (BMD) values, preoperative and postoperative 3-month and final follow-up Visual Analogue Scale (VAS) scores, Oswestry Disability Index (ODI) scores, disc height (DH), lumbar lordosis (LL) angle, and segmental lordosis (SL) angle were compared between the two groups. Logistic regression curves were used to study the correlations.

RESULTS

Among 102 patients, 23 patients (22.5%) experienced cage subsidence. The DH, LL, and SL in both groups increased postoperatively, at 3 months postoperatively, and at the final follow-up when compared to preoperative values (P<0.05). The EBQ score was 2.63±0.40 for the non-subsidence group and 3.40±0.49 for the subsidence group, with a significant difference between the two groups. The EBQ score is a predictive factor for cage subsidence after TLIF. A higher score significantly increases the risk of cage subsidence [odds ratio (OR) =74.95; 95% confidence interval (CI): 11.8-472; P<0.001]. On the receiver operating characteristic (ROC) curve, the area under the curve (AUC) for the EBQ score was 0.863 (95% CI: 0.815-0.924), and the optimal threshold for the EBQ score was 3.125 (sensitivity: 84.3%; specificity: 72.2%).

CONCLUSIONS

The EBQ score can predict the occurrence of cage subsidence in patients undergoing TLIF surgery. Patients with higher EBQ scores have a higher incidence of cage subsidence.