Level specific hounsfield unit thresholds as a predictor of subsidence following transforaminal and posterior lumbar interbody fusion

The purpose of this study was to (1) determine if Hounsfield unit values are associated with radiographic subsidence following transforaminal and posterior lumbar interbody fusion procedures, (2) determine clinically sensitive level-specific Hounsfield unit thresholds for subsidence, (3) evaluate which vertebral level is most predictive for radiographic subsidence, (4) identify if there is a difference in radiographic subsidence between expandable and non-expandable cages, and (5) determine if there is any association between Hounsfield units and patient reported outcome measures. A retrospective analysis was performed identifying patients who underwent single-level TLIF/PLIF from 2007-2022. The average Hounsfield units from L1-S1 were measured using Computed tomography taken within one year of surgery. Changes in segmental lordosis on intraoperative and postoperative lateral spinal radiographs were also measured. A total of 65 patients met inclusion criteria. Average follow-up time was 21.4 months. Nineteen (29%) patients had evidence of radiographic cage subsidence. On univariate analysis, L1-S1 were significantly associated with cage subsidence. Confounding was controlled with multivariate analysis demonstrating that L1-L3 and L5-S1 remained significantly associated with cage subsidence. Receiver operating characteristic curve analysis revealed that cutoff values of 215.5, 203.7, 174.2, 181.6, 205.5, 246.8 Hounsfield units at L1-S1 respectively had ideal sensitivities for screening. The results of the study suggest that there is a significant association between Hounsfield units and radiographic subsidence following lumbar interbody fusion. Measurements taken at the L5 vertebral level were found to be most predictive of radiographic subsidence. There is no difference in radiographic subsidence between expandable and non-expandable cages. These findings suggest that the use of Hounsfield units derived from computed tomography scans shows some promise as a tool to predict risk of cage subsidence. • Level-specific Hounsfield unit (HU) thresholds are significantly associated with cage subsidence after lumbar interbody fusion, with L5 and S1 being the most predictive vertebral levels. • HU cutoffs prioritizing sensitivity (>90%) may serve as a practical screening tool to identify patients at risk for subsidence using preoperative imaging. • No significant difference in subsidence was observed between expandable and static cages, suggesting cage type may not influence subsidence risk.