MRI versus CT: a retrospective investigation of the feasibility and agreeability in post-operative evaluation of screw position after posterior lumbar interbody fusion.

Many patients receive magnetic resonance (MR) and computed tomography (CT) scans post-operatively to review screw placement. Traditionally, CT is diagnostic but as metal artefact reduction sequences are advancing in MR, the necessity for both MR and CT scans is questionable. The objective is to establish the statistical agreeability of MR and CT for evaluation of adequate screw placement in posterior lumbar interbody fusion. This opportunistic retrospective study of 58 patients investigated 297 images of 296 implanted screws. Post-operative MR and CT images were scrutinised for depiction of lumbar pedicle screw position using a 5-point scale. Kappa value for statistical agreeability tested MR against CT. The 297 images of screws resulted in strong to near-perfect agreement between MR and CT (n = 297 k = 0.8042 p < 0.025). MRI resulted in high sensitivity (88.7%) and positive predictive value (78.3%). MRI demonstrated very high specificity (96.2%) and negative predictive value (98.2%). MR depicted screws (mean 12.6 mm diameter, mean 65.3 mm length) with 50% error in diameter and 30% in length from susceptibility artefact compared to manufacturer dimensions (6.5-7.5 mm diameter, 40-50 mm length). Adequate screw placement was high despite this (85.8%). On MR, the cortex border visibility was 60.7% and the spinal canal visibility was 74.6%. There is strong to near-perfect agreement between MR and CT for evaluating adequate screw position in PLIF surgery. MR alone is useful for analyzing screw placement and should be considered first-line imaging in uncomplicated cases with CT analysis reserved for cases of uncertainty. Level III retrospective cohort study.