Objective      To build a normal L3-S1 segment finite element model

and analyze the biomechanical difference before and after segment fusion in transforaminal lumbar interbody fusion (TLIF) model.       Methods      The normal L3-S1 segment CT data were used to rebuild the intact finite element (INT) model by software. The TLIF model was rebuild by right side facetectomy and discectomy

and a cage was inserted into the disc. Based on the TLIF model

an unfused model without fixation (M1)

an unfused model with bilateral pedicle screw fixation (M2)

a fused model with bilateral pedicle screw fixation (M3)

and a fused model without fixation (M4) were built. A 500 N preload with 7.5 N·m moment was applied to the surface of L3 segment to induce the flexion-extension

left-right lateral bending and left-right axial rotation. The L4-L5 segment range of motion (ROM) and the stress distribution of the pedicle screw-rod system and interbody fusion cage were observed.       Results      ROM was increased in the M1 model compared with the INT model

obviously in left axial rotation. ROM was similar to each other in the M2

M3 and M4 models

and was decreased compared with the INT model. The peak Von Mises stress of the pedicle screw-rod system in the M3 model was decreased compared with the M2 model. The peak Von Mises stress of the interbody infusion cage was similar in the M2

M3 and M4 models

but was obviously less than that in the M1 model.       Conclusion      A normal L3-S1 segment finite element model is successfully established. Pedicle screw-rod system instrumentation can effectively control the stability of the index segment and reduce the load applied to the interbody fusion cage. The peak Von Mises stress of posterior instrumentation is decreased after fusion. It is unnecessary to remove posterior instrumentation after fusion considering the limited biomechanical benefit and the surgery cost.