MRI of transforaminal lumbar interbody fusion: Imaging appearance with and without the use of human recombinant bone morphogenetic protein-2 (rhBMP-2)

Purpose: To describe the vertebral endplate and intervertebral disc space MRI appearance following TLIF, with and without the use of rhBMP-2, and to determine if the appearance is concerning for discitis/osteomyelitis. Materials and methods: After institutional review board approval, 116 TLIF assessments performed on 75 patients with rhBMP-2 were retrospectively and independently reviewed by five radiologists and compared to 73 TLIF assessments performed on 45 patients without rhBMP-2. MRIs were evaluated for endplate signal, disc space enhancement, disc space fluid, and abnormal paraspinal soft tissue. Endplate edema-like signal was reported when T1-weighted hypointensity, T2-weighted hyperintensity, and endplate enhancement were present. Subjective concern for discitis/osteomyelitis on MRI was graded on a five-point scale. Generalized estimating equation binomial regression model analysis was performed with findings correlated with rhBMP-2 use, TLIF level, graft type, and days between TLIF and MRI. Results: The rhBMP-2 group demonstrated endplate edema-like signal (OR 5.66; 95 % CI [1.58, 20.24], p=0.008) and disc space enhancement (OR 2.40; 95 % CI [1.20, 4.80], p=0.013) more often after adjusting for the TLIF level, graft type, and the number of days following TLIF. Both groups had a similar temporal distribution for endplate edema-like signal but disc space enhancement peaked earlier in the rhBMP-2 group. Disc space fluid was only present in the rhBMP-2 group. Neither group demonstrated abnormal paraspinal soft tissue and discitis/osteomyelitis was not considered likely in any patient. Conclusions: Endplate edema-like signal and disc space enhancement were significantly more frequent and disc space enhancement developed more rapidly following TLIF when rhBMP-2 was utilized. The concern for discitis/osteomyelitis was similar and minimal in both groups.