Devoted to diagnostic and interventional spine imaging and therapeutics

Library

Correlation of Pre-Treatment MRI and Clinical Outcomes in Patients with Osteoporotic Vertebral Body Compression Fractures Treated with Percutaneous Vertebroplasty 2004

Do, Huy, M.D., Arakawa, Hideki, M.D., Curtis, Lisa, M.D., Marks, Michael, M.D.

Purpose:
To assess the accuracy of MRI signal changes in predicting the clinical outcomes of patients with vertebral compression fractures (VCFs) who were treated with percutaneous vertebroplasty (PV).

Methods and Materials:
Thirty four patients with 37 painful osteoporotic VCFs were reviewed. All patients had pre-treatment MRI consisting of T1-weighted, T2-weighted, and STIR sequences. MRI signal intensities within the target vertebra were compared with adjacent normal appearing vertebrae. Pain relief was assessed by comparing the pre-treatment and 30 day post-treatment pain scores using the visual analog scale (VAS) and correlated with MRI signal changes. The sensitivities, specificities, positive and negative predictive values for each MRI sequence were calculated using Fisher’s exact test. Our hypothesis is that patients with VCFs exhibiting MRI signal changes consistent with vertebral bone marrow edema (T1 hypointensity, T2 and STIR hyperintensity) are most likely to experience pain relief with PV treatment.

Results:
Thirty of 37 treatments (81%) resulted in either full (24/37, 65%) or partial (6/37, 16%) pain relief; full pain relief means a VAS of zero and partial pain relief means a lower VAS after treatment than before treatment. Seven treatments (7/37, 19%) did not result in pain relief. No patients experienced worsening pain after treatment and there were no treatment related complications. We found no statistically significant value of MRI in predicting clinical pain relief. When assessing each sequence individually, 87.5% of patients with T1 hypointensity improved with PV compared with 75% of patients with T1 iso/hyperintensity. 75% of patients with T2 hyperintensity improved with PV compared with 85% of patients with T2 iso/hypointensity. 82.4% of patients with STIR hyperintensity improved with PV compared with 100% of patients with STIR isointensity. The sensitivities and specificities of T1 hypointensity, T2 hyperintensity, and STIR hyperintensity for any clinical improvement are 70%/50%, 41/4%/42.9%, and 77.8%/0%, respectively. There was no improvement in these values when the MRI sequences were assessed in combination.

Conclusion:
Proper patient selection is an important factor to achieve better clinical outcome for patients evaluated for possible PV treatment. This evaluation usually consists of detailed history, physical examination and review of imaging studies (X-rays, MRIs, CTs, and bone scans). However this study does not support th