Predictive value of pedicle involvement with MRI in spine metastases

OBJECTIVES The study aimed to retrospectively evaluate the accuracy and value of magnetic resonance imaging (MRI) in predicting pedicle involvement for patients with spine metastases. METHODS Forty-five patients with a vertebral metastasis encroaching at least one pedicle were studied using MRI before surgery and regularly after surgery. Patients were categorized on the basis of their numbers of pedicle involvement (Group 1: one pedicle was involved, n = 23; Group 2: two pedicles were involved, n = 22). The diagnostic accuracy was calculated, and comparisons of intraoperative blood loss and recurrence rate between the two groups were performed. RESULTS The overall performance of MRI in predicting the pedicle involvement was as follows: accuracy, 94.4%; sensitivity, 95.5%; and specificity, 91.3%. Less intraoperative blood loss was observed for Group 1 compared with Group 2 (1,661 ± 672 ml and 2,173 ± 790 ml, respectively, P = 0.024). Tumor relapse occurred in 8.7% (2/23) of Group 1 and in 22.7% (5/22) of Group 2 with median recurrence free survival time 14 and 9 months, respectively. CONCLUSIONS MRI is a reliable approach to assess pedicle involvement. It has potential for use in the evaluation of the clinical characteristics of patients with spine metastases.