[The role of T2*-weighted gradient-echo magnetic resonance sequences in the study of suspected dorsal-lumbosacral vertebral metastases].

INTRODUCTION

Magnetic resonance (MR) imaging showed high reliability in detecting spine metastases with spin-echo (SE) sequences, T1-weighted sequences being generally more sensitive than T/-weighted ones. We investigated the value of T2*-weighted gradient-echo (GE) sequences in studying spine metastases.

MATERIALS AND METHODS

Twenty patients with established diagnosis of primary carcinoma and clinically suspected thoracic and/or lumbosacral spine metastases underwent .5-T MR imaging and 99mTc-HDP bone scan. The disagreement of GET2*- versus SET2-weighted images as well as versus bone scan and the disagreement of total MR results versus bone scan results were evaluated by McNemar test. The agreement of GET2*- versus SET1-weighted images was evaluated by Cohen's kappa.

RESULTS

Of a total of 111 MR signal abnormalities consistent with metastasis, 109 (98.2%) were T2*-hyperintense, whereas only 50 (45.1%) were T2-hyperintense (p < .0001) and 51 (45.9%) were detected with bone scan (p < .0001). Of a total of 121 MR and/or bone scan findings consistent with metastasis, 111 (91.7%) were MR positive, with high disagreement with 61 (50.4%) positive at bone scan (p < .00001). T2*-hyperintensity associated with T1-hypointensity (with or without T2-hyperintensity) was the most frequent pattern (104/111), 93.7%).

CONCLUSIONS

T2*-weighted GE sequences seem to be more effective than T2-weighted SE sequences and as effective as T1-weighted SE sequences. MR imaging confirms its ability in detecting abnormalities consistent with spine metastases.