Accuracy of SWI sequences compared to T2*-weighted gradient echo sequences in the detection of cerebral cavernous malformations in the familial form

Gianvincenzo Sparacia,Claudia Speciale,Aurelia Banco,Francesco Bencivinni,Massimo Midiri
DOI: https://doi.org/10.1177/1971400916665376
2016-08-22
The Neuroradiology Journal
Abstract:Purpose The purpose of this study was to assess the accuracy of susceptibility-weighted imaging (SWI), compared with T2*-weighted gradient echo (GRE) imaging in assessing cerebral cavernous malformations. Materials and methods We retrospectively evaluated 21 patients with a familial form of cavernous malformation. Magnetic resonance (MR) protocol included non-enhanced and contrast-enhanced fast-spin echo (FSE) T1-weighted sequences, FSE T2-weighted sequences, fluid-attenuated inversion-recovery (FLAIR), GRE T2*-weighted and SWI sequences. Images were reviewed in consensus by two expert neuroradiologists to assess the location, number, size and conspicuity of the lesions on T2*-weighted GRE and SWI sequences. Statistical differences in the number, size and conspicuity of the lesions seen on the SWI images and the T2*-weighted GRE images were assessed with the nonparametric Wilcoxon signed rank test. Results The number of cavernous malformations was significantly higher ( p < .001) on the SWI images ( n = 152) than on T2*-weighted GRE images ( n = 56). Lesion size was significantly higher ( p < .001) on SWI images (mean: 0.4 cm, SD ± 0.55) than on T2*-weighted GRE sequences (mean: 0.2 cm, SD ± 0.51) and the differences were statistically significant ( p < .001). Lesion conspicuity was significant higher ( p < .001) on SWI than on T2*-weighted GRE images. In one patient who underwent a 2-month follow-up for the onset of neurologic symptoms related to cerebral hemorrhage, a cerebral hematoma was identified at the site of a cerebral cavernous malformation that was demonstrated only on the SWI images in the previous MR examination. Conclusions The SWI sequence, being more sensitive to substances which distort the local magnetic field than the GRE T2*W sequence, showed a higher sensitivity in identifying cerebral cavernous malformations. Thus, routine clinical neuroimaging protocol should contain SWI sequences to evaluate patients with (or suspected) cerebral cavernous malformations.
What problem does this paper attempt to address?