Objective: Normal facial nerve enhancement patterns derived from spin-echo (SE) sequences have not been systematically compared on contrast-enhanced 3D inversion recovery-prepared fast spoiled gradient-echo (IR-FSPGR) sequences, now in widespread use. We hypothesize that features unique to IR-FSPGR may engender differences in the appearance of the normal facial nerve, which may confound analysis of pathologic enhancement. We compared unenhanced and contrast-enhanced SE and IR-FSPGR sequences in a cohort of patients without facial nerve pathology.
Materials and methods: Twenty-three patients without facial nerve pathology were examined. Unenhanced and contrast-enhanced signal intensity (SI) of seven facial nerve segments was assessed on SE and IR-FSPGR by two neuroradiologists. SI was assigned a value of 0-3 (0, absent; 1, faint; 2, equivalent to brain; 3, equivalent to enhancing dural sinus). Statistically significant differences were assessed for each segment.
Results: Significantly higher unenhanced and contrast-enhanced SI was present in most facial nerve segments on IR-FSPGR compared with SE, including cisternal, canalicular, labyrinthine, and geniculate segments (p ≤ 0.01). Enhancement patterns were generally similar; however, significant enhancement of the labyrinthine segment was detected only on SE (p = 0.011). For unenhanced images, mean kappa statistic was 0.32, and for the contrast-enhanced images, mean kappa statistic was 0.04, implying fair and slight agreement between readers, respectively.
Conclusion: Significantly greater SI is observed in most facial nerve segments on both unenhanced and contrast-enhanced IR-FSPGR among healthy subjects and may be misinterpreted as pathologic when evaluated in the context of existing enhancement paradigms. Examiners should remain cognizant of normal deviations from expected enhancement patterns in IR-FSPGR imaging to avoid misdiagnosis and other interpretive pitfalls.