Faizy TD, Broocks G, Frischmuth I, Westermann C, Flottmann F, Schönfeld MH2, Nawabi J, Leischner H, Kutzner D2, Stellmann JP, Heesen C, Fiehler J, Gellißen S, Hanning U.

OBJECTIVES: The aim of the study is to compare coronal spectrally fat-suppressed 2D turbo spin-echo (TSE) with 2D short-tau inversion-recovery (STIR) sequences for the detection of optic nerve hyperintensities in patients with acute optic nerve neuritis (ON). METHODS: A retrospective review of patients with suspected unilateral ON and pathological visual evoked potentials, who received coronal TSE and STIR sequences with similar fast and clinically feasible acquisition times in addition to our standard imaging protocol. All images were evaluated and compared concerning the presence of optic nerve lesions, lesion lengths, and signal intensities in different anatomical parts of the optic nerves and CNR measures. A summary confidence score (CS) was calculated based on each reader’s subjective confidence regarding the scoring items. RESULTS: Interobserver agreements regarding the detection of optic nerve lesions were excellent for both sequences (TSE, κ = 0.89 and STIR, κ = 0.80). Greater extensions (17.4 ± 6.3 mm vs. 14.1 ± 5.8 mm), as well as higher numbers of optic nerve lesions in symptomatic nerves, were detected on TSE (49/52) compared with STIR (45/52) sequences (both p < 0.001). Overall CS were significantly (p < 0.001) higher for TSE (2.8) compared with STIR (2.1) sequences regarding the presence or absence of optic nerve lesions. CNR ratios of lesions‘ mean signal intensities vs. ipsilateral surrounding orbital fat and vs. signal intensity measurements from contralateral optic nerves were significantly higher on TSE compared with STIR (p < 0.001 for both comparisons). CONCLUSION: Spectrally fat-suppressed coronal 2D TSE sequences appear to be more sensitive for the detection of hyperintense optic lesions compared with 2D STIR sequences. KEY POINTS: • Spectrally fat-suppressed TSE sequences showed higher detection rates of hyperintense optic nerve lesions, as well as a higher reader confidence scores compared with STIR. • Optic nerve signal abnormalities on TSE sequences were brighter and showed a greater expansion along the optic nerve course. • CNR measures were significantly higher on TSE compared with STIR, when comparing the ratios of mean signal intensities of optic nerve lesions to ipsilateral orbital fat and to contralateral healthy optic nerves of both sequences.


Eur Radiol. 2019 Nov;29(11):6266-6274



Link to Pubmed