ABSTRACT
OBJECTIVES: Multi spin echo (MSE) sequences are often used for obtaining T2-relaxometry data as they provide defined echo times (TEs). Due to their time-consuming acquisition, they are frequently replaced by turbo spin echo (TSE) sequences that in turn bear the risk of systematic errors when analyzing small structures or lesions. With this study, we aim to test whether T2-relaxometry data derived from either dual-echo TSE or 12-echo MSE sequences are equivalent for quantifying peripheral nerve lesions. Hereditary transthyretin (ATTRv) amyloidosis was chosen as a surrogate disease, as it allows the inclusion of both asymptomatic carriers of the underlying variant transthyretin gene (varTTR) and symptomatic ATTRv amyloidosis patients. MATERIALS AND METHODS: Overall, 50 participants with genetically confirmed varTTR (20 clinically symptomatic ATTRv amyloidosis; 4 females, 16 males; mean age, 61.8 years; range, 33-76 years; and 30 asymptomatic varTTR-carriers; 18 females, 12 males; mean age, 43.1 years; range, 21-62 years), and 30 healthy volunteers (13 females, 17 males, mean age 41.3 years, range 22-73) were prospectively included and underwent magnetic resonance neurography at 3 T. T2-relaxometry was performed by acquiring an axial 2-dimensional dual-echo TSE sequence with spectral fat saturation (TE1/TE2, 12/73 milliseconds; TR, 5210 milliseconds; acquisition time, 7 minutes, 30 seconds), and an axial 2-dimensional MSE sequence with spectral fat saturation and with 12 different TE (TE1, 10 milliseconds to TE12, 120 milliseconds; ΔTE, 10 milliseconds; TR, 3000 milliseconds; acquisition time, 11 minutes, 23 seconds) at the right mid to lower thigh. Sciatic nerve regions of interest were manually drawn in ImageJ on 10 central slices per participant and sequence, and the apparent T2-relaxation time (T2app) and proton spin density (ρ) were calculated individually from TSE and MSE relaxometry data. RESULTS: Linear regression showed that T2app values obtained from the dual-echo TSE (T2appTSE), and those calculated from the 12-echo MSE (T2appMSE) were mathematically connected by a factor of 1.3 throughout all groups (controls: 1.26 ± 0.02; varTTR-carriers: 1.25 ± 0.02; symptomatic ATTRv amyloidosis: 1.28 ± 0.02), whereas a factor of 0.5 was identified between respective ρ values (controls: 0.47 ± 0.01; varTTR-carriers: 0.47 ± 0.01; symptomatic ATTRv amyloidosis: 0.50 ± 0.02). T2app calculated from both TSE and MSE, distinguished between symptomatic ATTRv (T2appTSE 66.38 ± 2.6; T2appMSE 84.6 ± 3.3) and controls (T2appTSE 58.1 ± 1.0, P = 0.0028; T2appMSE 72.8 ± 0.7, P < 0.0001), whereas differences between varTTR-carriers (T2appTSE 61.8 ± 1.5; T2appMSE 76.7 ± 1.3) and ATTRv amyloidosis were observed only for T2appMSE (P = 0.0082). The ρ value differentiated well between healthy controls (ρTSE 365.1 ± 7.2; ρMSE 170.4 ± 3.8) versus varTTR-carriers (ρTSE 415.7 ± 9.8, P = 0.0027; ρMSE 193.7 ± 5.3, P = 0.0398) and versus symptomatic ATTRv amyloidosis (ρTSE 487.8 ± 17.9; ρMSE 244.7 ± 13.1, P < 0.0001, respectively), but also between varTTR-carriers and ATTRv amyloidosis (ρTSEP = 0.0001; ρMSEP < 0.0001). CONCLUSIONS: Dual-echo TSE and 12-echo MSE sequences provide equally robust and reliable T2-relaxometry data when calculating T2app and ρ. Due to their shorter acquisition time and higher resolution, TSE sequences may be preferred in future magnetic resonance imaging protocols. As a secondary result, ρ can be confirmed as a sensitive biomarker to detect early nerve lesions as it differentiated best among healthy controls, asymptomatic varTTR-carriers, and symptomatic ATTRv amyloidosis, whereas T2app might be beneficial in already manifest ATTRv amyloidosis.
