Magic Angle Effect: A Relevant Artifact in MR Neurography at 3T?

Subjects, Hardware Prerequisites, and Positioning

This study was approved by the ethics committee of the Medical Faculty of the University of Heidelberg (reference S-057/2009). Twenty-five adult healthy volunteers (mean age, 25 years; range, 19–31 years; men = 5, women = 13) without any history of peripheral nerve disease and 14 patients (mean age, 51 years; range, 53–46 years; men = 5, women = 9) with confirmed neuropathic lesions participated in the study and gave written informed consent. All measurements were performed by using a 3T whole-body MR imaging scanner (Magnetom Tim Trio; Siemens, Erlangen, Germany). The integrated body coil was used as a radio-frequency transmitter; as a radio-frequency receiver, we used a 2 × 4 element surface-array coil (multipurpose 2 × 4 channel surface-array coil, Model MR 8300; Noras MRI Products, Hoechberg, Germany). The 2 coil arrays were firmly attached opposite each other on the anterior and posterior surfaces of the proximal right thigh. For an exact positioning of the thigh and, by implication, of the sciatic nerve at defined angle positions (Θ) relative to the main magnetic field, an MR imaging−compatible polyvinyl chloride construction built in-house was used. Its design consisted of a 10-cm-wide baseboard on top of which an isosceles triangle was formed by 1 adjustable thigh board and 1 additional support board. These 2 flexible boards could be raised from the horizontal position into 7 predefined distance locks at angles of Θ = 0°, 30°, 35°, 40°, 45°, 50°, or 55° between the thigh board and the baseboard, which in turn was aligned to B0 (Fig 1). The accuracy of the device for the adjustment of the desired angle positions was estimated by finding the true anatomic angle of the sciatic nerve with the thigh board (abbreviated by [I] in Fig 1) adjusted at the 55° position. The mean angular deviation of the sciatic nerve in the group of healthy subjects from the Θ = 55° position of the device was 4.6° ± 2.9°.

• Download figure
• Open in new tab
• Download powerpoint

Fig 1.

Construction chart of in-bore positioning aid. Distance locks (a–f) precisely define 7 angle positions between the thigh (I) and base board (II) relative to B₀: 0°, 30°, 35°, 40°, 45°, 50°, and 55°. This polyvinyl chloride device was firmly positioned under the gluteal and thigh region of the supine patient and aligned in the z-direction, parallel to the static magnetic field B0. Angulations were achieved by stepwise erection of the thigh board relative to the baseboard. The insert on the upper right displays an oblique top view of the device.

MR Imaging Protocol

After the survey scan in 3 planes, a T1-weighted TSE sequence in an axial orientation (perpendicular to the femur) was performed for a detailed anatomic overview of the sciatic nerve at the proximal thigh. For quantitative analysis of the change in intraneural T2, we used a DE-TSE sequence, and for the qualitative visual assessment of intraneural T2-weighted contrast, we used a T2-weighted TIRM sequence for fat saturation (TI = 180 ms). The latter is the most commonly used sequence in clinical diagnostic imaging of peripheral nerve disorders. Table 1 provides the applied pulse sequences and corresponding measurement parameters. Additionally, for selected angle positions, we performed T1-weighted TSE and T2-weighted TIRM measurements in a parasagittal orientation relative to the femur for assessment of the longitudinal course of the sciatic nerve in the sagittal plane (Fig 2). Table 2 shows the sequences performed at the respective angle positions Θ.

• Download figure
• Open in new tab
• Download powerpoint

Fig 2.

Parasagittal MR images of the sciatic nerve of 1 representative healthy subject at Θ = 55°. A, T1-weighted TSE image shows that the sciatic nerve (arrows) courses parallel to the femur among the adductor magnus, semitendinosus, and biceps femoris muscles. B, Sagittal T2-weighted TIRM image at the position of maximum occurrence of the MA at Θ = 55° shows hyperintense alteration of the nerve.

Calculation of Outcome Measures

Quantitative outcome measures and CNR were calculated from 2 predefined circular regions of interest: sROI, comprising only the connective tissue of the interfascicular compartment (interfascicular space), and lROI, including the interfascicular space and intraneural nerve fascicles.

Both regions of interest were consistently placed within the sciatic nerve without contact with its epineural borderzone to exclude any extraneural tissue. The sROI was always placed within the lROI, but care was taken to exclude intraneural nerve fascicles from the sROI—that is, to obtain only the connective tissue within the interfascicular space (interfascicular space) by placing the sROI between intraneural nerve fascicles. Finally, small and large regions of interest of equal size were placed in adjacent muscular (short head of the biceps femoris muscle) and fatty soft tissue (adjacent intermuscular fat). Measurements were performed only on axial 2D images (DE-TSE, T2-TIRM) with perpendicular orientation to the longitudinal course of the sciatic nerve to avoid significant partial volume effects.

The T2_app was determined for each region of interest from the DE-TSE sequence according to the following equation:

S₁ denotes the mean signal intensity of the first (T_E1 = 12 ms) and S₂ of the second echo (T_E2 = 69 ms). The term “apparent T2” reflects the fact that several echoes with different echo times contribute to the signal intensity of TSE sequences, from which a true T2 cannot be calculated.

In addition, the angle-dependent increase of T2, as would be expected from theoretic modeling based on the angle-dependent dipole-dipole interaction (equation 1), was calculated. For the given tissue-specific T2 relaxation time, estimated by this study for the sciatic nerve, further angle-dependent T2 change is given by equation 3,¹⁴ with A and K being tissue-specific constants:

A and K were estimated by fitting the curve as given by equation 3 to the measured T2 (Θ) by using a Levenberg-Marquardt algorithm. R² was calculated to measure the goodness-of-fit.

T2-weighted TIRM is the sequence of choice for a clinical MRN examination but does not allow calculation of quantitative tissue parameters. Its signal intensity is influenced by the filling factor of the coil or inhomogeneous radio-frequency attenuation, for example. Therefore we calculated CNR values according to equation 4 by using the signal intensities of the sciatic nerve (S_n) and muscle (S_m), respectively:

where σ_ref is the SD of the signal intensity outside the thigh.

Statistical Analyses

For group analysis, T2_app and CNR values were averaged over all obtained axial sections for each Θ. Nonparametric statistical analyses were then performed by using the Wilcoxon signed rank test.

Comparison with Clinical Findings

The CNR results obtained in healthy subjects were finally compared with those of patients with peripheral neuropathy involving neuropathic lesions localized by clinical and NCS/EMG examination. As in healthy subjects, CNR was obtained from identical intraneural and muscular regions of interest from the T2-weighted TIRM sequence on 5 contiguous sections. CNR values from patients are given in Table 3. In the patient comparison group, the measurement of T2_app was not accomplished because of constraints of measurement time and patient positioning. Only in an additional small sample of patients with confirmed cubital tunnel syndrome (n = 5) was the measurement of T2_app achievable. In this group, only mild or moderate degrees of disease severity, defined by the absence of muscle atrophy or higher grade muscle weakness (Medical Research Council score of >3 in all ulnar target muscles of the hand), were included.