Between July and August 2012, 15 healthy volunteers were prospectively included in the study. Written informed consent was obtained from all persons before entering the study. All subjects were aged 18 years or older and had no history of neurological disorders, amblyopia or diseases of the optic nerve. The study was approved by the local ethics committee of the Albert Ludwigs University Freiburg, Germany, and was performed in accordance with the ethical standards laid down in the 1964 Declaration of Helsinki.
Transbulbar sonography and MRI were carried out on the same day and both were repeated during a second visit 28 ± 11 (range 9 – 53) days later. For each volunteer, only the right eye was evaluated because two measurements on a paired organ are not independent and may bias the results.
Ultrasound examinations of the ONSD were carried out in B-mode using a Philips iU22 ultrasound system and a 9 – 3 MHz linear array transducer (Philips Medical Systems; Bothell, WA). After a resting time of five minutes volunteers were examined in supine position with the upper part of the body and the head elevated to 20-30°. For safety reasons of biomechanical side effects the mechanical index (MI) was reduced to 0.2, the thermal index (TI) to 0.0. The ultrasound probe was placed on the temporal part of the closed upper eyelid using a thick layer of ultrasound gel. The anterior part of the optic nerve was depicted in a transversal plane showing the papilla and the optic nerve in its longitudinal course. ONSD was assessed 3 mm and 5 mm behind the papilla, as described previously [9, 12, 13]. In order to gauge the ONSD, the distance between the external borders of the hyperechoic area surrounding the optic nerve was quantified. Each optic nerve sheath was examined three times and means were calculated and considered for further evaluation (Figure 1).
Measurements were carried out by two experienced sonographers (JB and AH). They performed investigations of both eyes of all study subjects independently and were blinded to the results of each other. In order to determine scan-rescan reproducibility observer 1 examined all subjects at both visits. For calculation of inter-observer variability, observer 2 quantified ONSD in two volunteers at the first visit and in 8 volunteers at the second visit. Five individuals were measured on a third visit by both sonographers. To evaluate the intra-observer variability observer 1 did repeated measurements offline using sonographic images that were recorded at the first visit after 75 ± 8 days (range 60 – 86).
MRI was performed on a 3 T whole-body scanner (Magnetom TIM Trio, Siemens Healthcare, Erlangen, Germany) using a 32-channel phased-array head coil. The employed MRI sequence protocols for ONSD measurements were based on the methodological setup described by Weigel et al. . Prior to the volunteer study MRI protocols were optimized for a 32-channel head coil regarding resolution and contrast versus signal-to-noise-ratio. Subjects were instructed to fixate on a target inside the scanner, with the right eye in straight gaze.
Two different variants of a T2-weighted turbo spin echo (TSE) sequence were employed [10, 14, 15]: (1) A fast T2-weighted overview TSE which provides good soft tissue contrast and morphological data for planning: TR = 4000 ms, TE = 130 ms, echo train length = 25, bandwidth = 120 Hz/pixel, ‘weak’ chemical fat saturation. The sequence was applied twice with nine contiguous slices in sagittal (FOV = 21 x 21 cm2, Matrix = 448 x 448) and transversal (FOV = 21 x 18 cm2, Matrix = 448 x 392) orientation leading to a nominal spatial resolution of 0.47 x 0.47 mm2 with slice thickness = 3 mm. The acquisition time was 1:06 min (transversal) and 1:10 min (sagittal), respectively (Figure 1). (2) A rapid T2-weighted half-Fourier acquisition single-shot turbo spin-echo (HASTE) sequence that was primarily optimized for quantification: TR = 1700 ms, TE = 129 ms, number of excitations = 1, bandwidth = 196 Hz/pixel, FOV = 19 x 16 cm2, Matrix = 448 x 378, phase encoding direction left to right, nominal spatial resolution = 0.42 x 0.42 mm2, slice thickness = 2 mm. Acquisition time was 1.7 s per slice. Two slices were acquired perpendicular to the optic nerve orientation within the intraorbital track, guided by the morphological TSE images: the first slice 3 mm behind the papilla, the second slice in a depth of 5 mm (Figure 1).
The ONSD was measured on coronary section HASTE images by drawing spherical regions of interest (ROIs) around the external border of the cerebro spinal fluid (Figure 1). ROI evaluations were performed by a radiologist (KE) and a physician with a one-year experience in neuroradiology (FS) using a state-of-the-art radiology workstation (IMPAX EE R20 VIII P1, Agfa HealthCare N.V., Mortsel, Belgium). For improved visualization, HASTE images were magnified 7.5-fold. As shown by Weigel et al. , the accuracy of measurement in the chosen sequence and protocol setup is limited by the reproducibility of ROIs rather than the nominal resolution. This is possible since the experimental setup employs the partial volume effect together with the a-priori knowledge of very high signal cerebro spinal fluid adjacent to low signal parenchymal structures.
Images created during the first visit were assessed by both readers independently. These scans were reanalysed by observer 1 in order to determine intra-observer agreement. For evaluation of the scan-rescan reproducibility observer 2 quantified the ONSD at the second visit.
Values were expressed as mean ± standard deviation. Reproducibility, intra- and inter-observer variability as well as method comparisons were analyzed using the approach by Bland-Altman by calculating the mean difference (d) and standard deviation of the difference. From these data, the limits of agreement were calculated (σd, 95% confidence intervals). Additionally, intra- and inter-observer variabilities were evaluated using the Kappa coefficient. Furthermore, correlation analyses were performed with Pearson’s correlation coefficient (r) to quantify the strength of agreement.