Acute radial nerve entrapment at the spiral groove: detection by DTI-based neurography

Transcription

1 Eur Radiol (2015) 25: DOI /s MAGNETIC RESONANCE Acute radial nerve entrapment at the spiral groove: detection by DTI-based neurography Suren Jengojan & Florian Kovar & Julia Breitenseher & Michael Weber & Daniela Prayer & Gregor Kasprian Received: 4 June 2014 /Revised: 24 November 2014 /Accepted: 9 December 2014 /Published online: 11 January 2015 # European Society of Radiology 2015 Abstract Objectives This study evaluated the potential of three-tesla diffusion tensor imaging (DTI) and tractography to detect changes of the radial (RN) and median (MN) nerves during transient upper arm compression by a silicon ring tourniquet. Methods Axial T2-weighted and DTI sequences (b=700 s/ mm 2, 16 gradient encoding directions) of 13 healthy volunteers were obtained. Fractional anisotropy (FA) and apparent diffusion coefficient (ADC) values of the MN and RN were measured at the spiral groove and further visualized in 3D by deterministic tractography (thresholds: FA=.15, angle change=27 ). Results Local/lesional RN FA values increased (p=0.001) and ADC values decreased (p=0.02) during a 20-min upper arm compression, whereas no significant FA (p=0.49) or ADC (p=0.73) changes of the MN were detected. There were no T2-w nerve signal changes or alterations of nerve trajectories in 3D. Conclusions Acute nerve compression of the RN leads to changes of its three-tesla DTI metrics. Peripheral nerve DTI provides non-invasive insights into the selective vulnerability of the RN at the spiral groove. Key Points DTI-based neurography detects nerve changes during acute nerve compression. Compression leads to a transient increase in local radial nerve FA values. DTI provides insights into radial nerve vulnerability at the spiral groove. Keywords Diffusion tensor imaging. MRI. Entrapment neuropathies. Peripheral nerves. Radial nerve Abbreviations ADC Apparent diffusion coefficient asnr Approximate signal-to-noise ratio EMG Electromyography FA Fractional anisotropy MIP Maximum intensity projection MN Median nerve ROI Region of interest RN Radial nerve T2-w T2-weighted Electronic supplementary material The online version of this article (doi: /s ) contains supplementary material, which is available to authorized users. S. Jengojan: J. Breitenseher : M. Weber : D. Prayer : G. Kasprian (*) Department of Biomedical Imaging and Image-guided Therapy, Division of Neuro- and Musculosceletal Radiology, Medical University of Vienna, Währinger Gürtel 18-20, 1090 Vienna, Austria gregor.kasprian@meduniwien.ac.at F. Kovar Department of Trauma-Surgery, Medical University of Vienna, Vienna, Austria Introduction Twenty-one percent of nerve entrapment neuropathies are caused by compression of the radial nerve (RN) in the spiral groove (Spiral Groove Syndrome) [1]. Thus, the RN is the most frequently affected nerve in tourniquet-induced paralyses of the upper arm. The degree of the chronicity and the severity of the entrapment as well as the location of the compression affect the magnitude of the motor and sensory deficit [2]. Electromyography (EMG), nerve conduction studies (NCS) or peripheral nerve ultrasound are generally used to

2 Eur Radiol (2015) 25: diagnose and grade peripheral nerve lesions [3]. Unfortunately, the site of investigation is often not accessible to these diagnostic tools. EMG and NCS are generally unable to determine the structural causes of nerve lesions and are limited in their potential to assess acute nerve trauma. Since fibrillation potentials may need ten days to six weeks to be detectable by EMG [4], muscle denervation signs are not present at the early stages of even clinically severe nerve injuries. The better signal-to-noise ratio and the higher spatial resolution of clinical MR imaging at three-tesla has improved the visualization of peripheral nerve injuries. Recently diffusion tensor imaging (DTI) achieved promising results in the evaluation of traumatic peripheral nerve injuries [5, 6], and in cases of nerve entrapment syndromes [1, 7 9]. However, the existing clinical data is limited by the lack of a standardized and controlled physiological model of short-term nerve compression and its visualization by conventional and advanced (DTI) three-tesla magnetic resonance imaging (MRI). The aim of this study was to evaluate the potential of DTI and tractography to detect and visualize RN and median nerve (MN) changes related to acute nerve compression, under controlled conditions (during 20 min of compression with 220±30 mmhg) in healthy subjects. Materials and methods This single-centre, randomized, prospective study was approved by the local Institutional Review Board. After obtaining written informed consent, 13 consecutive healthy volunteers underwent three-tesla MRI examinations (Philips Achieva, Best, The Netherlands). All subjects fit the inclusion criteria: years of age, clinically healthy, body mass index of <30, a systolic arterial blood pressure <190 mmhg, no neurovascular impairment or previous surgery on the investigated limb, and no rash or dermatological condition or tattoos on the placement arm. The mean age of the participants was 24.3 years (range 22 28); eight (62 %) were males and five (38 %) were females. All subjects were examined in the supine position. A flex medium surface coil was placed onto the upper arm, covering the region 4 cm proximal and distal to a line that bisected the distance between the acromion and the lateral epicondyle. This line systematically indicated the position of future tourniquet placement. Before compression an axial T2-weighted (T2-w, TR 4808 ms, TE 90 ms, flip angle 90, FOV mm, matrix , slice thickness 2.5 mm, gap 0.3 mm, reconstructed voxel size 0.23/0.23/2.5 mm, acquisition time 6 min 24 s) and an axial, echo-planar, single-shot DTI sequence (TR 6433 ms, TE 78 ms, b-value=0/700 s/ mm 2, FOV mm, matrix 88 84, 16 diffusion encoding directions, fat saturation mode SPAIR, number of acquisitions five, acquisition time 8 min 47 s) of the previously defined target region was acquired. Then, the elastic silicone ring (The HemaClearTM of OHK Medical Device, Haifa, Israel) was rolled up the arm and placed at the initially defined position. This sterile device provides sufficient pressure (220±30 mmhg) to block brachial arterial and venous flow [10]. Five minutes after application of the tourniquet both sequences were repeated consecutively, during which the tourniquet remained in place (overall compression time 20 min). Transient temporary reddening and numbness or pain of the upper extremity were reported by some participants during upper arm compression. In each case these complaints disappeared immediately after tourniquet removal. During a final check-up 30 min after the MR examination none of the compression-related symptoms were revealed. A trained radiologist, experienced in peripheral nerve MRI, optically assessed the T2-w signal intensities of the MN and RN. A partial or complete signal increase of the nerve crosssections within the compression zone was classified as nerve signal change. Two raters with experience in peripheral nerve imaging assessed the quality of the DTI source data (Fig. 1). No case with visible motion artefacts was identified. Eddy current induced distortion occasionally appeared on the very proximal imaging planes, but did not affect the measurement regions (Fig. 1). The DTI source data was post processed using the commercial Extended MR WorkSpace (Philips Medical Systems, Best, The Netherlands) software. After registration of EPI images for motion correction (through plane) the diffusion source data was used to identify the RN and MN (Fig. 1). Deterministic tractography of the MN and RN was performed after placing two regions of interest (ROIs) along the course of the respective nerves, using standard fibre reconstruction thresholds (FA=0.15, angle change=27 ) and a FACT (fibre assignment by continuous tracking) algorithm (Fig. 2). Both raters identified the RN and MN on axial diffusion source images (Fig. 1) at the compression zone and a control zone, which was consistently located 2.5 cm proximally to the upper rim of the tourniquet. The cross-sectional shape of the humerus and its topographic relationship to the RN served as additional landmarks (Fig. 1). For FA and ADC quantification equally sized regions of interest were placed at the nerves cross section. In order to prevent the inclusion of surrounding soft tissue both raters aimed to keep the defined voxels/rois within all previously generated nerve trajectories. The signal-to-noise ratio (SNR) was measured in all cases in the compressed condition. As recommended by others [11], we determined an approximate SNR (asnr) value using the following approach: The asnr is defined as the mean value of the signal of a circumscribed region (long head of the triceps muscle) on diffusion unweighted (b=0 s/mm 2 ) images divided by the standard deviation measured in the same region.

3 1680 Fig. 1 From left to right: Coronal maximum intensity projections (MIP), axial T2-w and DTI source images, ADC and FA maps before (upper row) and during (lower row) upper arm compression by a silicone ring Fig. 2 Coronal (a, b) and axial (c, d) T2-w images co-registered with DTI-based neurography results. The trajectories of the radial (yellow) and median (blue) nerves can be visualized in the non-compressed (a, c) and compressed (b, d) upper arm. ROIs (green=before and red= during compression) are placed at a similar cross-sectional plane, for exact comparison between compressed and non-compressed FA and ADC values. The angle bracket (<) shows the major compression zone. ROIs of the control regions (not visualized) were placed 2.5 cm proximally to this zone Eur Radiol (2015) 25: tourniquet. The position of the MN is indicated by arrows and the RN is outlined by arrowheads. Please note the circumferential compression of the upper arm (lower row, coronal MIP)

4 Eur Radiol (2015) 25: Statistical analysis Means±standard deviations were calculated for FA and ADC measurements. In order to compare ADC and FA values of different regions, paired Student s t-tests were used. A p- value 0.05 was considered to indicate a significant result. In order to avoid an increasing type-ii error, no multiplicity corrections were performed. Rater agreement in FA ROI analysis was assessed using intraclass correlation (ICC) for metric data. Results A significant difference (p=0.001, paired t-test, Table 1, Fig. 3) in local RN FA values between no compression (mean FA=0.522±0.125) and during compression (mean FA=0.711 ±0.096) was found. Compression caused a significant decrease (p=0.024, paired t-test, Table 1, Fig. 3) in the regional ADC values (1.264 s/mm 2 ±0.350 s/mm 2 vs s/mm 2 ± s/mm 2 ) of the RN. Unaffected nerve segments proximal to the compression zone did not show any mean ADC (no compression: s/mm 2 ±0.442 s/mm 2, during compression: s/mm 2 ±0.268 s/mm 2 ;p=0.306)ormeanfavalue (no compression: 0.579±0.180, during compression: 0.550± 0.132; p=0.659) changes. The local FA (p=0.488, paired t-test) and ADC (p=0.733, paired t-test) values of the MN remained unchanged after compression. ADC or FA values of unaffected nerve segments proximal to the compression zone did not change during MN compression (Table 2, ControlADC/FA). The interobserver agreement in FA quantification was very good (ICC=0.987). The asnr was 11.14±4.17 (mean±standard deviation). The trajectories of the depicted peripheral nerves appeared unchanged before and after compression (Fig. 2). Table 1 During upper arm compression by a silicon ring tourniquet, a significant increase in regional fractional anisotropy (FA) and decrease in apparent diffusion coefficient (ADC) values of the radial nerve (RN) was found at the site of compression. Unaffected nerve segments (Control FA and ADC) proximal to the compression zone did not show any changes in DTI (diffusion tensor imaging) metrics before and during compression Radial nerve No compression Compression p-value mean ±SD mean ±SD FA ADC CONTROL FA CONTROL ADC FA fractional anisotropy, ADC apparent diffusion coefficient No T2-w signal changes of the RN and MN between compressed and non-compressed conditions could be identified by visual inspection (Fig. 1). Discussion DTI studies in animal models of peripheral nerve injuries [6] and preliminary experiences in patients with neuropathies [1, 7 9] have shown promising results for the use of DTI as a non-invasive tool for the quantification [12, 13] and functional characterization of peripheral nerve lesions [6]. However, these studies mostly focused on chronic compression neuropathies [7 9]. Chronic nerve compression lasts for days or weeks and results in perineural oedema, macrophage recruitment and fibrosis, as well as demyelination, ultimately leading to perineural thickening [14]. This may be well detected by peripheral nerve ultrasound and MRI [3]. In addition to the increased T2-w signal intensity of the chronically compressed peripheral nerve, previous DTI studies have shown an increase in ADC and a reduction of FA values [1, 6 9] within the affected nerve segment. As opposed to chronic nerve compression, acute nerve compression lasts over a period of a few hours only, and leads to reduced intraneural blood flow and oedema [14]. In the presented controlled model of acute nerve compression, we were able to demonstrate that three-tesla MRI shows compression-associated changes in peripheral nerve DTI metrics. Surprisingly, these changes were only detectable in the RN, not the MN. Although the protocol of our study did not enable identification of the exact microstructural correlates of this observation, this diversity may be explained by the local peripheral nerve topography, and by the pathophysiology of nerve compression. As the radial nerve is compressed against the immediately adjacent humeral bone, the local transient pressure exerted on the neural tissue of the RN is greater compared to that of the MN, which is fully imbedded within surrounding soft tissue structures. Here, we observed an increase in RN anisotropy as a potential consequence of the local compression and the resulting reduction of the RN extracellular/extraaxonal space. In contrast, the surrounding soft tissue structures protect the median nerve, resulting in unchanged or minimally lower local FA values of the nerve. This is in accordance with the higher frequency of clinical radial nerve entrapment compared to median nerve compression neuropathy after acute compression of the upper arm [14]. As none of the subjects experienced functional deficits of the RN, the observed compression-related increase in RN FA values most likely reflects the transient consequences of a mechanically induced nerve tissue rearrangement, whereas microstructural alterations (demyelination, Schwann-cell necrosis, axonal degeneration) of peripheral nerves in cases of chronic nerve compression lead to reduced local FA values [1, 8, 9].

5 1682 Eur Radiol (2015) 25: Fig. 3 Line graphs of FA (left column) and ADC (right column) measurements before and during compression: note the consistent increase of RN FA values during upper arm compression The main limitation of our study is the rather small number of included subjects. However, the reported changes in RN FA values were consistently seen in all cases (Fig. 3). Therefore we consider our study sufficiently sized to support our Table 2 Regional fractional anisotropy (FA) and apparent diffusion coefficient (ADC) values for the compressed and unaffected (Control FA and ADC) segments of the median nerve (MN) did not change before and during upper arm compression by a silicon ring tourniquet Median nerve No compression Compression p-value mean ±SD mean ±SD FA ADC CONTROL FA CONTROL ADC FA fractional anisotropy, ADC apparent diffusion coefficient findings. The high variability in RN and MN DTI metrics in the assessed regions (Fig. 3) can be related to several factors: (1) the variability of FA and ADC along the examined nerves, as recently described [8, 9], (2) the generally low asnr, (3) artefacts due to insufficient fat suppression (using SPAIR) and (4) interindividual differences in ROI placement. The latter effect was minimal within the studied subjects as plane specific anatomical landmarks were used for ROI localization before and during compression (Figs. 1 and 2). Although the observed DTI changes were relatively small (Fig. 3), they cannot be explained by factors other than compression. As no visible changes in T2-w RN and MN signal intensities were found by subjective assessment, T2-w signal quantification [15] and the use of fat suppression may be more sensitive to acute nerve compression. Moreover, perfusion MR sequences may identify nerve ischaemia-related changes. However, these sequences could not be included in the study protocol due to compression related imaging time constraints.

6 Eur Radiol (2015) 25: Conclusion Three-tesla DTI detects changes in peripheral nerve integrity caused by transient upper arm nerve compression by a tourniquet device. The increased FA values of the RN and unchanged FA values of the MN during acute (20 min) nerve compression may be explained by the different anatomical course of these nerves, and consequently, the different grade of exposure to mechanical pressure at the upper arm. Thus, these imaging findings provide non-invasive insights into the selective vulnerability of the RN at the spiral groove. Acknowledgements The scientific guarantor of this publication is Univ. Prof. Dr. Daniela Prayer. The authors of this manuscript declare no relationships with any companies, whose products or services may be related to the subject matter of the article. The authors state that this work has not received any funding. Dr. Michael Weber kindly provided statistical advice for this manuscript. Institutional Review Board approval was obtained. Written informed consent was obtained from all subjects (patients) in this study. Methodology: prospective, case-control study, performed at one institution. References 1. Chhabra A, Deune GE, Murano E, Prince JL, Soldatos T, Flammang A (2012) Advanced MR neurography imaging of radial nerve entrapment at the spiral groove: a case report. J Reconstr Microsurg 28: Aho K, Sainio K, Kianta M, Varpanen E (1983) Pneumatic tourniquet paralysis. Case report. J Bone Joint Surg Brit 65: Linda DD, Harish S, Stewart BG, Finlay K, Parasu N, Rebello RP (2010) Multimodality imaging of peripheral neuropathies of the upper limb and brachial plexus. Radiographics 30: Willmott AD, White C, Dukelow SP (2012) Fibrillation potential onset in peripheral nerve injury. Muscle Nerve 46: Hiltunen J, Suortti T, Arvela S, Seppa M, Joensuu R, Hari R (2005) Diffusion tensor imaging and tractography of distal peripheral nerves at 3 T. Clin Neurophysiol 116: Takagi T, Nakamura M, Yamada M et al (2009) Visualization of peripheral nerve degeneration and regeneration: monitoring with diffusion tensor tractography. Neuroimage 44: Khalil C, Hancart C, Le Thuc V, Chantelot C, Chechin D, Cotten A (2008) Diffusion tensor imaging and tractography of the median nerve in carpal tunnel syndrome: preliminary results. Eur Radiol 18: Guggenberger R, Markovic D, Eppenberger P et al (2012) Assessment of median nerve with MR neurography by using diffusion-tensor imaging: normative and pathologic diffusion values. Radiology 265: Bäumer P, Pham M, Ruetters M et al (2014) Peripheral neuropathy: Detection with diffusion-tensor imaging radiology. Radiology 273: Ladenheim E, Krauthammer J, Agrawal S, Lum C, Chadwick N (2013) A sterile elastic exsanguination tourniquet is effective in preventing blood loss during hemodialysis access surgery. J Vasc Access 14: Dietrich O, Raya JG, Reeder SB, Reiser MF, Schoenberg SO (2007) Measurement of signal-to-noise ratios in MR images: influence of multichannel coils, parallel imaging, and reconstruction filters. J Magn Reson Imaging 26: Guggenberger R, Eppenberger P, Markovic D et al (2012) MR neurography of the median nerve at 3.0 T: optimization of diffusion tensor imaging and fiber tractography. Eur J Radiol 81: e775 e Kim B, Srinivasan A, Sabb B, Feldman EL, Pop-Busui R (2014) Diffusion tensor imaging of the sural nerve in normal controls. Clin Imaging 38: Rempel D, Dahlin L, Lundborg G (1999) Pathophysiology of nerve compression syndromes: response of peripheral nerves to loading. J Bone Joint Surg Am 81: Bäumer P, Dombert T, Staub F et al (2011) Ulnar neuropathy at the elbow: MR neurography - nerve T2 signal increase and caliber. Radiology 260: