Human dorsal-root-ganglion perfusion measured in ...

3 downloads 0 Views 2MB Size Report
Jul 14, 2016 - nerve cell bodies of the primary sensory neurons and are located ... pseudounipolar, sensory neurons, which are embedded in a dense.
NeuroImage 141 (2016) 81–87

Contents lists available at ScienceDirect

NeuroImage journal homepage: www.elsevier.com/locate/ynimg

Human dorsal-root-ganglion perfusion measured in-vivo by MRI Tim Godel a,⁎, Mirko Pham a, Sabine Heiland a, Martin Bendszus a, Philipp Bäumer a,b a b

Department of Neuroradiology, Neurological University Clinic, University of Heidelberg Medical Center, Im Neuenheimer Feld 400, 69120 Heidelberg, Germany Department of Radiology, German Cancer Research Institute, Im Neuenheimer Feld 280, 69120 Heidelberg, Germany

a r t i c l e

i n f o

Article history: Received 1 March 2016 Accepted 12 July 2016 Available online 14 July 2016 Keywords: Dorsal root ganglia DCE-MRI Perfusion Permeability Polyneuropathy

a b s t r a c t Purpose: To develop an in-vivo imaging method for the measurement of dorsal-root-ganglia-(DRG) perfusion, to establish its normal values in patients without known peripheral nerve disorders or radicular pain syndromes and to determine the physiological spatial perfusion pattern within the DRG. Methods: This prospective study was approved by the institutional ethics committee and written informed consent was obtained from all participants. 46 (24 female, 22 male, mean age 46.0 ± 15.2 years) subjects without known peripheral neuropathies or pain syndromes were examined by a 3 Tesla MRI scanner (Magnetom VERIO or TRIO, Siemens AG, Erlangen, Germany) with a VIBE (Volume-Interpolated-Breathhold-Examination) dynamic-contrast-enhanced (DCE) T1-w-sequence (TR/TE 3.3/1.11 ms; 24 slices; voxel resolution 1.3 × 1.3 × 3.0 mm3) covered the pelvis from the upper plate of the 5th lumbar vertebra to the 2nd sacral vertebra. Transfer-constant (Ktrans) and interstitial-volume-fraction (interstitial-leakage-fraction, Ve) were modeled for the DRG and spinal nerve by applying the Tofts-model. Statistical analyses included pairwise comparisons of L5/S1 DRG vs. spinal nerve. Furthermore, distinct physiological zones within the S1 DRG were compared (cell body rich area (CBRA) vs. nerve fiber rich area (NFRA)). Results: DRG showed a significantly increased permeability compared to spinal nerve (Ktrans 3.8 ± 1.5 10−3/min vs. 1.6 ± 0.9 10−3/min, p-value: b 0.001) combined with an increased interstitial leakage of contrast agent into the extravascular-extracellular-space (Ve 38.1 ± 19.2% vs. 17.3 ± 9.9%, p-value: b0.001). Parameters showed no statistically significant difference on DRG-level (L5 vs. S1; p-value: 0.62 (Ktrans); 0.17 (Ve)) and -side (left vs. right; p-value: 0.25 (Ktrans); 0.79 (Ve)). Female gender was associated with a significantly increased permeability (Ktrans female 4.3 ± 1.4 10−3/min vs. male 3.4 ± 0.9 10−3/min, p-value: b 0.05) but no statistically significant differences in interstitial leakage (Ve female 40.1 ± 14,1% vs. male 34.5 ± 17.4%, p-value: 0.24). DRG showed distinct spatial distribution patterns of perfusion: Ktrans and Ve were significantly higher in the CBRA than in the NFRA (Ktrans 4.4 ± 1.8 10−3/min vs. 1.7 ± 1.2 10−3/min, p-value: b0.001 and Ve 40.9 ± 21.3% vs. 15.1 ± 11.7%, p-value: b 0.001). Conclusion: Non-invasive and in-vivo measurement of human DRG perfusion by MRI is a feasible technique. DRG show substantially higher permeability and interstitial leakage than spinal nerves. Even distinct physiological perfusion patterns for different microstructural compartments could be observed within the DRG. The technique may become particularly useful for future research on the poorly understood human sensory neuropathies and pain syndromes. © 2016 The Authors. Published by Elsevier Inc. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

1. Introduction The dorsal-root-ganglia (DRG) represent dense accumulations of nerve cell bodies of the primary sensory neurons and are located pairwise in, or closely adjacent to the intervertebral foramina along the spine. The prime electrophysiological function of the DRG is to receive the bulk afferent sensory input from the PNS and relay it to the CNS. A nerve fiber rich area (NFRA), which is located more centrally, contains the incoming and outgoing nerve fibers (Fig. 1), whereas the peripheral regions of the DRG are denominated cell body rich area (CBRA) for containing the densely accumulated cell bodies of the pseudounipolar, sensory neurons, which are embedded in a dense

capillary network (Jimenez-Andrade et al., 2008). Blood supply to the DRG is provided by the spinal branch of the dorsal trunk of the segmental arteries. After penetrating the ganglionic capsule, the arteries give rise to a subcapsular capillary plexus prior to penetrating into the deeper part of the ganglion and branching into intraganglionic blood vessels (Kobayashi et al., 2010). The regulation of blood flow to the DRG is achieved by muscular sphincters at strategic locations along the arteriole trunks and their branching points, to adjust flow to varying functional and metabolic demand (Kobayashi et al., 2010). While the peripheral nervous system (PNS) generally has a tight blood-nerve interface similar to the blood-brain-barrier (Weerasuriya & Mizisin, 2011), the DRG are intensely vascularized organs with high permeability between blood and nervous tissue (Jimenez-Andrade

http://dx.doi.org/10.1016/j.neuroimage.2016.07.030 1053-8119/© 2016 The Authors. Published by Elsevier Inc. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

82

T. Godel et al. / NeuroImage 141 (2016) 81–87

Fig. 1. Dense vascularization of the peripheral CBRA compared to the central NFRA. Left Confocal micrograph of a mouse L4 DRG labeled with the endothelial cell marker CD31 showing the marked difference in the density of the vascular supply of the densely arranged sensory ganglia (CBRA, white arrows) as compared to the corresponding entering/exiting spinal nerve and ventral/dorsal roots (ventral root, dorsal root, spinal nerve: NFRA, white arrowheads). 280 optical sections acquired at 0.5 μm z-plane intervals; z stack thickness 140 μm. Scale bar 100 μm. Right Bright-field photomicrograph of a whole-mount L4 DRG for anatomical reference. This figure is used with permission of BioMed Central Ltd. (Jimenez-Andrade et al., 2008).

et al., 2008; Hirakawa et al., 2004; Jacobs, 1978; Anzil et al., 1976; Jacobs et al., 1976). This exception to the otherwise restricted permeability of the PNS is potentially of high clinical relevance, because of the vulnerability to low and high molecular weight neurotoxic substances (Cavanagh, 1973; Cho, 1977; London & Albers, 2007; Le Quesne & McLeod, 1977; Viaene et al., 1999; Ikeda et al., 1973) as well as to toxic metabolites in drug- or metabolic-induced neuropathies (Lee & Swain, 2006; Quasthoff & Hartung, 2002; Mielke et al., 2006; Pignata et al., 2006; Pardo et al., 2001; Pettersen et al., 2006; McArthur et al., 2005; Berger et al., 1993; Keswani et al., 2002; Nicholas et al., 2007a; Nicholas et al., 2007b; Rahman et al., 2016). In the CNS it is firmly established that an increased neural activity is accompanied by an elevated, measurable oxygenation, blood volume and flow (Logothetis & Pfeuffer, 2004). Therefore it is likely, that particularly the increased activity of DRG neurons, which may especially occur in pathological conditions of increased excitability and activation (Xiao et al., 2016; Hogan, 2007; Sapunar et al., 2005; Zhang et al., 2004), also leads to increasing demands in DRG perfusion. Furthermore, the DRG plays an increasingly important role in the treatment of pain syndromes and neuropathic pain as a target of electrical neurostimulation (Pei et al., 2015; Forget et al., 2015; Eldabe et al., 2015; Van Buyten et al., 2015; Liem et al., 2013; Koopmeiners et al., 2013). Current knowledge about DRG perfusion is limited and comes from experimental animal studies only (Jimenez-Andrade et al., 2008; Kobayashi et al., 2010). Despite the potential physiological and clinical relevance, to the best of our knowledge, in-vivo measurements in humans have not been reported so far. Dynamic-contrast-enhanced (DCE) methods of MRI perfusion have long been utilized for the clinical and functional imaging investigation of the CNS. These methods have only recently been tested for the PNS targeting the trunks of peripheral nerve at the proximal extremities (Baumer et al., 2014). DCE MRI can be used for the calculation of perfusion parameters such as the interstitial leakage of contrast agent (Ve) and blood-tissue permeability (Ktrans), providing functional in-vivo information that is otherwise not available in humans (Cuenod & Balvay, 2013; Sourbron, 2010). In this study we developed a DCE MRI method dedicated to the quantitative measurement of DRG perfusion and applied this method in patients without peripheral neuropathies or pain syndromes. 2. Patients and methods 2.1. Clinical and demographic patient data This study was approved by the institutional ethics board (S3982012) and written informed consent was obtained from all patients.

Overall, we included 46 patients (24 female, 22 male, age 46.0 ± 15.2 years) for whom a contrast-enhanced examination of the spine was indicated for reasons not pertaining to any peripheral nervous system disorder (spinal MRI to rule out metastases, spinal tumors, multiple sclerosis, spinal hemorrhage or fractures of a lumbar vertebra). Patients with symptoms suggestive of any peripheral neuropathy or radiculopathy were explicitly excluded from the study. 2.2. Imaging Examinations were conducted at a 3 Tesla MRI scanner (Magnetom VERIO or TRIO, Siemens AG, Erlangen, Germany) between 7/2015 and 12/2015. The position of the imaging slab was chosen so that its upper edge was aligned in parallel with the upper plate of the 5th lumbar vertebra. Before contrast agent administration, T1-weighted volumetricinterpolated-breathhold-examination (VIBE) sequences with 5°, 8°, 11°, 14° and 17° flip angle were acquired. Then, a T1-weighted, dynamic-contrast-enhanced (DCE) VIBE sequence was acquired with the following sequence parameters: TR/TE 3.3/1.11 ms; flip angle 15°; 24 slices; resolution 1.3 × 1.3 × 3.0 mm3. The contrast agent (Dotarem, Guerbet, France) was administered intravenously at the beginning of the third frame of the sequence at the standard concentration of 0.1 mmol/kg with a flow rate of 3.5 ml/s by automated injection. A total of 24 frames were recorded with a rate of 7.46 s/frame. A 15channel receive/transmit spine coil and an 8-channel receive body flex coil (Siemens) were used. 2.3. Quantitative image analysis Quantitative image analysis was performed using the commercially available software plug-in IB DCE 1.2 (Imaging Biometrics, Elm Grove, USA) to Osirix (Pixmeo, Bernex, Switzerland). Image masks were obtained by manual segmentation around the contour of the L5 and S1 DRG, which were reliably visible in the anatomical, T1-weighted nonenhanced image (Fig. 2). As internal control, the 5th lumbar spinal nerve on the right side at a presacral localization was likewise measured. This nerve was chosen because of its anatomical positioning in the same imaging slab as the L5 and S1 DRG. Additional regions-ofinterests (ROIs) were placed in the common iliac artery to obtain time-signal curves needed to define the arterial input function (AIF) and to detect the time point of bolus arrival in the lumbar region. Since it is not possible to differentiate the central (NFRA) and peripheral (CBRA) zone on conventional MR images, further masks were obtained by defining these areas on the Ktrans/Ve maps which were rendered voxel-wise as described below.

T. Godel et al. / NeuroImage 141 (2016) 81–87

83

2.4. Voxel-wise modeling of Ktrans and Ve

2.5. Statistical analysis

The Tofts model is the most common pharmacokinetic model assuming a tissue compartment in addition to a vascular compartment. Given the Gadolinium (Gd) concentration as a function of time, it models how the contrast agent distributes in the body and how this depends on characteristics of DRG and spinal nerve perfusion. Ktrans (expressed in min−1) is the most important and significant tissue dependent parameter in Tofts. It assesses either plasma flow in flowlimited scenarios or tissue permeability in permeability-limited scenarios for the uptake. In mixed scenarios it indicates a combination of the flow and permeability properties of the tissue. Since the applied contrast agent Dotarem is of small molecular size (559 Da, Gd-DOTA), its leakage from the capillaries into the extracellular, extravascular space (EES) is diffusive and hence reversible; it is therefore proportional to the difference in concentrations between the two compartments and Ktrans is the constant of this proportionality. Furthermore the modeling reveals the interstitial blood volume fraction Ve (interstitial leakage fraction, expressed in %). Preliminary evaluations of the time-dependent signal increase after contrast administration showed a relatively fast influx of contrast agent into the DRG tissue. This observation suggested use of the Tofts model for modeling a concentration time curve to the measured data. We calculated these parameters for the ROI within the DRG L5 and S1 and the spinal nerve L5.

Data visualization and statistical analyses were performed using Microsoft Excel. Mean values were calculated for Ktrans and Ve in each subject for DRG L5 and S1 as well as spinal nerve L5. Bar charts for Ktrans and Ve were charted in Microsoft Excel. CBRA and NFRA values of the S1 DRG were charted. Mean values were tested for statistical significance using a Student's t-Test, with a two-sided alpha level for significance of p-value: b0.05. Pearson correlation analysis was performed for age vs. Ktrans/Ve. 3. Results Perfusion parameter maps of the DRG and spinal nerves were generated (Fig. 3) and quantitatively assessed for a total of 46 participants (Figs. 4, 5). DRG Ktrans values were significantly increased compared to spinal nerve L5 (Ktrans = 3.8 ± 1.5 10− 3/min vs. 1.6 ± 0.9 10− 3/min, pvalue: b0.001). Likewise, Ve was significantly higher in the DRG compared to the spinal nerve L5 (Ve = 38.1 ± 19.2% vs. 17.3 ± 9.9%, pvalue: b 0.001, Fig. 4). Parameters showed no statistically significant difference on DRGlevel (L5 vs. S1; p-value: 0.62 (Ktrans); 0.17 (Ve)) and -side (left vs. right; p-value: 0.25 (Ktrans); 0.79 (Ve)). Females showed a statistically significant increased permeability compared to men (Ktrans = 4.3 ± 1.4 10−3/min vs. 3.4 ± 0.9 10−3/min, p-value: b0.05) whereas

Fig. 2. Image acquisition, segmentation and perfusion parameter calculation. Quantitative analysis of DRG perfusion. Regions of interest were defined manually comprising the DRG L5/S1 as well as the right spinal nerve L5 in axial T1-VIBE source images as illustrated in the upper left corner. Quantitative parameter maps were calculated using the Tofts model and yielded values for Ktrans and Ve.

84

T. Godel et al. / NeuroImage 141 (2016) 81–87

interstitial leakage fraction showed no statistically significant differences (Ve female 40.1 ± 14,1% vs. male 34.5 ± 17.4%, p-value: 0.24). Since peripheral nerves are known to show age-related degenerative changes, correlation analyses between perfusion parameters and age were performed. No significant correlations between age and Ktrans or Ve were observed. Quantitative readout parameters (Ktrans and Ve) of the peripheral zone (CBRA) were compared to the values at the central zone (NFRA) of the S1 DRG (Fig. 5). In this comparison, the DRG showed a distinct spatial pattern of distribution of Ktrans and Ve (Fig. 5). Theses parameters were significantly higher in the peripheral zone than in the central zone (Ktrans = 4.4 ± 1.8 10−3/min vs. 1.7 ± 1.2 10−3/min, p-value: b 0.001 and Ve = 40.9 ± 21.3% vs. 15.1 ± 11.7%, p-value: b 0.001). Ktrans (A) and Ve (B) were calculated using the Tofts model and are visualized in quantitative parameter maps. These parameter maps also allow relatively good differentiation between areas with high or low values for Ktrans and Ve within the DRG, representing the peripheral CBRA and the central NFRA. High Ktrans and Ve values are found in the connective tissue surrounding the DRG, containing the ganglions vascular plexus. 4. Discussion In this study we show that DCE MRI can be used to measure human DRG perfusion in-vivo by parameters of blood-tissue permeability (Ktrans ) and interstitial leakage fraction (Ve). To our knowledge this is the first in-vivo investigation of physiologic and metabolic human DRG function indicated by quantitative parameters of DRG perfusion and permeability. The assessed perfusion parameters of blood-tissue permeability (Ktrans) and leakage of contrast agent

into the interstitial space (Ve) were significantly increased in the DRG compared to the spinal nerve. Additionally, perfusion parameters in the peripheral zone of the DRG, containing cell bodies of the primary sensory neurons (CBRA), were significantly increased compared to the central NFRA, containing the incoming and outgoing nerve axons. Interestingly, the perfusion parameters of the spinal nerve were roughly similar compared to the central NFRA in DRG, which may be related to microstructural similarities and adjacency between these two segments of the PNS (spinal nerve trunk continuing partly as the fibers entering the DRG). Altogether, these observations are consistent with former invasive animal studies, using autoradiographical measurements and immunohistochemical methods, that revealed a three-fold increase of blood flow within the DRG compared to peripheral nerve (Sasaki et al., 1997) as well as a seven-fold blood-vessel-density in the CBRA compared to the NFRA (Jimenez-Andrade et al., 2008). Microvascular approaches in animals using light, immunofocal and electron-microscopy after intravenous injection of tracers have also previously shown that the blood vessels of the CBRA have large fenestrations and a relative lack of tight junction proteins compared to the peripheral nerve (Jacobs, 1978; Anzil et al., 1976; Jacobs et al., 1976), allowing large molecules to leak out of the CBRA supporting microvasculature while remaining intravascularly within the NFRA and the peripheral nerves (Hirakawa et al., 2004; Jacobs et al., 1976; Arvidson, 1979; Olsson, 1968). The blood-nerve-barrier in peripheral nerves forms a tight and highly regulated interface similar to the blood-brain-barrier (Weerasuriya & Mizisin, 2011). The specific purpose of the particular leaky structure of the DRG blood tissue border and the physiologically elevated leakage of its microvascular endothelium (Hirakawa et al.,

Fig. 3. Perfusion parameter maps in S1 DRG. Parameter maps for Ktrans and Ve at the level of the DRG S1 overlayed with axial T1-weighted images, showing DRG at high spatial resolution for visual recognition and anatomical orientation (C, D).

T. Godel et al. / NeuroImage 141 (2016) 81–87

85

Fig. 4. Perfusion parameters of the L5/S1 DRG compared to the spinal nerve L5. Statistical analysis of perfusion parameters of the L5/S1 DRG vs. spinal nerve L5. Ktrans (A) as well as Ve (B) of the DRG are significantly increased compared to the spinal nerve (***p-value: b0.001).

2004) is likely to satisfy the high metabolic demand of sensory cell bodies including those that need to support the longest axons innervating the most distal extremity regions (Devor, 1999). An increased perfusion coupled with dense vascularization, the lack of tight junction proteins

and the presence of large, endothelial fenestrations may synergistically contribute to an increased vulnerability of primary sensory neurons of the DRG to neurotoxic agents.

Fig. 5. Perfusion parameters of the CBRA compared to the NFRA of the S1 DRG. Statistical analysis of perfusion parameters of the CBRA vs. NFRA of the S1 DRG. Ktrans (A) as well as Ve (B) of the peripheral CBRA is significantly increased compared to the central NFRA (***p-value: b0.001).

86

T. Godel et al. / NeuroImage 141 (2016) 81–87

Drug induced neuropathic pain is one of the major side effects of chemotherapy for example in cancer or HIV-treatment (Pettersen et al., 2006; Fallon, 2013). It is characterized by numbness, tingling, burning sensation and cold allodynia (Pain Suppl., 1986). The risk of neuropathy is proportional to the dose and duration of drug administration (Lee & Swain, 2006; Seidman et al., 2008). While the cell bodies of alpha-motoneurons are housed within the tight blood-brain-barrier of the CNS, the sensory neuron bodies of the DRG are exposed to toxic substances circulating in the blood. This microstructural feature, specific for the DRG, likely explains why anti-neoplastic and anti-HIV agents preferably induce a peripheral sensory polyneuropathy frequently characterized by neuropathic pain. Interestingly, the in-vivo morphometric measurement of the DRG in oxaliplatin induced chemotoxic sensory and painful distally symmetric polyneuropathy revealed clear signs of structural degeneration during the course of disease (Bäumer et al., 2015). This observation further points towards a central and primary role of the DRG in the pathogenesis particularly of toxic metabolically induced painful polyneuropathies. Dose escalation of chemotherapeutic agents (e.g. cisplatin) is frequently necessary to prolong survival in a variety of oncologic conditions (Mielke et al., 2006; Aghajanian et al., 1998; Pasini et al., 2002; Cocconi et al., 1999). In contrast to neuropathy, other side effects such as drug-induced hypersensitivity or neutropenia represent treatable conditions (e.g. by treatment with antihistamines, steroids or granulocyte colony-stimulating factor (Einzig et al., 1998)). However, the symptoms and the electrophysiological and structural alterations of chemotherapy induced peripheral neuropathy unfortunately still represent complications which are currently untreatable, eventually dose limiting and may lead to discontinuation of chemotherapy (Lee & Swain, 2006; Quasthoff & Hartung, 2002; Mielke et al., 2006; Pignata et al., 2006). In our study female gender was associated with a significantly increased vascular permeability within the DRG. Consistent with this observation, various clinical studies have reported that females may be more sensitive to chronic and neuropathic pain than males (Ruau et al., 2012; Hurley & Adams, 2008; Fillingim et al., 2009). Gender differences in chemotherapy induced neuropathic pain have only been investigated in animal studies so far, showing contradictory results. While there was no gender difference in cisplatin induced cold and mechanical allodynia, female mice showed an increased sensitivity against cold allodynia, but not mechanical allodynia in paclitaxel induced polyneuropathy (Naji-Esfahani et al., 2016). In contrast, another animal study observed no gender difference in paclitaxel induced neuropathic pain (Hwang et al., 2012). The regulation of blood flow into and within the DRG is achieved by muscular sphincters at strategical locations along the arteriole trunks and their branching points, to adjust flow to varying functional and metabolic demand (Kobayashi et al., 2010). It is conceivable that DRG not only adapt vascular supply and blood flow or permeability to functional variability dependent on short-term activity but also to long-term changes in neuronal activity or excitability which particularly may occur in sensory neuropathies and neuropathic pain syndromes (Xiao et al., 2016; Hogan, 2007; Sapunar et al., 2005; Zhang et al., 2004). DCE MRI could potentially identify such alterations in DRG perfusion. The potential clinical relevance of measuring this is indicated by findings that increased neuronal excitability may cause or aggravate the symptoms of human sensory neuropathies including neuropathic pain (Obata & Noguchi, 2008; Zhao et al., 2015; Malet & Brumovsky, 2015), oxaliplatin induced toxic sensory neuropathy (Zhao et al., 2012), lumbar disc herniation (Yan et al., 2015), or inflammatory pain (Malet & Brumovsky, 2015). In contrast, metabolic dysfunction also plays a crucial role in the pathogenesis of diverse pain syndromes. In-vitro animal studies using electrostimulation demonstrated that DRG are highly sensitive to hypoxia induced by mechanical compression, causing abnormal sensations and pain in radiculopathy (Sugawara et al., 1996). The painful

symptoms of diabetic neuropathy, which often dominate during the early stages of this most prevalent human neuropathy, likely result from metabolic aberrations leading to biochemical, structural and functional changes in the DRG (Sima & Sugimoto, 1999; Tavee & Zhou, 2009). Animal studies suggest that tissue ischemia and lower ambient oxygen tension in the DRG due to hyperglycemia may lead to aberrant metabolic functions and sensory symptoms (Rahman et al., 2016). Given this wide spectrum of potential relevance, it can be anticipated that in-vivo methods of measuring aspects of human DRG perfusion such as the one presented here, will find application in the investigation of the time course and localization of highly prevalent but poorly understood PNS disorders such as drug- or metabolicinduced peripheral sensory neuropathies and neuropathic pain syndromes. Beyond these fields, changes in DRG perfusion most likely play an important role in low back pain and sciatica in patients with disc herniation and spinal canal stenosis. Our study has several limitations. First, it is a limitation of our study and of any DCE MRI perfusion technique that the absolute values of perfusion parameters as reported here needed to be modeled. We chose the Tofts model for this purpose because it is valid particularly for parameter estimation in tissues with high vascular permeability. A limitation of the Tofts model is the fact that plasma flow and tissue permeability cannot be estimated separately. Changes in Ktrans can be attributed to changes in plasma flow or tissue permeability only if the flow regime meets certain presuppositions. Second, sequence parameters in DCE-MRI are a compromise between coverage, temporal, and spatial resolution. Small variations in the arterial input function (AIF) due to a relatively low temporal resolution does not affect the intraindividual ratio of perfusion parameters between different levels (S1/ L5) or sides (right/left), but it results in some inter-individual variation of the absolute perfusion parameters. Third, the normal values presented here may depend to some extent on the sequence parameters chosen, and researchers adopting the technique will have to confirm their own control values on their scanner system. And fourth, since there is no possibility to distinguish the central (NFRA) and peripheral (CBRA) zones in the anatomical image, additional ROIs where drawn by defining these areas in the acquired Ktrans/Ve maps.

5. Conclusion The non-invasive and in-vivo measurement of human DRG perfusion is feasible by MRI and shows a substantially higher permeability and interstitial leakage in the DRG than spinal nerves. Even distinct physiological perfusion patterns could be observed for different microstructural compartments within the DRG. These advances may become particularly useful for future studies on the poorly understood human sensory neuropathies and pain syndromes.

Disclosure The authors report no disclosures relevant to the manuscript.

Acknowledgements P.B. was supported by a postdoctoral fellowship from the Medical Faculty of the University of Heidelberg. M.P. was supported by a memorial stipend of the Else-Kröner-Fresenius Foundation and received grants from the EFSD/JDRF/Novo Nordisk European Programme in Type 1 Diabetes Research and the Deutsche Forschungsgemeinschaft (SFB 1158 TPÄ3). S.H. and M.B. were supported by a grant from the German Research Council (SFB 1118).

T. Godel et al. / NeuroImage 141 (2016) 81–87

References Aghajanian, C., et al., 1998. Phase II study of “dose-dense” high-dose chemotherapy treatment with peripheral-blood progenitor-cell support as primary treatment for patients with advanced ovarian cancer. J. Clin. Oncol. 16 (5), 1852–1860. Anzil, A.P., Blinzinger, K., Herrlinger, H., 1976. Fenestrated blood capillaries in rat craniospinal sensory ganglia. Cell Tissue Res. 167 (4), 563–567. Arvidson, B., 1979. Distribution of intravenously injected protein tracers in peripheral ganglia of adult mice. Exp. Neurol. 63 (2), 388–410. Baumer, P., et al., 2014. Peripheral nerve perfusion by dynamic contrast-enhanced magnetic resonance imaging: demonstration of feasibility. Investig. Radiol. 49 (8), 518–523. Bäumer, P.X.,.A., Weiler, M., Heckel, A., Schwarz, D., Pham, M., Jäger, D., Bendszus, M., Apostolidis, L., 2015. Oxaliplatin-induzierte Polyneuropathie: detektion und quantitative Charakterisierung der Pathophysiologie mittels MR Neurografie [abstract]. Röntgenkongress Hamburg. Berger, A.R., et al., 1993. 2',3'-Dideoxycytidine (ddC) toxic neuropathy: a study of 52 patients. Neurology 43 (2), 358–362. Cavanagh, J.B., 1973. Peripheral neuropathy caused by chemical agents. CRC Crit. Rev. Toxicol. 2 (3), 365–417. Cho, E.S., 1977. Toxic effects of adriamycin on the ganglia of the peripherial nervous system: a neuropathological study. J. Neuropathol. Exp. Neurol. 36 (6), 907–915. Cocconi, G., et al., 1999. Mature results of a prospective randomized trial comparing a three-weekly with an accelerated weekly schedule of cisplatin in advanced ovarian carcinoma. Am. J. Clin. Oncol. 22 (6), 559–567. Cuenod, C.A., Balvay, D., 2013. Perfusion and vascular permeability: basic concepts and measurement in DCE-CT and DCE-MRI. Diagn. Interv. Imaging 94 (12), 1187–1204. Devor, M., 1999. Unexplained peculiarities of the dorsal root ganglion. Pain (Suppl. 6), S27–S35. Einzig, A.I., et al., 1998. Phase I study of paclitaxel (taxol) and granulocyte colony stimulating factor (G-CSF) in patients with unresectable malignancy. Investig. New Drugs 16 (1), 29–36. Eldabe, S., et al., 2015. Dorsal root ganglion (DRG) stimulation in the treatment of phantom limb pain (PLP). Neuromodulation 18 (7), 610–617. Fallon, M.T., 2013. Neuropathic pain in cancer. Br. J. Anaesth. 111 (1), 105–111. Fillingim, R.B., et al., 2009. Sex, gender, and pain: a review of recent clinical and experimental findings. J. Pain 10 (5), 447–485. Forget, P., et al., 2015. Clinical evidence for dorsal root ganglion stimulation in the treatment of chronic neuropathic pain. A review. Acta Anaesthesiol. Belg. 66 (2), 37–41. Hirakawa, H., et al., 2004. Regional differences in blood-nerve barrier function and tightjunction protein expression within the rat dorsal root ganglion. Neuroreport 15 (3), 405–408. Hogan, Q.H., 2007. Role of decreased sensory neuron membrane calcium currents in the genesis of neuropathic pain. Croat. Med. J. 48 (1), 9–21. Hurley, R.W., Adams, M.C., 2008. Sex, gender, and pain: an overview of a complex field. Anesth. Analg. 107 (1), 309–317. Hwang, B.Y., et al., 2012. Gender differences in paclitaxel-induced neuropathic pain behavior and analgesic response in rats. Korean J. Anesthesiol. 62 (1), 66–72. Ikeda, Y., et al., 1973. Long-term toxicity study of methylmercuric chloride in monkeys (first report). Toxicology 1 (4), 361–375. Jacobs, J.M., 1978. Vascular permeability and neurotoxicity. Environ. Health Perspect. 26, 107–116. Jacobs, J.M., Macfarlane, R.M., Cavanagh, J.B., 1976. Vascular leakage in the dorsal root ganglia of the rat, studied with horseradish peroxidase. J. Neurol. Sci. 29 (1), 95–107. Jimenez-Andrade, J.M., et al., 2008. Vascularization of the dorsal root ganglia and peripheral nerve of the mouse: implications for chemical-induced peripheral sensory neuropathies. Mol. Pain 4, 10. Keswani, S.C., et al., 2002. HIV-associated sensory neuropathies. AIDS 16 (16), 2105–2117. Kobayashi, S., et al., 2010. Microvascular system of the lumbar dorsal root ganglia in rats. Part I: a 3D analysis with scanning electron microscopy of vascular corrosion casts. J. Neurosurg. Spine 12 (2), 197–202. Koopmeiners, A.S., et al., 2013. Effect of electrical field stimulation on dorsal root ganglion neuronal function. Neuromodulation 16 (4), 304–311 discussion 310-1. Le Quesne, P.M., McLeod, J.G., 1977. Peripheral neuropathy following a single exposure to arsenic. Clincal course in four patients with electrophysiological and histological studies. J. Neurol. Sci. 32 (3), 437–451. Lee, J.J., Swain, S.M., 2006. Peripheral neuropathy induced by microtubule-stabilizing agents. J. Clin. Oncol. 24 (10), 1633–1642. Liem, L., et al., 2013. A multicenter, prospective trial to assess the safety and performance of the spinal modulation dorsal root ganglion neurostimulator system in the treatment of chronic pain. Neuromodulation 16 (5), 471–482 discussion 482. Logothetis, N.K., Pfeuffer, J., 2004. On the nature of the BOLD fMRI contrast mechanism. Magn. Reson. Imaging 22 (10), 1517–1531. London, Z., Albers, J.W., 2007. Toxic neuropathies associated with pharmaceutic and industrial agents. Neurol. Clin. 25 (1), 257–276.

87

Malet, M., Brumovsky, P.R., 2015. VGLUTs and glutamate synthesis-focus on DRG neurons and pain. Biomolecules 5 (4), 3416–3437. McArthur, J.C., Brew, B.J., Nath, A., 2005. Neurological complications of HIV infection. Lancet Neurol. 4 (9), 543–555. Mielke, S., Sparreboom, A., Mross, K., 2006. Peripheral neuropathy: a persisting challenge in paclitaxel-based regimes. Eur. J. Cancer 42 (1), 24–30. Naji-Esfahani, H., et al., 2016. Gender differences in a mouse model of chemotherapyinduced neuropathic pain. Lab. Anim. 50 (1), 15–20. Nicholas, P.K., et al., 2007a. Distal sensory polyneuropathy in the context of HIV/AIDS. J. Assoc. Nurses AIDS Care 18 (4), 32–40. Nicholas, P.K., et al., 2007b. Symptom management and self-care for peripheral neuropathy in HIV/AIDS. AIDS Care 19 (2), 179–189. Obata, K., Noguchi, K., 2008. Contribution of primary sensory neurons and spinal glial cells to pathomechanisms of neuropathic pain. Brain Nerve 60 (5), 483–492. Olsson, Y., 1968. Topographical differences in the vascular permeability of the peripheral nervous system. Acta Neuropathol. 10 (1), 26–33. Classification of chronic pain. Descriptions of chronic pain syndromes and definitions of pain termsPain Suppl. 3, S1–226 Prepared by the International Association for the Study of Pain, Subcommittee on Taxonomy. Pardo, C.A., McArthur, J.C., Griffin, J.W., 2001. HIV neuropathy: insights in the pathology of HIV peripheral nerve disease. J. Peripher. Nerv. Syst. 6 (1), 21–27. Pasini, F., et al., 2002. High-dose chemotherapy in small-cell lung cancer. Anticancer Res. 22 (6B), 3465–3472. Pei, B.A., et al., 2015. Pulsed electrical stimulation protects neurons in the dorsal root and anterior horn of the spinal cord after peripheral nerve injury. Neural Regen. Res. 10 (10), 1650–1655. Pettersen, J.A., et al., 2006. Sensory neuropathy in human immunodeficiency virus/acquired immunodeficiency syndrome patients: protease inhibitor-mediated neurotoxicity. Ann. Neurol. 59 (5), 816–824. Pignata, S., et al., 2006. Residual neurotoxicity in ovarian cancer patients in clinical remission after first-line chemotherapy with carboplatin and paclitaxel: the Multicenter Italian Trial in Ovarian cancer (MITO-4) retrospective study. BMC Cancer 6, 5. Quasthoff, S., Hartung, H.P., 2002. Chemotherapy-induced peripheral neuropathy. J. Neurol. 249 (1), 9–17. Rahman, M.H., et al., 2016. Pyruvate dehydrogenase kinase-mediated glycolytic metabolic shift in the dorsal root ganglion drives painful diabetic neuropathy. J. Biol. Chem. Ruau, D., et al., 2012. Sex differences in reported pain across 11,000 patients captured in electronic medical records. J. Pain 13 (3), 228–234. Sapunar, D., et al., 2005. Distinct membrane effects of spinal nerve ligation on injured and adjacent dorsal root ganglion neurons in rats. Anesthesiology 103 (2), 360–376. Sasaki, H., et al., 1997. Neuropathology and blood flow of nerve, spinal roots and dorsal root ganglia in longstanding diabetic rats. Acta Neuropathol. 93 (2), 118–128. Seidman, A.D., et al., 2008. Randomized phase III trial of weekly compared with every-3weeks paclitaxel for metastatic breast cancer, with trastuzumab for all HER-2 overexpressors and random assignment to trastuzumab or not in HER-2 nonoverexpressors: final results of cancer and Leukemia Group B protocol 9840. J. Clin. Oncol. 26 (10), 1642–1649. Sima, A.A., Sugimoto, K., 1999. Experimental diabetic neuropathy: an update. Diabetologia 42 (7), 773–788. Sourbron, S., 2010. Technical aspects of MR perfusion. Eur. J. Radiol. 76 (3), 304–313. Sugawara, O., et al., 1996. The effects of mechanical compression and hypoxia on nerve root and dorsal root ganglia. An analysis of ectopic firing using an in vitro model. Spine (Phila Pa 1976) 21 (18), 2089–2094. Tavee, J., Zhou, L., 2009. Small fiber neuropathy: a burning problem. Cleve. Clin. J. Med. 76 (5), 297–305. Van Buyten, J.P., et al., 2015. Stimulation of dorsal root ganglia for the management of complex regional pain syndrome: a prospective case series. Pain Pract. 15 (3), 208–216. Viaene, M.K., et al., 1999. Cadmium: a possible etiological factor in peripheral polyneuropathy. Neurotoxicology 20 (1), 7–16. Weerasuriya, A., Mizisin, A.P., 2011. The blood-nerve barrier: structure and functional significance. Methods Mol. Biol. 686, 149–173. Xiao, Y., et al., 2016. Decreased voltage-gated potassium currents in rat dorsal root ganglion neurons after chronic constriction injury. Neuroreport 27 (2), 104–109. Yan, J., et al., 2015. Hyperexcitability and sensitization of sodium channels of dorsal root ganglion neurons in a rat model of lumber disc herniation. Eur. Spine J. Zhang, X.F., et al., 2004. Differential action potentials and firing patterns in injured and uninjured small dorsal root ganglion neurons after nerve injury. Brain Res. 1009 (1–2), 147–158. Zhao, M., et al., 2012. Acute cold hypersensitivity characteristically induced by oxaliplatin is caused by the enhanced responsiveness of TRPA1 in mice. Mol. Pain 8, 55. Zhao, Y., et al., 2015. Chronic nerve injury-induced Mas receptor expression in dorsal root ganglion neurons alleviates neuropathic pain. Exp. Ther. Med. 10 (6), 2384–2388.