CT and MRI Evaluation of Nerve Sheath Tumors of the Cervical Vagus Nerve

Transcription

1 Neuroradiology/Head and Neck Imaging Original Research Anil and Tan CT and MRI of the Cervical Vagus Nerve Neuroradiology/Head and Neck Imaging Original Research Gopinathan Anil 1 Tiong-Yong Tan Anil G, Tan TY Keywords: cervical vagus nerve, CT, MRI, neurofibroma, schwannoma DOI: /AJR Received September 7, 2010; accepted after revision November 26, Both authors: Department of Radiology, Changi General Hospital, 2, Simei St 3, Pin Code , Singapore. Address correspondence to G. Anil (ivyanil10@gmail.com). AJR 2011; 197: X/11/ American Roentgen Ray Society CT and MRI Evaluation of Nerve Sheath Tumors of the Cervical Vagus Nerve OBJECTIVE. Nerve sheath tumors arising from the cervical vagus are rare. The purpose of this study was to evaluate the role of CT and MRI in the diagnosis of these uncommon neoplasms. MATERIALS AND METHODS. The CT and MR studies and clinical data of 11 patients with surgicopathologic evidence of a nerve sheath tumor of the cervical vagus (nine schwannomas [including two ancient schwannomas] and two neurofibromas) who had been referred to our institute from January 1999 through 2009 were retrospectively reviewed. The tumors were evaluated with respect to their location, number, morphology, attenuation and/or signal intensity, enhancement characteristics, and patterns of mass effect. RESULTS. The tumors were solitary and well circumscribed. On CT, eight tumors were hypodense with poor enhancement, two were predominantly isodense, and a single lesion had multiple cystic areas with enhancing solid components. On MRI, they were heterogeneously bright on T2-weighted images with intense, inhomogeneous postgadolinium enhancement. The split fat sign, entering and exiting nerve sign, fascicular sign, and hyperintense rim sign were seen in some patients. The internal or common carotid artery was displaced anteriorly in eight patients, maintained a neutral position in two patients, and was displaced posterolaterally in another patient. In all patients except two, the tumor separated the carotid artery from the internal jugular vein. Vagal schwannomas splayed the carotid bifurcation in three patients. CONCLUSION. In conclusion, we present the patterns of mass effect and a spectrum of CT and MRI characteristics of nerve sheath tumor of the cervical vagus including observations that are sparingly described in the published literature. T he vagus nerve exits the posterior cranial fossa through the jugular foramen; before entering the thorax, it runs its entire course in the neck within the carotid sheath. Tumors arising from this cervical segment of the vagus nerve are rare and are often accidentally detected at surgery, with most being asymptomatic [1]. The most common neoplasms arising in the cervical vagus are paragangliomas (50%), followed by nerve sheath tumors (schwannomas [31%] and neurofibromas [14%]) [1]. Since the first report of a schwannoma of the cervical vagus in 1926 by Sekigachi and Oijie, there have been trickling case reports and a few short surgical series on nerve sheath tumors of the cervical segment of the vagus nerve [2], whereas the cases described in the radiology literature are mostly clumped with parapharyngeal space masses or other neurogenic tumors of the neck [3]. The accepted treatment of a cervical nerve sheath tumor is surgical excision, and less often an observational approach is adopted. The benefits of surgical excision must be carefully weighed against the risk of postoperative neurologic deficit especially in view of the indolent course seen with many of these tumors. Hence, an accurate and prompt diagnosis of cervical nerve sheath tumors is crucial. By virtue of their location, clinical examination is of limited value and imaging techniques are frequently necessary to suspect the diagnosis. In this study we evaluated the clinical application of CT and MRI for the diagnosis of nerve sheath tumors of the cervical vagus nerve with an emphasis on their imaging characteristics and patterns of mass effect. We present a spectrum of findings including observations that are infrequently described in the literature published in the English language. AJR:197, July

2 Anil and Tan Materials and Methods We retrospectively reviewed the CT and MR images of 11 patients with a nerve sheath tumor of the cervical vagus nerve who were referred to our institute, a Joint Commission International accredited multispecialty hospital, from January 1999 through All 11 patients had intraoperative and pathologic confirmation of the diagnosis. There were six men and five women (mean age, 38.4 years; age range, years). According to a chart review, 10 of the 11 patients had presented with a painless, slowly growing lateral neck mass of variable duration (7 55 months) and the tumor was detected incidentally during CT evaluation for a suspected parotid neoplasm in the remaining patient. Besides the apparent mass, four patients had hoarseness of voice and one had both hoarseness of voice and dysphagia at presentation. A single patient had a history of neurofibromatosis type 1. No specific neurologic signs or deficits were recorded in any of these patients. All the tumors were surgically resected through a transcervical approach. The nerve of origin was determined by direct visualization at surgery. The cases in which the nerve of origin was equivocal were excluded at the outset. Histopathology results confirmed the diagnosis of benign schwannoma in nine patients and neurofibroma in two patients. Contrast-enhanced CT alone or contrastenhanced CT and MRI were performed in our department or at the primary referring institution. Both CT and MR images were available for review in five patients; only CT had been performed in the remaining six patients. Because these studies were performed at different points in time, at different institutions, and on different scanners, the imaging parameters varied. However, all the CT scans were contrast-enhanced studies with z- axis coverage from the skull base to the arch of the aorta that were obtained on a helical CT or MDCT scanner. The section thickness and pitch were in the range of 3 5 mm and 1 1.5, respectively. The CT images were reconstructed at a 3- to 5-mm thickness in both the axial and coronal planes. The MRI examination included multiplanar images obtained on a 1- or 1.5-T scanner. The minimum sequences included axial T1-weighted (TR range/te range, /9 30), T2-weighted ( /80 120), and coronal STIR (inversion time, ms). Gadolinium-enhanced images were obtained in at least two orthogonal planes. The section thicknesses varied from 3 to 5 mm and the intersection gap, from 1 to 2.5 mm. Each study was reviewed by two radiologists with 18 and 5 years of experience in head and neck imaging. The findings were recorded by consensus. The scans were evaluated as to tumor location, laterality, relation to the vessels in the carotid space, patterns of mass effect on the adjoining fat planes and structures, attenuation and signal intensity, and enhancement pattern. The lesion margins (well defined or ill defined), shape (round to oval or irregular), size (long- and short-axis diameters measured on axial images), and number (single or multiple) were recorded. The CT attenuation of the lesion was compared with that of the adjoining skeletal muscles. Contrast enhancement was subjectively graded as poor, moderate, or intense: Enhancement similar to skeletal muscle was considered moderate; that less than skeletal muscle, poor; and that approaching the attenuation of contrast material in the vessels, intense. The tumor was classified as being of high signal intensity on T2-weighted images if it had signal intensity greater than that of fat and as being of low signal intensity on T1-weighted images if it had signal intensity lower than that of muscle. Gadolinium-enhanced T1-weighted images were compared with unenhanced T1-weighted images and were graded in the same manner as on CT. The tumor texture was recorded as heterogeneous or homogeneous in attenuation and signal intensity. On MRI, tumor texture was based on the appearance of the lesion on T2-weighted imaging. Results The clinical profiles, final diagnoses, and imaging findings are summarized in Table 1. Location The tumors were located in the carotid space (Fig. 1). Only two of the tumors were in the infrahyoid neck, whereas the remaining nine either were located entirely in the suprahyoid neck or straddled both the supraand infrahyoid neck, with their bulk above the level of carotid bifurcation. Size, Shape, and Number All the tumors were solitary and well defined with an oval to round shape. Five were oval along their long axis to give a fusiform configuration. A single tumor, which was shown to be schwannoma on pathology, showed gentle surface lobulations. Both the neurofibromas had smooth contours just as the schwannomas did. The average long-axis diameter of the tumors was 3.72 cm (range, 3 6 cm), and the average short-axis diameter was 2.86 cm (range, cm). Patterns of Mass Effect The suprahyoid and predominantly suprahyoid tumors (i.e., 9/11 tumors) displaced the fat in the parapharyngeal space anteriorly. Medial displacement of the visceral space (Fig. 2) was seen in seven cases. Eight tumors displaced the internal carotid artery (ICA) anteriorly as compared with the contralateral side, five of which showed a simultaneous component of medial displacement (Fig. 1). In the two patients with infrahyoid nerve sheath tumors, the carotid arteries were in a neutral position as compared with the contralateral side (Fig. 3); in one, the common carotid artery (CCA) was anteromedial to the tumor, and in the other it was just medial to the tumor. In all but two patients, the nerve sheath tumor separated the ICA or CCA from the internal jugular vein (IJV) (Figs. 1 and 3). The two exceptions were as follows: In one patient, there was paradoxical simultaneous posterolateral displacement of both the ICA and IJV without separation (Fig. 4A), and in the other patient both the vessels were displaced together, anteriorly. In three patients, the vagal nerve sheath tumors splayed the ICA and the external carotid artery (ECA) (Fig. 4) along different oblique planes. In these three patients, we measured the angle of contact of the ICA and ECA with the tumor by drawing tangential lines toward the circumference of the tumor from the center of the artery on an axial image. The angle of contact between the tumor and artery was less than 180 in all three. The arteries were merely displaced by the tumor or were stretched around the tumor. The vagal nerve sheath tumors neither encased the arteries nor filled the crotch of the carotid bifurcation. Texture and Enhancement The CT attenuation characteristics of the tumors were as follows: Eight tumors were hypoattenuating with poor enhancement (Fig. 1); one tumor was isoattenuating; one tumor had a central hypoattenuating area of fluid density and an isoattenuating thick, irregular rind of solid tissue; and one tumor was markedly heterogeneous with several nonenhancing fluid-density areas of various sizes and moderately enhancing intervening solid components (Fig. 5A). At histopathology, the latter two nerve sheath tumors turned out to be ancient schwannomas (Fig. 5). CT did not show any specific features that could distinguish the two neurofibromas in this series from the schwannomas. None of the tumors showed calcification or fat attenuation. On MRI, the tumors were hypo- to isointense on T1-weighted images (Fig. 6A) and were heterogeneously hyperintense on T2- weighted images (Fig. 6B). Gadolinium enhancement was marked but heterogeneous (Fig. 6C). In the five patients with both CT and MRI available, MRI was clearly superior in showing the internal heterogeneity of the tumor. An MR study of one of the ancient 196 AJR:197, July 2011

3 CT and MRI of the Cervical Vagus Nerve TABLE 1: Clinical Profile, Imaging Findings, and Final Diagnosis of Patients With Nerve Sheath Tumors of the Cervical Vagus Nerve Patient Tumor Mass Effect Age Displacement of Separation of IJV (y) Sex Modality Location Size (cm) ICA or CCA and ICA or CCA Texture and Enhancement Diagnosis Schwannoma Splaying of ICA and ECA No F CT, MRI R, psh Anterior Present Present Hypoattenuating with poor enhancement on CT; 2 41 M CT, MRI L, psh 6 4 Anterior Present Absent hypo- to isointense on T1-weighted imaging; heterogeneously bright on T2-weighted imaging; Schwannoma 3 22 F CT L, SH Anterior Present Absent marked but heterogeneous enhancement on gadolinium-enhanced Neurofibroma imaging 4 36 F CT L, SH 3 2 Posterior a Absent Present Schwannoma 5 30 M CT R, IH 3 3 Neutral Present Absent Neurofibroma 6 44 F CT R, SH 3 2 Anterior Absent b Absent Schwannoma 7 23 M CT R, psh 4 3 Anterior Present Present Schwannoma 8 30 M CT, MRI L, SH 4 3 Anterior Present Absent Schwannoma Schwannoma 9 69 M CT, MRI L, IH 3 3 Neutral Present Absent Isodense on CT; isointense on T1-weighted imaging; heterogeneously bright on T2-weighted imaging; marked and heterogeneous enhancement on gadolinium-enhanced imaging Ancient schwannoma M CT, MRI R, SH Anterior Present Absent Marked heterogeneity on both CT and MRI with several cystic areas and heterogeneously enhancing solid areas Ancient schwannoma F CT L, psh 3 2 Anterior Present Absent Central hypodense area of fluid density and isodense peripheral rind Note ICA = internal carotid artery, CCA = common carotid artery, IJV = internal jugular vein, ECA = external carotid artery, R = right side of neck, psh = predominantly suprahyoid although these masses straddle both the supra- and infrahyoid neck, L = left side of neck, SH = suprahyoid neck, IH = infrahyoid neck. a Vagal schwannoma tumor brought about a paradoxical posterior displacement of both the IJV and ICA without separating them. b Vagal schwannoma displaced both the IJV and ICA anteriorly without separating them from each other. schwannomas showed several nonenhancing cystic areas of CSF signal intensity and intervening T2 hyperintense, hyperenhancing solid areas (Fig. 5). Special Signs The split fat sign, wherein a rim of fat is seen surrounding the nerve sheath tumor (Fig. 5B) was well appreciated in four of the five patients who had MR studies available for review. Proximal and distal tails representative of the entering and exiting nerve root were also identified in four patients on coronal MR images (Figs. 5B and 5C). In two patients with schwannoma, T2-weighted imaging showed a narrow hyperintense rim surrounding a solid area of relatively lower signal intensity (Fig. 7). In two of the MR studies, the vagal schwannomas showed multiple small, circular, relatively low-signal-intensity areas surrounded by a background of hyperintensity on T2-weighted imaging, resulting in a fascicular appearance (Fig. 6B). Discussion The cervical vagus nerve runs within the carotid sheath between the ICA or CCA and the IJV as a neurovascular bundle, with the former anteromedial to the nerve and the latter anterolateral to it [4, 5]. Some of the reported variations to this normal anatomy are the cervical vagus nerve being anterior to the carotid artery [5, 6], the cervical vagus nerve containing accessory parathyroid tissue, or the cervical vagus nerve terminating within a lobe of the thyroid gland with a small distal continuation [7]. Within the jugular foramen, the vagus bears a well-marked ganglionic enlargement (4 mm in diameter) that is referred to as the jugular or superior ganglion. The inferior vagal ganglion, or nodose ganglion, is larger in diameter; is located just below the skull base; and is seen as a 2.5-cm-long cylindric structure that blends with the vagal trunk below [8, 9]. Several authors believe that vagal schwannomas tend to arise near this nodose ganglion and that neurofibromas arise more randomly along the course of the cervical vagus nerve [1, 10, 11]. Although tumor location was not recorded intraoperatively, from the fact that eight of the nine vagal schwannomas in this study (almost 89%) were predominantly located in the suprahyoid neck, we are inclined to attribute this pattern of distribution to the aforementioned hypothesis of their origin being at the nodose ganglion. Benign peripheral nerve sheath tumors are usually divided into two groups: schwannomas (neurilemmomas) and neurofibromas. They are thought to arise from neuroectodermal elements of the nerve sheath, such as Schwann cells, perineural fibroblasts, or their more primitive precursors [12, 13]. Schwannomas, as descendants of Schwann cells, arise outside the involved fascicle and grow eccentrically [14]. At imaging, this eccentric growth is visible only in large nerves, whereas the smaller nerves are completely obliterated by the schwannoma [15]. At AJR:197, July

4 Anil and Tan Fig year-old woman with neurofibroma of left cervical vagus nerve (patient 3 in Table 1). Axial contrast-enhanced CT image shows left-sided, wellcircumscribed, homogeneously hypodense mass displacing internal carotid artery (ICA) and external carotid artery (ECA) anteromedially. This pattern of mass effect is typical for carotid space lesion. Lack of enhancement is reflection of hypovascular nature of tumor. Tumor splaying internal jugular vein (IJV) and ICA suggests its vagal origin. histopathology, schwannomas are characterized by the presence of specific hypercellular areas (Antoni type A areas), with frequent nuclear palisading arrangements (Verocay bodies), interspersed with less dense reticular areas (Antoni type B areas) with myxoid tissue and more water content. The likely precursor of neurofibromas that is, the perineural fibroblast is a more embryologically primitive neuroectodermal cell than the Schwann cell. These tumors intertwine themselves within several fascicles of origin. Removal of neurofibromas is therefore more difficult than resection of schwannomas and is more likely to result in functional loss [14]. Unlike schwannomas, neurofibromas do not contain Antoni A and B areas. They are composed of interlacing fascicles of wavy elongated cells that often contain abundant collagen. Myxoid and degenerative areas are also not as abundant in neurofibromas as in schwannomas. Neurofibromas may be localized, plexiform, or diffuse. However, more than 90% are of the localized variety [15] like the two encountered in this study. In the head and neck region, neurofibromas are much rarer than schwannomas [8, 12, 14]. The results of the current study endorse this trend with a 2:9 ratio of neurofibromas to schwannomas. We did not find any sex predilection in contrast to earlier studies that noted male [16] or female [17] predominance. A Fig year-old man with schwannoma of left cervical vagus nerve (patient 2 in Table 1). Contrastenhanced CT scan shows left-sided, wellcircumscribed, heterogeneously hypodense mass with few areas of patchy enhancement. Anterior displacement of internal carotid artery (ICA), external carotid artery (ECA), and parapharyngeal fat and medial displacement of visceral space (white arrow) are shown here. The internal jugular vein is compressed by tumor and hence is not visualized in this image. wide age range of years is reported for head and neck schwannomas, whereas neurofibromas are seen in a slightly younger age group of years [18]. This study shows a similar trend with most of the schwannomas presenting in the fifth decade and both neurofibromas presenting in the third decade of life. A Fig. 3 Axial contrast-enhanced CT image of 69-year-old man with schwannoma (S) of left cervical vagus nerve located entirely in infrahyoid neck (patient 9 in Table 1). Well-defined hypodense mass displaces internal carotid artery (ICA) medially and internal jugular vein (IJV) laterally. There is no anterior displacement of artery probably because of resistance from anteriorly overlapping sternocleidomastoid muscle (SCM). Nerve sheath tumors of the cervical vagus nerve are often asymptomatic or are associated with minimal symptoms, and referral is usually initiated by the presence of a visible neck mass [1, 14]. Neurologic deficits are rarely seen at presentation. When present, the tumors are usually located in more tightly confined spac- Fig. 4 Axial contrast-enhanced CT images of 36-year-old woman with schwannoma of left cervical vagus nerve (patient 4 in Table 1). A, Well-defined, poorly enhancing tumor (S) displaces internal carotid artery (ICA) and internal jugular vein (IJV) in posterolateral direction. This is unlike anterior displacement of ICA that is normally seen in carotid space masses. Again, failure to separate ICA from IJV is unusual for vagal tumor. Small separation of vessels that otherwise show paradoxical posterior displacement favors vagal origin of tumor. ECA = external carotid artery. B, Image obtained 2 cm below level of A shows tumor (S) is splaying carotid bifurcation. Arteries are splayed apart but are not encased by tumor, and there is less than 180 of contact between tumor and vessels. ICA = internal carotid artery, ECA = external carotid artery, IJV = internal jugular vein. B 198 AJR:197, July 2011

5 CT and MRI of the Cervical Vagus Nerve es. Hoarseness of voice is the most common symptom. Dysphonia, dyspnea, dysphagia, cough, syncopal episodes, tongue weakness and atrophy, and Horner syndrome and tenderness and coughing on exerting pressure on the tumor are some of the other symptoms documented in the literature [1, 12 14, 16, 17]. Total surgical resection remains the treatment of choice for nerve sheath tumors because they are relatively radioresistant and some previously benign tumors can undergo malignant change. However, because this disease is generally benign, all efforts are made to preserve nerve function and quality of life. Postoperative neurologic deficiency is common but is often transient. The severity and nature of postoperative neurologic deficit depend on the extent to which nerve fascicles are sacrificed. Function-conserving surgeries such as nerve resection with primary anastomosis or neural graft interposition, tumor enucleation between adjacent healthy nerve fibers, tumor emptying (preserving tumor capsule), and even tumor shelling out (leaving gross tumor inside the capsule) have been described [14, 16, 17, 19]. Most nerve sheath tumors are solitary [15, 20] as noted in this study. Our observations with regard to the average size, margins, and location within the lateral neck are concordant with those in the literature. Because vagal tumors are located within the carotid space, they cause anterior displacement of the fat in the parapharyngeal space and medial displacement of the visceral space [21]. The intuitive explanation for the anterior displacement of the ICA seen with the majority A Fig year-old man with ancient schwannoma of right cervical vagus nerve (patient 10 in Table 1). A, Axial contrast-enhanced CT scan shows marked heterogeneity of well-circumscribed tumor with several cystic areas within and patchy enhancement of intervening solid areas. B, Coronal T1-weighted image (TR/TE, 378/12) shows heterogeneously hypointense mass. Note sliver of fat (arrowheads) seen around tumor representing split fat sign. At cranial end of tumor, entering nerve (arrow) is seen. Split fat sign and entering and exiting nerve sign suggest neurogenic origin of lesion. C, Coronal fat-suppressed T2-weighted turbo spin-echo image (4000/99) exquisitely reflects marked internal heterogeneity and extensive cystic change, hallmark of ancient schwannoma. It also shows peripherally entering nerve (arrow) that favors schwannoma over neurofibroma. Differentiating this lesion from malignant disease (e.g., metastatic node) would be difficult without identifying split fat sign and entering and exiting nerve sign. of the nerve sheath tumors of the cervical vagus nerve (eight of 11 patients, almost 73%) is the posterolateral location of the nerve relative to the artery. This pattern of mass effect is well documented [8, 22]. However, infrahyoid tumors did not show a similar anterior displacement of the CCA. In the two patients with an infrahyoid nerve sheath tumor, the artery was in a neutral position as compared with the contralateral side but showed a greater degree of medial displacement than a suprahyoid tumor of comparable size. In both cases, the sternomastoid belly overlapped the tumor anteriorly by virtue of its location in the lower neck. Presumably the resistance from the anterior tough muscular structure precluded anterior displacement of the artery, given that the artery would rather be displaced more medially along a path of lower resistance. In approximately 82% of the patients in our study, the tumor separated the IJV from the ICA or CCA. Furukawa et al. [23] had originally proposed this pattern of separation of the vessels as a result of the location of the vagus between the vein and artery as a radiologic sign for distinguishing tumor of vagal origin from that arising from the cervical sympathetic chain. Because the sympathetic chain runs posteromedial to both the artery and vein, it is likely to displace them together instead of insinuating between them [23 25]. We noted two exceptions to the criteria of Furukawa and colleagues [23]. First, in patient 6 from Table 1, the bulk of the tumor was growing posteriorly with mild anterior displacement of the IJV and ICA without B splaying them apart. Because the ICA is anteromedial to the cervical vagus nerve and the IJV is anterolateral to it, anterior displacement of both structures without separating them is conceivable. This pattern of mass effect makes the tumor indistinguishable from a schwannoma of the cervical sympathetic chain. Second, in patient 4 from Table 1, both the ICA and IJV were displaced posterolaterally without separation (Fig. 4). A corollary to the Furukawa criteria based on a similar observation was added by Saito et al. [24]: In the presence of posterolateral displacement of both the ICA and IJV, any observable separation of the vessels should favor a vagal lesion. This corollary to the paradigm of Furukawa and colleagues was valid in this patient. A schwannoma of the cervical sympathetic chain splaying the carotid bifurcation is a rare but well-documented observation [25]. However, we could not find any previous reference in the published literature to vagal schwannomas effecting a similar splaying of the ICA and ECA, as seen in three patients in our study. This observation contradicts certain observations in the published literature that almost rule out such a possibility [24, 26]. It may be noted that glomus vagale tumors that arise from the same site (i.e., nodose ganglion) as most of the vagal schwannomas are known to splay the carotid bifurcation when sufficiently enlarged [8]. A similar mechanism may be responsible for the splaying of the carotid bifurcation by a vagal schwannoma as well. The hypo- to isoattenuation of most of the nerve sheath tumors in this study with vari- C AJR:197, July

6 Anil and Tan able texture (heterogeneous or homogeneous) and poor enhancement is consistent with the well-known CT appearances of these tumors [3, 22]. Nerve sheath tumors are hypovascular tumors, but one third of schwannomas may appear hyperenhancing on CT because of contrast stasis secondary to obstruction of venous drainage [8]. The heterogeneity of schwannomas is attributed to cystic degeneration, xanthomatous change, or areas of relative hypocellularity adjacent to densely cellular or collagenous regions [27]. On MRI, nerve sheath tumors are generally hypo- to isointense on T1-weighted images and hyperintense on T2-weighted images depending on cellularity. The T2 relaxation time diminishes with increasing cellularity; hence, the hypercellular Antoni type A areas are less bright than the fluid-rich Antoni type B areas. On routine gadoliniumenhanced studies, marked heterogeneous enhancement is almost a rule unlike the usual poor enhancement seen on CT [3, 18]. In the literature on musculoskeletal imaging, we encounter several MRI signs for peripheral nerve sheath tumors. During this study, we found that extrapolation of these radiologic signs to evaluate tumors of the cervical vagus nerve is feasible and relevant. Neurovascular bundles are surrounded by fat, so masses arising from the nerves within these bundles maintain a rim of fat around them (split fat sign) [15, 28]. This sign was depicted on most of the MR studies evaluated during A Fig year-old woman with schwannomas of right cervical vagus nerve. A, Axial T1-weighted image (TR/TE, 560/16) shows hypointense mass in right carotid space. B, On axial T2-weighted fast spin-echo image (3000/102), right carotid space mass is heterogeneously hyperintense. There are multiple small, circular low-signalintensity areas surrounded by background of mild hyperintensity (slightly to left of center of tumor and adjacent to its anterior circumference) representing fascicular appearance. This appearance on MRI is more commonly seen in schwannomas than in neurofibromas. C, On axial fat-suppressed contrast-enhanced T1-weighted image (420/16), right carotid space mass shows intense heterogeneous enhancement, once again typical for schwannoma. this study. A fusiform shape of the tumor with identification of an entering and exiting nerve is highly suggestive of a neurogenic tumor. We found this sign to be best depicted on coronal MR images. If the parent nerve enters and exits centrally, neurofibroma is the favored diagnosis; when the parent nerve is found to be eccentric, schwannoma is more likely [28]. Less often, nerve sheath tumors show a fascicular appearance described as small ringlike structures on T2-weighted images corresponding to the fascicular bundles within the nerve [15, 28]; these bundles were well seen on axial images of the neck. Neurofibromas with collagenous tissue in the center and myxoid tissue in the periphery may exhibit a target sign seen on T2-weighted imaging as a hypointense area in the center of an otherwise bright lesion [15]. Schwannomas with hypercellular Antoni type A areas in the center and more reticular Antoni type B areas along the periphery may project a similar appearance. However they are distinguished by the narrow width of the hyperintense rim. The two schwannomas that showed this hyperintense rim sign on T2-weighted imaging had a bright rim less than one fourth of the diameter of the lesion, as described earlier by Jee et al. [28]. The target sign was not encountered during this study. Ancient schwannoma is a subtype of schwannoma characterized by marked degenerative changes and is pathologically typified by perivascular hyalinization, calcification, ossification cystic necrosis, relative loss of B Antoni type A tissue, and degenerative nuclei that may be misinterpreted as sarcomatous pleomorphisms. The two ancient schwannomas from this study were drastically more heterogeneous than the other tumors. MRI is clearly superior to CT in depicting this heterogeneity. Cystic areas with pericystic enhancement, blood products, and calcification are some of the characteristics of this lesion [29]. Metastatic lymph nodes, paragangliomas, and schwannoma of the cervical sympathet- Fig year-old man with schwannoma of left cervical vagus nerve. Coronal fat-suppressed T2- weighted turbo spin-echo image (TR/TE, 3140/99) reveals mass with thin hyperintense rim. Such hyperintense rim, usually less than one fourth of width of lesion, is feature of schwannoma. Thicker rim would be representative of target sign, which is typically seen in neurofibromas. C 200 AJR:197, July 2011

7 CT and MRI of the Cervical Vagus Nerve ic chain are the usual differential diagnoses to be considered. Metastatic lymph nodes are often multiple with a known primary facilitating the diagnosis [8]. However, in the absence of periadenitis or extranodal extension of disease, it may be difficult to distinguish a solitary metastatic node from a heterogeneous schwannoma with cystic areas or from a typical ancient schwannoma. In such circumstances, some of the aforementioned radiologic signs specific for tumors of neurogenic origin may be useful. Paragangliomas show early arterial enhancement on CT; they are hypervascular lesions (whereas nerve sheath tumors are hypovascular) and show certain characteristic MRI appearances such as scattered flow voids (salt and pepper appearance), that usually enable the appropriate diagnosis [30]. Their pattern of splaying the carotid bifurcation is also distinct from that of vagal tumors. In most cases, separation of the IJV and ICA is useful to distinguish vagal tumors from sympathetic chain tumors [23]. In conclusion, we present the spectrum of CT and MRI findings of nerve sheath tumors of the cervical vagus nerve. These tumors tend to be well circumscribed, solitary, oval to round, and localizable to the carotid space. On CT, they are mostly hypo- to isoattenuated with poor-to-moderate heterogeneous contrast enhancement. They are heterogeneously bright on T2-weighted images with inhomogeneous but intense postgadolinium enhancement. In tumors with marked heterogeneity, ancient schwannoma should be suspected. Nerve sheath tumors of the cervical vagus routinely displace the ICA anteriorly or anteromedially, fat in the parapharyngeal space anteriorly, and the visceral space medially. However, tumors in the infrahyoid neck may not show similar anterior displacement of the CCA, probably because of resistance from an overlapping sternomastoid. Nerve sheath tumors of the cervical vagus nerve tend to separate the IJV from the ICA; however, exceptions to this rule are possible. If the IJV and ICA are displaced anteriorly without separation, then distinguishing them from a sympathetic trunk tumor is difficult. In case of the rare occurrence of simultaneous posterolateral displacement of the IJV and ICA, slight separation of the vessels should favor the diagnosis of a vagal tumor instead of a sympathetic chain tumor. Although not previously documented in literature, vagal schwannomas can splay the ICA and ECA without filling the crotch of the bifurcation or encasing the arteries. Radiologic signs of neurogenic tumors such as the split fat sign, entering and exiting nerves, fascicular appearance, hyperintense rim sign that are well documented in the literature on musculoskeletal imaging may be elicited to diagnostic advantage while evaluating these carotid space tumors. References 1. Green JD Jr, Olsen KD, DeSanto LW, Scheithauer BW. Neoplasms of the vagus nerve. Laryngoscope 1988; 98: Dalui D. Benign solitary schwannoma of the cervical vagus nerve. Indian J Otolaryngol Head & Neck Surg 1980; 32: Som PM, Sacher M, Stollman AL, Biller HF, Lawson W. Common tumours of the parapharyngeal space: refined imaging diagnosis. Radiology 1988; 169: Daniels DL, Williams AL, Haughton VM. Jugular foramen: anatomic and computed tomographic study. AJR 1984; 142: Giovagnorio F, Martinoli F. Sonography of the cervical vagus nerve: normal appearance and abnormal findings. AJR 2001; 176: Gibson A. Bilateral abnormal relationship of the vagus nerve in its cervical portion. J Anat Physiol 1915; 69: Tubbs RS, Loukas M, Shoja MM, et al. An unreported variation of the cervical vagus nerve: anatomical and histological observations. Folia Morphol (Warsz) 2007; 66: Som PM, Curtin HD. Parapharyngeal space. In: Som PM, Curtin HD, eds. Head and neck imaging, 3rd ed., vol. 2. St. Louis, MO: Mosby, 1996: Gray H. Anatomy of the human body. Philadelphia, PA: Lea & Febiger, 1918; Bartleby.com, Accessed September 6, St. Pierre S, Theriault R, Le Clerc J. Schwannomas of the vagus nerve in the head and neck. J Otolaryngol 1985; 14: Galli J, Almadori G, Paludetti G, Rosignoli M, Corina L, Ieraci A. Plexiform neurofibroma of the cervical portion of the vagus nerve. J Laryngol Otol 1992; 106: Kline DG, Hudson AR, eds. Tumors involving nerve. In: Nerve injuries: operative results for major nerve injuries, entrapments, and tumors. Philadelphia, PA: Saunders, 1995: Peetermans JFE, Van de Heyning PH, Parizel PM, et al. Neurofibroma of the vagus nerve in the head and neck: a case report. Head Neck 1991; 13: Gilmer-Hill HS, Kline DG. Neurogenic tumours of the cervical vagus nerve: report of four cases and review of the literature. Neurosurgery 2000; 46: Murphey MD, Smith WS, Smith SE, et al. From the archives of the AFIP: imaging of musculoskeletal neurogenic tumors radiologic-pathologic correlation. RadioGraphics 1999; 19: Leu YS, Chang KC. Extracranial head and neck schwannomas: a review of 8 years experience. Acta Otolaryngol 2002; 122: Zhang H, Cai C, Wang S, Liu H, Ye Y, Chen X. Extracranial head and neck schwannomas: a clinical analysis of 33 patients. Laryngoscope 2007; 117: Weber AL, Montandon C, Robson CD. Neurogenic tumors of the neck. Radiol Clin North Am 2000; 38: de Araujo CE, Ramos DM, Moyses RA, Durazzo MD, Cernea CR, Ferraz AR. Neck nerve trunks schwannomas: clinical features and postoperative neurologic outcome. Laryngoscope 2008; 118: Kransdorf MJ, Murphey MD. Neurogenic tumors. In: Kransdorf MJ, Murphey MD, Bralow L, eds. Imaging of soft tissue tumors. Philadelphia, PA: Saunders, 1997: Harnsberger HR, Osborn AG. Differential diagnosis of head and neck lesions based on their space of origin. Part 1. The suprahyoid part of the neck. AJR 1991; 157: Silver AJ, Mawad ME, Hilal SK, et al. Computed tomography of the carotid space and related cervical spaces. Part II. Neurogenic tumors. Radiology 1984; 150: Furukawa M, Furukawa MK, Katoh K, et al. Differentiation between schwannoma of the vagus nerve and schwannoma of the cervical sympathetic chain by imaging diagnosis. Laryngoscope 1996; 106: Saito DM, Glastonbury CM, El Sayed IH, et al. Parapharyngeal space schwannomas. Arch Otolaryngol Head Neck Surg 2007; 133: Anil G, Tan TY. Imaging characteristics of schwannoma of the cervical sympathetic chain: a review of 12 cases. Am J Neuroradiol 2010; 31: Wang CP, Hsiao JK, Ko JY. Splaying of the carotid bifurcation caused by a cervical sympathetic chain schwannoma. Ann Otol Rhinol Laryngol 2004; 113: Cohen LM, Schwartz AM, Rockoff SD. Benign schwannoma: pathological basis for CT inhomogeneities. AJR 1986; 147: Jee WH, Oh SN, McCauley T, et al. Extraaxial neurofibromas versus neurilemmomas: discrimination with MRI. AJR 2004; 183: Isobe K, Shimizu T, Akahane T, Kato H. Imaging of ancient schwannoma. AJR 2004; 183: Rao AB, Koeller KK, Adair CF. Paragangliomas of the head and neck: radiologic-pathologic correlation. RadioGraphics 1999; 19: AJR:197, July