Intraneural Perineurioma: A Systematic Review With Illustrative Cases

In the English language literature, intraneural perineurioma is loosely referred to as localized hypertrophic neuropathy/mononeuropathy, hypertrophic neurofibroma, hypertrophic neuropathy, intraneural neurofibroma, interstitial hypertrophic neuropathy, pseudo–onion bulb neuropathy/mononeuropathy and, occasionally, onion bulb neuropathy/mononeuropathy. These various names describe a unique clinical and radiologic dyad occurring primarily in adolescents or young adults, who present with a slow-growing, painless mononeuropathy that characteristically causes progressive loss of motor function.1 A concomitant sensory deficit is uncommon, and a history of preceding trauma is currently debatable. Imaging modalities frequently disclose fusiform, segmental enlargement of the affected nerve, which is histologically expanded predominantly or exclusively by concentric whorls of either spindle-shaped perineurial cells (pseudo–onion bulbs) or Schwann cells (true onion bulbs), with or without the presence of interspersed collagen bundles and/or occasional residual axon–Schwann cell complexes at the whorls' center. This obscure nomenclature creates difficulty in accurately distinguishing true cases of intraneural perineurioma from histologically similar Schwann cell– derived reactive/demyelinating processes or the inherited polyneuropathies of Charcot-Marie-Tooth and Dejerine-Sottas.2,3 

The gold standard to distinguish perineurial cells from Schwann cells is electron microscopy, whereby perineurial cells characteristically show thin, elongated cytoplasmic processes, an incomplete basal lamina, poorly formed tight junctions, and pinocytotic vesicles. These characteristic features were initially observed by Lallemand et al4 and subsequently verified by others.5–8 By immunohistochemistry, perineurial cells are readily distinguished from Schwann cells by their positive staining for epithelial membrane antigen and negative staining for S100 protein.9–11 Despite the advent of this newer technology, intraneural lesions composed predominantly or exclusively of concentric whorls of perineurial cells are still, albeit less commonly, being classified as localized hypertrophic mononeuropathy, etc—names that should be reserved for those histologically similar intraneural onion bulb lesions composed of Schwann cells.

In summary, our primary goal was to objectively and thoroughly review the English language literature and accurately identify and summarize the clinicopathologic characteristics of definitive cases of intraneural perineurioma. In addition to describing two pediatric cases of intraneural perineurioma involving the ulnar nerve and an unnamed nerve of the dorsolateral tongue, we discuss the histologic, immunohistochemical, cytogenetic/molecular, and ultrastructural characteristics and the differential diagnosis of this uncommon and unique neoplasm.

Epithelial membrane antigen (EMA; clone E29, dilution 1:150; Dako Inc, Carpinteria Calif), vimentin (VIM; clone 3B4, dilution 1:300; Dako), S100 protein (S100; clone DR96 & BC96, dilution 1: 500, Biocare Medical Inc, Concord, Calif), neurofilament protein (NFP; clone 2F11, dilution 1:25; Dako), and Ki-67 (clone MIB-1, dilution 1:500; Dako) were applied using a BioGenex i6000 automated instrument (BioGenex Inc, San Ramon, Calif). Heat-induced epitope retrieval (HIER), which consisted of steam for 20 minutes in citrate buffer, was performed for each antibody, except NFP, which did not undergo an epitope retrieval procedure.

Fresh tissue was placed in 4% glutaraldehyde and processed for electron microscopic analysis. Ultrathin sections were stained with uranyl acetate and lead citrate and examined with a JEM-1210 transmission electron microscope (JEOL USA, Peabody, Mass).

Definitive cases of intraneural perineurioma reported in the English language literature were identified through a Medline search (as of November 2006) using “perineurioma, intraneural perineurioma, perineurial, hypertrophic mononeuropathy, localized hypertrophic mononeuropathy, hypertrophic neurofibroma, hypertrophic neuropathy, intraneural neurofibroma, interstitial hypertrophic neuropathy, pseudo onion-bulb neuropathy, pseudo onion-bulb mononeuropathy, onion-bulb neuropathy, onion-bulb mononeuropathy, pseudo onion-bulb and onion-bulb” as key words. The original articles were obtained for all citations, and additional cases were identified from the cited references. In addition, the anatomic pathology databases of Baylor College of Medicine and its affiliated hospitals (Texas Children's Hospital, Ben Taub General Hospital, and Veterans Affairs Medical Hospital) in Houston, Tex, were queried for additional cases of intraneural perineurioma using the aforementioned key words.

The histopathologic features of each reported case were reviewed to identify definitive cases of intraneural perineurioma, despite the original author(s) classifying nomenclature. Inclusion criteria consisted of (1) histologic demonstration of nerve expansion predominantly or exclusively by concentric whorls of spindle-shaped perineurial cells having thin, elongated, eosinophilic, cytoplasmic processes that may or may not envelop residual axons and/or Schwann cells, and with or without interspersed collagen bundles; (2) immunohistochemical confirmation of perineurial cell lineage by demonstrating positive staining with EMA and negative staining with S100; and/or (3) ultrastructural confirmation of perineurial cell lineage by demonstrating thin cytoplasmic processes with an incomplete basal lamina, poorly formed tight junctions, and pinocytotic vesicles, all with or without interspersed collagen bundles.

A 6-year-old Caucasian girl presented with a 1-year history of a right-sided tongue mass that was unassociated with trauma, pain, motor, or sensory abnormalities. Her medical history was significant only for a febrile seizure at 1 year of age. There was no personal or family history of neurofibromatosis (NF) or other inheritable neuropathies. Physical examination disclosed a 1-cm, soft, nontender, pink mass on the right lateral border of the tongue. Lingual movement was normal, and there were no abnormalities with swallowing. Six months later the mass was surgically excised. Her postoperative course was unremarkable, and a 4-month follow-up visit showed no evidence of recurrence or motor or sensory deficits.

A 5-year-old Caucasian boy was noted to have difficulty extending his right fourth and fifth fingers. The duration of the problem was unknown, and there was no associated pain or history of trauma. Physical examination was significant for a flexure deformity of the right fourth and fifth proximal and distal interphalangeal joints, mild dysmorphic facies (cup-shaped prominent ears, a flat nasal bridge, and minor telecanthus), and a total of 4 café au lait spots on the left flank, right arm, and legs, the largest of which measured 1.5 cm in diameter. Ophthalmologic evaluation was negative for Lisch nodules or other ocular pathology. The patient did not meet criteria for NF. There was no family history of NF or other inheritable neuropathies. Of note, his maternal grandmother had “a few small café au lait spots,” and his older brother and maternal uncle had similar dysmorphic facies. His medical history was significant for asthma and repeated ear infections, requiring bilateral myringotomy and adenoidectomy by 2 years of age. Electromyography confirmed motor palsy of the right ulnar nerve; however, no sensory deficit was present. Full-body contrast-enhanced magnetic resonance imaging revealed a solitary, homogeneously enhancing 1.5-cm ovoid mass involving the right ulnar nerve and located just proximal to the medial condyle. There was no evidence of metastatic disease. Surgical resection with nerve grafting was performed 6 months later. He underwent physical therapy with some motor improvement (ie, increased strength and attenuation of the claw deformity), and there have been no associated sensory abnormalities. Multiple follow-up visits during the last 3 years have detected no evidence of local tumor recurrence. The patient is currently 8 years of age and is being followed annually until puberty to assess for the possible development of NF.

The specimen from case 1 was received in 10% neutral-buffered formalin and consisted of a 0.8 × 0.6 × 0.5-cm, glistening light tan–to-pink, rubbery nodule with a flattened base. The cut surface was uniformly light tan with peripherally located mucosa. The entire specimen was submitted for routine histologic processing.

The specimen from case 2 consisted of an enlarged, fusiform portion of resected light tan nerve (1.5 cm in length × 0.7 cm maximum diameter), which had a uniformly light tan cut surface. A representative cross section was submitted for frozen section, with a preliminary diagnosis of an “organized Schwann cell proliferation, await permanent sections for further classification.” The remaining specimen was submitted for routine histologic processing as well as cytogenetic and ultrastructural analysis.

Histologic and immunohistochemical analysis of both cases disclosed an intraneural perineurioma. Hematoxylin-eosin–stained sections from case 1 showed a plexiform pattern of expanded nerve fascicles embedded within a fibrovascular stroma, whereas case 2 showed an expanded nerve with alteration of normal fascicular compartmentalization. In both cases, the expansion was due to a proliferation of pseudo–onion bulb structures with interspersed collagen bundles, the former composed of concentric whorls of spindle-shaped cells with thin, elongated, eosinophilic, cytoplasmic processes (Figure 1, A and B). The spindle cells were of perineurial origin due to positive membranous staining with epithelial membrane antigen and negative staining for S100 protein (Figures 2, A and B, and 3, A and B). At the center of the concentric whorls, NFP highlighted occasional residual axons in case 1, and NFP and S100 highlighted several axon–Schwann cell complexes in case 2 (Figures 3, A and B, and 4, A and B). Ki-67 (MIB-1) was performed on case 2 and demonstrated less than 2% positive nuclear staining (image not shown). Ultrastructural analysis of case 2 confirmed that the pseudo–onion bulbs were concentric layers of perineurial cells, which enveloped residual axon–Schwann cell complexes. The elongated cytoplasmic processes demonstrated an incomplete basal lamina, poorly formed tight junctions, and pinocytotic vesicles, as well as interspersed collagen bundles (Figure 5, A through D). Cytogenetic analysis of case 2 was uninformative due to fibroblast overgrowth within cell culture.

Definitive cases of intraneural perineurioma included in our investigation were selected based on appropriate histology and verification of perineurial cell lineage by either immunohistochemical and/or ultrastructural analysis. Cases excluded from our study included the following: (1) those demonstrating classic histologic features of intraneural perineurioma but without verification of perineurial cell lineage by immunohistochemical or ultrastructural analysis12–17; (2) those demonstrating classic histologic features of intraneural perineurioma but with subsequent demonstration of Schwann cell lineage by immunohistochemistry and/or ultrastructural analysis18–26; (3) cases of extraneural (soft tissue) perineurioma27–63; and (4) cases in which the pathologic process was due or related to a ganglion cyst, trauma, and/or reactive/demyelinating processes.14,64–66 Of note, Gruen et al67 in one of the largest series of localized hypertrophic neuropathy to date, only demonstrated perineurial cell lineage by immunohistochemistry in “… six of the more recent cases …” of a total of 14. With this qualifying statement, we only included the 6 most recent cases from that study; the remaining 8 cases were excluded. Also, Ariza et al11 performed immunohistochemical analysis on 2 cases of intraneural perineurioma, which were previously described in detail by Bilbao et al68 in 1984; the clinicopathologic data of the latter is included in our investigation.

In summary, 51 definitive cases of intraneural perineurioma were identified from 29 articles within the English language literature, based upon our inclusion criteria.3,4,10,67–92 Additionally, 2 cases of intraneural perineurioma were identified from our computer database search. The clinicopathologic data from the published and 2 current cases, where available, are summarized in Tables 1 and 2. In brief, a total of 53 tumors were identified from 23 males and 29 females, with only one author describing synchronous bilateral intraneural perineuriomas of the ulnar nerves in a 36-year-old woman.71 Taking this case into account, the male-female ratio was 0.8. Tumor size was provided in all but 12 cases, with an average, median, and range of 5.4, 4.0, and 0.5 to 18 cm, respectively. A total of 37 cases reported no history of trauma preceding tumor development; the remaining 16 cases did not provide this information. The duration of symptoms prior to diagnosis was, on average, 53 months, with a median and range of 30 months and 2 to 300 months, respectively. Motor abnormalities were present in 43 cases (81%), absent in 5 cases (9.5%), and not reported in 5 cases (9.5%). Concomitant sensory abnormalities were present in 13 cases (25%), absent in 34 cases (64%), and not reported in 6 cases (11%).

Treatment modalities consisted of surgical resection with nerve grafting (15/53), surgical resection (11/53), excisional biopsy (11/53), excision (6/53), neurolysis (2/53), subtotal excision (1/53), surgical resection with primary anastomosis/coaptation (1/53), and none specified (6/53). A total of 35 cases provided information pertaining to the postoperative follow-up period (Table 1), with a mean, median, and range of 20, 12, and 1 to 72 months, respectively. There was no evidence of tumor recurrence, nor was there mention of metastatic disease in any of the 35 cases. Of the 16 cases treated by surgical resection with nerve grafting or primary anastomosis, 5 (31%) and 11 (69%) reported no improvement and some improvement in motor function, respectively. Personal and family history was available for 22 cases, with pertinent findings identified in only 4. One patient had Beckwith-Wiedemann syndrome,73 2 patients had café au lait spots but failed to meet diagnostic criteria for NF (current case 2),90 and 1 patient had a personal and family history of NF.87 

Patient age was available in each case, with an average, median, and range of 23, 20, and 2 to 64 years, respectively. When stratified by age and sex, there was no significant difference in the percentage of reported cases by sex; however, most cases (46/53; 87%) occurred in individuals 39 years of age or younger (Table 2). Only 1 of the 53 cases failed to disclose a site designation.3 Affected sites in order of decreasing frequency were the ulnar nerve (9/53; 17%), median nerve (6/53; 11%), peroneal nerve (5/53; 9%), posterior interosseous nerve (5/53; 9%), sciatic nerve (4/53; 8%), radial nerve (4/53; 8%), brachial plexus (4/53; 8%), femoral nerve (3/53; 6%), tongue (3/ 53; 6%), and tibial nerve (2/53; 4%). The remaining sites consisted of buccal mucosa, intratemporal facial nerve, mandibular dental nerve, the jugulocarotid region of the neck, oculomotor nerve, C8/T1 nerve root, volar aspect of the wrist, and an undesignated location; each site only contributed 1 reported case (1/53; 2% each).

Molecular analyses were performed on only 4 of the 53 reported cases. One case demonstrated the following abnormal karyotype: 45,XX,add(14)(p13),−22,add(22)(q11.2)[cp15]/46,XX,add(22)(q11.2)[1].79 Using a combined chromosome 14/22 centromeric probe, fluorescence in situ hybridization of paraffin-embedded tissue from a second case demonstrated 3 fluorescence in situ hybridization signals in 53% of tumor cells, indicating monosomy for the centromeric region of either chromosome 14 or 22.79 Using an LSI BCR/ABL locus-specific probe, a third case was shown to have a deletion of the BCR locus (22q11) in 75% of tumor cells.82 Routine cytogenetic studies were attempted in 1 of our 2 cases (case 2); however, the results were uninformative due to fibroblast overgrowth within cell culture.

Regarding the peripheral nervous system, the perineurium is composed of concentric layers of perineurial cells, which envelop myelinated and unmyelinated nerve fascicles and are continuous with the pia-arachnoid membrane at the level of the nerve roots based upon comparative histopathologic analyses.11,59–68,93–96 In contrast to their normal role within the perineurium, perineurial cells are uncommonly implicated in neoplastic processes. Perineurioma, according to the current World Health Organization classification of nervous system tumors, is a benign tumor of neoplastic perineurial cells.1 Histologically, perineurioma is subclassified as “intraneural” or “extraneural (soft tissue)” if the perineurial cell proliferation is confined within or unassociated with a nerve, respectively. The more common extraneural (soft tissue) form, with the exception of the sclerosing variant, usually arises within the limbs or trunk of middle-aged women, less commonly in the retroperitoneum, gastrointestinal tract, and paratesticular region43,46,47,97 and rarely within the ventricular system.42 The sclerosing variant of soft tissue perineurioma arises within the fingers or palm of young adults, with a male predilection.98 The rarely encountered intraneural perineurioma, however, typically presents in adolescents or young adults as a slow-growing, painless mononeuropathy that characteristically causes progressive loss of motor function and, uncommonly, a sensory deficit.1 Imaging studies frequently disclose fusiform, segmental enlargement of the affected nerve, usually less than 10 cm; however, the entire nerve may be affected.1 

Although perineurioma was initially deemed a reactive process, mounting evidence supports its true neoplastic nature. Initial molecular findings in two cases of intraneural perineurioma demonstrated abnormalities and/or monosomy of the long arm of chromosome 22,79 a region known to harbor molecular abnormalities in other nervous system tumors—meningioma and schwannoma.99 Subsequent studies of soft tissue perineurioma, focusing on the region 22q11-q13.1, including the NF-2 gene (22q12.2), have not only confirmed monosomy of the long arm of chromosome 22 but cryptic deletions and/or point mutations within this region as well.72,100–103 Lasota et al101 performed DNA sequencing on exons 1 through 15 of the NF-2 gene in 8 cases of soft tissue perineurioma and demonstrated 5-point mutations in 4 of the cases—2 in the 5′-untranslated region and 1 each in exons 3, 6, and 8. More recently, fluorescence in situ hybridization analysis of a single case of intraneural perineurioma that focused on the BCR locus (22q11) showed a deletion of this region in 75% of analyzed nuclei.82 Despite abnormalities within this region, to our knowledge there has been no definitive association between intraneural or soft tissue perineurioma and NF. Based on our review, only 1 patient included in our meta-analysis had NF, and although 2 other patients had multiple café au lait spots, they did not meet diagnostic criteria for NF. It should be noted, however, that personal and family history was only available in 22 of the 53 cases. To date, karyotypic findings are available for only 6 cases of perineurioma in the English language literature, with both soft tissue and intraneural perineurioma containing relatively few abnormalities.72,79,102 However, recent data suggest abnormalities of chromosome 10 may be unique to sclerosing perineurioma.72,102 

For practical purposes, the differential diagnosis for localized, fusiform enlargement of a peripheral nerve includes schwannoma, neurofibroma, traumatic neuroma, intraneural perineurioma, LHM and its synonyms and, less likely, chronic inflammatory demyelinating polyneuropathy (CIDP) or the inherited polyneuropathies of Charcot-Marie-Tooth (CMT) and Dejerine-Sottas (DS), the latter 3 of which should be distinguishable by clinical history, physical examination, and imaging modalities. By routine histology and immunohistochemistry, CIDP, CMT, DS, and LHM demonstrate similar-appearing S100-positive, EMA-negative onion bulbs in a background of collagen. The presence of macrophages is seen in LHM and CIDP, and the absence of macrophages is seen in DS and CMT.3 For the most part, neurofibroma, schwannoma, and traumatic neuroma lack a significant onion bulb component and should be distinguishable by histology; the latter is also preceded by a history of trauma. The pseudo–onion bulbs of intraneural perineurioma are EMA positive and S100 protein negative compared with other Schwann cell– derived S100 protein positive, EMA-negative onion bulb peripheral nerve sheath lesions. In either situation, if present, NFP and S100 protein will highlight residual axons and Schwann cells, respectively, at the center of the concentric whorls. Because perineurial cells can demonstrate weak EMA staining intensity, alternate antibodies, including CD57 (Leu-7), claudin-1, collagen IV, GLUT-1, laminin, and PGP 9.5 have been recently evaluated for their diagnostic utility in distinguishing perineurioma from neurofibroma and schwannoma. In brief, immunoreactivity (percent positive) for these neoplasms was (1) perineurioma: EMA (99%), S100 (3%), CD57 (0%), claudin-1 (42%), collagen IV (96%), GLUT-1 (90%), laminin (90%), and PGP 9.5 (100%)^*; (2) schwannoma: EMA (19%), S100 (98%), CD57 (75%), claudin-1 (60%), collagen IV (100%), GLUT-1 (90%), and PGP 9.5 (0%)^†; and (3) neurofibroma: EMA (15%), S100 (96%), CD57 (71%), claudin-1 (67%), GLUT-1 (100%), and PGP 9.5 (100%).^‡ Of special note, EMA, claudin-1, and GLUT-1 immunoreactivity within neurofibroma and schwannoma was confined to elongated spindle cells with delicate cytoplasmic processes predominantly at the periphery or in the capsule of each lesion, suggesting remnants of the perineurial sheath.95,105,106 Additionally, claudin-1–positive, perivascular dendritic cells were noted within neurofibroma and schwannoma, suggesting perineurial cells recapitulating the normal blood-nerve barrier.105 Immunoreactivity for CD57 (Leu-7) was present in neurofibroma and schwannoma and absent in perineurioma. Collagen IV was not evaluated in neurofibroma, and laminin was not evaluated in schwannoma or neurofibroma. Furthermore, to our knowledge, these antibodies (excluding EMA, CD57, and S100 protein) have not been evaluated in cases of S100-positive, EMA-negative Schwann cell–derived onion bulb lesions, due to reactive/ demyelinating processes or the inherited polyneuropathies of CMT or DS. Based on our immunohistochemical review, we feel that EMA and S100 should be initially used and, if necessary, followed by claudin-1 and GLUT-1, with the caveat of recognizing their potential diagnostic pitfall in neurofibroma and schwannoma, whereby claudin-1 and GLUT-1 immunoreactivity has been documented to highlight residual perineurial cells at the periphery or in the capsule of the tumor, and claudin-1 will furthermore highlight perivascular dendritic cells within the lesion. The immunoreactivity of CD57 may also serve a useful function in distinguishing perineurioma from neurofibroma and schwannoma, as CD57 immunoreactivity, to our knowledge, has not been reported in perineurioma. In any event, ultrastructural analysis of either fresh or paraffin-embedded tissue will provide the definitive answer.

In review of malignant peripheral nerve sheath tumors with perineurial cell differentiation from the English language literature, confirmed either by immunohistochemistry,117,118 ultrastructural analysis,119 or both immunohistochemistry and ultrastructural analysis,120–124 not a single reported case described a preceding benign perineurioma (soft tissue or intraneural) undergoing malignant transformation. The absence of molecular information on these cases precludes correlation with those associated with perineurioma (soft tissue or intraneural), thereby prohibiting inferences as to whether or not malignant peripheral nerve sheath tumors with perineurial cell differentiation are a continuum of benign perineurioma (soft tissue or intraneural). In conclusion, based upon our systematic review, intraneural perineurioma should be regarded as a benign tumor that does not recur or metastasize and has little if any association with neurofibromatosis.