Microcystic/Reticular Schwannoma of the Proximal Sigmoid Colon: Case Report With Review of Literature Open Access

A 61-year-old man undergoing screening for malignant neoplasm, was found on endoscopy to have a 12-mm polyp in the proximal sigmoid colon, which was resected and retrieved. The patient had his last clinical contact at this hospital in 2009 for unrelated complaints.

The formalin-fixed, paraffin-embedded tissues were analyzed for morphology. Mucin stains and immunohistochemistry were performed subsequently using the following antibodies: S100, cytokeratin, CD34, CD68, synaptophysin, chromogranin, CD56, smooth muscle actin, CD117, HMB45, and MIB1 (Table 1).

Electron microscopic examination was performed retrospectively on formalin-fixed, paraffin-embedded tissue, and thus was somewhat limited.

A 0.8 × 0.6 × 0.6-cm tan pink polyp was received in formalin; it was inked, bisected, and submitted in its entirety for histologic examination. The histology (Figure 1, A and B) showed a 7-mm-diameter polypoidal nodular proliferation composed of unremarkable colonic glands and underlying submucosal stroma forming a well-circumscribed nodule composed mainly of multivacuolated cells with a pseudo–signet ring/lipoblastic-like appearance organized in nests surrounded by small capillaries. This proliferation merged with the more usual spindle cell proliferation suggestive of benign spindle cell neoplasm and shared similar immunophenotype makeup. There was diffuse cytoplasmic staining for S-100 (Figure 1, C) antibody and focal (<20% of cells) CD34 positivity. The cells were negative for cytokeratin (Figure 1, D), synaptophysin, chromogranin, CD56, smooth muscle actin, CD117, and HMB 45 markers, thus ruling out the possibilities of signet ring cell carcinoma, neuroendocrine neoplasm, smooth muscle tumor, gastrointestinal stromal tumor (GIST), and melanoma. The lipoblast-like cells were also negative for histiocytic marker CD68, which highlighted occasional normal histiocytes within the stroma. Additionally, in support of the nonepithelial nature of this proliferation is the negativity for mucin stains: mucicarmine and Alcian blue with periodic acid–Schiff. Alcian blue pH 2.5 without and with hyaluronidase was positive, confirming that the extracellular substance was an acidic mucosubstance. Proliferative activity (assessed with MIB-1 immunohistochemical stain) was less than 2%.

No areas of necrosis, nuclear atypia, or mitosis were identified. The deep margin was uninvolved by the proliferation, and subsequently this lesion was considered completely excised.

Because this entity had not been previously described, in 2007 this case was signed out as a peripheral nerve sheath tumor with prominent pseudo–signet ring/lipoblastic-like cell features, consistent with a variant of schwannoma.

The ultrastructural findings are depicted in Figure 2. Lesional cells showed elongated nuclei with dispersed euchromatin and marginated heterochromatin. Spindle cells showed irregular cytoplasmic borders with some variable interdigitation. Rare nondesmosomal junctions were present. The interstitium contained banded collagen bundles; however there were no Luse bodies (wide fiber-spaced collagen). No definite external lamina was appreciated; however, there was some nonspecific condensation of extracellular matrix around tumor cells. The vacuolated signet ring areas showed some cytoplasmic clearing; however, these peculiar signet ring–like areas were primarily the result of accumulated flocculent extracellular matrix (histochemically acid mucosubstance). Cytoplasmic organelles were not well preserved, but both the spindle and vacuolated epithelioid cells revealed scattered, moderately dense, moderately pleomorphic intracellular granules. Because of loss of structural detail, it was difficult to determine if these were membrane bound. Chromogranin and synaptophysin immunostains were negative, militating against an interpretation of dense core neurosecretory granules. There were no recognizable lipid droplets, premelanosomes, fusiform densities, or micropinocytotic vesicles. These ultrastructural findings, though not specific, when taken in the context of the histology, histochemistry, and immunocytochemical evidence, were, however, supportive of an unusual form of spindle/epithelioid peripheral nerve sheath tumor. These findings may be limited because of suboptimal tissue processing.

Schwannomas are benign mesenchymal tumors that are derived from the cells of Schwann that form the neural sheath. In the gastrointestinal tract they arise from the Schwann cells of the neural plexus of the gastrointestinal wall.7 Though schwannomas are considered benign neoplasms, rare cases of malignant schwannoma have been reported.8 Indicators of malignant potential on histologic diagnosis include nuclear atypia, abnormal mitosis, and invasion.9 The microcystic/reticular variant of schwannoma is extremely rare and tends to localize in the gastrointestinal tract. It needs to be recognized, as both malignant and nonmalignant entities enter the differential diagnosis. Morphologically, it resembles a signet ring cell type of carcinoma or even a myxoid GIST. Even conventional gastrointestinal schwannomas occur sparingly and tend to localize in the stomach.3 The recently described microcystic/reticular variant is more common in the large intestine. Only 2 of the 12 cases reported so far have occurred in the stomach. In contrast to other sites, conventional Schwannomas arising in the gastrointestinal tract are unencapsulated with pushing margins, although infiltration between smooth muscle fibers can be seen.1,10 A discontinuous peripheral cuff of lymphocytes, sometimes with germinal center formation, is usually present within these tumors. Features like nuclear palisading and Verocay bodies are not usually present even in conventional gastrointestinal schwannomas. Liegl et al2 described the morphologic findings of microcystic/reticular pattern schwannoma in their 2008 paper. According to them, these schwannomas are composed of anastomosing and intersecting strands of spindle cells with ill-defined eosinophilic cytoplasm set in a myxoid and collagenous stroma. The nuclei were round, oval, and tapered and showed inconspicuous to small nucleoli. Scattered inflammatory cells were present, whereas nodular aggregates of lymphocytes were rare.2,6 In their group of 10 cases, 5 cases were located in the gastrointestinal tract. Of these, 4 cases were located in the submucosa and were circumscribed but unencapsulated. Our case was also well circumscribed but lacked a definite capsule, compared with classic schwannomas at other sites, which are well encapsulated. One of their cases showed infiltration between smooth muscle fibers and extended into the mucosa. Our case also shows the tumor cells infiltrating between the muscularis mucosae and between the crypts. All their cases involving the gastrointestinal tract were characterized by an absence of discontinuous cuff of lymphocytes with germinal centers, which was present in their cases presenting within the adrenal gland, the bronchus, and the deep soft tissue of the upper back. Our case involving the gastrointestinal tract was also characterized by an absent peripheral lymphocytic infiltrate. Kienemund et al6 recently reviewed 2 more cases, both arising in the sigmoid colon.

Table 2 shows the comparative clinical pathologic findings of the 11 cases of gastrointestinal microcystic/reticular schwannomas reported in the literature so far and our case (case 12).

Unlike most soft tissue tumors, schwannoma is characterized ultrastructurally by the presence of distinct linear, frequently duplicated external lamina.11 In our case no definite external lamina was appreciated; however, there was some nonspecific condensation of extracellular matrix around tumor cells. This could be because the specimen was compromised because of formalin fixation. Hou et al1 reviewed 33 cases of gastrointestinal schwannomas by electron microscopy. All their cases were composed of elongated, spindle-shaped cells with interdigitating or intertwining cellular processes that occasionally formed prominent parallel membranous structures. Many of the cell processes were surrounded by basement membranes. Our case showed similar features, though the morphology was not preserved.

The important differential diagnostic consideration, particularly on small biopsy specimens, is mucinous carcinoma. The bluish pink myxoid matrix and the microcystic structures can mimic small tubular glands, leading to an erroneous diagnosis of mucinous adenocarcinoma. Nuclear atypia and cytokeratin immunoreactivity are absent in microcystic/reticular schwannoma, allowing for this distinction. Another significant differential diagnosis in such cases can be a GIST with prominent myxoid changes. The diagnosis can be confirmed by immunostaining for CD117 or DOG-1 (for CD117-negative cases), which would be negative in schwannomas. S100 positivity can be seen in some cases of GIST, particularly in the small intestine.12 In such cases mutational analysis for KIT and PDGFRA genes13 can also be used to identify GIST as compared to a microcystic/reticular schwannoma. Intestinal perineuriomas can also be entertained as a differential diagnosis, as these tumors are composed of spindle cells with bland oval nuclei and bipolar cytoplasmic processes in a background of collagenous stroma. Perineurioma shows epithelial membrane antigen positivity and S100 and glial fibrillary acidic protein negativity, which helps in the differential diagnosis in such cases. Myxoid synovial sarcoma can show a reticular growth pattern. 14 Areas showing conventional fascicular pattern with greater nuclear atypia and immunohistochemical positivity for epithelial membrane antigen and cytokeratin and negative S100 immunoreactivity can help to differentiate the two. Also, the characteristic t(X;18) translocation can be useful to identify synovial sarcoma.7 

Microcystic/reticular schwannomas can be a problematic differential diagnosis even in cases of nonluminal gastrointestinal tract tumors like pancreas. This was highlighted by Liegl et al15 in their recent case report of a 5-cm mass in the pancreatic head. The fine-needle aspiration was interpreted as malignant tumor cells consistent with a moderately differentiated adenocarcinoma, whereas the histology of the excised mass was consistent with a microcystic/reticular schwannoma.

Outside the gastrointestinal tract, for a microcystic/reticular variant of schwannoma the main differential diagnosis of consideration are reticular perineurioma, myoepithelial tumors, and extraskeletal myxoid chondrosarcoma.

In summary, this is a rare variant of schwannoma demonstrating the peculiar pseudo–signet ring/lipoblastic-like pattern most likely due to degenerative changes of this lesion. It is important to be aware of it in practice, as it can simulate a signet ring cell carcinoma or a GIST with prominent myxoid change, particularly on small biopsy specimens. It is also useful to be aware of this entity in fine-needle aspiration of pancreatic lesions, as it can simulate a pancreatic adenocarcinoma. The follow-up data in literature emphasizes the fact that microcystic/reticular schwannomas are completely benign.