Colorectal polyps containing S-100-positive neural proliferations in the lamina propria that lack ganglion cells have been variously referred to as "neuromas" or "neurofibromas." However, these lesions have not been systematically examined, and whether they are associated with type 1 neurofibromatosis (NF1) or other inherited syndromes is unknown. The aim of this study was to evaluate the clinicopathologic and immunohistochemical features of these lesions, in comparison to colorectal neurofibromas from known NF1 patients. Morphologically similar lesions from 26 patients (mean age, 62 y; range, 46 to 88 y; male/female ratio, 10/16) were retrieved from surgical pathology and consult files. Clinical and endoscopic data were obtained, and immunohistochemistry for S-100 protein, glial fibrillary acidic protein, neurofilament protein (NFP), epithelial membrane antigen, claudin-1, CD34, smooth muscle actin, and KIT was performed. The findings were compared with those in mucosal biopsies of 5 submucosal neurofibromas from NF1 patients. All 26 polyps were sessile, ranging from 1 to 6 mm in size (mean, 2.5 mm). Most arose in the distal colon (15 rectosigmoid, 7 descending, 2 transverse, and 2 ascending), and were incidentally found at screening colonoscopy. After a mean follow-up of 6.5 years (range, 3 mo to 17.5 y), none of the patients developed other neural polyps, and none had evidence of NF1 or other inherited syndromes. Histologically, the lamina propria of the polyps contained a diffuse cellular proliferation of uniform bland spindle cells with elongated, tapering nuclei, abundant, dense eosinophilic cytoplasm, and indistinct cell borders, entrapping adjacent crypts. No nuclear atypia, pleomorphism, mitotic activity, or associated ganglion cells were observed. All showed strong staining for S-100 protein in essentially 100% of cells. NFP highlighted rare axons in 7 lesions. All other markers were negative. The 5 neurofibromas showed similar histologic features, but were generally less uniformly cellular, showed some intralesional heterogeneity, and showed less extensive staining for S-100 protein; all contained scattered NFP-positive axons. In summary, solitary colorectal polyps containing pure Schwann cell proliferations in the lamina propria are not associated with NF1. Distinguishing these lesions from NF1-associated neurofibromas is difficult based on histologic features; the presence of an underlying submucosal nodule or mass should be excluded endoscopically, and immunohistochemistry should be performed. Although their nature is uncertain, we propose the interim designation "mucosal Schwann cell 'hamartoma'" to avoid confusion with the neural lesions that have significant associations with inherited syndromes.
