Immunohistochemical staining for smoothelin in the duplicated versus the true muscularis mucosae of Barrett esophagus.

Background: The muscularis mucosa underlying the metaplastic mucosa of Barrett esophagus is frequently duplicated, with an intervening layer of lamina propria between the superficial or neomuscularis mucosa (NMM) and the deep/true muscularis mucosa (TMM). This duplication causes difficulties with accurate staging of superficially invasive carcinoma in biopsy specimens and endoscopic mucosal resections (EMRs), as invasion underneath the superficial muscle layers may be mistaken for submucosal invasion. Mucosal resections or other ablative nonsurgical therapies can be curative in patients with esophageal intramucosal carcinoma, whereas patients with submucosal invasion are recommended for esophagectomy. Therefore, the accurate staging of such specimens is crucial. Smoothelin is a novel smooth muscle protein expressed only by fully differentiated smooth muscle cells and not by proliferative or noncontractile smooth muscle cells and fibroblasts. It has been suggested that in the bladder, immunohistochemistry for smoothelin may help separate hyperplastic muscularis mucosa from the true muscularis propria. We hypothesized that in the esophagus, immunohistochemistry for smoothelin would differentiate the NMM from the TMM. Design: Thirty cases of EMRs for Barrett esophagus-related neoplasia were retrieved from the archives of the pathology department, St Michael's Hospital. Immunohistochemical staining for smoothelin was performed to evaluate differential staining in the TMM versus NMM. Fifteen cases were stained for smooth muscle actin and smooth muscle myosin. The staining score was evaluated on a scale from 0 to 3 according to the percentage and intensity of staining. Results: Immunohistochemical staining results for smoothelin were as follows: the NMM showed weak focal staining (+ 1) in 23 of 30 cases (82%), and moderate staining (+ 2) in 7 of 30 cases (12%), and the TMM showed very strong and diffuse staining (+ 3) in 30 of 30 cases. No cases showed negative (0) staining in the NMM. With smooth muscle actin and myosin, strong and diffuse staining was observed with similar intensity in both the TMM and NMM in 15 of 15 cases. Conclusions: In our study, smoothelin staining in the NMM is significantly weaker than that seen in the true/deep muscularis mucosa. This pattern is similar to that reported for the muscularis mucosae of the urinary bladder. Although smoothelin can readily distinguish the 2 layers, its value might be limited by the need to simultaneously compare the 2 layers. Although this might be of use in EMR specimens in which both layers are visible, use in biopsies may be limited as the residual staining in the NMM may inhibit definitive evaluation. This issue may be resolved by the use of appropriate standard controls, individual optimization of the antibody, and the use of an automated digital assessment.