SOX10 Expression in Primary Salivary Gland Tumors: A Meta-Analysis

Abstract

Abstract Purpose Primary salivary gland tumors form a heterogeneous group with diagnostic challenges. Sex-determining region Y-related homeobox 10 (SOX10) is a transcription factor studied in primary salivary gland tumors for over a decade. To date, no systematic review or meta-analysis has been conducted on its expression in salivary gland tumors. Methods A PubMed literature search identified all studies in which SOX10 was evaluated by immunohistochemistry in primary salivary gland tumors. Information on SOX10 positivity, staining pattern, percentage of positive cells, and information on the antibody source and clone were tabulated in Microsoft Excel. The meta-analysis was performed with OpenMeta Analyst. The meta-analysis adheres to PRISMA (Preferred Reporting Items for Systematic reviews and Meta-Analyses) guidelines. Results The meta-analysis included 35 studies with 1146 histological and 35 cytological samples immunostained with SOX10. The list of SOX10-positive tumors according to our meta-analysis includes pleomorphic adenoma (pooled meta-analytic estimate of SOX10 positivity 95.8%, 95% CI 91.7–99.9%), polymorphous adenocarcinoma (pooled estimate 97.4%, 95% CI 92.3–102.6%), canalicular adenoma (pooled estimate 98.9%, 95% CI 96.4–101.4%), basal cell adenoma (pooled estimate 91.3%, 95% CI 80.2–102.5%), adenoid cystic carcinoma (pooled estimate 96.6%, 95% CI 94.1–99.2%), myoepithelioma (pooled estimate 75.2%, 95% CI 52.0–98.4%), acinic cell carcinoma (pooled estimate 97.4%, 95% CI 94.1–100.6%), intercalated duct type of intraductal carcinoma (pooled estimate 90.8%, 95% CI 75.9–105.8%), and microsecretory carcinoma (pooled estimate 96.1%, 95% CI 87.2–105.0%). Limited evidence suggests that basal cell adenocarcinoma, epithelial–myoepithelial carcinoma, and myoepithelial carcinoma are also SOX10 positive. SOX10-negative entities include Warthin’s tumor (pooled estimate 2.3%, 95% CI 0–7.0%), apocrine type of intraductal carcinoma (pooled estimate 10.3%, 95% CI − 10.4–31.1%), salivary duct carcinoma (pooled estimate 3.7%, 95% CI − 0.6–8.1%), and oncocytoma (pooled estimate 8.0%, 95% CI − 2.7–18.7%). Polyclonal antibodies and clones EP268, N-20, and A-2 provided similar results. Clone BC34 stained a lower percentage of SOX10-positive tumors. Conclusion This meta-analysis lists SOX10-positive and -negative tumors. This demonstrates the necessity of subtype-specific reporting and stating the antibody clone used in immunohistochemical studies.