Epithelial Markers in Sinonasal Inverted Papilloma with Malignant Transformation: A Scoping Review and Proposal for an Integrative Model of Malignant Risk Stratification

Background/Objectives: Sinonasal inverted papilloma is a rare Schneiderian epithelial tumor, histologically benign in most cases, but characterized by locally aggressive behavior, a high recurrence rate and a recognized potential for malignant transformation, most frequently into sinonasal squamous cell carcinoma. The present article is a scoping review that maps and critically synthesizes the available evidence on epithelial, immunohistochemical and selected molecular markers in malignant-transformed sinonasal inverted papilloma, rather than a quantitative systematic review or meta-analysis. Methods: The literature search was performed according to PRISMA-ScR principles in PubMed/MEDLINE, Embase, Scopus/Web of Science and Google Scholar, using the search keywords sinonasal inverted papilloma, malignant transformation, markers, HPV, p63, p40, p16 and Ki-67, as well as complementary markers of the cell cycle and of the EGFR/CDKN2A/TP53 pathways. Results: Based on the available data, p40 and p63 have diagnostic value for defining and mapping the squamous phenotype, including invasive carcinomatous nests, but they are not standalone predictive markers of malignant transformation. p16 cannot be automatically used as a surrogate for active HPV infection in inverted papilloma, while the results regarding HPV remain inconclusive in relation to malignant transformation of sinonasal inverted papilloma. Ki-67, especially when interpreted topographically and in relation to dysplasia and invasion, may suggest proliferative acceleration, but prognostic thresholds have not yet been validated. Conclusions: The conclusion of our study is that the risk of malignant transformation cannot be reduced to a single marker; it must be interpreted through an integrative clinical, imaging, histopathological, immunohistochemical and molecular model, which requires prospective multicenter validation.