Pattern-Anchored Diagnosis of Oral Epithelial Dysplasia: A Position Paper From COST Action INTERCEPTOR Working Group 2
HEAD & NECK PATHOLOGY, cilt.20, sa.1, 2026 (ESCI, Scopus)
- Yayın Türü: Makale / Derleme
- Cilt numarası: 20 Sayı: 1
- Basım Tarihi: 2026
- Doi Numarası: 10.1007/s12105-026-01963-7
- Dergi Adı: HEAD & NECK PATHOLOGY
- Derginin Tarandığı İndeksler: Emerging Sources Citation Index (ESCI), Scopus, EMBASE, MEDLINE, Biomedical Reference Collection: Corporate Edition (EBSCO), Health Research Premium Collection (ProQuest)
- Açık Arşiv Koleksiyonu: AVESİS Açık Erişim Koleksiyonu
- Ankara Üniversitesi Adresli: Evet
Özet
Purpose Oral epithelial dysplasia (OED) is the principal histopathological marker for malignant risk stratification in oral potentially malignant disorders (OPMDs). Current World Health Organisation (WHO) three-tier and other binary grading systems are limited by modest inter-observer reproducibility and suboptimal prognostic discrimination, particularly for architecture-predominant lesions where cytological atypia is inconspicuous. The most clinically consequential failure of current practice is not mis-grading but non-recognition of dysplasia altogether, with direct implications for patient surveillance and management.Methods This position paper, arising from Working Group 2 of COST Action CA21140 INTERCEPTOR, synthesises evidence from inter-observer reproducibility studies, feature-based prognostic models, under-recognition studies, outcome data and computational pathology analyses to propose and contextualise a pattern-anchored, grade-retaining diagnostic framework for OED.Results Five recognisable histological patterns are proposed: conventional/basaloid, keratinising/differentiated, verrucous, papillomatous, and HPV-associated, functioning as diagnostic anchors that capture the full morphological spectrum of OED, including architecture-predominant phenotypes that are disproportionately under-recognised in routine practice. Feature-based prognostic models demonstrate that architectural traits characterising these patterns (bulbous rete ridges, AUROC 0.74; cohesion loss, AUROC 0.73) predict malignant transformation more effectively than grade alone. Independent computational pathology analyses identify corresponding immune microenvironment and architectural signals as the strongest prognostic variables, providing convergent support.Conclusion A two-step approach i.e. first identifying the dominant morphological pattern and then assigning grade within that pattern using existing WHO or binary criteria, offers a pragmatic, grade-retaining refinement compatible with current diagnostic frameworks and clinical pathways. The proposal is hypothesis-driven and requires prospective multicentre validation before wider adoption.