Oral epithelial dysplasia (OED) is characterized by abnormal structural and cellular changes in the oral mucosa epithelium. These changes are often detected histologically in potentially malignant disorders and are associated with an increased risk of malignant transformation into oral squamous cell carcinoma (OSCC). Currently, clinicians rely mainly on the pathologist's histological gradings (binary or tertiary grading systems) to determine the follow-up protocol for patients with OED. However, both histological gradings are rather subjective and not highly reliable. Therefore, there is a need for biomarkers to distinguish high-risk OED cases from low-risk cases. Twelve oral tongue dysplasia samples were collected for multiplex immunostaining. The samples were derived from 7 patients who did not develop OSCC within 5 y of their OED diagnosis (low-risk group, LR) and 5 patients who developed OSCC (high-risk group, HR). Ten antibodies targeting epithelial cells, immune cells, a proliferation marker, and immune checkpoints were used (PanCK, CD4, CD8, CD11b, CD68, CD56, FoxP3, Ki67, PD-1, PD-L1). Uniform manifold approximation and projection generated from multiplex staining revealed distinct differences in the distribution of immune cells and immune checkpoint expression between the LR and HR groups. The HR group showed a higher number of CD8+ cytotoxic T cells and CD68+ macrophages than the LR group. Notably, both PanCK+ epithelial cells and CD8+ cytotoxic T cells exhibited increased proliferation in the HR group, as indicated by Ki67 expression. Additionally, PD-1 expression was significantly elevated in the HR group. The logistics regression model further confirmed the predictive value of individual markers and identified a multimarker model with superior predictive performance. Our multiplex immunostaining revealed significant differences in the immune landscape between LR and HR cases of OED dysplasia. CD8+ cytotoxic T cells, CD68+ macrophages, and PD-1+ cells may serve as potential biomarkers for stratifying HR from LR oral tongue dysplasia cases.
Sandell et al. (2026) studied this question.