Background
Although nerve involvement can predict recurrence and prognosis in oral squamous cell carcinomas, there still have controversies and limitations regarding the standardization for its detection. In this study, we explore the impact of neural invasion in oral squamous cell carcinomas prognosis, comparing intraneural invasion (tumor cells inside nerve structure) and perineural invasion (cells involving the nerve, but not invading its sheath).

Methods
Surgical slides stained with hematoxylin and eosin from 235 patients with oral squamous cell carcinomas were carefully verified for the presence of intraneural invasion and perineural invasion. The location in the tumor (intratumoral vs. peritumoral) and number of foci (unifocal or multifocal) were also explored. Survival analyses for cancer-specific survival and disease-free survival were performed with Cox proportional model.

Results
Neural invasion was identified in 74 cases, 64.9% displayed intraneural invasion and 35.1% displayed perineural invasion. Univariate analysis revealed a significantly poorer cancer-specific survival, but not disease-free survival, in patients with intraneural invasion, in contrast to cases with perineural invasion that did not achieve significant association with both cancer-specific survival and disease-free survival. Further analyses revealed that the location in the tumor and number of foci had little impact on discriminatory ability of intraneural invasion. Multivariate analysis confirmed that intraneural invasion is significantly and independently associated with poor cancer-specific survival (hazard ratio: 2.50, 95% CI: 1.31–3.79, p = 0.003).

Conclusion
This study provides evidence that intraneural invasion, but not perineural invasion, is a relevant predictor of survival in patients with oral squamous cell carcinomas, suggesting that its association with other clinical and pathological prognostic factors should be consider in determining the optimal treatment protocol and prognosis of these patients.