Objectives The purpose of this study was to compare magnetic resonance imaging (MRI) maximum tumor diameter and depth of invasion with histopathology in oral tongue squamous cell carcinoma (OTSCC) patients in our Institute. Another objective was to compare recorded nodal status between MRI and histology. Material and methods MRI and pathological records of 45 patients diagnosed with T1-T3 OTSCC were reviewed retrospectively. Maximum tumor diameter and depth of invasion were measured and rechecked by oral radiologist and pathologist. Nodal status was recorded from both MRI and histopathology. Correlation analyses were performed using Pearson's correlation. Results Both maximum tumor diameter and depth of invasion correlated significantly between MRI and histology (rho = 0.874,p <.001;rho = 0.898,p <.001). Significant correlation was found between MRI and pathological dimensions in the MRI-based T-staged subgroups of T2 and T3 but not in T1. MRI sensitivity for detecting pathologically positive nodes was 60%. MRI specificity for detecting pathologically negative nodes was 83%. Moderate correlation was found between MRI and histological nodal status (rho = 0.44,p = .003). Conclusions MRI tumor dimensions correlate with histopathological data in OTSCC. Based on our Finnish patient material and results, MRI serves as an accurate tool in supporting OTSCC patient treatment in our Institute.

Background Tumour budding (B) and depth of invasion (D) have both been reported as promising prognostic markers in oral squamous cell carcinoma (OSCC). This meta-analysis assessed the prognostic value of the tumour budding and depth of invasion combination (BD model) in OSCC. Methods Databases including Ovid MEDLINE, PubMed, Scopus and Web of Science were searched for articles that studied the BD model as a prognosticator in OSCC. PICO search strategy was "In OSCC patients, does BD model have a prognostic power?" We used the reporting recommendations for tumour marker prognostic studies (REMARK) criteria to evaluate the quality of studies eligible for systematic review and meta-analysis. Results Nine studies were relevant as they analysed the BD model for prognostication of OSCC. These studies used either haematoxylin and eosin (HE) or pan-cytokeratin (PCK)-stained resected sections of OSCC. Our meta-analysis showed a significant association of BD model with OSCC disease-free survival (hazard ratio = 2.02; 95% confidence interval = 1.44-2.85). Conclusions The BD model is a simple and reliable prognostic indicator for OSCC. Evaluation of the BD model from HE- or PCK-stained sections could facilitate individualized treatment planning for OSCC patients.