Related Experiment Video
Updated: Aug 6, 2026
Intraoperative Assessment of Resection Margins in Oral Cavity Cancer: This is the Way
Published on: May 10, 2021
Real-World Correlation Between Radiologic and Pathological Depth of Invasion in Oral Squamous Cell Carcinoma: A
Onur Yildirim1, Danielli Matsuura2, Cristina Valero2
1Department of Radiology, Memorial Sloan Kettering Cancer Center, New York, New York, USA.
Purpose:
The prognostic significance of pathological depth of invasion (p-DOI) is widely acknowledged, leading to its inclusion in the clinical and pathological staging systems for Oral Squamous Cell Carcinoma (OSCC) in the 8th Edition of the AJCC Cancer Staging Manual. However, radiologic assessment of DOI (r-DOI) remains challenging, and limited evidence exists regarding its correlation with p-DOI, the current gold standard. This study aims to evaluate the correlation between r-DOI and p-DOI in a large, real-world cohort.
Method:
Following Institutional Review Board approval, we evaluated 261 patients with biopsy-proven OSCC who underwent primary surgery between 2010 and 2015. Radiological images were reviewed by a single subspecialty-trained neuroradiologist, blinded to clinicopathological information. Preoperative MRI, CT, and FDG-PET scans were analyzed; in cases of multiple imaging studies, the imaging modality that best depicted the tumor's invasive front in each individual case was selected for r-DOI measurement. r-DOI was calculated by measuring tumor depth from a horizontal reference line connecting the tumor to the adjacent normal mucosa. The correlation between r-DOI and p-DOI was assessed using Pearson's product-moment correlation and visualized via scatterplot analysis. Outliers were defined as cases with residuals ≥ 2 standard deviations from the regression line.
Results:
The median age at diagnosis was 61.8 years (59% male). The most common subsite was the oral tongue (55.9%), followed by the floor of mouth (11.5%), lower gum (14.9%), and other sites (combined 17.6%). Imaging modalities used for r-DOI determination included CT (n = 187, 71.6%), PET/CT (n = 50, 19.2%), and MRI (n = 24, 9.2%); 105 patients (40.2%) underwent both PET/CT and CT, 11 (4.2%) had MRI and CT, 12 (4.6%) had PET/CT and MRI, and 5 (1.9%) received all three modalities. The Pearson correlation coefficient between r-DOI and p-DOI, calculated as continuous variables, demonstrated a strong positive correlation (Pearson's r = 0.834, p < 0.001).
Conclusion:
Our study demonstrates that routinely used standard-of-care imaging studies may provide a clinically useful adjunct for preoperative estimation of depth of invasion, with reduced reliability in intermediate-depth (5-10 mm) lesions, where cautious interpretation is warranted.
