Validation of clinical tumor depth of invasion measure in patients with oral cavity cancer

Abstract

Objectives

The 8th edition of the American Joint Committee on Cancer (AJCC) for head and neck malignancies incorporated depth of invasion (DOI) for oral cavity squamous cell carcinoma (OCSCC). Currently, there is no standardized method to determine clinical DOI (cDOI). We aim to validate radiologic DOI (rDOI) utilizing contrast-enhanced computed tomography (CECT) imaging.

Materials and methods

Radiographic DOI was defined as the coronal measurement of the deepest margin parallel to the expected normal mucosal plane. To establish a valid, accurate and repeatable rDOI, a retrospective cohort at a single institution tertiary hospital underwent preoperative CECT. Radiographic DOI was independently assessed by 2 neuroradiologists. Correlation was determined between rDOI and pDOI. A prospective pilot study of 49 patients with newly diagnosed OCSCC underwent rDOI assessment compared to final pDOI to assess our rDOI method clinical applicability.

Results

Neuroradiologists demonstrated a high degree of predictability between rDOI and pDOI with as great as 90.91 % accuracy for readable CECTs. Inter-observer correlations were strong in the retrospective (Session 1: ICC = 0.956; Session 2: ICC = 0.932) and prospective cohorts (Session 1: ICC = 0.859; Session 2: ICC = 0.913) illustrating reliability. Intraobserver correlations were also strong (Retrospective Reader 1 ICC = 0.968, Reader 2 ICC = 0.941; prospective Reader 1 ICC = 0.965, Reader 2 ICC = 0.800, p < 0.001). T1 lesions were immeasurable 25–62.5 % of the time.

Conclusions

Using CECT, the coronal measurement of the deepest margin parallel to the expected normal mucosal plane can enhance determination of a reliable cDOI in OCSCC corresponding to pDOI without change in staging for T2 and greater lesions.