Accuracy of magnetic resonance imaging in evaluating the depth of invasion of tongue cancer. A prospective cohort study

[Display omitted] •MRI has an average 2.3 mm false-positive measurement of tongue cancer.•MRI has an average 1.92-mm false-positive measurement in two-dimensional judgement.•Depth of infiltration on MRI that could identify OS and DSS was both 11 mm.•Depth of infiltration on MRI that could identify nodal metastasis was 8 mm. This study compared the consistency of depth of invasion (DOI) measurements by magnetic resonance imaging (MRI) and intraoperative and postoperative pathological sections due to a lack of large sample studies. From April 2015 to December 2017, patients with squamous cell carcinoma of the tongue were included in the study. Different invasion depths were measured by MRI and on intraoperative and postoperative pathological sections. The differences between two-dimensional tumor margins were analyzed using Mimics 15.0 and Geomagic Control 16.0. Statistical analyses were performed using IBM SPSS software version 25.0 (IBM Corp., Armonk, NY). This study included 150 patients, the overall difference between MRI and postoperative pathological sections (DMP) and the overall difference between intraoperative and postoperative pathological sections (DIP) based on pathological specimens were 2.32 ± 1.68 mm and 0.68 ± 0.99 mm. The overall difference between MRI and intraoperative pathological sections (DMI) based on intraoperative specimens was 1.64 ± 1.32 mm. The tumor growth pattern and T stage were significantly correlated with measurement differences. The cutoff value of MRI depth that could identify nodal metastasis was 8 mm, and were both 11 mm for OS and DSS. Clinicians performing T staging on patients with tongue cancer based on MRI measurements must consider the false-positive mean depth of 2.3 mm as well as the growth pattern and specific infiltration depth. The prognostic MRI depths that enabled the identification of nodal metastasis, OS and DSS were 8 mm, 11 mm and 11 mm, respectively. Name: A Prospective, Observational, Real-world Study Based on the Register System of Oral and Maxillofacial Malignant Tumors. (ClinicalTrials.gov ID: NCT02395367)