What is the Role of Depth of Invasion With Site Specificity in Managing the Clinically Negative Neck for Patients With Stage I Oral Cavity Squamous Cell Carcinoma? (cT1N0).

Background: Guidelines for elective neck dissection (END) in oral cavity squamous cell carcinoma (OSCC) remain controversial with observation, sentinel lymph node biopsy or END in all subjects inconsistently proposed. The current study evaluated the performance of the (D)OI (depth of invasion) and (S)site to determine (E)lective (N)eck (D)issection (DSEND) algorithm in recommending END for cT1N0 OSCC subjects.

Study design, setting, sample: This is a retrospective case series of subjects treated between 2017 and 2023. All cT1N0 subjects treated with the DSEND algorithm were included. Exclusion criteria included subjects with recurrent disease, prior radiation therapy of the head and neck, or follow-up less than 1 year.

Predictor: The predictor variable was management of the neck-observation versus END per DSEND algorithm. Subjects who reached the site-specific threshold were treated with END. A secondary analysis compared preoperative depth and intraoperative depth to evaluate if one was more accurate compared to final pathology depth.

Outcome variable: The outcome variable was isolated neck recurrence (yes/no) during the follow-up period. The secondary analysis outcome variable was depth of invasion measured preoperative, intraoperative and on final pathology.

Covariates: The covariates were age, sex (male/female), smoking and alcohol status (yes/no), tumor size (cm), nodal status (positive/negative), anatomical subsite and final pathology depth of invasion (mm).

Analyses: We report descriptive statistics, accuracy of DSEND, comparison of preoperative versus intraoperative frozen section to final pathology depth of invasion using Student's t-test, setting P < .05 as statistically significant.

Results: The sample was composed of 88 subjects with a mean age of 63.7 (standard deviation 13.2). Females represented 51% (45 of 88) of the subjects. DSEND protocol was used for 80 subjects. Observation of the neck was recommended in 57.5% (46 of 80) of subjects with only 1 developing isolated neck recurrence (false negative 2% (1 of 46)). In patients managed with END 6% (2 of 33) experienced isolated neck recurrence. The difference between final pathology and intraoperative depth was significantly smaller than final pathology compared to preoperative depth.

Conclusions and relevance: The DSEND algorithm applied to cT1N0 OSCC allowed for nonsurgical management of the neck in 57.5% (46 of 80) of subjects with a 2% (1 of 46) false negative. Frozen section analysis was more accurate than preoperative depth of invasion from biopsy.