Recurrence Risk in Lateralized Floor of Mouth Carcinoma: Depth of Invasion Versus Midline Distance

BACKGROUND: Contralateral nodal recurrence in lateralized floor-of-mouth squamous cell carcinoma (OSCC) remains difficult to predict, and objective preoperative parameters to guide contralateral neck management in tumors without midline crossing are lacking. METHODS: This retrospective cohort study included 97 patients with lateral floor-of-mouth OSCC treated by surgical resection and neck dissection. Depth of invasion (DOI) was evaluated using endpoint-specific univariable Cox regression. Contralateral nodal disease was analyzed in patients without midline crossing (n = 68) using a composite endpoint of occult contralateral nodal metastasis and/or subsequent contralateral nodal recurrence. Tumor-to-midline distance was assessed on standardized multiplanar preoperative CT or MRI. RESULTS: DOI was associated with local recurrence (HR 1.322 per mm, 95% CI 1.131-1.545, p < 0.001) and ipsilateral regional recurrence (HR 1.199, 95% CI 1.052-1.367, p = 0.007), but not with contralateral recurrence (HR 1.030, 95% CI 0.834-1.272, p = 0.782). Among 68 patients without midline crossing, no occult contralateral metastases were identified in 19 clinically contralateral node-negative patients undergoing elective contralateral neck dissection. Four patients (5.9%) developed subsequent contralateral nodal recurrence. Midline distance was smaller in patients with an event (median 3.15 vs. 6.40 mm; p = 0.022; AUC 0.845). All four events occurred at < 5 mm (4/21 [19.0%] vs. 0/42 [0%]; Fisher p = 0.010). In time-to-event analysis, greater midline distance was associated with a lower hazard of contralateral recurrence (HR 0.622 per mm, 95% CI 0.395-0.981, p = 0.041). CONCLUSION: DOI showed an outcome-specific association with local and ipsilateral, but not contralateral, recurrence. In tumors without midline crossing, quantitative proximity to the midline was associated with subsequent contralateral nodal failure. The observed 5-mm threshold is exploratory and requires external validation before it can inform clinical decision-making.