Treatment strategy for cervical lymph node metastases from early-stage tongue and floor of the mouth squamous cell carcinoma using tumour budding and depth of invasion as predictors

Abstract

Objectives: This study aimed to determine whether elective neck dissection can help improve outcomes in early-stage tongue and floor squamous cell carcinoma (SCC) by statistically analysing the relationship between information obtained from biopsy specimens and the incidence and prognosis of cervical lymph node metastasis (CLM).

Materials and methods: Biopsy specimens of 103 patients diagnosed with early cT1-T2 cancer of the tongue and floor of the mouth were included.

Results: Multivariate analysis showed that the three parameters significantly correlated with CLM, and univariate analyses showed that budding score (BS) ≥ 5 and pathological depth of invasion (pDOI) ≥ 5 mm were independent risk factors for CLM. There were significant differences in the 5-year cumulative disease-specific survival between the BS < 5 and BS ≥ 5 groups, the pDOI < 5 mm and pDOI ≥ 5 mm groups, and the positive and negative budding and depth of invasion (BD) score groups.

Conclusion: In early-stage tongue and floor of the mouth cancers with maximum tumour diameter ≤ 20 mm, it may be necessary to treat occult CLM during initial surgery based on the following preoperative criteria: pDOI ≥ 5 mm or BS ≥ 5 in biopsy specimens and DOI ≥ 8 mm on imaging. The BD model exhibited the highest specificity and proved helpful for CLM prediction.

Clinical relevance: pDOI ≥ 5 mm and BS ≥ 5 were independent predictors of CLM and prognosis in early-stage tongue and floor of the mouth cancers with a maximum tumour diameter of 20 mm.

Keywords: Biopsy; Cervical lymph node metastasis; Depth of invasion; Early tongue and floor of the mouth squamous cell carcinoma; Tumour budding.

Fig. 1

Mode of invasion and tumour budding. TB seen in oral squamous cell carcinoma using haematoxylin–eosin staining (A, B) and immunostaining for cytokeratin (C, D). Both cases were classified as YK3 using haematoxylin–eosin staining, although BS was 2 and 8, respectively. TB: tumour budding, YK: Yamamoto–Kohama classification, BS: budding score

Fig. 2

Pathological depth of invasion. The pDOI (mm) is measured as the perpendicular distance from the basement membrane region to the deepest point of the infiltrative front of the tumour in the biopsy specimens. pDOI: pathological tumour depth of invasion

Fig. 3

Five-year cumulative survival rate. The 5y-OS of all 103 patients was 89.1%, and the 5y-DSS was 96.1% (A, B). The 5y-DSS of 84 CLM-negative patients was 100%, and that of 19 CLM-positive patients was 70.6%, showing a significant difference between the two groups (P < 0.001) (C). There were significant differences in the 5y-DSS between the BS < 5 (100%) and BS ≥ 5 (77.8%) groups (P < 0.001) (D), between the pDOI < 5 mm (100%) and pDOI ≥ 5 mm (76.8%) groups (P < 0.001) (E), and between the BD score negative groups (100%) and BD model score positive group (62.9%) (P < 0.001) (F). In the END group (n = 4) and the neck dissection group after late CLM (n = 15), the 5y-DSS of the END group was 100% better than that of the late metastasis group (58.9%), but the difference was not significant (P = 0.233) (G). 5y-OS: 5-year cumulative overall survival, 5y-DSS: 5-year cumulative disease-specific survival, CLM: cervical lymph node metastasis, pDOI: pathological depth of invasion, BS: budding score, BD model: budding and depth of invasion model, END: elective neck dissection