Tumor Budding as a Histologic Prognostic Parameter in Oral Squamous Cell Carcinoma: Comparison of Scoring Systems and Rationale for the Use of Immunohistochemistry

Context.—: Tumor budding (TB) is associated with nodal metastasis and worse outcomes in oral squamous cell carcinoma. However, there is a need for standardization of its assessment. Objective.—: To perform TB scoring using 2-tiered and 3-tiered systems, with and without pancytokeratin (CK) immunohistochemistry, and assess relationship with outcome parameters. Design.—: Tumor buds were scored by using 2-tiered (low: <5; high: ≥5) and 3-tiered (low: 0-4; intermediate: 5-9; high: ≥10) systems on hematoxylin-eosin- and CK-stained sections. Results.—: TB was evaluated in 400 oral squamous cell carcinomas, which were predominantly gingivobuccal (242; 61%), had aggressive worst patterns of invasion (WPOI) (WPOI-4: 59%; WPOI-5: 27%), and late stage (298; 75%). Two-year disease-free survival (DFS) was 56.3%, and 2-year overall survival (OS) was 58.7%. Using 3-tiered scoring, 220 cases (55%) had low, 101 (25%) had intermediate, and 79 (20%) had high TB. With 2-tiered scoring, 220 cases (55%) had low and 180 (45%) had high TB. Greater concordance was seen between hematoxylin-eosin- and CK-stained sections with 2-tiered scoring: 14% were up-scored versus 25% with 3-tiered scoring. Cases with high and intermediate TB had significantly shorter DFS and OS. TB on CK-stained sections scored by both systems was an independent predictor of nodal metastasis. With all 4 scoring models, high TB was an independent predictor of recurrence. All 4 models showed similar predictive values for OS and DFS. Conclusions.—: TB is a significant predictor of nodal metastasis and recurrence, irrespective of site and stage. Three-tiered scoring does not show improved prediction of outcomes over 2-tiered scoring.